UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Measurements of cell survival at low doses of radiation Brosing, Juliet Wain 1983

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

Item Metadata

Download

Media
831-UBC_1983_A1 B76.pdf [ 6.74MB ]
Metadata
JSON: 831-1.0085787.json
JSON-LD: 831-1.0085787-ld.json
RDF/XML (Pretty): 831-1.0085787-rdf.xml
RDF/JSON: 831-1.0085787-rdf.json
Turtle: 831-1.0085787-turtle.txt
N-Triples: 831-1.0085787-rdf-ntriples.txt
Original Record: 831-1.0085787-source.json
Full Text
831-1.0085787-fulltext.txt
Citation
831-1.0085787.ris

Full Text

MEASUREMENTS OF CELL SURVIVAL AT LOW DOSES OF RADIATION by JULIET WAIN BROSING B.Sc, Humboldt State University, 1976 M.Sc. , F l o r i d a State U n i v e r s i t y , 1978 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES THE DEPARTMENT OF PHYSICS We accept t h i s t h e s i s as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA A p r i l 1983 ( c ) J u l i e t Wain Brosing, 1983 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6 (3/81) i i ABSTRACT The effect of low closes of radiation is of primary importance if we wish to understand the basic mechanisms of radiation damage. In vitro experiments performed at clinically relevant doses can also lead to better understanding of radiotherapy protocols and fractionation regimes. The availability of accurate data at low doses can facilitate the examination of survival models which describe dose-effect relationships. Most cellular radiobiology experiments are performed at high doses (3 to 30 Cray). The errors in these experiments, while acceptable at high doses, are too large to allow determination of radiobiological parameters, such as oxygen enhancement ratio (OER) and relative biological effective-ness (RBE), at low doses (0 to 3 Cray). These experiments are limited in the low dose region because we are measuring only the surviving fraction, in a population of predominantly surviving cells, and because there is an uncertainty of 10 to 15% in the number of cells plated. We have developed a technique to assay cell survival at low doses. The exact number of cells plated is determined microscopically. After incubation, the number of killed cells and the number of surviving cells are both determined, by microscopic examination. While extremely labor intensive, this technique yields survival data, in the low dose region, which is much more accurate than the data obtained using classical meth-ods. This technique can be used to measure many radiobiological parame-ters. We have chosen to examine the effect of oxygen at low doses. Our results clearly demonstrate that, for asynchronous Chinese Hamster Ovary (CHO) cells, the radiosensitizing effect of oxygen is reduced at lower doses. A Picker X-ray source (280 kVp, HVL 1.7 mm Cu) was used for i i i these experiments. The choice of a surv iva l model has important implications in the low dose region. T h e predict ions of three d i f ferent surv iva l models regarding the effect of oxygen at low doses are d i scussed. Th i s technique can be used to complement the classical "h igh dose assay" to obtain data that encompasses a large dose range. Th i s will be especially va luable, for example, when attempting to ful ly descr ibe rad io-biological parameters, or when examining surv iva l models. i v TABLE OF CONTENTS ABSTRACT i i TABLE OF CONTENTS •". i v LIST OF TABLES v i i LIST OF FIGURES v i i i ACKNOWLEDGEMENTS . x 1.0 INTRODUCTION 1 1.1 RADIATION 1 1.2 LOW DOSES OF RADIATION 7 1.2.1 D e f i n i t i o n of shoulder 7 1.2.2 Su r v i v a l models 8 1.2.3 Importance of low doses 17 1.3 ENDPOINTS USED TO ASSAY LOW DOSES 21 1.4 THE TECHNIQUE 23 1.5 OXYGEN EFFECT AT LOW DOSES 28 2.0 MATERIALS AND METHODS . 32 2.1 CULTURE TECHNIQUE . . . . . 32 2.2 SYNCHRONY TECHNIQUE 32 2.3 OER TECHNIQUE 38 2.4 IRRADIATION OF CELLS ,44 3.0 RESULTS 47 3.1 DEVELOPMENT OF TECHNIQUE TO ASSAY CELL SURVIVAL AT LOW DOSES .47 3.1.1 Microscopic examination and elim i n a t i o n of errors . . . 47 V TABLE OF CONTENTS (continued) Page 3.1.2 Microtest p l a t e s 48 3.1.3 Endpoint 53 3.1.4 Doublets 53 3.1.5 Scoring 55 3.2 THE OXYGEN EFFECT AT LOW DOSES 62 3.2.1 OER-Asynchronous 62 3.2.2 OER-Synchronous 69 4.0 DISCUSSION 80 4.1 TECHNIQUE 80 4.1.1 Uses of Technique 80 4.1.2 I d e n t i f i c a t i o n of c e l l s 80 4.1.3 P l a t i n g E f f i c i e n c y . . . . . 81 4.1.3.1 Factors e f f e c t i n g the P l a t i n g E f f i c i e n c y . . . 81 4.1.3.2 Debris 83 4.1.3.3 Unattached c e l l s 85 4.1.4 Feeder c e l l s 86 4.1.5 Improvements and future plans 87 4.1.5.1 Automation 87 4.1.5.2 Synchronized experiments 87 4.1.5.3 Other p o s s i b i l i t i e s 88 4.1.6 Limitations 90 4.2 OXYGEN EXPERIMENTS . 91 4.2.1 The importance of OER at low doses 91 v i TABLE OF CONTENTS (continued) Page 4.2.2 Predictions for 0ER from s u r v i v a l models 92 4.2.2.1 SH-MT 93 4.2.2.2 RMR 97 4.2.2.3 L-Q 103 4.2.2.4 The importance of s u r v i v a l models at low doses.108 4.2.3 Predictions f o r OER from the mechanistic viewpoint . 110 4.2.4 Asynchronous Data 113 4.2.5 Synchronous Data 118 4.2.5.1 Limitations of Data 118 4.2.5.2 OER f o r synchronous c e l l s 118 4.2.6 Oxygen Concentration 119 5.0 CONCLUSIONS 121 6.0 BIBLIOGRAPHY . . •. . 122 vi i LIST OF TABLES Table T i t l e Page I Some Equations for Cell Survival 11 II LET Values for a few heavy ion beams 15 III Protocol for determining mitotic index 36 IV Protocol for TCA precipitation of DNA 37 V Specifications for petri dishes and microtest plates 49 VI Plating efficiencies for day 0 and day 2 59 VII Survival, with standard errors, using both assays 62 VIII Plating efficiencies for low and high dose assays 83 IX L-Q parameters for Koch's 1979 data 107 X Parameter ratios for data in Table IX 108 XI L-Q Parameters for asynchronous CHO cells 115 XII Parameter ratios for data in Table XI 115 v i i i LIST OF FIGURES Figure T i t l e Page 1 Schematic Diagram of DNA and electron tracks 6 2 T y p i c a l Dose-response Relationships 8 3 F r a c t i o n a l Errors i n S and K 25 4 The Relationship between the percentage error i n S and the number of colonies for both Poisson and binomial s t a t i s t i c s 27 5 C e l l number as a function of shake-off 35 6 R a d i o s e n s i t i v i t y as a function of c e l l cycle 43 7 I r r a d i a t i o n Set-up 46 8 Colonies i n Microtest Plates and P e t r i s 52 9 Survival of CHO c e l l s plated i n microtest p l a t e s and p e t r i dishes with etched g r i d 54 10 S u r v i v a l of CHO c e l l s with doublet error 56 11 The e f f e c t on s u r v i v a l of changing the i n i t i a l day of scoring 60 12 High dose s u r v i v a l of asynchronous CHO c e l l s 63 13 Low dose s u r v i v a l of asynchronous CHO c e l l s 64 14 Oxygen e f f e c t for asynchronous CHO c e l l s 66 15 Asynchronous CHO c e l l s 68 16 Survival curves for aerobic CHO c e l l s at low doses 70 17 Survival curves for hypoxic CHO c e l l s at low doses 71 18 Low dose s u r v i v a l of CHO c e l l s i n Gl/S stage of the c e l l cycle 73 19 Oxygen e f f e c t of Gl/S CHO c e l l s 74 20 CHO c e l l s i n Gl/S 75 21 Low dose s u r v i v a l of CHO c e l l s i n l a t e S stage of the c e l l cycle 76 22 Oxygen e f f e c t for CHO c e l l s i n late S 77 ix LIST OF FIGURES (continued) Figure T i t l e Page 23 Late S CHO c e l l s 78 24 OER as a function of dose for the SH-MT model 96 25 OER as a function of dose f o r the RMR model '102 26 OER as a function of dose for the L-Q model 105 27 OER as a function of dose for 3 d i f f e r e n t s u r v i v a l models 109 28 OER values and the L-Q equation 114 29 OER values and the SH-MT equation. 116 X A C K N O W L E D G E M E N T S I have d iscovered that a Ph .D . thesis is not accomplished in isolation. Th i s thesis has involved many people in many dif ferent ways, i t would not be poss ib le, without writ ing a complete thesis on the subject, to adequately explain the contr ibut ions and thank the many people who have helped me. However, I will attempt to name a few. / I was supported as a student on Dr s . Palcic and Skarsgard ' s NCI g rant and I would like to express my grat i tude to the National Cancer Institute of Canada and the B . C . Cancer Foundation for their financial support in the course of this work. I would like to thank all the staff at the B . C . Cancer Research Cent re for the ir encouraging words and f r iendly smiles. T h e y made the Research Cent re a v e r y pleasant place in which to work. I would like to thank D r . F r e d C ran s ton , my f i r s t phys i c s teacher i n ' un i ve r s i t y , for conv inc ing me years ago, when I was young and fool ish, that a Ph .D . in phys ic s was a worthwhile and possible endeavor. I would like to thank D r . Branko Palcic, my superv i so r , for the many hours he spent with me and all the help he unsel f i sh ly gave, but especial ly for being more of an eternal optimist than I was. I would like to express my grat i tude to Dr . L loyd Skar sgard for being so thorough and meticulous. Dr . K i rs ten Skov for patiently explaining DNA damage to me. D r . Dick Kornelson for thoroughly descr ib ing the phys ic s of radiation to me, David Noble for all his f r iendly l ib rary he lp, Bev Ersoy for teaching me how to use the word processor , and D r . Mladen Korbel ik and Stuart Berger for many interest ing d iscuss ions (some of which pertained to radiobio logy). x i My thanks to Diane Wurst and Isabel Harrison who both provided sanity maintainence when the going got rough and laughter for the lighter times. I would like to recognise my dear friend Ann Hanham (who was also completing her Ph.D. at the same time) for the many frustrations and achievements we shared. An enormous amount of credit belongs to Keith LeComte, my partner in life, without whom I could not have completed this work. He deserves thanks for the encouragement and assistance he gave me, for the many dinners he brought me, for all the times he was there when I needed him. I would like to thank my family, Keith's family and our many, many friends for never doubting that this thesis would someday be completed. Their optimism and encouragement helped me over many a difficult period. To all of the above people, but especially to Keith LeComte, I dedicate this thesis. x i i T H E SC I ENT I ST does not s tudy nature because it is useful -She studies it because she delights in it. A n d she del ights in it because it is beaut i fu l . If nature were not beaut i fu l , it would not be worth knowing. A n d if nature were not worth knowing, life would not be worth l iv ing. — A d a p t e d from a saying by Henri Poincare 1 1. I N T R O D U C T I O N 1 .1 RAD IAT ION Mank ind is cons tant l y bombarded with rad iat ion from natura l and man made sources . The natu ra l b a c k g r ound rad iat ion is composed p r i n c i pa l l y of (1) cosmic r a y s , (2) emiss ions from the d i s i n teg ra t i on of u r an i um, tho r i um, radium and other rad ioact ive elements in the ea r th ' s c r u s t , and (3) emiss ions from potass ium 40, carbon 14, and other rad ioact ive isotopes o c c u r r i n g na tu ra l l y in the body . The backg round dose in an average pe r son ' s l i fe span is e s sent ia l l y doubled by man made sources of r a -d i a t i on , c h i e f l y medical (Upton 1982). The ef fect s of rad iat ion on b i o l o g i -cal systems are the re fo re of persona l importance to each of us . T h i s is espec ia l l y important because we have no natu ra l wa rn ing system to i n d i -cate when danger from rad iat ion is p re sen t . Radiat ion can be d i v i ded into two ca tegor ie s , non - i on i z i ng and i o n -i z ing r ad i a t i on . Non - i on i z i ng r ad i a t i on , such as l i g h t , g i ve s enough energy to the e lect rons to exc i te molecules and rea r range chemical bonds . Ioniz ing rad iat ion (A lpha pa r t i c l e s . X - r a y s and pions are al l examples) causes damage by impart ing enough ene rgy to e lect rons in the i r r ad i a ted matter to allow them to ion ize. T h i s means that the e lect rons are removed from the nuc le i they were fo rmer ly at tached to , caus ing chemical bonds to be b roken and/or changed . The i n te rac t i on of these e lect rons w i th mole-cules of water creates long l i ved react i ve rad ica l s wh ich cause add i t iona l damage. The ionized e lect rons can also impart some of t he i r ene r gy to o ther e l e c t r on s , c rea t i ng a cha in react ion and caus ing more ion i zat ions , f ree ing more e lec t rons w h i c h , in t u r n , cause more damage. Th i s process cont inues un t i l all of the ene rgy ava i lab le is e xpended . A l t hough both 2 ionizing and non-ionizing radiation can cause damage to biological matter, there are distinct differences in the mechanisms and the type of damage inflicted by the two types. This thesis will be concerned only with ioniz-ing radiation. X-rays were discovered by W.K. Roentgen in 1895, and were quickly used for medical purposes. One of the first side effects that became apparent was reddening and blistering of the skin. By 1902 cancer was recognized as a delayed side effect of the radiation, usually occurring on the physicists' or other radiation workers' hands (Rona 1979). This cancer was associated with gross injury to the skin induced by fairly high radiation doses. The possibility of an association between low ra-diation doses and an increased incidence of cancers was recognized by E. B. Lewis of the California Institute of Technology, among others (e.g. Sacks & Seeman 1947, Damashek & Cunz 1957, Lewis 1957). Lewis demon-strated that groups of exposed people (the atomic bomb survivors in Japan, radiologists, and non-cancerous patients treated with X-rays) showed an increased incidence of leukemia. This obviously caused consid-erable concern. Subsequently many questions arose (and in fact remain with us today) concerning the cancer risk associated with low-level ra-diation (UNSCEAR 1958, NCRP No. 64 1980, UNSCEAR 1982). To answer these questions more knowledge is needed about the basic mechanisms of radiation damage. To obtain this knowledge, meaningful experiments using low doses of radiation must be performed. Due to technical difficulties and statistical limitations most cellular radiobiology experiments have been performed at relatively high radiation doses. Therefore, estimates of cell killing at low radiation doses have predominately come from the data obtained at high doses, using 3 ext rapo la t i on methods. Th i s a s sumpt ion, that low dose estimates of cel l k i l l i n g can be ex t rapo la ted from the data with h igh doses, has ra re l y been te s ted . Exper iments per formed with low doses of radiat ion d i rec ted at te s t ing th i s assumption would be inva luab le . A n ' endpo int ' is de f ined as the effect that is measured at the end of the exper iment . Many endpo int s have been used to s tudy the ef fects of rad ia t ion ( for example, see E l k i nd & Whitmore 1967), some of which are d i s cu s sed b r i e f l y in sect ion 1.3. The endpo int with wh ich th i s thes i s is mainly concerned is the loss of p r o l i f e r a t i ve capac i ty of mammalian ce l l s . T h i s is one of the ce l l f unc t i on s most sens i t i ve to the ef fect of r ad ia t i on . For normal ly d i v i d i n g ce l l populat ions rad iob io log i s t s genera l l y def ine v i a b i l i t y as the capac i ty of a ce l l to generate a clone of s imilar ce l l s . T h e r e f o r e the terms Moss of v i a b i l i t y ' , ' c e l l d e a t h ' , and ' loss of p r o l -i f e ra t i ve capac i t y ' are often used i n te r changeab l y . In th i s context the ' l e t ha l ' e f fect of rad iat ion is de f ined as that wh ich induces loss of p r o l i f e r a t i v e capac i t y . C o n v e r s e l y , a ' s u r v i v o r ' is de f ined as a cel l w h i c h , a f te r i r r a d i a t i o n , has reta ined its ab i l i t y to generate a clone of l i ke ce l l s . T h e r e f o r e , the ' s u r v i v i n g f r a c t i on ' (u sua l l y des ignated by S) is that f r ac t i on of the cel l populat ion wh ich ' s u r v i v e s ' the rad iat ion t r e a t -ment, those wh ich reta in the ab i l i t y to generate a clone of l ike ce l l s . These de f in i t i on s wi l l be used for the remainder of th i s the s i s . The term 'dose ' as used in th i s thes i s re fe r s to the absorbed dose; i . e . the ene rgy imparted (absorbed) per un i t mass. The Systeme I n t e rna -t ional (SI) un i t is the g r a y ( abbrev i a ted C y ) , wh ich co r re sponds to an ene rgy abso rp t ion of 1 joule per k i logram of i r r ad i a ted material (1 Cy = U / k g ) . P rev i ou s l y the un i t of dose was cal led the r a d , de f ined as an ene rgy abso rp t ion of 100 e rg s/g ram. S ince 1 J = 10 7 e r g s , the g r ay is equ i va len t to 100 rads . 4 When irradiating biological materials with X- or gamma rays we are primarily concerned with the effects of electrons (produced by photons interacting mainly through the Compton and photoelectric effects) in the irradiated material. As an electron traverses the material it causes many ionizations and excitations in the molecules. (For example, a 100 keV electron will produce approximately IO4 ionizations.) These, in turn, cause biological damage. The energy loss resulting from an electron traversing material can be described using the Bethe-Bloch equation: dE/dx = -{HirZ2ek I mV2) * N * ln(2mV 2/l) where dE/dx = energy loss per unit path length (stopping power) Z = charge on incident particle (1,2,3,. . ) V = velocity of incident particle N = number of electrons/unit volume of absorber I = mean ionization potential for electrons in absorber m = mass of electron e = charge of electron x = path length (g/cm2) For high energy electrons the In term must be modified to include relativistic effects. From the above expression it is evident that dE/dx increases with decreasing V. In other words, as an electron slows down, more energy is transferred, and the spacing between ionizations decreases. At the end of an electron's 'path' the ionizations are quite closely spaced. The maximum stopping power for electrons in water is 30.3 keV/ym. This is the stopping power of a lOOeV electron, corresponding to a range of about 4.5 nm (ICRU Report 16). As the electron's energy decreases below 100 eV, it's dE/dx (stopping power) actually decreases. Electrons 5 , with energies greater than 100 eV will gradually decrease in energy, passing through a point where the electron has a maximum stopping power (maximum density of ionizations), which is near the end of the electron's path. If an electron stops within the nucleus of a cell (a CHO cell is about 15 um in diameter, it's nucleus about 5 urn) it will deposit quite densely ionizing radiation. There is experimental evidence to suggest that a double strand break (DSB) in DNA may cause cell inactivation. The DNA helix is about 2 nm in diameter. A double strand break (DSB) in the DNA might well occur if an electron has the maximum stopping power (30.3 keV/um), which corresponds to the production of approximately 1 ion pair per nm. On the other hand, an electron with just enough energy to traverse the cell (a 30 keV electron has a range of about 17 um) will only produce about 1 ion pair every 20 nm, and is therefore not likely to cause a DSB in the DNA. The relative magnitude of the DNA and the density of the ionizations are shown schematically in figure 1. The X-rays used for the experiments in this thesis are produced with a 280 keV potential, and have a half value layer (HVL) of 1.7 mm Cu. The primary beam of photons is attenuated to 50% of its original intensity at a depth of approximately 4 cm in water. The electron spectrum from the photons is very complex. The electrons produced from the primary photon beam have energies from zero to about 146 keV, corresponding to a range in water from zero to about 0.2 mm. 6 p o r t i o n o f DNA h e l i x F i g u r e 1. S C H E M A T I C D I A G R A M O F D N A A N D E L E C T R O N T R A C K S T h e + a n d - s i g n s r e p r e s e n t i o n p a i r s w h i c h a r e p r o d u c e d a l o n g t h e e l e c t r o n ' s p a t h . T h e D N A h e l i x i s a b o u t 2 n m i n d i a m e t e r . O n l y a v e r y s m a l l p o r t i o n o f t h e D N A h e l i x i s r e p r e s e n t e d . 7 1.2 LOW DOSES OF RAD IAT ION 1.2.1 Def in i t ion of ' s hou lde r ' H i s t o r i c a l l y , re la t ionsh ips between ' s u r v i v a l ' and ' do se ' , or ' do se -response ' r e l a t i on sh ip s , were f i r s t obta ined for many s t ra in s of m ic roo rga -nisms ( v i r u s e s , b a c te r i a , y e a s t ) . I r rad iated v i r u s e s have exponent ia l inact ivat ion c u r ve s and some bacter ia also y i e ld th i s t ype of response. The f i r s t examples of non exponent ia l dose - response were i n t e rp re ted as ev idence for mult ip le t a r ge t s . Ano the r i n te rp re ta t i on was the bel ief that any nori exponent ia l dose response must be due to some a r t i f ac t of the expe r iment , for example, c lumping of ce l l s . Severa l decades a f te r obta in ing dose - response re la t ionsh ips for microorgan i sms, techn iques for s co r i ng v iab le mammalian cel l s were d e v e l -oped (Puck & Marcus 1956), a l lowing dose- response re la t ionsh ips to be obta ined fo r mammalian ce l l s . In cont ra s t to microorgan i sms, exponent ia l k i l l i n g is uncommon with mammalian ce l l s ; at low doses of rad iat ion there is p ropo r t i ona l l y less k i l l i n g per un i t dose than at h i ghe r doses. T h e r e -f o r e , s u r v i v a l c u r v e s for mammalian ce l l s , w i th the s u r v i v i n g f rac t ion p lot ted logar i thmica l l y and the dose l i n e a r l y , a re u sua l l y de s c r i bed as hav ing a ' s hou lde r ' (see f i g u r e 2 ) . The shou lder reg ion is genera l l y fol lowed by a reg ion of approx imate ly exponent ia l s u r v i v a l . The terms Mow dose ' and ' s hou lde r r eg i on ' are often used i n te rchangeab l y (and wi l l be in th i s thes i s ) to ind icate th i s por t ion of the cel l s u r v i v a l c u r v e . 8 Dose Figure 2. TYPICAL DOSE-RESPONSE RELATIONSHIPS. 1.2.2 Survival models There are many different models and theories (e.g. see Table I) to describe mammalian cell survival. Most of them are based upon some hypothesis as to the mechanism of radiation damage and some of them result in relatively simple algebraic expressions. These algebraic ex-pressions are useful descriptions of dose-response relationships (cell survival curves). Using these expressions, comparisons can easily be made between parameters describing different cell lines, or between parameters describing different radiation conditions for the same cell line. It must be emphasized that the validity of a model can not be determined solely on the basis of curve fitting. The interpretation of the data is also 9 compl icated by the fact that most models assume a cel l populat ion wh ich is homogeneous w i th re spec t to the ef fect of rad iat ion ( A l p e r 1979). Howev-e r , most exper iments are per formed with heterogeneous cel l populat ions ( fo r example, a s ynch ronous popu la t i on s ) . Tab le I l i s ts a few of the models most commonly used in the l i t e r a -t u r e , though it is by no means complete. The re are at least 20 d i f f e r en t models ava i l ab le , most of them wi th 3 or more parameters . However, even when exper iments are per formed with the utmost c a r e , there is not enough informat ion ava i lab le to spec i f y more than two parameters w i th s i gn i f i can t p rec i s i on (Chapman 1980). Most of the d i s cu s s i on in th i s thes i s wi l l t he re fo re be l imited to a few of the most commonly used two parame-te r models. Before d e r i v i n g some of the s u r v i v a l equat ions shown in Table I i t is i n s t r u c t i v e to cons ide r the s impl iest form of ta rget t heo r y . The process of i nac t i va t i ng the cel l s is a s ta t i s t i ca l problem i f the ion izat ions and the ce l l s are randomly d i s t r i b u t e d , and Poisson s t a t i s t i c s can be used to de sc r i be the inact i va t ion of the ce l l s : P a , n ' = ( a V a ) / n ! where P = p r obab i l i t y of n events o c c u r r i n g w i th in a c e l l , g i ven a is the a, n average number of event s (h i t s ) o c cu r i n g in the ce l l s . If we assume s i ng l e -h i t s i n g l e - t a r ge t t h e o r y , in o ther words a s ing le event is s u f f i c i en t to p roduce the measured e f f e c t , then on ly those ce l l s wh ich rece ived no h i t s wi l l be s u r v i v o r s . For large numbers of e ven t s , S, 10 the surviving fraction, is equal to P the probability that a cell will receive no hits; S = P = (a°e" a)/0! = e"a a ,u c ~a S = e The average number of hits, a, is proportional to the dose, D; a = kD Therefore, the expression for survival becomes _ -kD S = e On a semilogarithmic plot, this leads to a straight exponential, such as shown in figure 2. 11 TABLE I. SOME EQUATIONS FOR CELL SURVIVAL SURVIVAL EQUATION 2 parameters S=l-[l-exp(-D/D 0)] S=exp[-(aD+3Dz)] ABBREVIATION SH-MT L-Q DRA C-L REFERENCE Puck'S Marcus, 1956 Bender & Gooch, 1962 S i n c l a i r , 1966 K e l l e r e r & Rossi, 1972 Chadwick & Leenhouts, 1973 3 parameters S=exp(-cxD) [l+aD/e] 8^ S=exp(-D/D ) ( l - [ l - e x p ( - D / D 0 ) ] n ) RMR Tobias et a l , 1979 SH-MT Dutreix et a l , 1973 Bender & Gooch, 1962 The two parameter s i n g l e - h i t mu l t i - t a r ge t ( SH -MT) model y i e l d s a s u r v i v a l c u r v e w i th zero in i t i a l s lope, a shou lder r eg i on , and an e x p o n e n -t ia l terminal reg ion . T h i s model assumes a cel l has n t a r ge t s , a l l of wh ich must be inac t i va ted (h i t ) in o rde r for the cel l to be k i l l ed (to lose i t s ab i l i t y to p r o l i f e r a t e ) . Let -X be the inact i va t ion constant for each t a r ge t . Let D Q be the i nve r se of X. The p robab i l i t y of s u r v i v a l of any ta rget a f te r a dose D is then e x p ( - D / D Q ) , and there fo re the p robab i l i t y of inact i va t ion is 1 - e x p ( - D / D Q ) . The p robab i l i t y of i nac t i va t i ng a l l n ta rget s (and the re fo re k i l l i n g the cel l ) is (1 -exp( -D/Dg) ) n . The s u r v i v i n g f r a c -t ion is then g i ven by 12 S = 1 - [ 1 - e x p ( - D / D 0 ) ] n . Surv iva l cu rves which fit this model can be descr ibed by the two parameters n and D^. The final slope of the logarithm of surv iva l vs dose curve is equal to X, or D Q 1 . D Q is the dose that would, on the average, g ive one lethal event per target. It is the dose required to reduce cell surv iva l by a factor of 1/e, or 37%, in the exponential region. The parameter n, as descr ibed above, was original ly designated the target number. However, when it was determined that n was a function of the physiological state of the ce l l , of the stage of the cell cyc le , and of irradiation condit ions, the term 'extrapolation number' was adopted (A lper et al 1960). T h i s number can be found by extrapolating the terminal exponential region back onto the S axis. The shoulder reg ion, as interpreted by this model, represents that dose region where the cells sustain some injury which, by itself is not lethal (sublethal i n j u r y ) , but which renders the su rv i v ing cells more sensit ive to subsequent i rradiat ion. The popular i ty of this model is due, in par t , to the fact that Puck & Marcus (1956) used it to descr ibe the f i rst mammalian cell surv iva l cu r ve obtained with tissue cu l ture techniques. T h e y referred to n as the 'hit number ' , not a target number, and obtained a value of n=2 for the human cells they were us ing . The fact that this model contains only 2 parame-te r s , allowing comparisons to be made with relative ease, is also respons i -ble for its widespread use. It has become evident that this model does not adequately fit the data, especially at v e r y low doses, where it ove re s t i -mates the surv iva l (Chapman 1980). 13 Another two parameter model is the l inear-quadrat ic (hereafter referred to as L-Q) model. Th i s model y ie lds a non-zero initial slope, has a shoulder reg ion, and is continuously bending at high doses. It also assumes that the shoulder region is due to the necessity for more than one hit per target for inactivation to occur . T h i s equation has been proposed on empirical g rounds by Sinclair (1966). He found that the linear quadrat ic equation gave the best fit to his data. Kel lerer and Rossi (1972) der ived the same equation with their theory of dual radiation act ion. T h e y based their work on extensive analysis of the Relative Biological Effect iveness (RBE) of neutrons compared to X - r a y s . RBE is the ratio of doses of two d i f ferent radiations required to produce the same biological effect. R B E (of test radiation) Dose of ' s tandard ' radiation to cause a g iven effect Dose of 'test' radiation to g ive the same effect Kel lerer and Rossi studied a var iety of d i f ferent endpoints and examined the decrease in RBE with increasing neutron d o s e ( D n ) . They concluded that RBE - D n T h i s leads to the relationship between the doses of X - r a y s ( D ) and neutrons. D 2 = k D (since RBE = D /D ) x n x n To complete the formulation of their theory Kel lerer & Rossi examined the data with regards to the va ry ing linear energy transfer associated with the d i f ferent radiations. 14 Linear Energy Transfer, commonly called LET, represents the amount of energy released per unit path length of an ionizing particle. The units commonly used are keV/um. LET values are used to describe the ioniza-tion density which a particle produces along its path. LET is an important quantity because the amount of biological damage can sometimes be cor-related to the LET of the particles in the radiation beam. For instance, it is well established that high LET radiations reduce the radiosensitizing effect of oxygen (e.g. Barendsen et al 1965, Skarsgard et al 1980). LET, unfortunately, is not a unique designation for a beam of ra-diation. A radiation beam has a spread of LET values, but for simplicity a mean value is usually assigned. LETro is defined as "the total energy (on the average) lost by an ionizing particle per unit path length...it in-cludes losses by all collisions and has also been called the stopping power" (from Mortimer et al 1965). An example of the range of LET values can be seen by examining the neon-ion beam (which is designated as high LET) used at the Bevalac in Berkeley. It is usually labeled as having a mean LET^ = 234 keV/um. It has been estimated (Ngo et al v 20 1980b) that 75% of the dose is due to high-LET 1 QNe particles with a mean LET^ = 284 keV/um, approximately 22% of the dose is due to high- and medium-LET heavy nuclear fragments with a mean LET^ = 58 keV/um, and about 3% of the dose is due to relatively low-LET fast protons, helium ions, and fast neutrons. Although not unique, the value of LET assigned to a particle beam gives an indication of the ionization density which it produces. Table II contains mean LET values commonly assigned to various types of ra-diation. 15 TABLE I I . LET values f o r a few heavy ion beams* Radiation Energy (MeV/n) Residual Range (cm i n water) LET OO (keV/um) Carbon Carbon Neon Neon Argon Argon 470 70 670 60 570 8.3 31.9 1.3 32.0 00.6 13.4 0.015 1500 10 35 24 105 85 *from Ngo, et a l 1980a, page 98 Kel lerer and Rossi a rgued that, although inactivation by high L E T particles follows s ing le -h i t , s ing le-target theory (only one 'hit' is requ i red to inactivate a ce l l ) , it would take the interaction of two ' sub- les ions ' in order to inactivate a cell by particles of low L E T radiat ion. T h u s the damage from high L E T particles is d i rect ly proportional to dose, and the damage from low L E T particles is proportional to the square of the dose. The rate of cell kill by the single hit,mechanism is g iven by a, which equals the initial slope. The rate of cell kill by the double-h i t mechanism is g iven by 8. Another way of looking at this theory is to regard the linear dose term as descr ib ing intratrack events , and the squared dose term as descr ib ing intertrack events . Chadwick and Leenhouts (1973), in their 'Molecular Theory of Cell S u r v i v a l ' , a r r i ved at the same equation as the l inear-quadrat ic model based on quite d i f ferent assumptions. T h e y assumed that cell death is attr ibutable solely to the induction of unrepa i red doub le - s t rand breaks in the nuclear DNA. These breaks can occur either as a result of a single 16 event (and therefore would be direct ly proportional to dose) or from separate events result ing in two s ing le-s trand breaks at sites near each other on complementary strands of the DNA (proportional to the dose sguared) . The alpha term, which is the initial s lope, therefore represents the damage done by the single event double s trand breaks . The beta term represents the damage done by the interaction of two single strand breaks in the DNA. Mammalian surv iva l curves with no shou lders , such as those caused by high L E T radiations (Skarsgard et al 1967), can also be interpreted from the point of view of the Chadwick-Leenhouts model. The closely spaced ionizations of the high LET radiation are more likely to cause a double s t rand break in the DNA. A s ing le-event double s trand break is less l ikely to be repaired (because the complementary chain of DNA is also damaged), therefore , if the shoulder represents the sublethal damage that is repairable, with high L E T radiation no shoulder would be present. An example of a three parameter model is the Repair -Misrepair (RMR) model, developed by Dr . Tobias at Berke ley . The vers ion shown in Table I has a non-zero initial slope. His model uses the following parameters; 1) a equals the y ie ld of uncommitted (to repair or misrepair) lesions per rad (Note: this is not the same as the a in the L-Q model) 2) £ equals the repair ratio of linear repair to quadrat ic repair 3) <\> is the probabi l i ty that linear repair is eurepair (successful repa i r ) . The equa -tion for surv iva l shown in Table I is based on the following assumptions; 1) before irradiation there are no lesions 2) all quadrat ic repair results in lethal misrepair (the cell loses its abi l ity to pro l i ferate) . The most widely used 3 parameter model is the three parameter s ing le-hit mult i -target (hereafter re fer red to as SH-MT) model. Bender 17 and Cooch (1962) added a single hit component to the two parameter SH-MT model. This form of the equation has a non-zero initial slope. Since it has 3 parameters, instead of only 2, it gives a better fit. It has gained wide acceptance, particularly among those experimenters using heavily ionizing radiation (Barendsen et al 1963, Todd 1966, Wideroe 1966). ' It should be pointed out that at low doses the first approximation of any model can be transformed into a polynomial of the form InS = a + bD + cD 2 + dD 3 + By definition, a = 0 (data is normalized to S = 1 at zero dose). The fewer parameters required in a particular model the greater the possibility that the resolution of the experimental data will be sufficient to obtain accu-rate values for the parameters. The above equation, with a = 0 and only the first two parameters, is just the L-Q equation. The data obtained with the low dose technique will therefore be primarily fitted to the L-Q equation. It must be emphasized that a good fit to available data does not verify the validity of the assumptions upon which the model is based; it only says that the algebraic expression can be used to describe the data. 1.2.3 Importance of low doses The determination of the effect of low doses of radiation on biological systems is important from the point of view of basic mechanisms of radio-biology. In turn, better understanding of radiobiological mechanisms will lead to improvements in radiation protection, radiotherapy, and in testing of models to describe radiation effects. There are approximately 86,000 new cancer patients every year in Canada. That's about 4 in every 1000 people.' Of these more than half will receive radiation (perhaps in conjunction with other treatments) for 18 treatment of their disease. T h u s , radiotherapy is a major treatment affecting a large segment of the population. Radiobiologists, attempting to descr ibe a dose-response c u r v e , usually give a series of radiation doses to a population of cel l s , as 'acute' or ' s ingle-shot ' exposures . Conver se ly , a radiotherapist almost always fractionates the total dose requ i red , g iv ing a small dose dai ly, or a few times a week, for several weeks. It has been shown both in v ivo and m vi tro that cells i rradiated and then g iven time to repair before any subse -guent i r rad iat ion, are able to repair a good deal of the infl icted damage (E lk ind & Sutton 1959, E lk ind & Sutton 1960, McCul loch & T i l l 1962). T h u s , to a f i rst approximation, cells exposed to a fractionated series of doses are affected as if each dose fraction were inf l ict ing the damage character is t ic of the initial region of the cell surv iva l c u r v e . For this reason, the relevance to radiotherapy of predict ions obtained from e x p e r i -ments where large 'acute' doses were g iven is quest ionable. T h e most important aspect of cell ki l l ing for the pract ice of radiotherapy lies in the response of clonogenic cells to the f irst two or three g rays of radiation (A lper 1979). Radiation protection is concerned with the exposure of workers , and the general pub l i c , to ve ry low doses of radiat ion. The r i sks from low doses are those of causing gene mutations or of inducing cancer. In order for either of these to happen the affected cell must retain its prol i ferat ive capacity. If the radiation kills the cell then a cancer cannot develop, nor can a gene mutation be expressed. T h e r e f o r e , information concerning the relationship between radiation dose and loss of v iabi l i ty (cell surv iva l data) is important. 19 D a t a i n t h e low d o s e r e g i o n w o u l d a l s o b e v a l u a b l e i n o r d e r t o t e s t m o d e l s d e s c r i b i n g t h e b i o l o g i c a l e f f e c t s o f r a d i a t i o n . M o s t d o s e - e f f e c t m o d e l s m a k e p r e d i c t i o n s a t low d o s e s b u t , s i n c e l i t t l e c e l l s u r v i v a l d a t a i s a v a i l a b l e i n t h i s d o s e r a n g e , t h e v a l i d i t y o f t h e s e p r e d i c t i o n s h a s n o t b e e n t e s t e d . F o r e x a m p l e , d a t a d e t e r m i n i n g w h e t h e r t h e r e i s a t h r e s h o l d f o r r a d i a t i o n d a m a g e w o u l d b e i m m e n s e l y u s e f u l . T h e t w o p a r a m e t e r s i n g l e - h i t m u l t i - t a r g e t m o d e l p r e d i c t s t h a t t h e r e i s z e r o s l o p e a t z e r o d o s e , i n o t h e r w o r d s , t h a t t h e r e i s n o d a m a g e c a u s e d b y t h e f i r s t b i t o f d o s e . T h e l i n e a r - q u a d r a t i c m o d e l p r e d i c t s a s l o p e e q u a l t o a l p h a , t h e ' s i n g l e - h i t ' c o e f f i c i e n t , w h i c h i s g e n e r a l l y o f t h e o r d e r o f 10 1 g r a y 1, a t z e r o d o s e . T h e L - Q m o d e l t h e r e f o r e p r e d i c t s t h a t a n y d o s e c a u s e s s o m e d a m a g e , a l t h o u g h i t m a y b e v e r y s m a l l . C a r e f u l m e a s u r e m e n t s a t v e r y low d o s e s a r e n e c e s s a r y i n o r d e r t o d i s c r i m i n a t e b e t w e e n t h e s e t w o p r e -d i c t i o n s . A c c u r a t e c e l l s u r v i v a l d a t a w o u l d a l s o a l l o w d e t e r m i n a t i o n o f c e r t a i n m o d e l p a r a m e t e r s w h i c h a t p r e s e n t , d u e t o l a c k o f d a t a , c a n o n l y b e e s t i m a t e d . T h e L - Q m o d e l a g a i n p r o v i d e s u s w i t h a n e x a m p l e . T h e d a t a f r o m p r e v i o u s e x p e r i m e n t s i s o n l y a c c u r a t e e n o u g h t o a l l o w d e t e r m i n a t i o n t o w i t h i n ± 1 0 % f o r t w o i n d e p e n d e n t p a r a m e t e r s ( C h a p m a n 1 9 8 0 ) . A s i n g l e p a r a m e t e r , f o r e x a m p l e t h e a/6 r a t i o , c a n o f c o u r s e b e d e t e r m i n e d m o r e p r e c i s e l y . If we c o u l d d o a n i n d e p e n d e n t e x p e r i m e n t ( e . g . a t low d o s e s w i t h a d i f f e r e n t t e c h n i q u e ) t o a c c u r a t e l y d e t e r m i n e t h e v a l u e f o r a l p h a o r b e t a a n d c o m b i n e t h e r e s u l t s , t h e n b o t h a l p h a a n d b e t a c o u l d b e a c c u -r a t e l y d e t e r m i n e d . A s e c o n d e x a m p l e o f w h e r e low d o s e e x p e r i m e n t s w o u l d b e u s e f u l f o r p a r a m e t e r d e t e r m i n a t i o n i s p r o v i d e d b y T o b i a s ' s R M R m o d e l . T h e v a l u e o f e i n t h e v e r s i o n o f t h e R M R m o d e l s h o w n i n T a b l e I c a n b e o b t a i n e d b y 20 fitt ing cell surv iva l cu rve s . When c - 0, there is no repair , and a straight exponential is obtained. When E is ve ry large then linear eurepair is much more important than quadrat ic misrepair, and a surv iva l curve with a pronounced shoulder is obtained. Determination of e under d i f ferent irradiation conditions may add to our understanding of the processes involved in repair of radiation damage. The understanding of both basic radiobiology and the mechanisms of radiation damage would be great ly enhanced if we could assay for cell surv iva l at low doses. The exact mechanisms by which radiation damage is expressed are not ful ly understood (e.g . Radiation Biology in Cancer Research, 1980, edited by Meyn & Withers, pages 21-250). If we wish to determine the mechanisms of radiation damage at low doses then obviously we must do experiments at low doses. Th i s is especially true if, as s u g -gested by some authors (Planel et al 1976, Jenssen & Ramel 1978) mecha-nisms of damage are d i f ferent at low and high doses. Regardless of this however, the examination of the effect of the f irst increments of dose could lend valuable insight into the mechanisms of radiation damage. Th i s would g ive us a better understanding of basic radiobiology which might, in t u r n , lead to better treatment protocols and more rational protection procedures . It is evident that to answer questions at low doses of radiat ion, experiments must be performed at these doses. Th i s thesis is concerned with experiments on mammalian cell s u rv i va l . However, many other e n d -points, other than mammalian cell s u r v i v a l , have been used to quant i fy the effect of radiation at low doses. I will br ie f ly review these before d i scuss ing mammalian cell s u r v i v a l . 21 1 .3 ENDPOINTS USED T O A S S A Y LOW RADIAT ION DOSES Work with a common American garden plant, Tradescant ia (the spiderwort) can be quanti f ied satisfactori ly down to less than 0.003 gray of X - r ay radiation (NCRP Report 64 1980). The flower buds , normally blue or purp l i sh in color, contain large numbers of 'stamen hairs , ' each consist ing of a chain of about 25 individual cel ls . Radiation induces pink mutation events in the cel ls , which are easily scored against the normal blue color. Vicia faba, a bean root, is another plant system used to determine the low dose effect of radiation. The inhibition of root growth is the index of damage. In a s tudy by Winston, et al (1975) this effect was quanti f ied to as low as 0.16 g r ay s . Chromosome studies are also capable of quant i fy ing effects at low doses (e .g . Puck 1958, Bender 1960, Chu et al 1961, Bender 1969, Brewen & Luippold 1971, Pohl -Rul ing et al 1978, Evans et al 1979). The chromosomes of cells are microscopically examined to determine the amount of abnormalities (breaks , abnormal metaphases, micronuclei , etc.) as a function of dose. These experiments are usual ly quite large in scope. For example: to obtain statistically val id data at the 0.05-, 0.10-, 0.25-, and 0.50- g ray dose points, over 14,000 metaphases were scored in a study by Lloyd et al (1975) to determine the dose-response cu rve for human lymphocytes. The f requency of micronuclei has been successfu l ly used as an endpoint to determine the chromosome aberrat ion f requency in X - r a y treated human lymphocytes (Countryman & Heddie 1976) to as low as 0.50 g r ay . Micronucleus assays are based upon the consideration that a d irect correlation exists between chromosome breaks and the formation of 22 micronuclei (Evans et al 1959). It is believed that micronuclei arise from chromosomal fragments that are not incorporated into daughter nuclei at mitosis because these fragments lack a centromere. Recently studies have attempted to show a correlation between the f reguency of micronuclei and clonogenic surv iva l of irradiated cells of a certain ploidy (Midander and Revesz, 1980, Crote et al 1981a,b, Joshi et a l , 1982a,b,c). The work reported by Grote et al 8 Joshi et al showed a correlation between micronucleus f requency and colony forming ability to doses as low as 0.2 Gy . The method used to determine the colony forming ability was quite similar to the technique presented in this thesis with the following d i f ferences: 1) On ly cells i rradiated in the G l stage of the cell cycle are assayed, 2) a d i f ferent cell line is used (BHK 21 C13, a Syr ian hamster l ine) , 3) feeder cells are u sed , 4) the cells are d iv ided into 3 categories instead of 2. Th i s is just one example of the many attempts that are being made to correlate these var ious endpoints with the actual inactivation of cel ls. As ye t , no one has shown a definite relationship between any of the var ious endpoints and cell s u rv i va l . When s tudy ing the effect of radiation on the prol i ferat ive capacity of mammalian cel l s , results achieved with other endpoints and other cell lines must be kept in mind. Measurements of cell surv iva l in the low dose region are limited because one is faced with measuring the su rv i v ing fraction in a population of predominantly su rv i v ing cells. Even when the surv iva l experiment is performed with the utmost care , it is not possible to determine S, the fraction of su rv i v ing cel ls , with precis ion greater than approximately 10%. In a log S v s . D (dose) plot, the absolute er ror s in log S are of the same magnitude throughout the dose range. T h u s at low doses (S 23 between 1.0 and 0.5), such uncertainties do not permit determinations of radiobiological parameters (e.g . the oxygen enhancement ratio (OER ) , relative biological effectiveness (RBE) of d i f ferent radiation modalities, cell cycle dependence, effects of radiosensit izers and protectors) with suff icient precis ion to ascertain whether or not these parameters change at low dose as compared to higher doses. The uncertaint ies in such surv iva l measurements arise primari ly from uncertainties in the number of cells plated on the f irst day. E r ro r s in the counting of cells in su spen -s ion, multiple di lutions before plating and the plating itself are all c o n -tr ibut ing factors. 1 .4 T H E T E C H N I Q U E T O A S S A Y C E L L S U R V I V A L A T LOW DOSES A method is herein presented whereby surv iva l of cells is measured with s ignif icant ly less than ±10% e r r o r . It is part icu lar ly applicable at su rv i v ing fractions between 1.0 and 0.5, hence at dose levels of 0 to 3 C y . Rather than determining just the su rv i v ing fraction S, the fraction of ki l led ce l l s , K, was also measured. T h e su rv i v i ng fract ion can be calculated from the K measurement; S = 1 - K. K was determined by microscopically following the fate of s ingle cells (plated on day 0) through the incubation period of 7 days. 24 It has long been realized that if the exact number of cells plated is known, much more accurate data can be obtained. At the 6th L .H . C r a y Conference in London in 1 975 Joel Bedford and H. Gr iggs stated: "Two methods may be used for measuring responses in cell populations of exactly known numbers. One is to isolate cells individual ly with the aid of a micropipette, and count them one at a time as they are plated and the other is to inoculate a number of Petri dishes marked with g r i d s , and microscopically locate and record the coordinates of each cell within the g r i d . We have carr ied out one experiment to test the feasibil ity of the latter method, and although the process is tedious, it is poss i -b le . " If both the number of kil led cells and the number of surv iv ing cells are determined, then the exact number of cells plated is known. In the low dose reg ion, a large fractional e r ror in K translates to a small fractional e r ro r in S. We know S = 1 - K. The e r ro r in the s u r v i v -ing fraction is AS and the e r ro r in the kil led fraction is AK. The re lat ion-ship between the er ror s is AS = AK and the fractional e r ror in S is AS/S = AK/S = (AK/K) * (K/S) In the low dose reg ion, where K is less than S, most of the cells are su r v i vo r s . We are interested in minimizing AS/S, the fractional e r ro r in S. A s can be seen from the g raph in f igure 3, as long as K is less than S, AS/S is much smaller than AK/K. A simple mathematical example will demonstrate this point: At 80% surv iva l (S = 0.8, K = 0.2) if you misidentified 50% of the dead cells as live cells that would translate to only a 13% fractional e r ro r in S, the su rv i v ing fraction (see f igure 3). T h u s , in the low dose region we will have a smaller fractional e r ro r in our measurement of S than in our measurement of K. Notice this only applies in the low dose region. 25 F R A C T I O N A L E R R O R IN K ( A K / K ) F igure 3. F R A C T I O N A L ERRORS IN S AMD K. The f igure demonstrates the relationship between the fractional e r ro r s in S and K for d i f ferent surv iva l levels. 26 In actual fact we measure both the number of killed cells and the number of su rv i v ing cel ls. Th i s gives us an advantage statist ical ly. If we accurately know the number of su rv i v ing and non - surv i v ing cells we can use binomial statistics to ascertain our minimum e r r o r . If we only know the number of su rv i vo r s then we must use Poisson statistics in order to determine our minimum e r r o r s . From f igure 4 we can see that our minimum er ror is much less for binomial statistics than for Poisson statist ics, for each level of su rv i va l . There fore it is to our advantage if we can measure both the number of su rv i vo r s and the number of nonsur -v i vo r s . Th i s is poss ible, although labor intens ive, if we observe the cells microscopical ly. 27 4 0 0 8 0 0 1 2 0 0 1 6 0 0 N u m b e r of c o l o n i e s f o r S d e t e r m i n a t i o n F igure H. T H E RELAT IONSH IP BETWEEN T H E P E R C E N T A G E ERROR IN S AND T H E NUMBER O F COLON IES FOR B O T H POISSON AND BINOMIAL S T A T I S T I C S . The f igure demonstrates the decrease in both the percentage e r ro r in S and the number of colonies requ i red if both S and K are measured (binomial statistics) rather than measuring only S (Poisson stat i s t ics ) . 28 1 .5 O X Y G E N E F F E C T A T LOW DOSES We have used this technique to study the effect of oxygen on cell surv iva l at high and low doses. It has been known for many years that, when oxygen is present , l iving organisms are sensit ized to the lethal effects of radiation (Holthusen 1921, Dowdy et al 1950, Dewey 1960). Hypoxic cells (cells def ic ient in oxygen) have been shown to exist in almost every animal tumour that has been studied (Denekamp 1982). T h e r e are many who believe hypoxia to be a major limiting factor in radiation therapy (Thomlinson & Gray 1955, Evans & Naylor 1963, Urtasun et al 1976 and Henk & Smith 1973). In o rder to determine whether this is the case, the effect of oxygen at cl inical ly relevant doses must be exp lored. It wasn't unti l 1955 that a mechanism for hypoxia was proposed. Thomlinson 6 Gray (1955) correlated the pattern of necrosis within a tumour to the microenvironment around each blood vesse l . Cel ls near the blood vessel were well oxygenated. T h e amount of oxygen available to the cells decreased in proport ion to the distance from the blood vesse l . Cells located at a distance from the blood vessel that was much greater than the di f fus ion distance for oxygen (about 145 micrometers) were necrot ic. A layer of hypoxic viable cells 2 to 3 cells th ick exists between the necrotic cells and the oxygenated cel ls. It is these cells that can cause problems in radiotherapy (Franko et al 1979a,b). If hypoxic cells eventual ly die as a result of their continued n u t r i -tional depr ivat ion they do not present a problem to rad iotherapy. How-e v e r , this is usual ly not the case. A f t e r a dose of radiation suff ic ient to kill the well oxygenated cel l s , the blood and nutr ient supply may be increased to the hypox ic cel l s , allowing them to re -enter mitosis and regrow the tumour. Th i s is thought to be one of the mechanisms 29 responsible for the fai lure of local tumour contro l . Theoret ica l ly , this ' reoxygenat ion 1 can be used advantageously. Once reoxygenation occurs the cells are no longer radioresistant. A second dose of radiation, c o r r e -sponding to the time of maximum reoxygenat ion, can kill the previous ly hypoxic cells that are now oxygenated. There are, unfortunate ly, some problems in the practical application of this theory. The timing of reox-ygenation is likely to va ry from tumour to tumour and is crit ical to the success of fractionated radiotherapy (Fowler et al 1 976, Denekamp et al 1980, Suit & Wette 1966). The Oxygen Enhancement Ratio (OER) is def ined as the ratio of doses in the absence and the presence of oxygen , requ i red to produce equal biological ef fects. Cell surv iva l data for irradiat ion with X - r a y s at high doses (analogous to the total tumour dose, rather than the dose per fraction) show an OER that is constant with dose and equal to about 3. If the OER is independent of dose this implies that the OER for a f rac t ion-ated regime would also be 3. Th i s poses considerable problems for the radiotherapist. A n OER of 3 implies that the hypoxic cells are 3 times more resistant to radiation than the oxic cells. Normal healthy cel l s , which are ox ic , are therefore much more sensit ive to the radiation than the hypoxic tumour cel ls. It is therefore d i f f icult to get enough dose to the tumour to kill the hypoxic cells while still spar ing the oxic healthy t issues. Fractionation of the total tumour dose would be of value if hypoxic cells are slower to recover from radiation damage than oxic cel ls , as suggested by Koch and others (e .g . L i t tbrand & Revesz 1969, Hall & Cavanagh 1969, Hall 1972 , Koch et al 1977, Koch 1979). A fractionation scheme would then allow the healthy oxic cells a better chance to recover 30 from the radiation damage than the hypoxic tumour cel ls , decreasing the advantage that the hypox ic cells have. If the OER is lower at low doses in v i vo , then changing the f r a c -tionation scheme could minimize the problem of hypox ia . By del iver ing the total tumour dose in a larger number of smaller fractions the hypoxic cells have less advantage over the oxic cel ls, hn v i tro experiments concerning the effect of oxygen at cl inical ly relevant doses can be used to determine if the OER changes with dose. Actual data on the effect of oxygen at low radiation doses can also help eliminate unacceptable models that deal with dose effect relations. A model which predicts the same mechanism for damage in aerobic and hypoxic conditions is consistent with an OER that is independent of d o s e — i n other words that oxygen is 'dose modify ing ' . A model which postulates that under hypoxic conditions only high L E T events infl ict damage (e .g . Wideroe 1970) or that the effect of low-LET events is r e -duced relative to that of high L E T events at low doses would not accom-modate a dose modifying effect of oxygen (Winston et al 1975). A whole conference was devoted to the question of cell surv iva l after low doses of radiation (Sixth L .H . C r a y Conference, London, 1974, A l p e r , T . , ed.) where among other subjects the guestion of oxygen e n -hancement ratio at low doses was examined. Experiments us ing Ch lamy-domonas (Bryant and Lansley 1975) and Vicia faba roots (Winston et al 1975), showed no change in the oxygen effect at low doses. Experiments performed using the standard technigue of measuring cell surv iva l ob -tained confl ict ing results. Aga in , some researchers presented data and arguments suggest ing a diminished OER at low doses of ionizing radiation in mammalian cells i rradiated in v i t ro (Revesz et al 1975, Chapman et al 31 1 975b, McNally 1975, Pettersen et al 1975) while others presented data and arguments claiming that the OER is constant throughout the dose range (Phil l ips et al 1975 , Koch 1975). At the close of the conference the consensus seemed to be that in plants oxygen was dose modifying (constant with dose) but with regards to mammalian cells the effect of oxygen at low doses was not c lear—more work needed to be done. The debate has still not ended today, although more data, suppor t -ing both v iewpoints, are available. We have attempted to add to these data, with experiments using our new technique. 32 2. M A T E R I A L S AND METHODS 2.1 C U L T U R E T E C H N I Q U E Chinese hamster ovary (CHO) cells were grown in suspens ion. The cells were spun in special culture vessels (Bellco Hanging Bar Sp inner F lasks , Bellco Class Inc.) and maintained at 37°C (5% C 0 2 , 95% air atmosphere) in alpha medium (Flow) supplemented by 10% fetal calf serum ( F C S , C I B C O ) . Cells were di luted daily to maintain log-phase growth. T h e y were kept at a concentration between 0.7 X 10 5 and 3.8 X 10 5 ce l l s /ml. T h e cel l -doubl ing time was 12-13 hours . Twelve hours before each experiment with asynchronous cells the ceil population in the sp inner was di luted to 1 x 10 5 cel l s /ml. In twelve hours the cell population had doubled and reached a concentrat ion of 2 x 10 5 cel ls/ml. At this point cells for the experiment were removed from the sp inner . 2.2 S Y N C H R O N O U S T E C H N I Q U E For the experiments with synchronous ce l l s , cells were synchron ized us ing a s l ight variat ion of the mitotic shake-of f technique descr ibed by Terasima & Tolmach (1963). The pr incip le of mitotic shake-of f is th is : when cells growing attached to the bottom of a flask are about to d iv ide (go through mitosis) they ' r ound -up ' and are not f irmly attached. T h e r e -fore, if the flask is v igorous ly shaken the cells which are about to d iv ide are shaken off. To obtain our mitotic cells the following procedure was employed: Cel ls growing exponential ly in a sp inner flask were kept at a concen-tration between 0.8 and 3.0 X 10 5 cells/ml for 2 days pr ior to the day of p lat ing. Twelve hours before p lat ing , the cells were di luted to 1 X 10 5 33 cel ls/ml. At the time of plating the cells were at a concentration of approximately 2 X 10 5 cel l s /ml. T h e y were plated into large tissue cul ture f lasks (growth area of 75cm 2 ) at a concentration of 10 6 cel l s / f lask, with Ham's F-10 medium supplemented with 10% fetal calf serum (approximately 5 ml of cells and 10 ml of medium). Using this medium (rather than the standard alpha medium supplemented; with 10% fetal calf serum) a larger percentage of mitotic cells could be shaken off , as seen in f igure 5. The plated cells were placed in the incubator and fed after 24 hours with 15 ml fresh F-10 medium. Shake-of f occur red approximately 24 hours later, the time of the f i rst shake-of f being 46 to 50 hours after the plating of the f lasks , when the cells would be in exponential growth at a concentration of approximately 10 7 cells per f lask. T h e shake-of f procedure was conducted in a walk- in 37°C warm room. T h e f lasks were p laced, 6 at a time, on a horizontal shaker table and shaken for 5 minutes (128 v ib ra t ions . /m in , 4.5cm/stroke). Twelve f lasks were used for each experiment. T h e over ly ing medium was collected in chil led ( 4 ° C ) centr i fuge vessels and replaced with 10 ml of p H - and temperature- equi l ibrated F-10 medium. T h e f lasks were then placed in a 37°C water bath for 10 minutes pr ior to the next shake-off ; steri le technique was obse rved . Th i s was repeated 9 to 10 times. The cell number as a function of shake-of f is shown in f igure 5. T h e cells for the experiment were taken from the 9th shake-of f (and 10th if neces sa ry ) . These were spun down, at 4 ° C , in chi l led glass test tubes , 20 ml per tube, for 8 minutes at 700 rpm (approximately 110g, a S O R V A L L RC-3 general purpose centr i fuge was u sed ) . T h e medium above the pellet was aspirated unti l 1 ml remained, the cells were combined and resuspended in alpha medium 34 (Flow), supplemented with 10% fetal calf serum (FCS, CIBCO) but lacking sodium bicarbonate, at a concentration of 105 cells/ml. Removal of sodium bicarbonate allows the pH to be maintained at approximately 7.3 without the presence of 5% CG"2. This suspension was placed in a stoppered flask with a spin bar, in a 37°C water bath, in the 37°C warm room. To determine the degree of synchrony two techniques were routinely used, staining of the mitotic cells to determine the percentage of cells in mitosis and measuring the uptake of labelled thymidine throughout the cell cycle. Details of the procedures are given in Table III and IV, respec-tively. 35 I I I I I I I I I I 1 2 3 4 5 6 7 8 9 1 0 S H A K E - O F F Figure 5. CELL NUMBER AS A FUNCTION OF SHAKE-OFF. The figure demonstrates the increase in cell yield when using Ham's F-10 medium rather than the standard medium. Cells for the experiments were taken from the 9th shake-off (and 10th if necessary). Cells from previous shake-offs were discarded. The error bars represent the 10% standard error associated with the electronic Coulter Counter measurements of the cell concentration. 36 TABLE III. PROTOCOL FOR DETERMINING MITOTIC INDEX 1. Centrifuge 5ml cell suspension 5 minutes at 1000 rpm (lO5^ cells is plenty). 2. Resuspend with continuous agitation by drop-wise addition of cold (4°C) 0.1 M sucrose to a total volume of 5ml. 3. Hold in ice-bath for 3 minutes. 4. Recentrifuge 5 minutes at 1000 rpm. 5. Resuspend with continuous agitation by dropwise addition of cold (4°C) Carnoy to a total volume of 5ml. 6. Hold in ice-bath a minimum of 10 minutes. 7. Recentrifuge 6 minutes at 600 rpm. 8. Resuspend in 0.1ml cold Carnoy. 9. Apply to slide, air dry. 10. Stain for 12 minutes with Giemsa-Acetic-orcein, rinse with distilled water. 11. Count % mitotic cells microscopically. Notes. Stain: Mix 2ml Giemsa stock, 4ml Acetic-orcein, 4ml doubled distilled water. Filter. Make up fresh for each use. Acetic-orcein: dissolve 0.5gm orcein in 100ml boiling 45% acetic acid. Cool. Filter 3 times. Refilter before use if crystals form. At step 6, cells may be stored on ice and dealt with later if time does not permit immediate completion of the protocol. Mitotic cells can be determined by the appearance of dark masses in the nucleus indicating the condensation of chromosomes. Non mitotic nuclei are uniform in uptake of the stain. 37 TABLE IV. PROTOCOL FOR TCA PRECIPITATION OF DNA 1. Place 0.2 ml of cells (2X10V cells) in 0.5 ml of 10 microcurie/ml 3H Thymidine at 37°C for 10 minutes. 2. After 10 minutes put on ice, add 9 ml cold (4°C) PBS. 3. Spin 6 minutes at 600 rpm. 4. Pour off carefully, set on ice. 5. Add approximately 3 ml 5% TCA (U°C) to test tube, being sure to stir cells up at bottom of test tube. 6. Wet the filter paper with 5% TCA. 7. Pour cells and TCA from step # 5 through filter. 8. Wash test tube with 2-3 ml 5% TCA, pour through filter. 9. Pour approximately 3 ml 5% TCA through filter. 10. Pour approximately 3 ml 5% TCA through filter. 11. Pour approximately 3 ml 75% ethanol (U°C) through filter. 12. Place filters in scintillation vials, let dry overnight. 13. Add 10 ml scintillation cocktail, count. Notes. To filter, a vacuum system is employed. The filters used are Sartorius Membrane Filters, pore size 0.2 microns, diameter 25 mm. All TCA used is at U°C. 38 2.3 OER EXPER IMENTS As an example of the use of this technique the effect of oxygen at high and low doses was s tud ied. The f irst set of experiments were performed with an asynchronous population of CHO cells grown in a sp inner flask (Moore et al 1976). Cells were cooled to 4°C and were made hypoxic by flowing pur i f ied nitrogen ( N 2 , less than 5 ppm of present) over s t i r red cell suspensions for at least 45 minutes pr ior to the start of irradiation as well as dur ing i rradiat ion. Cells were kept at 4°C in special glass irradiat ion vessels (Parker et al 1969) in growth medium in a total volume of less than 20 ml; the cell concentration was kept at a p -proximately 1 .25 X 1 0 5 cel ls/ml. Aerobic cells were treated identically except that the flow of N^ was not employed. Al iquots of 1 ml were taken from the irradiat ion vessel after the requ i red doses and put in plastic dilution tubes fi l led with 9 ml of chil led ( 4 ° C ) alpha medium ( lack-ing sodium bicarbonate and fetal calf serum). Cells were plated from the dilution tubes into 5 cm plastic petri dishes (Fa lcon, with or without an etched g r id ) and 96-well plastic microtest plates (Nunc, Po lys tyrene) . For the high dose assays the petri dishes were prepared the day before the experiment. Sterile 60 x 15 mm petri dishes (Falcon) were fi l led with 5 ml alpha medium (Flow) supplemented with 10% fetal calf serum ( F C S , C l B C O ) and 10 5 feeder cel ls. Feeder cells are irradiated cells (60 C y ) that are still capable of metabolizing but can no longer d iv ide. T h e y condition the medium prov id ing optimum growth conditions for the cells which are being assayed for colony forming abi l i ty. These petri dishes were placed in the incubator and had reached pH and temperature egui l ibr ium by the following day. 39 The cell concentration in the dilution tubes was determined using a cell counter (e .g . Coulter Counter ) . The cells were plated from the dilution tubes so that approximately 150 colonies per petri dish emerged after 7 days , i rrespect ive of the atmosphere or dose of i rradiat ion. A f ter p lat ing, the dishes were returned to the incubator for 7 days. The medium was then careful ly poured out of the petri dishes and 4 ml of Methylene Blue stain was added to each d i sh . A f te r approximately 6 minutes the stain was poured off and the dishes r insed gently in cold running water. The stained colonies were counted after the dishes had d r i ed . For the low dose assay two plating methods were used. At the ve ry beginning of the development of this technique 60 x 15 mm petri dishes with an etched g r i d (Falcon) were used. The dishes were fil led with 5 ml of alpha medium (Flow) supplemented with 10% fetal calf serum ( F C S , C I B C O ) . No feeders were used. Approximately 400 cells were plated into each dish from the dilution tubes. The plated dishes were then returned to the incubator. Considerable problems with 'satellite' colonies arose with the use of the petri dishes with etched g r i d . 'Satellite' colonies were formed due to the action of the media removing cells from the colonies and start ing new colonies nearby. The use of microtest plates eliminated these problems and did not alter the radiation response (for more details see section 3.1 .2). The method used with the microtest plates was to plate the cells at an average density of 1-3 cells per well. The day before the experiment warm (37°C ) alpha medium supplemented with 10% fetal calf serum was f i l tered us ing a .20 micron pore steri l ization fi lter unit ( SYBRON/Na l ge ) . Sterile glass test tubes (25 ml total volume) were filled with 21 ml of this \ 40 ref i l tered medium. These were capped and stored in the refr igerator (4°C ) until the following day. A f ter the irradiation approximately 200 to 250 cells were taken from the dilution tubes and put in the glass test tubes. T h e irradiation vessels , dilution tubes, and glass test tubes were all kept at 4°C at all times. A f t e r addition of the cells the test tube solution was vortexed for 5 seconds. Th i s cell solution was poured into a 25 ml Erlenmeyer f lask. A 10 ml multiple d ispenser syr inge was used to take the cell solution from the flask and put 200yl into each well. The syr inge had to be refi l led once in order to fill all 96 wells. A n 18 gauge needle, 1? inches long (Yale, Luer -Lok Hub, 18G, 1£) was used. The f i rst attempts at d ispens ing equal multiple aliquots involved the use of a glass 10 ml Hamilton sy r inge . The syr inge was thoroughly r insed between the plating of each dose point with f i l tered sterile PBS (Phosphate Buf fered Sal ine). Several problems arose with this system. Due probably to the frequent autoclaving of the syr inge it did not remain air t ight. A i r bubbles would form in the sy r inge , causing nonuniform aliquots to be d ispensed into the wells. T h e largest problem by far was that contamination occasionally appeared in some of the wells due to the r ins ing with the PBS. To correct this the glass Hamilton syr inge was replaced by a multiple d ispens ing (200yl) syr inge holder which used disposable 10 ml syr inges (Ster iStep, NCS Diagnost ics, Inc. ) . A steri le needle (Yale, Luer -Lok Hub , 18G, H ) and sterile plastic syr inge (Becton, Dickinson & C o . , 10cc Sy r inge , Lue r -Lok T ip ) were used for each dose point. Th i s reduced contamination problems. To achieve uniform d is tr ibut ion of the cells the following procedure was obse rved . Before pour ing the cell solution in the test tubes into the Er lenmeyer f lask, each test tube was vortexed for 5 seconds. Immediately 41 after pour ing the contents into the 25 ml flask the f irst 10 ml was taken up into the sy r inge . The full syr inge was turned upside down and any air expel led. The solution was then d i spensed, 200 ul into each well. Before the syr inge was refi l led with the second 10 ml about 5 ml of the solution was taken up and then released back into the solution at a rate suff ic ient to st ir the solution but not so fast as to cause foam to form. Immediately after this the remaining 10 ml was taken up into the syr inge and d i spensed, 200 yl into each of the 48 remaining wells. Two microtest plates were always plated for each dose point. A f te r p lat ing, the cells were incubated at 37°C. A minimum of 20 minutes incubation was necessary to allow the cells to adhere to the bottom surface before location of the cells could be determined. Cells were located either on the day of plating or 2 or 3 days later. T h e day of the initial location d id not affect the radiation response (for more details see section 3.1.5). By noting the position of each ce l l , we could follow the fate of each cell for the incubation period (7 day s ) . The use of microtest plates proved to be advantageous in several respects and d id not alter the radiation response. The second set of experiments was performed with synchronous CHO cel ls. CHO cells have a doubl ing time of 12 to 13 hours . The cell cycle is commonly labeled as follows (the times g iven are for CHO cel ls , taken from Whitmore & Cu lyas 1980): M, or mitosis, represents the stage of the cell cyc le when the cell is phys ica l ly d iv id ing in two. Th i s stage lasts approximately i hour . G1 (C for gap) is the postmitotic phase preceding DNA synthes is which is 2 to 2\ hours long. T h e period dur ing which the cell is act ively engaged in synthes iz ing new DNA in preparat ion for the next mitosis is called the S phase. Th i s is the longest stage of 41 after pouring the contents into the 25 ml flask the f irst 10 ml was taken up into the syr inge. The full syr inge was turned upside down and any air expel led. The solution was then d i spensed, 200 pi into each well. Before the syr inge was refi l led with the second 10 ml about 5 ml of the solution was taken up and then released back into the solution at a rate suff icient to st ir the solution but not so fast as to cause foam to form. Immediately after this the remaining 10 ml was taken up into the syr inge and d i spensed, 200 yl into each of the 48 remaining wells. Two microtest plates were always plated for each dose point. A f te r plat ing,the cells were incubated at 37°C . A minimum of 20 minutes incubation was necessary to allow the cells to adhere to the bottom surface before location of the cells could be determined. Cells were located either on the day of plating or 2 or 3 days later. The day of the initial location did not affect the radiation response (for more details see section 3.1.5). By noting the position of each ce l l , we could follow the fate of each cell for the incubation period (7 day s ) . The use of microtest plates proved to be advantageous in several respects and did not alter the radiation response. The second set of experiments were performed with synchronous CHO cells. CHO cells have a doubl ing time of 12 to 13 hours . The cell cycle is commonly labeled as follows (the times g iven are for CHO cel l s , taken from Whitmore & Cu lyas 1980): M, or mitosis, represents the stage of the cell cycle when the cell is phys ica l ly d iv id ing in two. Th i s stage lasts approximately \ hour. G1 (G for gap) is the postmitotic phase preceding DNA synthesis which is 2 to 2| hours long. The period dur ing which the cell is actively engaged in synthes iz ing new DNA in preparat ion for the next mitosis is called the S phase. Th i s is the longest stage of 42 the CHO cell cyc le , approximately 8 hours long. C2 is the premitotic phase, following DNA synthes i s , lasting approximately 2 hours. Experiments to determine the effect of oxygen at low doses were conducted 3 hours and 8 hours after mitosis. These times correspond to CHO cells near the C1/S interface, a radiosensit ive stage of the cell cyc le , and mid to late S, a radioresistant stage of the cell cyc le , respec -t ive ly. The radiosensit iv ity as a function of cell cycle is shown in f igure 6. The cells were taken from the flask in the warm room at the des -ignated times and put into irradiat ion vessels . The remainder of the experiment was performed following the same procedure descr ibed for the asynchronous cells with the following except ion. Cells in the irradiation vessels were at a concentration of 5000 cel ls/ml. Th i s was due to the d i f f icu l ty in obtaining large quantit ies of synchron ized cel ls. 1.0 O o < DC LL-CS ± 0.1 > > DC Z> CO S y n c h r o n i z e d C H O ce l l s I s o d o s e cu rve s (in air) --...2 Gy , C H O . . . . ••<•)" '"•••••.7 Gy. CHO 2 4 6 8 10 12 HOURS AFTER S H A K E - O F F Figure 6. RAD IOSENS IT IV ITY AS A F U N C T I O N OF C E L L C Y C L E . The bottom cu rve represents data from the Berke ley laboratory, showing a typical cell cycle dependent radiosensit iv i ty (Blakely et al 1980). V-79 cells (which are v e r y similar to CHO cells) were irradiated with 8 Gy of X - r a y s . The two dashed cu rves are drawn through data obtained at two points in the cell cyc le with our CHO cel ls , as descr ibed in the captions of f igures 18 and 21. The important thing to note is that at the two points chosen (3 and 8 hours after shake -o f f ) , the cells have di f ferent radiosensit iv it ies. 44 2.4. I RRAD IAT ION I r rad ia t ion was performed us ing an X - r a y source ( P i c k e r , 280 k V p , H V L 1.7 mm C u ) . Samples were in specia l g lass i r r ad ia t i on vesse l s ( f i g u re 7 ) . The 'a rms ' of the vesse l were taped to the s ides of the p lex i g la s s conta iner i t sat in to p reven t the vesse l from t i l t i n g . Co ld water was added to a g r adua ted c y l i n d e r w i th 600 ml of c r u shed ice to b r i n g the total volume to 820 ml. Th i s was poured around the i r r ad ia t i on vesse l to keep it at 4°C . A s t i r motor was su spended above the ve s s e l . The X - r a y source was below the ve s se l . To obta in a dose rate of 1.6 gray/min the p lex i g la s s conta iner sat on top of the 20 x 20 cm co l l imator. Th i s dose rate was used for the h igh dose contro l po ints ( g reate r than 2 Gy in o x y g e n , or 3 Gy in n i t r o g e n ) . To ach ieve a lower dose rate (0.3 gray/min) the p lex i g l a s s conta iner was suppo r ted a round the edges by a wooden p lat form 49 cm above the 20 x 20 cm co l l imator. F r i c k e f e r r ou s su l fate dos imetry was used to determine the dose rates . The F r i c k e dos imeter u t i l i ze s the ox idat ion of f e r r ou s i ron to f e r r i c i ron in a c i d i c , oxygenated aqueous so lut ion to measure the dose ab so rbed . A f t e r i r r ad i a t i on of the so lut ion ( in the exact s e t - up that is used for the cel l exper iments ) the f e r r i c ion concent ra t ion was determined spect rophotometr i ca l l y by measur ing the absorbance at 304 nm. The absorbed dose was then ca lcu la ted u s ing the equat ion : Dose ( rads ) = (0.965 X 10 9 X A ^ ° p / [ e 3 Q J | * p * G ( F e + 3 ) ] where A 3 0 1 ! is the net absorbance of the i r r ad i a ted sample at 304 nm, net K found from the spectrophotometer g r a p h , e ^ n . is the molar ex t i n c t i on 45 coefficient of Fe 3 at 304 nm. At 20°C (which is the temperature at which these measurements were made) the value of £^04 ' s 2121 (see Fr icke & Hart, 1966, Table III, page 190). p is the density of the ferrous sulfate solution (1.024 ± 0.001 between 15 and 2 5 ° C ) . The C va lue, defined as the mean number of ferr ic ions produced by an imparted energy of 100 eV, is a function of radiation energy . The value of 14.7 ferr ic ions/100 eV was determined from ICRU Reports 14 and 17 for our part icular con -ditions ( X - r ay source: P icker , 280 k V p , H V L = 1.7 mm C u ) . 46 GAS IN when x = 0.0 cm dose rate = 1.6 Gy/min when x = 49.0 cm dose rate =0.3 Gy/min I ^ 20 X 20 cm collimator X-ray head Figure 7. IRRADIATION S E T - U P . T h e cell suspension was put into a special glass irradiat ion vessel (with a magnetic st ir bar) which was placed in a plexiglass container containing ice water, a st ir motor was suspended above the vesse l . To obtain a dose rate of 1.6 Gy/min the plexiglass container was placed d i rect ly on top of the 20 X 20 cm collimator. To achieve a lower dose rate of 0.3 Gy/min the plexiglass container was supported around the edges by a wooden p lat -form 49 cm above the 20 X 20 cm collimator. 47 3. RESULTS 3.1 DEVELOPMENT OF TECHNIQUE TO ASSAY CELL SURVIVAL AT LOW  DOSES < 3.1.1 Microscopic examination and elimination of errors The first priority in developing a technique that would be accurate at low doses was to obtain knowledge of the exact number of cells plated. The easiest way to obtain this knowledge was determination of both K, the killed fraction of cells and S, the surviving fraction of cells. The first advantage, statistically, of measuring K rather than just S is that, in the low dose region, a large fractional error in K translates to a small fractional error in S (see figure 3, page 25). It is the fractional error in S that we are trying to minimize. The second advantage is that, since we actually measure both K and S, our minimum error is determined by binomial statistics rather than Poisson, and is therefore smaller, as shown in figure 4 (page 27). In order to determine K the cells had to be ex-amined microscopically. Although labour intensive, this proved very advantageous. Due to this microscopic examination many of the sources of error that are present in the 'classical' technique are eliminated. A cell coun-ter, (e.g Coulter counter) is not used to determine the number of cells plated. Therefore any errors in dilution or plating as well as errors in the Coulter counter measurement do not contribute to errors in the num-ber of cells analysed. There is an error due to the biological variance present. This error we can minimize by assuring that all the conditions are duplicated each time an experiment is performed. The standard error calculated from the 48 experimental data points for each dose is about 2% greater than the minimum er ror calculated from the binomial statistics. Th i s indicates that the biological var iance is about 2%, which is quite acceptable. Identification of cells can also be a source of e r ro r . It is sometimes d i f f icu l t , when locating single cells on the day of p lat ing, to d ist inguish cell size debr i s from the actual cel ls. Th i s debr is has many sources. Some of it may come from lysed cells. In order to eliminate this debris and decrease injury to the cel ls , care is taken to keep vortex ing and spinning of the cells at a minimum. Debris can also be caused by filaments from the fi lter paper , pieces of glass from pipette t ips , non-disso lved i ng red i -ents in the medium, and many other undef ined sources. To minimize these sources of debr i s the medium is ref i l tered using a .20 micron pore steri l ization f i l ter unit ( SYBRON/Na l ge ) . Th i s does not eliminate all debr i s but it does decrease it substant ia l ly. T h e cells can be easily d ist inguished from debr i s if the scor ing is performed after all the cells have d i v i ded , since there is v e r y little debr i s that looks like doublets. Single cells and debr i s , on the other hand, are often quite similar in appearance. T h u s , e r ro r is introduced if cells are identif ied before they have d i v ided . However, if not all the cells divide (see Section 3.1.5), then by only ident i fy ing those cells which have d iv ided er ror is again in t roduced. The exact magnitude of these er ror s is d i f f icult to measure because they are so small. Until the system is automated we chose to score after s eve r -al cell d ivis ions had taken place, due to the ease with which identification could be made. 3.1.2 Microtest plates The use of 96-well microtest plates proved to be ve ry advantageous. We orig inal ly used petri dishes with etched g r id but had considerable 49 problems with 'satellite' colonies that formed due to the action of the media removing cells from the colonies and starting new colonies nearby. The satellite colonies caused problems because it is very d i f f icul t , in the petri d i shes, to determine from which colonies the satellite colonies o r i g -inated. There fore it was dif f icult to determine the total number of cells or ig inat ing from a single cell on the day of plat ing. We thus had to examine the dishes every day so that we could d ist inguish between the small satellite colonies and the non - su rv i vo r s . By observ ing every day however, we increased the likelihood of satellite colonies forming due to movement of the dishes. The use of microtest plates eliminated these problems. The depth of medium in the wells of the microtest plates is approximately twice as great (see Table V) as in the petri d i shes, with a much smaller surface area. Th i s leads to v i r tua l ly no dis turbance of the colonies by movement of the media at the surface where the cells are attached, eliminating most satellite colonies. TABLE V Sp e c i f i c a t i o n s f o r p e t r i dishes and microtest plates p e t r i dishes microtest plates (per well) height of medium volume of medium diameter of medium surface area of medium Ratios: volume/height surface/volume 0.25 cm 5ml = 5cm3 5 cm 19.64 cm2 19.64 cm2 3.93 cm"1 0.56 cm 0.2ml = 0.2cm3 *0.67 cm 0.36 cm2 0.36 cm2 1.76 cm *Note; the diameter at the top of each well i s 7mm, at the bottom i t i s 6.4mm. The wells are about h a l f f u l l (.2ml out of .396ml), therefore a diameter of 6.7mm i s assumed for t h i s c a l c u l a t i o n . 50 The satellite colonies cause a great deal of frustration for the exper-imenter. The recorded location of the cells in the petri dishes with etched grid is only approximate and therefore the large number of satellite colonies tended to mask the location of the recorded colonies, making identification difficult, if not impossible, on day 7. Several experiments had to be aborted due to this difficulty. While attempting the preliminary experiments with the petris with etched grid it was recognized that only the fate of those cells initially located should be recorded. If a recording was made of all the colonies observed on day 7 then the number of dead colonies recorded would be much larger than actual fact due to the number of small satellite colonies misidentified as dead colonies. Therefore, to eliminate this error the fate of only those cells initially located on the first day was recorded. Even with this change in technique the satellite colonies could still contribute to an error in the number of dead cells recorded. The satellite colonies can either decrease or increase the number of dead cells recorded. The error increases with dose, as more non-survi-vors are present, and therefore such an artifact causes a change in the shape of the survival curve. A surviving colony is defined (for CHO cells) as one containing 50 cells or more at 7 days. The largest error that satellite colonies cause is when they are formed at an early date, such as day 3 or day 4. For example, take a colony which would have had just over 50 cells (in 7 days), and would therefore be a survivor. If it loses some of its cells in forming a satellite colony on day 3 then two colonies would form, both with fewer than 50 cells. The total, in this hypothetical case, would be 51 over 50 cells. Yet, the original colony would be recorded as -a non-survivor since it has fewer than fifty cells. It it also possible that the presence of satellite colonies could cause non-survivors to be recorded as survivors, although the probability of this happening is much smaller. Suppose there is a colony which would be just under 50 cells (and therefore a non-survivor). If on day 3 a cell from a different colony attached adjacent to this non-survivor, and con-tinued to divide, forming a group of greater than 50 cells, this colony would be recorded as a survivor. The main difficulty in eliminating these small errors is that it is impossible, with the petri dishes, to determine which satellite colonies belong to which 'parent' colonies. Thus, it is not possible to determine the total number of cells originating from a single cell on the day of plating. The use of microtest plates eliminated completely the problem of satellite colonies. It is very rare to find a satellite colony in the microtest wells. An approximate number would be one satellite colony in every 3000 cells scored (one satellite in every 24 microtest plates). However, even when a satellite colony does appear it is often possible to determine the main colony from which it originated since the cells are plated at a densi-ty of 1 to 3 cells per well. This eliminates the error because the total number of cells in the colony can be determined. Photographs were taken to show the comparison of colonies after 7 days growth in wells and in petris. These are shown in figure 8. The colonies in the wells are quite tightly formed while those in the dishes are spread out. 52 Figure 8. COLONIES IN MICROTEST PLATES AND PETRIS. 53 The use of microtest plates did not affect the survival measurements. Figure 9 shows the oxic survival curves obtained with asynchronous cells plated in microtest plates and in petri dishes with etched grid. The error bars represent the standard error obtained from the microtest plate data, where there is a minimum of 4 independent experiments for each point shown. Within this experimental error the two curves are identical. 3.1.3 Endpoint The endpoint used was the same as in 'classical' cell survival exper-iments. In this case, with CHO cells, a survivor is defined as a cell that can multiply and divide so as to form 50 cells or more per colony in 7 days. Since we examine the colonies microscopically there is very little ambiguity in using this endpoint. Very few of the colonies actually re-quire a cell count, only those that have about 50 cells. Colonies with many more than 50 cells or many fewer are easily categorized. Most colonies formed had many more than 50 cells. A healthy unirradiated cell will form a colony of several thousand cells in 7 days, since the doubling time is approximately 12 hours. 3.1.4 Percentage of Doublets Doublets (from a single cell which has just divided) in the cell population at the time of irradiation will affect the survival measurements. These doublets will cause an error because if only one of the cells of the doublet receives a lethal 'hit' from the radiation the other cell could still multiply and form a colony. If a single cell (instead of a doublet) was lethally hit, it would be a nonsurvivor. But if it were a doublet it would likely be a survivor. This error is, of course, present in the 'classical' technique also, but it is insignificant compared to the other errors pre-sent. 5'I T T 1.00 A s y n c h r o n o u s C H O c e l l s ( o x i c c o n d i t i o n s ) o r -o < DC Li . 0.90 p e I r i s CD 0.80 > > CC Z) CO 0.70 m i c r o t e s t p l a t e s 0.60 0 0.5 1.0 1.5 2.0 D O S E ( G r a y ) F i q u r e 9. S U R V I V A L OF CHO C E L L S P L A T E D IN M I C R O T E S T P L A T E S A N D PETRI D ISHES WITH E T C H E D GR ID . Exponent i a l l y g row ing a s ynch ronous CHO cel l s were i r r ad i a t ed in d i lu te su spens ions at 0°C unde r aerob ic cond i t ions (1.25 x 1 0 5 cel ls/ml in g r owth medium; X - r a y source : P i c k e r , 280 k V p , H V L = 1 . 7 mm C u , dose rate 0.3 C y / m i n ) . Ce l l s were then d i l u t e d , counted and plated e i ther into pe t r i d i shes with etched g r i d (dotted l ine) at a concent ra t ion of approx imate ly 400 cel l s per d i sh or into 96-wel l microtest plates ( so l id l ine) at a concent ra t ion of 1-3 ce l l s per we l l . The g r owth of i nd i v i dua l ce l l s was fol lowed microscop ica l l y o ve r a 7 day p e r i o d , so that the f r a c -tion of k i l l ed c e l l s , K, as well as the s u r v i v i n g f rac t ion S, was measured. The l ines rep re sent the l ea s t - squa re s f i t to the L-Q s u r v i v a l model S = e x p ( - a D - 6 D 2 ) . The e r r o r bars r ep re sen t the s t anda rd e r r o r s associated with the ce l l s p lated into microtest p la tes . 55 We determined, by scoring on the day of plating, the percentage of doublets and found this to be about 6.1%. The error caused by the doublets will obviously increase with dose, as more cells are killed, and therefore more doublets are 'hit.' It is probable that the actual survival curves for the asynchronous cells are more sensitive at the higher doses than we measured, due to this doublet error. The magnitude of this error can be calculated by assuming that the distribution of doublets in the cell population is uniform and that the probability of any one cell getting 'hit' is random. This error is about 0.1% at 90% survival, and increases to about 1% at 68% survival. The survival curves for the asynchronous data, with and without the doublet correction, are plotted in figure 10. The error bars represent the standard errors associated with the measured (without doublet error) values. The curves are the same within experi-mental error. This doublet error does not affect the reproducibility of any one survival curve. It is a systematic error affecting only the absolute sur-vival measurements. This doublet error can be eliminated by scoring on the day of plating and discarding doublets. It can also be eliminated by using synchronous populations of cells. As our method becomes more refined, and therefore this doublet error a more significant fraction of the total error, these steps must be taken. 3.1.5 Scoring The very first experiments were performed by observing the cells daily over the 7 day period. After the first few experiments we de-termined that, even the cells destined to form colonies of fewer than 50 cells (non-survivors), over 98% divided at least once and generally 56 T I I I I I A s y n c h r o n o u s C H O c e l l s w i t h d o u b l e t e r r o r 0 0.5 1.0 1.5 2.0 2.5 3.0 D O S E ( G r a y ) Figure 10. SURVIVAL OF CHO CELLS WITH DOUBLET ERROR. Exponentially growing asynchronous CHO cells were irradiated in dilute suspensions at 0°C under aerobic (0_) or hypoxic (N„) conditions (1.25 x 105 cells/ml in growth medium; X-ray source: Picker, 280 kVp, HVL = 1.7 mm Cu, dose rate 0.3 Cy/min). Cells were then diluted and plated in 96-well microtest plates. The growth of individual cells was followed microscopically over a 7 day period, so that the fraction of killed cells, K, as well as the surviving fraction S, was measured. The top curve in each set represents the least-squares fit to the L-Q survival model S = exp(-aD -BD2). The error bars are the standard errors associated with this data. The oxic curve is the same as in figure 9. The bottom (dotted) curve in each set represents the means of the same data, corrected for the effect of 6.1% doublets in the cell population. This correction is computed assuming 1) a uniform distribution of the doublets in the cell population and 2) the probability of any one cell getting 'hit' is random. The lower curve in each set is the least-squares fit (using the corrected means) to the L-Q survival model. 57 several times at these low doses (0.3 - 3 G y ) . Th i s conclusion is s u p -ported by various reports in the l iterature (e .g . E lk ind et al 1963, Colombo & Marin 1963, Froese 1966, Thompson S Suit 1967, Hurwitz & Tolmach 1969, Thompson & Suit 1969, Hopwood & Tolmach 1979, Col lyn d'Hooghe et al 1980, Crote et al 1981a, Jung 1982). Grote et al (1981a) reported that in the dose range of 0 to 3.8 Gy less than 1% of hamster cells (BHK 21 C13) failed to div ide at least once. Elkind et al (1963) reported that after a dose of 2.17 Gy to Chinese hamster cells (V79-1) , the non - surv i vo r s were capable of 4 to 5 d iv is ions. They also stated that " large exposures are reguired to reduce the number of post irradiation divis ions to 1 or fewer. " Froese (1966), who also used V79-1 Chinese hamster cel ls , reported that 99% of the unirrad iated cells d i v ided . Of the cells which received a dose of 5 Gy 98% d i v ided . Th i s percentage dropped to 97% after a dose of 8 G y . (Our highest dose was 3 G y . ) Colombo 5 Marin (1963) stated that "only the highest dose tested (4.5 Gy) caused some cell to die before the 2nd day following i r rad iat ion, that is within the f i rst or possibly 2nd divis ion c yc l e . " E lk ind & Whitmore (1967) report that after a dose of 10 G y , non - su rv i v ing Chinese hamster cells will d iv ide at least once on the average. Our own observat ions are in agree-ment with the results in the l i terature. Work in our laboratory show that at 5 Gy less than 3% of the cells failed to d i v ide , at 2 Gy and at zero dose over 98% of the cells d i v ided . T h u s the f requency of 'colonies' with only one cell is extremely small, at these low doses. The small number of non-d iv ider s is not a function of dose, at these low doses. The re fo re , by ident i fy ing only those cells which have d iv ided l ittle, if any , e r ro r is in t roduced. In fact, the er ror due to misidentifying debr i s as cells is el iminated, as mentioned in section 3.1.1. 58 A second consideration when scoring only the cells that have div ided is radiation induced divis ion delay. Froese & Cormack (1968) reported that, for Chinese hamster cells (V-79) irradiated with 5 g rays of ^ C o gamma rays , the median delay of colony-forming cells is about 1 hour , and that for non-colony-forming cells about 4 hours. These delays are comparable to previous ly reported values by E lk ind et al (1963) (0.96 hours per g r a y ) ; Froese (1966) (0.85 hours per g r a y ) ; and Yu & Sinclair (1967) (about 0.5 hour per g ray ) for whole populations of irradiated V-79 cel ls. Our experiments were performed with Chinese hamster ovary (CHO) cel ls , which have a cell cycle time similar to that of V-79 cel ls, and a similar radiation induced divis ion delay would be expected. Our highest dose was 3 Gy so the maximum delay expected would be about 3 hours . The re fo re , if the scor ing for d iv ided cells was done at least 16 hours after plating (doubl ing time of 13 hours plus 3 hours due to divis ion delay) all the cells should have d i v ided . Even when scor ing on day 1 , 24 hours after p lat ing, all the cells affected by radiation induced divis ion delay would have d i v ided . For the data presented in this thesis the earliest day of scor ing was usual ly day 2, 48 hours after p lat ing, well after any radiation induced divis ion delay has been overcome. The r a -diation induced divis ion delay would not influence our choice of cells to score. The use of microtest plates and the knowledge that over 98% of the cells at these doses d iv ided at least once allowed us to eliminate the scor ing on all but two days . The f irst s cor ing , to record the location of the cel ls , was done on day 2 or day 3, when all the cells have d iv ided at least once and general ly several times. Th i s was much easier on the 59 experimenter and eliminated any error due to misidentifying the debris as cells. The second and final scoring was done on day 7. Figure 11 compares the survival curves obtained when the initial scoring was performed on different days. The two curves, obtained with the asynchronous oxic cells, are the same within the experimental errors. The solid triangles represent data at 3 doses from 2 experiments where the initial scoring was done on day 1 . It is clear that the values obtained when first scoring on day 1 fit the survival curves obtained when the initial scoring was done on day 2 or day 3. Table VI compares the plating efficiency obtained with asynchronous cells when the day of initial scoring changes. TABLE VI PLATING EFFICIENCIES FOR DAY 0 AND DAY 2 Day of I n i t i a l P l a t i n g E f f i c i e n c y with Scoring Standard Deviation 0 0.92 ± 0.03 2 0.96 ± 0.02 Microtest plates were plated with unirradiated cells following the procedure used during X-ray experiments. Some plates were initially scored on day 0, 3 hours or more after plating, and others were initially scored on day 2. When the scoring was done on day 0 the location of all cells, whether singlets or doublets, was recorded. When the scoring was done on day 2 only the location of those cells which had divided at least once was recorded. The plates were then scored on day 7 to determine the plating efficiency. Since over 98% of the cells divide, the difference 60 Figure 11. T H E E F F E C T ON S U R V I V A L OF C H A N C I N G T H E INITIAL DAY OF SCOR ING. Exponentia l ly growing asynchronous CHO cells were treated as descr ibed in the caption of f igure 9. Cells were plated into 96-well microtest plates. The initial microscopic locating of the cells was performed on day 1 (24 hours after p lat ing, shown as solid t r i ang les ) , day 2 (48 hours after p lat ing, bottom dashed c u r v e ) , or on day 3 (72 hours after p lat ing, top solid c u r v e ) . The lines represent the least-squares fit to the L-Q s u r -vival model S = exp( -aD - 6 D 2 ) . The two e r ro r bars shown are the s tan -dard e r ro r s associated with the last dose point when the initial scor ing was performed on day 2 and day 3. To allow the e r ro r bars to be d i s t i n -guished they are not superimposed (although they are computed from the same dose represented by the t r iang le ) . 61 in plating eff ic iency is mainly due to the error of misidentifying debr is as cel ls. Th i s i l lustrates that fact tha t—un les s a method of easily d ist inguishing between debr is and single cells is employed, or unless the damage being assayed prevents all cell d i v i s i on—scor ing should be performed after the cells have d i v ided . A simplification of the technigue is that we don't have to f ind and record each and every ce l l . We can chose which cells we identify and follow, prov ided the choice is random. For instance we can d i s regard cells that are too close to other cel ls , or cells that are at the edge of the well, and thus in a shadow when viewed microscopical ly. The main assumption is that this selection is random and not based on di f ferences between cells dest ined to become surv i vor s ver sus non - su rv i vo r s . If we do this selection on the day of p lat ing, before any cell divis ion takes place, this is certainly t rue . If we do this selection on subseguent days , after one or more cell d ivis ions have taken place, care must be taken not to d i s regard cells that are only 2 to a colony versus those that are 8-16 cel l s /co lony. 62 3.2 T H E O X Y G E N E F F E C T A T LOW AND HIGH DOSES 3.2.1 P E R - Asynchronous F igure 12 represents the results of 'classical ' experiments (high dose assay) performed in our laboratory with CHO cells dur ing the years 1979 to 1981. These experiments were performed as descr ibed in section 2.3. The shaded area represents the low dose region of interest. F igure 13 represents results in the low dose region (the shaded region of f igure 12) using the technique presented in this thesis. Each low dose experiment ( f igure 13) contained a high dose control . A f ter irradiation and removal of the aliquots for the low doses the cell solution was fur ther irradiated to high doses so that the OER could be determined at 1% s u rv i va l , using the high dose assay ( f igure 12). Th i s enabled us to determine the OER at both high and low doses with the same cells under the same experimental condit ions. The surv iva l at the highest of the low dose points (2 Gy in 0 2 and 3 Gy in Nj) was determined using both assays. These results are shown in Table VII. Th i s data demonstrates that, within experimental e r ror (which is smaller for the low dose assay) , the two assays yield the same result , in the common dose range. TABLE VII. SURVIVAL, WITH STANDARD ERRORS, USING BOTH ASSAYS Dose low dose assay high dose assay 2 Gy in 3 Gy in 0.68 + .02 0.65 ± .06 0.75 ± .03 0.70 ± .05 It is ev ident. by comparing f igures 12 and 13, that the separation between the O^ and N^ curves is reduced in the low dose region, indicating a smaller oxygen effect. From the data of f igure 13, the OER 6 3 1.0 0.1 z: o H o < DC U_ O z > > 3 0.001 CO 0.01 — i 1 1 CHO CELLS SUMMATION (1979-1981) \ o2 J I 1 L 0 5 10 15 20 25 30 D O S E ( G r a y ) Figure 12. HIGH DOSE S U R V I V A L OF A S Y N C H R O N O U S CHO C E L L S . Exponential ly growing CHO cells were i rradiated in dilute suspensions at 0°C under aerobic (0^) or hypoxic (N_) condit ions (2 x 10 5 cells/ml in growth medium; X - r a y source: P icker , "280 k V p , H V L = 1.7 mm C u , dose rate 2 Gy /min ) . Cells were then d i lu ted, counted and plated into petri d ishes. A f te r a 7 day incubation per iod, the su rv i v ing fraction S was determined. Solid lines t h r o L i g h the experimental points represent the least-squares fit to the L-Q surv iva l model S = exp( -aD - 6 D 2 ) . The shaded region represents the surv iva l /dose region shown in f igure 13. 6'I i 1 r i r A s y n c h r o n o u s C H O c e l l s 1.00 2 : o I— O < CC LL CD > > DC Z> CO 0.9 0 0.80 0.70 Kt 1 Y 0.60 J L J L 0 0.5 1.0 1.5 2.0 2.5 3.0 D O S E ( G r a y ) Figure 13. LOW DOSE S U R V I V A L OF A S Y N C H R O N O U S C E L L S . Cells were treated as descr ibed in f igure 12 except that after irradiation they were plated into microtest plates (and/or petri d i shes ) . The growth of individual cells was followed microscopically over a 7 day per iod, so that the fraction of ki l led cel ls, K, as well as the su r v i v i n g fraction S, was measured. Each point represents a minimum of 8 microtest plates (approximately 100 cells per plate) from at least 4 independent e x p e r i -ments. The cr i ter ion for su rv i v ing cells is the same as in f igure 12 (50 cells or more/colony in 7 d a y s ) . The data were again fitted to the q u a -dratic equation as descr ibed in the caption to f igure 12. 65 was determined at various doses using an iterative parametric analysis procedure (Lam et al 1979, Lam et al 1981) and the results appear in figure 14, plotted as a function of the. radiation dose delivered in the presence of Oj. The best fit of the equation OER = a + bD is used to transform the oxygen data to the nitrogen data. The two sets of data (the nitrogen data and the transformed oxygen data) are pooled and fitted with a 2nd degree polynomial. The sum of squares of these data points about this fitted polynomial is then calculated. The computer program then performs an iterative procedure to find the best a,b pair to transform the oxygen data so that the sum of squares of the combined data about the quadratic polynomial is minimized. Given the desired confidence limit (in this case, 63% confidence limit which is one standard deviation) an oval in the a,b plane is generated corresponding to the uncertainty in a and b. This procedure provides the best OER of the form a + bD, and the associated uncertainty. The data shown in figure 14 can be adequately represented by a straight line (on a linear dose scale) indicating an OER value at the lowest dose of approximately 1.5. At the highest dose measured (1.5 Gy in 0 2) the OER is approximately 2. The error bars represent one standard deviation. 66 0.2 0.4 0.6 0.8 1 2 DOSE (Gy) IN O, Figure 14. O X Y G E N E F F E C T FOR A S Y N C H R O N O U S CHO C E L L S . Solid c irc les represent the OER values calculated from the actual data points in f i gure 13. The data points for 0 are transformed using a dose-dependent OER descr ibed by the equat ion: OER = a + bD. The parameters a and b are fitted by an iterat ive procedure (see text) which minimizes the squared er ror s in S (least squares f itt ing) and which also gives the uncerta int ies associated with each OER value. T h e e r ro r bars represent one s tandard deviation (63% conf idence l imit). Fo r comparison, the OER value at h igher dose (at S = 0.01 and calculated from the data in f igure 12, open c i r c le , is shown. 67 Figure 15 represents the averages of the data in figure 13, plotted as ln(S/S^)/D versus radiation dose. This results in a graphical linear-ization of the L-Q equation; ln(S/S )/D = -a - BD. o The slope of each plot is the quadratic inactivation constant, B, and the intercept with the zero dose axis is a, the linear inactivation constant (Chapman 1980). This type of cell inactivation plot allows the reader to readily determine any changes of the independent parameters of cell inactivation. From figure 15 we can see that the affect of hypoxia is to cause a substantial decrease in the value of 8 . The linear inactivation constant, a, shows a slight decrease for the hypoxic cells. 68 Figure 15. A S Y N C H R O N O U S CHO C E L L S . Data points shown are the same as the data presented in f igure 13, plotted to y ie ld a graphica l l inearization of the L -Q equation S = exp( -aD - 6 D 2 ) . T h e slope of each plot is the quadrat ic inactivation constant, 6, and the intercept with the zero dose axis is a, the linear inactivation constant. 69 3 . 2 . 2 OER -Synchronous The OER was measured at 2 points in the cell cyc le ; 1) a radiosens i -tive stage at 3 hours after shake-off assumed to be the C l / S border and 2 ) a radioresistant stage at 8 hours after shake-of f when the cells are in late S. The isodose curve shown in f igure 6 (page 4 3 ) demonstrates the vary ing radiosensit ivit ies of cells as they pass through the cell cyc le. Each low dose experiment contained a high dose control. A f ter removing aliquots of cells at prescr ibed times (for the low dose assay) the same set of cells was fu r ther irradiated to h igher doses. Complete s u r -vival cu rves for the synchron ized cells were not obtained at high doses due to the d i f f icu l ty of obtaining a suff ic ient number of synchronized cells. Surv iva l was only obtained for a few dose points, which were at approximately the same surv iva l level in oxygen and n i t rogen, so that an approximate value for OER could be obtained. Th i s allowed us to de ter -mine whether the OER decreased at lower doses. The surv iva l curves for oxygenated cells i rradiated with low doses for a synchronous , G1/S , and late S cells are shown in f igure 16. As expected the G l IS cells are more sensit ive to the radiation than the late S cel ls. The re is essential ly no d i f ference, within experimental e r r o r , between the asynchronous cells and the cells in the G l / S phase of the cell cycle at these low doses. A similar relationship is shown for hypoxic conditions in f igure 17. For hypoxic cells the asynchronous and G1/S cells have identical r e -sponses, in the dose range shown. The late S cells a re, as expected, more resistant to the radiation. However, the d i f ferences are not as great between the curves as they were for oxygen. Th i s lends support to theories that propose d i f ferent repair mechanisms or d i f ferent lesions in hypoxia and oxygen . 7 0 1 . 0 0 O — 0 . 9 0 O < CC LL CD 21 > > rr Z> CO 0 . 8 0 U 0 . 7 0 0 . 6 0 C H O c e l l s ( o x i c ) L a t e S 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 DOSE (Gray) Figure 16. S U R V I V A L C U R V E S FOR A E R O B I C CHO C E L L S A T LOW DOSES. Th i s plot represents three d i f ferent populations of CHO cells i rradiated under oxic condit ions; asynchronous cel ls, cells in the late S stage of the cell cyc le , and cells in the G1/S interface stage of the cell cyc le . The lines represent the least squares fit of the L -Q equation to the e x p e r i -mental points. The data, with error bar s , for these curves can be seen in f igures, 13, 18, and 21. 71 o H O < rr LL. o > rr Z> CO 1 . 0 0 0 . 9 0 0 . 8 0 0 . 7 0 0 . 6 0 C H O c e l l s ( h y p o x i c ) 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 D O S E ( G r a y ) Figure 17. S U R V I V A L C U R V E S FOR HYPOX IC CHO C E L L S A T LOW DOSES. Th i s plot represents the same three populations of CHO cells shown in f igure 16, i rradiated under hypox ic condit ions. The s t i r red cell su s -pensions were gassed for 45 minutes pr ior to and dur ing irradiat ion with N 2 (less than 5ppm 0 2 ) . The lines represent the least squares fit of the L-Q equation to the experimental points. The data, with er ror bar s , for these curves can be seen in f i gures , 13, 18, and 21. 72 The curves shown in f igure 18 are the best fits of the L-Q eguation using the. nitrogen and oxygen data from the cells in the C1/S stage of the cell cyc le. The er ror bars represent standard e r ro r s . Al l points represent a minimum of two independent experiments. From the data of f igure 18, the OER was determined at various doses using an iterative parametric analysis procedure (Lam et al 1979, Lam et al 1981) and the results appear in f igure 19, plotted as a function of the radiation dose del ivered in the presence of O^. To obtain these results a quadrat ic relationship was assumed for the surv iva l curves ( L -Q model). The C1/S cells show an OER of 1.7 ± 0.2 at a dose of 1.4 Gy in oxygen . These same cells show an OER of 2.1 ± 0.2 at about 6 Gy in oxygen (about 10% surv iva l level). F igure 20 represents the averages of the data in f igure 18, plotted as ln(S/S )/D versus radiation dose. The data are more scattered than o the asynchronous data, due to the fact that fewer experiments were p e r -formed. Th i s makes it d i f f icult to draw conclusions. The a and 6 values for the hypoxic G1/S cells are the same as the values for the hypoxic asynchronous cel ls. Th i s is expected from f igure 17, which showed that the hypoxic surv iva l cu rves were identical. The a and 6 values for the aerobic G1/S cells have both decreased relative to the aerobic a s ynch ro -nous cells. The data for the late S cells are presented similarly in f igures 21, 22 and 23. The late S cells show an OER of 1.4 ± 0.3 at 1.4 Gy in oxygen and an OER of 2.3 i 0.2 at 6 Gy in oxygen (about 30% surv iva l level) . Th i s demonstrates a lower OER at low doses. The scatter of points is h igher for the late S cells than for the G1/S cel ls. Th i s is due, in part , to a s l ight loss in synchrony as the cells 7 3 o h-o < cc LL CD Z > > CC Z) 00 1.00 0.90 0.80 0.70 0.60 1 1 1 r C H O c e l l s ( G 1 / S ) J L 0 0.5 1.0 1.5 2.0 2.5 3.0 D O S E ( G r a y ) Figure 18. LOW DOSE S U R V I V A L OF CHO C E L L S IN C1/S S T A G E OF T H E C E L L C Y C L E . Exponential ly growing CHO cells were synchron ized us ing the mitotic shake-of f technique. Cells were taken from the synchron ized population 3 hours after shake-of f when they were near the G1/S interface stage of the cell cyc le and irradiated in dilute suspensions at 0°C under aerobic (0_) or hypox ic (N.) conditions (5 x 10 3 cells/ml in growth medium; X - r a y source: P i cker , 280 k V p , H V L = 1.7 mm C u , dose rate 0.3 Gy /min ) . T h e cells were then di luted and plated into 96-well microtest plates. The growth of indiv idual cells was followed microscopical ly over a 7 day per iod , so that the fraction of ki l led cel l s , K, as well as the surv i v ing fraction S, was measured. Each point represents a minimum of 2 independent experiments. Solid lines through the experimental points represent the least-squares fit to the L-Q su rv iva l model S = exp( -aD - 0 D 2 ) . The e r ro r bars are the computed s tandard e r r o r s . 74 cc Ul O T r 1 — i — i i i | CHO cells (Gl/S) T 1 T — I — I — r — r < rr r-Z til 2 UJ o z < X z ui z UJ o X o J — I I I • I I J I I l l l l 0.2 0.4 0.6 0.8 1 1 0 DOSE (Gy) I N Og Figure 19. OXYGEN EFFECT OF Gl/S CHO CELLS. Solid circles represent the OER values calculated from the actual data points in figure 18. The data points for 0_ are transformed using a dose-dependent OER described by the equation: OER = a + bD. The parameters a and b are fitted by an iterative procedure (see text) which minimizes the squared errors in S (least squares fitting) and which also gives the uncertainties associated with each OER value. The error bars represent one standard deviation (63% confidence limit). For comparison, the OER value at higher dose (at S = 0.10), open circle, is shown. 75 C H O c e l l s (G 1 / S ) I >v crj CD CM I o X Q CO C 0.6 1.2 1.8 2 . 4 3.0 3.6 D O S E (Gray) Figure 20. CHO C E L L S IN G l /S. Data points shown are the same as the data presented in f i gure 18, plotted to y ie ld a graphica l l inearization of the L -Q equat ion. T h e slope of each plot is the quadrat ic inactivation constant, 3, and the intercept with the zero dose axis is a , the linear inactivation constant. T h e arrow on the e r ro r bar for the f irst oxic point indicates that the e r ro r extends beyond the g raph border . 76 o r— o < c r L L > cc ZD CO 1.00 0.90 U 0.80 k 0.70 k 0.60 0 0.5 1.0 1.5 2.0 2.5 3.0 D O S E ( G r a y ) Figure 21. LOW DOSE S U R V I V A L OF CHO C E L L S IN L A T E S S T A G E OF T H E C E L L C Y C L E . Exponential ly growing CHO cells were synchron ized using the mitotic shake-of f technigue. Cells were taken from the synchron ized population 8 hours after shake-of f when they were in the late S stage of the cell cycle and irradiated in di lute suspensions at 0°C under aerobic (0^) or hypoxic (N_) conditions (5 x 10 3 cells/ml in growth medium; X - r a y source: P ick-er , 280 k V p , H V L = 1.7 mm C u , dose rate 0.3 Gy /m in ) . The cells were then di luted and plated into 96-well microtest plates. The growth of individual cells was followed microscopically over a 7 day per iod, so that the fraction of kil led cel l s , K, as well as the su rv i v ing fraction S, was measured. Each point represents a minimum of 2 independent experiments. Solid lines through the experimental points represent the least-sguares fit to the L-Q surv iva l model S = exp(-aD-(BD 2 ) . The er ror bars are the computed standard e r r o r s . 77 rr UJ O < rr t-2 UJ ui o 2 < I 2 Ul Ul CJ >• X o T 1 1 l l l l 1 1—I I I CHO cells (Late S) J -I I 1 1 1 1 -I I i - t i t 0 . 2 0 . 4 0 . 6 0 . 6 1 6 1 0 DOSE (Gy) IN 0 „ Figure 22. OXYGEN EFFECT FOR CHO CELLS IN LATE S. Solid circles represent the OER values calculated from the actual data points in figure 21. The data points for 02 are transformed using a dose-dependent OER described by the equation: OER = a + bD. The parameters a and b are fitted by an iterative procedure (see text) which minimizes the squared errors in S (least squares fitting) and which also gives the uncertainties associated with each OER value. The error bars represent one standard deviation (63% confidence limit). For comparison, the OER value at higher dose (at S = 0.30), open circle, is shown. 78 0.6 1.2 1.8 2.4 3.0 3.6 D O S E (Gray) Figure 23. LATE S CHO CELLS. Data points shown are the same as the data presented in figure 21, plotted to yield a graphical linearization of the L-Q equation. The slope of each plot is the quadratic inactivation constant, 8, and the intercept with the zero dose axis is a, the linear inactivation constant. The arrows on the error bar for the first oxic point indicate that the error extends beyond the graph border. 79 progress through the cell cycle. The fact that there are only 2 experi-ments for most dose points also contributes to the scatter of the data points. This scatter can clearly be seen in figure 23. This analysis is still valuable,' however, for although absolute numbers can not be deter-mined, general trends can be observed. It is doubtful that the value of the quadratic inactivation constant actually changes sign. This is probably just ah artifact of the scatter of the data. Within the error bars shown in figure 23 a positive $ can easily be obtained. The a and 8 values appear to be the same for both the hypoxic and aerobic cells in late S. The a value may be higher than any of the a values for either asynchronous cells or cells in the G1/S stage of the cell cycle, although the scatter of data makes this conclusion very tentative. Unlike the asynchronous or G1/S cells, hypoxia does not affect the 6 value for cells in the late S stage of the cell cycle. 80 4. DISCUSSION 4.1 T E C H N I Q U E 4.1.1 Uses of technique T h i s technique y ie lds cell su rv i va l data in the shoulder region of the cell radiation surv iva l cu rve which are reproducible to within 2-4%. The technique can be used to assay low level effects of carc inogens, var ious types of radiat ion, d r u g s , etc. For example, most of cel lular radiobiology could be examined at 'cl inical ly relevant ' doses by this technique. 4.1.2 Identification of Cells T h e major assumption involved in this technique is that identif ication of a cell can be done unambiguously. At low doses, radiation damage is not ev ident unt i l a f ter the f i r s t cell d iv i s ion. If scor ing is performed immediately after p la t ing , before any cell divis ion has o c c u r r e d , the identif ication of the cells will not be affected by the radiation damage. However, when scor ing single ce l l s , the probabi l i ty of misidentifying debr i s as cells is increased (see section 3.1.5). Th i s e r ro r is a systematic e r r o r , causing a decrease in the plating eff ic iency ( P E ) , and it is not dose dependent. Addit ional problems when scor ing s ingle cells are related to this problem of mis identifying debr i s as cel ls . When ident i fy ing single ce l l s , additional time is requ i red to locate the prescr ibed number of cel ls . T h e end result is that fewer cells can be located when scor ing single cel l s . T h e problem is compounded by the fact that it is necessary to score the complete experimental set before the majority of the cells have completed d iv i s ion . Since the cell sample is smaller the e r ro r is again increased. 81 Some of these problems could be minimized when, the automated system is completely implemented (see section 4.1.5.1). If the scoring is performed after the first cell division there is little problem in discerning between debris and cells. As discussed in section 3.1.5 reports in the literature (Elkind et al 1963, Froese 1966, Crote et al 1981a) suggest that at the doses we are using less than 1% of the cells fail to divide. Our own data demonstrate that the percentage is less than 2% and it is not dose dependent, at these low doses (see section 3.1.5). However, if asynchronous cells are used when scoring after the cells have divided, a very small error is introduced due to the percentage of doublets in the cell population. This error is less than 0.1% at 90% sur-vival and increases to only 1% at 68% survival (see figure 10, page 56). This error can be eliminated by scoring twice or by the use of syn-chronized populations of cells (see section 3.1.4). 4.1.3 Plating Efficiency 4.1.3.1 Factors affecting the plating efficiency The plating efficiency (PE) with the low dose assay is much higher than that obtained in the classical survival experiment (high dose assay) as demonstrated in Table VIII. The cells for each assay are treated exactly the same until the time of plating. For the control and equivalent dose points the cell samples are plated from the same test tube. The only difference after irradiation is that some of the cells are plated into petri dishes with feeders (high dose assay) and some into microtest plates without feeders (low dose assay). The difference in PE thus cannot be due to the handling of the cells. It also is not due to the feeders. Studies have showtr that, if the feeders affect the PE at all, it will in-crease it slightly (Elkind & Whitmore 1967). 82 The following are possible reasons for the difference in PE between the low and high dose assays: 1) The presence of a large number of non-dividing cells. This was discussed earlier is sections 3.1.5. and 4.1.2. (for more details see these sections). Reports in the literature and our own observations conclude that at doses between 0 and 3 Cy less than 2% of the cells fail to divide. It is not a dose dependent phenome-non ,/ The difference in PE is obviously not due to the presence of non-dividing cells. 2) The amount of debris present. The Coulter Counter (used in the high dose assay to determine the number of cells) counts some of the debris as cells; this leads to a lower PE. This is the most likely explanation and is discussed in more detail in the next section. 3) The presence of a large number of unattached cells. This would actually have the same effect as debris. Measurements in our laboratory have determined that less than 1% of cells fail to adhere to the bottom. This could not account for the large differences in PE, although it could contribute to it. In section 4.1.3.3. this is discussed in much more detail. 4) Dilution and plating errors in the high dose assay. It is doubtful that these errors could be high enough to account for the dif-ference in PE. In the first place any dilution and plating errors would be random and therefore could not explain the consistently lower PE for the high dose assay. Secondly for the asynchronous experiments the cells in the dilution tubes were at a relatively low concentration (6.5 X 103 cells/ml) so that it was not necessary to plate extremely small volumes (22 - 45 yl were the volumes plated). This minimizes the error in the dilution and plating. The concentration in the dilution tubes of the synchronous cells was even lower (1.5 X 103 cells/ml), therefore a much larger volume was plated (150 - 200 yl). The dilution and plating errors for the 83 synchronous cells were therefore lower than for the asynchronous cells. However, the difference in PE between the high and low dose assays is much greater for the synchronous cells. It is therefore unlikely that the dilution and plating errors are responsible for the difference in PE be-tween the two assays. TABLE VIII. ~ ~ ~ — PLATING EFFICIENCIES FOR LOW AND HIGH DOSE ASSAYS low dose assay high dose assay Asynchronous Cells in a i r 0.94 ± 0.01 0.78 ± 0.03 in nitrogen 0.93 ±0.01 0.74 ± 0.03 Gl/S Cells in a i r 0.95 ± 0.01 0.48 ± 0.06 in nitrogen 0.95 + 0.01 0.56 ±0.04 Late S Cells in a i r 0.92 ± 0.01 0.43 ± 0.15 in nitrogen 0.91 ± 0.004 0.52 ± 0.04 4.1.3.2 Debris The principle of counting cells using the electronic Coulter Counter is best described by quoting from the instruction manual (Instruction manual for the Coulter Counter, Model Zg, Coulter Electronics, Inc., Haleah, Florida, page 2-1.). Particles or cells, suspended in an electrolyte, can be sized and counted by passing them through an orifice (aperture) with a specific path of current flow for a given length of time. As particles or cells pass through the aperture and dis-place an equal volume of electrolyte, the resistance in the path of current changes. This results in corresponding current and voltage changes. The quantity (magnitude) of this change is directly pro-portional to the volumetric size of the particle or cell. 84 The number of changes within a specific length of time is proportional to the number of particles or cells within the suspension. It is possible to set a 'window' of acceptable voltage and current values, this corresponds to a window of acceptable volume. Any particle which is within this volume (e.g. cells, cell size debris, etc.), is counted as a cell. By examining the cell suspension microscopically it is evident that there is a great deal of debris, that, on the basis of size alone, would conceivably be counted as cells. The difference in PE between the high and low dose assays demonstrates that this error may be of the order of 15% or so. The PE difference between the low and high dose assays is much greater when synchronized cells are employed. I believe this is a further indication that the Coulter Counter is counting debris as cells. Since the synchronized cells are handled much more vigorously than the asynchro-nous cells (besides the shake-off they remain in a stirred suspension for up to 8 hours) there is a lot more debris present in the cell suspension. This has been verified when examining the cell suspension microscopical-ly. When determining the cell concentration for the high dose assay, some of this debris is counted by the Coulter Counter. However, it is mainly ignored when determining the cell number for the low dose assay, espe-cially when the initial day of scoring is performed after the cells have divided. This could account for the large difference in PE between the high and low dose assays. The difference in PE is a systematic effect and thus does not affect the survival results or the measurement of radiobiological parameters such as OER. 85 4.1.3.3.Unattached cells. The Coulter Counter also counts cells that will not attach when plated while the low dose assay does not. If these cells were present then a difference in plating efficiency (PE) between the high and low dose assays would be expected. This is not a radiation effect since the nor-malized survivals obtained using both the low and high dose assays are the same (at the same dose, see Table VII). To determine the number of unattached cells two methods were used. The most accurate method, which was very tedious, involved plating cells into 96-well microtest plates, 100 ul of cell suspension per well. The microtest plates were examined microscopically and all attached and unattached cells were recorded. Both unirradiated cells and cells irradiated with 2 Cy were examined. The total volume of medium in each well was examined microscopically and no unattached cells were detected floating in the medium. Since only a limited number of cells were examined (129 cells - 0 dose, 144 cells - 2 Cy) we can estimate that the unattached cells represent less than 1% of the cell population. A second method (described in the next paragraph), which had a 6% error, confirmed that there were no unattached cells in the unirradiated cell suspension. The cell concentration of a suspension was determined using a hema-cytometer. A known volume of cells was then plated and allowed to adhere to the bottom of a petri dish (with 2 mm etched grid) for 1 hour. The number of cells which had attached was then observed microscopically. There was an error of approximately 6% in both the measurements of cell concentration (using the hemacytometer) and number of attached cells (using the petri dishes with etched grid) due to statistical fluctuations and sample size. The difference between the number of cells plated (using 86 the hemacytometer) and the number of attached cells (using the petri dishes with etched grid), which is the number of unattached cells, was about 3%. Within the 6% experimental error this is insignificant. This confirms the result obtained by examining the plated cell suspension microscopically. The difference in PE between the low and high dose assays is not due to unattached cells. 4.1.4 Feeder cells Feeder cells have been irradiated at high doses (60 Gy) so that they won't divide (or adhere to the bottom), but they continue to metabolize, providing a better environment (hopefully) for the irradiated cells to grow in. Due to the fact that we microscopically identify our cells intro-duction of feeder cells would cause many difficulties in cell identification. Feeder cells often increase the average diameter of control colonies (Elkind & Whitmore 1967) and, as well, can increase the plating efficiency for cell lines with very low PE. At present, both our colony size and plating efficiency are quite satisfactory. However, if the introduction of feeder cells was desirable (for example, to reduce the differences in plating between the high and low dose assays, or to further increase our already high PE), there are two possibilities. One is to use feeders of a different cell type so that there would be no problem in identifying the cells being assayed. Another possibility is to use preconditioned medium. To condition the medium, feeder cells are placed in medium for a prede-termined period of time. The medium is then filtered before being used to plate the cells in, eliminating the actual feeder cells. However, if too many feeder cells are used, the medium becomes spent, and the plating efficiency will decrease rather than increase (Elkind & Whitmore 1967). 87 Therefore complete control experiments would be necessary if the situation warranted the use of feeder cells. Currently, the introduction of feeder cells would be an unnecessary complication, as both our colony size and plating efficiency are quite satisfactory. 4.1.5 Improvements and future plans • 4.1.5.1 Automation A degree of automation of the technique is underway in our labora-tory and will make the system much less labor intensive. A computer controlled stage is used in conjunction with the microscope. The observer examines the plated dish microscopically. When a cell is located the coor-dinates are recorded by the computer. Subsequent observations of the cells are greatly facilitated since, as the computer controlled stage brings o each cell into the field of view, the observer simply examines the cell and the observation (e.g. survivor or non-survivor) is recorded by the computer. The image of the cell can also be projected onto a TV screen, facilitating the examination of the cells. Automation will enable a single experimenter to examine a larger cell sample, which, in turn, yields better statistics and smaller standard errors. 4.1.5.2 Synchronized Experiments Cells in different stages of the cell cycle have different radiosensi-tivities (see figure 6, page 43). Therefore, when irradiating asynchro-nous cells the resultant survival curve is a composite of many different radiosensitivities, which contributes to the scatter in the data. A popu-lation of cells which is homogeneous with respect to the effects of ra-diation, would be extremely useful when examining mechanisms of damage and/or low level effects. To obtain such a population it is necessary to 88 synchronize the cells. We have begun experiments with cells synchronized using the mitotic shake-off technique. The technique used to obtain the data in this thesis is extremely labor intensive, particularly when synchronized cells are used. To obtain 4 more measurements for each of the 8 dose points shown in figures 18 and 21 would take approximately 24 more successful experiments (at 40 hours per experiment). Until the tech-nique becomes automated and other improvements are made it is extremely difficult for one person to acquire enough data with synchronous cells to draw meaningful conclusions. For any serious work with synchronized cells an automatic syn-chronizing method is a necessity. A method that would acquire large quantities of well synchronized cells (with a minimum of experimenter participation) would greatly enhance our ability to acquire sufficient data. 4.1.5.3 Other Possibilities There is a wealth of information available with this technique because the cells are examined microscopically. This information is inaccessible when using macroscopic techniques. For example, growth rate studies, where the cells are studied daily to determine colony size, would be very feasible with this technique, especially once the system becomes automated. Both the survivors and non-survivors exhibit a range of colony sizes. We were careful in these studies to use the same endpoint as with the high dose assay (50 cells/ colony/ 7 days = survivor). However, it would be worthwhile to further analyze the data. In preliminary investigations we have chosen to divide the non-survivors into two distinct groups; 1) those that we feel are truely dead and will never attain 50 cells or more per colony, these are called D or dead cells and 2) 89 those that are only dead by definition, if given enough time they will form colonies of more than 50 cells, these are called QD or guote dead cells. There are many guestions we could attempt to answer if the number of D and QD cells was known for different conditions: Is the ratio of D/QD a function of dose? Is the ratio a function of LET? Can the D cells be attributed to single hit kill [a effect)? Can the QD cells be the result of multiple-hit cell kill (g effect)? Does the D/QD ratio change when radiation sensitizers are used? The answers to these questions could be used to test (or to formulate) hypotheses upon which some survival models are based. We believe there is much useful information here. It is conceivable that when the automated system is completely imple-mented different observers can examine the cells for different endpoints and such information as mentioned above could be easily obtained from ongoing experiments. Separating subpopulations of surviving and non-surviving cells into different classifications is just one of many possibilities for further analy-sis and new information that this technique could yield. A study of the morphology of a dying cell would also be of interest. On the day of plating, before any cell division takes place, no difference is discernible between cells that will become survivors and those that will be non-survivors (at these low doses). However on day 3 there are considerable differences. If these differences could be quantified and shown to be of predictive value, the possibility of shortening the assay is obvious. 90 4.1.6 Limitations This technique has been developed for use at low doses of radiation and a word of caution is in order. As can be seen from the plot in figure 3 the percentage error in the surviving fraction remains small only when S is greater than K. Once the doses are no longer small and the surviving fraction has decreased, the advantage is gone and the error increases. Also, at very high doses (e.g. 20 Gy), injury to the cell is evident even before the first division, and the error in identifying cells is greater. There are two limitations to beware of when applying this technique. The first one is that the cells have to be able to adhere to the surface of the culture vessel. If the cells don't adhere (for example, after some chemical treatment), they cannot be located microscopically. The advan-tage of being able to measure both the killed and surviving fractions is then lost. The second limitation is that we are only able to measure low level effects. Once the surviving fraction drops much below 50% most of the advantages of this technique are lost. The two assays compared in this thesis should be used to complement each other. The low dose assay should be used when the survival level is above about 50% and the high dose assay at lower survival levels. This combination of techniques will be especially valuable when attempting to describe radiobiological parameters, examine models, determine the effect of radiation sensitizers, etc., over a complete dose range. 91 4.2 OXYGEN EXPERIMENTS 4.2.1 The importance of OER at low radiation doses Many investigators believe that the radioresistance of hypoxic cells is a major limiting factor in radiotherapy (see section 1.5, pages 28-30). To determine the importance of hypoxic cells to radiation therapy the effect of oxygen at clinically relevant doses (below 3 Cy) should be explored. There is a large amount of data describing the oxygen enhancement ratio (OER) at high doses. However, whether or not the results of these data (from high doses) can be extrapolated to determine the effect of oxygen at low, clinically relevant doses is a subject which has been much debated. The lack of sufficient and accurate cell survival data at low doses contributes to the continuation of this debate. There are currently 3 major viewpoints on the oxygen enhancement ratio (OER) as a function of dose: 1) that the OER is constant with dose, 2) that the OER decreases at low doses and 3) that the OER is less than unity at very low doses and thus has a protective effect. The resolution of this debate has important implications; for example it applies to the fractionation regimes in radiotherapy (e.g. Douglas 1.982a,b). A whole conference was devoted to the question of cell survival after low doses of radiation (Sixth L.H. Cray Conference, London, 1974) where among other subjects the question of oxygen enhancement ratio at low doses was examined. This question is still not resolved today. As an indication of the interest that is still apparent in this problem, there is a workshop addressing the question of the inactivation of cells at low doses at the 1983 Radiation Research annual meeting. As mentioned in the introduction (page 25) at the Gray conference in 1974 some researchers presented data and arguments suggesting a diminished OER at low doses of ionizing 92 radiation in mammalian cells irradiated jn vitro (Revesz et al 1975, Chapman et al 1975b, McNally 1975, Pettersen et al 1975) while others presented data and arguments claiming that the OER is constant throughout the dose range (Phillips et al 1975, Koch 1975). All these results were obtained by the standard technique of measuring cell sur-vival. / Some researchers have reported a protective effect of oxygen at very low doses. Littbrand & Revesz (1969) claim to have obtained an OER of less than unity at very low doses. Cullen et al (1980) demonstrated that this effect could be the result of inhomogeneities in the dose re-ceived at the cell surface, with the different gases used. Nias et al (1973) obtained survival curves with human cells (HeLa cell line), that demonstrate a similar crossover at low doses. Although the best fit to the data gave an OER of less than unity the difference was not considered to be significant (Nias 1974). It is not possible to completely rule out an OER of less than unity (at low doses) because of the large amount of scatter in survival mea-surements. This is another demonstration of the need for accurate cell survival data at low doses. 4.2.2 Predictions for OER from survival models The majority of the OER data available is at high doses. We must therefore rely extensively on survival models to make predictions at low doses. This discussion will be limited to 3 of the most widely used mod-els, the 2 parameter single hit-multi target model (SH-MT), the 3 para-meter repair-misrepair model (RMR), and the linear-quadratic model (L-Q). The equations for these were all shown in Table I (page 11). The predictions with regard to the OER of these 3 models will be examined. 93 We will then d iscuss the experimental evidence in light of these p r e -dict ions. The OER is def ined as the ratio of the doses, in the absence and presence of oxygen , requ i red to reduce the cell surv iva l to a certa in level. To ar r i ve at the express ion for OER for each surv iva l model we simply equate the express ions for surv iva l in the absence and presence of oxygen and determine the ratio of the doses. 4.2.2.1 The SH-MT model In the following equations D Q and n are the SH-MT parameters for the aerobic surv iva l c u r v e , D ' and n' are the SH-MT parameters for the o ^ hypoxic surv iva l c u r v e , D is the dose in oxygen at the prescr ibed surv iva l level ( S Q ) , and is the dose in nitrogen (or hypoxia) at the same surv iva l level (S ). n S = l-(l-exp(-D /D ) ) n = S = l-(l-exp(-D /D')) n' o x o n y o ( l -exp(-D x/D o)) n = l-exp(-D y/D;)) n' ( l - e x p(-D x/D o)) n / n' = l-exp(-Dy/D^) expC-Dy/D^) = l - ( l - e x p(-D x/D o)) n / n' -Dy/D^ = l n [ l - ( l - e x p(-D x/D o)) n / n'] OER = D /D = y x -(D;/Dx) * l n [ l - ( l - e x p(-D x/D o)) n / n'] ( 1 ) When n = n 1 the express ion within the square brackets becomes; [l-(l-exp(-D /D ))*] = exp(-D /D ) X O X o and the OER is reduced to: OER(n=n') = -(D'/D ) * (-D /D ) = D'/D o x x o o o D q is the inverse of the slope of the exponential portion of the cell surv iva l c u r v e . Proponents of the SH -MT model often define the OER as the ratio of the slopes of the exponential portions of the surv iva l cu rves 94 (this leads to an OER which is constant with dose). However this is only true if the extrapolation numbers are the same. To report a difference in the extrapolation number for hypoxic and aerobic cells and also to state that the OER is constant and equal to the ratio of the slopes is not only inconsistent but wrong. To determine the limiting values of the OER from equation 1 the following approximations are made: as x approaches 0, exp(-x)=1-x, 1-exp(-x)=x, ln(1-x)=(-x) As D x approaches 0; (l-exp(-D x/D o)) n / n , = (D x/D o) n / n' l n [ l - ( D x / D o ) n / n ' = -(D x/D o) n / n' 0ER(as D approaches 0) = -(D'/D )*(-D /D ) n / n ' OER (as D -•<».= (D'/D n / n') * (D IC*/1*')-1!) x o o X If n = n' this reduces to the same relationship we derived earlier when n = n': 0ER(as D 0, n=n') = D'/D . x o o If n is greater than n1 the exponent for is positive and therefore as D x approaches zero the OER approaches zero: 0ER(as D x -*• 0, n>n') •*• 0 If n is less than n' the exponent for D is negative and the OER approaches infinity as the dose approaches zero: 0ER(as D * 0, n<n') +' « x In the limit as the dose approaches infinity the OER relationship that we obtained with no approximations (equation 1) is undefined. We therefore start again by equating the two survival equations: S o=l-(l-exp(-D x/D o)) n = S n=l-(l-exp(-D y/D;)) n' (l-exp(-D x/D o)) n - (l-exp(-D y/D (;)) n , 95 Take the In of each side; (n)*ln[l-exp(-D /D )] = (n')*ln[l-exp(-D /D')] x o y o As the dose, D, approaches infinity, exp(-D/DQ) approaches zero, which allows the use of the following approximation; ln[1-x] = -x, if x is very small. Therefore, as D approaches infinity; ln[l-exp(-D/D )] = -exp(-D/D ) o o The above equation becomes; (n)*exp(-Dx/DQ) = (n')*exp(-Dy/D^) Next, the ln is taken of both sides, and the resulting equation is solved for D : y ln[n] - D /D = ln[n'] - D /D' x o y o D = D'(ln[n'] - ln[n] + D /D ) y o x o D /D = (D*/D )(ln[n'] - ln[n] + D /D ) y x o x x o OER = (D'/D )*ln[n'] - (D'/D )*ln[n] + D'/D o x o x o o As D x approaches infinity, "WDx goes to zero so that the above equation reduces to: OER(as D •*• ») = D' / D o o This relationship is independent of the values of the extrapolation num-bers. It is immediately obvious that the only condition which leads to a constant OER throughout the dose range is when the extrapolation num-bers are equal. The relationships we have just derived mathematically are demonstrated graphically in Figure 24. This is a plot of OER vs dose using the OER equation for the SH-MT model that has just been derived (no approximations). As can be seen from this figure the OER is only constant if the extrapolation numbers are the same. The OER is higher 96 D O S E (Gray ) IN 0 2 F igure 24. OER AS A F U N C T I O N O F DOSE FOR T H E S H - M T MODEL. T h e f igure demonstrates the OER as a function of dose for three d i f ferent hypothetical condit ions: a) where n < n ' , b) n = n' and c) n > n 1 . n is the extrapolation number for cells i r rad iated under aerobic conditions and n' the equivalent for i rradiat ion under hypox ic condit ions. For the case a) the ratio of n :n ' = 0.83, for the case c) the ratio of n :n ' = 2.4. It is ev ident that the only condit ion that leads to an O E R independent of dose is when n = n 1 . 97 at low doses if n is less than n1, and lower at low doses if n is greater than n'. Therefore, when assuming a SH-MT model, if investigators report a reduced extrapolation number for cells irradiated in hypoxia (n1 less than n) they are in fact reporting a reduced OER at lower doses. An extrapolation number of one indicates a straight exponential survival curve with no shoulder. This implies no repair of sublethal damage (Quintiliani 1979, Littbrand S Revesz 1969). Likewise a reduction of n implies a reduction in the amount of recovery from sublethal damage. Therefore, if n does not equal n1, this indicates that the repair capability is different for cells irradiated under oxic and hypoxic conditions, which leads to an OER which is not constant with dose. A reduced extrap-olation number in hypoxia has been reported by several investigators (e.g. Belli et al 1967, Elkind et al 1968, Hall et al 1966, Humphrey et al 1963, Pettersen et al 1973, Revesz & Littbrand 1964, Millar et al 1978, Nias et al 1973). If the extrapolation numbers are the same, the repair capabilities are similar and a constant OER can be predicted. As discussed in the introduction (pages 11 and 12) the original interpretation of n was that it represented the number of targets necessary to 'hit' in order to inactivate a cell. According to this interpretation if the number of targets is different in the absence and presence of oxygen, an OER which is not constant with dose would be expected. 4.2.2.2 RMR model The OER can also be examined from the point of view of Tobias's RMR model (Tobias et al 1979). The procedure is the same; we thus equate the two survival equations for hypoxic and aerobic conditions and determine the ratio of the doses. In the following equations a represents 98 the yield of uncommitted lesions per rad, e is the repair ratio (of linear and quadratic repair), and <(> represents the probability that linear repair is successful (eurepair). The subscripts and refer to the aerobic ^ o n (oxic) and hypoxic (in nitrogen) conditions respectively. D x is the dose in oxygen at the prescribed survival level ( S Q ) and D is the corre-sponding dose in nitrogen at the same survival level ( S N ) . Let us first consider what the RMR yields if the only effect of oxygen is to change the initial number of lesions, everything else (re-pair, type of damage, etc.) is the same when the cells are irradiated under hypoxic and aerobic conditions. In this case a / a o n Equating the survival levels for hypoxic and aerobic cells we find: S = exp(-ct D ) * (1 + a D / E ) e * = o o x o x S = exp(-a D ) * (1 + a D /e) e < f > n n y n y The next step is to take the In of each side; -a D + (e6) * l n [ l + a D /e] = o x o x -a D + (ed,) * l n [ l + a D /e] (2) n y n y As the dose, D, approaches infinity; ctD » e<f>ln[l + aD/e] so that the above equation reduces to a D = a D o x n y OER = D /D = a /a y x o n OER (as D - * - » a ^ a ) s a / a o n o n To determine the limit at very low doses we return to equation 2, before any approximations were made, and expand the In term, taking only the first term in the expansion. This approximation is valid for small-doses. The expansion for the In term is: In x = 2[ (x-l)/(x+l) + ((x-l)/(x+l)) V3 + . . . ] , x > 0 99 - a D + < J . ( 2 a D / ( 2 + aD/e)) = o x o x o x - a D + (j)(2a D / ( 2 + a D /e)) n y n y n y O E R = D /D = [-a + 2<j>a / (2 + a D /e ) ] / y x o o o x [-a +•2<fra / ( 2 + a D / e ) ] n n n y OER = [(-2e<fra - D i>a2 + 2c$2a )*(2e«|> + D <ba )]/ o x o o y n [(-2e<(1a - D <j>a2 + 2e<t2a ) * (2e6 + D Act ) ] n y n n x o As the dose approaches zero, each term in the above expression with D x or goes to zero, and the OER becomes: OER=[ (-2eAcx + 2e<f>2ct ) * (2e<j>) ] / [ (-2e<t>a + 2e$*a )*(2e<}))] o o n n = [a (•-l)]/[o (<|>-1)] o n OER (as D + 0 , o ^ a ) s a / a o n o n This is the same result we obtained in the limit at very high doses. Therefore, if the only effect of the presence of oxygen during irradiation is to change the initial number of lesions (the repair and the type of damage are the same) then the RMR model predicts a constant OER. There is experimental evidence to suggest that the presence of oxygen affects more than just the initial amount of damage; this will be discussed shortly. We can also examine the predictions of the RMR model if repair is different in the absence and presence of oxygen. The assumption is made that the probability of misrepair occurring in the linear repair process is not the same t ^r)' ^ " n e r e ' s a m P ' e experimental evidence to suggest this, particularly at the molecular level (e.g. Roots & Smith 1975, Koch & Painter 1975, Skov et al 1982, Berger 1982). Damage inflicted under hypoxic conditions does not appear to repair as well as damage inflicted under aerobic conditions (d> < 4 ). The mathematics that follows is similar n o to the previous example. 100 S = exp(-a D ) * (1 + a D /e) e<*>o o o x o x = S = exp(-a D ) * (1 + a D /e) £ < t >n n n y n y -a D + (ed. ) * l n [ l + a D /e] o x o o x = -a D + (ed) ) * lf i . f l + a D /e] n y n ... n y as D approaches infinity aD >> e<j>ln [1+aD/e] and thus the equation above reduces to; a D = a D o x n y OER = D /D = a /a y x o n OER (as D •-»•«», A f <j> ) = a /a o n o n To determine the OER at very low doses we proceed as before, expanding the In term and taking only the first term -a D + <f> (2a D /(2 + a D /e)) o x To o x o x = -a D + d> (2a D /(2 + a D /e)) n y n n y n y OER = D /D = [-a + 2d. a /(2 + a D /e) ] / y x o o o o x [-a + 2<j> a /(2 + a D /e) ] n n n n y OER = f(-2ea - D a 2 + 2ed> a )*(2e + D a ) ] / o x o o o y n [ (-2ea - D a 2 + 2ed> a )*(2e + D a )] n y n n n x o As the dose approaches zero, each term in the above expression with D x or D aoes to zero, and the OER becomes: y -OER = [ (-2ea + 2ed> a ) * (2e) ] [ (-2ea + 2ed> a ) * (2e) ] o o o n n n OER(as D -»- 0, (f> ? <(, ) = [a (<j> - l ) ] / [ a (()> -1)] o n o o n n Let us now compare the OER values at high and low doses, keeping in mind that d> is less than 1. If d> • o Y n a /a > a (d> - l ) / a (<j> -1) o n o o n n < which implies that the OER is higher at low doses. 101 If <j) > A then the reverse is true Yo n a /a < a ( A - l ) / a (A -1) o n o o n n resulting in a lower OER at low closes. If A q = <j>n then the OER is con-stant. Therefore, the RMR model predicts an OER which is not constant for all doses if the repair of initial lesions is different under aerobic and hypoxic conditions. This is demonstrated graphically in figure 25. There are numerous reports in the literature that demonstrate this difference in repair (see page 99). 102 rr U J O < or ui UJ O 2 < I 2 HI 2 ' U J o >-X o 3 V-2 H 1 I-0. 1 10 100 D O S E (Gray ) IN O, F igure 25. OER AS A F U N C T I O N OF DOSE FOR T H E RMR MODEL. T h e f igure demonstrates the effect of the repa ir of the damage inf l icted under aerobic or hypox ic condit ions on the OER . If there is a d i f ference between the repairs of the respective damages then the OER is not constant with dose. T h e r e is experimental ev idence that shows a d i f ference in repair (see tex t ) . T h e only s ituation that leads to an O E R which is constant with dose, shown by the dotted l ine, is when the repair is the same regardless of the condit ions under which the damage was in f l i c ted. The dotted line also represents the situation d i scussed in the text when the only d i f ference between irradiat ion under aerobic and hypox ic condit ions is the initial number of lesions. T h e 1 following parameter values were used (all a values are in units of r a d " ): a) for the A = <f> c u r v e ; e = 30, A = A =0.93, a = 0.016, a = 0.005, b) for the A° < c u r v e ; e = 30, A° = o"86, A = 0°93. a = 0.^15, a = 0.010, c) foP the A > A c u r v e ; e = 30, A = (T.93, A = 0°.86, a = o /o i5 , a = 0.005. 0 n ° 103 1.2.2.3 L-Q model The expression for OER using the L-Q model can be determined by following the same procedure used in the previous examples. In the following equations a Q and B Q are the L-Q parameters for the aerobic survival curve, ctn and B p are the L-Q parameters for the hypoxic sur-vival curve. is the dose in oxygen at the prescribed survival level (S ) and D is the dose in nitrogen at the same survival level (S ). o y 3 n S = exp(-a D -6 D 2) = S = exp(-a.D -B D 2) o o x o x n n y n y a D + 6 D 2 = a D + 6 D 2 o x o x n y n y a + 6D = a (D /D ) + 8 (D 2/D 2)D o o x n y x n y x x (note: OER = D /D ) y x 8 D (OER) 2 + a (OER) - (a + 8 D ) = 0 n x n o o x This can be solved using the solution to the quadratic equation: OER ='[-a + ( a 2 + 43 D (a + 6 D ))*]/2B D n n n x o o x n x This expression is used to draw the OER curves shown in figure 26. To discuss the OER at very low and very high doses a much simpler expression is useful. To arrive at these expressions we return to the 2nd step in the above set of equations: a D + 8 D 2 = a D + g D 2 o x o x n y n y D ( a + 8 D ) = D (a + g D ) x o o x y n n y OER = D / D = ( a + g D ) / ( c t + B D ) y x o o x n n y In the limit at very low doses; as D approaches 0, a >> BD, and this leads to; OER (as D + 0) = a /a o n 104 In the limit at very high closes; as D approaches infinity, a << BD, and this leads to; OER = D /D = g D /B D y x o x n y (D /D ) 2 = B /B y x o n OER (as D -*• ») £ (B /B )* o n • , The OER as a function of dose is shown in figure 26. If oxygen is dose modifying (constant OER with dose) the following relationship is found; a /a = (B /B )* o n o n This, implies that the two parameters are not independent (if OER is constant). They are restricted to values that satisfy the above equation. This has never been experimentally verified; on the contrary there is ample experimental evidence to show that the two parameters are in-dependent. Therefore, on the basis of the L-Q model and experimental evidence a constant OER would not be predicted. The L-Q model has become widely used by several investigators (e.g. Wideroe 1970, Chapman 1980, Douglas 1982a, Kellerer S Rossi 1972, Chadwick & Leenhouts 1973). A growing volume of experimental evidence (Chapman 1980, Wideroe 1966) suggests that mammalian cells are killed by ionizing radiation through at least two independent mechanisms. The L-Q model provides a mathematical basis to explain two mecha-nisms. The alpha effect is linearly proportional to dose. The beta effect is proportional to the dose squared. The two effects carry equal weight at a dose equal to a/B. The a effect predominates at low doses and the B effect is responsible for most of the cell kill at high doses. 105 rr LU O w 4 < rr H 2 UJ Z> 111 O 2 < I 2 UJ 2 UJ O >-X o Linear Q u a d r a t i c Model 0.1 10 1 0 0 DOSE (Gray) IN CL Figure 26. OER AS A FUNCTION OF DOSE FOR THE L-Q MODEL. The figure demonstrates the effect on OER of the relative magnitudes of the a and /& ratios obtained for cells irradiated in the presence and absence of oxygen. The only situation which leads to an OER which is constant with dose, shown by the dotted line, is when the two parameters are not independent. This has never been observed experimentally (see text)._The following parameter values were used (all a values are in units of Gy , all & values are in units of Gy 2 ) : a) for the case where a la = /B //B ; a = 0.96, B = 0.307, a = 0.30, 8 = 0.03, b) for the case" where a"/a °< /B //B ; a = 0.30, B = 0.27, 8 = 0.20, 6 = 0.03, c) for the Case where" a 7a £ /B //B ; a° = 0.30. B°= 0.1875, a = 0.0652, „ „„„ o n n n n r> n B n = 0.030. 106 Single-hit kill (alpha effect) results from energy dense events (Wideroe (1977) gives the threshold as 12 keV/um) which show little or no repair. In a typical X-ray or gamma ray spectrum these energy dense events come from the high LET component of the low LET radiation. These high LET events could be due to Bragg peaks (Porter 1965) or electron track ends (Tobleman & Cole 1974, Cole et al 1974). At low doses : this high LET component of the radiation would cause a greater propor-tion of the lethal damage than at higher doses. As the LET of the ra-diation increases, the OER tends to unity and the single event component of cell inactivation increases (Barendsen et al 1963, Kellerer & Rossi 1971). Chapman (1980) therefore predicts that the single event kill would exhibit an OER of 1.6-1.7. Multiple-hit cell kill (beta effect) can be described as resulting from at least two independent energy events. It shows complete repair of sublethal damage under appropriate conditions and exhibits an OER of 3.0-3.5 (Chapman 1980). This difference in OER for the alpha and beta effects implies that oxygen operates mainly by en-hancing the production of double events; it mainly affects the & in the L-Q model (Chapman et al 1975a, 1975c). This measurement of a differential oxygen effect for single and double events would lead to a lower OER at low doses. The exact amount of influence that oxygen has on inactivation by the single-hit component of radiation (the alpha effect) is a controversial subject (Wideroe 1970, Revesz and Littbrand 1967, Barendsen et al 1966). The relatively large uncertainty due to the scatter of the survival mea-surements, as well as poor statistics, different cell lines, etc., make direct comparisons difficult. 107 Koch (1979) has consistently stated that he believes oxygen to be dose-modifying (constant OER) and that any evidence to the contrary is due to artifacts from the experimental technique. He presents data in air, argon, and nitrogen and shows that the hypoxic survival curves (argon and nitrogen) can be superimposed onto oxic (air) survival curves if each nitrogen dose is divided by 3. The oxic curves start at 2 Gy (6 Gy for the hypoxic data). The dose range used in the oxic survival curves obtained for this thesis (and showing an oxygen effect which is not dose-modifying) is 0 - 2 Gy. Unfortunately these two sets of data cannot be directly compared, since they are in different dose ranges. However Koch does fit his data to the L-Q model and gives the values for a and (3 (see Table IX). As discussed earlier, the alpha and beta values can be used to determine the OER values at very low and very high doses. If we do this with Koch's values we see that in two of the three sets of data (early and late S) we obtain a lower OER at low doses (see Table X). For the mitotic cells the ratios are again not equal, indicating that the OER is not constant throughout the dose range. TABLE IX L-Q PARAMETERS FOR KOCH'S 1979 DATA cells a a o -1 -1 n n X 10 gray ;-2 -2 gray mitotic early S late S 2.2 0.68 1.1 0.58 0.28 0.46 3.0 2.8 0.28 0.41 0.26 0.028 108 TABLE X Parameter Ratios for data i n Table IX c e l l s a /a (6/6 )* o n o n m i t o t i c 3.79 2.71 early S 2.43 3.28 lat e S 2.39 3.16 4.2.2.4 The Importance of S u r v i v a l Models at low doses A n example of how d i f f i c u l t i t is ( i f not imposs ible) to p red i c t the OER at low doses from the data at h i gher doses can be demonstrated by d i s cu s s i n g the work of M i l l a r et al (1978). The author s d i s cu s s the i n t e r -p re ta t ion of cel l s u r v i v a l data (at h igh doses) u s ing 3 d i f f e r en t e q u a -t i on s ; the s i n g l e - h i t mu l t i - t a r ge t (2 parameter S H - M T ) , the mu l t i - t a r ge t w i th in i t i a l s lope (3 parameter S H - M T ) , and the l i n ea r - quad r a t i c ( L -Q ) model. These th ree equat ions are shown in Tab le I (page 11). The au tho r s state that " . . . . a l l th ree models f i t the data ext remely well and we cannot inva l idate one of the models with the data p re sented h e r e . " While the models may be i nd i s t i ngu i shab le on the bas i s of t he i r f i t s to the da t a , the p red i c t i on s at low doses , wh ich we can determine from the g i v en parameters , a re qu i te d i f f e r e n t . Us ing the parameters fo r each of the models (wh ich the author s p r o v i d e , from the h igh dose data) we can determine the OER as a f unc t i on of dose. T h i s is shown in f i g u re 27. A t h i gh doses the OER is e xac t l y the same for all t h ree models (which is e x p e c t e d ) . However , at low doses there is cons ide rab le d i f f e r ence . T h i s v i v i d l y demonstrates that s u r v i v a l models wh ich are i nd i s t i ngu i shab le at h igh doses can y i e l d v a s t l y d i f f e r en t p red i c t i on s in the low dose reg i on . 109 Figure 27. OER AS A FUNCTION OF DOSE FOR 3 DIFFERENT SURVIVAL MODELS. The sets of parameters for these curves were taken from a paper by Millar et al 1978. Although the authors concluded that "...all three models fit the data extremely wel|..." the figure demonstrates that the predi-ctions at low doses of the three models (which were indistinguishable at high doses) are vastly different. 110 An experimenter is introducing a very large bias to the interpre-tation at low doses when a survival model is chosen. It is important to stress that the bias at low doses is introduced despite the fact that at high doses there is no bias; the models all predict the same effects. Since the majority of cell survival data is obtained at high doses extreme care must be taken when extrapolating the results to lower doses. 4.2.3 Predictions for OER from the mechanistic viewpoint If only the amount of damage changes due to the presence of oxy-gen, but the type of damage and ability to repair remain the same, then a constant OER with dose is expected (Koch 1979, Winston et al 1975). In this interpretation oxygen is said to be dose modifying since, if the doses for the hypoxic points are all multiplied by the same factor, the two survival curves can be superimposed. This implies that the shapes of the survival curves do not change due to the presence of oxygen. If the type of damage, or the repair capability, is different in oxygen, then no relationship would necessarily exist between the two survival curves (Koch 1979). In this case the OER could be dose dependent. There have been numerous reports in the literature (e.g. Belli et al 1967, Littbrand & Revesz 1969, Nias et al 1973, Siracka et al 1975, Hall 1972) that show a decrease in the ability of hypoxic cells to repair suble-thal damage. If the repair capability of hypoxic cells is different from aerobic cells, then an OER which is not constant with dose would be expected. This was predicted by both the SH-MT and RMR models. 111 Koch (1979) lists 3 radiochemical reasons why the damage in oxygen should be different from that in nitrogen. These are: (1) Many more organic radicals will be oxidized by 0 2 in the presence of oxygen. Thus it is almost impossible that the type of radiation damage, at the molecular level, will be identi'-cal in the presence or absence of oxygen, unless the radicals are very long-lived and form adducts at the 'end of the experi-ment' when the cells are invariably put back into an aerobic environment. (2) The population of initial radicals and reactive species formed after aerobic radiation will be different than after hy--='" poxic radiation (e.g. the yield of 0 2" and H 2 0 2 ) . 3) Autooxidations (of unsaturated lipids) may be consid-erably reduced if oxygen is not present immediately after irradiation. By his own reasoning, this implies an OER which is not constant with dose. There is considerable experimental evidence at the molecular level, from our laboratory and others, that show that the initial number of DNA breaks is greater in the presence of oxygen than in hypoxia (Palcic and Skarsgard 1972, Hohman et al 1976, Skov et al 1979). The repair of this initial damage to DNA, inflicted under oxic and hypoxic conditions, in-volves two distinct kinetics, a fast and a slow component of repair. In oxygen, a larger percentage of the damage is repairable by the fast component (Berger 1982 , Skov et al 1982). The rates of the fast and slow components of repair appears to be the same for both oxic and hypoxic conditions. There is also experimental evidence which indicates that the chemical nature of the breaks is different in the absence and presence of oxygen. It appears that the contribution of hydroxyl radicals to damage assayed by cell survival techniques is higher in cells irradiated in air (oxic conditions) than in hypoxia (Chapman et al 1973). This is in agreement with molecular results that show that the contribution of hydroxyl radicals 112 to DNA damage is greater in cells irradiated in air than in hypoxia (Chapman et al 1975a, Skov et al 1981, Roots et al 1982 , Skov 1 983). Revesz and colleagues (Deschavanne et al 1981, Revesz 1981, Midander et al 1982) have also shown that there are differences in both the radiation induced DNA damage and in the repair mechanisms for cells irradiated in aerobic and anoxic conditions. This evidence supports the theory of a lower OER at lower doses. As discussed earlier, if only the initial amount of damage (in this case the number of DNA breaks) is different, then a constant OER with dose would be expected. However, if there is a differ-ence in the repair or in the type of damage then there would not neces-sarily be a simple relationship between the survival curves. An OER which is dose dependent would then be a possibility. Littbrand & Revesz (1969) claim that a dose dependent OER is a result of the fact that oxygen can both protect and sensitize cells to the lethal effects of radiation. They claim that oxygen can protect cells by permitting the oxidative metabolism which probably provides the energy reguired by the repair processes. At the same time oxygen sensitizes the cell by enhancing the radiation damage. Thus they claim that oxygen may play a dual role in regard to radiation survival and whether it pro-tects or sensitizes is dependent on the radiation dose. At low doses an OER value below unity or around unity would be expected because the contribution to recovery is much larger than the enhancement contribu-tion. At high doses the contribution of oxygen to the recovery processes is proportionally much less than the enhancement of radiation damage and an OER of 3 would be expected. There is no experimental evidence that clearly demonstrates an OER of less than unity, but neither can the possibility be completely negated. 113 The large amount of scatter in survival measurements and the lack of accurate data at very low doses prevents the resolution of this debate. 4.2.4 Asynchronous Data The data presented in this thesis for asynchronous CHO cells clearly show a lower OER at low doses. This is in agreement with previous molecular work from this laboratory which demonstrated a lower OER at low doses (Palcic et al 1978). The OER values, as described in the caption of figure 14, were computed with no assumption about the shape of the cell survival curve. Figure 28 shows the same OER values as figure 14. In addition, the OER curves generated from 3 sets of alpha/beta values are shown. These alpha/beta values are shown in Table XI. The first two sets represent the alpha/beta values obtained when the asynchronous data is fitted to the L-Q model using the method of least squares, using only the low dose data, then only the high dose data. The survival curves with these values were shown in figures 12 and 13. The third set of alpha/beta values represent the best alpha/beta values ob-tained when using an iterative process that utilizes both data sets and the alpha value from low doses. In addition the ratios of the alpha/beta values, which represent the OER values at very low and very high doses, are shown in Table XII for all 3 sets. In each case the OER is lower at lower doses. It is interesting that from Figure 12, which shows the asyn-chronous survival curves in nitrogen and oxygen at high doses, the OER appears constant. However if we examine the alpha/beta values that we obtain for the best fits of these data, they lead to a lower OER at lower doses, i.e. 2.02 vs 3.26 at higher doses. This can also be seen in figure 28. This figure demonstrates that the OER values generated from the alpha/beta values obtained using an iterative process and both data 11<J D O S E (G ray ) in 0 2 Figure 28. OER VALUES AND THE L-Q EQUATION. The OER values, represented by solid circles, are the same as those shown in figure 11. The a and B values used are shown in Table XI. The dotted line represents the values obtained from the data at low doses. The dashed line represents the values obtained from the data at high doses. The solid line, which gives the best fit through the total dose range, represents the values obtained using both data sets and an itera-tive procedure. 115 sets closely match the experimental OER values. The other two sets of L-Q parameters each give a good fit only in the dose range corresponding to the actual data points. TABLE XI L-Q PARAMETERS FOR ASYNCHRONOUS CHO CELLS x I O " 1 x i o - 2 Gy" 2 a o a n e o S n from low doses (0 - 3 Gy) .944 .585 4. 96 0. 991 from high doses (3 - 30 Gy) 1.69 .835 2. 85 0. 268 from combination 1 .21 . 757 3 . 33 0. 300 TABLE XII ASYNCHRONOUS CHO CELLS RATIOS OF PARAMETERS IN TABLE XI a /a o n /B //B o n from low doses (0 - 3 Gy) 1 .59 2. 25 from high doses (3 - 30 Gy) 2 .02 3 . 2 6 from combination 1 .60 3 .33 We found no evidence in our data of an OER less than unity at very low doses. This does not prove that an OER less than one is impossible. Our lowest measurement was at 0.35 grays in oxic conditions, and, of course, the error is larger at smaller doses. Figure 29 shows attempts to fit the SH-MT model to the OER values obtained with the asynchronous data. No satisfactory fit was obtained. This is in accordance with many other investigators who have stated 116 cc UJ o < cc UJ UJ o < z UJ UJ (D > X o 1 1 — I — I I I I I I A s y n c h r o n o u s C H O c e l l s S H - M T M o d e l • „ • = i a 0 r « d • 1 1 1 1 I I I I T 1 1 1 I I I I \y y 0= * S 0 r a d • ' ' I 1 1 I I I ' ' ' I l l l l ' ' I l l l l 0.1 10 1 0 0 DOSE (Gray) in O, Figure 29. OER VALUES AND THE SH-MT EQUATION. The OER values, represented by solid circles, are the same as those shown in figure 14. The four curves represent attempts to fit the SH-MT model to the OER data. The values for the parameters are shown on the curves. No satisfactory fit could be obtained. 117 that the 2 parameter SH-MT model does not fit low dose data satisfactorily (Koch 1979, Chapman 1980). The closest approximation to the data in-volves a reduced extrapolation number for hypoxic cells. This implies that the repair capabilities of the cells are affected by the presence of oxy-gen. This has been well documented by many workers, as discussed earlier. It is -important to emphasize that the excellent fit we obtained with the L-Q model does not imply that this model is necessarily valid. The first approximation, at low doses, of any survival model results in the linear-quadratic equation, as discussed in the introduction (page 13). The excellent fit we obtained implies that the L-Q model is a satisfactory description of the data. The lower OER at lower doses that we obtained is not dependent on any particular survival model. A lower OER at lower doses implies that either the type of damage or the repair capabilities of the cells (or both) are different in aerobic cells from those in hypoxic cells. It also lends support to the theory that the initial region of the cell survival curve is due to single hit killing which involves high LET events. A differential effect of oxygen on the alpha and beta components of cell kill is also supported (but not proven) by our result of a lower OER at lower doses. 118 4.2.5 Synchronous Cells 4.2.5.1 .Limitations of the Data It is evident from figures 20 and 23 that the scatter in the data is greater than one would like, making it difficult to draw definite con-clusions. This is mainly attributable to insufficient data and also to problems with obtaining and maintaining synchronized cell populations (see section 4.1.5.2 for further discussion). Despite the fact that there are insufficient data for a thorough analysis, these data may be representative of the results obtainable with many more experiments. Therefore, it is worthwhile to see what these results suggest, bearing in mind that these are only tentative conclusions subject to verification when more experiments are completed. It is with this understanding that the following 'conclusions' are suggested. 4.2.5.2 OER in synchronous cells For the cells in late S a lower OER at lower doses is clearly demon-strable. However, for the cells in C1/S the mean OER at low doses is lower than at high doses, but within the experimental error, the values could be the same. The greater difference between the OER at low and high doses for the cells in Gl/S and late S may be partly explainable by the different radiosensitivities of the cells. Cells irradiated in Gl/S and plated, immedi-ately duplicate their DNA. This allows no time (or very little time) to repair any of the damage inflicted at the time of irradiation before it is 'fixed' by duplication. This is the usual explanation for the radiosensi-tivity of cells in G1/S. On the other hand, cells in late S have already duplicated their DNA, and a complete cell cycle must be traversed before duplication is again performed (and the damage is 'fixed'). Cells in late 119 S therefore have a great deal of time to repair any damage inflicted at the time of irradiation. A lower OER at lower doses would imply two things; 1) that the type of damage inflicted when the cells were irradiated under aerobic or hy-poxic conditions is different and/or 2) that the repair of the damage inflicted when the cells were irradiated under aerobic or hypoxic con-ditions is different. Since cells in Gl/S are radiosensitive these differ-ences may be masked by the overwhelming sensitivity of the cells to any damage. If the first situation is the case, for example the type of damage inflicted under oxic conditions is more lethal, this may not be evident for the cells in C1/S because there is no time to repair the 'more repairable' type of damage inflicted in the absence of oxygen. For the same reason any difference in the repairability of the damage produced in hypoxic or aerobic conditions (the second case) would not be as evident for the cells in G1/S because there is no time for repair of either damage. So a lower OER at low doses would not be very evident in radiosensitive cells which have little time to repair. Cells in late S, which are radioresistant, and therefore more able to repair damage, would be more affected by the difference in the damage inflicted under aerobic and hypoxic conditions. So a lower OER at lower doses is more evident for the cells in late S. 4.2.6 Oxygen Concentration It has been pointed out (Koch 1975) that a decreased OER at low doses may be the result of the presence of small amounts of oxygen at low doses in the hypoxic samples, which is then depleted by radiation, yielding an apparent increase in OER at higher doses. At the present time we can not totally exclude this possibility as we can not measure the exact 0 2 concentration in our samples. Nevertheless, the explanation 120 that the observed OER at lower doses is solely due to some traces of oxygen at low doses is unlikely. For CHO cells, one obtains full OER (at high doses) when the oxygen concentration approaches 0.1 - 0.2 uM in solution (e.g. Millar et al 1979, Whillans & Hunt 1982). If the observed OER of 1.5 at 0.5 Cy is solely due to traces of oxygen present, the oxygen concentration would have to be approximately 3 uM. Whillans and Rauth (1980) showed that under identical conditions: Type II vessels, gassing purified M2 of less than 9 ppm (in our case less than 5 ppm) for over 45 minutes, the oxygen concentration in solutions is less than 0.2 uM. Thus, we should be getting a full oxygen effect even at these low doses. 121 5. CONCLUSIONS We have developed a technique to assay cell survival at low doses. The 'classical' technique/ used at high doses, measures (macroscopically) only the number of surviving cells. The largest error in the high dose assay comes from the uncertainty in the number of cells plated, which is of the order of 10 to 15%. In our low dose assay the exact number of plated cells is determined microscopically. After incubation, the number of killed cells and the number of surviving cells are both determined, by microscopic examination. While extremely labour intensive, this low dose assay yields survival data, in the low dose region, which is much more accurate than the data obtained using classical methods. This technique can be used to measure many radiobiological parame-ters. We have chosen to examine the effect of oxygen at low doses. Our results clearly demonstrate that, for asynchronous CHO cells irradiated under the experimental conditions described in this thesis, the radio-sensitizing effect of oxygen (OER) is reduced at lower doses. It has been suggested that a decreased OER at low doses may be the result of the presence of small amounts of oxygen in our hypoxic samples. Recent work in our laboratory indicates that the experimental conditions employed lead to a full oxygen effect. This technique can be used to complement the classical 'high dose assay' to obtain data that encompasses a large dose range. This will be especially valuable, for example, when attempting to fully describe radio-biological parameters, or when examining survival models. 122 6. BIBLIOGRAPHY Alper,T., Gillies,N.E. & Elkind.M.M. (1960) The sigmoid survival curve in radiobiology. Nature, 186, 1062-1063. Alper, T. (1975) Cell, Survival after Low Doses of Radiation: Theoretical and Clinical  Implications, Proceedings of the Sixth L H Gray Conference, held at Bedford College, London, 16-21 September 1974., T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. Alper,T. (1979) Cellular Radiobiology. Cambridge University Press, Cambridge. Barendsen,G.W. , Walter, H.M. D., Fowler,J.F. & Bewley,D.K. (1963) Effects of Different Ionizing Radiations on Human Cells in Tissue Culture. Radiat. Res., 1_8, 106-119. Barendsen,C.W., Koot.C.J., van Kersen,C.R., Bewley,D.K., Field,S.B. & Parnell,C.J. (1965) The effect of oxygen on the impairment of the proliferate capacity of human cells in culture by ionizing radiation of different LET. Int. J.  Radiat. Biol., 10, 317-327. Bedford,J.S. & Griggs,H.G. (1975) The estimation of survival at low doses and the limits of resolution of the single-cell-plating technigue. In; Cell Survival after Low Doses  of Radiation: Theoretical and Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. 34-39. Belli,J.A., Discus,G.J. & Bonte,F.J. (1967) Radiation Response of Mammalian Tumor Cells. I. Repair of Sublethal Damage jn Vivo. J. Natl. Cancer Inst. , 38, 673-682. Bender,M.A. (1960) X-ray-Induced Chromosome Aberrations in Mammalian Cells |n vivo and in vitro, in Immediate and Low Level Effects of Ionizing Ra-diations; Proc. Symp. held at Venice on June 22-26, 1959; under the joint sponsorship of U. N. E. S. CO., I.A.E.A. and C.N.R.N. and published as a supplement to the Int. J. Radiat. Biol. , 103-118. Bender,M.A. & Gooch, P.C. (1962) The kinetics of X-ray survival of mammalian cells jn vitro. Int. J.  Radiat. Biol., 5, 133-145. Bender,M.A. (1969) Human radiation cytogenetics. In: Adv. in Radiat. Biol., 3, L.G. Augenstein, R. Mason & M. Zeile, eds. Academic Press, New York. 215-275. Berger,S. (1982) Repair of DNA damage in mammalian cells after low doses of ionizing radiation. M.Sc. Thesis, Univ. of British Columbia. 123 Blakely,E.A., Ngo,F.Q. H . , Chang,P.Y., Lommel, L. , Kraft-Weyrather ,W. , Kraft,G., & Tobias,C.A. (1980) Heavy Ion Cell-Cycle Response and Progression Effects. In: Biological and Medical Research with Accelerated Heavy Ions at the  Bevalac, 1977-1980., M. C. Pirruccello & C.A.Tobias, eds. LBL-1 1220, UC-48. 125-135. BrewenJ.G. & Luippold,H.E. (1971) Radiation induced human chromosome aberrations: |n vitro dose rate studies. Mut. Res., V2, 305-314. ~~ • Bryant,P.E. & Lansley,!. (1975) Survival-curve shape and oxygen enchancement ratio at low doses for Chlamydomonas. In; Cell Survival after Low Doses of Radiation:  Theoretical and Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. 107-114. Chadwick,K.H. & Leenhouts,H. P. (1973) A molecular theory of cell survival. Physics Med. Biol., 18, 78-87. Chapman,J.D., Reuvers, A. P. , Borsa,J. & Greenstock,C. L. (1973) Chemical Radioprotection and Radiosensitization of Mammalian Cells Growing [n Vitro. Radiat. Res., 56, 291-306. Chapman, J. D., Dugle,D.L., Reuvers,A. P., Gillespie,C. J . & Borsa,J. (1975a) Chemical radiosensitization studies with mammalian cells growing in  vitro. (Proceedings of the 5th Int. Congress of Radiation Research, Seattle) In: Radiation Research Biomedical, Chemical and Physical  Perspectives. O.F.Nygaard, H.I.Adler, W.K.Sinclair, eds. Academic Press, Inc., New York. 752-760. Chapman,J.D., Gillespie, C.J., Reuvers, A.P., & Dugle, D.L. (1975b) Radioprotectors, radiosensitizers, and the shape of the mammalian cell survival curve. In; Cell Survival after Low Doses of Radiation:  Theoretical and Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. 135-140. Chapman ,J. D., Gillespie, C.J., Reuvers, A.P., £ Dugle, D.L. (1975c) The inactivation of Chinese Hamster Cells by X-Rays: The effects of Chemical Modifiers on Single- and Double- Events. Radiat. Res., 64, 365-375. Chapman,J.D. (1980) Biophysical Models of Mammalian Cell Inactivation by Radiation. In: Radiation Biology in Cancer Research, R.E. Meyn & H.R. Withers, eds. Raven Press, New York. 21-32. Chu,E.H.Y., Giles,N.H., Passano,K. (1961) Types and Frequencies of Human Chromosome Aberrations Induced by X-rays., Proc. Natl. Acad. Sci. (U.S.), 47, 830-839. 124 Cole,A., Cooper,W.G., Shonka,F., Corry,P.M., Humphrey,R.M. & Ansevin,A.J. (1974) DNA scission in hamster cells and isolated nuclei studied by low-voltage electron beam irradiation. Radiat. Res., 60, 1-33. Collyn d'Hooghe,M., Hemon,D., Valleron,A. J . S Malaise,E. P. (1980) Comparative Effects of Ionizing Radiations on Cycle Time and Mitotic Duration. A Time-Lapse Cinematography Study. Radiat. Res., 81 , 384-392. Colombo,C. & Marin,G. (1963) Cytological Analysis of Colonies Developed From Mammalian Cells Irradiated Ln Vitro. Exp. Cell Res. , 29, 268-277. Countryman,P. I. S Heddle, J.A. (1976) The production of micronuclei from chromosome aberration in irradi-ated cultures of human lymphocytes. Mut. Res., 41 , 321-331. Cullen,B.M., Evans,N.T.S., Walker,H.C., Emery, E.W. & BoagJ.W. (1980) Cell survival at low oxygen tension and dose build-up in argon. Int.  J. Radiat. Biol., 37(1 ), 19-32. Damashek,W. & Cunz, F.W. (1957) Diagnostic and Therapeutic X-ray Exposure and Leukemia. J. Am.  Med. Assoc., 163, 838-840. DenekampJ. McNally ,N. J., Fowler,J.F. & Joiner,M.C. (1980) Misonidazole in fractionated radiotherapy; little and often? Br. J.  Radiol. , 53, 981-990. Denekamp, J. (1982) Cell Kinetics and Cancer Therapy. Charles C. Thomas Publisher, Springfield, Illinois. Deschavanne,P.J., Midander,J., Edgren,M., Larsson,A., Malaise,E.P. & Revesz,L. (1981) Oxygen enchancement of radiation induced lethality is greatly reduced in glutathione deficient human fibroblasts. Biomedicine, 35, 35-37. Dewey,D.L. (I960] Effect of Oxygen and Nitric Oxide on the Radiosensitivity of Human Cells in Tissue Culture, Nature, 186, 780-782. Douglas, B.G. (1982a) Implications of the quadratic cell survival curve and human skin radiation "tolerance doses" on fractionation and superfractionation dose selection. Int. J. Radiat. Oncol. Biol. Phys. , 8, 1135-1142. Douglas, B.G. (1982b) Superfractionation: Its Rationale and Anticipated Benefits. Int. J.  Radiat. Oncol. Biol. Phys., 8, 1 143-1 153. 125 Dowdy, A. H., Bennett,L. R. , & Chastain ,S.M. (1950) Protective action of anoxic anoxia against total body roentgen irra-diation of mammals. Radiology , 55 , 879-885. Dutreix,J., Wambersie,A. & Bounik,C. (1973) Cellular Recovery in Human Skin Reactions: Applications to Dose Fraction Number Overall Time Relationship in Radiotherapy. Eur. J.  Cancer, 9, 159-167. Elkind,M.M. & Sutton,H. (1959) X-ray Damage and Recovery in Mammalian Cells in Culture. Nature, 184, 1293-1295. Elkind,M.M. & Sutton,H. (1960) Radiation Response of mammalian cells grown in culture. 1 . Repair of X-ray damage of surviving Chinese hamster cells. Radiat. Res., 13, 556-593. Elkind,M.M., Han,A. & Volz,K.W. (1963) Radiation Response of Mammalian Cells Crown in Culture. IV. Dose Dependence of Division Delay and Postirradiation Growth of Surviv-ing and Nonsurviving Chinese Hamster Cells. J. Natl. Cancer Inst., 30, 705-721. Elkind, M.M. & Whitmore, G.F. (1967) The Radiobiology of Cultured Mammalian Cells. Gordon and Breach, New York. Elkind,M.M., Withers,H.R. £ Belli,J.A. (1968) Intracellular Repair and the Oxygen Effect in Radiobiology and Radiotherapy. In; Frontiers of Radiation Therapy, 3_, J.M. Vaeth, ed. S. Karger, Basel. 55-87. Evans,H.J., Neary,G.J. £ Williamson,F.S. (1959) The relative biological efficiency of single doses of fast neutrons and gamma-rays on Vicia faba roots and the effect of oxygen. Part II. Chromosome damage; the production of micronuclei. Int. J. Rad. Biol. , 3, 216-229. Evans,H.J., Buckton,K .E. , Hamilton,G. E. £ Carothers,A. (1979) Radiation-induced chromosome aberrations in nuclear-dockyard work-ers. Nature, 277, 531-534. Evans,N.T.S. , £ Naylor,P. F.D. (1963) The effect of oxygen breathing and radiotherapy upon the tissue oxygen tension of some human tumors. Brit. J. Radiol., 36, 418-425. Fowler,J.F., Sheldon,P.W., DenekampJ. £ Field,S.B. (1976) Optimum fractionation of the C3H mouse mammary carcinoma using X-rays, the hypoxic cell sensitizer Ro-07-0582 or fast neutrons. Int.  J. Radiat. Oncol. Biol. Phys. , 1, 579-592. 126 Franko,A.J. & Sutherland,R.M. (1979a) Oxygen diffusion distance and development of necrosis in multicellu-lar spheroids. Radiat. Res., 79, 439-453. Franko,A.J. & Sutherland , R.M. (1979b) Radiation survival of cells from spheroids grown in different oxygen concentrations. Radiat. Res., 79, 454-467. Fricke,H. & Hart,E.J. (1966) Chemical Dosimetry. In: Radiation Dosimetry, Vol II. Instrumenta- tion , F.H.Attix & W. C. Roesch, eds. Academic Press, New York. 126-239. Froese, C. (1966) Division delay in HeLa cells and Chinese Hamster cells. A time-lapse study. Int. J. Radiat. Biol. , 10(4), 353-367. Froese,G. & Cormack.D.V. (1968) A correlation between division delay and loss of colony-forming ability in Chinese hamster cells irradiated jn vitro. Int. J. Radiat.  Biol. , 14(6), 589-592. Grote,S.J., Joshi'G.P., Revell,S.H. & Shaw,C.A. (1981a) A method for the scrutiny of live mammalian cells in culture and for the measurement of their proliferative ability after x-irradiation. Int.  J. Radiat. Biol., 39(4), 377-394. Grote,S.J., Joshi,G.P., Revell,S.H. & Shaw,C.A. (1981b) Observations of radiation-induced chromosome fragment loss in live mammalian cells in culture, and its effect on colony-forming ability. Int. J. Radiat. Biol. , 39(4), 395-408. Hall,E.J., Bedford,J.S. & Oliver,R. (1966) Extreme hypoxia; its effect on the survival of mammalian cells irradi-ated at high and low dose-rates. Brit. J. Radiol., 3_9, 302-307. Hall,E.J. & CavanaghJ. (1969) The effect of hypoxia on recovery of sublethal radiation damage in Vicia seedlings. Brit. J. Radiol., 42, 270-277. Hall, E.J. (1972) The Effect of Hypoxia on the Repair of Sublethal Radiation Damage in Cultured Mammalian Cells. Radiat. Res., 49, 405-415. HenkJ.M. & Smith,C.W. (1973) Unequivocal clinical evidence for the oxygen effect. Brit. J. Radiol., 46, 146. Hohman,W.F., Palcic,B. & Skarsgard, L. D. (1976) The effect of nitroimidazole and nitroxyl radiosensitizers on the post-irradiation synthesis of DNA. Int. J. Radiat. Biol. , 30, 247-261. 127 Holthusen, H. (1921) Beitrage zur biologie der strahlenwirking. Pfluger's Archiv fur die Gesamte Physiologie 187, 1-24. Hopwood.L.E. & Tolmach,L.J. (1979) Manifestations of damage from ionizing radiation in mammalian cells in the post irradiation generations. In: Adv. in Radiat. Biol. , 8, J.T. Lett £ H. Adler, eds.. Academic Press, New York. 317-362. Humphrey,R.M., Dewey,W.C. & Cork,A. (1963) Effect of Oxygen in Mammalian Cells Sensitized to Radiation by Incorporation of 5-Bromodeoxyuridine into the DNA. Nature, 198, 268-269. Hurwitz,C. & Tolmach,L.J. (1969) Time-lapse cinemicrographic studies of X-irradiated HeLa S3 cells. I. Cell progression and cell disintegration. Biophys. J., 9, 607-633. ICRU Report 14 (1969) Radiation Dosimetry: X Rays and Gamma Rays with Maximum Photon Energies Between 0.6 and 50 MeV. International Commission on  Radiation Units and Measurements Report 14. ICRU Report 16 (1970) Linear Energy Transfer. International Commission on Radiation Units  and Measurements Report 16. ICRU Report 17 (1970) Radiation Dosimetry: X Rays Generated at Potentials of 5 to 150 kV. International Commission on Radiation Units and Measurements Report TT. ; Jenssen, D. & Ramel,C. (1978) Factors Affecting the Induction of Micronuclei at Low Doses of X-rays, MMS and Dimethylnitrosamine in Mouse Erythroblasts. Mut Res, 58, 51-65. Joshi,G.P., Nelson,W.J., Revell,S.H. & Shaw,C.A. (1982a) X-ray-induced chromosome damage in live mammalian cells, and improved measurements of its effects on their- colony-forming ability. Int. J. Radiat. Biol., 41 (2), 161-181. Joshi,G.P., Nelson,W.J., Revell,S.H. £ Shaw,C.A. (1982b) Division probability and division delay in diploid Syrian hamster cells following a range of X-ray doses. Int. J. Radiat. Biol. , 41(4), 443-448. Joshi,C.P., Nelson,W.J., Revell,S.H. £ Shaw,C.A. (1982c) Discrimination of slow growth from non-survival among small colonies of diploid Syrian hamster cells after chromosome damage induced by a range of X-ray doses. Int. J. Radiat. Biol., 42(3h 283-296. 128 Jung.H. (1982) Postirradiation Growth Kinetics of Viable and Nonviable CHO cells. Radiat. Res., 89, 88-98. Kellerer,A.M. & Rossi,H.H. (1971) RBE and the Primary Mechanism of Radiation Damage. Radiat. Res., 47, 15-34. Kellerer,A.M. & Rossi,H.H. (1972) The theory of dual radiation action. Current Topics Radiat. Res., 8, 85-158. Koch, C.J. (1975) Measurements of very low oxygen tensions in liquids; does the extrapolation number of mammalian survival curves decrease after X-irradiation under anoxic conditions? In; Cell Survival after Low  Doses of Radiation: Theoretical and Clinical Implications, T. Alper, ecT The Institute of Physics and John Wiley & Sons, London. 167-173. Koch,C.J. S Painter,R.B. (1975) The Effect of Extreme Hypoxia on the Repair of DNA Single-Strand Breaks in Mammalian Cells. Radiat. Res., 64, 256-269. Koch,C.J., MenesesJ.J., Harris,J.W. (1977) The Effect of Extreme Hypoxia and Glucose on the Repair of Poten-tially Lethal and Sublethal Radiation Damage by Mammalian Cells. Radiat. Res., 70, 542-551. Koch, C.J. (1979) The Effect of Oxygen on the Repair of Radiation Damage by Cells and Tissues. Adv. in Radiation Biol., 8, 273-315. Lam,G.K.Y. , Henkelman,R.M., Douglas,B.G. & Eaves,C.J. (1979) Method of analysis to derive cell survival from observation of tissue damage following fractionated irradiation. Radiat. Res., 77, 440-452. Lam,G.K.Y., Henkelman ,R.M., Douglas,B.G. & Eaves,C.J. (1981) Dose dependence of pion RBE values for mouse foot skin reactions. Int. J. Radiat. Oncol. Biol. Phys., 7, 1689-1694. Lewis,E.B. (1957) Leukemia and Ionizing Radiation. Science, 1 25(3255), 965-972. Littbrand, B. £ Revesz, L. (1969) The effect of oxygen on cellular survival and recovery after ra-diation. Brit. J. Radiol. , 42, 914-924. Lloyd,D.C., Purrott,R.J. , Dolphin,G.W. , Bolton,D., Edward,A.A., & Corp,M.J. (1975) The relationship between chromosome aberrations and low-LET ra-diation dose to human lymphocytes. Int. J. Radiat. Biol., 28, 75-90. 129 McCulloch.E.A. £ Til l , J . E. (1962) The Sensitivity of Cells from Normal Mouse Bone Marrow to Gamma Radiation jn vitro and m vivo. Radiat. Res., 1 6 , 822-832. McNallv, N.J. (1975) The effect .of repeated small doses of radiation on recovery from sublethal damage by Chinese hamster cells irradiated in oxic or hypoxic conditions in the plateau phase of growth. In; Cell Survival  after Low Doses of Radiation: Theoretical and Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley £ Sons, London. 119-125. Meyn,R.E. £ Withers,H.R. (1979) Radiation Biology in Cancer Research. Raven Press, New York. Midander,J., Deschavanne,P.J. , Malaise,E.P. £ Revesz.L. (1982) Survival Curves of Irradiated Glutathione-deficient Human Fibroblasts: Indication of a Reduced Enhancement of Radiosensitivity by Oxygen and Misonidazole. Int. J. Radiat. Oncol. Biol. Phys. , 8, 443-446. Midander, J. £ Revesz, L. (1980) The frequency of micronuclei as a measure of cell survival in irradi-ated cell populations. Int. J. Radiat. Biol. , 38, 237-242 Millar,B.C., Fielden,E.M. £ MillarJ.L. (1978) Interpretation of Survival-curve data, for Chinese Hamster Cells, line V-79 using the multi-target, multi-target with initial slope, and ot,6 equations. Int. J. Radiat. Biol. , 33(6), 599-603. Millar, B.C., Fielden, E.H., 6 Steele, J. (1979) A biphasic radiation survival response of mammalian cells to molecular oxygen. Int. J. Radiat. Biol. , 36, 177-180. Moore, B.A., Palcic, B. £ Skarsgard, L.D. (1976) Radiosensitizing and toxic effects of the 2-nitroimidazole Ro-07-0582 in hypoxic mammalian cells. Radiat. Res. , 67, 459-473. Mortimer,R., Brustad,T. £ Cormack,D.V. (1965) Influence of Linear Energy Transfer and Oxygen Tension on the Effectiveness of Ionizing Radiations for Induction of Mutations and Lethality in Saccharomyces cerevisiae. Radiat. Res. , 26, 465-482. NCRP Report No. 64. (1980) Influence of Dose and Its Distribution in Time on Dose-Response  Relationships for low-LET Radiations. National Council on Radiation Protection and Measurements, Washington, D.C. Ngo,F.Q.H. , Blakely,E.A. , Yang,T.C.H., Yezzi,M.J. £ Tobias,C.A. (1980a) Cellular Damage and Repair Following Heavy-Ion Irradiation. In: Biological and Medical Research with Accelerated Heavy Ions at the  Bevalac, 1977-1980., M. C. Pirruccello £ C.A.Tobias, eds. LBL-1 1220 , UC-48. 103-112. 130 Ngo.F.Q.H., Blakely,E.A. s Tobias, C A . (1980b) Effects of Combined Low- and High-LET radiations. In; Biological  and Medical Research with Accelerated Heavy Ions at the Bevalac,  1977-1980., M.C.Pirruccello & CA.Tobias, eds. LBL-1 1 220 , UC-48. 89-101. Nias,A.H.W., Swal low, A. J ., KeeneJ.P. & Hodgson,B.W. (1973) Absence of a fractionation effect in irradiated HeLa cells. Int. J.  Radiat. Biol., 23(6), 559-569. Nias,A.H.W. (1974) The Clinical Significance of Cell Survival Curves. In: The Biological  and Clinical Basis of Radiosensitivity. M. Friedman, ed., Charles C. Thomas Publisher., Springfield, III., 156-169. Palcic,B. & Skarsgard,L.D. (1972) The effect of oxygen on DNA single-strand breaks produced by ionizing radiation in mammalian cells. Int. J. Radiat. Biol. , 21(5), 417-433. Palcic,B., Hohman,W.F. & Skarsgard, L. D. (1978) Variation of oxygen enchancement ratio with radiation dose. Radiat.  Res., 74, 497. (Abstract) Parker, L. , Skarsgard, L.D. & Emmerson, P.T. (1969) Sensitization of anoxic mammalian cells to X-rays by triacetoneamine N-oxy. Survival and toxicity studies. Radiat. Res. , 38, 493-500. Pettersen,E.O. , Oftebro,R. & Brustad,T. (1973) X-ray inactivation of human cells in tissue culture under aerobic and extremely hypoxic conditions in the presence and absence of TMPN. Int. J. Radiat. Biol. , 24(3), 285-296. Pettersen,E.O., Wibe,E., Lovhaug,D., Oftebro,R. & Brustad,T. (1975) Effects of oxygen and TMPN on the initial part of the dose-effect curves in human cells in culture. In; Cell Survival after Low Doses  of Radiation: Theoretical and Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. 150-157. Phillips,T.L. , Fu,K.K. & Lane,L.J. (1975) The effect of hypoxia and dose-rate on the shape of survival curves for EMT6 tumor and B14FAF cells jn vitro. In; Cell Survival after  Low Doses of Radiation: Theoretical and Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. 158-166. Planel.G. Soleilhavoup, J. P. Tixador,R., Croute,F. & Richoilley ,C. (1976) Demonstration of a Stimulating Effect of Natural Ionizing Radiation and of very Low Radiation Doses on Cell Multiplication. In; Biological  and Environmental Effects of Low-Level Radiation. , International Atomic Energy Agency, Vienna, Vol. 1., 127-140. 131 Pohl-Ruling,J. , Fischer,P. & Pohl,E. (1978) The low level shape of dose response for chromosome aberrations. In Late Biological Effects of Ionizing Radiation, Vol II, IAEA/STI/PUB/489. International Atomic Energy Agency, Vienna. Porter,E.H. (1965) Dose-rate and Survival Curves. Brit. J. Radiol., 38, 607-612. Puck, T.T. (1958) Action of Radiation on Mammalian Cells. III. Relationship Between Reproductive Death and Induction of Chromosome Anomalies by X-irradiation of Euploid Human Cells jri vitro. Proc. Natl. Acad. Sci.  (U.S. ) , 44, 772-780. Puck,T.T. & Marcus,P.I. (1956) Action of X-rays on mammalian cells. J. Exp. Med. , 103, 653-666. Quintiliani,M. (1979) Modification of Radiation Sensitivity: The Oxyqen Effect. Int. J.  Radiat. Oncol. Biol. Phys., 5, 1069-1076. ~ Revesz,L. & Littbrand,B. (1964) Variation of the relative sensitivity of closely related neoplastic cell lines irradiated in culture in the presence or absence of oxygen. Nature, 203, 742-744. Revesz,L. & Littbrand,B. (1967) Qualitative differences between cellular radiation damage suffered under oxic and anoxic conditions; the dual effect of oxygen. In; Proc. Int. Conf. Radiat. Biol. Cancer, Kyoto, 1966, 67-76. Revesz, L., Littbrand, B., Midander, J. & Scott, O.C.A. (1975) Oxygen effects in the shoulder region of the cell survival curves. In; Cell Survival after Low Doses of Radiation: Theoretical and  Clinical Implications, T. Alper, ed. The Institute of Physics and John Wiley & Sons, London. 141-149. Revesz, L. (1981) Radiobiological Oxygen Effects and Cellular Sulfhydryls. studia  biophysica., 86 , 17-18. Rona.E. (1979) Laboratory Contamination in the Early Period of Radiation Research. Health Physics, 37, 723-727. Roots,R. & Smith,K.C. (1975) Rejoining of DNA single-strand breaks in mammalian cells incubated in buffer or in medium after aerobic or anaerobic X-irradiation. Int. J. Radiat. Biol., 27(6), 596-602. Roots,R., Chatterjee.A., Blakely,E., Chang,P., Smith,K. & Tobias,C. (1982) Radiation Responses in Air-, Nitrous Oxide- and Nitrogen- Saturated Mammalian Cells. Radiat. Res., 92, 245-254. 132 Sacks,M.S. & SeemanJ. (1947) A Statistical Study of Mortality From Leukemia. Blood, 2, 1-14. Sinclair, W.K. (1966) The shape of radiation survival curves of mammalian cells cultured in vitro. In; Biophysical Aspects of Radiation Quality,. I.A.E.A., Vienna. 21-48. Siracka,E., Littbrand , B. , Clifton,K.H. & Revesz,L. (1975) Variations in Sensitivity of Synchronized Chinese Hamster Cells to Oxic and Anoxic X-ray Exposures. Neoplasma, 22(6), 647-653. Skarsgard,L.D., Kihlman,B. A., Parker,L. , Pujara,C.M. & Richardson ,S. (1967) Survival, Chromosome Abnormalities, and Recovery in Heavy-lon-and X-lrradiated Mammalian Cells. Radiat. Res. Suppl., 7 , 208-221 . Skarsgard,L.D. , Palcic,B. £ Lam,G. K . Y. (1980) RBE and OER values for the TRIUMF biomedical pion beam. Radiat.  Res. , 83, 471. (Abstract) Skov,K.A., Palcic,B. & Skarsgard,L.D. (1979) Radiosensitization of Mammalian Cells by Misonidazole and Oxygen: DNA Damage Exposed by Micrococcus luteus Enzymes. Radiat. Res., 79, 591-600. Skov,K.A., Palcic,B. & Skarsgard, L. D. (1981) Enhancement of indirect action by sensitizers: Studies on DNA damage. Radiat. Res. , 87, 472. (Abstract) Skov,K.A., Berger,S., Adomat,H., Palcic,B. & Skarsgard,L.D. (1982) Cellular Repair of Radiation-Induced Breaks in DNA. Radiat. Res. , 91_, 388. (Abstract) Skov,K.A. (1983) The Contribution of Hydroxyl Radical to Radiosensitization: A Study of DNA damage. Radiat. Res., in preparation. Suit,H. & Wette,R. (1966) Radiation dose fractionation and tumour control probability. Radiat.  Res., 29, 267-281. Terasima,T. & Tolmach,l I. (1963) Growth and nuclei acid synthesis in synchronously dividing popu-lations of HeLa cells. Exp. Cell. Res. , 30, 344-362. Thomlinson,R.H. & Gray,L.H. (1955) The histological structure of some human lung cancers and the possible implications for radiotherapy. Brit. J. Cancer, 9, 539-549. Thompson, L.H. & Suit, H.D. (1967) Proliferation Kinetics of X-irradiated mouse L cells studied with time-lapse photography. I. Experimental methods and data analysis. Int. J. Radiat. Biol. , 13, 391-397. 133 Thompson, L.H. & Suit, H.D. (1969) Proliferation Kinetics of X-irradiated mouse L cells studied with time-lapse photography. II. Int. J. Radiat. Biol., 15, 347-362. Tobias.C.A., Blakely ,E. A. , Ngo,F.Q.H. & Yang,T.C.H. (1979) The Repair-Misrepair Model of Cell Survival. In; Radiation Biology in  Cancer Research, R.E.Meyn & H.R.Withers, eds. Raven Press, New York. 195-230. Tobleman,W.T. & Cole,A. (1974) Repair of Sublethal Damage and Oxygen Enhancement Ratio for Low-Voltage Electron Beam Irradiation. Radiat. Res., 60, 355-360. Todd,P. (1966) Reversible and irreversible effects of densely ionizing radiations upon the reproductive capacity of cultured mammalian cells. Med.  College Virginia Q., 1_, 2-14. UNSCEAR (1958) Report of the United Nations Scientific Committee on the Effects of Atomic Radiation, General Assembly document, 13th Session, Supple-ment No. 17 (A/3838). United Nations, New York. UNSCEAR (1982) Report of the United Nations Scientific Committee on the Effects of Atomic Radiation, General Assembly document, 37th Session, Supple-ment No. 45 (A/37/45). United Nations, New York. Upton,A.C. (1982) The Biological Effects of Low-Level Ionizing Radiation. Scientific  American, ~246(2), 41-49. Urtasun,R., Band,P., Chapman ,J . D. , Feldstein ,M. L. , Mielke,B. & Fryer,C. (1976) Radiation and High-Dose Metronidazole in Supratentorial Glioblastomas N. Engl. J. Med. , 2^ 94, 1364-1367. Whillans, D.W. & Hunt, J.W. (1980) A rapid mixing comparison of radiosensitization by oxygen and misonidazole in CHO cells. Radiat. Res., 90, 126-141. Whillans, D.W. & Rauth, A.M. (1980) An experimental and analytical study of oxygen depletion in stirred cell suspensions. Radiat. Res. , 84, 97-1 14. Whitmore,G.F. & Gulyas,S. (1980) Lethal and sublethal effects of misonidazole under hypoxic con-ditions. In: Radiation Sensitizers, Their Use in the Clincal Manage- ment of Cancer, L.W.Brady, ed. , Masson Publishing, New York. 99-106. Wideroe, R. (1966) High-energy electron therapy and the two-component theory of radiation. Acta Radiol. , 4, 257-278. 133 Thompson, L . H . & Su i t , H.D. (1 969) P ro l i f e ra t i on K ine t i c s of X - i r r a d i a t e d mouse L ce l l s s tud ied with t ime- lapse pho tog raphy . II. Int. J . Radiat. B i o l . , 15, 347-362. T o b i a s , C . A . , B l ake l y ,E. A . , N g o . F . Q . H . & Y a n g . T . C . H . (1979) The R e p a i r - M i s r e p a i r Model of Ce l l S u r v i v a l . In; Radiat ion B io logy in Cance r Re sea r ch , R .E .Meyn & H .R .Wi the r s , eds. Raven P re s s , New Y o r k . 195-230. Tob leman.W.T. & C o l e , A . (1974) Repa i r of Sub letha l Damage and Oxygen Enhancement Ratio for Low-Vo l tage E lect ron Beam I r rad ia t i on . Rad iat . Res. , 60, 355-360. T o d d , P . (1966) Reve r s i b le and i r r e v e r s i b l e ef fects of densely i on i z ing rad iat ions upon the r ep roduc t i v e capac i ty of c u l t u r e d mammalian ce l l s . Med.  Col lege V i r g i n i a Q . , 1_, 2-14. U N S C E A R (1958) Report of the Un i ted Nations Sc ien t i f i c Committee on the E f fect s of Atomic Rad ia t i on , Cenera l As sembly document, 13th Se s s i on , S u p p l e -ment No. 17 (A/3838). Un i ted Nat ions , New Y o r k . U N S C E A R (1982) Report of the Un i ted Nations Sc ien t i f i c Committee on the E f fect s of Atomic Rad i a t i on , Genera l As sembly document, 37th Se s s i on , S u p p l e -ment No. 45 (A/37/45). Un i ted Nat ions , New Y o r k . U p t o n , A . C . (1982) The B io log ica l E f fect s of Low-Leve l Ionizing Rad ia t i on . Sc ien t i f i c  Amer ican , 246(2) , 41-49. U r t a s o n . R . , B a n d , P . , Chapman ,J . D. , Fe lds te in ,M. L. , M i e l k e , B . & F r y e r , C . (1976) Radiat ion and H igh-Dose Metron idazole in Sup ra ten to r i a l Gl ioblastomas N. E n g l . J . Med. , 294, 1364-1367. Whi l lans , D.W. & Hunt , J.W. (1980) A rap id mix ing comparison of rad iosens i t i zat ion by oxygen and . misonidazole in CHO ce l l s . Rad iat . Re s . , 90, 126-141. Whi l lans, D.W. & R a u t h , A . M . (1980) A n exper imenta l and ana ly t i ca l s t udy of o xygen dep let ion in s t i r r e d cel l s u spens ions . Rad iat . Res. , 84, 97-114. Wh i tmore ,G.F . & C u l y a s . S . (1980) Lethal and sub letha l e f fect s of misonidazole unde r h ypox i c c o n -d i t i on s . In: Radiat ion S e n s i t i z e r s , T h e i r Use in the C l i n ca l Manage- ment of C a n c e r , L.W. B r a d y , ed . , Masson P u b l i s h i n g , New Y o r k . 99-106. Wideroe, R. (1966) H i g h - e n e r g y e lect ron t he rapy and the two-component theory of r ad ia t i on . Ac ta R a d i o l . , 4, 257-278. 1 34 Wideroe, R. (1970) A discuss ion of the oxygen effect based upon the two-component theory of radiation. Radiol. C l i n . (Basel ) , 3_9, 20-31. Wideroe,R. (1977) Theor ies about radiobiological effects. Kerntechn ik , 19 , 237-240. Winston,B.M. , B e r r y , R . J . & O l i ve r ,R . (1975) Oxygen enhancement ratio for small X - r a y doses measured in Vicia  faba roots. In; Cell Surv iva l after How Doses of Radiation: Theore t - ical and Clinical Implications, T . A lpe r , ed. The Institute of Phy s -ics and John Wiley & Sons, London. 115-118. Y u . C . K . & S inc la i r ,W.K. (1967) Mitotic Delay and Chromosomal Aberrat ions Induced by X - r ay s in Synchron ized Chinese Hamster Cells In V i t ro . J . Natl. Cancer Inst., 39, 619-632. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0085787/manifest

Comment

Related Items