UBC Faculty Research and Publications

Cord blood hematopoietic cells from preterm infants display altered DNA methylation patterns de Goede, Olivia M; Lavoie, Pascal M; Robinson, Wendy P Apr 20, 2017

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

Item Metadata

Download

Media
52383-13148_2017_Article_339.pdf [ 1.75MB ]
Metadata
JSON: 52383-1.0366287.json
JSON-LD: 52383-1.0366287-ld.json
RDF/XML (Pretty): 52383-1.0366287-rdf.xml
RDF/JSON: 52383-1.0366287-rdf.json
Turtle: 52383-1.0366287-turtle.txt
N-Triples: 52383-1.0366287-rdf-ntriples.txt
Original Record: 52383-1.0366287-source.json
Full Text
52383-1.0366287-fulltext.txt
Citation
52383-1.0366287.ris

Full Text

RESEARCH Open AccessCord blood hematopoietic cells frompreterm infants display altered DNAmethylation patternsOlivia M. de Goede1,2, Pascal M. Lavoie1,3 and Wendy P. Robinson1,2*AbstractBackground: Premature infants are highly vulnerable to infection. This is partly attributable to the preterm immunesystem, which differs from that of the term neonate in cell composition and function. Multiple studies have founddifferential DNA methylation (DNAm) between preterm and term infants’ cord blood; however, interpretation of thesestudies is limited by the confounding factor of blood cell composition. This study evaluates the epigenetic impact ofpreterm birth in isolated hematopoietic cell populations, reducing the concern of cell composition differences.Methods: Genome-wide DNAm was measured using the Illumina 450K array in T cells, monocytes,granulocytes, and nucleated red blood cells (nRBCs) isolated from cord blood of 5 term and 5 preterm(<31 weeks gestational age) newborns. DNAm of hematopoietic cells was compared globally across the450K array and through site-specific linear modeling.Results: Nucleated red blood cells (nRBCs) showed the most extensive changes in DNAm, with 9258differentially methylated (DM) sites (FDR < 5%, |Δβ| > 0.10) discovered between preterm and term infantscompared to the <1000 prematurity-DM sites identified in white blood cell populations. The direction ofDNAm change with gestational ageat these prematurity-DM sites followed known patterns of hematopoietic differentiation, suggesting that termhematopoietic cell populations are more epigenetically mature than their preterm counterparts. Consistentshifts in DNAm between preterm and term cells were observed at 25 CpG sites, with many of these siteslocated in genes involved in growth and proliferation, hematopoietic lineage commitment, and thecytoskeleton. DNAm in preterm and term hematopoietic cells conformed to previously identified DNAmsignatures of fetal liver and bone marrow, respectively.Conclusions: This study presents the first genome-wide mapping of epigenetic differences in hematopoietic cells acrossthe late gestational period. DNAm differences in hematopoietic cells between term and <31 weeks were consistent withthe hematopoietic origin of these cells during ontogeny, reflecting an important role of DNAm in their regulation. Due tothe limited sample size and the high coincidence of prematurity and multiple births, the relationship between cause ofpreterm birth and DNAm could not be evaluated. These findings highlight gene regulatory mechanisms at both cell-specific and systemic levels that may be involved in fetal immune system maturation.Keywords: DNA methylation, Cord blood, Preterm birth, Illumina 450K array, Epigenetics, Nucleated red blood cells,Epigenetic clock, Gestational age* Correspondence: wrobinson@cfri.ca1BC Children’s Hospital Research Institute, Room 2082, 950W 28th Avenue,Vancouver, BC V5Z 4H4, Canada2Department of Medical Genetics, University of British Columbia, Vancouver,BC V6T 1Z3, CanadaFull list of author information is available at the end of the article© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.de Goede et al. Clinical Epigenetics  (2017) 9:39 DOI 10.1186/s13148-017-0339-1BackgroundPreterm birth (PTB), defined as birth prior to 37 weeks ges-tational age (GA), occurs in approximately 11% of livebirths and accounts for over half of infant mortality casesworldwide [1]. If a premature infant survives the immediatepostnatal period, they face increased risk of developingmajor short- and long-term health problems includingcerebral palsy, chronic lung disease, visual and hearingimpairments, and adult metabolic diseases [2–6]. This ele-vated risk is attributable to organ immaturity, as well as anincreased risk of medical complications linked to oxidativestress and inflammation during the neonatal period [7–9].The immune system is not spared from the effects ofPTB. The composition and function of hematopoietic cellpopulations change dramatically throughout gestation asthe embryonic and fetal immune system mature. Prematureinterruption of the immunologically protected intrauterineenvironment results in an extremely fragile infant whoseimmune system is unprepared for the microbe-ridden ex-ternal environment. A variety of systemic and cell-specificalterations in immune function have been identified in pre-term infants that greatly increase their vulnerability toinfection [10–12].The importance of DNA methylation (DNAm) in theprocess of hematopoietic cell lineage commitment is wellestablished [13, 14], and multiple studies have found differ-ential methylation between cord blood of preterm and terminfants [15–17]. However, these studies have used wholeblood, which is a mixed-cell sample in which overallDNAm levels are influenced by cell composition [18, 19].As a result, these studies cannot distinguish prematurity-associated DNAm patterns due to differences in cellcomposition from DNAm patterns reflecting developmen-tal changes in immune function.Using the Illumina Infinium Human Methylation 450Bead Chip (450K array), we provide genome-wide DNAmprofiles of T cells, monocytes, granulocytes, and nucleatedred blood cells (nRBCs) collected from cord blood of in-fants born at term or highly preterm (<31 weeks GA).These DNAm profiles were compared between cell typesand across GA to evaluate an epigenetic basis for alteredneonatal immune function with prematurity. This studyprovides important insights into the role of DNAm inearly hematopoietic system maturation in humans.MethodsStudy participants and sample collectionEthics approval for this study was obtained from theUniversity of British Columbia Children’s and Women’s(C&W) Research Ethics Board (certificate numbers H07-02681 and H04-70488). Written informed parental con-sent to participate was obtained. Individual patient datais not reported. Cord blood was collected from neonatesdelivered by caesarean section at the C&W HealthCentre of BC (Vancouver, Canada). A total of 10 infantswere involved in the study: 5 preterm (GA range 26–30 weeks) and 5 term (GA 38 weeks) (Table 1). None ofthe subjects had histological evidence of chorioamnioni-tis in the placenta. All term births were singleton, andthe caesarean section was performed in the absence oflabor. The preterm births had more variable clinicalcharacteristics, including one case of preeclampsia, fourbirths from multiple pregnancies, and a case of laborpreceding the caesarean section (Table 1). In the cases ofmultiple pregnancies, only one subject was used andother siblings were excluded. Since the preterm birthswere all <31 weeks, immune function is expected to besignificantly altered compared to term births regardlessof the cause of prematurity.T cells, monocytes, and nRBCs were collected from cordblood by fluorescence-activated cell sorting (FACS). Thesesorting methods were designed to prevent erythrocyte-white blood cell (WBC) cross-contamination, a commonoccurrence in cord blood [20] and are described in detailTable 1 Subject characteristics and cell types collected from each subjectSex GA (weeks) Multiple birth Presence of labor Indication for PTB Cells collectedterm_1 M 38 No No n/a allterm_2 M 38 No No n/a allterm_3 F 38 No No n/a allterm_4 M 38 No No n/a allterm_5 M 38 No No n/a allpreterm_A M 26 No No Preeclampsia T cells, nRBCspreterm_B F 29 Yes No Placental insufficiency T cells, gran., mono.preterm_C M 30 Yes No Placental insufficiency allpreterm_D F 30 Yes No Placental insufficiency T cells, mono., nRBCspreterm_E M 30 Yes Yes Twin-to-twin transfusion syndrome allFor the column “Cells collected”: all T cells, granulocytes, monocytes, and nRBCs; gran. granulocytes; mono. monocytes; n/a not applicablede Goede et al. Clinical Epigenetics  (2017) 9:39 Page 2 of 13in the Additional file 1. Granulocytes were collected bydensity gradient centrifugation and hypotonic red bloodcell lysis. All cell populations were collected from all termsubjects; however, due to small sample volumes and vari-ability in blood cell counts, some cell populations couldnot be collected from some preterm subjects (Table 1).DNA extraction and DNA methylation data collectionDNA was extracted from all samples using standardprotocols and purified with the DNeasy Blood &Tissue Kit (Qiagen, MD, USA). DNA was bisulphite-converted using the EZ DNA Methylation Kit (ZymoResearch, CA, USA) before amplification andhybridization to the 450K array following manufac-turer’s protocols (Illumina, CA, USA). Samples wererandomly distributed across four 450K array chips, asshown in Additional file 1: Figure S1. 450K arraychips were scanned with a HiScan reader (Illumina).Raw intensity data for all hematopoietic cells werebackground corrected in GenomeStudio (Illumina).Quality control was performed using the 835 controlprobes included in the array. The intensity data werethen exported from GenomeStudio and were con-verted into M values using the lumi package [21] inR software [22]. Sample identity and quality wereevaluated as described in Additional file 1. The 450Karray targets 485577 DNAm sites, but probe filteringwas performed as described in Additional file 1 toproduce a final dataset of 429765 sites. Red-greencolor bias was corrected for using the lumi package[21], and the data were normalized by subset within-array quantile normalization [23].DNA methylation data analysisUnsupervised Euclidean clustering of the samplesbased on DNAm β values and principal componentanalysis based on DNAm M values were performedas exploratory global analysis steps. DNAm was thenevaluated at subsets of the 450K array based on sur-rounding CpG density. These subsets are detailed inAdditional file 1. Median DNAm (β values) of theseCpG site groups were compared between all celltypes using ANOVA followed by Tukey’s honest sig-nificant difference test, using a multiple comparison-adjusted p value threshold of 0.005. DNAm-basedestimates of GA for the samples were calculatedusing a method developed by Knight et al. [24] incord blood.Differential methylation based on cell type and birthgroup (preterm or term) was assessed by linear mod-eling using the R package limma [25]. The samemodel was used to assess both PTB-associated andcell type-specific DNAm: the interaction of cell typeand birth group was the variable of interest, and sexwas included in the model as a covariate. Since eachcell type was collected from the same set of individ-uals and the sample size was small, DNAm may havebeen influenced by inter-individual differences. To adjustfor this, the model included a within-individual consensuscorrelation estimated using the duplicateCorrelation()function in limma [25]. Resulting p values were adjustedfor multiple comparisons by the Benjamini and Hochberg[26] false detection rate (FDR) method.For the comparison between preterm and termsamples, statistically significant sites (“prematurity-as-sociated DM sites”) were limited to those with anFDR < 5% and a |Δβ| > 0.10. Prematurity-associatedDM sites were identified separately for each cellpopulation. For cell type-specific DNAm, statisticallysignificant sites (“cell type-DM sites”) were limited tothose with an FDR < 5% and a |Δβ| > 0.20. Cell-typeDM sites were identified separately within the twobirth groups. ErmineJ was used to evaluate enrich-ment of gene ontology (GO) terms in genes associ-ated with the cell-type and prematurity-associatedDM sites [27].Several other studies have performed similar evalu-ations of DNAm differences between preterm andterm births, using whole cord blood instead of iso-lated cell populations [15–17]. The PTB-associatedCpG sites discovered in those studies (29 CpG sitesfrom Parets et al. [17]; 1347 CpG sites fromFernando et al. [16]; and 1555 CpG sites fromCruickshank et al. [15]) were overlapped with theprematurity-associated and the cell-type DM CpGsites identified in this study. Several subject charac-teristics varied between these studies, includingcause of prematurity, ethnicity, and maternal agerange. Overlapping sites are thus likely to be thosepresent in multiple cell populations and also be un-related to ethnicity or to PTB etiology. The specificsof these overlaps are described in Additional file 1.To assess how prematurity-associated DNAm mightreflect hematopoietic origin, DNAm patterns were com-pared between the birth groups and the cell types at aset of previously identified CpG sites that showed differ-ential methylation between erythroblasts derived fromfetal liver and erythroblasts derived from adult bonemarrow (“source-DM sites”) [28]. These source-DM siteswere divided into two groups: the top 100 CpG siteshypomethylated in adult BM erythroblasts (“BM-hypo-methylated sites”) and the top 100 CpG sites hypo-methylated in FL erythroblasts (“FL-hypomethylatedsites”), with ranking and selection based on Lessardet al.’s β values. Median DNAm (β value) at thesesource-DM sites was compared between cell types andbirth groups by ANOVA followed by Tukey’s honestde Goede et al. Clinical Epigenetics  (2017) 9:39 Page 3 of 13significant difference test, with a multiple comparison-adjusted p value threshold of 0.005.ResultsGlobal DNA methylation profiles of preterm and termhematopoietic cellsCell type was the dominant influence when DNAmprofiles of term and preterm cell populations werecompared by array-wide Euclidean clustering (Fig. 1a).Prematurity also had an observable impact on epigen-etic relationships between the samples, with somesamples clustering by birth group (preterm or term)within each cell type. However, these GA subgroupswere not perfect, with some preterm samples cluster-ing more closely with their term counterparts.Evaluating genome-wide DNAm by β value densitydistributions suggests that the effect of prematurity islargest in nRBCs. Term nRBCs were hypomethylatedrelative to preterm nRBCs, whereas all of the WBCpopulations showed similar distributions betweenterm and preterm samples (Fig. 1b).None of the cord blood hematopoietic cell popula-tions differed in median array-wide DNAm betweenpreterm and term infants; although term nRBCs werenotably hypomethylated compared to preterm nRBCs,this difference was not significant (Fig. 1c). To iden-tify genomic regions where the association betweenDNAm and prematurity is strongest, subsets of the450K array based on CpG density were evaluated. InWBCs, no significant differences were observed be-tween preterm and term samples at any of the CpGdensity subgroups. In nRBCs, term cells displayed hy-pomethylation relative to their preterm counterpartsat the intermediate and low CpG density regions,however, these differences did not pass the multiple-test corrected significance threshold (p > 0.005).Fig. 1 Genome-wide DNAm comparisons between term and preterm hematopoietic cells. a 450K array-wide Euclidean clustering of DNAm data for eachpreterm and term subject. b DNAm β value density distributions; each line represents the mean of that birth group/cell type combination. c Comparisonof median DNAm between GA groups and cell types array-wide, and in CpG sites grouped by CpG density. *p< 0.05, **p< 0.005de Goede et al. Clinical Epigenetics  (2017) 9:39 Page 4 of 13Prematurity-associated differentially methylated CpG sitesLinear modeling was performed within each cell type toidentify cell-specific prematurity-associated DM sites(FDR < 5%, |Δβ| > 0.10). nRBCs showed the greatest differ-ence between preterm and term samples, with 9258prematurity-associated DM sites; more than tenfoldgreater than observed in granulocytes, monocytes, and Tcells (Table 2; Additional file 2). The majority ofprematurity-associated DM sites were specific to a singlecell type, making it unlikely that these changes weredriven by chance genetic differences between the samples(Additional file 1: Figure S2). Twenty-five of theprematurity-associated DM sites were identified across allcell types, 17 of which increased in DNAm and 8 of whichdecreased in DNAm with GA (Additional file 2). All cellpopulations had the highest number of their prematurity-DM sites in the gene body and intergenic regions (Fig. 2a),which is consistent with the representation of these generegions on the 450K array (33 and 24% of CpG sitesassayed, respectively). In nRBCs, the TSS-upstream and 5′UTR gene regions were also highly represented inprematurity-DM sites. This likely reflects the global natureof erythrocyte demethylation with maturity [29, 30].The direction of DNAm change at prematurity-associated DM sites in each cell type paralleled their pat-terns of DNAm upon terminal differentiation (Table 2).For T cells, the high percentage of prematurity-associatedDM sites with increased DNAm in term samples (72%) isconsistent with increased DNAm with terminal differenti-ation of these cells [13, 14, 31]. For granulocytes andmonocytes, the majority of prematurity-associated DMsites were hypomethylated in term samples (69 and 61%,respectively) in keeping with the documented loss ofDNAm in myeloid cells [13, 14, 31]. For nRBCs, the vastmajority of prematurity-associated DM sites (94%) showedreduced DNAm in term samples. Given that terminalerythroid differentiation is associated with global demeth-ylation [29, 30], this change likely reflects an increasingproportion of mature erythroblasts in the nRBC popula-tion at term. Overall, these data suggest an epigeneticbasis for the increased maturity of cord bloodhematopoietic cell populations during fetal development.GO pathway analysis of the prematurity-associated DMsites revealed enrichment of distinct sets of genes for eachcell type (Additional file 3). The two significant GO termsin granulocytes (FDR < 10%) related to negative regulationof the ERK1 and ERK2 cascades, and Ras guanyl-nucleotideexchange factor activity. In T cells, the only significant GOterm (FDR < 10%) was embryonic placenta development.Monocyte prematurity-associated DM sites were associatedwith eight significant GO terms (FDR < 10%), all of whichwere related to epidermal and hair growth anddevelopment. Evaluating the nRBC prematurity-associatedDM sites revealed 152 significantly enriched GO terms(FDR < 10%); recurring themes in this list included Ras-and Rho-related activity, the cytoskeleton, and terms relatedto renal, muscle, and neuronal processes.Cell-specific DNA methylation patterns differ betweenpreterm and term infantsAfter establishing the cell-specific DNAm differences be-tween preterm and term births, we next investigatedwhether prematurity affects cell-type differences inDNAm (Additional file 4). Linear modeling revealed thatnRBCs were the most distinct cell type in term samples,consistent with our previous findings [20, 32], but inter-estingly, T cells were the most distinct cell type of thepreterm samples (Table 3). The relatively low number ofmonocyte- and granulocyte-DM sites in both GA groupswas likely because these cell types are both of the mye-loid lineage and thus epigenetically similar, in contrast toT cells and nRBCs, which are the only representatives oftheir respective hematopoietic lineages. In the WBCpopulations, the number of cell type-DM sites did notchange drastically between preterm and term samples(Table 3). In contrast, the number of nRBC-DM sitesnearly tripled between the preterm and term samples.This large change coincides with the increased hypome-thylation in term nRBCs relative to their preterm coun-terparts (Fig. 1b), which made term nRBCs more distinctfrom term WBCs.When the cell-type DM sites were compared betweenpreterm and term samples based on gene region and thecell type of interest’s relative DNAm—that is, whetherthe unique cell population has DNAm that is higher,lower, or in between the DNAm levels of the other celltypes—there was little difference in genomic representa-tion or direction of DNAm change, particularly withinWBCs (Fig. 2b). In nRBCs, the increase in the numberof CpG sites hypomethylated relative to WBCs occurredin all gene regions, but most dramatically in the genebody and intergenic regions. Overall, this indicates thatprematurity does not have a major impact on WBCs’epigenetic relationships to each other. In contrast,Table 2 Number of prematurity-associated DM sites for each cell type (FDR < 5%, |Δβ| > 0.10)T cells Granulocytes Monocytes nRBCsTotal 273 987 692 9258DNAm decreases with GA 76 (28%) 679 (69%) 425 (61%) 8731 (94%)DNAm increases with GA 197 (72%) 308 (31%) 267 (39%) 527 (6%)de Goede et al. Clinical Epigenetics  (2017) 9:39 Page 5 of 13nRBCs become more epigenetically distinct from WBCsas gestation progresses, adopting their uniquely hypo-methylated profile [20, 32, 33]. The representation ofthese changes across all gene regions is likely reflectiveof nRBC demethylation being a global and largely pas-sive process [29, 30].Comparison to other epigenetic studies of preterm birthPrevious studies have also identified distinctive DNAmpatterns between preterm and term infants [15–17].However, these studies were performed on either wholecord blood samples or the buffy coat and were not ableto distinguish systematic prematurity-associated changesFig. 2 Prematurity-associated and cell type-DM sites grouped by gene region and changes in DNAm. a Prematurity-associated DM sites (FDR < 5%,|Δβ| > 0.10). b Cell type-DM sites (FDR < 5%, |Δβ| > 0.20). TSS 1500 & 200 1500 or 200 bp upstream from transcriptional start site, UTRuntranslated regionTable 3 Number of cell type-DM sites (FDR < 5%, | Δβ| > 0.20)T cells Granulocytes Monocytes nRBCsPreterm 12974 1410 1665 9056Term 12662 1900 1508 26176Common 10991 (85%, 87%) 1201 (85%, 63%) 1221 (73%, 81%) 7645 (84%, 29%)Percentages of cell type-DM sites in common between the two GA groups are reported relative to the number of preterm DM sites first, then number of termDM sitesde Goede et al. Clinical Epigenetics  (2017) 9:39 Page 6 of 13from those caused by shifts in cell composition acrossgestation. We compared our prematurity-associated andcell type-DM sites with the CpG sites found to be sig-nificantly associated with PTB (FDR < 5%) by Paretset al. [17] (29 CpG sites), Fernando et al. [16] (1347 CpGsites), and Cruickshank et al. [15] (1555 CpG sites).For all three of the comparison studies, approxi-mately 30% of their DM sites were also discovered inat least one of our sets of prematurity-associated DMsites (9/29, 369/1347, and 427/1555 replicated DMsites) (Fig. 3a). When our identified cell type-DMsites were overlapped with the three comparison stud-ies, the overlap was much lower than that with theprematurity-associated DM sites (Fig. 3b). The great-est overlap by number of sites occurred at T cell-specific and nRBC-specific DM sites (Fig. 3b). How-ever, a notable proportion of Parets et al.’s [17] 29differentially methylated CpG sites in PTB were asso-ciated with monocyte-specific DNAm in our data, atrend not seen in comparison with the other twostudies. This could be due to a chance difference inaverage monocyte proportions between their pretermand term subjects or it could have come from Paretset al.’s [17] use of the buffy coat rather than thewhole blood. Additionally, a subset of 196 of Fer-nando et al.’s [16] prematurity-associated CpG sitesthat were associated only with the state of being pre-mature, and not with GA, showed almost no overlapwith our cell type-DM sites (Fig. 3b).A CpG site in MYL4, encoding myosin light chain 4,was the only DM site identified by Fernando et al. [16],Cruickshank et al. [15], and Schroeder et al. [34]; how-ever, it was not observed in any of our cell populations.We also did not find any prematurity-associated DMsites in ESR1, encoding the estrogen receptor, in any ofour cell populations, despite this gene being identifiedby both Fernando et al. [16] and Schroeder et al. [34].However, we did replicate some of Fernando et al.’s [16]top findings of differential methylation in NCOR2,DNAJC17, PYCR2, ATP6V0A1, RARA, FBLN7, IGF2BP1,Fig. 3 Overlap of prematurity-associated and cell type-DM sites with prematurity-associated CpG sites identified in previous studies. Proportion ofprematurity-associated CpG sites found by Cruickshank et al., Fernando et al., and Parets et al. [15–17] also represented in (a) the prematurity-associatedDM sites (FDR < 5%); and (b) the cell type-DM sites (FDR < 5%, |Δβ| > 0.20) identified in this study. The numbers beside bars are the number of overlappingCpG sites between the two listsde Goede et al. Clinical Epigenetics  (2017) 9:39 Page 7 of 13and ATP2B2, as well as differential methylation observedby Cruickshank et al. [15] in NFIX, OXT, DNMT3A,RUNX1, and AIRE. We also found prematurity-associated DNAm in ADORA2A and GABBR1, whichwas identified by both of these studies. Of the 54prematurity-associated DM sites we observed across allcell types, seven were also identified by both Fernandoet al. [16] and Cruickshank et al. [15]. Of these sharedCpG sites, two are located in the gene body of WWTR1,and two are located in the 5′UTR of CLIP2; the otherthree are intergenic.To further explore how the cell-specific DNAmchanges we observed compared to trends in whole cordblood, we applied the recently published epigenetic clockfor GA to our data [24]. This GA-epigenetic clock wasdesigned using cord blood samples and, unlike the epi-genetic clock designed for adult samples [35], was onlyvalidated in cord blood. This is unsurprising, since cordblood is the most frequently studied tissue in studies ofthe fetus or neonate. However, we were curious to seehow this whole blood-based algorithm would performon its constituent cell types. In all preterm cell popula-tions, estimated DNAm GA was an overestimate of ac-tual GA (Additional file 5). In term samples, the GAestimates were more accurate: when estimated GA wasaveraged across all cell types within an individual, noneof the term individuals had estimates over 1 week differ-ent than their actual GA (Additional file 5). There werealso some intriguing cell type-specific trends in GA esti-mates: for example, T cells had the highest GA estimatesin preterm individuals, but one of the lowest in term in-dividuals (Fig. 4). Additionally, monocytes in term indi-viduals were consistently estimated as the “oldest” cellpopulation, whereas nRBCs had low GA estimates re-gardless of birth group.DNA methylation associated with prematurity may reflecthematopoietic originWhile changes in DNAm may in part reflect an aging“clock” [24], our cell-specific GA-epigenetic clock ana-lyses above suggest that other factors can modify thistrend. One such factor may be the predominanthematopoietic organ, which shifts from the liver to thebone marrow early in the third trimester of gestation[36]. We hypothesized that the preterm samples used inthis study, which range from 26 to 30 weeks GA, have agreater proportion of liver-derived cells than the termsamples. Hematopoietic source-related methylation dif-ferences with PTB were evaluated using CpG sites previ-ously associated with liver- or bone marrow-specificDNAm in ex vivo-derived nRBCs [28]. From these 5937“source-DM sites”, two groups of CpG sites wereFig. 4 DNAm-based estimates of gestational age (GA) by cell type and birth group. Dashed lines reflect mean actual GA for the birth group.Estimates calculated using methods published by Knight et al. [24]de Goede et al. Clinical Epigenetics  (2017) 9:39 Page 8 of 13assessed: the top 100 sites that were hypomethylated inadult bone marrow-derived nRBCs relative to fetal liver-derived nRBCs (“BM-hypomethylated sites”) and the top100 sites that displayed the opposite pattern (“FL-hypo-methylated sites”). Only one of these 200 CpG sitesoverlapped with the 148 CpG sites used in theGA-epigenetic clock [24]. In our samples, all cell typesdisplayed the same trend, with preterm samples lessmethylated at FL-hypomethylated sites and term samplesless methylated at BM-hypomethylated sites (Fig. 5).This difference at BM-hypomethylated sites was signifi-cant in all cell types except T cells.Overlapping our prematurity-associated DM sites withLessard et al.’s hematopoietic source-DM sites providedfurther support for a relationship between DNAm andhematopoietic origin of cord blood cell types (Table 4).CpG sites that increased in DNAm with increasing GAoverlapped almost exclusively with fetal liver-hypomethylated CpG sites, likely reflecting the reducedcontribution of the liver to hematopoiesis as gestationprogresses. In contrast, many of the CpG sites that de-creased in DNAm with increasing GA were also associ-ated with hypomethylation in bone marrow-derivedhematopoietic cells, corresponding with this organ be-coming the primary source of hematopoietic cells towardsthe end of the third trimester.DiscussionPrevious DNAm studies using cord blood have identifiedsignificant differences between preterm and term infants[15–17]; however, interpretation of these studies is lim-ited by the confounding factor of cord blood cell com-position. Granulocytes and T cells are the two mostabundant cell types in whole blood and thus are themost likely to influence overall DNAm, but cell typeproportions show considerable inter-individual variabil-ity and also change with gestational age [12]. SomeDNAm changes previously associated with prematuritymay simply reflect these changes in cell compositionwith GA. This study is the first evaluation of the epigen-etic impact of PTB in hematopoietic cell populationsisolated from the same individuals.An important question is the functional role of theseDNAm changes in hematopoietic cell populations duringontogeny. There was a notable difference in the numberof cell-specific prematurity-associated DM sites in eachcell type, ranging from 273 in T cells to 9258 in nRBCs(Table 2). The number of prematurity-associated DMsites in a given cell type may relate to the magnitude ofphenotypic differences between preterm and term popu-lations. For example, we and others have reported majorfunctional differences in dendritic cells and macrophagesbetween preterm and term infants [37–40]. In contrast,fewer functional differences have been observed betweenpreterm and term T cells [41, 42]. For granulocytes,much less is known regarding gestational differences.The high number of prematurity-associated DM sites weobserved in granulocytes (987) suggests a more dynamicmaturation across late gestation than for either mono-cytes or T cells. Alternatively, it is possible that theseDNAm changes may reflect differences in the compos-ition of granulocyte subsets, including a mixture of eo-sinophils, basophils, and mast cells, between age groups.However, this is less likely given that our granulocyteswere overwhelmingly represented by neutrophils in bothpreterm and term samples (>95%; data not shown).Given the extent of DNAm differences between pretermand term granulocytes, functional studies may providenew insight into the limitations of the preterm immunesystem.Our findings showed moderate overlap with previousstudies of prematurity-associated DNAm in whole cordblood, with approximately 30% of the DM sites fromeach comparison study also discovered in at least one ofour cell types (Fig. 3a). This is a greater amount of over-lap than the 161 of 1347 CpG sites Fernando et al. [16]found in common with Cruickshank et al. [15] and an-other study not evaluated in this paper [34]. ThisFig. 5 Comparison of median DNAm between GA groups and cell types at CpG sites associated with hematopoietic origin. DNAm was compared atthe top 100 CpG sites hypomethylated in erythroblasts derived from adult bone marrow stem cells (left) and the top 100 CpG sites hypomethylated inerythroblasts derived from fetal liver stem cells, as identified by Lessard et al. [28]. *p < 0.05, **p < 0.005de Goede et al. Clinical Epigenetics  (2017) 9:39 Page 9 of 13increased overlap with other studies may reflect reducednoise our data due to eliminating variation due to cellcomposition differences. Alternatively, since we com-pared four sets of prematurity-associated DM sites (oneper cell type), all of which were of fairly large size, wemay have increased our chance of overlap just by havinga greater number of hits.When the cell type-DM sites discovered in this studyare overlapped with the three comparison studies, thehighest numbers of overlapping sites were observed withT cell- and nRBC-specific DM sites, the two cell typeswith the strongest cell-specific DNAm patterns (Fig. 3b).We additionally compared our cell type-DM sites to asubgroup of 196 CpG sites identified by Fernando et al.to be associated with PTB but not GA—and thusthought to reflect systematic differences due to prema-turity rather than cell composition—and found almostno overlap (Fig. 3b). This supports their assertion thatthose 196 CpG sites are more likely related to themolecular mechanisms of PTB than cell composition,compared to the 1151 CpG sites they identified as asso-ciated with both PTB and GA. Thus, subsequent studiesof PTB may be able to work around concerns of variabil-ity in cell composition by considering prematurity as aseparate variable from GA.Two CpG sites highlighted by Fernando et al. as being ofpotential interest for preterm delivery were not replicatedin our study. One site, in MYL4 (myosin light chain 4), wasthe only DM site identified by Fernando et al. [16],Cruickshank et al. [15], and Schroeder et al. [34]; the other,ESR1 (estrogen receptor) was observed in both Fernandoet al. [16] and Schroeder et al. [34]. Fernando et al.suggested that these sites may be related to the laborprocess, since MYL4 activity is involved in the myometrialcontraction pathway [43], and upregulation of ESR1 leadsto the increase in estrogen activity required for contractions[44]. The lack of replication in our data may be a conse-quence of all of our subjects being born by caesarean sec-tion, whereas all of the comparison studies included at leastsome subjects born by vaginal delivery. Notable genes inwhich we replicated the differential methylation found byother studies include ADORA2A, which has been associ-ated with the inflammatory pathway in the myometrium[45], and GABBR1, which encodes a gamma-aminobutyricacid receptor and has been associated with chemotaxis incord blood-derived stem cells [46]. These two genes wereidentified as differentially methylated in both Fernandoet al. [16] and Cruickshank et al. [15].GO pathway analyses of the prematurity-associatedDM sites highlighted potentially important differencesin gene regulation that are unique to each cell type(Additional file 3). For instance in granulocytes,prematurity-associated DM sites were significantlyenriched for genes associated with the Ras-Raf-MEK-ERK cascade. Defects in this pathway have beenassociated with impaired neutrophil extracellular trapformation and with respiratory burst in neutrophils[47, 48], both of which are also deficient in preterminfants [49, 50]. The prevalence of prematurity-associ-ated DM sites in genes associated with these functionscould reflect either reduced functional ability in pretermneutrophils or a low proportion of neutrophils within thepreterm granulocyte population. Some of our findings pointtowards novel pathways potentially involved in the matur-ation of hematopoietic cells, such as the enrichment forprematurity-associated DM sites in genes associated withplacental development in T cells and with dermal develop-ment in monocytes. In nRBCs, prematurity-associatedDNAm changes were widespread and associated with GOterms related to the cytoskeleton, membrane compositionand cell-cell junctions, and motility. This may reflect thelarge-scale structural changes that occur in erythroblasts asthey mature and prepare to extrude their nucleus.Based on gestational age differences, DNAm conformedto the epigenetic profile of the dominant hematopoieticorgan when evaluated in source-DM sites [28]: the liver inmid-gestation, and the bone marrow in late gestation(Fig. 5). Despite these candidate sites being identified exclu-sively in nRBCs derived ex vivo from hematopoietic stemcells, the DNAm trends in this study were consistent acrossboth nRBCs and WBCs. Thus, our findings indicate thathematopoietic sources have epigenetic signatures that areshared across multiple cell lineages derived from that organ.Additionally, our analysis of these hematopoietic source-DM sites revealed that nRBCs actually gain DNAm with in-creasing GA in functionally relevant regions of the genome,specifically the fetal liver-hypomethylated sites. This is arare occurrence, as the overwhelming trend is for nRBCs toTable 4 Overlap between cell-specific prematurity-associated DM sites (FDR < 5%, |Δβ| > 0.10) and Lessard et al.’s source-DMsites [25]DNAm decreases with GA DNAm increases with GAT cell Gran. Mono. nRBC T cell Gran. Mono. nRBCTotal 76 679 425 8731 197 308 267 527Overlap with BM-hypo. sites 25, 32.9% 197, 29.0% 213, 50.1% 895, 10.3% 1, 0.5% 0, 0.0% 1, 0.4% 1, .02%Overlap with FL-hypo. sites 0, 0.0% 1, 0.1% 0, 0.0% 2, 0.0% 22, 11.2% 74, 24.0% 70, 26.2% 89, 16.9%Gran. granulocytes, mono. monocytesde Goede et al. Clinical Epigenetics  (2017) 9:39 Page 10 of 13become demethylated both during erythropoiesis [29, 30]and as the fetus approaches term. This novel observation isimportant to our understanding of hematopoiesis duringontogeny since it indicates that although nRBC demethyla-tion is largely global and passive [29, 30], it also has someselectivity, with certain CpG sites protected from the wide-spread DNAm loss.The main limitation in our study is the small sample size.Other studies evaluating prematurity-associated DNAmhad sample sizes ranging from 22 [16] to 50 [17]. With onlyten subjects, our study had reduced power to detectchanges in DNAm. Considering the large epigenetic differ-ences between cell lineages, we expect that our study wassufficiently powered to compare cell types. However, differ-ential methylation associated with prematurity is expectedto be of a smaller scale than cell type differences, so thismay have led to an underestimation of prematurity-associated DM sites. There was also an increased chancethat genetic factors impacted our findings, since some CpGsites are methylation quantitative trait loci (mQTLs), orsites where DNAm is more strongly associated with indi-viduals than cell type [20, 51]. We mitigated this concernby performing an additional probe filtering step to removesuspected mQTLs, as described in the SupplementaryMethods (Additional file 1).It is possible that heterogeneity in our subjects’ clinicalcharacteristics reduced our ability to detect prematurity-associated differential methylation. All births were caesar-ean sections with no indications of infection; however, onepreterm case was attributed to preeclampsia, and four ofthe five preterm births were multiples (Table 1). Thisraises the concern that multiplicity in the preterm subjectsmay have confounded our results. There is limited infor-mation on how the immune system differs with multiplebirths: it has been shown that intrauterine infection occursmore often in preterm births with dizygotic twins com-pared to monozygotic twins or singletons, but no differ-ences in postnatal outcome have been associated withzygosity [52]. Additionally, CD4+ T cell activity has beenobserved to be significantly lower in preterm dizygotictwins than in preterm singletons [53]. For DNAm studies,the effect of twin births has only been assessed within twinpairs [54], not between twins and singletons. For ourstudy, considering the extreme difference in GA of ourpreterm and term cases (<31 versus >38 weeks), we expectthat prematurity will have a much greater effect on DNAmthan differences due to multiplicity. This is in keeping withFernando et al.’s [16] observation of more distinct cluster-ing of DNAm in extreme PTB cases compared to inter-mediate PTB cases. Despite these limitations, we identifiedprematurity-DM sites that showed reasonable overlap withprior studies [15–17] and cell-specific DNAm patterns thatwere consistent with our previous findings in cord bloodcell populations [20, 32].ConclusionsThe preterm immune system differs from that of theterm neonate in both cell composition and function,resulting in heightened vulnerability to infection in pre-term infants. We identified epigenetic markers ofimmune system differences with prematurity by compar-ing DNAm of major cord blood hematopoietic cellpopulations across gestation. Changes in DNAm be-tween preterm and term hematopoietic cells in our studylikely reflect a shift from the liver to the bone marrow asthe predominant hematopoietic source with advancinggestational age. Granulocytes were identified as a candi-date cell population of particular interest in preterm in-fants’ susceptibility to infection, due to their relativelyhigh number of prematurity-associated DM sites and theenrichment of these sites for GO terms related to theRas-Raf-MEK-ERK cascade. Our findings provide im-portant insights into the epigenetic regulation ofhematopoietic cell-specific functions during fetal devel-opment. These data may have clinical implications, asthey highlight gene regulatory mechanisms on both cell-specific and systemic levels that are involved in neonatalimmune system maturity. Larger samples will be re-quired to determine the potential impact of cause ofPTB (such multiple gestations or preeclampsia) on theseepigenetic profiles.Additional filesAdditional file 1: Supplementary methods and figures. (DOCX 207 kb)Additional file 2: “Location and genomic context of prematurity-DMsites”. Each sheet is a table summarizing prematurity-DM sites (FDR < 5%,|Δβ| > 0.10) for each cell type, as well as one for the DM sites commonacross all cell types. (XLSX 1890 kb)Additional file 3: Significantly enriched GO terms (corrected p value <0.10)from ErmineJ analysis of prematurity-DM sites for each cell type, ordered bycorrected p value. (XLSX 18 kb)Additional file 4: “Location and genomic context of cell type-DM sites”.Each sheet is a table summarizing cell type-DM sites (FDR < 5%, |Δβ| > 0.20)for each birth group/cell type combination. (XLSX 11523 kb)Additional file 5: DNAm-based estimates of gestational age (GA) usingKnight et al.’s [24] methods. (XLSX 10 kb)Abbreviations450K array: Illumina Infinium Human Methylation 450 BeadChip; BM: Bonemarrow; DM: Differentially methylated; DNAm: DNA methylation; FDR: Falsedetection rate; FL: Fetal liver; GA: Gestational age; GO: Gene ontology;nRBC: Nucleated red blood cell; PTB: Preterm birth; WBC: White blood cellAcknowledgementsWe thank the BC Children’s and Women’s Hospital staff for their help withsubject recruitment. We also thank Paul Villeneuve for his contributions to thecell sorting method; Kristi Finlay for obtaining consent for preterm subjects inthis study; Ruby Jiang, Mihoko Ladd, Dr. Maria Peñaherrera for their work insample processing; and Dr. Lisa Xu for flow cytometer operation.FundingThis research was funded by grants from the Canadian Institutes of HealthResearch (CIHR; MOP-123478 to PML and MOP-49520 to WPR). OMdG issupported by a CIHR Frederick Banting and Charles Best Graduatede Goede et al. Clinical Epigenetics  (2017) 9:39 Page 11 of 13Scholarship–Master’s Award. PML is supported by a Clinician-Scientist Awardfrom the BC Children's Hospital Research Institute and a Career InvestigatorAward from the Michael Smith Foundation for Health Research. WPR is sup-ported by a Scientist Award from the BC Children's Hospital ResearchInstitute.Availability of data and materialsThe dataset supporting the conclusions of this article is available in the NCBIGene Expression Omnibus repository, GSE82084.Authors’ contributionsOMdG sorted cord blood cells for DNA methylation analyses, performedDNA methylation data analyses, and wrote the manuscript. PML and WPRsupervised and designed the research, interpreted the data, and co-wrotethe manuscript. All authors read and approved the final manuscript.Competing interestsThe authors declare that they have no competing interests.Consent for publicationNot applicable.Ethics approval and consent to participateEthics approval for this study was obtained from the University of BritishColumbia Children’s and Women’s Research Ethics Board (certificate numbersH07-02681 and H04-70488). Written informed parental consent to participatewas obtained. Individual patient data is not reported.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1BC Children’s Hospital Research Institute, Room 2082, 950W 28th Avenue,Vancouver, BC V5Z 4H4, Canada. 2Department of Medical Genetics, Universityof British Columbia, Vancouver, BC V6T 1Z3, Canada. 3Department ofPediatrics, University of British Columbia, Vancouver, BC V6T 1Z3, Canada.Received: 26 October 2016 Accepted: 8 April 2017References1. Blencowe H, Cousens S, Chou D, Oestergaard M, Say L, Moller AB, et al.Born too soon: the global epidemiology of 15 million preterm births.Reprod Health. 2013;10 Suppl 1:S2.2. Barker DJ, Gelow J, Thornburg K, Osmond C, Kajantie E, Eriksson JG.The early origins of chronic heart failure: Impaired placental growthand initiation of insulin resistance in childhood. Eur J Heart Fail.2010;12(8):819–25.3. Koupil I, Leon DA, Lithell HO. Length of gestation is associated withmortality from cerebrovascular disease. J Epidemiol Community Health.2005;59(6):473–4.4. O’Connor AR, Wilson CM, Fielder AR. Opthalmological problems associatedwith preterm birth. Eye (Lond). 2007;21(10):1254–60.5. O’Reilly M, Sozo F, Harding R. Impact of preterm birth and bronchopulmonarydysplasia on the developing lung: long-term consequences for respiratoryhealth. Clin Exp Pharmacol Physiol. 2013;40(11):765–73.6. Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet. 2008;371(9608):261–9.7. Lavoie PM, Lavoie JC, Watson C, Rouleau T, Chang BA, Chessex P.Inflammatory response in preterm infants is induced early in life by oxygenand modulated by total parenteral nutrition. Pediatr Res. 2010;68(3):248–51.8. Strunk T, Currie A, Richmond P, Simmer K, Burgner D. Innate immunity inhuman newborn infants: prematurity means more than immaturity. JMatern Fetal Neonatal Med. 2011;24(1):25–31.9. Takala TI, Makela E, Suominen P, Matomaki J, Lapinleimu H, Lehtonen L, etal. Blood cell and iron status analytes of preterm and full-term infants from20 weeks onwards during the first year of life. Clin Chem Lab Med. 2010;48(9):1295–301.10. Dowling DJ, Levy O. Ontogeny of early life immunity. Trends Immunol.2014;35(7):299–310.11. Kan B, Razzaghian HR, Lavoie PM. An immunological perspective on neonatalsepsis. Trends Mol Med. 2016;22(4):290–302.12. Sharma AA, Jen R, Butler A, Lavoie PM. The developing human pretermneonatal immune system: a case for more research in this area. ClinImmunol. 2012;145(1):61–8.13. Cedar H, Bergman Y. Epigenetics of haematopoietic cell development. NatRev Immunol. 2011;11(7):478–88.14. Ji H, Ehrlich LI, Seita J, Murakami P, Doi A, Lindau P, et al. Comprehensivemethylome map of lineage commitment from haematopoietic progenitors.Nature. 2010;467(7313):338–42.15. Cruickshank MN, Oshlack A, Theda C, Davis PG, Martino D, Sheehan P, et al.Analysis of epigenetic changes in survivors of preterm birth reveals theeffect of gestational age and evidence for a long term legacy. GenomeMed. 2013;5(10):96.16. Fernando F, Keijser R, Henneman P, van der Kevie-Kersemaekers AMF, MannensMM, van der Post JAM, et al. The idiopathic preterm delivery methylation profilein umbilical cord blood DNA. BMC Genomics. 2015;16:736.17. Parets SE, Conneely KN, Kilaru V, Fortunato SJ, Syed TA, Saade G, et al. FetalDNA methylation associates with early spontaneous preterm birth andgestational age. PLoS One. 2013;8(6), e67489.18. Lam LL, Emberly E, Fraser HB, Neumann SM, Chen E, Miller GE, et al. Factorsunderlying variable DNA methylation in a human community cohort. ProcNatl Acad Sci U S A. 2012;109 Suppl 2:17253–60.19. Reinius LE, Acevedo N, Joerink M, Pershagen G, Dahlen SE, Greco D, etal. Differential DNA methylation in purified human blood cells:implications for cell lineage and studies on disease susceptibility. PLoSOne. 2012;7(7), e41361.20. de Goede OM, Razzaghian HR, Price EM, Jones MJ, Kobor MS, Robinson WP,et al. Nucleated red blood cells impact DNA methylation and expressionanalyses of cord blood hematopoietic cells. Clin Epigenetics. 2015;7(1):95.21. Du P, Kibbe WA, Lin SM. lumi: a pipeline for processing Illumina microarray.Bioinformatics. 2008;24(13):1547–8.22. R Core Team. R: A language and environment for statistical computing.Vienna: the R Foundation for Statistical Computing; 2014.23. Maksimovic J, Gordon L, Oshlack A. SWAN: Subset-quantile within arraynormalization for Illumina Infinium HumanMethylation450 BeadChips.Genome Biol. 2012;13(6):R44. -2012-13-6-r44.24. Knight AK, Craig JM, Theda C, Bækvad-Hansen M, Bybjerg-Grauholm J,Hansen CS, et al. An epigenetic clock for gestational age at birth based onblood methylation data. Genome Biol. 2016;17:206.25. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powersdifferential expression analyses for RNA-sequencing and microarray studies.Nucleic Acids Res. 2015;43(7):e47.26. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practicaland powerful approach to multiple testing. J R Stat Soc. 1995;57:289–300.27. Gillis J, Mistry M, Pavlidis P. Gene function analysis in complex data setsusing ErmineJ. Nat Protoc. 2010;5(6):1148–59.28. Lessard S, Beaudoin M, Benkirane K, Lettre G. Comparison of DNAmethylation profiles in human fetal and adult red blood cell progenitors.Genome Med. 2015;7(1):1. -014-0122-2 . eCollection 2015.29. Shearstone JR, Pop R, Bock C, Boyle P, Meissner A, Socolovsky M. GlobalDNA demethylation during mouse erythropoiesis in vivo. Science. 2011;334(6057):799–802.30. Yu Y, Mo Y, Ebenezer D, Bhattacharyya S, Liu H, Sundaravel S, et al.High resolution methylome analysis reveals widespread functionalhypomethylation during adult human erythropoiesis. J Biol Chem.2013;288(13):8805–14.31. Alvarez-Errico D, Vento-Tormo R, Sieweke M, Ballestar E. Epigenetic controlof myeloid cell differentiation, identity and function. Nat Rev Immunol.2015;15(1):7–17.32. de Goede OM, Lavoie PM, Robinson WR. Characterizing thehypomethylated DNA methylation profile of nucleated red blood cells fromcord blood. Epigenomics. 2016;8(11):1481–94.33. Bakulski KM, Feinberg JI, Andrews SV, Yang J, Brown S, McKenney S L, et al.DNA methylation of cord blood cell types: applications for mixed cell birthstudies. Epigenetics. 2016;11(5):354–62.34. Schroeder JW, Conneely KN, Cubells JC, Kilaru V, Newport DJ, Knight BT, etal. Neonatal DNA methylation patterns associate with gestational age.Epigenetics. 2011;6(12):1498–504.35. Horvath S. DNA methylation age of human tissues and cell types. GenomeBiol. 2013;14:R115.de Goede et al. Clinical Epigenetics  (2017) 9:39 Page 12 of 1336. Holt PG, Jones CA. The development of the immune system duringpregnancy and early life. Allergy. 2000;55(8):688–97.37. Sharma AA, Jen R, Kan B, Sharma A, Marchant E, Tang A, et al. ImpairedNLRP3 inflammasome activity during fetal development regulates IL-1betaproduction in human monocytes. Eur J Immunol. 2015;45(1):238–49.38. Weatherstone KB, Rich EA. Tumor necrosis factor/cachectin and interleukin-1secretion by cord blood monocytes from premature and term neonates.Pediatr Res. 1989;25(4):342–6.39. Strunk T, Prosser A, Levy O, Philbin V, Simmer K, Doherty D, et al. Responsivenessof human monocytes to the commensal bacterium Staphylococcus epidermidisdevelops late in gestation. Pediatr Res. 2012;72(1):10–8.40. Lavoie PM, Huang Q, Jolette E, Whalen M, Nuyt AM, Audibert F, et al.Profound lack of interleukin (IL)-12/IL-23p40 in neonates born early ingestation is associated with an increased risk of sepsis. J Infect Dis.2010;202(11):1754–63.41. Currie AJ, Curtis S, Strunk T, Riley K, Liyanage K, Prescott S, et al. Preterminfants have deficient monocyte and lymphocyte cytokine responses togroup B streptococcus. Infect Immun. 2011;79(4):1588–96.42. Melville JM, Moss TJ. The immune consequences of preterm birth. FrontNeurosci. 2013;7:79.43. Salomonis N, Cotte N, Zambon AC, Pollard KS, Vranizan K, Doniger SW, et al.Identifying genetic networks underlying myometrial transition to labor.Genome Biol. 2005;6(2):R12.44. Hirota Y, Cha J, Dey SK. Prematurity and the puzzle of progesteroneresistance. Nat Med. 2010;16(5):529–31.45. Lee Y, Sooranna SR, Terzidou V, Christian M, Brosens J, Huhtinen K, et al.Interactions between inflammatory signals and the progesterone receptorin regulating gene expression in pregnant human uterine myocytes. J CellMol Med. 2012;16(10):2487–503.46. Zangiacomi V, Balon N, Maddens S, Tiberghien P, Versaux-Botteri C,Deschaseaux F. Human cord blood-derived hematopoietic and neural-likestem/progenitor cells are attracted by the neurotransmitter GABA. StemCells Dev. 2009;18(9):1369–78.47. Tortorella C, Stella I, Piazzolla G, Simone O, Cappiello V, Antonaci S.Role of defective ERK phosphorylation in the impaired GM-CSF-inducedoxidative response of neutrophils in elderly humans. Mech Ageing Dev.2004;125(8):539–46.48. Hakkim A, Fuchs TA, Martinez NE, Hess S, Prinz H, Zychlinsky A, et al.Activation of the Raf-MEK-ERK pathway is required for neutrophilextracellular trap formation. Nat Chem Biol. 2011;7(2):75–7.49. Yost CC, Cody MJ, Harris ES, Thornton NL, McInturff AM, Martinez ML, et al.Impaired neutrophil extracellular trap (NET) formation: a novel innateimmune deficiency of human neonates. Blood. 2009;113(25):6419–27.50. Peden DB, VanDyke K, Ardekani A, Mullett MD, Myerberg DZ, VanDykeC. Diminished chemiluminescent responses of polymorphonuclearleukocytes in severely and moderately preterm neonates. J Pediatr.1987;111(6 Pt 1):904–6.51. Yuen RK, Avila L, Penaherrera MS, von Dadelszen P, Lefebvre L, Kobor MS, etal. Human placental-specific epipolymorphism and its association withadverse pregnancy outcomes. PLoS One. 2009;4(10), e7389.52. Spiegler J, Hartel C, Schulz L, von Wurmb-Schwark N, Hoehn T, Kribs A, et al.Causes of delivery and outcomes of very preterm twins stratified byzygosity. Twin Res Hum Genet. 2012;15(4):532–6.53. Aquilano G, Capretti MG, Nanni F, Corvaglia L, Aceti A, Gabrielli L, et al.Altered intracellular ATP production by activated CD4+ T-cells in verypreterm infants. J Immunol Res. 2016;2016:8374328.54. Gordon L, Joo JE, Powell JE, Ollikainen M, Novakovic B, Li X, et al. NeonatalDNA methylation profile in human twins is specified by a complex interplaybetween intrauterine environmental and genetic factors, subject to tissue-specific influence. Genome Res. 2012;22(8):1395–406.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:de Goede et al. Clinical Epigenetics  (2017) 9:39 Page 13 of 13

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:
http://iiif.library.ubc.ca/presentation/dsp.52383.1-0366287/manifest

Comment

Related Items