Study setting

The primary focus of the study analyzed here was to investigate the ability of preexisting maternal antibodies to inhibit congenital CMV transmission. The study consisted of three groups of monkeys: a control group that received no hyperimmune globulin (HIG) pretreatment (n = 5, 3 of which were historical controls [27]), a “standard” pretreatment group (n = 3), and a “high-potency” pretreatment group (n = 3). All eleven pregnant RhCMV-seronegative dam monkeys were first depleted of CD4⁺ T cells using an anti-CD4 monoclonal antibody, and one week later intravenously inoculated with RhCMV, as previously described [9]. The three dams in the standard HIG pretreatment group each received a single dose of a standard HIG preparation 1 hour prior to viral inoculation. The three dams in the high-potency pretreatment group each received an initial dose 1 hour prior to viral inoculation and a second dose 3 days later. Both doses in the high-potency group used a high-potency HIG preparation by screening serum donor monkeys for serum RhCMV neutralizing activity, as described in [9].

The RhCMV inoculum was a mixture of three different strains: UCD52, UCD59, and 180.92, at relative frequencies of 25%, 25%, and 50% (by infectious viral titer), respectively. Both UCD52 and UCD59 are derived from serial propagation on primary monkey kidney epithelial cells, which express epithelial cell morphology and cell-specific markers [31] and contain a full-length UL128-UL131 coding capacity [32]. Previous studies have shown that inoculation of RhCMV-naïve rhesus macaques with UCD52 and UCD59 results in similar patterns of viremia and shedding in bodily fluids, comparable to those observed in colony-reared macaques naturally infected with circulating strains of endemic RhCMV [32–34]. On the other hand, CMV strain 180.92 was derived from serial propagation on rhesus fibroblasts, and is known to be a mixed virus with the dominant strain containing deletions in the UL/b’ region while retaining an intact UL128-UL131 locus. Experimental inoculation of rhesus macaques with 180.92 showed rapid emergence of a minor wild-type like variant but only limited tissue dissemination and viral excretion of the defective UL/b' virus strain [35].

For each of the 11 dams studied, samples were taken from maternal blood plasma, urine, saliva, and amniotic fluid at multiple time points following infection. Sample availability varied across dams for reasons such as early fetal loss or low sample volume, previously described in [9]. A subset of the available samples had virus populations that were successfully sequenced and form the basis of our analysis (S1 Table). The remainder of these samples did not have successful viral sequencing due to either low viral loads or inadequate sample quality prior to library construction. In addition, virus populations in placental tissue from one control group monkey and two standard pretreatment group monkeys, as well as tissues from one congenitally infected fetus (from a standard pretreatment group dam) were successfully sequenced (S2 Table).

RhCMV viral load was quantified from each sample using qPCR, as described in [9]. For all samples, multiple viral load measurements (3 to 18) were taken to ensure that samples with relatively low levels of virus present were identified as being positive for RhCMV. Viral load on the log₁₀ scale was calculated as the mean of the individual log₁₀ viral load sample measurements. When viral load was below the limit of detection (100 viral copies per ml for plasma and amniotic fluid and 100 viral copies per total DNA μg for urine and saliva), we set its value to half of the detection limit.

Animal study ethics statement

The animal protocol titled “Maternal immune correlates with protection against congenital cytomegalovirus transmission in rhesus monkeys” was approved by the Tulane University and the Duke University Medical Center Institutional Animal Care and Use Committees (IACUC) under the protocol numbers P0285 and A186-15-06, respectively. Indian-origin rhesus macaques were housed at the Tulane National Primate Research Center and maintained in accordance with institutional and federal guidelines for the care and use of laboratory animals, specifically the USDA Animal Welfare regulations, PHS Policy on Humane Care and Use of Laboratory Animals[36], the NIH/NRC Guide for the Care and Use of Laboratory Animals, Association for Assessment and Accreditation of Laboratory Animal Care accreditation guidelines, as well as Tulane University and Duke University IACUC care and use policies. Tulane National Primate Research Center has strict policies to minimize pain and distress. The monkeys were observed on a daily basis and were administered tiletamine/zolazepam (Telazol), or ketamine if they showed signs of discomfort, pain or distress. In case of illness, the protocol involved analgesics administration and supplemental nutritional support and/or fluid therapy as needed.

Housing conditions were determined by the time and type of RhCMV inoculation, aiming to avoid horizontal transmission of RhCMV from other colony members, where RhCMV is endemic. RhCMV-seronegative pregnant macaques were housed in pairs after RhCMV inoculation if inoculated concurrently with the same viral isolate. Otherwise, single housing in BL2 containment facilities was required. The monkeys were maintained in a standard environment enrichment setting which included manipulable items, swings, food supplements (fruit, vegetables, treats), task-oriented feeding methods as well as human interaction with caretakers and research staff. Dams were released into the colony after 2 or 3 weeks following C-section.

Anesthesia was considered for all procedures considered to cause more than slight pain in humans, including routine sample collection. The agents used included: ketamine, butorphanol, Telazol, buprenorphine, carprofen, meloxicam, and midazolam as needed. The criteria for end-point was defined as loss of 25% of body weight from maximum body weight during protocol, major organ failure or medical conditions unresponsive to treatment and surgical complications unresponsive to immediate intervention. Policies stated that animals deemed at endpoint would be euthanized by overdose of pentobarbital under the direction of the attending veterinarian, consistent with the recommendations of the American Veterinary Medical Association guidelines on euthanasia.

PCR amplification, viral sequencing, and analysis pipeline

We PCR-amplified two variable regions within the genes encoding glycoprotein B (gB) and glycoprotein L (gL) of RhCMV for subsequent next-generation sequencing. The gB amplicon was 408 nucleotides long and the gL amplicon 399 nucleotides long. The primer sequences for each amplicon are specified in [9]. The gB and gL regions sequenced for this study were selected to amplify all three strains present in the inoculum (UCD52, UCD59, and 180.92) in an unbiased manner. Indeed, we previously confirmed the absence of primer bias against these strains [9]. By sequencing two regions of approximately 400 nucleotides, we were able to identify gB and gL haplotypes instead of characterizing variants across a number of sites without knowledge of their linkage. For each of a given sample’s two amplified loci, our goal was to process two technical replicates. A small number of samples (6 in total), however, only had a single successfully sequenced replicate, while several samples had more than two successfully sequenced replicates (S1 and S2 Tables). As described previously [9], each replicate sample was independently PCR-amplified and sequenced following library preparation. All plasma samples required only a single round of PCR, whereas the samples from urine, saliva, amniotic fluid, placental tissues, and fetal tissues all required nested PCR. All replicates were sequenced on an Illumina MiSeq platform, using paired end reads of 300 bases.

To identify viral haplotypes and quantify their frequencies, we first used PEAR [37] to reconstruct (for each available technical replicate) the targeted locus by merging the paired-end reads corresponding to each sequenced fragment. The fused reads were then filtered using the extractor tool from the SeekDeep pipeline [38], which filters sequences according to their length, overall quality scores, and presence of primer sequences. Haplotype reconstruction for a given technical replicate was performed on the filtered sequences using the qluster tool from SeekDeep, which performs an iterative process of removing spurious, low abundance sequence groups by adding them to more abundant, genetically similar sequence groups when the genetic mismatch between groups occurs at nucleotide positions with low quality.

To obtain a set of haplotypes and their frequencies for a given sample, we combined identified haplotypes across technical replicates. Specifically, for a haplotype to be considered present in a sample, we required it to be detected in both sample replicates. Haplotypes that did not meet this criterion were merged with their genetically-closest haplotype in the sample, and the count of this genetically closest haplotype in the sample was increased accordingly. When only a single replicate was available, we could not perform this step and therefore kept all identified haplotypes present in the single available replicate. When more than two technical replicates were available, we restricted our analyses to the two replicates that were the most similar to one another genetically, based on correlation of haplotype frequencies (see below).

Quality assurance and error reduction in sequencing data

We performed additional tests and required additional criteria to be met to ensure the quality of each sample that would undergo subsequent analysis. First, to reduce the number of spurious haplotypes in a given sample, we set a frequency threshold that sample haplotypes were required to exceed. This threshold was set to 0.436% based on analysis of plasmid controls. Specifically, we constructed two synthetic plasmids, one containing the gB gene and the other containing the gL gene. Two technical replicates from each plasmid were sequenced using the same protocol as for the RhCMV samples. Plasmids were run separately from the monkey samples. Because a single haplotype should be present in these plasmid control populations, any shared low-frequency haplotype is likely a product of PCR amplification error or sequencing error. We found 19 minor haplotypes in the gB plasmid control sample after merging technical replicates. These haplotypes ranged in frequency from 0.01% to 0.59% (S1 Fig). We found 29 minor haplotypes in the gL plasmid control sample after merging technical replicates. These haplotypes ranged in frequency from 0.03% to 0.42% (S1 Fig). Our chosen frequency threshold of 0.436% was set at the 0.95 quantile of the combined minor haplotype distributions from the gB and gL plasmids.

As a second quality assurance step, we performed chimera detection on the haplotypes in each merged sample. A haplotype was classified as a chimera if there was a combination of partial alignments to two observed (and higher frequency) haplotypes in the same sample. Detected chimeras were discarded. These chimeras contributed to only a small fraction of the total reads in each sample (ranging from 1.95% to 7.82% of the reads across all samples).

As a third quality assurance step, we restricted our analysis to those samples that had a Pearson correlation score exceeding 0.70 between the frequencies of the shared haplotypes across technical replicates on the log₁₀ scale. For those samples with only a single technical replicate, we could not perform this step and instead included the sample in our analysis only if the read count exceeded 5000.

S1 Table shows the final set of maternal tissue and amniotic fluid samples that were included in our analyses, for both the gB and the gL loci. S2 Table shows the set of samples from the maternal-fetal interface (other than the amniotic fluid samples) and from fetal samples that were included in our analyses. In addition to these samples, the genetic composition of the inoculum was analyzed. Each of the three viral stocks comprising the inoculum (UCD52, UCD59, 180.92) was independently sequenced. Two successfully sequenced replicates were available for each of the three stock samples.

Strain classification and nucleotide diversity calculations

Each identified haplotype in a sample was classified as belonging to one of the three strains that comprised the inoculum (UCD52, UCD59, or 180.92) based on its genetic distance to the reference sequences of these three strains. The reference sequences of the targeted gB and gL regions were obtained from [32] for strains UCD52 and UCD59 and from [39] for strain 180.92. Nucleotide diversity π present in a sample was calculated for each strain independently using the commonly used Nei-Gojobori equation, as described in [40]. All identified haplotypes, across all sequenced samples, are listed in S1 Appendix and S2 Appendix. The frequencies of these haplotypes in each of the sequenced samples, including the inoculum, are given in S3 Appendix and S4 Appendix.

Statistical analysis and software

Data processing, analysis, and visualization were performed in R. Pairwise comparisons between groups were performed using non-parametric tests as indicated. For the visualization of the haplotype networks, we employed the R package RCy3 version 1.2.0 that interfaces R 3.4 with Cytoscape.

Data and code availability

Sequence data in fastq format from all the samples are available in SRA under the Bioproject PRJNA386504. All the R code used in the analysis of the sequence data is publicly available on GitHub: dverac/SNAPP.

Maternal viral load dynamics, congenital transmission, and strain dominance

As previously described in [9], dams in the high-potency HIG pretreatment group had reduced peak viral loads in maternal plasma relative to dams in the control group following primary maternal infection (S2A Fig). Viral kinetics in the saliva and urine were also delayed in the high-potency pretreatment group compared to the control group (S2C and S2D Fig) [9]. Interestingly, and as previously noted [9], only dams with a peak plasma viral load exceeding 5.0 log₁₀ viral copies/mL transmitted the virus to the amniotic fluid compartment. This included all 5 dams in the control group, 2 out of 3 dams in the standard pretreatment group, but none of the 3 dams in the high-potency pretreatment group. Viral dynamics in the amniotic fluid, when detectable, did not appear to differ between the control group monkeys and the standard HIG pretreatment group monkeys (S2B Fig).

Of the three viral strains used in the RhCMV inoculum, UCD52 became dominant in the overwhelming majority of tissue compartments, regardless of pretreatment group status (S3 Fig) [9]. Because of this, it is likely that UCD52 has higher viral fitness in vivo, relative to both UCD59 and 180.92. The reason for UCD52 dominance over UCD59 is unknown: RhCMV-naïve macaques inoculated with UCD52 and UCD59 have similar patterns of viremia and shedding in bodily fluids [32–34]. That strain 180.92 did not become dominant is not surprising given previous findings of only low-level viral shedding and limited tissue dissemination of this strain in inoculated RhCMV-naïve macaques [35]. Given the dominance of the UCD52 strain in the overwhelming majority of samples, we focused our remaining analyses on haplotypes that were classified as belonging to the dominant UCD52 strain.

Minor RhCMV haplotypes and levels of genetic diversity during acute maternal infection

Across the majority of analyzed samples, we found that the dominant in vivo UCD52 haplotype was the canonical UCD52 reference haplotype of the viral inoculum. This was the case both for the gB locus and the gL locus, and across all groups and compartments studied.

Our analysis of amplified sequences from the gB locus identified 7 minor haplotypes in the UCD52 inoculum, as well as a large number of minor haplotypes in maternal and fetal compartments (S3 Appendix, Fig 1, S4–S11 Figs). The number of haplotypes identified in a sample was not positively correlated with the sample’s viral load (S12 Fig), indicating that the number of observed haplotypes was not restricted by sample viral load. The identified minor haplotypes differed from the canonical UCD52 gB haplotype by typically only a single (either synonymous or nonsynonymous) nucleotide mutation. These minor haplotypes ranged in frequency from just above the sequencing error cut-off frequency of 0.436% up to 43.27%, with a median frequency of 0.80%. Maternal samples differed in the number of identified gB haplotypes they contained, ranging from 1 to 33, with a median of 5 haplotypes per sample. Minor haplotypes identified in the maternal samples that were not identified in the UCD52 inoculum potentially arose through intrahost de novo mutation. Alternatively, these minor haplotypes may have been present in the inoculum at frequencies below 0.436%. Within individual dams, we observed that some of the minor haplotypes were shared across timepoints from the same compartment and/or across compartments (Fig 1, S4–S11 Figs). This finding indicates that some of these minor haplotypes persist over a timespan of weeks in a given compartment and that some of these minor haplotypes are likely transmitted across anatomic compartments. Of the minor haplotypes that were shared across compartments, most were shared between the plasma and one other compartment (Fig 1, Fig 2, S4–S11 Figs). This pattern may be due to plasma being a source of viral haplotypes for other compartments; alternatively, it may simply be due to a larger number of plasma samples being successfully sequenced relative to those from other compartments (S1 Table). Interestingly, in 6 out of the 8 monkeys that had both urine and saliva sequences available, there were also minor gB haplotypes that appeared to be shared exclusively between urine and saliva samples. These haplotypes were generally found first in urine and then in a later week in the saliva, suggesting potential oral auto-inoculation from virus shed in urine.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. UCD52 haplotype networks for the gB locus across sampled tissues from three representative monkeys.

Haplotype networks are shown for one control group dam (C4), one standard pretreatment group dam (S2), and one high-potency pretreatment group dam (HP3). Unique haplotypes are shown as circles (nodes). Node sizes scale with haplotype frequency. Green and red lines (edges) connect haplotypes that differ by a single nucleotide. Green edges depict synonymous mutations and red edges depict nonsynonymous mutations. Samples are labeled by collection week. Blue lines connect shared haplotypes across samples. The UCD52 inoculum stock is included in the haplotype network of each monkey.

https://doi.org/10.1371/journal.ppat.1007968.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The number of minor UCD52 haplotypes at the gB locus that are either shared or unique across compartments, by monkey.

Here, the set of minor haplotypes for a given compartment includes all timepoint samples from that compartment. Patterns of minor haplotype sharing for (A) control group monkeys, (B) standard pretreatment group monkeys, and (C) high-potency pretreatment group monkeys. Compartments are color-coded as in Fig 1.

https://doi.org/10.1371/journal.ppat.1007968.g002

To assess whether the number of identified gB haplotypes differed by pretreatment group, we calculated the median number of minor UCD52 haplotypes in each available tissue for each of the 11 dams. This quantity reflects a measure of haplotype richness. We found no significant differences in the median number of minor gB haplotypes by tissue across any pair of pretreatment groups (all Mann-Whitney U tests, p > 0.1; Fig 3A–3D). We further calculated the proportion of minor gB haplotypes that differed from the canonical haplotype by at least one nonsynonymous mutation, by tissue and monkey. No significant differences between these proportions were found between tissues, indicating no clear trend of positive selection or relaxed purifying selection in certain tissues over others at the gB locus. No significant differences between these proportions were found between pretreatment groups on a tissue-by-tissue basis, again indicating no clear trend of positive selection or relaxed purifying selection in one group over others at the gB locus.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Median number of minor haplotypes, by locus, tissue, and pretreatment group.

The panels A,B,C and D show the median number of gB minor haplotypes; panels E,F,G and H show the median number of gL minor haplotypes. Rows show tissues: plasma (A,E), amniotic fluid (B,F), saliva (C,G), and urine (D,H). Marker symbols correspond with those in S2 Fig. We found no significant differences in the median number of minor haplotypes across pretreatment groups for any tissue, at either locus (Mann Whitney test; p > 0.1).

https://doi.org/10.1371/journal.ppat.1007968.g003

The UCD52 haplotype patterns observed at the gL locus are consistent with those at the gB locus. Specifically, minor gL haplotypes generally differed from one of the two dominant gL haplotypes present in the inoculum by a single nucleotide (S13–S22 Figs). Similar to the frequencies observed for gB haplotypes, minor gL haplotypes were present at frequencies as low as 0.44% and up to 48.16%, with a median frequency of 1.05%. Samples differed in the number of identified minor gL haplotypes they contained, ranging from 2 to 29 with a median of 6 minor haplotypes per sample. Again, no correlation was found between the number of haplotypes identified in a sample and the sample’s viral load (S23 Fig). Some of the identified minor gL haplotypes appeared to persist within the same tissue over time, and some were shared across tissue compartments. Similar to our findings at the gB locus, most of the minor haplotypes that were shared across compartments were shared between the plasma and one other compartment (S24 Fig). Minor gL haplotypes shared between urine and saliva compartments again suggested auto-inoculation. Finally, consistent with the findings from the gB locus, the median number of gL minor haplotypes observed in any tissue did not differ between pretreatment groups (Fig 3). We again found no significant differences between tissues or pretreatment groups in the proportion of minor gL haplotypes that differed from the canonical reference haplotype by at least one nonsynonymous mutation, consistent with the lack of pattern at the gB locus.

We next assessed whether HIG pretreatment had an impact on RhCMV genetic diversity, as measured by pairwise nucleotide diversity π for each sample’s UCD52 viral population. Levels of viral genetic diversity varied significantly between monkeys, compartments, and across weeks (Fig 4 for gB, S25 Fig for gL). Despite this variation, median levels of gB viral genetic diversity did not differ by pretreatment group for any tissue (all Mann-Whitney U tests, p > 0.1) besides the amniotic fluid. In this compartment, median levels of gB viral genetic diversity appeared to be slightly higher in standard pretreatment group monkeys than in control animals (Mann-Whitney U test, p = 0.095). Median levels of gL viral genetic diversity did not differ by pretreatment group for any tissue (all Mann-Whitney U tests > 0.1; S25 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Pairwise genetic diversity π over time, by tissue, for the gB locus.

Levels of UCD52 genetic diversity in each monkey varied over time, for each tissue. Marker symbols correspond with those in S2 Fig.

https://doi.org/10.1371/journal.ppat.1007968.g004

Because error rates may be variable across samples due to differences in PCR protocol (in particular, single versus nested PCR), and because the plasmid controls may underestimate the error rate of the samples due to higher template copy numbers in the plasmid controls, we reanalyzed the sequence data at both the gB and the gL loci using a more conservative threshold of 0.88% (double that of the 0.436% threshold). At this more conservative threshold, we identified fewer minor haplotypes and arrived at lower levels of genetic diversity. However, the conclusions we arrived at using the 0.436% threshold were robust to the more conservative threshold of 0.88%. Specifically, there was evidence for some minor haplotypes persisting across time in a given tissue, and for some minor haplotypes being shared between compartments (S5 Appendix). The extent to which minor haplotypes were shared across samples from the same dam was reduced, but still noticeable, particularly at the gL locus. There were no significant differences in levels of RhCMV genetic diversity by HIG pretreatment group or by tissue compartment, with the exception of viral genetic diversity in the amniotic fluid being slightly higher in the standard pretreatment group monkeys than those of the control group (Mann-Whitney U test, p = 0.095). As a further sensitivity analysis, we reanalyzed the sequence data at both the gB and the gL locus using a less conservative threshold of 0.22% (half that of the 0.436% threshold). At this 0.22% threshold, we identified a larger number of minor haplotypes and arrived at slightly higher levels of genetic diversity. However, the conclusions we arrived at using the 0.436% threshold were again robust at the less conservative threshold of 0.22%. There was extensive evidence for minor haplotypes persisting across time in a given tissue, and for some minor haplotypes being shared between compartments (S5 Appendix). Again, there were no significant differences in levels of RhCMV genetic diversity by HIG pretreatment group or by tissue compartment, with the exception of viral genetic diversity in the amniotic fluid being slightly higher in the standard pretreatment group monkeys than those of the control group (p = 0.095).

Genetic diversity and compartmentalization of maternal RhCMV variants identified in placenta and amniotic fluid

We next sought to determine the extent to which minor UCD52 haplotypes were shared between maternal compartments and compartments comprising the maternal-fetal interface (amniotic fluid and placental tissues). As reported above, we found that some minor gB and gL UCD52 haplotypes were shared between maternal plasma samples and amniotic fluid samples (Figs 1 and 2; S3 Appendix and S4 Appendix). Similarly, some of the minor gB and gL UCD52 haplotypes found in placental tissue samples were also present in maternal plasma samples (Fig 2). Specifically, between placental tissues and plasma samples, we observed 6 shared minor gL haplotypes in C4 (S15 Fig), 3 shared minor gB haplotypes in S2 (Fig 1), 1 shared minor gL haplotype in S2 (S18 Fig), and 1 shared minor gB haplotype in S3 (S9 Fig). As these minor shared haplotypes are mostly present at marginal frequencies in maternal tissues (median frequency in plasma for shared haplotypes: 1.28%, minimum 0.47%, in gB, S2; maximum 24.05% in gL, S2) (S3 Appendix and S4 Appendix), the bottleneck between mother and placental tissues is likely relatively large. This reasoning is based on the observation that transmission of low-frequency minor haplotypes would only be evident if the number of virions transmitting is large [41,42], that is, when the transmission bottleneck is loose. Since multiple low-frequency minor haplotypes appear to be shared between maternal plasma and placental tissue samples, and each of these haplotypes is unlikely to transmit between tissues unless the bottleneck is loose, this strongly indicates that the number of virions reaching placental tissues is considerably high. Interestingly, a large number of minor gB and gL UCD52 haplotypes were found in amniotic fluid samples (Fig 2, S24 Fig; S3 Appendix and S4 Appendix), most of which were not observed in maternal tissues. This indicates that de novo viral mutations may occur in the fetus and subsequently be shed into the amniotic fluid. In the one case in which placental plasma was available for analysis (dam S2 in Fig 1; S18 Fig), we found considerably more minor haplotypes in both gB and gL gene regions in this sample compared with the paired placental tissue and many of these minor haplotypes were not observed in maternal plasma.

Genetic diversity and compartmentalization of fetal RhCMV variants

Congenital infection was confirmed in two of three dams in the standard pretreatment group and in all five control dams, whereas congenital infection did not occur in any of the three high-potency pretreatment group dams. All five control dams experienced fetal loss within 2–3 weeks of maternal infection and fetal tissues were often not recovered. In standard pretreatment group dam S3, nearly all the fetal tissues harvested at 6 weeks post-RhCMV infection tested positive for RhCMV, including fetal lung, brain, kidney, spleen, heart, placenta, amniotic fluid, and amniotic membrane. Similar to our observation in the maternal tissue compartments, a single major UCD52 haplotype (the canonical reference haplotype) was present in all fetal tissue samples. Multiple minor UCD52 haplotypes were also detected in these samples (S9 Fig; Fig 5). Intriguingly, a second, minor haplotype was found across all the fetal tissues, present at low frequencies of ≤1%. This haplotype was also observed in one of the paired dam’s three amniotic fluid samples (week 3, frequency of 0.8%), placenta (frequency of 0.85%), and in two of the paired dam’s plasma samples (at weeks 3 and 6; frequencies of 0.58% and 0.66%, respectively) (Fig 5). Of the remaining 21 minor UCD52 haplotypes in fetal tissues, 4 were also present in amniotic fluid samples and 5 in plasma samples of paired dam S3 (Fig 5). Plasma haplotypes detected as late as weeks 5 and 6 post-inoculation contribute to those shared haplotypes. In an attempt to gauge whether the minor fetal haplotypes may have all already been present in the viral inoculum, we determined the overlap between minor haplotypes in the inoculum and those in the fetal samples: none of the minor fetal haplotypes were identified in the UCD52 inoculum. Moreover, we found that only 0–5 minor haplotypes were shared between the fetal tissues and non-paired dams. In comparison, 10 minor haplotypes were shared between fetal tissues and maternal plasma/amniotic fluid samples from the paired dam S3. These findings indicate that the minor haplotypes identified in fetal tissues likely originated de novo, either in the fetal tissues themselves or in maternal tissues.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Minor UCD52 haplotypes found in fetal tissues, and their presence in maternal compartments of dam S3.

Each row depicts a haplotype found in at least one fetal tissue (purple), harvested at 6 weeks post-RhCMV infection. Rows are ordered by their mutated nucleotide position when compared to the major haplotype, which is located in the bottom row. A square denotes that a given haplotype was found in a given sample.

https://doi.org/10.1371/journal.ppat.1007968.g005

To determine whether patterns of genetic diversity were similar between fetal tissues and other tissues, we then calculated pairwise genetic diversity π from each available fetal tissue. We observed lower diversity in the fetal tissues compared to that in both the amniotic fluid (p = 0.025) and the plasma at late weeks post-infection (Weeks 4 to 6, p = 0.095). We further observed higher diversity in the fetal tissues compared to that in plasma during the first three weeks post-infection (Weeks 1 to 3, p = 0.024). These observations together again suggest that the maternal viral population contributes to viral diversity in the fetus and that congenital transmission may be subject to a loose bottleneck. Transmission of RhCMV from maternal plasma to fetal tissues may also occur over an extended period of time, as minor low-frequency haplotypes identified in fetal tissues were shared with those identified in maternal plasma samples between weeks 3 and 6 (S9 Fig).