Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Polyphosphate : a novel negative regulator of complement and its therapeutic potential in age-related… Ocariza, Linnette Mae 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2015_november_ocariza_linnette.pdf [ 12.29MB ]
JSON: 24-1.0167824.json
JSON-LD: 24-1.0167824-ld.json
RDF/XML (Pretty): 24-1.0167824-rdf.xml
RDF/JSON: 24-1.0167824-rdf.json
Turtle: 24-1.0167824-turtle.txt
N-Triples: 24-1.0167824-rdf-ntriples.txt
Original Record: 24-1.0167824-source.json
Full Text

Full Text

POLYPHOSPHATE:	  	  A	  NOVEL	  NEGATIVE	  REGULATOR	  OF	  COMPLEMENT	  	  AND	  ITS	  THERAPEUTIC	  POTENTIAL	  	  IN	  AGE-­‐RELATED	  MACULAR	  DEGENERATION	  by	  	  LINNETTE	  MAE	  OCARIZA	  	  BMLSc,	  The	  University	  of	  British	  Columbia,	  2012	  	  A	  THESIS	  SUBMITTED	  IN	  PARTIAL	  FULFILLMENT	  OF	  THE	  REQUIREMENTS	  FOR	  THE	  DEGREE	  OF	  	  MASTER	  OF	  SCIENCE	  	  in	  	  THE	  FACULTY	  OF	  GRADUATE	  AND	  POSTDOCTORAL	  STUDIES	  	  (Pathology	  and	  Laboratory	  Medicine)	  	  THE	  UNIVERSITY	  OF	  BRITISH	  COLUMBIA	  	  (Vancouver)	  October,	  2015	  	  ©	  Linnette	  Mae	  Ocariza,	  2015	  	  	   ii	  Abstract	  The	   innate	   mammalian	   response	   to	   injury	   involves	   the	   initiation	   of	   activation	   of	   two	   major	  blood-­‐borne	   proteolytic	   systems;	   coagulation	   and	   complement.	   Recent	   studies	   have	   revealed	  that	   there	   is	   considerable	   crosstalk	  and	   interplay	  between	   these	   two	   systems.	  Polyphosphate	  (polyP)	   is	   a	   naturally	   occurring	   inorganic	   linear	   polymer	   that	   co-­‐regulates	   these	   two	   systems,	  acting	  as	  a	  promoter	  of	  coagulation	  and	  an	  inhibitor	  of	  complement.	  This	  thesis	  aims	  to	  further	  characterize	  the	  mechanisms	  by	  which	  polyP	  regulates	  the	  complement	  system,	  and	  to	  test	  its	  physiological	  relevance	  in	  a	  model	  of	  human	  disease,	  age-­‐related	  macular	  degeneration	  (AMD),	  the	  pathogenesis	  of	  which	  involves	  excess	  complement	  activation	  and	  oxidative	  stress.	  Based	  on	  data	   from	   our	   lab	   and	   studies	   in	   bacteria	   that	   polyP	   dampens	   complement	   activation	   and	  interferes	  with	  oxidative	  stress,	   I	  hypothesized	  that	  polyP	  would	  protect	  against	  AMD.	  To	  test	  this	   hypothesis,	   I	   used	   hemolytic	   assays	   to	  measure	   the	   complement	   activity	   in	   response	   to	  polyP,	  in	  vitro	  studies	  with	  AMD-­‐associated	  cell	  lines	  to	  examine	  protective	  properties	  of	  polyP,	  and	  an	  in	  vivo	  model	  of	  AMD	  to	  evaluate	  the	  therapeutic	  efficacy	  of	  polyP.	  I	  showed	  that	  polyP	  dampens	  complement	  activation	  by	  interfering	  with	  the	  terminal	  pathway	  of	  complement,	  and	  that	  it	  also	  interferes	  with	  oxidative	  stress-­‐induced	  cellular	  damage.	  The	  mechanisms	  by	  which	  it	  exerts	  this	  effect	  have	  not	  yet	  been	  determined.	  However,	  in	  vivo,	  in	  rodent	  models	  of	  AMD,	  polyP	  protects	  against	  laser-­‐induced	  choroidal	  neovascularization	  (CNV),	  a	  feature	  of	  AMD,	  with	  reduced	   deposition	   of	   complement.	   An	   agent	   such	   as	   polyP,	   that	   simultaneously	   suppresses	  complement	  activation	  and	  protects	  against	  oxidative	  stress,	  holds	  potential	  therapeutic	  value.	  The	  findings	  in	  this	  thesis	  raise	  awareness	  of	  the	  potential	  importance	  of	  a	  ubiquitous,	  naturally	  occurring	   inorganic	  compound	  that	  has	   largely	  been	  overlooked.	  Most	   important,	   the	  findings	  reveal	  a	  promising	  use	   for	  polyP	  as	  a	   treatment	   for	  AMD,	  a	  common	  and	  devastating	  disease	  that	  affects	  millions	  of	  people	  worldwide.	  	  	  	  	   iii	  Preface	  Chapter	  1.3.3	  was	  published	  as	  Wat,	  J.,	  Foley,	  J.H.,	  Krisinger,	  M.J.,	  Lei,	  V.,	  Ocariza,	  L.M.,	  Wasney,	  G.A.,	   Lameignere,	   E.,	   Strynadka,	   N.C.,	   Smith,	   S.A.,	   Morrissey,	   J.H.,	   and	   Conway,	   E.M.	   (2014).	  Polyphosphate	   suppresses	   complement	   activity	   via	   the	   terminal	   pathway.	   BLOOD,	   Jan	  30;123(5):768-­‐76.	   I	  performed	  the	  gel	   filtration	  studies	  and	  the	  membrane	  binding	  assays	  and	  wrote	   the	  methods	   of	   the	   latter,	   included	   in	   the	   Supplementary	  Materials	   section.	   This	   was	  completed	  at	  The	  University	  of	  British	  Columbia.	  I	  conceptualized	  the	  study	  designs	  (except	  for	  the	  gel-­‐filtration	  study,	  which	  was	  conceptualized	  by	  a	  previous	  graduate	  student),	  and	  all	  experiments	  were	  also	  performed	  and	  analyzed	  by	  me	  (except	   for	   the	   rat	   in	   vivo	   study,	   which	   was	   performed	   by	   a	   Research	   Associate).	   The	   cell	  proliferation	  studies	  were	  performed	  by	  a	  summer	  student	  whom	  I	  was	  supervising.	  	  The	  UBC	  Animal	  Care	  Committee	  examined	  and	  approved	  the	  use	  of	  animals	  for	  experiments	  as	  described	   in	   this	   thesis	   under	   Application	   Number:	   A14-­‐0135.	   Studies	   with	   humans	   were	  approved	  by	  the	  UBC	  Clinical	  Research	  Ethics	  Board	  (CREB)	  under	  protocol	  number:	  H12-­‐02508.	  	  	  	  	  	  	  	  	  	   iv	  Table	  of	  Contents	  Abstract	  .......................................................................................................................	  ii	  Preface	  ........................................................................................................................	  iii	  Table	  of	  Contents	  ........................................................................................................	  iv	  List	  of	  Figures	  .............................................................................................................	  vii	  List	  of	  Abbreviations	  .................................................................................................	  viii	  Acknowledgements	  .....................................................................................................	  xi	  Dedication	  ..................................................................................................................	  xii	  Chapter	  1:	   Introduction	  .............................................................................................	  1	  1.1	   The	  Coagulation	  Cascade	  ............................................................................................................	  2	  1.2	   Polyphosphate	  (PolyP)	  ................................................................................................................	  4	  1.2.1	   Polyphosphate	  and	  Coagulation	  .........................................................................................	  5	  1.3	   The	  Complement	  System	  ............................................................................................................	  8	  1.3.1	   Regulation	  of	  Complement	  Activation	  ..............................................................................	  11	  1.3.2	   Role	  of	  Polyphosphate	  in	  Complement	  .............................................................................	  13	  1.4	   Age-­‐Related	  Macular	  Degeneration	  (AMD)	  ..............................................................................	  14	  1.4.1	   Structure	  of	  the	  Retina	  and	  Choroid	  .................................................................................	  15	  1.4.2	   	  Pathogenesis	  of	  AMD	  .......................................................................................................	  18	  1.4.3	   Role	  of	  Complement	  in	  AMD	  .............................................................................................	  19	  1.4.4	   Role	  of	  the	  MAC	  in	  AMD	  ...................................................................................................	  21	  1.4.5	   Role	  of	  Oxidative	  Stress	  in	  AMD	  ........................................................................................	  22	  1.4.6	   Complement	  and	  Oxidative	  Stress	  ....................................................................................	  22	  1.4.7	   Current	  Therapies	  for	  AMD	  ...............................................................................................	  23	  Chapter	  2:	   Hypothesis	  ..............................................................................................	  25	  	  	   v	  Chapter	  3:	   Overall	  Goal	  and	  Objectives	  ....................................................................	  26	  3.1	   Overall	  Goal	  ..............................................................................................................................	  26	  3.2	   Specific	  Objectives	  ....................................................................................................................	  26	  Chapter	  4:	   Materials	  and	  Methods	  ..........................................................................	  27	  4.1	   Reagents	  ...................................................................................................................................	  27	  4.2	   Cell	  Culture	  ...............................................................................................................................	  27	  4.3	   Terminal	  Pathway	  Hemolytic	  Assay	  ..........................................................................................	  28	  4.4	   Platelet	  Releasate	  Hemolytic	  Assay	  ..........................................................................................	  28	  4.5	   Stability	  of	  PolyP≥1000	  in	  Serum	  ..............................................................................................	  29	  4.6	   Gel	  Filtration	  .............................................................................................................................	  29	  4.7	   Measuring	  Effect	  of	  PolyP≥1000	  on	  C5b-­‐7	  and	  C5b-­‐8	  Binding	  to	  Erythrocyte	  Membranes	  .....	  30	  4.8	   In	  vivo	  rodent	  model	  of	  laser-­‐induced	  choroidal	  neovascularization	  (CNV)	  ............................	  31	  4.9	   Quantification	  of	  C5b-­‐9	  (MAC)	  Deposited	  on	  ARPE-­‐19	  (RPE)	  Cells	  and	  RF/6A	  (CEC)	  ...............	  33	  4.10	   Quantification	  of	  SC5b-­‐9	  ........................................................................................................	  34	  4.11	   Quantification	  of	  Nuclear	  Integrity	  .........................................................................................	  34	  4.12	   Effect	  of	  PolyP	  on	  the	  Integrity	  of	  H2O2	  ................................................................................	  35	  4.13	   Catalase	  Assay	  ........................................................................................................................	  35	  4.14	   qRT-­‐PCR	  Catalase	  Gene	  Expression	  Analysis	  ..........................................................................	  36	  4.15	   VE-­‐cadherin	  Expression	  Assay	  ................................................................................................	  36	  4.16	   EA.hy	  926	  Cell	  Proliferation	  Assay	  ..........................................................................................	  37	  4.17	   BrdU	  Assay	  for	  Cell	  Proliferation	  ............................................................................................	  37	  4.18	   Statistics	  ..................................................................................................................................	  38	  4.19	   Ethics	  ......................................................................................................................................	  38	  Chapter	  5:	   Results	  ....................................................................................................	  39	  5.1	   Mechanisms	  by	  which	  PolyP	  Suppresses	  Complement	  Activation	  ...........................................	  39	  5.1.1	   PolyP	  Interacts	  Directly	  with	  C5b,6	  ...................................................................................	  40	  5.1.2	   PolyP≥1000	  Interferes	  with	  Binding	  of	  C5b-­‐7	  and	  C5b-­‐8	  Complexes	  to	  Erythrocyte	  Membranes	  .......................................................................................................................	  42	  5.1.3	   Human	  Platelet	  Releasates	  Suppress	  Complement	  Activation	  via	  the	  Terminal	  Pathway	  44	  5.2	   In	  vivo	  Validation	  of	  Therapeutic	  Utility	  of	  PolyP	  in	  a	  Complement-­‐mediated	  Disease	  ...........	  46	  	  	   vi	  5.2.1	   Comparing	  the	  Effect	  of	  PolyP130	  and	  PolyP≥1000	  on	  Complement	  Activation	  ..............	  46	  5.2.2	   PolyP≥1000	  Suppresses	  Complement-­‐mediated	  Damage	  in	  a	  Rodent	  Model	  of	  Wet	  AMD	   47	  5.2.3	   PolyP≥1000	  is	  Functionally	  Stable	  for	  at	  least	  10	  days	  in	  Serum	  ......................................	  51	  5.3	   Cellular	  Effects	  of	  PolyP	  ............................................................................................................	  52	  5.3.1	   PolyP130	  Suppresses	  C5b-­‐9	  Deposition	  on	  CEC	  and	  RPE	  Cells	  ..........................................	  53	  5.3.2	   Effects	  of	  PolyP130	  on	  Release	  of	  SC5b-­‐9	  from	  CEC	  and	  RPE	  Cells	  ...................................	  56	  5.3.3	   PolyP130	  and	  Oxidative	  Stress	  ..........................................................................................	  58	  5.3.4	   Morphologic	  Evidence	  that	  PolyP	  Protects	  Cells	  Against	  Oxidative	  stress	  ........................	  58	  5.3.5	   PolyP	  does	  not	  alter	  the	  Integrity	  of	  H2O2	  .......................................................................	  61	  5.3.6	   Effect	  of	  of	  PolyP130	  and	  monoP	  on	  Cellular	  Expression	  of	  Catalase	  ...............................	  63	  5.3.7	   PolyP	  Protects	  Nuclear	  Integrity	  of	  RPE	  cells	  and	  CEC	  from	  Oxidative	  Stress	  ...................	  65	  5.3.8	   Effect	  of	  PolyP130	  on	  Oxidative	  Stress-­‐Induced	  Changes	  in	  Junctional	  Protein	  Integrity	  .	  66	  5.3.9	   Effect	  of	  PolyP130	  of	  Endothelial	  Cell	  Proliferation	  ..........................................................	  69	  Chapter	  6:	   Discussion	  and	  Future	  Directions	  ............................................................	  71	  6.1	  	   PolyP:	  Complex	  Roles	  in	  Coagulation	  and	  Complement	  ..........................................................	  71	  6.2	  	   Mechanisms	  by	  Which	  PolyP	  Suppresses	  Complement	  ..........................................................	  72	  6.3	  	   Cellular	  Models	  of	  AMD:	  Limitations	  and	  Advantages	  ............................................................	  74	  6.4	  	   Role	  of	  PolyP	  and	  MonoP	  as	  an	  Anti-­‐oxidant	  ..........................................................................	  75	  6.5	  	   Lack	  of	  Proliferative	  Effect	  on	  Endothelial	  Cells	  ......................................................................	  78	  6.6	  	   Therapeutic	  Potential	  for	  AMD	  ................................................................................................	  78	  6.7	   Strengths	  and	  Weaknesses	  of	  the	  AMD	  Model	  of	  Laser-­‐induced	  CNV	  ....................................	  79	  6.8	  	   Future	  In	  vivo	  Studies	  with	  PolyP	  .............................................................................................	  80	  Chapter	  7:	   Conclusion	  ..............................................................................................	  82	  Bibliography	  ...............................................................................................................	  83	  	  	  	  	  	   vii	  List	  of	  Figures	  Figure	  1:	  The	  coagulation	  cascade..	  ..........................................................................................	  4	  Figure	  2:	  Chemical	  composition	  of	  polyphosphate	  ...................................................................	  5	  Figure	  3:	  Function	  of	  polyP	  in	  the	  coagulation	  cascade	  .............................................................	  7	  Figure	  4:	  The	  complement	  cascade.	  .......................................................................................	  10	  Figure	  5:	  PolyP	  suppresses	  complement	  activation.	  ...............................................................	  12	  Figure	  6:	  Reduced	  visual	  acuity	  in	  AMD.	  .................................................................................	  15	  Figure	  7:	  Pictorial	  representation	  of	  the	  normal	  macula	  and	  pathophysiological	  changes	  in	  AMD	  .........................................................................................................................	  17	  Figure	  8:	  PolyP	  suppresses	  the	  terminal	  pathway	  of	  complement	  in	  a	  concentration-­‐and	  size-­‐dependent	  manner.	  ..................................................................................................	  40	  Figure	  9:	  Gel	  filtration	  of	  complement	  proteins	  to	  assess	  interaction	  with	  polyP≥1000.	  ............	  41	  Figure	  10:	  Effect	  of	  PolyP	  on	  membrane	  binding/integration	  of	  C5b-­‐7	  and	  C5b-­‐8	  ...................	  43	  Figure	  11:	  Platelet	  releasates	  suppress	  complement	  activation	  via	  the	  terminal	  pathway	  ......	  45	  Figure	  12:	  Terminal	  pathway	  hemolytic	  assay	  to	  compare	  polyP130	  and	  polyP≥1000	  .................	  47	  Figure	  13:	  C5b-­‐9	  deposition	  and	  neovascularization	  after	  laser	  injury	  in	  rats	  ..........................	  49	  Figure	  14:	  C5b-­‐9	  deposition	  and	  neovascularization	  after	  laser	  injury	  in	  mice	  ........................	  50	  Figure	  15:	  Time-­‐dependent	  change	  in	  function	  of	  polyP≥1000	  in	  serum.	  ...................................	  52	  Figure	  16:	  Flow	  cytometry	  to	  detect	  C5b-­‐9	  deposition	  on	  CEC	  and	  RPE	  cells.	  ..........................	  55	  Figure	  17:	  	  Measurement	  of	  SC5b-­‐9	  formation	  with	  ELISA.	  .....................................................	  57	  Figure	  18:	  Cell	  morphologic	  changes	  in	  response	  to	  oxidative	  damage	  and	  polyP130	  ...............	  60	  Figure	  19:	  Effect	  of	  polyP130	  on	  functional	  integrity	  H2O2.	  .......................................................	  62	  Figure	  20:	  Effect	  of	  polyP130	  and	  monoP	  on	  catalase	  activity.	  .................................................	  64	  Figure	  21:	  Effect	  of	  polyP130	  and	  monoP	  on	  catalase	  gene	  expression	  .....................................	  64	  Figure	  22:	  Protective	  effect	  of	  polyP130	  on	  nuclear	  integrity	  following	  cell	  exposure	  to	  oxidative	  stress.	  ....................................................................................................................	  66	  Figure	  23:	  Effect	  of	  polyP130	  on	  VE-­‐cadherin	  expression.	  ........................................................	  68	  Figure	  24:	  Effect	  of	  polyP130	  and	  monoP	  on	  cell	  proliferation.	  ................................................	  70	  	  	   viii	  List	  of	  Abbreviations	  ADP	   	   adenosine	  diphosphate	   	  aHUS	  	   atypical	  hemolytic	  uremic	  syndrome	  AMD	  	  	   age-­‐related	  macular	  degeneration	  APC	  	   	   activated	  protein	  C	  apoA1/A2/E	  	   apoliproprotein	  A1/A2/E	  ARVO	  	   The	  Association	  for	  Research	  in	  Vision	  and	  Ophthalmology	  AT	  	   	   antithrombin	  ATP	   	   adenosine	  triphosphate	  BCA	   	   bicinchoninic	  acid	  BrdU	  	  	   bromodeoxyuridine	  C1INH	  	   C1	  esterase	  inhibitor	  cDNA	  	   complementary	  deoxyribonucleic	  acid	  CEC	  	   	   choroidal	  endothelial	  cells	  CFH	  	   	   complement	  factor	  H	  CNV	   	   choroidal	  neovascularization	  CR1	   	   complement	  receptor	  1	  cRBC	  	  	   chicken	  red	  blood	  cell	  CRP	  	   	   C-­‐reactive	  protein	  DAF	   	   decay	  accelerating	  factor	  FACS	  	  	   fluorescence	  activated	  cell	  sorting	  FB	   	   factor	  B	  	  	   ix	  FH	  	   	   factor	  H	  FI	  	   	   factor	  I	  GWAS	  	   genome	  wide	  association	  study	  GVB	  	   	   gelatin	  veronal	  buffer	  HI-­‐FBS	  	   heat	  inactivated	  fetal	  bovine	  serum	  HMWK	   high	  molecular	  weight	  kininogen	  iNKT	  cells	  	   invariant	  natural	  killer	  T	  cells	  Kal	  	   	   Kallikrein	  MAC	   	   membrane	  attack	  complex	  MASP	  	   MBL-­‐associated	  serine	  protease	  MBL	  	   	   mannose-­‐binding	  lectin	  monoP	  	   monophosphate	  Oligo-­‐dT	  	   oligomeric	  deoxy	  thymine	  nucleotides	  PBS	  	   	   phosphate	  buffered	  saline	  PGE	  	   	   prostaglandin	  E	  PK	  	   	   prekallikrein	  PNH	  	   	   paroxysmal	  nocturnal	  hemoglobinuria	  polyP	  	   polyphosphate	  PPK1	  	   polyphosphate	  kinase	  1	  PPK2	  	  	   polyphosphate	  kinase	  2	  qRT-­‐PCR	  	   quantitative	  real-­‐time	  polymerase	  chain	  reaction	  ROS	  	   	   reactive	  oxygen	  species	  RPE	  	   	   retinal	  pigment	  epithelial	  	  	   x	  RT-­‐PCR	   reverse	  transcription	  polymerase	  chain	  reaction	  SAP-­‐A	   serum	  amyloid	  A	  SC5b-­‐9	  	   soluble	  (or	  S-­‐protein)	  C5b-­‐9	  TER	  	   	   transepithelial	  resistance	  TF	   	   tissue	  factor	  TFPI	  	   	   tissue	  factor	  pathway	  inhibitor	  VE-­‐cadherin	  	   vascular	  endothelial	  cadherin	  VEGF	  	   vascular	  endothelial	  growth	  factor	  VWF	  	  	   von	  Willebrand	  Factor	  	  	  	  	  	  	  	  	  	  	  	  	  	   xi	  Acknowledgements	  Completion	  of	  this	  thesis	  would	  not	  have	  been	  possible	  without	  the	  efforts	  of	  many	  who	  have	  shared	  in	  this	  endeavor	  with	  me,	  some	  of	  whom	  deserve	  special	  mention.	  Thanks	  to	  Ed,	  the	  best	  supervisor	  a	  grad	  student	  can	  hope	  for.	  You’ve	  encouraged	  me	  to	  think	  creatively,	  and	  believed	  in	  my	  abilities	  when	  I	  myself	  was	  doubtful.	  I	  can’t	  thank	  you	  enough	  for	  the	  many	  opportunities	  you’ve	  provided	  for	  me	  to	  grow	  professionally,	  and	  especially	  for	  your	  dedication	  and	  guidance	  while	  writing	  this	  thesis.	  	  Thank	   you	   also	   to	  my	   supervisory	   committee	  members;	   Dr.	   Kelly	  McNagny,	   Dr.	   Jaychandran	  Kizhakkedathu,	  past	   committee	   chair	  Dr.	  Colby	   Zaph,	   and	  present	   committee	   chair	  Dr.	  Haydn	  Pritchard,	   for	   providing	   invaluable	   insight	   to	   my	   project,	   and	   challenging	   me	   to	   be	   a	   better	  student	  and	  scientist.	  This	  would	   not	   have	   been	   as	   enjoyable	   of	   an	   experience	  without	  my	   lab	  mates,	  whom	   I	   am	  happy	   to	   also	   be	   able	   to	   call	   friends.	   Victor,	   thanks	   for	   guaranteeing	   there	   is	   never	   a	   dull	  moment	  in	  the	  lab.	  Alice,	  thanks	  for	  being	  an	  exceptional	  mouse-­‐whisperer	  and	  partner	  in	  the	  in	  vivo	   project.	   Jovian,	   thanks	   for	   helping	  me	   start	   off	   on	   the	   right	   foot.	   My	   summer	   students	  Jasmine	  and	  Cameron,	  thanks	  for	  your	  diligence	  in	  the	  experiments.	  Past	  teammates	  (Jonathan,	  Erica,	   Sahana,	  Mike,	   Stephanie,	   Kent,	   Jaclyn,	   Josh,	   Jesi)	   and	   current	   teammates	   (Emilie,	   Tara,	  Yolanda,	   Piyush,	   Jenny),	   thanks	   for	   giving	   me	   support,	   guidance,	   and	   many	   memorable	  moments.	  	  I	  would	  not	  be	  able	  to	  achieve	  anything	  without	  the	  unwavering	  love	  and	  support	  of	  my	  family	  and	  relatives.	  My	  Mom	  and	  Dad,	  for	  the	  sacrifices	  they’ve	  made	  and	  for	  the	  care	  they	  give	  so	  I	  can	   do	  my	   best.	   Kuya,	   for	  making	  me	  want	   to	   be	   smarter.	   Jeno,	   for	   never	   failing	   to	   provide	  comic	  relief.	  	  Thanks	  also	  to	  Ate,	  for	  being	  my	  number	  one	  cheerleader	  from	  far	  away.	  For	  all	  my	  friends	  outside	  the	  lab,	  thank	  you	  for	  your	  support	  and	  for	  helping	  me	  live	  a	  balanced	  life.	   	   Special	   thanks	   to	  Echo,	  who	  has	  been	  a	  patient	  and	  encouraging	  best	   friend	   throughout	  this	  entire	  process.	  	  Last	  but	  not	  least,	  to	  God,	  for	  giving	  science	  its	  meaning	  and	  beauty.	  	  	  	   xii	  Dedication	  To	  my	  parents,	  who	  were	  the	  first	  ones	  to	  teach	  me	  the	  value	  of	  education.	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   1	  Chapter	  1: Introduction	  Organisms	   have	   developed	   means	   to	   contain	   wounds	   by	   limiting	   bleeding	   and	   eliminating	  pathogens	   and	   damaged	   host	   cells	   via	   the	   recruitment	   of	   innate	   defense	   mechanisms,	  particularly	  within	  the	  vascular	  system	  [1].	  Disease	  emerges	  when	  there	  is	  unchecked	  activation	  of	   innate	   immune	   and/or	   coagulation	   responses.	   These	   include,	   for	   example,	   atherosclerosis,	  stroke,	  venous	  thromboembolic	  disease,	  arthritis,	  atypical	  hemolytic	  uremic	  syndrome,	  and	  age-­‐related	  macular	   degeneration.	   Understanding	   the	  mechanisms	   that	   regulate	   innate	   immunity	  and	  coagulation	  will	  uncover	  strategies	  to	  reduce	  tissue	  damage	  in	  these	  diseases.	  The	   focus	   of	   the	   Conway	   lab	   is	   to	   explore	   molecular	   interactions	   between	   the	   coagulation	  cascade,	   and	   another	   major	   blood-­‐borne	   proteolytic	   cascade,	   complement,	   which	   is	   a	   key	  component	   of	   innate	   immunity,	   in	   the	   hopes	   of	   gaining	   novel	   therapeutic	   insights.	   The	  complement	  and	  coagulation	  systems	  are	  highly	  regulated	  proteolytic	  cascades,	  which	  result	  in	  the	  containment	  of	  infection	  and	  bleeding,	  respectively.	  This	  is	  achieved	  through	  the	  production	  of	  final	  effectors.	  In	  the	  complement	  system,	  the	  final	  effector	  is	  the	  membrane	  attack	  complex	  (MAC),	   which	   causes	   lysis	   and	   opsonization	   of	   pathogenic	   organisms	   or	   damaged	   host	   cells.	  Byproducts	   of	   the	   enzymatic	   reactions	   are	   anaphylatoxins,	   which	   initiate	   a	   range	   of	   pro-­‐inflammatory	  events.	  Analogous	  to	  the	  complement	  activation	  pathway,	  the	  final	  effector	  in	  the	  coagulation	  system	  is	  thrombin,	  a	  serine	  protease	  that	  cleaves	  fibrinogen	  to	  fibrin	  to	  generate	  a	  fibrin	  clot.	  Recent	  studies	  have	  revealed	  significant	  overlap	  between	  these	  two	  systems.	  This	  is	  best	  exemplified	  by	  the	  dramatic	  success	  of	  the	  inhibitor	  of	  complement	  activation,	  eculizumab,	  in	   preventing	   the	   thrombotic	   manifestations	   of	   two	   diseases	   that	   are	   caused	   by	   excess	  complement	  activation	  –	  atypical	  hemolytic	  uremic	  syndrome	  (aHUS)	  and	  paroxysmal	  nocturnal	  hemoglobinuria	  (PNH)	  [2,	  3].	  Work	  initiated	  in	  the	  Conway	  lab,	  prior	  to	  my	  arrival,	  uncovered	  a	  factor	  -­‐	  polyphosphate	  (polyP)	  -­‐	   that	   co-­‐regulates	   complement	   and	   coagulation.	   This	   thesis	   aims	   to	   further	   characterize	   the	  mechanisms	   by	   which	   polyP	   functions,	   and	   to	   test	   its	   physiological	   relevance	   in	   a	   model	   of	  human	  disease.	  	  	   2	  In	  this	  Introduction,	  I	  will	  first	  provide	  a	  brief	  background	  of	  coagulation,	  highlighting	  how	  polyP	  participates.	  This	  will	  be	  followed	  by	  a	  discussion	  of	  the	  complement	  cascade	  and	  its	  regulation.	  A	  more	  detailed	  description	  of	  the	  origins	  and	  functions	  of	  polyP	  will	  be	  provided,	  followed	  by	  a	  review	   of	   the	   findings	   that	   led	   to	   this	   thesis,	   i.e.	   the	   discovery	   of	   how	   polyP	   regulates	  complement.	   Finally,	   I	   will	   discuss	   the	   clinical	   course	   and	   pathogenesis	   of	   a	   common,	  complement-­‐mediated	   cause	   of	   blindness	   -­‐	   age	   related	   macular	   degeneration	   –	   setting	   the	  stage	  for	  in	  vivo	  investigations	  with	  polyP.	  1.1	   The	  Coagulation	  Cascade	  Coagulation	   is	   a	  major	   proteolytic	   cascade	   in	   the	   blood,	   designed	   to	   prevent	   excess	   bleeding	  upon	   injury.	   It	   has	   two	   initiating	   pathways	   –	   extrinsic	   and	   intrinsic.	   The	   extrinsic	   pathway	   is	  triggered	   by	   exposure	   of	   tissue	   factor	   (TF)	   to	   circulating	   blood,	   either	   by	   damage	   to	   the	  endothelium	   or	   activation	   of	   circulating	   and/or	   subendothelial	   cells	   (e.g.	   monocytes,	  macrophages,	  pericytes,	  smooth	  muscle	  cells)	  by	  inflammatory	  stimuli	  [4].	  TF	  forms	  a	  complex	  with	   small	   amounts	   of	   constitutively	   circulating	   factor	   VIIa,	   thereby	   forming	   the	   so-­‐called	  extrinsic	   factor	   tenase	   complex.	   This	   complex,	   situated	   on	   cell	   surfaces	   at	   the	   site	   of	   injury,	  activates	  coagulation	  factors	  IX	  and	  X	  to	  generate	  factors	  IXa	  and	  factor	  Xa,	  respectively.	  Factor	  Xa	   produced	   from	   the	   tenase	   complex,	   cleaves	   prothrombin	   to	   generate	   the	   key	   enzyme	   in	  coagulation,	  factor	   IIa	  (thrombin).	  Like	  most	  coagulation	  steps,	  this	  requires	  the	  presence	  of	  a	  cofactor	  –	  in	  this	  case,	  factor	  Va	  –	  as	  well	  as	  a	  phospholipid	  surface	  and	  calcium	  ions	  [4].	  	  With	  thrombin	  generated	  beyond	  a	  threshold,	  it	   is	  believed	  to	  feed	  back	  to	  activate	  factor	  XI	  of	  the	  intrinsic	  pathway,	  thereby	  amplifying	  its	  own	  production.	  Thrombin	  cleaves	  fibrinogen	  to	  yield	  fibrin	  monomers.	  Formation	  of	  a	  stable	  fibrin	  clot	  is	  achieved	  by	  cross-­‐linking	  of	  the	  monomers	  by	  factor	  XIIIa	  that	  is	  itself	  generated	  by	  thrombin-­‐mediated	  activation	  of	  factor	  XIII	  [5]	  (Figure	  1).	  	  The	  intrinsic	  pathway	  is	  initiated	  by	  activation	  of	  factor	  XII	  (Hageman	  Factor)	  to	  factor	  XIIa	  upon	  contact	   with	   a	   negatively	   charged	   surface	   [5,	   6].	   Factor	   XIIa	   promotes	   clot	   formation	   by	  activating	  factor	  XI,	  which	  in	  turn	  activates	  factor	  IX.	  In	  the	  presence	  of	  factor	  VIIIa	  (carried	  by	  	  	   3	  von	  Willebrand	  Factor,	  VWF),	  factor	  IXa	  activates	  factor	  X,	  thereby	  channeling	  into	  the	  common	  pathways,	  with	  subsequent	  generation	  of	   thrombin.	  Factor	  XIIa	  also	  cleaves	  prekallikrein	   (PK),	  which	  is	  bound	  to	  high	  molecular	  weight	  kininogen	  (HMWK)	  on	  the	  surface	  of	  the	  membrane.	  Cleavage	  of	  PK	  generates	  the	  serine	  protease	  kallikrein	  (Kal),	  which	  activates	  HMWK	  to	  HMWKa.	  Kal	  is	  also	  able	  to	  further	  cleave	  factor	  XII	  bound	  on	  the	  surface	  of	  the	  membrane,	  resulting	  in	  feedback	  amplification	  of	  the	  pathway	  [6].	  An	  important	  negative	  regulator	  of	  factors	  XIa,	  XIIa	  and	  kallikrein	  is	  C1	  esterase	  inhibitor	  (C1INH)	  [5,	  7,	  8].	  	  Patients	  with	   factor	   XII	   deficiency	  do	  not	   bleed.	  However,	   lack	  of	   factor	   XII	   in	   rodent	  models	  protects	  against	  thrombosis	  and	  immune	  defects,	  and	  elevated	  factor	  XI	  and	  factor	  XII	  levels	  in	  humans	   increase	  the	  risk	  of	  atherothrombosis.	  This	  relatively	  recent	  recognition	  of	  the	  clinical	  importance	   of	   the	   contact	   pathway	   has	   resulted	   in	   increased	   efforts	   to	   understand	   how	   it	   is	  initiated.	   In	  fact,	  the	  mechanism	  of	  triggering	  activation	  of	  factor	  XII	  remained	  a	  mystery	  until	  approximately	   10	   years	   ago,	   when	   a	   physiologic	   in	   vivo	   anionic	   surface	   on	   which	   factor	   XII,	  prekallikrein	  and	  HMWK	  assemble	  for	  activation	  of	  factor	  XII,	  was	  identified	  as	  polyphosphate	  (polyP)	  [9].	  	  This	  discovery	  of	  a	  role	   for	  polyP	  has	  not	  only	  opened	  the	  door	  for	  the	  development	  of	  novel	  therapeutic	  interventions	  to	  prevent	  vascular	  disease	  and	  coagulation	  abnormalities,	  but	  it	   led	  to	   studies	   in	   the	   Conway	   lab	   to	   examine	   the	   role	   of	   polyP	   in	   the	   other	   major	   blood-­‐borne	  proteolytic	  cascade,	  complement,	  thereby	  setting	  the	  stage	  for	  this	  thesis.	  	  	   4	  	  Figure	   1:	   The	   coagulation	   cascade.	   Activation	   of	   coagulation	   is	   triggered	   by	   injury	   to	   a	   cell	   surface	  (endothelial	   in	   the	   above	   example),	   exposing	   initiators	   of	   the	   intrinsic	   (HMWK-­‐PK,	   factor	   XII)	   and	  extrinsic	  (TF-­‐VIIa)	  pathways.	  These	  converge	  with	  transformation	  of	  factor	  X	  to	  Xa,	  and	  prothrombin	  (II)	  to	   thrombin	   (IIa).	   The	   cascade	   culminates	   with	   the	   formation	   of	   a	   stable	   fibrin	   clot.	   Key	   negative	  regulators	  of	  the	  cascade	  that	  prevent	  excess	  thrombin	  generation	  and	  clot	  formation	  (not	  discussed	  in	  the	   thesis	   text)	   are	   shown	   in	   red	   (AT	   –	   antithrombin;	   APC	   –	   activated	   protein	   C;	   TFPI	   –	   tissue	   factor	  pathway	  inhibitor).	  	  	  	  1.2	   Polyphosphate	  (PolyP)	  Polyphosphate	   (polyP)	   is	   an	   inorganic	   linear	   polymer	   comprising	   orthophosphate	   monomers	  which	  are	  linked	  by	  high-­‐energy	  phosphoanhydride	  bonds	  [10]	  (Figure	  2).	  PolyP	  is	  ubiquitously	  expressed,	  found	  in	  virtually	  all	  organisms,	  but	  with	  variable	  polymer	  lengths	  [11,	  12].	  Bacteria	  produce	  longer	  lengths	  of	  the	  polymer	  (>300	  to	  1000	  residues),	  while	  mammals	  produce	  shorter	  	  	   5	  forms	  (from	  5-­‐800	  residues)	  [11,	  13,	  14].	   In	  the	   last	  3-­‐4	  decades,	  bacterial	  polyP	  has	  been	  the	  best	   characterized.	   It	   is	   essential	   for	   the	   pathogenicity	   of	  most	   prokaryotes,	   playing	   a	   crucial	  role	   in	  metabolism,	   immune	  protection,	  as	  a	  source	  of	  energy,	  and	  as	  an	  anti-­‐oxidant	  [15-­‐17].	  The	  major	   enzyme	   that	   synthesizes	   polyP	   in	   bacteria	   is	   polyP	   kinase	   1	   (PPK1),	   loss	   of	   which	  results	  in	  decreased	  virulence	  [15,	  18].	  PolyP	  kinase	  2	  (PPK2)	  is	  the	  other	  bacterial	  enzyme	  that	  synthesizes	  polyP,	  and	  may	  also	  play	  a	  role	  in	  phosphorylating	  ADP	  to	  ATP,	  allowing	  the	  bacteria	  to	  use	  polyP	  as	  an	  energy	  reserve	  [16].	  PolyP	  can	  also	  be	  synthesized	  in	  vitro,	  and	  is	  used	  for	  a	  variety	   of	   industrial	   purposes	   such	   as	   flame-­‐retardants,	   fertilizers,	   water	   treatment	   reagents,	  and	  food	  additives	  [19].	  	  	  	  Figure	   2:	   Chemical	   composition	   of	   polyphosphate.	   Polyphosphate	   (polyP)	   consists	   of	   a	   string	   of	  orthophosphate	  monomers	  (bracketed).	  It	  is	  found	  in	  varying	  lengths	  in	  all	  living	  organisms,	  ranging	  from	  tens	  to	  more	  than	  a	  thousand	  units	  long.	  	  	  1.2.1	   Polyphosphate	  and	  Coagulation	  Extensive	   studies	   in	   the	   last	   decade	   have	   revealed	   that	   polyP	   plays	   an	   important	   role	   in	  coagulation	  [20,	  21].	  Although	  ubiquitously	  expressed,	  polyP	  is	  found	  at	  high	  concentrations	  in	  the	   dense	   granules	   of	   platelets,	   which	   are	   analogous	   to	   acidocalcisomes	   of	   bacteria	   [9,	   22].	  When	   platelets	   are	   activated,	   polyP	   with	   a	   mean	   chain	   length	   of	   ~70	   orthophosphate	   units	  (range	  ~60-­‐100)	   is	   released	   [23].	   It	   has	  not	  been	  established	  what	  happens	   to	   the	  polyP,	   i.e.,	  	  	   6	  whether	   it	   binds	   to	   the	   surface	  of	   cells	   (e.g.,	   platelets,	   endothelial	   cells),	   or	   if	   it	   is	   cleared	  by	  degradation,	  the	  kidney	  or	  the	  reticuloendothelial	  system.	  PolyP	  enhances	  coagulation	  at	  different	  steps	  in	  the	  cascade	  [23],	  and	  the	  site	  of	  action	  depends	  partly	   on	   the	   length	   of	   the	   polymer	   (Figure	   3).	   For	   primary	   hemostasis,	   the	   concentration	   of	  polyP	  released	  from	  platelets	  was	  sufficient	  to	  initiate	  fibrin	  formation	  [23].	  PolyP	  of	  at	  least	  45	  orthophosphate	   units,	   at	   a	   concentration	   of	   >1	   µg/mL	   is	   able	   to	   activate	   factor	   XII	   and	  prekallikrein	  in	  human	  or	  murine	  plasma	  [23].	  PolyP	  reduces	  TF-­‐induced	  clotting	  time	  almost	  6-­‐fold	  in	  normal	  plasma,	  and	  accelerates	  TF-­‐triggered	  clotting	  in	  factor	  XII-­‐deficient	  human	  plasma	  ~1.3	  fold	  compared	  to	  TF	  alone	  [23].	  PolyP	  (~25-­‐75	  orthophosphate	  units	  long)	  achieves	  this	  by	  abrogating	   the	   anti-­‐coagulant	   effect	   of	   tissue	   factor	   pathway	   inhibitor	   (TFPI)	   [9].	   PolyP	   also	  accelerates	  factor	  V	  activation	  by	  thrombin	  [9].	  Further	  downstream,	  polyP	  at	  a	  concentration	  of	  as	  little	  as	  0.5	  µg/mL	  dramatically	  accelerates	  clotting	  in	  the	  presence	  of	  exogenous	  factor	  Va	  [23].	   In	   the	   intrinsic	  pathway,	  polyP	   is	   able	   to	  decrease	   the	   time	  and	   increase	   the	  amount	  of	  thrombin	   generation	   [23].	   In	   addition,	   platelet-­‐sized	   polyP	   accelerates	   factor	   XI	   activation	   by	  thrombin,	  potentially	  acting	  as	  a	  cofactor	  [24].	  These	  pro-­‐coagulant	  activities	  are	  dependent	  on	  the	   presence	   of	   factor	   XII	   [23].	   	   PolyP	   also	   affects	   fibrin	   formation,	   increasing	   clot	   integrity,	  which	  results	  in	  prolonged	  fibrinolysis	  time	  [23,	  25].	  This	  function,	  unlike	  polyP’s	  pro-­‐coagulant	  effects	  upstream	  of	  thrombin	  generation,	  is	  independent	  of	  factor	  XII	  activation	  [23].	  In	  platelet-­‐poor	  plasma,	  polyP	  (~75	  orthophosphate	  units)	  reduces	  the	  clotting	  time	  more	  than	  5-­‐fold	  and	  delays	  fibrinolysis	  ~1.6-­‐fold	  [9].	  	  	  	  	   7	  	  Figure	  3:	  Function	  of	  polyP	   in	   the	  coagulation	  cascade.	  The	  role	  of	  polyP	  in	  the	  coagulation	  cascade	  is	  partly	   dependent	  on	   the	   length	  of	   the	  polymer,	   depicted	  here	   simply	   as	   “long-­‐chain”	   (A)	   or	   “platelet-­‐size”	  (B)	  polyP.	  Long-­‐chain	  polyP	  (up	  to	  several	  thousand	  units)	   is	  most	  effective	  at	  activating	  factor	  XII	  (1).	   Both	   long	   and	   short	   polyP	   chains	   (~60-­‐100	   units	   long)	   enhance	   thrombin-­‐mediated	   activation	   of	  factor	  V	  (2),	  increase	  clot	  integrity	  (3),	  and	  accelerate	  back-­‐activation	  of	  factor	  XI	  by	  thrombin	  (4).	  Figure	  adapted	  from	  Morrissey	  et	  al.	  2012,	  Blood.	  [20]	  	  The	  molecular	  mechanisms	  underlying	  polyP’s	  procoagulant	  properties	  are	  not	  yet	  well-­‐defined.	  PolyP	   binds	   tightly	   to	   thrombin	   but	   does	   not	   bind	   to	   prothrombin	   [26].	   It	   also	   binds	   to	  fibrinogen,	   factors	   XI	   and	   XII,	   prekallikrein,	   HMWK,	   and	   von	   Willebrand	   factor	   [25,	   27].	  Intravenous	   infusion	   of	   polyP	   into	   mice	   induces	   thrombosis	   [23].	   Consistent	   with	   the	  biochemical	  studies,	  mice	  lacking	  factor	  XII	  are	  protected	  against	  polyP-­‐induced	  thrombosis	  [23,	  28].	  	  	  	  	   8	  Overall,	  there	  are	  ample	  data,	  in	  vitro	  and	  in	  vivo	  in	  mouse	  models,	  to	  support	  the	  notion	  that	  polyP	   promotes	   coagulation.	  With	   increasing	   evidence	   of	   interplay	   between	   coagulation	   and	  innate	  immunity,	  our	  lab	  examined	  the	  role	  of	  polyP	  in	  a	  key	  blood-­‐borne	  component	  of	  innate	  immunity,	  the	  complement	  system.	  1.3	   The	  Complement	  System	  The	  complement	  system	  is	  a	  proteolytic	  system	  that	  comprises	  over	  30	  soluble	  and	  membrane-­‐bound	   proteins	   that	   contribute	   to	   innate	   and	   adaptive	   immunity	   [29].	   Analogous	   to	   the	  coagulation	   system,	   complement	   activation	   is	   tightly	   regulated	   to	   ensure	   a	   rapid,	   highly	  localized	  and	  temporally	  restricted	  response	  to	  promote	  clearance	  of	  the	  offending	  pathogen	  or	  injured	  cells,	  while	  limiting	  damage	  to	  surrounding	  tissue,	  and	  facilitating	  healing	  [1].	  	  There	   are	   three	   pathways	   by	  which	   the	   complement	   system	   is	   initiated;	   classical,	   lectin,	   and	  alternative	  [30]	  (Figure	  4).	  The	  classical	  pathway	  is	  triggered	  by	  the	  binding	  of	  the	  C1	  complex	  to	  a	  pathogen	  surface,	  directly	  or	  via	  an	  IgG	  or	  IgM	  antibody	  [31].	  The	  C1	  complex	  comprises	  a	  C1q	  molecule	  and	  two	  each	  of	  C1r	  and	  C1s	  [32].	  C1q	  has	  a	  collagen	  tail	  and	  6	  globular	  heads	  which	  are	  responsible	  for	  attaching	  to	  pathogens	  [31].	  When	  more	  than	  one	  globular	  head	  has	  bound	  to	  a	  pathogen,	  the	  C1r2C1s2	  complex	  undergoes	  a	  conformational	  change,	  activating	  each	  C1r	  to	  cleave	  one	  C1s.	  This	  transforms	  C1s	  to	  a	  serine	  protease,	  which	  then	  cleaves	  C4	  to	  C4a	  and	  C4b	  [31].	  C4b	  attaches	  to	  a	  pathogen,	  recruiting	  circulating	  C2,	  bringing	  it	   into	  close	  proximity	  for	  cleavage	  by	  C1s	  into	  C2a	  and	  C2b.	  C4b	  and	  C2b	  form	  the	  C4bC2b	  complex,	  also	  known	  as	  C3	  convertase	  [33].	  (Note	  that	  we	  use	  current	  nomenclature	  for	  C2a	  and	  C2b,	  the	  latter	  being	  the	  larger	  fragment).	  The	   lectin	  pathway,	  also	  known	  as	   the	  mannose-­‐binding	   lectin	   (MBL)	  pathway,	   is	   triggered	  by	  the	   interaction	   of	   MBL	   to	   a	   mannose-­‐containing	   polysaccharide	   on	   pathogen	   surfaces.	   This	  results	   in	   binding	   of	  MBL-­‐associated	   serine	   protease	   (MASP)	   1	   and	   2	   to	   the	  MBL.	   The	  MBL-­‐MASP1-­‐MASP2	  complex	  results	  in	  activation	  of	  the	  MASPS,	  cleaving	  C4	  into	  C4a	  and	  C4b,	  in	  an	  identical	  fashion	  as	  for	  C1s,	  yielding	  again	  the	  C4b2b	  C3	  convertase	  [34,	  35].	  	  	  	   9	  The	  alternative	  pathway	  is	  unlike	  the	  other	  two	  pathways	  in	  that	  it	  does	  not	  require	  a	  pathogen	  surface	  or	  antibody	  to	  be	  initiated,	  i.e.,	  it	  is	  constitutively	  active	  at	  low	  levels,	  but	  may	  be	  rapidly	  amplified,	   wherein	   it	   does	   require	   exposure	   to	   non-­‐host	   surfaces.	  Moreover,	   the	   alternative	  pathway	   is	  highly	  dependent	  on	   the	  presence	  of	   ionic	  calcium	  and	   ionic	  magnesium,	  whereas	  the	   classical	   and	   lectin	   pathways	   are	   only	   dependent	   on	   ionic	   calcium	   [36].	   The	   alternative	  pathway	  relies	  on	  the	  fact	  that	  circulating	  C3	  is	  constantly	  being	  hydrolyzed	  to	  a	  C3b-­‐like	  moiety	  known	  as	  C3(H2O2),	  to	  which	  the	  active	  fragment	  of	  factor	  B,	  Bb,	  may	  bind,	  thereby	  forming	  a	  solution	  phase	  C3(H2O2)Bb	  C3	  convertase.	  Most	  of	  this	  is	  inactivated	  by	  negative	  regulators	  (see	  below).	   In	   a	   setting	   requiring	   recruitment	   of	   an	   innate	   immune	   response,	   this	   C3	   convertase	  may	   further	   cleave	   C3	   to	   generate	   C3a	   +	   C3b,	   the	   latter	  which	   binds	   to	   a	   pathogen	   surface,	  providing	   a	   binding	   site	   for	   factor	   B.	   Factor	   B	   is	   cleaved	   by	   Factor	   D	   (FD),	   yielding	   Bb	   and	  liberating	  Ba.	  Bb	  binds	   to	  C3b,	   forming	  C3bBb,	   the	  alternative	  pathway	  C3	  convertase.	  This	   is	  relatively	   unstable	   compared	   to	   the	  C3	   convertase	  of	   the	   lectin	   and	   classical	   pathway,	   and	   is	  stabilized	  by	  properdin	  (Factor	  P)	  [37].	  As	  can	  be	  seen	  from	  the	  above,	  the	  three	  pathways	  converge	  with	  formation	  of	  a	  C3	  convertase,	  which	  cleaves	  C3	  to	  C3a	  and	  C3b.	  C3a	  is	  released	  into	  the	  circulation	  and	  is	  a	  relatively	  potent	  anaphylatoxin,	  promoting	   the	   recruitment	  of	   inflammatory	   cells	   to	   the	   site	  of	   infection/injury	  [38].	   As	   the	   density	   of	   C3b	   deposited	   on	   the	   target	   surface	   increases	   and	   binds	   to	   the	   C3	  convertases,	  the	  substrate	  specificity	  of	  the	  convertase	  shifts	  to	  C5,	  and	  thus,	  C4b2b(C3b)n	  and	  C3bBb(C3b)n	  become	  C5	  convertases	  [39].	  C5	  convertase	  cleaves	  C5	  into	  C5a	  and	  C5b.	  C5a	  is	  an	  anaphylatoxin	   that	   is	   more	   potent	   that	   C3a	   [40].	   C5b	   immediately	   binds	   to	   circulating	   C6,	  forming	  a	  strong	  C5b-­‐6	  complex.	  This	  allows	  C7	  to	  bind	  via	  C6,	  forming	  the	  C5b-­‐7	  complex	  that	  integrates	  partially	   into	   the	  membrane	  of	   the	  target	  cell	  or	  pathogen.	  C8	  then	  binds	   to	  C5b-­‐7	  and	  this	  C5b-­‐8	  complex	  integrates	  completely	  into	  (but	  not	  through)	  the	  membrane.	  Numerous	  C9	  molecules	   then	  attach	   (up	   to	  ~20	  per	  C5b-­‐8)	  and	  extend	   to	   form	  a	  pore-­‐like	  complex	   that	  inserts	  into	  the	  membrane,	  causing	  damage	  and	  lysis	  of	  the	  target	  cell	  or	  pathogen	  [41,	  42].	  This	  C5b-­‐9	  complex	   is	  often	  referred	  to	  as	   the	  membrane	  attack	  complex	   (MAC)	   [43].	  Assembly	  of	  the	   MAC	   from	   C5b	   is	   known	   as	   the	   terminal	   pathway.	   It	   occurs	   spontaneously,	   in	   an	   ion-­‐	  	   10	  independent	  manner,	  and	  without	   the	  necessity	  of	  enzymes	   [43].	  Some	  C5b-­‐9	  may	  not	   insert	  into	   the	   membrane	   or	   may	   be	   forced	   out	   by	   membranous	   negative	   regulators.	   This	   is	   then	  referred	   to	   as	   soluble	   C5b-­‐9,	   S-­‐protein	   C5b-­‐9	   or	   SC5b-­‐9	   [44].	   Unlike	  membrane-­‐bound	   C5b-­‐9	  which	   is	  amphiphilic,	  SC5b-­‐9	   is	  hydrophilic,	  which	  allows	   it	   to	  be	  measured	   in	  the	  plasma	  [44,	  45].	  Although	  SC5b-­‐9	   is	  non-­‐cytolytic,	   it	   is	  known	  to	  have	  pro-­‐inflammatory	  and	  procoagulant	  properties	  [44,	  46].	  	  	  Figure	   4:	   The	   complement	   cascade.	   Complement	   is	   initiated	   via	   the	   lectin,	   classical,	   or	   alternative	  pathways.	   These	   converge	   to	   form	   a	   C3	   convertase,	   cleaving	   C3	   to	   C3a	   and	   C3b.	   With	   more	   C3b	  deposited	   on	   the	   target	   surface,	   the	   specifity	   of	   the	   so-­‐formed	   C3bBbC3b	   convertase	   shifts	   to	   C5,	  thereby	  resulting	  in	  the	  cleavage	  of	  C5	  to	  C5a	  and	  C5b.	  C6	  binds	  to	  C5b,	  yielding	  C5b,6,	  the	  starting	  point	  of	   the	   terminal	   pathway.	  With	   subsequent	   binding	   of	   C7,	   C8	   and	  multiple	   C9	  molecules,	   the	   terminal	  pathway	   ends	   with	   the	   assembly	   of	   the	   C5b-­‐9	   membrane	   attack	   complex	   (MAC),	   causing	   lysis	   of	   or	  damage	  to	  the	  pathogen	  or	  target	  cell.	  Specific	  negative	  regulators	  of	  complement	  activation,	  discussed	  in	  the	  text,	  are	  shown	  in	  red.	  This	  figure	  is	  based	  on	  a	  version	  provided	  by	  Jovian	  Wat	  [47].	  	  	  	   11	  1.3.1	   Regulation	  of	  Complement	  Activation	  As	  with	   all	   biological	   systems,	   a	   delicate	   balance	   is	   required	   to	   optimize	   the	   function	   of	   the	  complement	   system.	   Inadequate	   complement	   activation	   results	   in	   reduced	   clearance	   of	  pathogens	  and	  risk	  of	  overwhelming	  infection,	  while	  excess	  activation	  may	  cause	  tissue	  damage	  to	  the	  host	  [48].	  Thus,	  there	  are	  several	  regulatory	  mechanisms	  in	  place	  at	  various	  points	  in	  the	  cascade	  that	  keep	  the	  system	  in	  check,	  several	  of	  which	  are	  described	  in	  the	  following.	  	  The	   need	   for	   a	   membrane	   surface	   for	   most	   complement	   proteins	   localizes	   complement	  activation	  where	  it	  is	  needed.	  Unbound,	  most	  complement	  components	  are	  not	  stable	  in	  their	  active	   form.	   There	   are	   therefore	   many	   negative	   regulators	   that	   participate	   to	   prevent	  autoimmunity.	   C1	   esterase	   inhibitor	   (C1INH)	   is	   a	   circulating	   serine	   protease	   inhibitor	   (serpin)	  that	   dampens	   activation	   via	   the	   lectin	   and	   classical	   pathways.	   It	   dissociates	   the	   activated	  (C1rC1s)2	  complex	  from	  C1q,	  rendering	  it	  unable	  to	  generate	  the	  C4bC2b	  C3	  convertase.	  It	  also	  neutralizes	   C1s	   and	   the	   MASPs,	   rendering	   them	   unable	   to	   cleave/activate	   C2	   and	   C4	   [49].	  Deficiency	   of	   C1INH	   in	   humans	   results	   in	   hereditary	   angioneurotic	   edema,	   an	   inflammatory	  disease	  that	  is	  associated	  with	  excessive	  swelling	  around	  the	  airways	  [50].	  	  The	   plasma	   serine	   protease,	   Factor	   I	   (FI)	   dampens	   complement	   activation	   primarily	   via	   the	  alternative	  pathway,	  although	  it	  may	  also	  affect	  the	  other	  pathways	  [29].	  FI	  cleaves	  C3b	  and	  C4b	  bound	  on	  host	  cell	  membranes,	   rendering	  them	  inactive	   in	  terms	  of	   forming	  a	  C3	  convertase.	  C3b	  is	  cleaved	  to	  iC3b,	  and	  subsequently	  to	  C3dg,	  while	  C4b	  is	  cleaved	  to	  C4c	  and	  then	  to	  C4d	  [30].	  FI	  requires	  a	  cofactor	  to	  cleave/inactivate	  either	  C3b	  or	  C4b.	  These	  include	  soluble	  factor	  H	  (FH)	   and	  C4b-­‐binding	   protein	   (C4BP),	   and	  membrane-­‐bound	   cofactors	   CD46	   and	   complement	  receptor	  1	  (CR1)	  [50,	  51]	  (Figure	  4).	  	  FH	  has	  multiple	  mechanisms	  by	  which	   it	  negatively	  regulates	  complement.	  Although	  a	  soluble	  protein,	  FH	  attaches	  to	  polyanionic-­‐rich	  host	  cell	  surfaces,	  and	  there	  binds	  most	  effectively	  to	  C3b,	  where	   it	  acts	  as	  a	  cofactor	   for	  FI-­‐mediated	  C3b	   inactivation	   [51].	  Bacterial	   cell	  walls	   lack	  polyanion	   sialic	   acid,	   and	   thus	   complement	   activation	   is	   allowed	   to	   proceed,	   as	   effective	  	  	   12	  inactivation	  of	  C3b	  and	  C4b	  by	  FI/FH	  does	  not	  occur	  [50].	  FH	  also	  interferes	  with	  binding	  of	  B/Bb	  to	  C3b,	  thereby	  preventing	  formation	  of	  the	  alternative	  pathway	  C3	  convertase	  [52,	  53].	  It	  also	  destabilizes	  the	  C3	  convertase.	  The	  importance	  of	  FH	  is	  highlighted	  by	  the	  diseases	  that	  result	  from	   its	   deficiency,	   including	   atypical	   hemolytic	   uremic	   syndrome	   (aHUS)	   and	   age-­‐related	  macular	  degeneration,	  both	  of	  which	  feature	  excess	  complement	  activation	  [54].	  Similar	  to	  FH,	  Complement	  receptor	  1	  (CR1)	  and	  decay	  accelerating	  factor	  (DAF,	  also	  known	  as	  CD55)	   compete	   with	   Bb	   binding	   to	   C3b,	   and	   can	   cause	   Bb	   to	   dissociate	   from	   the	   C3bBb	   C3	  convertase	  complex	  [50]	  (Figure	  4).	  CD59,	  also	  referred	  to	  as	  protectin,	  inhibits	  the	  final	  step	  of	  the	  terminal	  pathway	  –	  the	  binding	  of	  C9	  to	  the	  C5b-­‐8	  complex.	  CR1,	  DAF,	  and	  CD59	  are	  found	  almost	   exclusively	   on	   host	   cells	   [55,	   56].	   A	   deficiency	   in	   CD59	   and	   DAF	   is	   associated	   with	  paroxysmal	   nocturnal	   hemoglobinuria	   (PNH),	   wherein	   red	   blood	   cells	   are	   susceptible	   to	  episodes	   of	   spontaneous	   lysis	   by	   complement	   [50].	   The	   diseases	   that	   arise	   from	   regulatory	  protein	   deficiencies	   emphasize	   the	   importance	   of	   regulating	   complement	   activation,	   both	  spatially	  and	  temporally.	  	  	  Figure	   5:	   PolyP	   suppresses	   complement	   activation.	   Hemolytic	   assays	   were	   performed	   in	   serum	  (conducted	  by	  Jovian	  Wat)	  and	  show	  suppression	  of	  the	  terminal	  pathway	  (a)	  by	  polyP≥1000	  (	  	  ),	  compared	  to	  monophosphate	  (	   	  ).	  The	  mechanism	  was	  partially	  elucidated	  by	  the	  discovery	  that	  suppression	  only	  occurs	   before	   the	   addition	   of	   C7	   when	   this	   assay	   is	   conducted	   in	   a	   purified	   protein	   system	   (b).	   This	  suggests	  an	  interaction	  of	  polyP	  with	  any	  one	  or	  all	  of	  C5b,6,	  C6,	  or	  C7.	  Figure	  adapted	  from	  Wat	  et	  al.	  2014,	  Blood	  [57].	  	  	  	   13	  1.3.2	   Role	  of	  Polyphosphate	  in	  Complement	  Coagulation	   and	   complement	   are	   coordinately	   activated	   to	   protect	   the	   organism	   from	   injury	  induced	   bleeding	   and	   infection.	   Although	   polyP	   reportedly	   interferes	   with	   complement	  activation	  in	  bacteria	  [58],	  its	  profound	  pro-­‐coagulant	  properties	  in	  mammalian	  systems	  led	  us	  to	   predict	   that	   polyP	   would	   similarly	   enhance	   complement	   activation.	   Surprisingly,	   however,	  studies	   from	  the	  Conway	   lab,	  published	   in	  Blood	   in	  2014	   [57],	   revealed	  that	  polyP	  suppresses	  complement	   activation.	   This	  was	   determined	  using	  well-­‐established	   in	   vitro	   hemolytic	   assays,	  measuring	  complement	  activation	  in	  human	  serum	  and	  then	  with	  purified	  proteins.	  The	  assays	  are	   based	   on	   the	   increased	   sensitivity	   of	   rabbit,	   chicken	   or	   sheep	   erythrocytes	   to	   human	  complement,	  and	  the	  ion-­‐dependence	  of	  the	  different	  pathways.	  By	  incubating	  the	  erythrocytes	  with	   human	   serum,	   complement	   activation	   proceeds,	   causing	   the	   red	   blood	   cells	   to	   lyse,	  releasing	  hemoglobin	  which	  is	  measurable	  with	  a	  spectrophotometer	  at	  an	  absorbance	  at	  405	  nm.	  	  As	   compared	   to	   monoP,	   polyP	   caused	   a	   concentration-­‐dependent	   reduction	   in	   total	  complement-­‐mediated	   hemolysis	   (Figure	   5).	   The	   effect	   was	   also	   polymer	   length-­‐dependent,	  with	   increasing	   lengths	   eliciting	   greater	   suppression.	   Further	   studies	   from	   the	   Conway	   lab	  showed	  that	  polyP	  suppresses	  complement	  most	  prominently	  via	  the	  terminal	  pathway,	  either	  at	   the	  C5b,6,	   C7,	   or	   C5b-­‐7	   step	  of	   the	   cascade	   [57]	   (Figure	   5).	   This	  was	  observed	  not	   only	   in	  serum,	   but	   also	   in	   a	   purified	   system,	   indicating	   that	   polyP	  must	   directly	   interact	  with	   one	   or	  more	   complement	   components.	  Notably,	   once	   the	  C5b-­‐7	   complex	  was	   formed,	   polyP	  had	  no	  dampening	  effect.	  These	   unique	   observations	   led	   to	   further	   investigations	   of	   the	   mechanisms	   by	   which	   polyP	  interferes	   with	   complement	   activation	   via	   the	   terminal	   pathway,	   allowing	   me	   to	   add	   new	  important	  knowledge	  and	  providing	  a	  foundation	  for	  this	  thesis.	  	  	   14	  The	  finding	  that	  this	  naturally	  occurring	  inorganic	  polymer,	  polyP,	  could	  suppress	  complement	  activation,	  also	  raised	  the	  question	  as	  to	  whether	  polyP	  could	  be	  used	  as	  a	  therapy.	  Based	  on	  work	   by	   other	   groups,	   intravenous	   administration	   of	   polyP	   is	   expected	   to	   be	   complicated	   by	  thrombosis	   (see	   above).	   Therefore,	   for	   validation	   of	   the	   therapeutic	   efficacy	   of	   polyP,	   we	  identified	   a	   complement-­‐mediated	   disease	   that	   was	   relatively	   protected	   from	   the	   systemic	  vasculature,	  i.e.,	  age-­‐related	  macular	  degeneration	  (AMD).	  1.4	   Age-­‐Related	  Macular	  Degeneration	  (AMD)	  AMD	  is	  a	  world-­‐wide	  leading	  cause	  of	  impaired	  vision	  or	  blindness	  for	  people	  over	  the	  age	  of	  50	  [59].	  With	   the	   steady	   increase	   in	   the	   aging	   population,	   the	   number	   of	   affected	   individuals	   is	  expected	  to	  continue	  to	  climb.	  AMD	  is	   therefore	  not	  only	  a	  challenge	  for	  the	  affected	  patient	  and	  his/her	  family,	  but	  also	  a	  major	  public	  health	  concern,	  as	  well	  as	  a	  huge	  economic	  burden.	  In	   Canada	   alone,	   AMD	   has	   an	   estimated	   $2.6	   billion	   negative	   impact	   on	   the	   gross	   domestic	  product	  [60].	  	  AMD	   affects	   the	   macula,	   the	   part	   of	   the	   eye	   that	   is	   responsible	   for	   detailed	   central	   vision.	  Individuals	  who	  suffer	   from	  AMD	  have	  trouble	  with	  common	  tasks,	   such	  as	   recognizing	   faces,	  reading	  small	  print	  and	  signs,	  and	  independently	  mobilizing	  (Figure	  6).	  They	  are	  more	  likely	  to	  suffer	  from	  depression,	  and	  have	  an	  overall	  reduced	  quality	  of	  life	  [60,	  61].	  	  Two	  forms	  of	  advanced	  AMD	  exist;	  the	  Dry	  or	  atrophic	  form	  accounts	  for	  ~90%	  of	  cases,	  while	  the	  Wet	  or	  exudative	  or	  vascular	  form	  accounts	  for	  the	  rest.	  However,	   it	   is	  the	  Wet	  form	  that	  results	   in	   ~90%	   of	   cases	   of	   blindness	   [62].	   	   Thus	   far,	   the	   only	   specific	   therapies	   available	   for	  AMD	  are	  those	  that	  target	  the	  Wet	  form,	  and	  these	  are	  not	  uniformly	  effective.	  Thus,	  there	  is	  an	  urgent	  need	  for	  new	  therapies,	  preferably	  those	  that	  combat	  both	  forms.	  This	  will	  require	  more	  knowledge	   of	   the	   underlying	  mechanisms	   of	   AMD.	  Although	   the	   pathogenesis	   of	   AMD	   is	   not	  entirely	   understood,	   oxidative	   stress	   and	   excess	   activation	   of	   the	   complement	   system	   are	  considered	  key	  contributors	  [63].	  Thus,	  AMD	  represents	  a	  highly	   localized	  disorder,	   in	  need	  of	  new	  therapies.	  	  We	  predicted	  that	  polyP	  might	  be	  such	  a	  therapy.	  	  	  	   15	  	  	  Figure	  6:	  Reduced	  visual	  acuity	   in	  AMD.	  The	  images	  depict	  scenes	  as	  might	  be	  viewed	  by	  an	  individual	  with	   normal	   vision	   (a)	   and	   by	   a	   person	   suffering	   from	  AMD	   (b).	   Loss	   of	   visual	   acuity	   (detailed	   central	  vision)	   is	  apparent.	   Images	  adapted	  from	  the	  National	  Eye	   Institute,	  “Facts	  About	  Age-­‐Related	  Macular	  Degeneration”	  [64].	  	  	  1.4.1	   Structure	  of	  the	  Retina	  and	  Choroid	  AMD	  primarily	   affects	   the	   retinal	   and	   choroid	   layers	  of	   the	  eye.	   The	   retina	   comprises	   several	  layers	   of	   neurons.	   Light	   enters	   the	   eye	   and	   the	   photons	   are	   absorbed	   by	   the	   photoreceptor	  cells,	   i.e.,	   rods	  and	  cones.	  The	  macula	  contains	  the	  densest	  population	  of	  photoreceptor	  cells.	  The	  absorbed	  photons	  are	  converted	  to	  an	  electrical	  signal	   that	   is	   transferred	  to	  bipolar	  cells.	  The	  signal	  is	  then	  transmitted	  to	  the	  ganglion	  cells,	  which	  relay	  an	  electrical	  impulse	  to	  the	  optic	  nerve	   [65].	   A	   single	   layer	   of	   melanin-­‐containing	   cells	   termed	   the	   “retinal	   pigment	   epithelial	  (RPE)”	   cells	   is	   positioned	  posterior	   to	   the	  photoreceptor	   cells	   (Figure	  7a).	   These	   cells	   support	  the	   photoreceptor	   cells	   by	   phagocytosing	   old	   photoreceptor	   disks,	   recycling	   photopigment	  molecules,	  and	  supplying	  nutrients	  [65,	  66].	  Adjacent	  and	  exterior	  to	  the	  RPE	  layer,	  is	  a	  barrier	  known	  as	  Bruch’s	  membrane.	  This	  contains	  the	  basement	  membrane	  of	  the	  RPE	  layer	  and	  the	  underlying	   choroid	   layer,	   as	  well	   as	   an	   outer	   collagenous	   zone,	   an	   elastic	   zone,	   and	   an	   inner	  collagenous	  zone	  [67].	  	  	  	   16	  Posterior	  to	  Bruch’s	  membrane	  is	  the	  choroid	  layer	  (Figure	  7a).	  The	  choroid	  consists	  of	  smooth	  muscle	   cells,	   secretory	   cells,	   neurons,	   and	   an	   abundance	   of	   blood	   vessels	   lined	   by	   choroidal	  endothelial	  cells	  (CEC).	  There	  are	  various	  functions	  of	  the	  choroid,	  including	  secretion	  of	  growth	  factors,	  adjusting	  the	  position	  of	  the	  retina,	  supplying	  the	  retina	  with	  nutrients	  and	  oxygen,	  and	  removing	  waste	  [68].	  Growth	  factors	  secreted	  from	  the	  choroid	  participate	  in	  altering	  the	  shape	  of	  the	  eye	  in	  pathologies	  such	  as	  hyperopia	  and	  myopia	  [69].	  Adjustment	  of	  retinal	  position	  is	  achieved	   by	   thickening	   and	   thinning	   of	   the	   choroid,	   and	   this	   in	   turn	   is	   accomplished	   by	   the	  smooth	  muscle	  cells,	  although	  the	  exact	  mechanisms	  are	  unknown	  [68].	  Of	  all	  its	  properties,	  the	  choroid	  is	  most	  important	  as	  the	  major	  vascular	  supply	  for	  the	  retina	  [68].	  Lined	   by	   choroidal	   endothelial	   cells,	   the	   choroidal	   blood	   vessels	   form	   a	   network	   of	  choriocapillaris.	  These	  vessels	  are	  fenestrated	  and	  20-­‐30	  μm	  in	  diameter,	  which	  is	  ~75%	  wider	  than	  the	  capillaries	  in	  the	  retina,	  thereby	  allowing	  for	  a	  slower	  rate	  of	  blood	  flow.	  Permeability	  to	   proteins	   [62],	   owing	   to	   the	   fenestrations	   and	   slow	   flow	   rate,	   facilitates	   the	   exchange	   of	  nutrients,	   oxygen,	   and	   wastes	   with	   the	   retina.	   The	   choriocapillaries	   also	   play	   a	   role	   in	  thermoregulation	  of	  the	  eye	  [68].	  	  	  	  	  	  	   17	  	  	  Figure	  7:	  Pictorial	  representation	  of	  the	  normal	  macula	  and	  pathophysiological	  changes	  in	  AMD.	  	  The	  normal	  macula	  consists	  of	  organized	   layers	  of	  photoreceptors,	  a	  monolayer	  of	  RPE	  cells,	  and	  a	  choroid	  vasculature	  that	  is	  separated	  from	  the	  retina	  by	  Bruch’s	  membrane	  (a).	  In	  Dry	  AMD,	  drusen	  accumulate	  in	  the	  sub-­‐retinal	  space,	  and	  this	  is	  associated	  with	  atrophy	  of	  the	  RPE	  and	  photoreceptor	  cells	  (b).	  Wet	  AMD	  is	  characterized	  by	  proliferation	  and	  migration	  of	  choroidal	  blood	  vessels	  into	  the	  retinal	  layers	  (c).	  The	  Dry	  form	  of	  AMD	  accounts	  for	  most	  cases.	  However,	  it	  is	  the	  Wet	  form	  that	  more	  commonly	  leads	  to	  blindness.	   The	  Wet	   form	  can	  develop	   from	   the	   intermediate	   stage	  or	  progress	   from	   the	  Dry	   form	   (d).	  Figures	  a-­‐c	  adapted	  from	  Brightfocus®	  Foundation.	  Illustration	  by	  Bob	  Morreale	  [70].	  	  	  	  	   18	  1.4.2	   	  Pathogenesis	  of	  AMD	  AMD	  is	  classified	  into	  early,	  intermediate,	  and	  late	  stages,	  based	  on	  specific	  abnormalities	  and	  by	  the	  associated	  symptoms.	  In	  the	  early	  stage,	  molecular	  byproducts	  deposit	  at	  the	  interface	  of	  the	   RPE	   layer	   and	   Bruch’s	   membrane.	   These	   are	   termed	   drusen,	   and	   consist	   of	   lipids	   and	  proteins	   which	   are	   largely	   immune-­‐related,	   and	   include	   inflammatory	   proteins,	  immunoglobulins,	   and	   several	   complement	   components	   and	   complexes	   [71].	   Also	   present	   is	  apolipoprotein	  E,	  RPE	  cell	  debris,	   lipofuscin,	  and	  melanin	  which	  are	  products	  of	  RPE	  cell	   injury	  [72,	  73].	  Drusen	  that	  form	  in	  the	  early	  stage	  of	  AMD	  are	  small-­‐	  to	  medium-­‐sized	  (<63-­‐124	  μm	  in	  diameter),	   well-­‐marginalized	   deposits	   known	   as	   “hard	   drusen”.	   At	   this	   stage,	   vision	   may	   be	  normal	   or	   only	  minimally	   affected,	  with	  mild	   blurriness,	   reduced	   contrast	   perception,	   and/or	  difficulty	   adjusting	   to	   brightness	   and	   darkness	   [74].	   The	   presence	   of	   only	   a	   few,	   small,	   hard	  drusen	  is	  considered	  a	  normal	  part	  of	  aging,	  and	  is	  not	  diagnostic	  of	  AMD	  [74].	  	  The	   intermediate	   stage	   is	   distinguished	   by	   the	   presence	   of	   at	   least	   one	   larger,	   more	   diffuse	  druse	  termed	  “soft	  drusen”,	  and	  an	  increase	  in	  the	  number	  of	  medium-­‐sized	  drusen	  [75].	  Vision	  loss	  is	  more	  prominent,	  and	  may	  be	  associated	  with	  a	  decreased	  ability	  to	  stabilize	  a	  focal	  image	  [65].	  	  The	  intermediate	  stage	  can	  lead	  to	  the	  advanced	  stage	  of	  AMD,	  of	  which	  there	  are	  two	  forms.	  Dry	   AMD	   is	   characterized	   by	   atrophy	   of	   the	   retina	   in	   the	   macular	   area,	   also	   known	   as	  geographical	   atrophy	   (Figure	  7b).	   This	   is	   linked	   to	   the	  enlargement	   and	   confluence	  of	   drusen	  and	   consequent	   thickening	   of	   Bruch’s	  membrane,	   blocking	   the	   flow	   of	   oxygen	   and	   nutrients	  from	   the	   choroid	   to	   the	   retina,	   as	   well	   as	   the	   flow	   of	   waste	   from	   the	   retina	   to	   the	  choriocapillaries	   [76,	   77].	   The	  RPE	   cells	   are	   the	   first	   to	  be	  affected,	   and	   their	   loss	  of	   function	  results	  in	  secondary	  degeneration	  of	  photoreceptor	  cells	  [78].	  Loss	  of	  visual	  function	  associated	  with	  Dry	  AMD	  occurs	  very	  gradually	  with	  symptoms	  that	  can	  progress	  in	  a	  span	  of	  years	  before	  severe	   visual	   impairment	   manifests.	   There	   are	   no	   treatments	   for	   this	   form,	   although	   a	   few	  therapies	  are	  currently	  undergoing	  clinical	  trial	  [79,	  80].	  	  	   19	  The	  other	  10%	  of	  advanced	  AMD	  cases	  are	  classified	  as	  Wet	  AMD.	  The	  name	  derives	  from	  the	  hallmark	  choroidal	  neovascularization	  (CNV)	  that	  occurs,	  as	  Bruch’s	  membrane	  is	  disrupted	  [81]	  (Figure	  7c).	  CNV	   is	  believed	   to	  be	   the	   result	  of	   chronic	   inflammation	   caused	  by	  drusen	  and	  a	  damaged	   RPE	   layer,	   although	   it	   is	   likely	   that	   the	   CEC	   also	   participate	   in	   a	   major	   way	   [81].	  Chronic	   inflammation	   causes	   an	   influx	   of	   macrophages,	   dendritic	   cells,	   and	   iNKT	   cells.	   In	  addition,	  normal	  aging	  of	  the	  retina	  and	  choroid	  alters	  the	  function	  of	  the	  retinal	  and	  choroidal	  cells,	   such	   that	   they	   become	   activated,	   and	   express	   more	   complement	   proteins	   as	   well	   as	  chemoattractants	  [82].	  The	  consequence	  is	  an	  influx	  of	  immune	  cells,	  which	  amplify	  and	  sustain	  an	   inflammatory	   response.	   This	   is	   associated	  with	   increased	  expression	  of	   angiogenic	   factors,	  most	   notably,	   vascular	   endothelial	   growth	   factor	   (VEGF).	   Blood	   vessels	   that	   form,	   emanating	  primarily	   from	  the	  choroid,	   initially	  start	  as	  capillaries,	  and	  may	  progress	   to	  become	   irregular,	  large	  and	   leaky	  blood	  vessels	   [83].	  The	  danger	   in	  CNV	  is	  that	  the	  vessels	  will	   leak	  their	  serous	  contents	  into	  the	  retina	  [84].	  This	  damages	  the	  RPE	  cells	  and	  photoreceptors,	  causing	  them	  to	  become	  dysfunctional,	  and	   results	   in	   loss	  of	  vision	   [85,	  86].	  Wet	  AMD	  may	  progress	   from	  the	  Dry	  form,	  or	  arise	  directly	  from	  intermediate	  AMD	  (Figure	  7d)	  [87].	  1.4.3	   Role	  of	  Complement	  in	  AMD	  Several	   lines	   of	   evidence	   support	   the	   notion	   that	   heightened	   activation	   of	   complement	  contributes	  to	  the	  pathogenesis	  and	  progression	  of	  AMD.	  Most	  notably,	  there	  are	  mutations	  in	  several	   genes	   that	   encode	   complement	   components	   that	   confer	   a	   higher	   risk	   of	   the	  development	  of	  AMD.	  The	  strongest	  association	  is	  with	  complement	  factor	  H	  (FH)	  [88,	  89].	  In	  a	  genome-­‐wide	  association	  study	  (GWAS),	  a	  single	  nucleotide	  polymorphism	  at	  amino	  acid	  402	  of	  FH	  was	   identified	   that	   increases	   the	   likelihood	  of	  AMD	  development	  by	  7.4-­‐fold	   [88,	  89].	  This	  Y402H	  (tyrosine	  replaced	  with	  histidine	  at	  amino	  acid	  residue	  402)	  mutation	  occurs	  at	  the	  site	  at	  which	   FH	   binds	   to	   C-­‐reactive	   protein	   (CRP)	   and	   glycosaminoglycans	   (e.g.,	   heparin)	   [52].	   The	  latter	   localizes	  FH	  to	  the	  surface	  of	   the	  host	  cell	  which,	  as	  previously	  discussed,	  enhances	  the	  capacity	  of	  FH	  to	  act	  as	  a	  cofactor	   for	  FI-­‐mediated	   inactivation	  of	  C3b	  on	  host	  cells.	  CRP	   is	  an	  inflammatory	  protein	  that	   is	  deposited	   in	  high	  amounts	   in	  the	  eyes	  of	  patients	  with	  AMD	  [90,	  91],	   [92].	   In	   patients	   with	   this	   risk	   variant	   of	   FH,	   complement	   component	   deposition	   on	   the	  	  	   20	  choroid	   vasculature	   was	   significantly	   increased,	   and	   associated	   with	   increased	   leakage	   of	  plasma	  contents	  into	  Bruch’s	  membrane	  [89].	  Although	  not	  definitively	  shown,	  the	  data	  suggest	  that	   the	   mutant	   form	   of	   FH	   is	   responsible	   for	   excess	   complement	   activation	   that,	   in	   these	  patients,	  leads	  to	  AMD.	  There	  are	  other	  polymorphisms	  in	  genes	  for	  complement	  components	  that	  have	  also	  been	  linked	  to	  AMD	  [93].	  Some	  of	  these	  are	  described	  below.	  A	  rare	  missense	  mutation	  of	  complement	  Factor	  I	  (FI)	  that	  has	  been	  implicated	  in	  increasing	  the	  risk	  of	  AMD	  [94,	  95]	  results	  in	  the	  substitution	  of	  glycine	  to	  arginine	  (G119R)	  [95].	  This	  variant	  form	  of	  FI	  is	  able	  to	  cleave	  and	  inactivate	  C3b,	  but	  to	  a	  lesser	  extent	  than	  the	  wild-­‐type	  protein,	  resulting	  in	  less	  complement	  inhibition	  [95].	  	  Common	  variants	  of	  C3	  have	  also	  been	  linked	  to	  AMD.	  A	  substitution	  of	  lysine	  by	  glutamine	  at	  amino	  acid	  155	  (K155Q)	  results	  in	  a	  C3/C3b	  that	  is	  relatively	  resistant	  to	  cleavage	  by	  FI/FH	  [94].	  In	  a	  model	  of	   laser	   induced	  CNV	  for	  Wet	  AMD,	  mice	  with	  heterozygous	  defects	  of	  C3	  develop	  smaller	  CNV	   lesions	   [96],	   supporting	   the	  notion	   that	   complement	  plays	  an	   important	   role.	  C9	  has	  also	  been	  implicated;	  a	  substitution	  of	  proline	  for	  serine	  at	  amino	  acid	  167	  (P167S)	  of	  C9	  is	  linked	   to	   an	   increased	   risk	   of	   AMD,	  while	   a	   nonsense	  mutation	   of	   arginine	   at	   amino	   acid	   95	  (R95X)	  may	  confer	  protection	  against	  AMD	  in	  Japanese	  carriers	  [94],	  [96].	  	  The	  R95X	  C9	  variant	  is	  one	  of	  the	  few	  complement	  component	  mutations	  that	  are	  associated	  with	  protection	  (also	  known	   as	   an	   ‘inverse	   association’)	   against	   AMD	  development.	   A	   Factor	   B	   polymorphism	   that	  results	  in	  a	  substitution	  of	  arginine	  for	  glutamine	  at	  amino	  acid	  32	  (R32Q)	  is	  also	  associated	  with	  decreased	   susceptibility	   to	   AMD,	   based	   on	   2	   separate	   cohort	   studies	   [97,	   98].	   In	   addition,	   a	  Factor	  B	  mutation	  that	  substitutes	   leucine	  for	  histidine	  at	  amino	  acid	  9	  (L9H)	  confers	  a	  similar	  protective	  effect	  [98].	  In	  the	  same	  studies,	  the	  C2	  polymorphism	  that	  substitutes	  glutamic	  acid	  for	  an	  aspartic	  acid	  at	  amino	  acid	  318	   (E318D)	  was	  also	   found	   to	  have	  an	   inverse	  association	  with	  AMD	  [45,	  98].	  	  The	   link	   between	   complement	   and	   AMD	   is	   further	   strengthened	   by	   evidence	   of	   enhanced	  activation	  of	  complement	  –	  particularly	   the	  alternative	  pathway	   -­‐	   in	   the	  eyes	  of	  patients	  with	  AMD	  [99,	  100].	  This	  results	  in	  the	  recruitment	  of	  inflammatory	  cells	  via	  augmented	  local	  release	  	  	   21	  of	  inflammatory	  mediators,	  with	  subsequent	  amplification	  of	  complement	  activation	  [98].	  This	  cycle	   is	   responsible	   for	   chronic,	   local	   inflammation	   that	   is	   a	   feature	   of	   AMD.	   Deposits	   of	  complement	   components	   and	   inflammatory	  mediators	   are	  often	   found	   in	  AMD	  affected	  eyes	  [88].	  Within	   drusen,	   C5	   and	   C5b-­‐9	   are	   abundant	   [48,	   72,	   101].	   C3a,	   C3d,	   all	   of	   the	   terminal	  pathway	  components,	  complement	  downregulators	  (clusterin	  and	  vitronectin),	  apolipoproteins	  (apoA1,	   apoA4,	   and	   apoE),	   serum	   amyloid	   A	   (SAP-­‐A),	   serum	   amyloid	   P	   (SAP-­‐P),	   and	  thrombospondin	   have	   also	   been	   detected	   in	   drusen	   [88,	   102,	   103].	   There	   is	   considerable	  controversy	  as	  to	  the	  role	  of	  C3a	  and	  C5a	  in	  the	  damage	  associated	  with	  AMD.	  C3a	  and	  C5a	  in	  drusen	  may	  induce	  expression	  of	  VEGF,	  leukocyte	  recruitment,	  and	  CNV	  [103].	  However,	  there	  is	  evidence	   that	  critical	  amounts	  of	  C3a	  and	  C5a	  are	   required	   for	   survival	  and	   integrity	  of	   the	  RPE	   cells	   [104].	   The	   latter	   may	   explain	   why	   the	   complement	   inhibitor,	   eculizumab	   (which	  interferes	  with	  the	  generation	  of	  both	  the	  MAC	  and	  C5a)	  was	  not	  effective	  in	  reducing	  drusen	  volume	  in	  clinical	  trials	  [105].	  1.4.4	   Role	  of	  the	  MAC	  in	  AMD	  Formation	   of	   the	   MAC	   is	   believed	   by	   many	   to	   be	   a	   key	   causative	   factor	   in	   AMD,	   directly	  damaging	  the	  RPE	  [106].	  MAC	  deposition	  has	  been	  correlated	  with	  RPE	  loss	  and	  AMD	  severity	  [107].	  As	  previously	  discussed,	  formation	  of	  a	  stable	  and	  functional	  MAC	  is	  negatively	  regulated	  via	  the	  terminal	  pathway	  by	  CD59,	  clusterin	  and	  vitronectin,	  each	  of	  which	  functions	  by	  binding	  to	   terminal	   pathway	   components	   and	   interfering	   with	   MAC	   assembly.	   CD59	   is	   a	  glycosylphosphatidylinositol-­‐linked	   protein	   that	   prevents	   C9	   molecules	   from	   polymerizing.	  Although	  no	  genetic	   studies	  have	   linked	  CD59	   to	  AMD,	  drusen	  components	   suppress	  CD59	   in	  RPE	  [108].	  Both	  clusterin	  and	  vitronectin	  are	  found	  in	  drusen,	  although	  their	  roles	  there	  are	  not	  known	  [109].	  Identification	  of	  polyP	  as	  a	  negative	  regulator	  of	  the	  terminal	  pathway	  is	  rationale	  to	  explore	  whether	  it	  has	  protective	  properties	  against	  AMD.	  	  	   22	  1.4.5	   Role	  of	  Oxidative	  Stress	  in	  AMD	  The	   highly	   vascular	   retina	   is	   continuously	   exposed	   to	   photons	   from	   UV	   light	   and	   it	   is	   thus	  susceptible	   to	   chronic	   oxidative	   stress	   [110].	   Oxidative	   stress	   is	   defined	   by	   an	   imbalance	  between	  reactive	  oxygen	  species	  (ROS)	  production	  and	  the	  body’s	  ability	  to	  detoxify	  these	  ROS.	  Generation	  of	  ROS	  is	  a	  physiologically	  normal	  process,	  formed	  naturally	  during	  the	  metabolism	  of	  oxygen.	  ROS	  play	  important	  roles	  in	  cellular	  function,	  but	  in	  excess,	  cause	  significant	  damage	  that	  may	  result	  in	  cell	  death,	  and	  a	  pro-­‐inflammatory	  response,	  with	  recruitment	  of	  innate	  and	  adaptive	  immune	  responses.	  Several	  natural	  defense	  mechanisms	  therefore	  exist	  to	  neutralize	  ROS.	  H2O2-­‐producing	  enzymes	  are	  compartmentalized	   in	   specialized	   lysosomes,	   referred	   to	  as	  peroxisomes	  [111].	  These	  also	  contain	  catalase,	  an	  enzyme	  that	  degrades	  H2O2	  into	  water	  and	  oxygen	   [111].	   Glutathione	   peroxidase	   and	   superoxide	   dismutase	   are	   other	   important	  antioxidant	   enzymes.	   From	   the	   perspective	   of	   AMD,	   the	   most	   prominent	   antioxidant	   in	   the	  human	  retina	  is	  believed	  to	  be	  catalase	  [112],	  and	  patients	  with	  AMD	  reportedly	  have	  reduced	  catalase	  activity	  in	  their	  retinas	  [113].	  The	  major	   risk	   factor	   for	  AMD	   is	   advanced	   age,	   and	   this	   is	   also	   associated	  with	   a	   diminished	  capacity	  to	  defend	  against	  the	  damaging	  effects	  of	  ROS.	  This	  is	  further	  exacerbated	  by	  exposure	  to	  smoking	  and/or	  second-­‐hand	  smoke,	  which	  increases	  the	  risk	  of	  AMD	  by	  two-­‐	  to	  three-­‐fold	  [114].	  Studies	  suggest	  that	  oxidative	  stresses	  in	  the	  eye	  modify	  otherwise	  normal	  proteins	  and	  lipids	  in	  the	  retina,	  and	  these	  initiate	  the	  formation	  of	  drusen	  [113].	  Based	  on	  large	  scale	  clinical	  studies,	   oral	   antioxidant	   supplements	   are	   therefore	   recommended	   for	   patients	   with	  early/intermediate	  AMD	  to	  prevent	  progression	  to	  advanced	  AMD	  [115-­‐117].	  The	  data	  are	  not	  strong,	  but	  the	  risk-­‐benefit	  profile	  favours	  taking	  these	  medications	  rather	  than	  not.	  1.4.6	   Complement	  and	  Oxidative	  Stress	  Chronic	  excess	  complement	  activation	  and	  oxidative	  stress	  cooperate	  to	  augment	  the	  damage	  associated	  with	  AMD	  [88,	  118,	  119],	  and	  not	  surprisingly,	  most	  often	  coexist.	  This	  likely	  reflects	  the	   fact	   that	   AMD	   occurs	   in	   older	   people	   who	   acquire,	   through	   exposure	   to	   environmental,	  	  	   23	  genetic	  and/or	  epigenetic	  factors,	  disruptions	  in	  the	  mechanisms	  that	  protect	  against	  host	  cell	  damage.	  	  Moreover,	  there	  are	  positive	  feedback	  loops	  wherein	  activation	  of	  complement	  leads	  to	  increased	  generation	  of	  inflammatory	  mediators,	  including	  C3a	  and	  C5a	  [120],	  as	  well	  as	  ROS,	  which	  in	  turn	  further	  augment	  the	  inflammatory	  response,	  [98,	  120],	  with	  further	  activation	  of	  complement.	  In	  AMD,	  these	  escalating,	  damaging	  pathways	  escape	  natural	  negative	  regulatory	  mechanisms,	   resulting	   in	   cellular	  damage,	  decreasing	   the	   capacity	  of	   the	  RPE	   to	  maintain	   the	  health	  of	  photoreceptors	   [88].	   It	   is	   reasonable	  to	  consider	  that	  the	  most	  effective	  therapeutic	  strategies	  might	   include	   simultaneous	   suppression	  of	   the	  generation	  of	  ROS	  and	  complement	  activation.	  1.4.7	   Current	  Therapies	  for	  AMD	  Current	   treatments	   for	  AMD	  are	  suboptimal	  and	  costly	   for	   the	  Wet	   form,	  and	  entirely	   lacking	  for	  the	  Dry	  form	  [121-­‐126].	  Treatments	  for	  the	  Wet	  or	  vascular/exudative	  form	  are	  principally	  designed	   to	   suppress	   vessel	   growth	   by	   interfering	   with	   vascular	   endothelial	   growth	   factor	  (VEGF)-­‐related	  pathways.	  For	  example,	  monoclonal	  antibodies	  that	  neutralize	  VEGF	  are	  widely	  used.	  Pegaptanib	  is	  an	  aptamer	  that	  targets	  a	  soluble	  form	  of	  VEGF	  and	  has	  been	  shown	  to	  have	  efficacy	   over	   a	   2	   year	   period	   [127].	   Bevacuzimab	   and	   ranibizumab	   are	   both	   humanized	  antibodies	  against	  VEGF	  that	  are	  able	  to	  inhibit	  all	  isoforms	  of	  VEGF,	  with	  ranibizumab	  having	  a	  3-­‐6	   fold	   higher	   affinity	   [128-­‐130].	   VEGF	   pathway	   inhibitors	   are	   injected	   into	   the	   eye	  (intravitreous)	  ~once	  per	  month,	  and	  overall	  result	  in	  improved	  vision	  in	  ~30%	  of	  patients,	  and	  arrest	   progression	   in	   ~90%	   [131-­‐134].	   However,	   up	   to	   20%	   of	   patients	   experience	   mild-­‐to-­‐serious	   side	   effects	   [135],	   and	   long-­‐term	   use	   may	   be	   complicated	   by	   development	   of	  widespread	  atrophy	  of	  the	  RPE	  cells	  [136].	  There	  are	  also	  concerns	  about	  potential	  detrimental	  effects	  on	  the	  retinal	  vasculature	  [137].	  Complement	  pathway	  inhibitors	  have	  been	  evaluated	  in	  preclinical	  and	  early	  clinical	  trials,	  mostly	  for	  Wet	  AMD,	  and	  these	  have	  targeted	  factor	  B,	  factor	  D,	   C3	   and	  C5	   (eculizumab).	   Limited	   information	   is	   available	  on	   the	  progress	  of	  most	  of	   these	  trials,	  although	  eculizumab	  failed	  to	  reduce	  drusen	  size	  in	  Dry	  AMD.	  We	  believe	  that	  polyP	  holds	  greater	  therapeutic	  promise,	  as	  it	  dampens	  generation	  of	  the	  MAC,	  while	  sparing	  generation	  of	  C3a	   and	   C5a,	   both	   of	   which	   may	   be	   protective	   to	   retinal	   health.	   Moreover,	   at	   least	   in	  	  	   24	  prokaryotes,	   polyP	   has	   anti-­‐oxidant	   properties.	   This	   raises	   the	   possibility	   that	   polyP	  might	   be	  capable	   of	   simultaneously	   suppressing	   the	   two	  major	   pathways	   that	   promote	   AMD,	   a	   strong	  rationale	  for	  further	  study.	  It	  is	  estimated	  that	  the	  prevalence	  of	  AMD	  will	  exceed	  150	  million	  people	  worldwide	  by	  the	  year	  2020	  [138].	  Although	  progress	  has	  been	  made	  in	  developing	  effective	  treatments	  for	  Wet	  AMD,	  there	   remains	   an	   urgent	   need	   for	   the	   development	   of	   alternative	   and	   more	   effective	  therapeutic	  strategies	  for	  both	  Wet	  and	  Dry	  AMD.	  	  	  	  	  	  	  	  	  	  	  	  	   25	  Chapter	  2: Hypothesis	  Based	   on	   our	   findings	   that	   polyP	   dampens	   activation	   of	   the	   complement	   system,	   we	  hypothesized	  that	  polyP	  will	  confer	  protection	  against	  AMD-­‐associated	  stresses.	  	  	  	   26	  Chapter	  3: Overall	  Goal	  and	  Objectives	  3.1 Overall	  Goal	  To	   increase	  our	  understanding	  of	   the	  mechanisms	  by	  which	  polyP	   regulates	   complement	  and	  cellular	   function	   and	   to	   apply	   this	   new	   knowledge	   in	   a	   rodent	  model	   of	   age-­‐related	  macular	  degeneration	  (AMD).	  3.2 Specific	  Objectives	  1. To	  uncover	  the	  mechanims	  by	  which	  polyP	  regulates	  complement	  a. Hemolytic	  assays	  to	  verify	  the	  effect	  of	  polyP	  on	  complement	  b. Determine	  the	  biochemical	  interaction	  of	  polyP	  with	  complement	  components	  2. To	  evaluate	  the	  therapeutic	  efficacy	  of	  polyP	  in	  a	  model	  of	  AMD	  a. In	  vivo,	  with	  a	  laser-­‐induced	  choroidal	  neovascularization	  rodent	  model	  in	  i. Rats	  ii. Mice	  	  b. In	  vitro,	  using	  retinal	  pigmented	  epithelial	  and	  choroidal	  endothelial	  cell	  lines,	  in	  response	  to:	  i. Complement-­‐mediated	  stress	  • Measure	  formation	  of	  complement	  final	  effectors	  ii. Oxidative	  stress	  • Elucidate	  possible	  mechanisms	  by	  which	  polyP	  may	  confer	  protection	  o Assess	  anti-­‐oxidant	  enzyme	  activity	  and	  expression	  in	  response	  to	  polyP	  o Test	  the	  effects	  of	  polyP	  on	  cell-­‐junction	  integrity	  in	  response	  to	  oxidative	  stress	  o Test	  the	  effect	  of	  polyP	  on	  cell-­‐proliferation	  	   27	  Chapter	  4: Materials	  and	  Methods	  4.1 Reagents	  Synthesized	  polyP≥1000	  was	  a	  generous	  gift	  from	  Dr.	  James	  H.	  Morrissey	  (Urbana,	  Illinois,	  USA).	  Synthesized	   polyP130	   was	   from	   Regenetiss	   Inc.	   (Tokyo,	   Japan).	   All	   complement	   proteins	   and	  human	   sera	   were	   purchased	   from	   Complement	   Technology,	   Inc.	   (Tyler,	   Texas,	   USA).	   Unless	  otherwise	  stated,	  antibodies	  were	  purchased	  from	  Life	  Technologies	  (Carlsbad,	  California,	  USA).	  	  4.2 Cell	  Culture	  Cell	   culture	  media	   and	   reagents	  were	   purchased	   from	   Life	   Technologies	   (Carlsbad,	   California,	  USA).	  RF/6A	  cells,	  a	  choroidal	  endothelial	  cell	  line	  derived	  from	  monkeys,	  and	  ARPE-­‐19,	  a	  retinal	  pigment	  epithelial	  cell	  line	  derived	  from	  humans,	  were	  purchased	  from	  American	  Type	  Culture	  Collection	  (Manassas,	  Virginia,	  USA).	  EA.hy	  926	  cells,	  a	  hybridized	  human	  umbilical	  vein	  cell	  line,	  was	  a	  gift	  of	  Dr.	  Cora-­‐Jean	  C.	  Edgel	  (University	  of	  North	  Carolina,	  USA).	  	  EA.hy	   926,	   RF/6A,	   and	   ARPE-­‐19	   cells	   were	   cultured	   in	   Dulbecco’s	   Modified	   Eagle’s	   Media,	  Modified	   Eagle’s	  Media,	   and	  Dulbecco’s	  Modified	   Eagle’s	  Media-­‐F12	   (Sigma-­‐Aldrich	   Corp.,	   St.	  Louis,	   Missouri,	   USA),	   respectively,	   all	   supplemented	   with	   10%	   heat-­‐inactivated	   fetal	   bovine	  serum	  (HI-­‐FBS),	  L-­‐glutamine	  (2	  mM),	  Na	  pyruvate	  (1	  mM),	  and	  penicillin	  and	  streptomycin	  (50	  U/mL).	  From	  frozen	  aliquots,	  cells	  were	  thawed	  for	  2	  minutes	  in	  a	  37°C	  water	  bath	  and	  directly	  added	  to	  polystyrene	  cell-­‐culture	  dishes	  (Corning	  Inc.,	  Corning,	  NY,	  USA)	  at	  1:10	  dilution	  in	  their	  growth	  media.	  They	  were	  allowed	  to	  adhere	  overnight	  in	  a	  humidified	  37°C	  incubator	  with	  5%	  CO2,	  after	  which	  the	  media	  was	  changed	  to	   fresh	  growth	  media.	  The	  cells	  were	  subsequently	  passaged	  every	  3-­‐5	  days.	  Passage	  number	  did	  not	  exceed	  25,	   after	  which	  a	   fresh	  aliquot	  was	  used.	  For	  propagation,	  cells	  at	   low	  passage	  were	  frozen	  with	  50%	  cryoprotective	  media	  (Basal	  Eagle’s	  Medium	  with	  Hank’s	  BSS	  and	  15%	  DMSO	  (Lonza,	  Walkersville,	  Maryland,	  USA))	  and	  50%	  HI-­‐FBS	  and	  stored	  in	  liquid	  nitrogen.	  	  	  	  	  	   28	  4.3 Terminal	  Pathway	  Hemolytic	  Assay	  Chicken	   red	   blood	   cells	   (cRBCs)	   were	   isolated	   from	   chicken	   whole	   blood	   (Colorado	   Serum	  Company,	  Denver,	   CO)	   by	   centrifugation	   at	   1000	   x	   g,	   and	   subsequently	  washed	   3	   times	  with	  gelatin	  veronal	  buffer	  (GVB)	  (Sigma-­‐Aldrich	  Corp.)	  with	  10	  mM	  EDTA.	  The	  EDTA	  chelates	  cations	  to	   prevent	   activation	   of	   upstream	   complement	   pathways,	   which	   are	   ion-­‐dependent.	   Prior	   to	  activation	   of	   the	   terminal	   pathway,	   varying	   concentrations	   of	   polyP	   are	   added	   to	   3.33	   x	   107	  cells/mL	   of	   cRBCs	   in	   a	   96-­‐well	   plate.	   The	   terminal	   pathway	  was	   activated	   by	   the	   addition	   of	  purified	  C5b,6	  (the	  concentration	  to	  achieve	  ~80%	  hemolysis	  was	  determined	  on	  the	  same	  day)	  to	  3.33	  x	  107	  cells/mL	  of	  cRBCs,	   followed	  by	   the	  addition	  of	  normal	  human	  serum	  at	  2%	   final	  concentration	  in	  a	  final	  volume	  of	  300	  µL.	  All	  dilutions	  were	  conducted	  in	  GVB-­‐EDTA.	  Reaction	  mixtures	   were	   incubated	   at	   37°C	   for	   30	   minutes,	   causing	   hemolysis	   of	   the	   cRBCs	   and	  consequent	   release	  of	  hemoglobin	  due	   to	   complement-­‐mediated	   lysis.	   The	   intact	   cRBCs	  were	  pelleted	  by	  centrifugation	  at	  600	  x	  g	  for	  3	  minutes.	  100	  µL	  of	  the	  supernatant	  was	  transferred	  to	  a	  new	  96-­‐well	  plate,	  and	  diluted	  1:1	  with	  dH2O.	  The	  absorbance	  of	  the	  supernatant	  samples	  at	  405	  nm	  was	  determined	  with	  the	  Spectramax	  384	  Plus	  Microplate	  Reader	  (Molecular	  Devices,	  Sunnyvale,	   CA,	   USA).	   The	   absorbance	   of	   the	   sample	   reflects	   the	   amount	   of	   hemolysis	   that	  occurred	  due	  to	  complement-­‐mediated	  lysis.	  	  Similarly,	   hemolytic	   assays	   using	   purified	   terminal	   pathway	   complement	   proteins	   were	  conducted,	   eliminating	   the	   need	   for	   normal	   human	   serum	   as	   well	   as	   EDTA,	   as	   the	   terminal	  pathway	   is	   ion-­‐independent.	   In	   these	   purified	   systems,	   purified	   C5b,6	   (the	   concentration	   to	  achieve	   ~80%	   hemolysis)	   is	   combined	   with	   a	   mixture	   of	   purified	   C7,	   C8,	   and	   C9	   (final	  concentrations	  of	  15	  nM,	  10	  nM,	  and	  25	  nM,	  respectively).	  4.4 Platelet	  Releasate	  Hemolytic	  Assay	  Using	  a	  16-­‐gauge	  needle,	  blood	  was	  drawn	  from	  a	  human	  volunteer	  (with	  Ethics	  Approval	  from	  UBC)	   using	   the	   straight-­‐drip	   technique	   into	   4.5	   mL	   citrated	   Vacutainer®	   tubes	   (BD,	   Franklin	  Lakes,	   New	   Jersey,	   USA)	   containing	   2	   µM	   prostaglandin	   E	   (PGE).	   A	   total	   of	   12	   tubes	   were	  	  	   29	  collected,	   with	   the	   first	   tube	   being	   discarded	   to	   avoid	   using	   blood	   in	   which	   the	   coagulation	  and/or	   complement	   pathways	   were	   activated	   through	   the	   trauma	   of	   the	   needle	   insertion.	  Whole	   blood	   obtained	   was	   centrifuged	   at	   100	   x	   g	   for	   10	   minutes.	   Platelet	   rich	   plasma	   was	  collected	  and	  transferred	  to	  1	  mL	  microcentrifuge	  tubes.	  These	  were	  centrifuged	  at	  200	  x	  g	  to	  obtain	   a	   platelet	   pellet,	  which	  was	   then	   re-­‐suspended	   in	   1	  mL	   CGS	   buffer	   (10	  mM	   trisodium	  citrate,	  30	  mM	  dextrose,	  1.2	  mM	  sodium	  chloride	  at	  pH	  6.5).	  The	  platelets	  were	  centrifuged	  at	  200	  x	  g	  and	  re-­‐suspended	  in	  CGS	  buffer	  twice	  more,	  followed	  by	  another	  centrifugation	  at	  200	  x	  g.	  One	  sample	  was	  re-­‐suspended	  in	  200	  µL	  CGS	  buffer,	  while	  another	  sample	  was	  re-­‐suspended	  in	  10	  nM	  thrombin	  diluted	  in	  CGS	  buffer	  to	  activate	  the	  platelets.	  The	  “activated”	  samples	  were	  mixed	  on	  a	  rocker	  for	  15	  minutes,	  after	  which	  platelet	  aggregates	  formed.	  The	  aggregates	  were	  centrifuged	   at	   200	   x	   g	   for	   10	  minutes	   and	   the	   supernatant	   containing	   the	   platelet	   releasate	  (which	  would	  also	  contain	  polyP	  released	  from	  the	  dense	  granules)	  was	  transferred	  to	  a	  clean	  tube.	   A	   terminal	   pathway	   hemolytic	   assay	  was	   then	   performed	  with	   the	   supernatants	   of	   the	  non-­‐activated	  and	  activated	  samples.	  4.5 Stability	  of	  PolyP≥1000	  in	  Serum	  The	  terminal	  pathway	  hemolytic	  assay	  was	  modified	  to	  test	  the	  stability	  of	  polyP≥1000	  in	  terms	  of	  its	   ability	   to	   dampen	   complement	   activation.	   100%	   normal	   human	   serum	   and	   200	   µM	   final	  concentration	  of	  polyP≥1000	  were	  co-­‐incubated	  at	  37°C	  for	  varying	  periods	  of	  time,	  ranging	  from	  30	  minutes	  to	  10	  days.	  The	  serum	  and	  polyP≥1000	  mixture	  was	  then	  added	  simultaneously	  to	  the	  hemolytic	   assay,	   to	   a	   final	   concentration	  of	   2%	   serum	  and	  300	  µM	  polyP≥1000.	   These	   samples	  were	  compared	  to	  serum	  without	  polyP≥1000	  but	  incubated	  at	  the	  same	  time	  points.	  	  4.6 Gel	  Filtration	  C5b,6	  was	  stored	  in	  a	  buffer	  consisting	  of	  10	  mM	  4-­‐(2-­‐hydroxyethyl)-­‐1-­‐piperazineethanesulfonic	  acid	  (HEPES),	  120	  mM	  NaCl,	  pH	  7.2	  (HBS).	  C7	  was	  stored	  in	  10	  mM	  Na3PO4,	  145mM	  NaCl,	  pH	  7.3	  (PBS).	   The	   buffers	   were	   filtered	   with	   a	   0.22	   μm	   pore	   diameter	   filter	   (Stericup	   Millipore	  Express™PLUS),	  under	  a	  Class	  2A	  Biosafety	  Cabinet.	  Proteins	  were	  thawed	  in	  a	  37ᵒC	  water	  bath	  	  	   30	  for	   10	   minutes.	   To	   exclude	   protein	   aggregates	   from	   being	   loaded	   onto	   the	   chromatography	  column,	  samples	  (proteins	  and	  polyP/monoP)	  were	  centrifuged	  at	  20,800	  x	  g	  for	  15	  minutes	  and	  the	  supernatants	  were	  removed	  for	  incubation	  with	  polyP≥1000	  or	  monoP	  and	  gel	  filtration.	  The	  following	  conditions	  were	  evaluated:	  	  a. C5b,6	  alone	  b. C5b,6	  with	  polyP≥1000	  c. C5b,6	  with	  monoP	  (Na3PO4)	  d. C7	  alone	  e. C7	  with	  polyP≥1000	  20	  μg	  of	  each	  protein	  in	  100	  μL	  of	  their	  respective	  buffers	  was	  prepared.	  polyP≥1000	  and	  monoP	  were	  used	  at	  a	   final	   concentration	  of	  9.71	  mM.	  The	  gel	   filtration	  column	   (G.E.	  Superose	  6	  PC	  3.2/30)	  was	  equilibrated	  with	  filtered	  HBS	  for	  C5b,6	  or	  PBS	  for	  C7	  for	  at	  least	  1	  hour,	  at	  50	  μL	  per	  minute.	  The	  detector	  was	  set	  for	  absorbance	  of	  280	  nm.	  After	  equilibration,	  samples	  were	  loaded	   onto	   the	   column	   and	   50	   μL	   fractions	   were	   collected.	   Data	   for	   each	   curve	   were	  normalized	  to	  the	  time	  point	  when	  the	  peak	  starts.	  	  4.7 Measuring	  Effect	  of	  PolyP≥1000	  on	  C5b-­‐7	  and	  C5b-­‐8	  Binding	  to	  Erythrocyte	  Membranes	  Chicken	  red	  blood	  cells	  (cRBC)	  were	  obtained	  by	  washing	  chicken	  whole	  blood	  (Colorado	  Serum	  Company,	  Denver,	  CO)	  with	  gelatin	  veronal	  buffer	  (GVB),	  followed	  by	  centrifugation	  at	  1000	  x	  g	  for	  10	  minutes.	   The	  cells	  were	  washed	  4	   times.	  The	  concentration	  of	   the	   cells	  was	  measured	  using	   Advia	   120	   Hematology	   Analyzer	   from	   Bayer	   (Leverkusen,	   Germany).	   The	   final	  concentration	  of	  cells	  used	  in	  each	  experiment	  was	  3.00	  –	  3.15	  x	  109	  cells/mL.	  	  Varying	   concentrations	  of	   polyP≥1000	   (0	  μM	  –	  10	  mM)	  were	   added	   to	   the	   cell	   suspension	   and	  incubated	  for	  5	  minutes.	  C5b,6	  was	  then	  added	  to	  the	  reaction	  mixtures	  (final	  concentration	  2.5	  nM)	   and	   incubated	   for	   5	   minutes.	   Finally,	   C7	   and	   C8	   or	   C7	   alone	   were	   added	   at	   final	  concentrations	  of	  2.5	  nM	  and	  incubated	  for	  a	  further	  5	  minutes.	  The	  cells	  were	  then	  pelleted	  by	  	  	   31	  centrifugation	  at	  300	  x	  g	  for	  3	  minutes	  and	  40	  μL	  of	  the	  supernatant	  was	  transferred	  to	  another	  microfuge	   tube.	   This	   fraction	  was	   again	   centrifuged	   at	   300	   x	   g	   for	   3	  minutes	   to	   remove	   any	  remaining	  cRBCs.	  32	  μL	  of	  this	  fraction	  was	  mixed	  with	  8	  μL	  of	  Laemmli	  loading	  buffer	  (with	  β-­‐mercaptoethanol)	  for	  separation	  by	  SDS-­‐PAGE	  using	  a	  10%	  acrylamide	  gel.	  After	  transfer	  of	  the	  gel,	  Western	  blotting	  was	  performed	  using	  a	  goat-­‐anti-­‐human	  C5	  primary	  antibody	  and	  680RD	  donkey-­‐anti-­‐goat	   secondary	   antibody	   from	   LI-­‐COR	   Biosciences.	   Following	   detection,	  densitometry	  was	  performed	  on	  the	  scanned	  blot	   images	  using	  the	  Odyssey	  Software	  from	  LI-­‐COR	   Biosciences	   (Lincoln,	   Nebraska,	   USA).	   Values	  were	   normalized	  to	  the	   experimental	  conditions	   in	   which	   cRBC	   were	   incubated	   with	   maximal	   concentrations	   of	   polyP≥1000,	  but	  without	  C7	   or	  C8.	  The	   amount	   of	   unbound	   C5b	   under	   these	   conditions	  was	   considered	   to	   be	  100%.	  Three	  independent	  experiments	  were	  averaged.	  4.8 In	  vivo	  rodent	  model	  of	  laser-­‐induced	  choroidal	  neovascularization	  (CNV)	  In	   collaboration	   with	   Dr.	   Joanne	  Matsubara	   (Eye	   Care	   Centre,	   Vancouver,	   BC),	   laser-­‐induced	  CNV	   experiments	   with	   rats	   were	   performed.	   All	   animal	   studies	   were	   approved	   by	   the	   UBC	  Animal	  Ethics	  Committee.	  The	  methods	  are	  as	  previously	  described	  [139].	  Briefly,	  eleven-­‐week-­‐old	  female	  Long	  Even	  (LE)	  rats	  (Charles	  River	  Laboratory,	  Wilmington,	  MA)	  weighing	  256	  to	  315	  g	  were	  used	  and	  handled	   in	  accordance	  with	   institutional	  guidelines	  and	  the	  ARVO	  Statement	  for	   the	  Use	   of	  Animals	   in	  Ophthalmic	   and	  Vision	  Research.	   The	   LE	   rats	  were	   anesthetized	  by	  intramuscular	   injection	  of	   a	   1	  mL/kg	  mixture	   (1:1)	   of	   ketamine	  hydrochloride	   (40	  mg/kg)	   and	  xylazine	   hydrochloride	   (10	   mg/ml),	   and	   their	   pupils	   were	   dilated	   with	   tropicamide	   (0.5%	  Mydrin).	  The	  eyes	  were	  pressed	  against	  a	  22	  mm	  x	  22	  mm	  glass	  coverslip	   to	  act	  as	  a	  contact	  lens.	  A	   slit-­‐lamp	  biomicroscope	  and	  a	  diode	   red	   laser	   (OcuLight	   SLx;	   Iris	  Medical	   Instruments,	  Mountain	  View,	  California,	  USA)	  set	  at	  650	  nm,	  150	  mW	  intensity,	  100	  ms	  duration,	  and	  75	  µm	  spot	   size	  was	  used	   to	   inflict	   localized	  damage	   to	  Bruch’s	  membrane	   in	   the	   retina	  of	   the	   rats.	  Damage	   was	   confirmed	   by	   the	   appearance	   of	   a	   central	   bubble.	   4	   roughly	   equidistant	   laser	  injuries	  around	  the	  optic	  nerve	  of	  each	  eye	  were	  administered	  to	  each	  rat.	  	  	  	   32	  Rats	  were	  administered	  5%	  isoflurane	  immediately	  after	  the	  laser	  injury,	  and	  maintained	  at	  2%	  isoflurane	  for	  the	   intravitreal	   injection	  (occurring	  within	  10	  minutes	  of	  the	   laser	   injury).	  Pupils	  were	  dilated	  using	  eye	  drops	  -­‐	  0.5%	  tropicamide	  and	  2.5%	  phenylephrine	  hydrochloride.	  0.4%	  benoxinate-­‐HCl	  was	  also	  applied	  topically	  to	  the	  eyes	  as	  an	  additional	  local	  anesthetic,	  followed	  by	   a	   drop	   of	   0.5%	   levofloxacin	   ophthalmic	   antibiotic	   solution.	   Under	   a	   Stereo	   dissection	  microscope	  (SMZ	  1000;	  Nikon,	  Tokyo,	  Japan),	  the	  eye	  was	  punctured	  in	  the	  limbus	  region	  with	  a	  fine	  28-­‐gauge	  needle.	  This	  allowed	  the	  penetration	  of	  a	  32-­‐guage	  Hamilton	  syringe	  needle,	  used	  to	  slowly	  inject	  5	  µL	  of	  monoP	  or	  polyP≥1000	  diluted	  in	  sterile	  water	  into	  the	  vitreous	  of	  the	  eye,	  to	   yield	   a	   final	   estimated	   concentration	   of	   ~200	   µM.	   After	   injection,	   the	   needle	   was	   held	   in	  place	   inside	  the	  eye	  for	  30-­‐60	  seconds	  and	  slowly	  withdrawn,	  to	  minimize	  fluid	   leakage.	  After	  intravitreal	  injection,	  the	  rats	  were	  monitored	  in	  an	  enclosed	  container	  to	  ensure	  recovery.	  	  5	  days	  post-­‐injury,	   rats	  were	  euthanized	  with	  CO2	  asphyxiation,	  and	   the	  eyes	  were	  harvested	  and	   fixed	   in	   4%	   paraformaldehyde	   for	   24	   hours.	   Each	   eye	   was	   dissected	   under	   a	   dissecting	  microscope.	   The	   anterior	   segment	   and	   crystalline	   lens	   were	   removed,	   and	   the	   retina	   was	  detached	  and	  cut	  from	  the	  optic	  nerve	  using	  fine	  curved	  scissors.	  The	  remaining	  choroidal/RPE	  eye	  cups	  were	  washed	  with	  PBS	  and	  subsequently	  immunostained	  with	  a	  1:100	  dilution	  of	  500	  μg/mL	   solution	   of	   FITC-­‐labeled	   Graffonia	   lectin	   (IB4)	   (Griffonia	   simplicifolia,	   Alexa	   Fluor®	   488	  Conjugate,	  Life	  technology,	  Burlington,	  ON)	  to	  detect	  endothelial	  cells,	  and	  a	  1:250	  dilution	  of	  a	  1	  mg/mL	  solution	  of	  rabbit	  IgG	  polyclonal	  anti-­‐C5b-­‐9	  (Bioss,	  Woburn,	  MA)	  to	  detect	  deposition	  of	   C5b-­‐9	   (MAC).	   With	   fine	   curved	   scissors,	   radial	   cuts	   were	   made	   towards	   the	   optic	   nerve,	  around	   the	   lesions.	   The	   cut	   eye	   cups	   were	   flatmounted	   on	   glass	   slides	   and	   visualized	   on	   a	  confocal	   microscope	   (Zeiss-­‐LSM	   510	   META,	   Thornwood,	   NY,	   USA),	   and	   the	   images	   were	  processed	  with	   the	   ImageJ	   program	   (1.47v,	   National	   Institute	   of	   Health,	   USA).	   Each	   eye	  was	  considered	  as	  n=1.	  Lesion	  size	  and	  MAC	  deposition	  for	  each	  rat	  eye	  was	  determined	  from	  the	  average	  of	  the	  4	  lesions	  on	  that	  eye.	  Similar	  methods	  were	  used	  to	  study	  CNV	  in	  C57Bl6	  wild-­‐type	  mice	  (age	  6-­‐10	  weeks).	  Equal	  final	  concentrations	  of	  polyP	  and	  monoP	  were	  injected	  intravitreally	  into	  eyes.	  As	  the	  total	  vitreous	  volume	  of	  mice	  is	  ~7	  µL	  compared	  to	  ~54	  µL	  in	  rats,	  1	  µL	  of	  solution	  was	  injected	  into	  the	  eyes	  	  	   33	  of	  mice.	  After	   sacrifice,	   the	  eyes	  were	  extracted,	  14	  days	  post-­‐injury.	   The	   same	  antibodies	  as	  with	  the	  rats	   (above)	  were	  used	  to	  detect	  the	  CNV	   lesions,	  with	  the	  fluorochrome	   label	  being	  changed	  to	  Cy3.	  C5b-­‐9	  was	  detected	  with	  a	  rabbit	  anti-­‐human	  polyclonal	  antibody	  for	  SC5b-­‐9,	  followed	  by	  a	  secondary	  antibody	  detected	  with	  a	  FITC-­‐labelled	  anti-­‐rabbit	  IgG.	  Volumes	  of	  CNV	  lesions	   and	   C5b-­‐9	   deposition	   were	   determined	   following	   imaging	   with	   a	   Nikon	   confocal	  microscope	  (Nikon	  Instruments,	  Melville,	  NY,	  USA).	  	  4.9 Quantification	  of	  C5b-­‐9	  (MAC)	  Deposited	  on	  ARPE-­‐19	  (RPE)	  Cells	  and	  RF/6A	  (CEC)	  Cells	  were	  grown	  to	  100%	  confluence	  in	  100	  mm	  cell-­‐culture	  dishes,	  and	  washed	  with	  PBS,	  after	  which	  serum-­‐free	  growth	  media	  was	  added.	  The	  cells	  were	  incubated	  overnight	  (~16	  hours)	  in	  the	  serum	  starvation	  media,	  which	  was	  then	  removed.	  Cells	  were	  washed	  3x	  with	  PBS	  and	  were	  lifted	   after	   incubation	   for	   10	   minutes	   with	   2.5	   mL	   of	   StemPro®	   Accutase®	   Cell	   Dissociation	  Solution.	  7.5	  mL	  of	  their	  respective	  growth	  media	  were	  used	  to	  neutralize	  the	  cell	  dissociation	  solution,	  and	  the	  cells	  were	  centrifuged	  at	  250	  x	  g	  for	  8	  minutes.	  The	  supernatant	  was	  removed	  and	   the	  cell	  pellets	  were	   resuspended	   in	  PBS,	   transferred	   into	  sterile	  1.5	  ml	  microfuge	   tubes,	  and	  pelleted	  at	  500	  x	  g	   for	  10	  minutes,	  after	  which	  the	  supernatant	  was	  removed.	  A	  range	  of	  concentrations	  of	  polyP130	  or	  monoP	  diluted	  in	  PBS	  was	  added	  to	  the	  cell	  pellets,	  and	  incubated	  for	  1	  hour	  at	  37°C.	  Normal	  human	  serum	  was	  added	  to	  the	  cells	  at	  a	  final	  concentration	  of	  5%	  (for	  CEC)	  or	  25%	  (for	  RPE)	  for	  another	  2	  hours	  to	  induce	  complement	  activation	  on	  the	  cells.	  The	  cells	  were	  once	  again	  pelleted	  at	  500	  x	  g	  for	  10	  minutes	  and	  the	  supernatant	  from	  each	  sample	  was	   transferred	   to	  microfuge	   tubes	  and	   frozen	  at	   -­‐80°C	   for	   subsequent	  analysis	  by	  ELISA	   (see	  below).	   Cells	  were	  washed	  once,	   resuspended	   in	   100	  µL	   of	   FACS	   buffer	   (1%	  BSA	   in	   PBS)	   and	  centrifuged	  at	  500	  x	  g,	  removing	  the	  supernatant	  after	  spinning,	  before	  addition	  of	  the	  primary	  antibody.	  A	  rabbit	  anti-­‐mouse	  C5b-­‐9	  antibody	  diluted	  in	  FACS	  buffer	  was	  added	  to	  the	  tubes	  at	  a	  final	   concentration	  of	  1	  μg/mL	  and	   incubated	  at	  4°C	   for	  40	  minutes.	   Excess	  primary	  antibody	  was	  removed	  by	  washing	  the	  cells	  once	  as	  previously	  described.	  Secondary	  antibody	  (goat	  anti-­‐mouse	   IgG	  Alexa	  Fluor	  488)	  diluted	   in	  PBS	  was	  added	  at	  a	   final	  concentration	  of	  5	  μg/mL	  and	  the	  cells	  were	  incubated	  for	  another	  30	  minutes	  at	  4°C.	  Cells	  were	  washed	  and	  each	  sample	  was	  	  	   34	  resuspended	   in	  500	  µL	  of	  FACS	  buffer	  with	  1	  µg/mL	   (final	   concentration)	  of	  propidium	   iodide	  and	  transferred	  into	  tubes	  for	  FACS	  analysis	  (BD	  LSRII,	  BD	  Biosciences,	  San	  Jose,	  CA,	  USA).	  Data	  analysis	   was	   conducted	   using	   FlowJo	   software	   (vX.0.7,	   FlowJo,	   LLC,	   Ashland,	   OR,	   USA)	   and	  GraphPad	  Prism.	  	  4.10 Quantification	  of	  SC5b-­‐9	  The	  MicroVue	  SC5b-­‐9	  Plus	  Enzyme	  Immunoassay	  Kit	  (Quidel,	  San	  Diego,	  CA,	  USA)	  was	  used	  to	  quantify	   the	   amount	   of	   SC5b-­‐9	   according	   to	   the	  manufacturer’s	   instructions.	   Briefly,	   96-­‐well	  plates	   coated	  with	  mouse	  monoclonal	   anti-­‐human	   SC5b-­‐9	   antibodies	   were	   washed	  with	   PBS	  containing	   0.05%	   Tween-­‐20®,	   and	   Proclin®	   300	   (wash	   buffer).	   Samples	   and	   controls	   were	  incubated	   in	   wells	   at	   room	   temperature	   for	   one	   hour.	   Unbound	   proteins	   were	   removed	   by	  washing	  the	  plate	  5x	  with	  wash	  buffer.	  Horseradish	  peroxidase-­‐conjugated	  polyclonal	  goat	  anti-­‐SC5b-­‐9	  antibodies	  were	  added	  to	  each	  well	  and	  incubated	  at	  room	  temperature	  for	  30	  minutes.	  The	  wells	  were	  washed	  as	  before,	  and	  a	  peroxide	  substrate	  and	  3,3’,5,5’-­‐teramethylbenzidene	  (TMB)	  –	  a	  chromogen	  –	  was	  added.	  H2SO4	  was	  used	  to	  stop	  the	  reaction	  after	  15	  minutes.	  The	  absorbance	  of	  the	  samples	  was	  read	  at	  450	  nm	  with	  a	  Spectramax	  384	  Plus	  Microplate	  Reader,	  with	  the	  level	  of	  absorbance	  correlating	  with	  the	  amount	  of	  SC5b-­‐9	  in	  the	  sample.	  	  4.11 Quantification	  of	  Nuclear	  Integrity	  RF/6A	  and	  ARPE-­‐19	  cells	  were	  seeded	  in	  Falcon	  TC-­‐treated	  black	  96-­‐well	  plates	  (BD	  Biosciences,	  San	  Jose,	  California,	  USA)	  at	  a	  concentration	  of	  104	  cells	  per	  well	  and	  cultured	  in	  their	  respective	  growth	  media	   for	   24	   hours	   at	   37°C.	   The	   cells	  were	   then	   simultaneously	   treated	  with	   varying	  concentrations	   and	   combinations	   of	   H2O2	   and	   polyP130	   or	   monoP	   diluted	   in	   their	   respective	  growth	  media,	  with	  final	  concentrations	  ranging	  from	  0-­‐1650	  μM	  and	  0-­‐2	  mM,	  respectively,	  at	  37°C	  for	  24	  hours.	  An	  Eclipse	  TS100	  brightfield	  phase-­‐contrast	  microscope	  (Nikon	  Instruments,	  Melville,	   New	   York,	   USA)	   was	   used	   to	   visualize	   changes	   in	   cellular	   morphology	   (confluence,	  intercellular	  interactions,	  and	  opacity	  of	  nuclei).	  Hoechst	  dye,	  a	  nuclear	  stain,	  was	  added	  to	  the	  cells	   at	   a	   final	   concentration	   of	   333	   ng/mL,	   and	   incubated	   for	   30	   minutes	   at	   37°C.	   An	  	  	   35	  ArrayScan™	  VTI	  High-­‐Content	  System	  Reader	  (Thermo	  Fisher	  Scientific,	  Burlington,	  ON,	  Canada)	  was	  used	  to	  quantify	  the	  intensity	  of	  nuclear	  staining	  and	  the	  number	  of	  cells	  with	  detectable	  nuclei.	  	  4.12 Effect	  of	  PolyP	  on	  the	  Integrity	  of	  H2O2	  The	  integrity	  of	  H2O2	  was	  assessed	  by	  monitoring	  changes	  in	  absorbance	  of	  1	  mM	  H2O2	  at	  240	  nm	  in	  the	  presence	  of	  polyP	  and/or	  catalase	  over	  different	  time	  periods	  as	  noted.	  Catalase	  was	  used	  as	  a	  positive	  control	  to	  demonstrate	  that	  its	  degradation	  of	  H2O2	  caused	  a	  rapid	  (within	  2-­‐3	  minutes)	  decrease	  in	  absorbance	  [140].	  In	  this	  case,	  the	  absorbance	  of	  H2O2	  was	  followed	  for	  10	  minutes	  after	  addition	  of	  catalase	  to	  the	  cuvette	  of	  the	  spectrophotometer.	  The	  addition	  of	  1	  mM	  polyP130	  instead	  of	  catalase	  had	  no	  effect	  on	  the	  baseline	  absorbance	  of	  H2O2.	  4.13 Catalase	  Assay	  ARPE-­‐19	   and	   RF/6A	   cells	   were	   grown	   to	   90%	   confluence	   in	   6-­‐well	   plates	   in	   their	   respective	  growth	  media	   and	  washed	  with	   PBS.	   Varying	   concentrations	   of	   polyP130	   or	  monoP	   diluted	   in	  growth	  media	  were	  added	   to	   the	   cells	   and	   incubated	   for	  24	  hours.	   The	   cells	  were	   lysed	  with	  RIPA	   buffer	   (30mM	   Tris-­‐HCl,	   150mM	   NaCl,	   1%	   Igepal,	   0.5%	   deoxycholate,	   2mM	   EDTA,	   0.1%	  EDTA,	   pH	   7.4)	   and	   centrifuged	   at	   14,000	   x	   g	   for	   15	  minutes.	   The	   pellets	  were	   discarded	   and	  supernatant	   lysates	   (80-­‐100	   µl	   volumes)	   were	   removed	   and	   stored	   at	   -­‐80°C	   for	   subsequent	  measurement	  of	  catalase	  activity.	  	  Total	   protein	   content	   of	   each	   sample	   (cleared	   lysate)	   was	   measured	   using	   the	   Pierce™	   BCA	  Protein	  Assay	  kit	  (Rockford,	  Illinois,	  USA)	  according	  to	  the	  manufacturer’s	  instructions.	  Catalase	  in	   each	   sample	  was	  measured	   using	   the	   catalase	   assay	   kit	   from	  Cell	   Biolabs,	   Inc.	   (San	  Diego,	  California,	   USA)	   according	   to	   the	  manufacturer’s	   instructions.	   Briefly,	   H2O2	  was	   added	   to	   the	  lysates	  for	  3	  minutes,	  allowing	  the	  catalase	  in	  the	  samples	  to	  degrade	  H2O2	  to	  O2	  and	  H2O.	  The	  reaction	   was	   quenched	   after	   3	   minutes	   with	   sodium	   azide.	   The	   leftover,	   undegraded	   H2O2	  allows	   the	   coupling	   reaction	   of	   3,5-­‐dichloro-­‐2-­‐hydroxy-­‐benzenesulfonic	   acid	   (DHBS)	   with	   4-­‐aminophenazone	   (4-­‐aminoantipyrene,	   AAP)	   to	   proceed,	   catalyzed	   by	   horseradish	   peroxidase	  	  	   36	  (HRP).	  After	  30	  minutes,	  the	  coupling	  reaction	  produces	  a	  quinoneimine	  dye	  that	  was	  measured	  at	  540	  nm	  with	  the	  Spectramax	  384	  Plus	  Microplate	  Reader.	  The	  absolute	  amount	  of	  catalase	  was	   determined	   from	   a	   standard	   curve	   of	   0-­‐100	   U/mL	   that	   was	   generated	   with	   purified	  catalase,	  provided	  in	  the	  kit.	  4.14 qRT-­‐PCR	  Catalase	  Gene	  Expression	  Analysis	  Treatment	  of	  the	  cells	   followed	  an	   identical	  procedure	  as	  that	  for	  the	  catalase	  assay.	  A	  kit	   for	  converting	  the	  RNA	  from	  cells	  to	  cDNA	  was	  used,	  according	  to	  the	  manufacturer’s	  instructions	  (RNeasy	  Mini®	  Kit,	  Applied	  Biosystems,	  Foster	  City,	  California,	  USA).	  Briefly,	  cells	  were	  lysed	  with	  lysis	  buffer	  (provided	  in	  kit)	  and	  β-­‐mercaptoethanol,	  followed	  by	  addition	  of	  70%	  ethanol.	  These	  were	   centrifuged	   in	   spin	   columns	   at	   ≥8000	   x	   g	   and	   underwent	   multiple	   washes	   with	   the	  supplied	   washing	   buffer.	   The	   RNA	   was	   finally	   eluted	   with	   RNase-­‐free	   water.	   qScript™	   cDNA	  Synthesis	   Kit	   from	  Quanta	   Biosciences	   (Gaithersburg,	  Maryland,	  USA)	  was	   used	   to	   synthesize	  cDNA	  from	  the	  isolated	  RNA.	  A	  master	  mix	  consisting	  of	  reverse	  transcriptase,	  dNTPs,	  random	  primers,	  magnesium,	  and	  oligo(dT)	  was	  added	  to	  400	  ng	  of	  the	  isolated	  RNA.	  The	  samples	  were	  run	  in	  a	  thermal	  cycler	  program	  according	  to	  the	  manufacturer’s	  instructions.	  A	  TaqMan®	  Fast	  Advanced	   Master	   Mix	   (Applied	   Biosystems,	   Carlsbad,	   California,	   USA)	   was	   added	   to	   the	  synthesized	  cDNA,	  with	  a	  catalase	  gene	  primer	  from	  the	  same	  company.	  GAPDH	  was	  used	  as	  a	  housekeeping	  gene	  control,	  using	  primers	  from	  the	  same	  company.	  The	  samples	  were	  again	  run	  and	  analyzed	  using	  the	  StepOnePlus	  System™	  (Applied	  Biosystems)	  qPCR	  machine	  and	  program.	  	  	  4.15 VE-­‐cadherin	  Expression	  Assay	  RF/6A	   were	   seeded	   at	   104	   cells/well	   in	   black	   96-­‐well	   plates	   and	   grown	   to	   90%	   confluence.	  Treatments	   of	   varying	  H2O2	   and	   polyP130	  were	   added	   and	   incubated	   for	   48	   hours.	   Cells	  were	  fixed	  with	  2%	  PFA	  for	  5	  minutes	  and	  blocked	  with	  5%	  BSA	  in	  PBS	  for	  1	  hour.	  Rabbit	  anti-­‐human	  VE-­‐cadherin	   polyclonal	   antibody	   at	   a	   final	   concentration	   of	   2	   µg/mL	   was	   incubated	   for	   40	  minutes,	  followed	  by	  a	  20	  minute	  incubation	  with	  AlexaFluor	  488	  nm	  goat	  anti-­‐rabbit	  secondary	  antibody	   at	   a	   final	   concentration	   of	   2	   µg/mL.	   Hoechst	   dye	   (100	   ng/mL)	   was	   added	   and	  	  	   37	  incubated	   for	   20	   minutes	   at	   37°C.	   The	   plate	   was	   scanned	   on	   a	   Cellomics	   ArrayScan	   Target	  Activation	  program,	  at	  15	  fields	  per	  well	  with	  a	  20x	  objective	  lens.	  4.16 EA.hy	  926	  Cell	  Proliferation	  Assay	  EA.hy	  926	  cells	  were	  seeded	  at	  a	  density	  of	  2	  x	  104	  cells/mL	  in	  a	  96-­‐well	  plate	  in	  the	  appropriate	  growth	  media	   (see	  Materials),	   and	  were	   incubated	  at	  37°C	  and	  5%	  CO2	  overnight.	   The	  media	  was	  removed	  and	  replaced	  with	  different	  dilutions	  of	  polyP130	  and	  monoP	  in	  EA.hy	  926	  growth	  media.	  At	  0,	  18,	  24,	  42,	  48,	  66,	  72	  hours,	  media	  was	  removed	  and	  replaced	  with	  30	  µL	  of	  0.5%	  trypsin	  to	  detach	  the	  cells	  from	  the	  plates.	  Trypsinization	  was	  stopped	  with	  30	  µL	  growth	  media.	  The	  cell	  suspension	  was	  mixed	  with	  Trypan	  Blue	  at	  a	  1:1	  dilution,	  and	  the	  number	  of	  cells	  was	  counted	  with	  a	  hemocytometer,	  as	  previously	  described	  [141].	  Averages	  were	  obtained	  from	  2	  counts	  of	  each	  sample,	  in	  triplicate.	  4.17 BrdU	  Assay	  for	  Cell	  Proliferation	  EA.hy	  926	  cells	  were	  seeded	  in	  96	  well	  plates	  at	  2	  cell	  densities	  (2.5	  x	  103	  or	  5	  x	  103	  cells/well).	  The	  cells	  were	  grown	   in	  media	   containing	  1%	  FBS	   for	  5.5	  hours,	   followed	  by	   treatments	  with	  varying	  concentrations	  of	  polyP130	  or	  monoP	  in	  growth	  media.	  A	  	  (bromodeoxyuridine)	  BrdU	  cell	  proliferation	   assay	   kit	   from	   Abcam	   (Toronto,	   ON,	   Canada)	   was	   used	   to	   quantify	   cell-­‐proliferation,	  performed	  according	  to	  the	  manufacturer’s	  instructions.	  Briefly,	  20	  µL/well	  of	  the	  BrdU	  labelling	  solution	  was	  added	  at	  20	  hours	  following	  addition	  of	  the	  polyP130	  or	  monoP	  and	  incubated	  a	   further	  4	  hours,	   to	  allow	  the	  BrdU	  to	  be	   incorporated	   into	  the	  newly	  synthesized	  DNA.	   Media	   and	   labeling	   solution	   were	   removed	   from	   the	   wells,	   and	   the	   cells	   were	   fixed	  (fixative	   solution	   provided	   in	   kit)	   for	   30	   minutes	   at	   room	   temperature.	   After	   washing,	   100	  µL/well	  of	   the	  anti-­‐BrdU	  antibody	  was	  added	  and	   incubated	   for	  1	  hour	  at	   room	  temperature,	  followed	  by	  addition	  of	  a	  peroxidase	  goat	  anti-­‐mouse	  IgG	  conjugate	  (100	  µL/well).	  A	  3,3’,	  5,5”-­‐tetramethylbenzidine	  (TMB)	  peroxidase	  substrate	  solution	  was	  added	  and	  incubated	  in	  the	  dark	  at	   room	  temperature.	  100	  µL	  of	   stop	   solution	  was	  added	  and	   the	  absorbance	  of	   the	   samples	  was	  measured	  at	  450	  nm	  with	  the	  Spectramax	  384	  Plus	  Microplate	  Reader.	  	  	   38	  4.18 Statistics	  Statistical	  analyses	  were	  conducted	  using	  GraphPad	  Prism	   (version	  5.0,	   La	   Jolla,	  CA,	  USA)	  and	  Microsoft	  Excel.	  Where	  appropriate,	  student’s	  t-­‐tests	  and	  one	  way	  analysis	  of	  variance	  (ANOVA)	  were	  conducted	  between	  samples,	  with	  P	  ≤	  0.05	  being	  considered	  significant.	  Unless	  otherwise	  noted,	  standard	  errors	  of	  the	  mean	  are	  shown	  in	  all	  results.	  4.19 Ethics	  All	  animal	  protocols	  were	  approved	  by	  the	  UBC	  Animal	  Ethics	  Committee,	  application	  number	  A14-­‐0135.	   Studies	   with	   humans	   were	   approved	   by	   the	   UBC	   Clinical	   Research	   Ethics	   Board	  (CREB)	  under	  protocol	  number	  H12-­‐02508.	  	   	  	  	   39	  Chapter	  5: Results	  5.1	   Mechanisms	  by	  which	  PolyP	  Suppresses	  Complement	  Activation	  Previous	  work	  in	  our	  lab	  revealed	  that	  polyP	  suppresses	  complement	  activation	  via	  the	  terminal	  pathway.	  Here,	  I	  first	  confirmed	  those	  findings,	  using	  the	  terminal	  pathway	  hemolytic	  assay	  (see	  Methods)	   in	   which	   cRBCs	   are	   incubated	   with	   varying	   concentrations	   of	   polyP	   or	   monoP,	  followed	  by	  addition	  of	  EDTA-­‐treated	  serum,	  and	  a	  quantity	  of	  purified	  C5b,6	  to	  achieve	  ~80%	  red	   blood	   cell	   lysis	   after	   a	   30	   minute	   incubation.	   As	   seen,	   polyP	   suppresses	   complement	  activation	  in	  a	  concentration-­‐dependent	  manner	  (Figure	  8a).	  With	  the	  same	  assay,	  Jovian	  Wat	  (a	  previous	  MSc	  student	  in	  the	  Conway	  lab)	  also	  showed	  that	  this	  suppression	  is	  polymer	  length-­‐dependent,	  and	  that	  monoP	  has	  no	  effect	  at	  equivalent	  orthophosphate	  concentrations	  (Figure	  8b)	   [57].	   Similar	   results	   were	   obtained	   with	   purified	   complement	   factors	   C7,	   C8	   and	   C9	  (replacing	   serum),	   indicating	   that	   the	   polyP	   interacts	   directly	   with	   one	   or	   more	   of	   the	  complement	  proteins,	   rather	   than	  mediating	   its	  effects	   solely	   through	   interactions	  with	  other	  serum	  component(s).	  	  	  	  	  	  	  	  	   40	  	  Figure	   8:	   PolyP	   suppresses	   the	   terminal	   pathway	   of	   complement	   in	   a	   concentration-­‐and	   size-­‐dependent	   manner.	   (a)	   Increasing	   concentrations	   of	   polyP≥1000	   resulted	   in	   greater	   suppression	   of	   the	  terminal	  pathway	  of	   complement.	  This	  was	  consistent	  with	  what	  was	  previously	   shown	  by	   Jovian	  Wat	  [57].	  Moreover,	  monoP	  does	  not	  affect	   the	  terminal	  pathway	  of	  complement,	  excluding	  the	  possibility	  that	   suppression	   occurs	   via	   an	   ionic	   effect	   (b,	   P1).	   Rather,	   with	   increasing	   lengths	   of	   polyP,	   a	   greater	  suppression	  is	  observed	  (b).	  The	  amount	  of	  C5b,6	  used	  was	  selected	  to	  result	  in	  ~80%	  hemolysis,	  and	  this	  was	  designated	  as	   the	  maximum	  (100%	   in	  a,	  1.0	   in	  b)	  hemolysis	  under	   these	  experimental	   conditions.	  n=3	  for	  a;	  n=4	  for	  b.	  *Comparisons	  to	  control	  sample	  (no	  phosphate).	  *P	  ≤	  0.001.	  	  	  5.1.1	   PolyP	  Interacts	  Directly	  with	  C5b,6	  The	   mechanisms	   by	   which	   polyP	   suppresses	   the	   terminal	   pathway	   of	   complement	   were	  examined.	  Evidence	  provided	  by	  Jovian	  Wat	  suggested	  that	  polyP	  is	  active	  only	  in	  early	  steps	  in	  the	  assembly	  of	  the	  C5b-­‐9	  membrane	  attack	  complex,	  i.e.	  prior	  to	  addition	  of	  C7	  to	  the	  C65b,6	  complex	   (Figure	   8b).	   We	   therefore	   tested	   whether	   polyP≥1000	   binds	   to	   C5b,6	   or	   C7	   by	   gel	  filtration.	  The	  addition	  of	  polyP≥1000	  to	  C7	  did	  not	  cause	  any	  shift	  in	  the	  gel	  filtration	  profile	  of	  C7	  (Figure	   9a),	   suggesting	   that	   there	  was	   no	   interaction	   between	   polyP≥1000	   and	   C7	   under	   these	  chromatography	   conditions.	   In	   contrast,	   pre-­‐incubation	   of	   polyP≥1000	   with	   C5b,6	   caused	   the	  C5b,6	   gel	   filtration	   profile	   to	   shift	   left,	   indicating	   a	   size	   increase,	   and	   suggesting	   a	   direct	  interaction	  between	  polyP≥1000	  and	  C5b,6	  (Figure	  9b).	  The	  broadness	  of	  the	  peak	  suggested	  that	  	  	   41	  polyP	  may	  cause	  C5b,6	  to	  form	  multimers.	  These	  findings	  were	  in	  line	  with	  thermal	  shift	  assays	  (aka	   differential	   scanning	   fluorimetry	   (DSF))	   performed	   by	   Jovian	   Wat,	   which	   revealed	   that	  polyP	  binds	  to	  and	  destabilizes	  C5b,6	  in	  a	  concentration-­‐dependent	  manner.	  In	  the	  same	  assay,	  polyP≥1000	  had	  no	  effect	  on	  C7	  [57].	  	  Figure	   9:	   Gel	   filtration	   of	   complement	   proteins	   to	   assess	   interaction	   with	   polyP≥1000.	   (a)	   C7	   and	  polyP≥1000	  were	  gel	  filtered	  individually	  or	  in	  combination.	  The	  retention	  time	  of	  C7	  off	  the	  gel	  filtration	  column	   did	   not	   change	   with	   the	   addition	   of	   polyP≥1000.	   (b)	   The	   addition	   of	   polyP≥1000,	   shortened	   the	  retention	  time	  of	  C5b,6.	  The	  broader	  peak	  can	  be	  attributed	  to	  the	  heterogeneity	  of	  polyP	  lengths	  in	  the	  sample.	  These	  findings	  indicate	  that	  polyP≥1000	  binds	  to	  C5b,6	  and	  causes	  it	  to	  have	  a	  higher	  oligomeric	  state.	   The	   ion	   dependence	   of	   this	   effect	   was	   excluded	   by	   the	   observation	   that	   the	   retention	   time	   of	  C5b,6	  was	  unchanged	  by	  the	  addition	  of	  an	  equivalent	  molar	  concentration	  of	  monoP.	  	  	  	   42	  5.1.2	   PolyP≥1000	  Interferes	  with	  Binding	  of	  C5b-­‐7	  and	  C5b-­‐8	  Complexes	  to	  Erythrocyte	  Membranes	  The	  preceding	  results	  indicated	  that	  polyP	  binds	  to	  C5b,6	  and	  alters	  its	  structure.	  We	  predicted	  that	  the	  so-­‐formed	  C5b,6	  would	  alter	  the	  ability	  of	  downstream	  terminal	  complexes,	  C5b-­‐7	  and	  C5b-­‐8,	   from	   binding	   to	   and/or	   integrating	   into	   the	   target	  membrane,	   thereby	   explaining	   the	  reduced	  red	  cell	  lysis	  in	  the	  terminal	  pathway	  assay	  (Figure	  8).	  To	  test	  that	  hypothesis,	  varying	  concentrations	  of	  polyP≥1000	  were	   incubated	   for	  5	  minutes	  with	  cRBC,	   followed	  by	  a	  5	  minute	  incubation	  with	  C5b,6,	  and	  5	  minutes	   later	  by	   the	  addition	  of	  equimolar	  concentrations	  of	  C7	  and	  C8	  (Figure	  10a)	  or	  C7	  alone	  (Figure	  10b)	  for	  a	  further	  5	  minutes.	  The	  amount	  of	  unbound	  C5b-­‐7	   or	   C5b-­‐8	   was	   determined	   by	   quantifying	   the	   amount	   of	   C5b	   in	   the	   supernatant,	   as	  detected	  by	  Western	  blot	  and	  immunodetection	  of	  C5/C5b	  (Figure	  10).	  As	  can	  be	  seen	  with	  the	  densitometry	   results,	   when	   no	   C7	   or	   C8	   is	   added	   (right	   bars),	   there	   is	   no	   binding	   to	   the	  membrane,	   and	   100%	   of	   the	   C5b	   is	   recovered.	   polyP≥1000	   interferes	   with	   binding	   of	   C5b-­‐7	  (Figure	   10a)	   and	   C5b-­‐8	   (Figure	   10b)	   in	   a	   concentration-­‐dependent	   manner.	   These	   findings,	  reported	   in	   the	   journal	  Blood	   [57],	   support	   the	   notion	   that	   destabilization	   of	   C5b,6	   by	   polyP	  either	   reduces	   formation	   of	   the	   downstream	   complexes	   (C5b-­‐7,	   C5b-­‐8,	   C5b-­‐9)	   and/or	   alters	  their	  structures	  so	  that	  binding	  to	  the	  membrane	  is	  reduced.	  	  	   43	  	  Figure	   10:	   Effect	   of	   PolyP	   on	   membrane	   binding/integration	   of	   C5b-­‐7	   and	   C5b-­‐8.	   	   With	   increasing	  concentrations	  of	  polyP≥1000,	  the	  amount	  of	  unbound	  C5b-­‐7	  increases	  (a),	  indicating	  that	  polyP	  interferes	  with	  the	  binding	  of	  this	  complement	  protein	  complex	  to	  the	  membrane.	  Similarly,	  when	  C8	  was	  added	  to	  the	  mixture	  (b),	  the	  reduction	  of	  membrane	  binding/integration	  was	  still	  observed.	  From	  these	  findings	  we	  can	  infer	  that	  destabilization	  of	  C5b,6	  by	  polyP≥1000,	  as	  previously	  found	  by	  Jovian	  Wat	  [57],	  results	  in	  the	  suppression	  of	  binding	  and/or	  integration	  of	  C5b-­‐7	  and	  C5b-­‐8,	  a	  key	  step	  in	  formation	  of	  a	  functional	  MAC.	  The	  results	  are	  representative	  of	  3	  experiments,	  each	  performed	  in	  triplicate.	  Error	  bars	  represent	  standard	   deviation.	   One	  way	   ANOVA	  was	   used	   to	   assess	   significance,	   comparing	   the	   results	   to	   0	   µM	  polyP.	  *P	  ≤	  0.05.	  	  	  	  	  	  	   44	  5.1.3	   Human	  Platelet	  Releasates	  Suppress	  Complement	  Activation	  via	  the	  Terminal	  Pathway	  Platelets	  are	  a	  major	   source	  of	  polyP,	  where	   they	  are	  primarily	   located	   in	   the	  dense	  granules	  [22].	   Platelet	   polyP	   has	   a	   size	   in	   the	   range	   of	   60-­‐100	   orthophosphate	   units.	   PolyP60-­‐100	   is	  released	  into	  the	  circulation	  and/or	  onto	  cell	  surfaces	  when	  the	  platelets	  are	  activated	  [20,	  22].	  Isolation	   of	   polyP	   from	   platelets	   is	   difficult,	   as	   the	   polyP	   is	   prone	   to	   degradation	   (personal	  communication,	   Dr.	   James	   Morrissey).	   Nonetheless,	   we	   attempted	   to	   validate	   that	   polyP	  released	   from	   platelets	   also	   dampens	   complement	   activation.	   Platelet	   releasates	   were	  therefore	  prepared	  from	  human	  platelets	  that	  were	  stimulated	  with	  thrombin.	  Aggregation	  of	  the	   platelets	   provided	   confirmation	   that	   the	   thrombin	   was	   functional.	   The	   effect	   of	   the	  supernatants	   from	   non-­‐activated	   and	   thrombin-­‐activated	   platelets	   on	   terminal	   pathway	  complement	  activation,	  using	  purified	  terminal	  pathway	  complement	  components,	  was	  tested.	  As	   seen	   in	   Figure	   11	   (published	   in	   Blood	   2014	   123:768-­‐76)	   [57],	   platelet	   releasate	   from	  thrombin-­‐activated	   platelets	   significantly	   suppressed	   complement	   activation	   as	   compared	   to	  the	   supernatant	   from	   non-­‐activated	   platelets.	   This	   effect	   appeared	   to	   be	   concentration-­‐dependent.	  Since	  we	  cannot	  directly	  measure	  polyP	  in	  the	  blood	  or	  in	  the	  releasates,	  we	  cannot	  be	   certain	   that	   it	   is	   the	   polyP	   that	   is	   the	   active	   constituent.	   However,	   it	   is	   likely	   that	   polyP	  participates,	   and	   thus	   these	   studies	   support	   the	   notion	   that	   polyP	   released	   from	   activated	  platelets	  dampens	  complement	  activation	  via	  the	  terminal	  pathway.	  	  	  	  	   45	  	  Figure	  11:	  Platelet	  releasates	  suppress	  complement	  activation	  via	  the	  terminal	  pathway.	  The	  left	  side	  of	  the	  figure	  shows	  the	  concentration-­‐dependent	  suppression	  of	  complement	  activation	  via	  the	  terminal	  pathway,	   using	   synthesized	   polyP≥1000.	   100%	   lysis	   was	   achieved	  with	   dH2O.	   The	   beige	   bars	   serve	   as	   a	  point	   of	   comparison	   for	   the	   suppressive	   effect	   of	   the	   releasates	   on	   complement	   activation	   via	   the	  terminal	  pathway.	  On	  the	  right	  side	  of	  the	  figure,	  the	  releasate	  from	  platelets	  activated	  with	  thrombin	  (red	   bars)	   significantly	   reduced	   hemolysis	   compared	   to	   the	   supernatant	   from	   non-­‐activated	   platelets	  (grey	   bars).	   Two	   concentrations	   of	   releasate/supernatant	  were	   used.	   The	   findings	   support	   the	   notion	  that	   polyP	   that	   is	   released	   from	   activated	   platelets	   dampens	   complement	   activation.	   n=3	   for	   each	  condition.	  Student’s	  t-­‐test	  was	  used	  to	  assess	  significance.	  *P	  ≤	  0.05,	  **P	  ≤	  0.01.	  	  	  	  	  	   46	  5.2	   In	  vivo	  Validation	  of	  Therapeutic	  Utility	  of	  PolyP	  in	  a	  Complement-­‐mediated	  Disease	  5.2.1	   Comparing	  the	  Effect	  of	  PolyP130	  and	  PolyP≥1000	  on	  Complement	  Activation	  Prior	   to	   initiating	   in	   vivo	   studies	   to	   test	   the	   clinical	   efficacy	   of	   polyP	   in	   a	   model	   of	   human	  disease,	   we	   considered	   which	   polyP	   preparation	   to	   use.	   We	   appreciated	   the	   possibility	   that	  different	   length	   polymers	   may	   have	   differential	   effects,	   but	   within	   the	   limited	   time	   of	   this	  thesis,	   we	   decided	   to	   select	   one	   size	   range	   to	   evaluate.	   In	   the	   pilot	   studies	   in	   rats	   (below),	  studies	  were	  performed	  with	  polyP≥1000,	  as	  this	  was	  the	  only	  formulation	  available	  at	  that	  time.	  However,	  polyP≥1000	  is	  not	  normally	  found	  in	  mammals	  [13,	  142].	  Moreover,	  the	  preparations	  of	  polyP≥1000	   are	   highly	   heterogeneous,	   and	   this	   length	   is	   known	   to	   have	   the	   most	   profound	  prothrombotic	  effects	  in	  vivo	  when	  administered	  intravenously	  [142].	  The	  vitreous	  is	  not	  known	  to	   contain	   procoagulant	   proteins.	   In	   fact,	   little	   is	   known	   of	   the	   composition	   of	   the	   vitreous.	  Nonetheless,	  our	  lab	  secured	  a	  steady	  source	  of	  a	  homogeneous	  preparation	  of	  polyP130	  (a	  gift	  from	  Regenetiss,	  Inc.,	  Japan).	  This	  length	  is	  in	  the	  range	  that	  mammals	  naturally	  produce	  (5-­‐800	  orthophosphate	  units).	  We	  therefore	  verified	  that	  this	  preparation	  is	  active	  and	  comparable	  to	  the	  polyP≥1000	   that	  was	  provided	  by	  Dr.	   James	  Morrissey.	   This	  was	  achieved	  using	   the	   serum-­‐based	  hemolytic	  assay	  to	  directly	  compare	  the	  effect	  of	  polyP130	  and	  polyP≥1000	  on	  the	  terminal	  pathway.	   As	   seen	   in	   Figure	   6,	   polyP130	   was	   very	   similar	   to	   polyP≥1000	   in	   suppressing	   terminal	  pathway	   complement	   activation.	   Thus,	   polyP130	  was	   used	   for	   the	   in	   vivo	   studies	   in	  mice	   (see	  below)	  and	  most	  of	  the	  subsequent	  in	  vitro	  experiments,	  unless	  otherwise	  stated.	  	  	   47	  	  Figure	  12:	  Terminal	   pathway	   hemolytic	   assay	   to	   compare	   polyP130	   and	   polyP≥1000.	   The	  effect	  of	   each	  form	  of	  polyP	  on	  the	  terminal	  pathway	  of	  complement	  was	  examined	  in	  a	  serum-­‐based	  hemolytic	  assay.	  The	  fact	  that	  the	  two	  forms	  came	  from	  different	  sources	  and	  were	  thus	  synthesized	  by	  slightly	  varying	  techniques	  may	  account	  for	  the	  slightly	  different	  response	  between	  the	  two	  polyP	  forms.	  Nevertheless,	  polyP130	  closely	  mimicked	  the	  suppressive	  effect	  of	  polyP≥1000.	  n	  =	  3.	  	  5.2.2	   PolyP≥1000	  Suppresses	  Complement-­‐mediated	  Damage	  in	  a	  Rodent	  Model	  of	  Wet	  AMD	  A	   pilot	   study	   of	   the	   protective	   effects	   of	   polyP	   against	   AMD	   was	   conducted	   using	   a	   laser-­‐induced	  choroidal	  neovascularization	  (CNV)	  model	   in	  rats.	  The	  model	  mimics	  the	  pathology	  of	  Wet	  AMD,	   in	   that	   Bruch’s	  membrane	   is	   disrupted	  by	   the	   laser	   injury,	   and	   this	   is	   followed	  by	  neovascularization	  into	  the	  sub-­‐retinal	  space	  [143].	  Laser-­‐induced	  CNV	  is	  the	  most	  widely	  used	  and	  accepted	  research	  model	  to	  test	  the	  efficacy	  of	  novel	  AMD	  therapies	  and	  to	  help	  elucidate	  the	  pathogenesis	  of	  AMD	  [143,	  144].	  The	  model	  may	  be	  used	  in	  different	  animals,	  from	  rodents	  to	  non-­‐human	  primates.	  Current	  therapies	  for	  Wet	  AMD	  that	  target	  VEGF	  or	  its	  receptors,	  were	  developed	  and	  approved	  based	  on	  findings	  using	  this	  model	  [143].	  	  	  	   48	  While	  I	   learned	  the	  techniques,	  the	  following	  studies	  in	  rats	  were	  performed	  by	  Dr.	  Jing	  Cui,	  a	  Research	  Associate	  in	  the	  lab	  of	  Dr.	  Joanne	  Matsubara	  (Eye	  Care	  Centre,	  UBC).	  I	  performed	  the	  quantitative	  analyses	  by	  confocal	  microscopy.	  Lesions	   in	  the	  eyes	  of	   the	  rats	  were	   induced	  by	  laser,	   and	   this	   was	   followed	   by	   intravitreal	   injection	   of	   either	   polyP≥1000	   or	   monoP	   (final	  concentration	   of	   ~200	   µM).	   After	   5	   days,	   the	   rats	   were	   euthanized,	   and	   the	   choroids	   were	  isolated,	   fixed,	  whole	  mounted	   and	   stained	   to	   detect	   C5b-­‐9	   deposition	   and	   lesion	   vascularity	  using	  isolectin	  B4.	  	  Results	  from	  the	  4	  lesions	  from	  each	  eye	  were	  averaged	  to	  yield	  areas	  of	  neovascularization	  and	  C5b-­‐9	  deposition.	  Overall,	  there	  was	  a	  trend	  for	   less	  vessel	  formation	  in	  the	  polyP≥1000-­‐treated	  eyes.	  Notably,	  the	  CNV	  lesions	  from	  the	  polyP≥1000-­‐treated	  eyes	  had	  notably	  less	  intense	  C5b-­‐9	  staining,	   findings	   that	   are	   consistent	   with	   the	   suppressive	   effect	   of	   polyP	   on	   complement	  activation	   (Figure	   13).	   These	   preliminary	   results	   supported	   our	   hypothesis	   that	   polyP	   confers	  protection	  against	  AMD	  by	  reducing	  complement-­‐mediated	  injury.	  	  The	  experiments	  set	  the	  stage	  for	  validating	  the	  findings	  through	  more	  extensive	  in	  vivo	  studies.	  Thus,	  with	  the	  help	  of	  Alice	  O’byrne,	  the	  Research	  Assistant,	   I	  performed	  similar	  studies	  using	  mice	   and	   polyP130	   or	   monoP	   at	   a	   final	   concentration	   of	   200	   µM.	   The	   number	   of	   mice	   was	  increased	  (n=17	  and	  15	  for	  polyP	  and	  monoP,	  respectively),	  and	  they	  were	  sacrificed	  for	  analysis	  at	  day	  14	  after	  laser	  injury.	  Confocal	  imaging	  allowed	  us	  to	  measure	  volumes	  of	  stained	  lesions,	  rather	   than	   areas	   and	   fluorescence	   intensities,	   thereby	   providing	   a	   more	   accurate	  representation	  of	  the	  impact	  of	  the	  therapeutic	   intervention.	  As	  seen	  in	  Figure	  14,	   intravitreal	  injection	  of	  polyP130	  resulted	  in	  a	  significant	  reduction	  in	  the	  volume	  of	  the	  neovascularity,	  with	  a	  trend	  for	  a	  reduction	  in	  C5b-­‐9	  deposition.	  In	  no	  case,	  was	  the	  amount	  of	  C5b-­‐9	  in	  the	  monoP-­‐treated	  eyes,	  greater	  than	  in	  the	  polyP130-­‐treated	  eyes	  (Figure	  14).	  	  Overall,	  these	  in	  vivo	  studies	  provide	  exciting	  confirmation	  that	  polyP	  exerts	  protection	  against	  laser-­‐induced	   CNV	   in	   rodents,	   and	   highlights	   the	   importance	   of	   further	   evaluating	   the	  underlying	  mechanisms.	  	  	   49	  	  Figure	  13:	  C5b-­‐9	  deposition	  and	  neovascularization	  after	   laser	  injury	  in	  rats.	  The	  left-­‐side	  panels	  in	  (a)	  show	   representative	   lesions	   from	   the	   monoP-­‐treated	   eyes,	   and	   the	   right-­‐side	   panels	   show	  representative	  lesions	  from	  the	  polyP≥1000-­‐treated	  eyes	  of	  rats	  (a).	  Red	  staining	  reflects	  C5b-­‐9	  deposition.	  Green	  staining	   indicates	  new	  vessel	   formation	   (CNV).	  A	   total	  of	  10	   lesions	   for	   the	  monoP	  group	  and	  8	  lesions	  for	  the	  polyP≥1000	  group	  were	  averaged	  for	  the	  fluorescence	  intensity	  of	  C5b-­‐9	  deposition	  (b)	  and	  CNV	   (c).	   The	   results	   from	   these	   preliminary	   studies	   suggested	   that	   polyP	   dampens	   complement	  activation	  and	  lesion	  size	  in	  this	  model,	  and	  were	  rationale	  for	  more	  extensive	  study.	  	  	  	   50	  	  	  Figure	  14:	  C5b-­‐9	  deposition	  and	  neovascularization	  after	  laser	  injury	  in	  mice.	  The	  left	  side	  panels	  of	  (a)	  show	   representative	   lesions	   from	   the	   monoP-­‐treated	   eyes,	   and	   the	   right-­‐side	   panels	   show	  representative	   lesions	   from	   the	   polyP130-­‐treated	   eyes	   (a).	   Red-­‐staining	   indicates	   new	   vessel	   formation	  (CNV).	  Green-­‐staining	   indicates	   C5b-­‐9	   deposition.	   A	   total	   of	   15	   eyes	   (3-­‐4	   lesions	   each)	   for	   the	  monoP	  group	   and	   17	   eyes	   (3-­‐4	   lesions	   each)	   for	   the	   polyP130	   group	   were	   averaged	   for	   the	   volume	   of	   C5b-­‐9	  deposition	  (b)	  and	  CNV	  (c).	  3D	  projections	  of	  the	  z-­‐stack	  images	  are	  depicted	  for	  both	  the	  monoP	  (d)	  and	  polyP130	  (e)	  lesions.	  Student’s	  t-­‐test	  was	  used	  to	  assess	  significance.	  *P	  ≤	  0.05.	  	  	   51	  5.2.3	   PolyP≥1000	  is	  Functionally	  Stable	  for	  at	  least	  10	  days	  in	  Serum	  	  The	  preceding	  in	  vivo	  studies	  raised	  the	  possibility	  of	  using	  polyP	  as	  a	  treatment	  for	  AMD.	  Many	  challenges	  remain,	  of	  course,	  before	  bringing	  polyP	  to	  the	  clinic.	  One	  of	  these,	  which	  we	  partly	  address	  here,	   involves	  a	  question	  of	   the	  stability	  of	  polyP.	   It	  was	  previously	  reported	  that	   the	  half-­‐life	  of	  polyP	   in	  plasma	  and	  serum	  ex	  vivo	   is	  ~90	  minutes	  [9,	  142].	  However,	  these	  studies	  only	   examined	   the	   stability	   of	   polyP	   in	   terms	   of	   the	   integrity	   of	   the	   polymer,	   rather	   than	   its	  function	  [9].	  We	  therefore	  used	  the	  terminal	  pathway	  of	  complement	  to	  measure	  the	  residual	  complement	  inhibitory	  function	  of	  polyP≥1000	  over	  time,	  following	  incubation	  of	  the	  polyP≥1000	  in	  serum	  ex	  vivo.	  As	  seen	  in	  Figure	  15,	  only	  after	  ~6	  hours,	  the	  polyP≥1000	  appear	  to	  lose	  some	  of	  its	  capacity	   to	   suppress	   complement	   activation	   via	   the	   terminal	   pathway.	   Even	   at	   10	   days	   (240	  hours),	  there	  was	  still	  substantial	  residual	  complement	  inhibiting	  activity.	  We	  did	  not	  evaluate	  the	  structural	  integrity	  of	  the	  polyP≥1000	  during	  the	  course	  of	  this	  study.	  Nor	  did	  we	  examine	  the	  stability	   of	   polyP	   in	   vitreous,	   an	   important	   step.	   However,	   the	   findings	   provide	   an	   important	  starting	  point.	  	  	  	  	   52	  	  Figure	  15:	  Time-­‐dependent	  change	  in	  function	  of	  polyP≥1000	  in	  serum.	  After	  incubation	  in	  normal	  human	  serum	   for	   varying	   amounts	   of	   time,	   200	   μM	   of	   polyP≥1000	   or	   buffer	   alone	   (no	   polyP)	   was	   added	   to	   a	  hemolytic	   assay	   to	   quantify	   the	   effect	   of	   any	   residual	   polyP	   on	   complement	   activation.	   When	   no	  polyP≥1000	  was	  added,	  lysis	  in	  the	  assay	  ranged	  from	  ~40%	  to	  ~25%	  over	  the	  48	  hours	  of	  the	  study.	  The	  polyP≥1000	  that	  was	  added	  to	  the	  serum	  almost	  totally	  suppressed	  complement	  activation	  via	  the	  terminal	  pathway,	  and	  this	  suppressive	  activity	  was	  only	  partly	  lost	  over	  the	  ensuing	  10	  days.	  n=3.	  	  	  	  5.3	   Cellular	  Effects	  of	  PolyP	  The	   preceding	   in	   vivo	   pilot	   studies	   provided	   strong	   evidence	   that	   local	   intravitreal	  administration	   of	   polyP	   suppresses	   complement	   activation	   in	   a	   rodent	   model	   of	   AMD.	   As	  discussed	  in	  the	  Introduction,	  retinal	  pigment	  epithelial	  (RPE)	  cells	  and	  choroid	  endothelial	  cells	  (CEC)	  are	  generally	  accepted	  to	  be	  the	  major	  cellular	  targets	   involved	  in	  the	  pathogenesis	  and	  progression	  of	  AMD.	  We	  therefore	  sought	   to	  determine	  whether	  polyP130	  protects	   these	  cells	  from	   stresses	   known	   to	   be	   associated	   with	   AMD,	   i.e.,	   excess	   complement	   activation	   and	  exposure	   to	   reactive	   oxygen	   species.	   Primary	   RPE	   cells	   and	   CEC	   are	   not	   easily	   obtained	   or	  cultured.	   However,	   human	   RPE	   and	  monkey	   CEC	   cell	   lines	   (ARPE-­‐19	   and	   RF/6A,	   respectively)	  	  	   53	  have	  been	  used	  extensively	  in	  several	  studies	  that	  investigate	  the	  mechanisms	  underlying	  AMD	  [145-­‐150].	  As	  these	  cells	  are	  commercially	  available	  and	  relatively	  easy	  to	  grow	   in	  culture,	  we	  used	  these	  to	  evaluate	  the	  protective	  properties	  of	  polyP.	  We	  herein	  refer	  to	  these	  cell	  lines	  as	  CEC	  and	  RPE.	  5.3.1 PolyP130	  Suppresses	  C5b-­‐9	  Deposition	  on	  CEC	  and	  RPE	  Cells	  We	  first	  assessed	  whether	  polyP130	  could	  protect	  the	  cells	  from	  complement-­‐mediated	  damage.	  Cells	  were	  pre-­‐incubated	  with	  varying	  concentrations	  of	  polyP130	  or	  equimolar	  concentrations	  of	  monoP	   (based	  on	  orthophosphate	  units),	  after	  which	  normal	  human	  serum	  was	  added	  to	   the	  cells	   in	   PBS	   at	   a	   final	   concentration	   of	   5%	   for	   CEC	   and	   25%	   cells	   for	   RPE	   cells	   to	   sufficiently	  induce	  complement	  activation.	  After	  1	  hour,	  the	  cells	  were	  washed,	  and	  the	  amount	  of	  C5b-­‐9	  deposition	  on	  the	  surface	  of	  the	  cells	  was	  quantified	  by	  flow	  cytometry.	  	  Figure	  16	  is	  representative	  of	  several	  experiments,	  each	  performed	  in	  triplicate.	  As	  can	  be	  seen,	  heat-­‐inactivated	   serum	   (HIS)	   did	   not	   induce	   C5b-­‐9	   deposition.	  When	   C7-­‐deficient	   serum	  was	  used,	   no	   C5b-­‐9	   was	   detected.	   Additional	   negative	   controls	   included	   a	   non-­‐specific	   isotype-­‐matched	  primary	  antibody	  (Figure	  16b,	  d).	  	  As	  predicted,	  polyP130	  suppressed	  C5b-­‐9	  deposition	  on	  the	  cells	   in	  a	  concentration-­‐dependent	  manner.	   MonoP	   also	   reduced	   C5b-­‐9	   deposition	   in	   a	   dose-­‐dependent	   manner,	   but	   not	   to	   a	  significant	   extent.	   A	   suppressive	   effect	   of	   monoP	   on	   total	   hemolytic	   activity	   would	   not	   be	  surprising,	   as	   this	  was	  previously	   observed	   in	   our	   study	   reported	   in	  Blood	   [57]	   particularly	   at	  concentrations	   exceeding	   ~250	   µM,	   at	   a	   level	   where	   critical	   complement-­‐dependent	   cations	  (Ca2+,	  Mg2+)	  may	  be	  chelated.	  In	  any	  case,	  the	  suppressive	  effect	  of	  polyP130	  on	  C5b-­‐9	  deposition	  was	   significantly	   greater	   than	   with	   monoP	   for	   the	   CECs	   at	   most	   concentrations,	   and	   at	   the	  second	  highest	   concentration	   in	   the	  RPE	   cells.	   Based	  on	  our	   findings,	   the	  protective	  effect	  of	  polyP130	  in	  this	  assay	  system	  was	  more	  prominent	  in	  the	  CEC	  than	  in	  the	  RPE	  cells.	  	  	  	   54	  	  	  (Legend	  for	  above	  panels	  of	  Figure	  16a,	  b	  on	  the	  next	  page)	  	  	  	  	  	  	  	  	  	  	   55	  	  Figure	  16:	  Flow	   cytometry	   to	  detect	  C5b-­‐9	  deposition	  on	  CEC	  and	  RPE	   cells.	  C5b-­‐9	  deposition	  on	  cell	  membranes	  was	  quantified	  on	  CEC	   (a,	  b)	  and	  RPE	   (c,	  d)	   cell	   lines	  after	  pre-­‐incubation	  with	  polyP130	  or	  monoP	  followed	  by	  exposure	  to	  human	  serum.	  The	  FACS	  profiles	  (a,	  c)	  provide	  a	  visual	  representation	  of	  the	  relative	  abundance	  of	  C5b-­‐9	  deposition	  on	  the	  cells.	  Shown	  here	  are	  representative	  profiles	  from	  the	  highest	  polyP130	  and	  monoP	  concentrations.	  For	  both	  cell	  lines,	  treatment	  with	  polyP130	  results	  in	  a	  shift	  to	  the	   left	  compared	  to	  the	  monoP	  and	  serum	  only	  control	  samples,	   indicating	  that	  polyP130	   interferes	  with	  C5b-­‐9	  deposition	  and	   thus	  complement	  mediated	  attack.	  The	  bar	  graphs	   (b,	  d)	   show	  the	  average	  intensity	  of	  FITC,	  the	  marker	  for	  C5b-­‐9	  on	  the	  cell	  surface,	  for	  all	  the	  samples	  tested.	  Grey	  bars	  represent	  the	  negative	  (isotype	  control,	  HIS:	  heat	  inactivated	  serum,	  and	  C7-­‐depleted	  serum)	  and	  positive	  (serum	  only)	   controls,	   while	   coloured	   bars	   represent	   treated	   samples.	   Both	   cell-­‐lines	   show	   a	   concentration-­‐dependent	   suppression	   of	   C5b-­‐9	   deposition	   with	   polyP130	   and	   monoP	   treatment,	   although	   polyP130	  treatments	  result	   in	  overall	  greater	  suppression	  for	  all	  concentrations.	  Error	  bars	  reflect	  standard	  error	  of	  the	  mean,	  with	  n=3.	  One	  way	  ANOVA	  was	  used	  to	  assess	  significance.	  *	  P	  ≤	  0.05;	  **	  P	  ≤	  0.01;	  ***	  P	  ≤	  0.001.	  	  	  	  	  	   56	  5.3.2 Effects	  of	  PolyP130	  on	  Release	  of	  SC5b-­‐9	  from	  CEC	  and	  RPE	  Cells	  Soluble	   C5b-­‐9	   (or	   SC5b-­‐9)	   is	   commonly	   used	   as	   a	   marker	   of	   complement	   activation	   and	  complement-­‐mediated	   cell	   damage.	   Plasma	   levels	   of	   SC5b-­‐9	   may	   be	   elevated	   in	   diseases	  associated	   with	   excess	   complement	   activation.	   SC5b-­‐9	   is	   readily	   measured	   by	   a	   sensitive	  commercial	  sandwich	  ELISA	  in	  which	  the	  detecting	  antibody	  identifies	  an	  epitope	  that	  is	  specific	  for	   the	   SC5b-­‐9	   complex.	   Released	   SC5b-­‐9	   is	   known	   to	   induce	   membrane	   changes	   and	  intracellular	   signaling	   in	   platelets	   and	   endothelial	   cells	   that	   promote	   inflammation	   and	  coagulation.	  Based	  on	  the	  preceding	  results,	  we	  predicted	  that	  polyP	  would	  dampen	  the	  generation	  of	  SC5b-­‐9	  in	  cells	  exposed	  to	  serum.	  This	  was	  tested	  as	  above	  by	  pre-­‐incubating	  the	  CEC	  and	  RPE	  cells	  with	   polyP130	   or	   monoP,	   and	   then	   exposing	   them	   to	   serum	   for	   1	   hour.	   Supernatants	   were	  collected	  and	  SC5b-­‐9	  was	  measured	  (Figure	  17).	  Release	  of	  SC5b-­‐9	   into	  the	  media	  was	  readily	  detected	  in	  cells	  exposed	  to	  normal	  serum.	  Exposure	  of	  the	  cells	  to	  heat-­‐inactivated	  serum	  or	  serum	  deficient	  in	  C7	  resulted	  in	  the	  release	  of	  minimal	  (almost	  undetectable)	  amounts	  of	  SC5b-­‐9.	  Pre-­‐incubation	  of	  RPE	  cells	  with	  polyP130	  resulted	  in	  only	  a	  slight	  reduction	  in	  SC5b-­‐9	  that	  was	  not	  different	  when	  compared	  to	  monoP	  (Figure	  17).	  The	  findings	  were	  similar	  with	  the	  CEC	   in	  that	   there	   was	   no	   change	   in	   the	   release/generation	   of	   SC5b-­‐9	   (Figure	   17)	   following	   pre-­‐incubation	  with	  monoP	  or	  polyP130.	  Overall,	  although	  polyP130	  caused	  a	  significant	  reduction	  in	  serum-­‐induced	   cell	   surface	   deposition	   of	   C5b-­‐9,	   generation	   and	   release	   of	   SC5b-­‐9	   was	   not	  affected.	  	  	   57	  	  Figure	  17:	  	  Measurement	  of	  SC5b-­‐9	  formation	  with	  ELISA.	  Following	  exposure	  of	  CEC	  (a)	  and	  RPE	  cells	  (b)	   to	   serum,	   SC5b-­‐9	   released	   into	   the	   media	   was	   quantified.	   Experiments	   performed	   in	   duplicate,	  consisted	  of	  varying	  concentrations	  of	  polyP130	  (red)	  and	  monoP	  (beige),	  as	  well	  as	  negative	  and	  positive	  controls	  as	  noted	  (grey	  –	  C7-­‐depleted	  serum;	  HIS-­‐	  heat	  inactivated	  serum).	  From	  these	  studies,	  we	  could	  not	  discern	  any	  change	  in	  the	  release	  of	  SC5b-­‐9	  in	  response	  to	  polyP130	  or	  monoP	  at	  the	  concentrations	  tested.	  a"b"	  	   58	  5.3.3 PolyP130	  and	  Oxidative	  Stress	  In	   addition	   to	  excess	   complement	  activation,	  oxidative	   stress	   is	  believed	   to	  play	  a	   key	   role	   in	  promoting	  the	  cellular	  damage	  that	   is	  associated	  with	  the	  development	  of	  AMD.	  Although	  we	  have	  evidence	  that	  polyP	  suppresses	  complement	  activation	  and	  protects	  against	  complement	  mediated	  cellular	  (RPE,	  CEC)	  damage,	  we	  have	  not	  excluded	  the	  possibility	  that	  polyP	  has	  other	  protective	   effects	   that	   may	   translate	   into	   in	   vivo	   protection.	   We	   therefore	   explored	   the	  possibility	  that	  polyP	  protects	  CEC	  and	  RPE	  cells	  against	  oxidative	  stress.	  This	  was	  accomplished	  by	  exposing	  cells	  to	  different	  concentrations	  of	  H2O2	  for	  varying	  periods	  of	  time,	  in	  the	  presence	  or	  absence	  of	  a	  range	  of	  concentrations	  of	  polyP130	  or	  monoP.	  	  Three	  independent	  approaches	  were	  used,	  assessing	  different	  endpoints	  for	  each.	  Serum	  was	  not	  present	  in	  these	  experimental	  setups,	  thereby	  excluding	  the	  participation	  of	  complement.	  5.3.4 Morphologic	  Evidence	  that	  PolyP	  Protects	  Cells	  Against	  Oxidative	  stress	  RPE	  cells	  were	  exposed	  to	  a	  range	  of	  concentrations	  of	  polyP130	   for	  24	  hours,	   in	  the	  presence	  (Figure	   18a,	   top)	   or	   absence	   (Figure	   18a,	   bottom)	   of	   H2O2.	   Changes	   in	   morphology	   were	  observed	   under	   a	   phase-­‐contrast	   microscope.	   In	   the	   absence	   of	   polyP130	   or	   with	   the	   lowest	  concentration	   of	   polyP130,	   the	   cells	   exhibited	   major	   morphologic	   changes,	   including	   cell	  shrinkage,	   rounding,	   loss	   of	   intercellular	   interactions,	   blebbing	   of	   the	   cytoplasm,	   and	  transparent	   nuclei	   (Figure	   18a,	   top	   right-­‐most	   panel).	   When	   the	   cells	   were	   simultaneously	  treated	  with	   polyP130,	   these	   apparently	   damaging	   effects	  were	   almost	   entirely	   abrogated,	   an	  effect	  that	  was	  dose-­‐dependent	  (Figure	  18a,	  top	  panels).	   In	  the	  absence	  of	  H2O2,	  polyP130	  had	  no	   detectable	   effects	   on	   the	   cells	   (Figure	   18a,	   bottom	   panels).	   Interestingly,	   monoP	   at	  equivalent	   concentrations	   as	   the	   polyP130	   also	   exhibited	   a	   dose-­‐dependent	   protective	   effect	  against	  H2O2-­‐induced	  oxidative	  stress	  (Figure	  18b,	  top).	  	  CEC	  were	  also	  exposed	  to	  H2O2-­‐induced	  oxidative	  stress	  in	  the	  presence	  or	  absence	  of	  polyP130	  and	   monoP	   (Figure	   18d	   and	   18e,	   top)	   as	   with	   the	   RPE	   cells.	   Although	   the	   formation	   of	   a	  precipitate	  (likely	  from	  the	  concentrated	  polyP130)	  slightly	  obscured	  the	  view	  of	  the	  cells	  at	  the	  	  	   59	  higher	  concentrations,	  we	  were	  able	  to	  detect	  a	  similar	  response	  of	  the	  CEC	  to	  the	  polyP130	  and	  monoP,	  as	  with	  the	  RPE	  cells	  (Figure	  18d	  and	  18e,	  bottom).	  Overall,	   the	   preceding	   studies	   revealed	   qualitative	   evidence	   that	   both	   polyP130	   and	   monoP	  protect	  CEC	  and	  RPE	  cells	  from	  H2O2-­‐induced	  oxidative	  stress.	  We	  further	  sought	  to	  validate	  the	  findings	  via	  the	  following	  more	  quantitative	  approaches.	  	  	  	  	  (Legend	  for	  above	  Figure	  18a-­‐c	  on	  next	  page)	  	  	   60	  	  	  	  	  Figure	   18:	   Cell	   morphologic	   changes	   in	   response	   to	   oxidative	   damage	   and	   polyP130.	   Representative	  images	  of	  RPE	  cells	  (ARPE-­‐19)	  (a,	  b)	  and	  CEC	  (RF/6A)	  (d,	  e)	  that	  were	  exposed	  for	  24	  hrs	  to	  500	  μM	  H2O2	  are	   shown.	   For	   both	   cell	   lines,	   H2O2	   induces	   dramatic	   cell	   morphologic	   changes,	   with	   rounding,	   less	  adherence	   to	   the	   surface,	   and	   loss	   of	   cell-­‐cell	   interactions	   (c	   and	   f,	   left	   panels),	   compared	   with	   the	  untreated	  cells	   (c	  and	  f,	   right	  panels).	  With	  the	  addition	  of	   increasing	  concentrations	  of	  polyP130	   in	  the	  presence	  of	  the	  H2O2,	  these	  morphologic	  changes	  were	  notably	  reduced	  (a,	  top	  panels),	  and	  cells	  took	  on	  their	   normal	   appearance,	   i.e.,	   similar	   to	  without	  H2O2	  (a,	   bottom	  panels).	  Neither	  monoP	   nor	   polyP130	  alone	  (without	  H2O2)	  had	  an	  effect	  on	  the	  RPE	  (a,	  b,	  bottom	  panels).	  The	  highest	  polyP130	  concentrations	  resulted	   in	   the	   formation	   of	   a	   precipitate	   on	   the	   CEC	   (d),	  making	   it	   difficult	   to	   detect	   changes	   in	   the	  morphology	  of	  the	  cells.	  However,	  all	  polyP130	  concentrations	  equal	  to	  or	  above	  500	  μM	  exhibited	  clear	  a	  protective	   effect	   (d,	   top	   panel).	   The	   monoP	   treatments	   with	   the	   CEC	   also	   revealed	   evidence	   of	  protection	  against	  H2O2-­‐induced	  cell	  damage	  (e,	  top	  panels),	  similar	  to	  the	  effect	  in	  RPE.	  The	  images	  are	  representative	  of	  triplicates	  from	  each	  treatment,	  and	  are	  shown	  at	  the	  same	  magnification.	  	  	   61	  5.3.5 PolyP	  does	  not	  alter	  the	  Integrity	  of	  H2O2	  The	  preceding	  studies	  suggested	  that	  polyP	  and	  monoP	  have	  anti-­‐oxidant	  properties.	  This	  may	  be	   achieved	   via	   several	   mechanisms,	   including	   induction	   of	   anti-­‐oxidants,	   such	   as	   catalase,	  glutathione,	  peroxidase,	  and/or	  superoxide	  dismutase	  [151,	  152].	  There	  is	  a	  precedent	  for	  this,	  at	  least	  in	  bacteria,	  where	  polyP	  induces	  expression	  of	  the	  antioxidant,	  catalase	  [17].	  However,	  it	  is	  also	  possible	  that	  the	  polyP	  and	  monoP	  are	  simply	  degrading	  or	  neutralizing	  the	  H2O2	  prior	  to	  inducing	  cellular	  damage.	  	  We	  therefore	  used	  an	  in	  vitro	  approach	  to	  test	  whether	  polyP130	  disrupts	  the	  integrity	  of	  H2O2.	  We	   first	   showed	   that	   H2O2	   has	   a	   measurable	   absorbance	   at	   240	   nm	   that	   is	   only	   minimally	  altered	  by	  the	  presence	  of	  polyP130	  (Figure	  19a).	   	  The	  integrity	  of	  H2O2	  could	  then	  be	  assessed	  by	   monitoring	   the	   reduction	   in	   absorbance	   when	   it	   is	   incubated	   with	   catalase,	   thereby	  converting	  it	  to	  H2O	  +	  O2.	  We	  confirmed	  that	  this	  occurs	  within	  ~2-­‐3	  minutes	  (Figure	  19a).	  The	  effect	  of	  1	  mM	  polyP130	  on	  1	  mM	  H2O2	  over	  5	  minutes	  was	  similarly	  studied	  by	  monitoring	  the	  change	  in	  absorbance.	  polyP130	  has	  no	  effect	  on	  the	  absorbance	  of	  the	  H2O2	  over	  a	  period	  of	  5	  minutes,	  and	  up	  to	  24	  hours	  (not	  shown).	  The	  residual	  integrity	  of	  the	  H2O2	  after	  24	  hours	  was	  confirmed	  by	   incubating	   the	  polyP130/H2O2	  with	  purified	   catalase	   (b).	   At	   that	   time,	   there	  was	  again	   a	   rapid	   decrease	   in	   absorbance	   (Figure	   19b),	   as	   the	   H2O2	   was	   converted	   to	   H2O	   with	  release	   of	   O2.	   The	   findings	   exclude	   the	   possibility	   that	   polyP130	   is	   directly	   destabilizing	   or	  inactivating	   the	   H2O2,	   and	   suggested	   that	   polyP130	   (and	  monoP)	   have	   anti-­‐oxidant	   properties	  mediated	  via	  other	  mechanisms.	  	  	  	   62	  	  Figure	  19:	  Effect	  of	  polyP130	  on	  functional	   integrity	  H2O2.	  (a)	  The	  absorbance	  at	  240	  nM	  of	  polyP130	  and	  H2O2	   was	  measured	   over	   time.	   Co-­‐incubation	   of	   H2O2	   with	   polyP130	   over	   24	   hours	   did	   not	   cause	   any	  change	  in	  absorbance	  (only	  5	  minutes	  shown).	  As	  a	  positive	  control,	  the	  addition	  of	  catalase	  to	  the	  H2O2	  (a:	  red),	  caused	  a	  rapid	  decrease	  in	  absorbance,	  indicative	  of	  the	  conversion	  of	  H2O2	  to	  H2O	  and	  O2.	  	  (b)	  After	   24	   hours,	   when	   the	   absorbance	   of	   the	   combined	   H2O2+polyP130	   had	   not	   changed,	   catalase	   was	  added	   to	   the	  mixture.	   The	  absorbance	  again	  dropped	   rapidly	   (b:	   purple	   line),	   indicating	   that	   the	  H2O2	  remained	  intact	  after	  24	  hours	  while	  in	  the	  presence	  of	  the	  polyP130.	  	  	  	  	   63	  5.3.6 Effect	  of	  of	  PolyP130	  and	  monoP	  on	  Cellular	  Expression	  of	  Catalase	  We	  examined	  the	  effect	  of	  polyP130	  on	  the	  expression	  of	  catalase,	  a	  major	  cellular	  anti-­‐oxidant	  enzyme	   that	   degrades	   H2O2	   [152].	   RPE	   cells	   were	   exposed	   to	   a	   range	   of	   concentrations	   of	  polyP130	  for	  24	  hours.	  Cell	  lysates	  were	  prepared	  and	  equal	  amounts	  were	  assayed	  for	  catalase	  activity,	   quantified	   from	  a	   standard	   curve	  of	   purified	   catalase.	   The	   cells	  were	  not	   exposed	   to	  H2O2,	   so	   as	   to	   avoid	   confounding	   the	   interpretation	   of	   the	   assays	   results.	   There	   was	   no	  consistent	  change	  in	  catalase	  activity	  in	  response	  to	  polyP	  as	  compared	  to	  the	  untreated	  control	  (Figure	  20a).	  MonoP	  at	  the	  highest	  concentration	  as	  that	  of	  polyP	  also	  had	  no	  apparent	  effect	  on	   catalase	   generation.	   Similarly,	   the	   same	   treatments	   for	   the	   CEC	   cells	   did	   not	   induce	   any	  change	   in	  catalase	  activity	   (Figure	  20b)	   for	  any	  of	   the	   treatment	  concentrations,	  compared	  to	  the	  untreated	  control	  sample.	  Unlike	  the	  RPE	  cells,	  however,	  the	  absolute	  catalase	  activity	  level	  was	  much	  lower	  for	  the	  CEC	  than	  for	  the	  RPE	  cells	  by	  ~2.5	  fold.	  Again,	  monoP	  also	  had	  no	  effect.	  These	  results	  were	  verified	  by	  measuring	  catalase	  gene	  expression	  by	  qRT-­‐PCR	  (Figure	  21).	  Cells	  were	  treated	  in	  an	  identical	  manner	  as	  for	  catalase	  activity.	  The	  qRT-­‐PCR	  results	  revealed	  that	  polyP130	  had	  no	  apparent	  effect	  on	  catalase	  gene	  expression	  for	  either	  RPE	  cells	  or	  CEC.	  	  	  	  	  	   64	  	  Figure	  20:	  Effect	  of	  polyP130	  and	  monoP	  on	  catalase	  activity	  (CAT).	  The	  values	  in	  the	  above	  graphs	  were	  determined	   from	   a	   standard	   curve	   using	   purified	   catalase	   with	   known	   catalytic	   activities,	   and	   the	  samples	  were	  normalized	  for	  total	  protein	  content.	  There	  was	  no	  consistent	  polyP-­‐dependent	  change	  in	  catalase	  activity	  in	  either	  the	  RPE	  cells	  (a)	  or	  the	  CEC	  (b).	  A	  high	  concentration	  of	  monoP	  (grey	  bar)	  does	  not	  induce	  catalase	  activity	  in	  these	  cells.	  Experiments	  were	  performed	  in	  duplicate.	  	  	  Figure	  21:	  Effect	   of	   polyP130	   and	  monoP	   on	   catalase	   gene	   expression.	   qRT-­‐PCR	  was	  used	   to	  measure	  catalase	  gene	  expression	  following	  exposure	  of	  RPE	  cells	  (a)	  and	  CEC	  (b)	  to	  polyP130	  and	  monoP.	  Values	  were	  normalized	  to	  samples	  without	  polyP130	  (untreated),	  designated	  a	  relative	  gene-­‐expression	  of	  1.	  No	  significant	   change	   in	   expression	   was	   observed	   for	   either	   the	   polyP130	   or	   monoP	   treated	   cells.	  Experiments	  were	  performed	  in	  duplicate.	  	  	  	   65	  5.3.7 PolyP	  Protects	  Nuclear	  Integrity	  of	  RPE	  cells	  and	  CEC	  from	  Oxidative	  Stress	  Direct	  visualization	  of	  the	  cells	  by	  phase-­‐contrast	  microscopy	  provided	  evidence	  that	  polyP	  and	  monoP	   exert	   a	   cytoprotective	   effect	   that	   is	   independent	   of	   complement.	   However,	   these	  studies	  were	  not	  quantitative.	  Due	  to	  the	  apparent	  difference	  in	  nuclear	  transparency	  between	  the	  treatments,	  nuclear	  integrity	  by	  nuclear	  staining	  was	  chosen	  as	  a	  means	  of	  quantifying	  the	  effects	  of	  polyP130	  and	  monoP	  on	  H2O2-­‐induced	  cellular	  damage.	  Hoechst	  dye	  was	  chosen	  as	  the	  stain	  of	   choice	  as	   it	   is	   able	   to	   cross	   the	   lipid	  bilayer	  of	   cells	   and	   to	  bind	   to	   intact	  DNA	   in	   the	  nuclei.	  Although	  the	  Hoechst	  dye	  does	  not	  distinguish	  between	  live	  and	  dead	  cells,	   it	  provides	  an	  indirect	  indication	  of	  cell	  viability,	  as	  oxidative	  stress	  is	  known	  to	  cause	  cell	  death,	  largely	  by	  damaging	  the	  DNA	  via	  reactive	  oxygen	  species	  [153,	  154].	  As	  shown	  in	  Figure	  22a,	  increasing	  concentrations	  of	  H2O2	  resulted	  in	  decreased	  nuclear	  staining	  of	  CEC,	   findings	  that	  are	  consistent	  with	  decreased	  cell	  viability	   [155].	  At	  all	  concentrations	  of	  H2O2,	  co-­‐incubation	  of	  the	  cells	  with	  polyP130	  augmented	  nuclear	  staining,	  and	  this	  appeared	  to	  occur	   in	  a	  dose-­‐dependent	  manner.	  The	  results	  are	   in	   line	  with	  what	  was	  observed	  under	  the	  phase-­‐contrast	  microscope.	  However,	  in	  contrast	  with	  the	  morphology	  findings	  shown	  in	  Figure	  18,	  monoP	  at	  the	  highest	  concentration	  (1000	  µM)	  did	  not	  protect	  the	  cells	  in	  terms	  of	  nuclear	  integrity	   (not	   shown).	   Similar	   findings	   were	   obtained	   with	   RPE	   (Figure	   22b).	   Interestingly,	  nuclear	  staining	  with	  the	  polyP130-­‐exposed	  cells	  appeared	  to	  be	  increased,	  at	  least	  for	  the	  CEC,	  suggesting	  that	  polyP130	  may	  induce	  cell	  proliferation	  and/or	  prevent	  cell	  death.	  Overall,	  these	  data	   are	   evidence	   that	   polyP130	   protects	   the	   nuclear	   integrity	   of	   the	   cells	   from	  H2O2-­‐induced	  oxidative	  damage.	  	  	  	   66	  	  Figure	  22:	  Protective	  effect	  of	  polyP130	  on	  nuclear	  integrity	  following	  cell	  exposure	  to	  oxidative	  stress.	  Cells	  were	  exposed	  to	  H2O2	  and	  polyP130	  as	  described	  in	  the	  text.	  Without	  polyP130	  treatment	  (a),	  nuclear	  integrity	   in	  CEC	  was	   reduced	  with	   increasing	  H2O2	  concentrations	   (beige	   line).	  The	  addition	  of	  polyP130	  reversed	   this	   effect	   in	   a	   concentration-­‐dependent	   manner,	   with	   the	   higher	   polyP	   concentrations	  resulting	   in	  protection	  of	  nuclear	   integrity.	   The	   response	  of	  RPE	   cells	   to	  polyP130	   (b)	  was	   similar.	   Error	  bars	   are	   not	   shown.	   Analyses	   were	   performed	   with	   a	   One	   Way	   ANOVA	   (significance	   not	   shown	   for	  samples	  treated	  with	  ≥	  500	  μM	  H2O2).	  n	  =	  3,	  **	  P	  ≤	  0.01;	  ***	  P	  ≤	  0.001.	  	  5.3.8 Effect	  of	  PolyP130	  on	  Oxidative	  Stress-­‐Induced	  Changes	  in	  Junctional	  Protein	  Integrity	  	  The	   integrity	  of	   the	  RPE	  monolayer	   is	  vital	   to	   the	  blood-­‐retinal	  barrier,	  as	   the	   latter	   regulates	  paracellular	  diffusion	  of	  nutrients	  and	  waste	  with	  the	  choroid	  [156,	  157].	  Disruption	  of	  the	  RPE	  monolayer	   due	   to	   disruption	   of	   intercellular	   junctions	   allows	   CNV	   to	   proceed,	   facilitates	   an	  inflammatory	   response	  with	   invasion	   of	  macrophages,	   and	   allows	   leakage	   of	   choroidal	   fluids	  into	  the	  sub-­‐retinal	  space	  [158,	  159].	  The	  choriocapillaries,	  comprising	  CEC,	  are	  fenestrated	  to	  facilitate	   exchange	   of	   nutrients	   and	   waste	   with	   the	   photoreceptors	   via	   the	   RPE	   [160].	   Tight	  intercelluar	  junctions	  are	  also	  present	  in	  CEC,	  and	  although	  not	  as	  well	  characterized	  as	  those	  of	  RPE	   cells,	   they	   are	   believed	   to	   play	   a	   role	   in	   AMD	   [160,	   161].	   Oxidative	   stress	   is	   known	   to	  damage	  intercellular	  junctions	  [158,	  162].	  Our	  observations	  of	  cellular	  protection	  with	  polyP130	  	  	   67	  provided	  reason	  to	  suspect	  that	  it	  may	  also	  protect	  the	  intercellular	  junctions	  of	  these	  cells	  from	  H2O2-­‐induced	  damage.	  	  We	   tested	   the	   response	  of	  CEC	  expression	  of	  VE-­‐cadherin	   to	  H2O2-­‐induced	  oxidative	   stress	   in	  the	  presence	  and	  absence	  of	  polyP130	  and	  monoP.	   	  For	  quantitative	  purposes,	  we	   imaged	  the	  cells	  with	  the	  Cellomics	  ArrayScan	  high	  throughput	  confocal	  microscope,	  which	  was	  optimized	  for	  quantifying	  VE-­‐cadherin	  on	  CEC.	  As	  seen	  in	  a	  representative	  experiment	  shown	  in	  Figure	  23,	  CEC	  were	  treated	  with	  varying	  concentrations	  of	  polyP130	  or	  monoP,	  in	  the	  presence/absence	  of	  H2O2,	   after	   which	   the	   cells	   were	   fixed	   and	   stained	   for	   expression	   of	   VE-­‐cadherin	   using	   a	  fluorescent-­‐tagged	   secondary	   antibody.	   The	   Cellomics	   ArrayScan	   high	   content	   confocal	  microscope	  was	   thus	  used	   to	  quantify	   the	   fluorescence	   in	  each	  sample.	  All	  experiments	  were	  performed	  in	  duplicate,	  and	  for	  each	  sample	  15	  fields	  were	  quantified,	  from	  which	  an	  average	  signal	  was	  obtained.	  In	   the	  absence	  of	  H2O2,	  polyP130	   induced	  expression	  of	  VE-­‐cadherin	  at	  a	  concentration	  of	  100	  µM.	  With	  the	  addition	  of	  10	  µM	  H2O2	  and	  in	  the	  absence	  of	  polyP,	  there	  was	  no	  evidence	  of	  an	  effect	  on	  VE-­‐cadherin.	  However,	  polyP	  at	  concentrations	  of	  10	  µM,	  100	  µM	  and	  500	  µM,	  caused	  an	  increase	  in	  VE-­‐cadherin	  expression,	  although	  it	  was	  only	  significantly	  augmented	  at	  100	  µM.	  A	  similar	  trend	  was	  observed	  with	  100	  and	  500	  µM	  H2O2.	  At	  1	  mM	  H2O2,	  increasing	  amounts	  of	  polyP130	  resulted	  in	  increased	  VE-­‐cadherin	  expression	  as	  well.	  Under	  all	  H2O2	  conditions,	  monoP	  at	  the	  high	  concentration	  used,	  had	  little	  or	  no	  effect	  on	  VE-­‐cadherin	  expression.	  	  	  	   68	  02004006008001000*0 0.01 0.1 0.5 10µM P13010 µM PolyP100 µM PolyP500 µM PolyP1000 µM PolyP1000 µM MonoP*[H2O2] mMVE-cadherin (Avg FITC intensity per cell)	  Figure	  23:	  Effect	  of	  polyP130	  on	  VE-­‐cadherin	  expression.	  VE-­‐Cadherin	  immunofluorescent	  staining	  on	  CEC	  cells	   was	   quantified	   by	   high	   content	   confocal	   microscopy.	   polyP130	   exhibits	   varying	   effects	   on	   VE-­‐cadherin	   expression	   following	   exposure	   of	   cells	   to	   oxidative	   stress.	   At	   1	   mM	   of	   H2O2,	   increasing	  concentrations	  of	  polyP130	  (red	  bars)	  protects	  against	  VE-­‐cadherin	  loss	  per	  cell.	  At	  lower	  concentrations	  and	  at	  no	  H2O2,	  there	  is	  an	  induction	  of	  VE-­‐cadherin	  expression	  up	  to	  100µM	  polyP130,	  and	  a	  subsequent	  decrease	   in	  expression	  at	  higher	  concentrations.	  MonoP	  (blue	  bars)	  at	  a	  high	  concentration	  (1000	  µM)	  had	  no	  apparent	  effect	  on	  VE-­‐cadherin	  expression,	  compared	  to	  the	  untreated	  control	  (white	  bar).	  One	  Way	  ANOVA	  was	  used	  to	  analyze	  significance	  per	  H2O2	  concentration	  group.	  *P	  ≤	  0.05.	  	  	  	  	  	  	  	   69	  5.3.9 Effect	  of	  PolyP130	  of	  Endothelial	  Cell	  Proliferation	  Although	  the	  mechanisms	  are	  not	  fully	  delineated,	  the	  preceding	  studies	  provide	  evidence	  that	  polyP	  (and	  possibly	  monoP)	  exhibit	  cytoprotective	  properties	  under	  some	  stress	  conditions	  that	  are	   linked	   to	  AMD.	   These	  data	   suggested	   that	   polyP	  may	   also	   induce	   cell	   proliferation	  under	  non-­‐stress	  conditions.	  To	  test	  this,	  we	  used	  EA.hy	  926	  cells,	  a	  human	  cell	  line	  that	  is	  a	  hybrid	  of	  human	  umbilical	  vein	  endothelial	  cells	  and	  A549	   lung	  cancer	  epidermal	  cells.	  The	   line	  exhibits	  phenotypic	   features	   that	  most	   closely	   resemble	   endothelial	   cells.	   They	   are	   robust,	   and	   grow	  easily	   in	   culture	   [163].	   We	   selected	   this	   cell	   line,	   rather	   than	   the	   CEC,	   as	   we	   were	   having	  difficulties	  maintaining	  the	  latter	  cell	  line	  at	  the	  time	  of	  these	  studies.	  We	   assessed	   the	   growth	   response	   of	   these	   cells	   to	   different	   concentrations	   of	   polyP130	   and	  monoP	  by	  culturing	  them	  from	  low	  density	  and	  allowing	  them	  to	  grow	  in	  growth	  media	  for	  up	  to	  4	  days.	  Cells	  were	  directly	  visually	  counted	  with	  a	  haemocytometer.	  Although	  the	  cell	  counts	  did	  indeed	  increase	  over	  time,	  there	  were	  no	  significant	  differences	  between	  those	  exposed	  to	  monoP,	  polyP130	  or	  media	  only	  (Figure	  24a).	  A	  similar	  effect	  was	  found	  with	  CEC	  in	  a	  preliminary	  study	  (not	  shown).	  We	  also	  quantified	   EA.hy	   926	   cell	   proliferation	  by	  measuring	  BrdU	   incorporation	   after	   24	  hrs	  exposure	  to	  polyP130	  (100	  µM	  -­‐	  1	  mM)	  or	  buffer	  alone.	  Again,	  we	  could	  not	  detect	  a	  significant	  effect	   of	   polyP130	   on	   cellular	   DNA	   incorporation	   of	   BrdU	   (Figure	   24b).	   Overall,	   under	   these	  experimental	  conditions,	  polyP130	  had	  no	  effect	  on	  cell	  proliferation.	  	  	  	  	  	  	  	   70	  	  	  Figure	  24:	  Effect	  of	  polyP130	  and	  monoP	  on	  cell	  proliferation.	  (a)	  EA.hy	  926	  cells	  were	  cultured	  in	  96-­‐well	  plates	  in	  the	  presence	  of	  varying	  concentrations	  of	  monoP	  or	  polyP130.	  Neither	  polyP130	  nor	  monoP	  had	  any	  effect	  on	  cell	  number	  over	  time.	  (b)	  BrdU	  incorporation	  for	  detection	  of	  newly	  synthesized	  DNA	  was	  measured	   in	   cells	   cultured	   with	   varying	   concentrations	   of	   polyP130	   and	   seeded	   at	   2	   densities.	   Again,	  polyP130	  did	  not	  exhibit	  any	  effect	  on	  cell	  proliferation.	  One	  way	  ANOVA	  was	  used	  to	  assess	  significance,	  n=3.	  Error	  bars	  not	  shown	  for	  (a).	  b"	  	   71	  Chapter	  6: Discussion	  and	  Future	  Directions	  In	  this	  thesis,	  I	  describe	  novel	  findings	  on	  the	  role	  of	  polyphosphate	  as	  a	  negative	  regulator	  of	  complement	   and	   an	   anti-­‐oxidant,	   and	   apply	   this	   new	   knowledge	   in	   rodent	  models	   toward	   a	  treatment	   for	   a	   common	   and	   serious	   complement-­‐mediated	   cause	   of	   blindness,	   age	   related	  macular	  degeneration	  (AMD).	  6.1	  	   PolyP:	  Complex	  Roles	  in	  Coagulation	  and	  Complement	  In	   response	   to	   injury,	   coagulation	   and	   complement	   are	   simultaneously	   activated	   to	   restrict	  bleeding	  and	  to	  discard	  invading	  pathogens	  and	  damaged	  cells,	  thereby	  protecting	  the	  organism	  from	  death,	  and	  facilitating	  healing.	  These	  two	  systems	  therefore	  act	  in	  concert,	  temporally	  and	  spatially	  [1,	  164,	  165].	  The	  findings	  presented	  in	  my	  thesis	  appear	  to	  conflict	  with	  this	  paradigm,	  with	  polyP	  acting	  as	  a	  promoter	  of	  coagulation	  and	  an	  inhibitor	  of	  complement.	  The	  explanation	  for	  these	  findings	  is	  not	  clear	  at	  this	  time,	  and	  we	  can	  only	  speculate	  as	  to	  why	  polyP	  has	  these	  diverse	   properties.	   However,	   polyP	   is	   not	   alone	   in	   exhibiting	   complex,	   and	   sometimes	  apparently	   opposing	   activities.	   The	   best	   example	   is	   thrombin,	   which	   is	   known	   to	   be	   a	  procoagulant,	  anti-­‐coagulant,	  and	  pro-­‐inflammatory	  [166-­‐169].	  Thrombin	  achieves	  this	  through	  allosteric	  modifications	  in	  its	  conformation	  and	  distinct	  interactions	  with	  its	  different	  substrates,	  influenced	   by	   the	   demands	   of	   homeostasis	   [166,	   167].	   Thus,	   thrombin	   promotes	   clotting	   by	  activating	  platelets	  and	  endothelial	  cells,	  and	  through	  cleavage	  of	  factor	  V,	  factor	  XI,	  factor	  XIII	  and	   fibrinogen.	   It	   also	   acts	   as	   an	   anticoagulant	   by	   binding	   to	   thrombomodulin	   and	   activating	  protein	  C,	  as	  well	  as	  promotes	   inflammation	  by	  augmenting	  expression	  of	  adhesion	  molecules	  on	  endothelial	  cells	  and	  inducing	  vascular	  permeability.	  Mast	  cells,	  one	  of	  the	  first	  responders	  of	  inflammation,	   have	   also	   been	   found	   to	   switch	   from	   an	   initially	   fibrinolytic	   phenotype	   to	   a	  prothrombotic	  phenotype	  in	  response	  to	  C5a	  influx	  [170].	  Overall,	  the	  response	  to	  injury	  must	  be	  able	  to	  adapt	  dynamically	  to	  temporal	  changes,	  and	  this	  might	  be	  most	  efficiently	  achieved	  by	  cells,	  proteins	  and	  other	  factors	  that	  possess	  diverse	  properties.	  	  	  	   72	  PolyP	  exhibits	  procoagulant	  and	  complement	  inhibitory	  properties,	  and	  that	  may	  depend	  in	  part	  on	  the	  local	  concentration	  and	  length	  of	  the	  polymer.	  Its	  differential	  function	  also	  will	  likely	  vary	  depending	  on	  environmental	   factors,	   and	   its	   particular	   protein	  partner	   (e.g.,	   thrombin	   versus	  C5b6).	   The	   relative	   affinities	   of	   varying	   lengths	   of	   polyP	   for	   C5b,6	   versus	   thrombin	   have	   not	  been	  studied.	  We	  also	  do	  not	  know	  how	  the	  complement	  inhibitory	  properties	  of	  polyP	  would	  be	   affected	   in	   the	   setting	   of	   a	   procoagulant	   environment,	   versus	   a	   primarily	   inflammatory	  environment.	   We	   could	   imagine,	   however,	   that	   polyP	   may	   serve	   simultaneously	   as	   a	  procoagulant	  and	  anti-­‐complement	  factor:	  following	  platelet	  activation,	  polyP	  is	  released	  from	  dense	   granules,	   whereupon	   it	   may	   coat	   the	   cell,	   protecting	   it	   from	   complement-­‐mediated	  destruction.	   At	   the	   same	   time,	   on	   the	   platelet	   surface,	   it	   may	   enhance	   the	   platelet’s	  procoagulant	  effects.	  Clearly,	  elucidating	  the	  complex	  role	  of	  polyP	   in	  vivo	  will	   require	   further	  study,	  which	  will	  hopefully	  result	   in	  the	  revelation	  of	  new	  therapeutic	  targets	   in	  the	  pathways	  that	  polyP	  affects.	  	  	  6.2	  	   Mechanisms	  by	  Which	  PolyP	  Suppresses	  Complement	  In	  the	  past	  10	  years,	  major	  advances	  have	  been	  made	  in	  elucidating	  the	  mechanisms	  by	  which	  different	  lengths	  of	  polyP	  promote	  coagulation.	  Until	  the	  studies	  in	  the	  Conway	  lab,	  essentially	  nothing	  was	  known	  about	  the	  role	  of	  polyP	  in	  complement.	  Such	  studies	  were	  a	  challenge,	  due	  to	   polyP	   being	   strongly	   anionic,	   complicating	   experiments	   in	   the	   complement	   system,	   since	  many	   of	   the	   steps	   in	   the	   classical,	   lectin,	   and	   alternative	   pathways	   (but	   NOT	   the	   terminal	  pathway)	   are	   dependent	   on	   calcium	   and	   magnesium	   ions.	   Thus,	   it	   has	   been	   difficult	   to	  distinguish	   the	   effect	   of	   polyP	   due	   to	   direct	   interactions	   with	   complement	   proteins	   from	   its	  chelation	  effect.	  	  Nonetheless,	  work	  by	   a	  Research	  Associate	   in	   our	   lab,	  Dr.	   Emilie	   Lameignere,	   has	   resulted	   in	  progress	   in	   this	   endeavor	   as	   she	   uncovered	   an	   ion-­‐independent	   mechanism	   by	   which	   polyP	  modulates	   the	   classical	   pathway	   of	   complement	   (unpublished).	   PolyP	   was	   found	   to	   interact	  directly	   with	   C1s	   and	   the	   C1-­‐esterase	   inhibitor	   (C1-­‐INH),	   augmenting	   the	   activity	   of	   C1-­‐INH,	  thereby	  dampening	  cleavage	  of	  C4	  and	  C2,	  and	  limiting	  the	  formation	  of	  the	  classical	  pathway	  	  	   73	  C4bC2a	  C3	  convertase.	  These	  studies	  are	  ongoing,	  but	  further	  support	  the	  conclusion	  from	  my	  thesis	   that	   polyP	   dampens	   activation	   of	   complement	   and	   justify	   continuing	   investigations	   to	  elucidate	  the	  mechanisms.	  	  We	   established	   that	   polyP	   binds	   directly	   to	   C5b,6,	   and	   in	   doing	   so,	   reduces	   binding	   and/or	  integration	   of	   the	   downstream	   complexes,	   C5b-­‐7	   and	   C5b-­‐8,	   to	   the	   target	   membrane.	   This	  would	  explain	  our	  finding	  that	  polyP	  reduces	  the	  generation	  of	  the	  lytic	  MAC.	  In	  the	  formation	  of	   the	   C5b-­‐7	   complex,	   C7	   binds	   mostly	   to	   the	   C6	   component	   of	   C5b,6	   [171].	   We	   have	   not	  determined	   the	  mechanism	   of	   action	   of	   polyP	   in	   terms	   of	   how	   it	   affects	   C7	   interaction	  with	  C5b,6.	   PolyP	   may	   reduce	   C7	   binding	   to	   C5b,6,	   or	   it	   may	   simply	   alter	   the	   structure	   of	   C5b,6	  (without	   affecting	   the	   quantitative	   binding	   to	   C7),	   rendering	   the	   C5b-­‐7	   complex	   (and	  consequently,	   C5b-­‐8	   and	   C5b-­‐9	   complexes)	   less	   functional	   [171].	   Interestingly,	   polyP	   had	   no	  effect	  on	  the	  formation	  of	  SC5b-­‐9.	  SC5b-­‐9	  is	  considered	  a	  product	  of	  defective	  assembly	  of	  the	  MAC,	  when	  terminal	  complement	  protein	  complexes	   take	  on	  altered	  structures	  and	  bind	   to	  S	  protein	   (aka	  vitronectin)	   instead	  of	   the	  cell	  membrane,	  exposing	  a	  heparin-­‐binding	   site	   [172].	  The	  S-­‐protein	  is	  incorporated	  into	  the	  complement	  complex	  as	  C5b,6	  and	  C7	  bind	  to	  each	  other,	  and	  is	  still	  susceptible	  to	  binding	  of	  C8	  and	  some	  C9	  molecules	  [173].	  PolyP,	  which	  destabilizes	  C5b,6,	   may	   still	   allow	   the	   formation	   of	   C5b,6	   to	   C7,	   making	   it	   less	   membranolytic,	   but	   also	  allowing	   for	   the	   incorporation	   of	   S-­‐protein.	   Thus,	   although	   the	   formation	   of	   C5b-­‐9	   on	  membranes	  is	  suppressed,	  the	  formation	  of	  SC5b-­‐9	  is	  largely	  unaffected.	  As	  discussed	  in	  the	  Introduction,	  host	  mammalian	  cells	  are	  endowed	  with	  negative	  regulators	  of	  complement,	  several	  of	  which	  target	  the	  terminal	  pathway.	  These	  include	  vitronectin,	  clusterin	  and	  CD59.	  It	  will	  be	  interesting	  to	  examine	  whether	  polyP	  interacts	  with	  these	  proteins,	  thereby	  further	  modulating	  generation	  of	  a	  functional	  MAC.	  	  Addressing	   these	   fundamental	   questions,	   and	   understanding	   how	   polyP	   affects	   direct	  interactions	  with	  C5b,6	  and	  other	  complement	  components	  (e.g.	  C1-­‐INH,	  FH,	  C1s,	  vitronectin),	  will	   ultimately	   allow	   for	   the	   design	   of	   specific	   therapies	   to	  modulate	   complement	   at	   distinct	  steps	   in	   the	  cascade.	  Thus,	  a	  modified	   form	  of	  polyP	  or	  a	  compound	  that	  binds	  specifically	   to	  	  	   74	  C5b,6,	  could	  turn	  off	  the	  terminal	  pathway	  and	  MAC	  formation,	  without	  affecting	  any	  other	  step	  in	  the	  cascade,	  thereby	  preserving	  the	  generation	  of	  other	  biologically	  important	  products	  (e.g.,	  C5a	  and	  C3a,	  which	  act	  as	  anaphylatoxins).	  This	  may	  be	  of	  particular	  interest	  for	  diseases	  such	  as	  AMD,	  where	  the	  MAC	  is	  believed	  to	  be	  the	  most	  damaging	  to	  the	  RPE	  and	  CEC.	  	  6.3	  	   Cellular	  Models	  of	  AMD:	  Limitations	  and	  Advantages	  With	  all	  models	  –	   in	  vitro	  and	   in	  vivo	  –	   there	  are	  advantages	  and	  disadvantages	   that	  must	  be	  recognized,	  so	  that	  conclusions	  from	  experiments	  can	  be	  appropriately	  drawn.	  We	  used	  ARPE-­‐19	   and	   RF/6A	   cell	   lines	   that	   represent,	   respectively	   RPE	   and	   CEC,	   to	   examine	   the	   effects	   of	  polyP.	  These	  cell	   lines	  have	  been	  well-­‐characterized	  and	  are	  widely	  used	  in	  the	  field.	  They	  are	  readily	   cultured,	   grow	   steadily,	   and	   with	   little	   change	   in	  morphology	   over	  multiple	   passages	  [145,	  148,	  150,	  174,	  175].	  The	  ARPE-­‐19	  cells	  are	  particularly	  well-­‐suited	  for	  our	  studies	  on	  AMD.	  These	  cells	  that	  spontaneously	  arose	  from	  human	  RPE	  cells,	  have	  a	  normal	  karyotype,	  express	  RPE-­‐specific	  markers	  CRALBP	  and	  RPE65,	  can	  polarize	  on	  laminin-­‐coated	  transwells,	  and	  develop	  tight	   junctions	  with	   relatively	   high	   transepithelial	   resistance	   (TER)	   that	   is	   reportedly	   found	   in	  vivo	  [145].	  There	  are	  however,	  some	  limitations.	  Depending	  on	  the	  subline	  of	  ARPE-­‐19	  studied,	  the	   TER	   of	   the	   cells	   gradually	   decline	   during	   passage,	   with	   corresponding	   reduction	   in	   the	  integrity	  of	  the	  tight	  junctions.	  We	  did	  not	  test	  for	  TER	  or	  expression	  of	  the	  specific	  markers;	  nor	  did	  we	  grow	   the	   cells	  under	   conditions	   in	  which	   they	  would	  polarize.	   The	   latter	   in	  particular,	  may	  alter	  their	  response	  to	  the	  stresses	  and	  to	  polyP.	  The	  RF/6A	  cells	  were	  also	  derived	  spontaneously	  from	  the	  choroid	  endothelial	  cells,	  but	  from	  a	  rhesus	  macaque	  fetus	  [150].	  These	  cells	  exhibit	  the	  cobblestone	  appearance	  of	  endothelial	  cells	  in	  culture	  and	  have	  Weibel	  Palade	  bodies	  and	  express	  von	  Willebrand	  factor	  (VWF),	  consistent	  with	  them	  being	  endothelial	  in	  origin.	  However,	  the	  expression	  of	  VWF	  gradually	  reduces	  with	  passage,	   and	  we	  determined	  by	  Western	  blot	   (not	   shown)	   that	   expression	  of	   the	   endothelial	  glycoprotein,	  thrombomodulin,	  is	  very	  low.	  Thus,	  there	  are	  some	  limitations	  to	  these	  cells,	  and	  they	  therefore	  may	  not	  be	  fully	  representative	  of	  their	  primary	  counterparts	  from	  the	  choroid	  endothelium.	  Finally,	  for	  some	  studies,	  we	  also	  used	  EA.hy	  926,	  which	  are	  phenotypically	  most	  	  	   75	  similar	  to	  human	  umbilical	  vein	  endothelial	  cells	  [163].	  These	  have	  been	  widely	  used,	  but	  there	  are	  many	  others	  that	  could	  be	  tested	  for	  validation	  purposes.	  6.4	  	   Role	  of	  PolyP	  and	  MonoP	  as	  an	  Anti-­‐oxidant	  Thus	   far,	   studies	   on	   the	   role	   of	   polyP	   in	   mammalian	   systems	   have	   been	   limited	   to	   the	  coagulation	   and	   complement	   systems.	   The	   anti-­‐oxidant	   properties	   of	   polyP	   demonstrated	   in	  this	   thesis	   present	   a	   potentially	   novel	   mammalian	   function.	   This	   finding	   is	   not	   entirely	  unprecedented,	  as	  polyP	   is	   known	   to	  play	  a	   role	   in	  bacterial	   survival	   in	   response	   to	  oxidative	  stress.	   In	   studies	   in	   prokaryotes,	   the	   overproduction	   of	   an	   exopolyphosphatase	   resulted	   in	  decreased	  levels	  of	  polyP,	  which	  rendered	  E.	  coli	  significantly	  more	  sensitive	  to	  H2O2-­‐mediated	  death	   [17].	   This	   was	   attributed	   at	   least	   in	   part,	   to	   enhanced	   expression	   of	   the	   anti-­‐oxidant	  catalase	  HPII,	  a	  catalase	  in	  E.	  coli	  that	  is	  reportedly	  dependent	  on	  the	  presence	  of	  polyP,	  at	  least	  in	   the	   stationary	   growth	   phase.	   A	   separate	   study	   found	   that	   an	   E.	   coli	   mutant	   lacking	   the	  polyphosphate	  kinase	  enzyme	  suffered	  from	  decreased	  viability	  when	  challenged	  with	  oxidative	  stress	   [176].	  The	  mechanism	  of	   these	  anti-­‐oxidant	  protective	  properties	   is	  believed	   to	   involve	  gene	  expression	  regulation	  of	  cyclic	  AMP	  receptor	  protein	  and	  RNA	  polymerase,	  sigma	  S	  [177].	  Cyclic	  AMP	  is	  an	  important	  messenger	  of	  many	  cell-­‐signaling	  pathways,	  while	  RNA	  polymerase,	  sigma	  S	  is	  a	  regulator	  of	  transcription	  of	  stationary	  phase	  bacterial	  genes.	  This	  gene	  expression	  regulatory	   function	   may	   be	   similar	   to	   the	   way	   inositol	   polyphosphates	   play	   a	   role	   in	   gene	  regulation	  and	  nucleic	  acid	  break	  repair	  [178-­‐184].	  In	   mammalian	   cell	   culture	   systems,	   we	   demonstrated	   that	   polyP	   exhibits	   protective	   effects	  against	   H2O2-­‐induced	   oxidative	   stress,	   as	   evidenced	   by	   cell	  morphology,	   retention	   of	   nuclear	  integrity,	   and	   expression	   of	   the	   intercellular	   junctional	   protein,	   VE-­‐cadherin.	   Interestingly,	  monoP	   also	   exhibited	   protective	   properties,	   but	   this	   was	   only	   evident	   in	   the	   qualitative,	   cell	  morphology	  studies.	  MonoP	  did	  not	  appear	  to	  protect	  against	  oxidative	  stress-­‐induced	  changes	  in	   nuclear	   integrity	   or	   VE-­‐cadherin	   expression.	   This	   may	   reflect	   differential	   cytoprotective	  mechanisms	  of	  monoP	  versus	  polyP,	  or	   alternatively,	   varying	   sensitivities	   to	   the	  assays.	   Thus,	  the	  stress-­‐induced	  changes	   in	  nuclear	   integrity	  and	  VE-­‐cadherin	  were	  not	  sensitive	   to	  monoP,	  	  	   76	  whereas	  cytoskeletal	  changes	  that	  may	  affect	  morphology,	  may	  be	  more	  sensitive.	  Even	  though	  the	   monoP	   and	   polyP	   were	   compared,	   based	   on	   equivalent	   molar	   concentrations	   of	  orthophosphate	  units,	  differences	  in	  response	  may	  also	  be	  attributed	  to	  slightly	  different	  ionic	  strengths,	  caused	  partly	  by	  folds	  in	  polyP,	  hiding	  anionic	  groups	  from	  exposure.	  The	   mechanism(s)	   underlying	   the	   cytoprotection	   of	   polyP	   and	   monoP	   has	   not	   yet	   been	  elucidated.	  In	  spite	  of	  some	  suggestive	  data,	  we	  have	  not	  reliably	  demonstrated	  that	  polyP	  (or	  monoP)	  induces	  gene	  or	  protein	  expression	  of	  catalase,	  as	  was	  found	  in	  the	  bacteria.	  This	  is	  not	  entirely	   surprising,	   as	   protective	   mechanisms	   for	   mammalian	   cells	   and	   bacteria	   would	   be	  expected	   to	   differ.	   Bacterial	   expression	   of	   catalase	   was	   dependent	   on	   growth	   phase,	   which	  cannot	  be	  extrapolated	  to	  mammalian	  cells,	  [176,	  185].	  As	  noted	  previously,	  our	  cells	  were	  not	  of	   primary	   origin,	   and	   were	   not	   grown	   to	   fully	   represent	   the	   in	   vivo	   situation,	   i.e.,	   lacking	  polarization	  and	  a	  subcellular	  matrix.	  We	  used	  polyP130	  to	  assess	  catalase	  expression,	  which	  is	  a	  much	   shorter	   form	   of	   the	   polymer	   than	   exists	   in	   bacteria.	   Overall,	   it	   is	   possible	   that	   under	  different	  conditions,	  catalase	  expression	  might	  be	  enhanced	  by	  polyP.	  These	  theories	  remain	  to	  be	  determined.	  	  Beyond	  catalase,	   it	   is	  worth	  considering	   that	  other	  protective	  mechanisms	  may	  participate	   to	  protect	   RPE	   cells	   and	   CEC	   from	   the	   H2O2-­‐induced	   oxidative	   stress.	   The	   other	   key	   antioxidant	  enzymes	   in	  mammalian	  cells	  are	  superoxide	  dismutase	  and	  glutathione	  peroxidase,	  with	  each	  one	   having	   a	   distinct	   mechanism	   of	   action,	   distinguished	   from	   that	   of	   catalase	   [152].	  Glutathione	  peroxidase	  reduces	  hydroperoxides	  (including	  H2O2)	  to	  alcohols	  [186].	  This	  reaction	  is	  dependent	  on	  the	  presence	  of	  selenium	  [187].	  Unlike	  catalase,	  which	   is	  not	  produced	   in	  all	  cell-­‐types	   and	   which	   has	   specific	   activity	   for	   H2O2,	   glutathione	   peroxidase	   is	   ubiquitously	  expressed,	  and	   is	   the	  main	  detoxification	  mechanism	  to	  counter	   low	   levels	  of	  oxidative	  stress	  induced	  by	  various	  types	  of	  hydroperoxides	  [152].	  Five	   isoforms	  of	  glutathione	  peroxidase	  are	  found	  in	  mammalian	  cells,	  with	  expression	  levels	  varying	  with	  the	  tissue	  type	  [152].	  Superoxide	  dismutase	  eliminates	  highly	  reactive	  O2-­‐	  anion	  radicals	  by	  catalyzing	  their	  formation	  into	  H2O2,	  which	  are	  then	  hydrolyzed	  by	  catalase	  or	  glutathione	  peroxidase	   [188].	  Mice	   lacking	  the	  gene	  for	   superoxide	   dismutase	   develop	   multi-­‐organ	   damage	   from	   continuous	   and	   unregulated	  	  	   77	  oxidative	   stress	   [189-­‐191].	   PolyP	  more	   likely	   affects	   glutathione	   peroxidase,	   which	   is	   able	   to	  directly	  degrade	  H2O2,	   the	   stressor	   that	  was	  used	   in	  our	   studies.	   It	   is	   also	  possible	   that	  polyP	  alters	  the	  manner	  in	  which	  these	  antioxidant	  enzymes	  function	  in	  concert.	  Further	  studies	  could	  include	   examining	   the	   ability	   of	   these	   three	   enzymes	   to	   degrade	   H2O2	   using	   purified	   protein	  systems	   in	   the	   presence	   or	   absence	   of	   polyP.	   Such	   an	   approach	   would	   allow	   one	   to	   assess	  whether	   polyP	   acts	   directly	   on	  one	  or	  more	  of	   these	   enzymes,	   or	   is	   involved	   in	   intermediary	  steps	  (e.g.,	  gene	  expression	  or	  cellular	  signaling).	  	  In	   addition	   to	  a	  direct	   anti-­‐oxidant	  enzyme	  mechanism,	   the	   chemical	  properties	  of	   the	  highly	  anionic	   polyP	   (and	   monoP)	   may	   also	   explain	   the	   protection	   against	   H2O2-­‐induced	   cellular	  damage.	  Generation	  of	  reactive	  radicals	  and	  reactive	  oxygen	  species	  from	  hydrogen	  peroxide	  is	  catalyzed	  by	   the	  presence	  of	  metals,	   such	   as	   iron	   and	   copper	   [192-­‐194].	   Through	   the	   Fenton	  reaction,	   iron	   accelerates	   oxidative	   damage	   by	   converting	   hydrogen	   peroxide	   to	   a	   hydroxyl	  radical	   and	   ferryl	   iron	   [195,	   196].	   These	  metals	  were	  undoubtedly	   present	   in	   our	   in	   vitro	   cell	  culture	  models	   in	  which	  we	   examined	   the	   effects	   of	   polyP	   and	  monoP,	   raising	   the	  possibility	  that	   polyP	   acts	   at	   least	   in	   part	   by	   chelating	   the	   free	   iron	   or	   copper,	   thereby	   reducing	   the	  production	  of	  damaging	   reactive	  oxygen	   species.	   This	  would	  be	   in	   line	  with	  our	   findings	  with	  monoP,	   the	   latter	   of	   which	   also	   provided	   some	   protection,	   at	   least	   in	   the	   qualitative	  morphology	  studies.	  As	  already	  discussed,	   the	  monoP,	  however,	  had	   little	  effect	   in	  protecting	  against	   H2O2-­‐induced	   changes	   in	   nuclear	   integrity	   and	   in	   VE-­‐cadherin	   expression,	   suggesting	  that	  chelation	  of	  metal	  ions	  is	  not	  entirely	  responsible	  for	  the	  protective	  effects	  of	  polyP	  [145,	  197,	  198].	  Overall,	  no	  matter	  what	  the	  mechanism,	  it	  is	  intriguing	  that	  polyP	  exhibits	  anti-­‐oxidant	  effects,	  as	   well	   as	   complement-­‐inhibiting	   effects.	   In	   terms	   of	   AMD,	   the	   combination	   of	   excess	  complement	  activation	  and	  oxidative	  stress	  are	  considered	  the	  key	  determinants	  in	  promoting	  disease	   progression.	   That	   polyP	   simultaneously	   interferes	  with	   both,	   is	   therefore	   of	   potential	  clinical	  value	  from	  a	  therapeutic	  point	  of	  view.	  If	  indeed,	  monoP	  also	  protects	  against	  oxidative	  stress,	  this	  may	  be	  of	  further	  value,	  providing	  longer	  lasting	  therapeutic	  benefit	  as	  the	  polyP	  is	  naturally	  hydrolyzed	  into	  the	  monomeric	  form.	  	  	   78	  6.5	  	   Lack	  of	  Proliferative	  Effect	  on	  Endothelial	  Cells	  In	  our	   in	  vitro	   studies,	  data	  without	  oxidative	   stress	   suggested	   that	  polyP	  may	   induce	  cellular	  proliferation.	   The	   implications	   of	   such	   a	   finding	   would	   extend	   beyond	   AMD,	   providing	   a	  potential	   therapeutic	   target	   for	   several	   conditions,	   including	   for	   example,	   wound-­‐healing,	  pathologic	  angiogenesis,	  inflammation	  and	  cancer.	  For	  that	  reason,	  we	  more	  carefully	  examined	  the	  effect	  of	  polyP	  on	  endothelial	  cell	  growth.	  Using	  2	  independent	  approaches	  (cell	  count	  and	  BrdU	  incorporation),	  we	  could	  not	  discern	  any	  significant	  effects	  of	  polyP	  on	  the	  proliferation	  of	  EA.hy	  926	  cells	  under	  non-­‐stress	  conditions.	  MonoP	  also	  had	  no	  effect.	  In	  smaller	  scale	  studies,	  similar	  findings	  were	  obtained	  with	  RPE	  cells.	  Further	  studies	  are	  needed	  to	  test	  whether	  polyP	  affects	  the	  cells	  when	  under	  stress.	  Nonetheless,	   this	   lack	  of	   an	  effect	  on	  endothelial	   cell	   proliferation,	  was	   in	   fact	   reassuring.	  As	  previously	  discussed,	  a	  major	  feature	  of	  wet	  AMD	  is	  excess	  growth	  of	  the	  choroid	  vasculature,	  which	   ultimately	   leads	   to	   leakage	   and	   bleeding	   and	   damage	   to	   the	   photoreceptors.	   PolyP-­‐induced	  proliferation	  of	  endothelial	  cells	  might	  therefore	  be	  expected	  to	  exacerbate	  the	  growth	  of	  CNV	  lesions,	  and	  that	  would	  severely	  limit	  the	  value	  of	  polyP	  as	  a	  potential	  therapeutic	  agent	  for	   AMD.	   It	   is	   intriguing,	   however,	   that	   in	   spite	   of	   extensive	   studies	   on	   the	   role	   of	   polyP	   on	  coagulation,	   little	   is	   known	   about	   its	   effect	   on	   endothelial	   cells.	   Studies	   to	   examine	   that	  important	  question	  are	  now	  in	  progress.	  	  6.6	  	   Therapeutic	  Potential	  for	  AMD	  Our	   in	  vivo	  studies	  with	  rats	  and	  mice	  support	  the	  notion	  that	  polyP	  exerts	  a	  protective	  effect	  against	  the	  development	  of	  Wet	  AMD,	  and	  provide	  an	  exciting	  novel	  potential	  use	  of	  polyP	   in	  mammalian	   systems.	   The	   two	   independent	   experiments	   indicate	   that	   both	   long-­‐chain	   polyP	  (polyP≥1000)	   as	   well	   as	   shorter-­‐chain	   polyP	   (polyP130)	   are	   able	   to	   abrogate	   the	   progression	   of	  AMD-­‐associated	  damage	  at	  different	  time	  points	  (5	  days	  and	  14	  days).	   In	  spite	  of	   in	  vitro	  data	  suggesting	  that	  monoP	  has	  anti-­‐oxidant	  properties,	  monoP	  injections	  had	  no	  beneficial	  effect	  in	  the	  laser-­‐induced	  CNV	  model.	  	  	   79	  6.7	   Strengths	  and	  Weaknesses	  of	  the	  AMD	  Model	  of	  Laser-­‐induced	  CNV	  We	  are	  optimistic	  of	  the	  potential	  clinical	  importance	  of	  our	  findings,	  based	  on	  the	  fact	  that	  this	  model	  has	  been	  widely	  and	  repeatedly	  validated.	  Laser-­‐induced	  CNV	  has	  been	  used	  in	  rodents,	  rabbits	   and	   non-­‐human	   primates	   as	   the	   standard	  model	   for	   validating	   drugs	   that	   target	   the	  VEGF	  pathway,	  currently	   in	  use	   in	  clinics	   throughout	   the	  world.	  The	  advantages	  of	   this	  model	  are	  several.	  It	  induces	  a	  highly	  reproducible,	  localized	  CNV	  lesion,	  with	  which	  one	  can	  measure	  the	  efficacy	  of	  potential	  therapies	  such	  as	  polyP.	  The	  acute	  injury	  provides	  a	  rapid	  turnaround	  time	  for	  results,	  and	  the	  localization	  of	  the	  injury	  as	  well	  as	  the	  method	  of	  drug	  delivery	  allows	  one	  to	  avoid	  systemic	  effects	  that	  intravenous	  administration	  of	  polyP	  might	  otherwise	  cause.	  	  In	   spite	   of	   the	   above,	   observations	   made	   with	   this	   model	   should	   be	   interpreted	   with	   some	  caution.	   The	   physiology	   of	   rodent	   eyes	   differs	   from	   those	   of	   humans	   in	   that	   rodents	   do	   not	  possess	  a	  macula,	  which	  is	  the	  structure	  of	  the	  retina	  that	   is	  most	  affected	  in	  AMD	  [199].	  The	  macula	   has	   anatomical	   features	   distinct	   from	   the	   rest	   of	   the	   retina,	   including	   a	   greater	  concentration	  of	  photoreceptors.	  This	  would	  suggest	  that	  the	  pathological	  consequences	  of	  the	  laser-­‐induced	  injury	  would	  differ	  from	  the	  pathology	  associated	  with	  AMD	  in	  humans.	   In	  turn,	  the	  response	  to	  polyP	  administration	  would	   likely	  also	  differ	  between	  the	  species	   [200].	  From	  that,	   it	   is	   remarkable	   (and	   fortunate)	   that	   agents	   used	   in	  mice	   appear	   to	   also	   be	   effective	   in	  human	  adults.	  We	  do	  not	  currently	  know,	  of	  course,	  if	  this	  applies	  to	  polyP.	  The	  nature	  of	  the	  injury	  that	  results	  from	  laser-­‐induced	  CNV	  also	  differs	  considerably	  from	  that	  found	   in	   human	   AMD.	   The	   rodent	   model	   is	   associated	   with	   a	   physical	   injury	   to	   Bruch’s	  membrane,	  which	  then	  triggers	  acute	  inflammation	  and	  consequently,	  the	  development	  of	  CNV	  lesions	   [143].	   In	   contrast,	   AMD	   in	   humans	   is	   a	   chronic,	   progressive	  multi-­‐stage	   inflammatory	  disease.	   Instead	   of	   a	   single	   trigger,	   the	   progression	   of	   AMD	   is	   influenced	   by	   numerous	  exogenous	   factors	   (e.g.,	   smoking),	   genetic	   predisposition	   (e.g.	   the	   genotype	   CFH	   Y402H),	  chronic	   oxidative	   stress,	   age	   (increased	   in	   Alzheimer’s	   Disease),	   and	   a	   range	   of	   other	  environmental	   factors	   (e.g.	   light	   exposure),	  many	   of	  which	   have	   not	   been	   identified	   yet	   [59,	  113,	   201].	   These	   factors	   are	   not	   accounted	   for	   in	   our	   rodent	  model.	   As	  with	  most	   approved	  	  	   80	  therapeutics,	   we	   can	   only	   hope	   that	   polyP	   would	   be	   effective	   in	   treating	   patients	   with	   or	  without	  underlying	  risk	  factors	  and	  varying	  degrees	  of	  pathology.	  	  6.8	  	   Future	  In	  vivo	  Studies	  with	  PolyP	  Further	  validation	  of	  our	  in	  vivo	  findings	  is	  required,	  with	  an	  expansion	  of	  the	  mouse	  model	  to	  include	  different	  lengths	  and	  concentrations	  of	  polyP	  treatment,	  and	  a	  post-­‐treatment	  period	  of	  7	   days	   instead	   of	   14	   days.	   The	   reason	   for	   altering	   this	   time	   period	   is	   that	   there	   have	   been	  reports	  that	  at	  14-­‐21	  days,	  the	  CNV	  injury	  starts	  to	  spontaneously	  regress	  [143,	  144].	  Ideally,	  the	  most	  reliable	  indication	  of	  a	  protective	  effect	  will	  be	  achieved	  at	  the	  time	  that	  the	  injury	  is	  at	  its	  worst.	  	  Further	  analyses	  will	  also	  require	  histologic	  studies	  to	  assess	  the	  localization	  of	  the	  polyP	  and	  to	  examine	   for	   toxicity	   –	   local	   and	   systemic.	   Techniques	   to	   probe	   for	   polyP	   are	   only	   recently	  emerging.	   Dr.	   Jim	   Morrissey	   is	   currently	   using	   different	   approaches	   to	   develop	   probes	   that	  specifically	   detect	   different	   length	   polyP.	   These	   are	   based	   on	   the	   existence	   of	  exopolyphosphatase	  binding	  sites	  on	  polyP.	  Other	  approaches	  include	  DNA-­‐binding	  probes	  (i.e.,	  DAPI),	   due	   to	   the	   similarity	   of	   polyP	   to	   the	   DNA	   sugar-­‐phosphate	   backbone	   [202-­‐204].	   The	  development	   of	   a	   sensitive	   and	   specific	   polyP	   probe	   will	   allow	   for	   simultaneous	   imaging	   or	  immunohistochemical	   detection	   of	   CNV,	   complement	   components,	   and	   polyP.	   This	   will	   help	  determine	  how	  and	  where	  polyP	  is	  acting,	  and	  whether	  it	  remains	  intact.	  	  In	  this	  thesis,	  I	  report	  that	  long	  chain	  polyP	  retains	  much	  of	  its	  complement	  suppressive	  function	  even	  after	  10	  days	  in	  normal	  human	   serum.	  This	   finding	   conflicts	   somewhat	  with	  a	  previous	   report,	   but	   there	   they	  measured	  the	  stability	  of	  polyP75	  in	  terms	  of	  its	  polymer	  length,	  and	  not	  its	  function	  [9].	  In	  any	  case,	   the	   stability	  of	  polyP	   in	   the	  vitreous	   is	  more	   relevant	   for	  our	  purposes	   [205],	   and	   these	  studies	  will	  be	  performed	  in	  vivo,	  with	  the	  availability	  of	  good	  detection	  methods.	  	  	  	   81	  One	  area	  of	   concern	  with	  administering	  polyP	  would	  naturally	  be	   thrombosis.	   The	   vitreous	   is	  not	   known	   to	   be	   associated	   with	   thrombotic	   disease,	   with	   clots	   only	   found	   there	   when	   the	  retinal	  vessels	  bleed.	  The	  fluid	  in	  the	  highly	  viscous	  vitreous	  does	  not	  turn	  over,	  but	  also	  has	  not	  been	   well-­‐characterized	   in	   terms	   of	   coagulation	   protein	   constituents.	   The	   question	   as	   to	  whether	  polyP	  induces	  thrombosis	  of	  the	  choroid	  vasculature	  will	  be	  determined	  histologically.	  It	  is	  not	  clear	  whether	  this	  would	  be	  detrimental,	  particuarly	  if	  it	  occurred	  only	  locally	  at	  the	  site	  of	  the	  laser	  injury.	  This	  would	  not	  cause	  loss	  of	  vision.	  Indeed,	  it	  may	  inhibit	  exudation	  of	  serous	  contents	  into	  the	  retina,	  which	  is	  the	  effect	  desired	  for	  a	  drug.	  In	  fact,	  photodynamic	  therapy,	  the	  standard	  treatment	  used	  prior	  to	  the	  introduction	  of	  anti-­‐VEGF	  drugs,	  exerted	  its	  effects	  by	  inducing	   localized	   thrombosis	   of	   the	   CNV	   lesions	   [206].	   Only	   if	   the	   polyP	   caused	   more	  widespread	  thrombosis,	  would	  there	  be	  concern.	  Even	  if	  the	  polyP	  distributed	  into	  the	  systemic	  circulation,	   the	   amounts	   would	   be	   very	   low,	   and	   not	   likely	   to	   cause	   harm.	   Indeed,	   polyP	  normally	  circulates	  in	  the	  plasma	  at	  micromolar	  concentrations	  and	  apparently	  does	  no	  harm.	  We	  have	  limited	  the	  current	  studies	  to	  the	  wet	  form	  of	  AMD.	  As	  noted,	  Dry	  AMD	  is	  much	  more	  common,	  and	  also	  believed	   to	  be	  mediated	  by	  excess	  complement	  and	  exposure	   to	  oxidative	  stress.	  Unfortunately,	  in	  spite	  of	  attempts	  by	  several	  groups,	  validated	  rodent	  models	  to	  study	  Dry	  AMD	  are	  lacking.	  This	  may	  be	  changing.	  A	  rodent	  model	  of	  Dry	  AMD	  with	  some	  promise	  has	  recently	   been	   reported.	   This	   model	   requires	   the	   intravitreal	   injection	   of	   carboxyethylpyrrole	  adducted	  proteins,	  a	  product	  of	  oxidative	  stress	   that	   is	   found	   in	  drusen.	  Mice	  do	  not	  develop	  CNV,	  but	  do	  apparently	  develop	  features	  of	  Dry	  AMD	  over	  many	  months,	  and	  this	  is	  associated	  with	  increased	  deposition	  of	  complement	  proteins	  [207].	  	  In	  summary,	  the	  in	  vivo	  findings	  described	  in	  this	  thesis	  are	  rationale	  for	  further	  study	  of	  the	  use	  of	  polyP	  to	  treat	  Wet	  AMD,	  able	  to	  simultaneously	  target	  both	  complement	  deposition	  and	  CNV	  development.	   The	  mechanistic	   details	   of	   this	   effect	   as	   well	   as	   potential	   side-­‐effects	   warrant	  further	  studies,	  beyond	  the	  scope	  of	  this	  thesis.	  	  	   82	  Chapter	  7: Conclusion	  In	   this	   thesis,	   I	   have	   further	   clarified	   the	  mechanism	  by	  which	  polyP	   suppresses	   complement	  activation	  via	  the	  terminal	  pathway,	  an	  effect	  that	  spares	  generation	  of	  C5a,	  while	  reducing	  the	  production	   of	   the	   damaging	   membrane	   attack	   complex	   (MAC).	   These	   studies	   have	   further	  revealed	   that	  polyP	  not	  only	  dampens	  complement	  activation,	  but	   that	   it	   also	   interferes	  with	  oxidative	   stress-­‐induced	   cellular	   damage.	   The	  mechanisms	  by	  which	   it	   exerts	   this	   effect	   have	  not	   yet	   been	   determined.	   However,	   an	   agent	   that	   simultaneously	   suppresses	   complement	  activation	  and	  protects	  against	  oxidative	  stress	  holds	  potential	  therapeutic	  value.	  Indeed,	  I	  have	  shown	   that	   in	   vivo,	   in	   rodent	  models	   of	   AMD,	   that	   polyP	   protects	   against	   laser-­‐induced	   CNV	  with	  reduced	  deposition	  of	  complement.	  	  There	  remains	  much	  work	  to	  be	  done.	  Nonetheless,	  the	  findings	  in	  this	  thesis	  raise	  awareness	  of	  the	   potential	   importance	   of	   a	   ubiquitous,	   naturally	   occurring	   inorganic	   compound	   that	   has	  largely	   been	   overlooked.	  Most	   important,	   the	   findings	   reveal	   a	   promising	   use	   for	   polyP	   as	   a	  treatment	  for	  AMD,	  a	  common	  and	  devastating	  disease.	  	  	  	  	  	  	  	  	  	  	   83	  Bibliography	  1.	   Markiewski,	  M.M.,	  et	  al.,	  Complement	  and	  coagulation:	  strangers	  or	  partners	  in	  crime?	  Trends	  Immunol,	  2007.	  28(4):	  p.	  184-­‐92.	  2.	   Hillmen,	   P.,	   et	   al.,	   Effect	   of	   the	   complement	   inhibitor	   eculizumab	   on	   thromboembolism	   in	  patients	  with	  paroxysmal	  nocturnal	  hemoglobinuria.	  Blood,	  2007.	  110(12):	  p.	  4123-­‐8.	  3.	   Cofiell,	   R.,	   et	   al.,	   Eculizumab	   reduces	   complement	   activation,	   inflammation,	   endothelial	  damage,	  thrombosis,	  and	  renal	  injury	  markers	  in	  aHUS.	  Blood,	  2015.	  125(21):	  p.	  3253-­‐62.	  4.	   Adams,	   R.L.	   and	   R.J.	   Bird,	   Review	   article:	   Coagulation	   cascade	   and	   therapeutics	   update:	  relevance	   to	   nephrology.	   Part	   1:	   Overview	   of	   coagulation,	   thrombophilias	   and	   history	   of	  anticoagulants.	  Nephrology	  (Carlton),	  2009.	  14(5):	  p.	  462-­‐70.	  5.	   Gailani,	   D.	   and	   T.	   Renne,	   Intrinsic	   pathway	   of	   coagulation	   and	   arterial	   thrombosis.	  Arterioscler	  Thromb	  Vasc	  Biol,	  2007.	  27(12):	  p.	  2507-­‐13.	  6.	   Vogler,	  E.A.	  and	  C.A.	  Siedlecki,	  Contact	  activation	  of	  blood-­‐plasma	  coagulation.	  Biomaterials,	  2009.	  30(10):	  p.	  1857-­‐69.	  7.	   Schreiber,	   A.D.,	   A.P.	   Kaplan,	   and	   K.F.	   Austen,	   Inhibition	   by	   C1INH	   of	   Hagemann	   factor	  fragment	   activation	   of	   coagulation,	   fibrinolysis,	   and	   kinin	   generation.	   J	   Clin	   Invest,	   1973.	  52(6):	  p.	  1402-­‐9.	  8.	   Cullmann,	   W.,	   et	   al.,	   Complement,	   coagulation	   and	   fibrinolytic	   parameters	   in	   hereditary	  angioedema	  (HAE).	  Clin	  Exp	  Immunol,	  1982.	  49(3):	  p.	  618-­‐22.	  9.	   Smith,	  S.A.,	  et	  al.,	  Polyphosphate	  modulates	  blood	  coagulation	  and	  fibrinolysis.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  2006.	  103(4):	  p.	  903-­‐8.	  10.	   Kornberg,	  A.,	   Inorganic	  polyphosphate:	  a	  molecule	  of	  many	  functions.	  Prog	  Mol	  Subcell	  Biol,	  1999.	  23:	  p.	  1-­‐18.	  11.	   Harold,	   F.M.,	   Inorganic	   polyphosphates	   in	   biology:	   structure,	   metabolism,	   and	   function.	  Bacteriol	  Rev,	  1966.	  30(4):	  p.	  772-­‐94.	  12.	   Bioscience,	   B.S.A.M.	  Conferences	  and	  events	   |	  Biochemical	  Society	   -­‐	   Inorganic	  Polyphosphate	  (polyP)	   Physiology	   2015	   29	   Jun	   2015	   [cited	   2015	   29	   Jun	   2015];	   Available	   from:	  13.	   Kumble,	  K.D.	  and	  A.	  Kornberg,	  Inorganic	  polyphosphate	  in	  mammalian	  cells	  and	  tissues.	  J	  Biol	  Chem,	  1995.	  270(11):	  p.	  5818-­‐22.	  14.	   Moreno,	   S.N.	   and	   R.	   Docampo,	   Polyphosphate	   and	   its	   diverse	   functions	   in	   host	   cells	   and	  pathogens.	  PLoS	  Pathog,	  2013.	  9(5):	  p.	  e1003230.	  15.	   Brown,	   M.R.	   and	   A.	   Kornberg,	   The	   long	   and	   short	   of	   it	   -­‐	   polyphosphate,	   PPK	   and	   bacterial	  survival.	  Trends	  Biochem	  Sci,	  2008.	  33(6):	  p.	  284-­‐90.	  16.	   Nocek,	   B.,	   et	   al.,	   Polyphosphate-­‐dependent	   synthesis	   of	   ATP	   and	   ADP	   by	   the	   family-­‐2	  polyphosphate	  kinases	  in	  bacteria.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  2008.	  105(46):	  p.	  17730-­‐5.	  17.	   Shiba,	  T.,	  et	  al.,	  Inorganic	  polyphosphate	  and	  the	  induction	  of	  rpoS	  expression.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  1997.	  94(21):	  p.	  11210-­‐5.	  18.	   Kim,	   K.S.,	   et	   al.,	   Inorganic	   polyphosphate	   is	   essential	   for	   long-­‐term	   survival	   and	   virulence	  factors	  in	  Shigella	  and	  Salmonella	  spp.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  2002.	  99(11):	  p.	  7675-­‐80.	  19.	   Kulakovskaya,	   T.V.,	   V.M.	   Vagabov,	   and	   I.S.	   Kulaev,	   Inorganic	   polyphosphate	   in	   industry,	  agriculture	  and	  medicine:	  Modern	  state	  and	  outlook.	  Process	  Biochemistry,	  2012.	  47(1):	  p.	  1-­‐10.	  20.	   Morrissey,	   J.H.,	   S.H.	   Choi,	   and	   S.A.	   Smith,	   Polyphosphate:	   an	   ancient	   molecule	   that	   links	  platelets,	  coagulation,	  and	  inflammation.	  Blood,	  2012.	  119(25):	  p.	  5972-­‐9.	  	  	   84	  21.	   Smith,	   S.A.	   and	   J.H.	   Morrissey,	   Polyphosphate	   as	   a	   general	   procoagulant	   agent.	   J	   Thromb	  Haemost,	  2008.	  6(10):	  p.	  1750-­‐6.	  22.	   Ruiz,	   F.A.,	   et	   al.,	   Human	   platelet	   dense	   granules	   contain	   polyphosphate	   and	   are	   similar	   to	  acidocalcisomes	  of	  bacteria	  and	  unicellular	  eukaryotes.	  J	  Biol	  Chem,	  2004.	  279(43):	  p.	  44250-­‐7.	  23.	   Muller,	  F.,	  et	  al.,	  Platelet	  polyphosphates	  are	  proinflammatory	  and	  procoagulant	  mediators	  in	  vivo.	  Cell,	  2009.	  139(6):	  p.	  1143-­‐56.	  24.	   Choi,	   S.H.,	   S.A.	   Smith,	   and	   J.H.	   Morrissey,	   Polyphosphate	   is	   a	   cofactor	   for	   the	   activation	   of	  factor	  XI	  by	  thrombin.	  Blood,	  2011.	  118(26):	  p.	  6963-­‐70.	  25.	   Smith,	   S.A.	   and	   J.H.	   Morrissey,	   Polyphosphate	   enhances	   fibrin	   clot	   structure.	   Blood,	   2008.	  112(7):	  p.	  2810-­‐6.	  26.	   Mutch,	  N.J.,	   et	   al.,	  Polyphosphate	  binds	  with	  high	  affinity	  to	  exosite	   II	  of	   thrombin.	   J	   Thromb	  Haemost,	  2010.	  8(3):	  p.	  548-­‐55.	  27.	   Montilla,	  M.,	  et	  al.,	  Polyphosphate	  binds	  to	  human	  von	  Willebrand	  factor	  in	  vivo	  and	  modulates	  its	  interaction	  with	  glycoprotein	  Ib.	  J	  Thromb	  Haemost,	  2012.	  10(11):	  p.	  2315-­‐23.	  28.	   Renne,	   T.,	   et	   al.,	  Defective	   thrombus	   formation	   in	  mice	   lacking	   coagulation	   factor	  XII.	   J	   Exp	  Med,	  2005.	  202(2):	  p.	  271-­‐81.	  29.	   Dunkelberger,	   J.R.	   and	  W.C.	   Song,	   Complement	   and	   its	   role	   in	   innate	   and	   adaptive	   immune	  responses.	  Cell	  Res,	  2010.	  20(1):	  p.	  34-­‐50.	  30.	   Zipfel,	  P.F.	  and	  C.	  Skerka,	  Complement	  regulators	  and	  inhibitory	  proteins.	  Nat	  Rev	   Immunol,	  2009.	  9(10):	  p.	  729-­‐40.	  31.	   Kishore,	   U.	   and	   K.B.	   Reid,	   C1q:	   structure,	   function,	   and	   receptors.	   Immunopharmacology,	  2000.	  49(1-­‐2):	  p.	  159-­‐70.	  32.	   Ziccardi,	  R.J.,	  The	  first	  component	  of	  human	  complement	  (C1):	  activation	  and	  control.	  Springer	  Semin	  Immunopathol,	  1983.	  6(2-­‐3):	  p.	  213-­‐30.	  33.	   Loos,	   M.	   and	   H.P.	   Heinz,	   Generation	   of	   the	   classical	   pathway	   C3	   convertase	   (EAC4b2a)	   by	  proteolytic	  enzymes.	  Acta	  Pathol	  Microbiol	  Immunol	  Scand	  Suppl,	  1984.	  284:	  p.	  67-­‐74.	  34.	   Matsushita,	   M.	   and	   H.	   Okada,	  Alternative	   complement	  pathway	  activation	  by	  C4b	  deposited	  during	  classical	  pathway	  activation.	  J	  Immunol,	  1986.	  136(8):	  p.	  2994-­‐8.	  35.	   Vorup-­‐Jensen,	   T.,	   et	   al.,	  Distinct	  pathways	  of	  mannan-­‐binding	   lectin	   (MBL)-­‐	  and	  C1-­‐complex	  autoactivation	   revealed	   by	   reconstitution	   of	   MBL	   with	   recombinant	   MBL-­‐associated	   serine	  protease-­‐2.	  J	  Immunol,	  2000.	  165(4):	  p.	  2093-­‐100.	  36.	   Sandberg,	   A.L.	   and	   A.G.	   Osler,	   Dual	   pathways	   of	   complement	   interaction	   with	   guinea	   pig	  immunoglobulins.	  J	  Immunol,	  1971.	  107(5):	  p.	  1268-­‐73.	  37.	   Thurman,	   J.M.	   and	   V.M.	   Holers,	   The	   central	   role	   of	   the	   alternative	   complement	   pathway	   in	  human	  disease.	  J	  Immunol,	  2006.	  176(3):	  p.	  1305-­‐10.	  38.	   Ember,	   J.A.	   and	   T.E.	   Hugli,	   Complement	   factors	   and	   their	   receptors.	   Immunopharmacology,	  1997.	  38(1-­‐2):	  p.	  3-­‐15.	  39.	   Kinoshita,	   T.,	   et	   al.,	   C5	   convertase	   of	   the	   alternative	   complement	   pathway:	   covalent	   linkage	  between	   two	   C3b	   molecules	   within	   the	   trimolecular	   complex	   enzyme.	   J	   Immunol,	   1988.	  141(11):	  p.	  3895-­‐901.	  40.	   el-­‐Lati,	   S.G.,	   C.A.	   Dahinden,	   and	   M.K.	   Church,	   Complement	   peptides	   C3a-­‐	   and	   C5a-­‐induced	  mediator	  release	  from	  dissociated	  human	  skin	  mast	  cells.	   J	   Invest	  Dermatol,	  1994.	  102(5):	  p.	  803-­‐6.	  41.	   Kitamura,	   H.,	   M.	   Tsuboi,	   and	   K.	   Nagaki,	   C3-­‐independent	   immune	   haemolysis:	  mechanism	  of	  membrane	  attack	  complex	  formation.	  Immunology,	  1986.	  59(1):	  p.	  147-­‐51.	  42.	   Podack,	  E.R.,	  Molecular	  composition	  of	  the	  tubular	  structure	  of	  the	  membrane	  attack	  complex	  of	  complement.	  J	  Biol	  Chem,	  1984.	  259(13):	  p.	  8641-­‐7.	  43.	   The	  Mannose-­‐Binding	  Lectin	  (MBL)	  Pathway.	  2010(March	  4,	  2015).	  	  	   85	  44.	   Yang,	   C.,	   L.	   Yang,	   and	   Y.	   Liu,	   Soluble	   complement	   complex	   C5b-­‐9	   promotes	   microglia	  activation.	  J	  Neuroimmunol,	  2014.	  267(1-­‐2):	  p.	  16-­‐9.	  45.	   Podack,	   E.R.	   and	   H.J.	   Muller-­‐Eberhard,	   SC5b-­‐9	   complex	   of	   complement:	   formation	   of	   the	  dimeric	  membrane	  attack	  complex	  by	  removal	  of	  S-­‐protein.	  J	  Immunol,	  1980.	  124(4):	  p.	  1779-­‐83.	  46.	   Hugo,	  F.,	  et	  al.,	  Quantitative	  measurement	  of	  SC5b-­‐9	  and	  C5b-­‐9(m)	  in	  infarcted	  areas	  of	  human	  myocardium.	  Clin	  Exp	  Immunol,	  1990.	  81(1):	  p.	  132-­‐6.	  47.	   Wat,	  J.,	  POLYPHOSPHATE	  AS	  A	  MODULATOR	  OF	  THE	  COMPLEMENT	  SYSTEM,	  in	  Experimental	  Medicine.	  2013,	  The	  University	  of	  British	  Columbia:	  p.	  102.	  48.	   Markiewski,	   M.M.	   and	   J.D.	   Lambris,	   The	   role	   of	   complement	   in	   inflammatory	   diseases	   from	  behind	  the	  scenes	  into	  the	  spotlight.	  Am	  J	  Pathol,	  2007.	  171(3):	  p.	  715-­‐27.	  49.	   Caliezi,	  C.,	  et	  al.,	  C1-­‐Esterase	  inhibitor:	  an	  anti-­‐inflammatory	  agent	  and	  its	  potential	  use	  in	  the	  treatment	  of	  diseases	  other	  than	  hereditary	  angioedema.	  Pharmacol	  Rev,	  2000.	  52(1):	  p.	  91-­‐112.	  50.	   Charles	   A	   Janeway,	   Jr.,	   et	   al.,	   The	   complement	   system	  and	   innate	   immunity.	   2001:	   Garland	  Science.	  51.	   Pangburn,	  M.K.,	  et	  al.,	  Polyanion-­‐induced	  self-­‐association	  of	  complement	  factor	  H.	   J	   Immunol,	  2009.	  182(2):	  p.	  1061-­‐8.	  52.	   Rodriguez	   de	   Cordoba,	   S.,	   et	   al.,	   The	   human	   complement	   factor	  H:	   functional	   roles,	   genetic	  variations	  and	  disease	  associations.	  Mol	  Immunol,	  2004.	  41(4):	  p.	  355-­‐67.	  53.	   Gordon,	   D.L.,	   et	   al.,	   Identification	   of	   complement	   regulatory	   domains	   in	   human	   factor	   H.	   J	  Immunol,	  1995.	  155(1):	  p.	  348-­‐56.	  54.	   Tichaczek-­‐Goska,	   D.,	   Deficiencies	   and	   excessive	   human	   complement	   system	   activation	   in	  disorders	  of	  multifarious	  etiology.	  Adv	  Clin	  Exp	  Med,	  2012.	  21(1):	  p.	  105-­‐14.	  55.	   Liszewski,	  M.K.,	  C.J.	  Fang,	  and	   J.P.	  Atkinson,	   Inhibiting	  complement	  activation	  on	  cells	  at	  the	  step	  of	  C3	  cleavage.	  Vaccine,	  2008.	  26	  Suppl	  8:	  p.	  I22-­‐7.	  56.	   Ollert,	  M.W.,	  et	  al.,	  Classical	  complement	  pathway	  activation	  on	  nucleated	  cells.	  Role	  of	  factor	  H	  in	  the	  control	  of	  deposited	  C3b.	  J	  Immunol,	  1995.	  155(10):	  p.	  4955-­‐62.	  57.	   Wat,	  J.M.,	  et	  al.,	  Polyphosphate	  suppresses	  complement	  via	  the	  terminal	  pathway.	  Blood,	  2014.	  123(5):	  p.	  768-­‐76.	  58.	   Zhang,	   Q.,	   Y.	   Li,	   and	   C.M.	   Tang,	   The	   role	   of	   the	   exopolyphosphatase	   PPX	   in	   avoidance	   by	  Neisseria	  meningitidis	  of	  complement-­‐mediated	  killing.	  J	  Biol	  Chem,	  2010.	  285(44):	  p.	  34259-­‐68.	  59.	   Facts	  About	  Age-­‐Related	  Macular	  Degeneration.	  	  2015	  June	  2015	  [cited	  2015;	  Available	  from:	  60.	   Brown,	   M.M.,	   et	   al.,	   Age-­‐related	   macular	   degeneration:	   economic	   burden	   and	   value-­‐based	  medicine	  analysis.	  Can	  J	  Ophthalmol,	  2005.	  40(3):	  p.	  277-­‐87.	  61.	   Hassell,	   J.B.,	  E.L.	  Lamoureux,	  and	   J.E.	  Keeffe,	   Impact	  of	  age	  related	  macular	  degeneration	  on	  quality	  of	  life.	  Br	  J	  Ophthalmol,	  2006.	  90(5):	  p.	  593-­‐6.	  62.	   Institute,	  W.E.,	  Age-­‐Related	  Macular	  Degeneration	  |	  Williamson	  Eye	  Institute	  (765)	  477-­‐2020.	  2010.	  2015.	  63.	   Tezel,	   G.,	   et	   al.,	   Oxidative	   stress	   and	   the	   regulation	   of	   complement	   activation	   in	   human	  glaucoma.	  Invest	  Ophthalmol	  Vis	  Sci,	  2010.	  51(10):	  p.	  5071-­‐82.	  64.	   Institute,	   N.E.	   Facts	   About	   Age-­‐Related	   Macular	   Degeneration.	   	   2015	   	   [cited	   2015;	   This	  information	  was	  developed	  by	  the	  National	  Eye	  Institute	  to	  help	  patients	  and	  their	  families	  search	   for	   general	   information	   about	   age-­‐related	   macular	   degeneration.	   An	   eye	   care	  professional	  who	   has	   examined	   the	   patient’s	   eyes	   and	   is	   familiar	  with	   his	   or	   her	  medical	  history	   is	   the	   best	   person	   to	   answer	   specific	   questions.].	   Available	   from:	  	  	   86	  65.	   Williams,	  D.P.,	  et	  al.,	  The	  Retina,	  in	  Neuroscience.	  2001,	  Sinauer	  Associates.	  66.	   Strauss,	   O.,	   The	   retinal	   pigment	   epithelium	   in	   visual	   function.	   Physiol	   Rev,	   2005.	   85(3):	   p.	  845-­‐81.	  67.	   Ugarte,	  M.,	  A.A.	  Hussain,	  and	  J.	  Marshall,	  An	  experimental	  study	  of	  the	  elastic	  properties	  of	  the	  human	   Bruch's	   membrane-­‐choroid	   complex:	   relevance	   to	   ageing.	   Br	   J	   Ophthalmol,	   2006.	  90(5):	  p.	  621-­‐6.	  68.	   Nickla,	  D.L.	  and	  J.	  Wallman,	  The	  multifunctional	  choroid.	  Prog	  Retin	  Eye	  Res,	  2010.	  29(2):	  p.	  144-­‐68.	  69.	   Wallman,	   J.	   and	   J.	  Winawer,	  Homeostasis	  of	  eye	  growth	  and	  the	  question	  of	  myopia.	  Neuron,	  2004.	  43(4):	  p.	  447-­‐68.	  70.	   Morreale,	   B.,	   What	   Is	   Macular	   Degeneration.	   2015,	   BrightFocus	   Foundation:	   p.	   The	   Normal	   Macula	   Compared	   to	   Wet	   and	   Dry	   Macular	  Degeneration.	  71.	   Anderson,	  D.H.,	  et	  al.,	  A	  role	  for	  local	  inflammation	  in	  the	  formation	  of	  drusen	  in	  the	  aging	  eye.	  Am	  J	  Ophthalmol,	  2002.	  134(3):	  p.	  411-­‐31.	  72.	   Umeda,	   S.,	   et	   al.,	  Molecular	   composition	   of	   drusen	   and	   possible	   involvement	   of	   anti-­‐retinal	  autoimmunity	  in	  two	  different	  forms	  of	  macular	  degeneration	  in	  cynomolgus	  monkey	  (Macaca	  fascicularis).	  FASEB	  J,	  2005.	  19(12):	  p.	  1683-­‐5.	  73.	   Johnson,	  L.V.,	  et	  al.,	  Cell	  culture	  model	  that	  mimics	  drusen	  formation	  and	  triggers	  complement	  activation	  associated	  with	  age-­‐related	  macular	  degeneration.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  2011.	  108(45):	  p.	  18277-­‐82.	  74.	   Jager,	   R.D.,	   W.F.	   Mieler,	   and	   J.W.	   Miller,	   Age-­‐related	  macular	   degeneration.	   N	   Engl	   J	   Med,	  2008.	  358(24):	  p.	  2606-­‐17.	  75.	   Sarks,	  J.P.,	  S.H.	  Sarks,	  and	  M.C.	  Killingsworth,	  Evolution	  of	  soft	  drusen	  in	  age-­‐related	  macular	  degeneration.	  Eye	  (Lond),	  1994.	  8	  (	  Pt	  3):	  p.	  269-­‐83.	  76.	   Pauleikhoff,	  D.,	   et	   al.,	  Aging	  changes	  in	  Bruch's	  membrane.	  A	  histochemical	  and	  morphologic	  study.	  Ophthalmology,	  1990.	  97(2):	  p.	  171-­‐8.	  77.	   Holz,	  F.G.,	  et	  al.,	  Pathogenesis	  of	  lesions	  in	  late	  age-­‐related	  macular	  disease.	  Am	  J	  Ophthalmol,	  2004.	  137(3):	  p.	  504-­‐10.	  78.	   Hageman,	   G.S.,	   Age-­‐Related	   Macular	   Degeneration	   (AMD)	   by	   Gregory	   S.	   Hageman	   â“	  Webvision.	  2015.	  2015.	  79.	   Karmel,	  M.	  (2013)	  Dry	  AMD	  Research	  Moves	  Ahead:	  Clinical	  Trials	  and	  New	  Leads.	  80.	   Bowes	   Rickman,	   C.,	   et	   al.,	   Dry	   age-­‐related	  macular	   degeneration:	   mechanisms,	   therapeutic	  targets,	  and	  imaging.	  Invest	  Ophthalmol	  Vis	  Sci,	  2013.	  54(14):	  p.	  ORSF68-­‐80.	  81.	   Campos,	  M.,	   et	   al.,	  A	  novel	   imaging	  technique	  for	  experimental	  choroidal	  neovascularization.	  Invest	  Ophthalmol	  Vis	  Sci,	  2006.	  47(12):	  p.	  5163-­‐70.	  82.	   Chen,	  H.,	   et	   al.,	  The	  aged	  retinal	  pigment	  epithelium/choroid:	  a	  potential	  substratum	  for	  the	  pathogenesis	  of	  age-­‐related	  macular	  degeneration.	  PLoS	  One,	  2008.	  3(6):	  p.	  e2339.	  83.	   Ambati,	   J.,	   J.P.	  Atkinson,	  and	  B.D.	  Gelfand,	  Immunology	  of	  age-­‐related	  macular	  degeneration.	  Nat	  Rev	  Immunol,	  2013.	  13(6):	  p.	  438-­‐451.	  84.	   Ferrara,	   N.,	  Vascular	  endothelial	  growth	   factor	  and	  age-­‐related	  macular	  degeneration:	   from	  basic	  science	  to	  therapy.	  Nat	  Med,	  2010.	  16(10):	  p.	  1107-­‐11.	  85.	   Tadayoni,	   R.,	   Choroidal	   Neovascularization	   Induces	   Retinal	   Edema	   and	   its	   Treatment	  Addresses	  this	  Problem.	  J	  Ophthalmic	  Vis	  Res,	  2014.	  9(4):	  p.	  405-­‐6.	  86.	   Kesen,	   M.R.C.S.W.,	   Choroidal	   Neovascularization:	   Practice	   Essentials,	   Background,	  Pathophysiology,	   in	  The	  Retina	  and	  its	  Disorders.	  2010,	  Academic	  Press:	  San	  Diego,	  CA,	  USA.	  p.	  87-­‐95.	  87.	   de	   Jong,	   P.T.V.M.,	   Age-­‐Related	   Macular	   Degeneration.	   New	   England	   Journal	   of	   Medicine,	  2006.	  355(14):	  p.	  1474-­‐1485.	  	  	   87	  88.	   Zipfel,	  P.F.,	  N.	  Lauer,	  and	  C.	  Skerka,	  The	  role	  of	  complement	  in	  AMD.	  Adv	  Exp	  Med	  Biol,	  2010.	  703:	  p.	  9-­‐24.	  89.	   Klein,	   R.J.,	   et	   al.,	   Complement	   factor	   H	   polymorphism	   in	   age-­‐related	  macular	   degeneration.	  Science,	  2005.	  308(5720):	  p.	  385-­‐9.	  90.	   Pangburn,	   M.K.,	   M.A.	   Atkinson,	   and	   S.	   Meri,	   Localization	   of	   the	   heparin-­‐binding	   site	   on	  complement	  factor	  H.	  J	  Biol	  Chem,	  1991.	  266(25):	  p.	  16847-­‐53.	  91.	   Bhutto,	   I.A.,	   et	   al.,	  C-­‐reactive	  protein	  and	  complement	   factor	  H	   in	  aged	  human	  eyes	  and	  eyes	  with	  age-­‐related	  macular	  degeneration.	  Br	  J	  Ophthalmol,	  2011.	  95(9):	  p.	  1323-­‐30.	  92.	   Thompson,	   D.,	   M.B.	   Pepys,	   and	   S.P.	  Wood,	  The	  physiological	   structure	   of	   human	  C-­‐reactive	  protein	  and	  its	  complex	  with	  phosphocholine.	  Structure,	  1999.	  7(2):	  p.	  169-­‐77.	  93.	   Chen,	  L.J.,	  et	  al.,	  Association	  of	  complement	  factor	  H	  polymorphisms	  with	  exudative	  age-­‐related	  macular	  degeneration.	  Mol	  Vis,	  2006.	  12:	  p.	  1536-­‐42.	  94.	   Seddon,	   J.M.,	  et	  al.,	  Rare	  variants	  in	  CFI,	  C3	  and	  C9	  are	  associated	  with	  high	  risk	  of	  advanced	  age-­‐related	  macular	  degeneration.	  Nat	  Genet,	  2013.	  45(11):	  p.	  1366-­‐70.	  95.	   van	  de	  Ven,	   J.P.,	   et	   al.,	  A	  functional	  variant	  in	  the	  CFI	  gene	  confers	  a	  high	  risk	  of	  age-­‐related	  macular	  degeneration.	  Nat	  Genet,	  2013.	  45(7):	  p.	  813-­‐7.	  96.	   Nishiguchi,	   K.M.,	   et	   al.,	   C9-­‐R95X	   polymorphism	   in	   patients	   with	   neovascular	   age-­‐related	  macular	  degeneration.	  Invest	  Ophthalmol	  Vis	  Sci,	  2012.	  53(1):	  p.	  508-­‐12.	  97.	   Spencer,	   K.L.,	   et	   al.,	  Protective	   effect	   of	   complement	   factor	  B	   and	   complement	   component	   2	  variants	  in	  age-­‐related	  macular	  degeneration.	  Hum	  Mol	  Genet,	  2007.	  16(16):	  p.	  1986-­‐92.	  98.	   Gold,	   B.,	   et	   al.,	   Variation	   in	   factor	   B	   (BF)	   and	   complement	   component	   2	   (C2)	   genes	   is	  associated	  with	  age-­‐related	  macular	  degeneration.	  Nat	  Genet,	  2006.	  38(4):	  p.	  458-­‐62.	  99.	   Bradley,	  D.T.,	  P.F.	  Zipfel,	  and	  A.E.	  Hughes,	  Complement	  in	  age-­‐related	  macular	  degeneration:	  a	  focus	  on	  function.	  Eye	  (Lond),	  2011.	  25(6):	  p.	  683-­‐93.	  100.	   Anderson,	   D.H.,	   et	   al.,	   The	   pivotal	   role	   of	   the	   complement	   system	   in	   aging	   and	   age-­‐related	  macular	  degeneration:	  hypothesis	  re-­‐visited.	  Prog	  Retin	  Eye	  Res,	  2010.	  29(2):	  p.	  95-­‐112.	  101.	   Wang,	   L.,	   et	   al.,	  Abundant	   lipid	  and	  protein	  components	  of	  drusen.	   PLoS	  One,	   2010.	  5(4):	   p.	  e10329.	  102.	   Mullins,	  R.F.,	  N.	  Aptsiauri,	  and	  G.S.	  Hageman,	  Structure	  and	  composition	  of	  drusen	  associated	  with	   glomerulonephritis:	   implications	   for	   the	   role	   of	   complement	   activation	   in	   drusen	  biogenesis.	  Eye	  (Lond),	  2001.	  15(Pt	  3):	  p.	  390-­‐5.	  103.	   Nozaki,	   M.,	   et	   al.,	   Drusen	   complement	   components	   C3a	   and	   C5a	   promote	   choroidal	  neovascularization.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  2006.	  103(7):	  p.	  2328-­‐33.	  104.	   Hoh	   Kam,	   J.,	   et	   al.,	   Complement	   component	   C3	   plays	   a	   critical	   role	   in	   protecting	   the	   aging	  retina	  in	  a	  murine	  model	  of	  age-­‐related	  macular	  degeneration.	  Am	  J	  Pathol,	  2013.	  183(2):	  p.	  480-­‐92.	  105.	   Garcia	  Filho,	  C.A.,	  et	  al.,	  Change	  in	  drusen	  volume	  as	  a	  novel	  clinical	  trial	  endpoint	  for	  the	  study	  of	   complement	   inhibition	   in	   age-­‐related	   macular	   degeneration.	   Ophthalmic	   Surg	   Lasers	  Imaging	  Retina,	  2014.	  45(1):	  p.	  18-­‐31.	  106.	   Bora,	   P.S.,	   et	   al.,	   Role	   of	   complement	   and	   complement	   membrane	   attack	   complex	   in	   laser-­‐induced	  choroidal	  neovascularization.	  J	  Immunol,	  2005.	  174(1):	  p.	  491-­‐7.	  107.	   Hageman,	  G.S.,	  et	  al.,	  An	  integrated	  hypothesis	  that	  considers	  drusen	  as	  biomarkers	  of	  immune-­‐mediated	  processes	  at	  the	  RPE-­‐Bruch's	  membrane	  interface	  in	  aging	  and	  age-­‐related	  macular	  degeneration.	  Prog	  Retin	  Eye	  Res,	  2001.	  20(6):	  p.	  705-­‐32.	  108.	   Ebrahimi,	   K.B.,	   et	   al.,	  Decreased	  membrane	   complement	   regulators	   in	   the	   retinal	  pigmented	  epithelium	  contributes	  to	  age-­‐related	  macular	  degeneration.	   J	   Pathol,	   2013.	  229(5):	   p.	   729-­‐42.	  109.	   Johnson,	  L.V.,	   et	   al.,	  Complement	  activation	  and	  inflammatory	  processes	  in	  Drusen	  formation	  and	  age	  related	  macular	  degeneration.	  Exp	  Eye	  Res,	  2001.	  73(6):	  p.	  887-­‐96.	  	  	   88	  110.	   Nowak,	   J.Z.,	   Age-­‐Related	   Macular	   Degeneration	   (AMD):	   Pathogenesis	   and	   Therapy.	  Pharmacological	  Reports,	  2006.	  58:	  p.	  353-­‐363.	  111.	   Bonekamp,	   N.A.,	   et	   al.,	   Reactive	   oxygen	   species	   and	   peroxisomes:	   struggling	   for	   balance.	  Biofactors,	  2009.	  35(4):	  p.	  346-­‐55.	  112.	   Tate,	  D.J.,	  Jr.,	  M.V.	  Miceli,	  and	  D.A.	  Newsome,	  Zinc	  induces	  catalase	  expression	  in	  cultured	  fetal	  human	  retinal	  pigment	  epithelial	  cells.	  Curr	  Eye	  Res,	  1997.	  16(10):	  p.	  1017-­‐23.	  113.	   Beatty,	   S.,	   et	   al.,	   The	   role	   of	   oxidative	   stress	   in	   the	   pathogenesis	   of	   age-­‐related	   macular	  degeneration.	  Surv	  Ophthalmol,	  2000.	  45(2):	  p.	  115-­‐34.	  114.	   Thornton,	   J.,	   et	   al.,	   Smoking	  and	  age-­‐related	  macular	  degeneration:	   a	   review	  of	   association.	  Eye	  (Lond),	  2005.	  19(9):	  p.	  935-­‐44.	  115.	   Association,	  A.O.,	  Antioxidants	  &	  Age-­‐Related	  Eye	  Disease.	  2015.	  2015.	  116.	   A	  randomized,	  placebo-­‐controlled,	  clinical	  trial	  of	  high-­‐dose	  supplementation	  with	  vitamins	  C	  and	   E	   and	   beta	   carotene	   for	   age-­‐related	   cataract	   and	   vision	   loss:	   AREDS	   report	   no.	   9.	   Arch	  Ophthalmol,	  2001.	  119(10):	  p.	  1439-­‐52.	  117.	   A	  randomized,	  placebo-­‐controlled,	  clinical	  trial	  of	  high-­‐dose	  supplementation	  with	  vitamins	  C	  and	  E,	   beta	   carotene,	   and	   zinc	   for	  age-­‐related	  macular	  degeneration	  and	   vision	   loss:	  AREDS	  report	  no.	  8.	  Arch	  Ophthalmol,	  2001.	  119(10):	  p.	  1417-­‐36.	  118.	   Reuter,	  S.,	   et	  al.,	  Oxidative	  stress,	  inflammation,	  and	  cancer:	  how	  are	  they	  linked?	  Free	  Radic	  Biol	  Med,	  2010.	  49(11):	  p.	  1603-­‐16.	  119.	   Thurman,	   J.M.,	   et	   al.,	   Oxidative	   stress	   renders	   retinal	   pigment	   epithelial	   cells	   susceptible	   to	  complement-­‐mediated	  injury.	  J	  Biol	  Chem,	  2009.	  284(25):	  p.	  16939-­‐47.	  120.	   Monsinjon,	   T.,	   et	   al.,	   Regulation	   by	   complement	   C3a	   and	   C5a	   anaphylatoxins	   of	   cytokine	  production	  in	  human	  umbilical	  vein	  endothelial	  cells.	  FASEB	  J,	  2003.	  17(9):	  p.	  1003-­‐14.	  121.	   Damico,	   F.M.,	   et	   al.,	  New	   approaches	   and	   potential	   treatments	   for	   dry	   age-­‐related	  macular	  degeneration.	  Arq	  Bras	  Oftalmol,	  2012.	  75(1):	  p.	  71-­‐6.	  122.	   Scaldaferri,	   F.,	   et	   al.,	   VEGF-­‐A	   links	   angiogenesis	   and	   inflammation	   in	   inflammatory	   bowel	  disease	  pathogenesis.	  Gastroenterology,	  2009.	  136(2):	  p.	  585-­‐95	  e5.	  123.	   Gerber,	   H.P.,	   et	   al.,	   VEGF	   couples	   hypertrophic	   cartilage	   remodeling,	   ossification	   and	  angiogenesis	  during	  endochondral	  bone	  formation.	  Nat	  Med,	  1999.	  5(6):	  p.	  623-­‐8.	  124.	   Ferrara,	   N.,	   et	   al.,	  Heterozygous	   embryonic	   lethality	   induced	   by	   targeted	   inactivation	   of	   the	  VEGF	  gene.	  Nature,	  1996.	  380(6573):	  p.	  439-­‐42.	  125.	   Chintalgattu,	  V.,	  D.M.	  Nair,	  and	  L.C.	  Katwa,	  Cardiac	  myofibroblasts:	  a	  novel	  source	  of	  vascular	  endothelial	   growth	   factor	   (VEGF)	  and	   its	   receptors	  Flt-­‐1	  and	  KDR.	   J	  Mol	   Cell	   Cardiol,	   2003.	  35(3):	  p.	  277-­‐86.	  126.	   Reichardt,	   L.F.	   and	   K.J.	   Tomaselli,	   Extracellular	   matrix	   molecules	   and	   their	   receptors:	  functions	  in	  neural	  development.	  Annu	  Rev	  Neurosci,	  1991.	  14:	  p.	  531-­‐70.	  127.	   Gragoudas,	  E.S.,	  et	  al.,	  Pegaptanib	  for	  neovascular	  age-­‐related	  macular	  degeneration.	  N	  Engl	  J	  Med,	  2004.	  351(27):	  p.	  2805-­‐16.	  128.	   Emerson,	   M.V.	   and	   A.K.	   Lauer,	   Current	   and	   emerging	   therapies	   for	   the	   treatment	   of	   age-­‐related	  macular	  degeneration.	  Clin	  Ophthalmol,	  2008.	  2(2):	  p.	  377-­‐88.	  129.	   Ferrara,	   N.,	   K.J.	   Hillan,	   and	   W.	   Novotny,	   Bevacizumab	   (Avastin),	   a	   humanized	   anti-­‐VEGF	  monoclonal	   antibody	   for	   cancer	   therapy.	   Biochem	   Biophys	   Res	   Commun,	   2005.	   333(2):	   p.	  328-­‐35.	  130.	   Rosenfeld,	  P.J.,	  et	  al.,	  Ranibizumab	  for	  neovascular	  age-­‐related	  macular	  degeneration.	  N	  Engl	  J	  Med,	  2006.	  355(14):	  p.	  1419-­‐31.	  131.	   Malgorzata,	   F.	   and	   A.	   Stankiewicz,	   Effectiveness	   of	   ranibizumab	   intravitreal	   injections	   for	  exudative	  age-­‐related	  macular	  degeneration	   treatment:	  12-­‐month	  outcomes.	  Med	   Sci	  Monit,	  2011.	  17(9):	  p.	  CR485-­‐90.	  132.	   Lucentis	  Vs	  Avastin	  |	  Macular	  Degeneration	  Treatment	  Options	  |	  AMD.	  2015.	  2015.	  	  	   89	  133.	   Ratnapriya,	  R.	  and	  E.Y.	  Chew,	  Age-­‐related	  macular	  degeneration-­‐clinical	  review	  and	  genetics	  update.	  Clin	  Genet,	  2013.	  84(2):	  p.	  160-­‐6.	  134.	   Kaiser,	   P.K.,	  Emerging	  therapies	   for	  neovascular	  age-­‐related	  macular	  degeneration:	  drugs	   in	  the	  pipeline.	  Ophthalmology,	  2013.	  120(5	  Suppl):	  p.	  S11-­‐5.	  135.	   Nuzzi,	  R.	  and	  F.	  Tridico,	  Local	  and	  systemic	  complications	  after	  intravitreal	  administration	  of	  anti-­‐vascular	  endothelial	  growth	  factor	  agents	  in	  the	  treatment	  of	  different	  ocular	  diseases:	  a	  five-­‐year	  retrospective	  study.	  Semin	  Ophthalmol,	  2015.	  30(2):	  p.	  129-­‐35.	  136.	   Martin,	   D.F.,	   et	   al.,	  Ranibizumab	  and	  bevacizumab	   for	   treatment	   of	   neovascular	  age-­‐related	  macular	  degeneration:	  two-­‐year	  results.	  Ophthalmology,	  2012.	  119(7):	  p.	  1388-­‐98.	  137.	   Kurihara,	  T.,	   et	   al.,	  Targeted	  deletion	  of	  Vegfa	  in	  adult	  mice	  induces	  vision	  loss.	   J	  Clin	   Invest,	  2012.	  122(11):	  p.	  4213-­‐7.	  138.	   Wong,	  W.L.,	  et	  al.,	  Global	  prevalence	  of	  age-­‐related	  macular	  degeneration	  and	  disease	  burden	  projection	   for	   2020	   and	   2040:	   a	   systematic	   review	   and	   meta-­‐analysis.	   Lancet	   Glob	   Health,	  2014.	  2(2):	  p.	  e106-­‐16.	  139.	   Espinosa-­‐Heidmann,	  D.G.,	  et	  al.,	  Age	  as	  an	  independent	  risk	  factor	  for	  severity	  of	  experimental	  choroidal	  neovascularization.	  Invest	  Ophthalmol	  Vis	  Sci,	  2002.	  43(5):	  p.	  1567-­‐73.	  140.	   Cohn,	   C.A.,	   et	   al.,	   Quantifying	   hydrogen	   peroxide	   in	   iron-­‐containing	   solutions	   using	   leuco	  crystal	  violet.	  Geochemical	  Transactions,	  2005.	  6(3):	  p.	  47.	  141.	   Cell	  Counting	  with	  a	  Hemocytometer:	  Easy	  as	  1,	  2,	  3.	  2014.	  142.	   Smith,	   S.A.,	   et	   al.,	   Polyphosphate	   exerts	   differential	   effects	   on	   blood	   clotting,	   depending	   on	  polymer	  size.	  Blood,	  2010.	  116(20):	  p.	  4353-­‐9.	  143.	   Lambert,	   V.,	   et	   al.,	   Laser-­‐induced	   choroidal	   neovascularization	   model	   to	   study	   age-­‐related	  macular	  degeneration	  in	  mice.	  Nat	  Protoc,	  2013.	  8(11):	  p.	  2197-­‐211.	  144.	   Giani,	  A.,	  et	  al.,	  In	  vivo	  evaluation	  of	  laser-­‐induced	  choroidal	  neovascularization	  using	  spectral-­‐domain	  optical	  coherence	  tomography.	  Invest	  Ophthalmol	  Vis	  Sci,	  2011.	  52(6):	  p.	  3880-­‐7.	  145.	   Dunn,	   K.C.,	   et	   al.,	   ARPE-­‐19,	   a	   human	   retinal	   pigment	   epithelial	   cell	   line	   with	   differentiated	  properties.	  Exp	  Eye	  Res,	  1996.	  62(2):	  p.	  155-­‐69.	  146.	   Lornejad-­‐Schäfer,	  R.S.C.S.H.F.J.,	  ARPE-­‐19	  retinal	  pigment	  epithelia	  cells	  as	  an	  in	  vitro	  model	  for	  the	  examination	  of	  light	  induced	  cell	  damage	  in	  the	  eye.	  2015,	  zet-­‐centre	   for	  alternative	  and	  complementary	  methods	  to	  animal	  testing	  (Linz)	  (AT).	  147.	   Jin,	   J.,	   et	  al.,	  Vascular	  endothelial	  growth	  factor	  regulates	  primate	  choroid-­‐retinal	  endothelial	  cell	   proliferation	   and	   tube	   formation	   through	   PI3K/Akt	   and	  MEK/ERK	   dependent	   signaling.	  Mol	  Cell	  Biochem,	  2013.	  381(1-­‐2):	  p.	  267-­‐72.	  148.	   Amrite,	   A.C.	   and	   U.B.	   Kompella,	  Celecoxib	   inhibits	  proliferation	  of	   retinal	  pigment	   epithelial	  and	   choroid-­‐retinal	   endothelial	   cells	   by	   a	   cyclooxygenase-­‐2-­‐independent	   mechanism.	   J	  Pharmacol	  Exp	  Ther,	  2008.	  324(2):	  p.	  749-­‐58.	  149.	   Rusovici,	   R.,	   C.J.	   Patel,	   and	   K.V.	   Chalam,	   Bevacizumab	   inhibits	   proliferation	   of	   choroidal	  endothelial	  cells	  by	  regulation	  of	  the	  cell	  cycle.	  Clin	  Ophthalmol,	  2013.	  7:	  p.	  321-­‐7.	  150.	   Lou,	   D.A.	   and	   F.N.	   Hu,	   Specific	   antigen	   and	   organelle	   expression	   of	   a	   long-­‐term	   rhesus	  endothelial	  cell	  line.	  In	  Vitro	  Cell	  Dev	  Biol,	  1987.	  23(2):	  p.	  75-­‐85.	  151.	   Wei,	  N.,	   et	  al.,	  Oxidative	  stress	  diverts	  tRNA	  synthetase	  to	  nucleus	  for	  protection	  against	  DNA	  damage.	  Mol	  Cell,	  2014.	  56(2):	  p.	  323-­‐32.	  152.	   Mates,	  J.M.,	  C.	  Perez-­‐Gomez,	  and	  I.	  Nunez	  de	  Castro,	  Antioxidant	  enzymes	  and	  human	  diseases.	  Clin	  Biochem,	  1999.	  32(8):	  p.	  595-­‐603.	  153.	   Daniel	  B.	  Menzel,	  R.V.,	  Bart	  Ziegler.	   Imaging	  of	  Reactive	  Oxygen	  Species	   in	  Living	  Lung	  Cells	  Exposed	   to	   Oxidizing	   Air	   Pollutants.	   in	   Oxidative	   Damage	   and	   Repair.	   1990.	   Pasadena,	  California:	  Pergamon	  Press.	  154.	   Cooke,	  M.S.,	  et	  al.,	  Oxidative	  DNA	  damage:	  mechanisms,	  mutation,	  and	  disease.	  FASEB	  J,	  2003.	  17(10):	  p.	  1195-­‐214.	  	  	   90	  155.	   Greiber,	  S.,	  et	  al.,	  Effects	  of	  lipoprotein(a)	  on	  mesangial	  cell	  proliferation	  and	  viability.	  Nephrol	  Dial	  Transplant,	  1996.	  11(5):	  p.	  778-­‐85.	  156.	   Zhou,	   J.,	   et	   al.,	  Neutrophils	  compromise	  retinal	  pigment	  epithelial	  barrier	  integrity.	   J	  Biomed	  Biotechnol,	  2010.	  2010:	  p.	  289360.	  157.	   Rizzolo,	  L.J.,	  Development	  and	  role	  of	  tight	  junctions	  in	  the	  retinal	  pigment	  epithelium.	  Int	  Rev	  Cytol,	  2007.	  258:	  p.	  195-­‐234.	  158.	   Narimatsu,	  T.,	  et	  al.,	  Disruption	  of	  cell-­‐cell	  junctions	  and	  induction	  of	  pathological	  cytokines	  in	  the	  retinal	  pigment	  epithelium	  of	  light-­‐exposed	  mice.	  Invest	  Ophthalmol	  Vis	  Sci,	  2013.	  54(7):	  p.	  4555-­‐62.	  159.	   Vinores,	   S.A.,	   et	   al.,	   Cellular	   mechanisms	   of	   blood-­‐retinal	   barrier	   dysfunction	   in	   macular	  edema.	  Doc	  Ophthalmol,	  1999.	  97(3-­‐4):	  p.	  217-­‐28.	  160.	   Lutty,	  G.A.,	  et	  al.,	  Development	  of	  the	  human	  choriocapillaris.	  Eye	  (Lond),	  2010.	  24(3):	  p.	  408-­‐15.	  161.	   Nagy,	  A.R.	  and	  T.E.	  Ogden,	  Choroidal	  endothelial	  junctions	  in	  primates.	  Eye	  (Lond),	  1990.	  4	  (	  Pt	  2):	  p.	  290-­‐302.	  162.	   Bailey,	  T.A.,	  et	  al.,	  Oxidative	  stress	  affects	  the	  junctional	  integrity	  of	  retinal	  pigment	  epithelial	  cells.	  Invest	  Ophthalmol	  Vis	  Sci,	  2004.	  45(2):	  p.	  675-­‐84.	  163.	   Edgell,	  C.J.,	  C.C.	  McDonald,	  and	  J.B.	  Graham,	  Permanent	  cell	  line	  expressing	  human	  factor	  VIII-­‐related	  antigen	  established	  by	  hybridization.	  Proc	  Natl	  Acad	  Sci	  U	  S	  A,	  1983.	  80(12):	  p.	  3734-­‐7.	  164.	   Carter,	  A.M.,	  Complement	  activation:	  an	  emerging	  player	  in	  the	  pathogenesis	  of	  cardiovascular	  disease.	  Scientifica	  (Cairo),	  2012.	  2012:	  p.	  402783.	  165.	   Conway,	  E.M.,	  Reincarnation	  of	  ancient	  links	  between	  coagulation	  and	  complement.	  J	  Thromb	  Haemost,	  2015.	  13	  Suppl	  1:	  p.	  S121-­‐32.	  166.	   Di	  Cera,	  E.,	  Thrombin	  as	  procoagulant	  and	  anticoagulant.	  J	  Thromb	  Haemost,	  2007.	  5	  Suppl	  1:	  p.	  196-­‐202.	  167.	   Lechtenberg,	  B.C.,	  S.M.	  Freund,	  and	   J.A.	  Huntington,	  An	  ensemble	  view	  of	  thrombin	  allostery.	  Biol	  Chem,	  2012.	  393(9):	  p.	  889-­‐98.	  168.	   Drake,	  W.T.,	   et	   al.,	  Thrombin	   enhancement	  of	   interleukin-­‐1	  and	   tumor	  necrosis	   factor-­‐alpha	  induced	  polymorphonuclear	  leukocyte	  migration.	  Lab	  Invest,	  1992.	  67(5):	  p.	  617-­‐27.	  169.	   Krisinger,	  M.J.,	  et	  al.,	  Thrombin	  generates	  previously	  unidentified	  C5	  products	  that	  support	  the	  terminal	  complement	  activation	  pathway.	  Blood,	  2012.	  120(8):	  p.	  1717-­‐25.	  170.	   Wojta,	   J.,	   K.	  Huber,	   and	  P.	   Valent,	  New	  aspects	   in	   thrombotic	  research:	  complement	   induced	  switch	   in	   mast	   cells	   from	   a	   profibrinolytic	   to	   a	   prothrombotic	   phenotype.	   Pathophysiol	  Haemost	  Thromb,	  2003.	  33(5-­‐6):	  p.	  438-­‐41.	  171.	   DiScipio,	   R.G.,	   et	   al.,	   The	   structure	   of	   human	   complement	   component	   C7	   and	   the	   C5b-­‐7	  complex.	  J	  Biol	  Chem,	  1988.	  263(1):	  p.	  549-­‐60.	  172.	   Preissner,	  K.P.,	  E.R.	  Podack,	  and	  H.J.	  Muller-­‐Eberhard,	  SC5b-­‐7,	  SC5b-­‐8	  and	  SC5b-­‐9	  complexes	  of	   complement:	   ultrastructure	   and	   localization	   of	   the	   S-­‐protein	   (vitronectin)	   within	   the	  macromolecules.	  Eur	  J	  Immunol,	  1989.	  19(1):	  p.	  69-­‐75.	  173.	   Podack,	  E.R.,	  W.P.	  Kolb,	   and	  H.J.	  Muller-­‐Eberhard,	  The	  SC5b-­‐7	  complex:	  formation,	  isolation,	  properties,	  and	  subunit	  composition.	  J	  Immunol,	  1977.	  119(6):	  p.	  2024-­‐9.	  174.	   Du,	   S.,	   et	   al.,	   Decorin	   inhibits	   angiogenic	   potential	   of	   choroid-­‐retinal	   endothelial	   cells	   by	  downregulating	   hypoxia-­‐induced	   Met,	   Rac1,	   HIF-­‐1alpha	   and	   VEGF	   expression	   in	   cocultured	  retinal	  pigment	  epithelial	  cells.	  Exp	  Eye	  Res,	  2013.	  116:	  p.	  151-­‐60.	  175.	   Vadlapatla,	   R.K.,	   et	   al.,	   Ritonavir	   inhibits	   HIF-­‐1alpha-­‐mediated	   VEGF	   expression	   in	   retinal	  pigment	  epithelial	  cells	  in	  vitro.	  Eye	  (Lond),	  2014.	  28(1):	  p.	  93-­‐101.	  176.	   Rao,	   N.N.	   and	   A.	   Kornberg,	   Inorganic	   polyphosphate	   supports	   resistance	   and	   survival	   of	  stationary-­‐phase	  Escherichia	  coli.	  J	  Bacteriol,	  1996.	  178(5):	  p.	  1394-­‐400.	  	  	   91	  177.	   Cheng,	   Y.	   and	   B.	   Sun,	   Polyphosphate	   kinase	   affects	   oxidative	   stress	   response	   by	  modulating	  cAMP	   receptor	   protein	   and	   rpoS	   expression	   in	   Salmonella	   typhimurium.	   J	   Microbiol	  Biotechnol,	  2009.	  19(12):	  p.	  1527-­‐35.	  178.	   Alcazar-­‐Roman,	   A.R.	   and	   S.R.	  Wente,	   Inositol	   polyphosphates:	   a	   new	   frontier	   for	   regulating	  gene	  expression.	  Chromosoma,	  2008.	  117(1):	  p.	  1-­‐13.	  179.	   Hanakahi,	   L.A.	   and	   S.C.	   West,	   Specific	   interaction	   of	   IP6	   with	   human	   Ku70/80,	   the	   DNA-­‐binding	  subunit	  of	  DNA-­‐PK.	  EMBO	  J,	  2002.	  21(8):	  p.	  2038-­‐44.	  180.	   Steger,	   D.J.,	   et	   al.,	   Regulation	   of	   chromatin	   remodeling	   by	   inositol	   polyphosphates.	   Science,	  2003.	  299(5603):	  p.	  114-­‐6.	  181.	   Alcazar-­‐Roman,	   A.R.,	   T.A.	   Bolger,	   and	   S.R.	   Wente,	   Control	   of	  mRNA	   export	   and	   translation	  termination	  by	  inositol	  hexakisphosphate	  requires	  specific	  interaction	  with	  Gle1.	   J	  Biol	  Chem,	  2010.	  285(22):	  p.	  16683-­‐92.	  182.	   Odom,	   A.R.,	   et	   al.,	   A	   role	   for	   nuclear	   inositol	   1,4,5-­‐trisphosphate	   kinase	   in	   transcriptional	  control.	  Science,	  2000.	  287(5460):	  p.	  2026-­‐9.	  183.	   York,	  J.D.,	  et	  al.,	  A	  phospholipase	  C-­‐dependent	  inositol	  polyphosphate	  kinase	  pathway	  required	  for	  efficient	  messenger	  RNA	  export.	  Science,	  1999.	  285(5424):	  p.	  96-­‐100.	  184.	   Sivakumaran,	  T.A.,	  et	  al.,	  A	  32	  kb	  critical	  region	  excluding	  Y402H	  in	  CFH	  mediates	  risk	  for	  age-­‐related	  macular	  degeneration.	  PLoS	  One,	  2011.	  6(10):	  p.	  e25598.	  185.	   Crooke,	  E.,	  et	  al.,	  Genetically	  altered	  levels	  of	  inorganic	  polyphosphate	  in	  Escherichia	  coli.	  J	  Biol	  Chem,	  1994.	  269(9):	  p.	  6290-­‐5.	  186.	   Ursini,	  F.	  and	  A.	  Bindoli,	  The	  role	  of	  selenium	  peroxidases	  in	  the	  protection	  against	  oxidative	  damage	  of	  membranes.	  Chem	  Phys	  Lipids,	  1987.	  44(2-­‐4):	  p.	  255-­‐76.	  187.	   Baker,	  R.D.,	  et	  al.,	  Selenium	  regulation	  of	  glutathione	  peroxidase	  in	  human	  hepatoma	  cell	  line	  Hep3B.	  Arch	  Biochem	  Biophys,	  1993.	  304(1):	  p.	  53-­‐7.	  188.	   McCord,	   J.M.	   and	   I.	   Fridovich,	  Superoxide	  dismutase.	  An	  enzymic	  function	  for	  erythrocuprein	  (hemocuprein).	  J	  Biol	  Chem,	  1969.	  244(22):	  p.	  6049-­‐55.	  189.	   Sentman,	   M.L.,	   et	   al.,	   Phenotypes	   of	   mice	   lacking	   extracellular	   superoxide	   dismutase	   and	  copper-­‐	  and	  zinc-­‐containing	  superoxide	  dismutase.	  J	  Biol	  Chem,	  2006.	  281(11):	  p.	  6904-­‐9.	  190.	   Li,	  Y.,	  et	  al.,	  Dilated	  cardiomyopathy	  and	  neonatal	  lethality	  in	  mutant	  mice	  lacking	  manganese	  superoxide	  dismutase.	  Nat	  Genet,	  1995.	  11(4):	  p.	  376-­‐81.	  191.	   Elchuri,	  S.,	  et	  al.,	  CuZnSOD	  deficiency	  leads	  to	  persistent	  and	  widespread	  oxidative	  damage	  and	  hepatocarcinogenesis	  later	  in	  life.	  Oncogene,	  2005.	  24(3):	  p.	  367-­‐80.	  192.	   Galaris,	  D.	  and	  K.	  Pantopoulos,	  Oxidative	  stress	  and	  iron	  homeostasis:	  mechanistic	  and	  health	  aspects.	  Crit	  Rev	  Clin	  Lab	  Sci,	  2008.	  45(1):	  p.	  1-­‐23.	  193.	   Ferrali,	   M.,	   et	   al.,	   Protection	   against	   oxidative	   damage	   of	   erythrocyte	   membrane	   by	   the	  flavonoid	  quercetin	  and	  its	  relation	  to	  iron	  chelating	  activity.	  FEBS	  Lett,	  1997.	  416(2):	  p.	  123-­‐9.	  194.	   Rodrigo,	  R.	   and	  C.	  Bosco,	  Oxidative	  stress	  and	  protective	  effects	  of	  polyphenols:	  comparative	  studies	   in	  human	  and	  rodent	  kidney.	  A	  review.	   Comp	  Biochem	  Physiol	  C	  Toxicol	  Pharmacol,	  2006.	  142(3-­‐4):	  p.	  317-­‐27.	  195.	   Touati,	  D.,	  Iron	  and	  oxidative	  stress	  in	  bacteria.	  Arch	  Biochem	  Biophys,	  2000.	  373(1):	  p.	  1-­‐6.	  196.	   Winterbourn,	  C.C.,	  Toxicity	  of	   iron	  and	  hydrogen	  peroxide:	  the	  Fenton	  reaction.	   Toxicol	  Lett,	  1995.	  82-­‐83:	  p.	  969-­‐74.	  197.	   Turowski,	   P.,	   et	   al.,	   Basement	   membrane-­‐dependent	   modification	   of	   phenotype	   and	   gene	  expression	  in	  human	  retinal	  pigment	  epithelial	  ARPE-­‐19	  cells.	  Invest	  Ophthalmol	  Vis	  Sci,	  2004.	  45(8):	  p.	  2786-­‐94.	  198.	   Mandell,	  K.J.,	  et	  al.,	  Expression	  of	  JAM-­‐A	  in	  the	  human	  corneal	  endothelium	  and	  retinal	  pigment	  epithelium:	   localization	  and	  evidence	   for	  role	   in	  barrier	   function.	   Invest	  Ophthalmol	   Vis	   Sci,	  2007.	  48(9):	  p.	  3928-­‐36.	  	  	   92	  199.	   Huber,	  G.,	  et	  al.,	  Novel	  rodent	  models	  for	  macular	  research.	  PLoS	  One,	  2010.	  5(10):	  p.	  e13403.	  200.	   1:	  Interpretation	  of	  Stereo	  Ocular	  Angiography	  :	  Retinal	  and	  Choroidal	  Anatomy.	  2015.	  201.	   Wu,	   T.,	   J.T.	   Handa,	   and	   J.D.	   Gottsch,	   Light-­‐induced	   oxidative	   stress	   in	   choroidal	   endothelial	  cells	  in	  mice.	  Invest	  Ophthalmol	  Vis	  Sci,	  2005.	  46(4):	  p.	  1117-­‐23.	  202.	   Kalow,	  J.,	  A	  Conjugated	  Polymer	  Fluorogenic	  Probe	  for	  Inorganic	  Polyphosphate.	  2015.	  203.	   Saito,	   K.S.,	   et	   al.,	   Direct	   Labeling	   of	   Polyphosphate	   at	   the	   Ultrastructural	   Level	   in	  Saccharomyces	   cerevisiae	   by	   Using	   the	   Affinity	   of	   the	   Polyphosphate	   Binding	   Domain	   of	  Escherichia	  coli	  Exopolyphosphatase.	  2005.	  204.	   Aschar-­‐Sobbi,	  R.,	  et	  al.,	  High	  sensitivity,	  quantitative	  measurements	  of	  polyphosphate	  using	  a	  new	  DAPI-­‐based	  approach.	  J	  Fluoresc,	  2008.	  18(5):	  p.	  859-­‐66.	  205.	   Bishop,	  P.,	  The	  biochemical	  structure	  of	  mammalian	  vitreous.	  Eye	  (Lond),	  1996.	  10	  (	  Pt	  6):	  p.	  664-­‐70.	  206.	   Donati,	   G.,	   A.D.	   Kapetanios,	   and	   C.J.	   Pournaras,	   Principles	   of	   treatment	   of	   choroidal	  neovascularization	  with	   photodynamic	   therapy	   in	   age-­‐related	  macular	   degeneration.	   Semin	  Ophthalmol,	  1999.	  14(1):	  p.	  2-­‐10.	  207.	   Pennesi,	   M.E.,	   M.	   Neuringer,	   and	   R.J.	   Courtney,	   Animal	   models	   of	   age	   related	   macular	  degeneration.	  Mol	  Aspects	  Med,	  2012.	  33(4):	  p.	  487-­‐509.	  	  	  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items