Abstract
Despite successful suppression of plasma HIV replication by antiretroviral therapy (ART), some women living with HIV (WLHIV) can still experience genital HIV shedding (discordant shedding). Female genital tract (FGT) bacterial and viral microbiome (bacteriome and virome) community dynamics during long-term ART in WLHIV are poorly understood but might contribute to discordant HIV shedding, as the bacteriome and virome are known to influence FGT health. Here, using metagenomic next-generation sequencing, we characterize the bacteriome and virome in 125 cervicovaginal specimens collected over two years from 31 WLHIV in Lima, Peru, and show that FGT bacteriome instability is associated with discordant HIV shedding, while longitudinal changes in FGT virome composition are associated with ART duration. Intrapersonal bacteriome variation is higher in discordant HIV shedders compared to non-shedders. Cervicovaginal virome composition changes over time, particularly in non-shedders. Specifically, anellovirus relative abundance is inversely associated with ART duration and CD4 counts. Our results suggest that discordant HIV shedding is linked with FGT bacteriome instability, and immune recovery during ART influences FGT virome composition.
Similar content being viewed by others
Introduction
With the development of effective antiretroviral therapies (ART), life expectancy for people living with HIV has increased significantly, and HIV infection can now be treated as a chronic disease1,2. Of the estimated 20 million women aged 15 and older living with HIV worldwide in 2022, an estimated 82% were taking ART3. While ART is effective in suppressing HIV replication, some women living with HIV (WLHIV) with undetectable plasma viral loads can still experience genital HIV shedding, described as discordant shedding4,5,6,7,8,9,10. The microbiome has been implicated in female genital tract (FGT) health and disease11. However, the relationship between discordant shedding and the FGT bacterial and viral microbiome (bacteriome and virome) during long-term ART in WLHIV remains poorly understood.
The vaginal bacteriomes of reproductive-aged women can typically be assigned to one of a few distinct community state types (CSTs), consisting mainly of different Lactobacillus species or diverse anaerobes including Gardnerella, Prevotella, Sneathia, Fannyhessea, Megasphaera, Dialister, Mobiluncus, Peptonophilus, Shuttleworthia, Parvimonas, and other genera11,12,13,14,15. Lactobacilli contribute to vaginal health by producing lactic acid, which lowers vaginal pH and creates a less favorable environment for pathogenic bacteria, and by competing with pathogens for space11. Cervicovaginal microbiome alterations have been associated with inflammation16, preterm birth17, changes to the mucosal epithelial barrier18,19, cervical cancer20,21, and sexually transmitted infections14,21,22,23. Although higher Lactobacillus relative abundance has been associated with vaginal health, many asymptomatic women have diverse vaginal microbiomes dominated by bacteria other than lactobacilli12,13,14,15. Longitudinal studies have demonstrated that the vaginal microbiome is stable over time in some women, while in others, vaginal microbiome composition fluctuates more frequently13,16.
The FGT microbiome is particularly important in HIV infection. Specific microbiome CSTs and bacterial species have been associated with increased HIV prevalence23 and acquisition14. In WLHIV, cervicovaginal microbiota have been associated with genital HIV shedding5,23,24. While pre-exposure prophylaxis (PrEP) has not been shown to significantly influence the vaginal microbiome in women without HIV25,26,27, multiple lines of evidence suggest that vaginal bacteria influence metabolism and concentrations of antiretroviral drugs in the FGT25,28,29,30,31,32,33. Whether long-term ART influences the vaginal microbiome in WLHIV is poorly understood. One study found no microbiome changes after six months of ART34, another reported higher Lactobacillus relative abundance in women with viral suppression after a year of ART35, and a third reported an association between bacterial vaginosis and ART with protease inhibitors compared to regimens without protease inhibitors36. Further study is needed to clarify the relationship between ART and the FGT microbiome over years or decades. The relationship between the microbiome and discordant genital HIV shedding is also unclear. Discordant HIV shedding, defined as detection of HIV RNA in the genital tract with an undetectable plasma viral load due to ART suppression, is incompletely understood, but has been associated with genital inflammation37, coinfections37,38, trauma39, and ulcers10, and has also been attributed to expression of integrated provirus from proliferating cells4. The importance of the microbiome in FGT health led us to consider the microbiome as a potential factor promoting discordant HIV shedding.
Compared to the bacteriome, less is known about the role of the virome in FGT health. Viruses found in the FGT include papillomaviruses, anelloviruses, herpesviruses, and bacteriophages40,41,42,43,44,45,46,47,48. Vaginal virome community structure has been linked with the bacteriome41,42,43. Vaginal virome alterations have been associated with inflammation42, cervical lesions44,45, and preterm birth46. Human papillomaviruses are the major cause of cervical cancer, the fourth most-diagnosed cancer in women worldwide in 202049. Anelloviruses, a recently discovered, prevalent viral family50, have been associated with FGT inflammation in women without HIV42 and CD4 T cell levels in WLHIV45. Little is known about the relationship between the FGT virome and long-term ART or discordant HIV shedding.
To investigate the relationship between the cervicovaginal bacteriome and virome and discordant shedding over time in WLHIV taking ART, we analyzed 125 cervicovaginal specimens collected over a two-year period from 31 women living with HIV-1 in Lima, Peru, nine of whom experienced discordant HIV shedding. We used metagenomic next-generation sequencing (NGS) to profile the bacteriome and virome in these specimens, and analyzed microbial community composition and diversity in relation to discordant shedding events (HIV detectable in the genital tract but not the plasma while on ART) vs concordant samples (all other timepoints), overall shedder or non-shedder status, and duration of ART and immune recovery. Our findings suggest that FGT bacteriome instability is associated with discordant HIV shedding, while immune recovery during ART influences the FGT virome.
Results
Shedders’ bacteriomes are less stable over time than non-shedders’ bacteriomes
To investigate the roles of both the bacteriome and virome in discordant HIV shedding, we analyzed 125 cervicovaginal specimens collected longitudinally from 31 women living with HIV-1 in Lima, Peru (Table 1, Fig. 1A, and Supplementary Data 1). Specimens were collected quarterly from participants over a period of two years as part of a larger study characterizing discordant shedding in people living with HIV4. In this study, we analyzed an average of 4 specimens per participant: specimens from the initial visit; specimens collected after 24 months; discordant shedding timepoints and immediate pre- or post-shedding timepoints for shedders; and the corresponding (equivalent) timepoints for matched non-shedders. Most specimens (n = 120) were cervicovaginal lavages (CVL). When CVL samples were not available, vaginal swabs (n = 2), cytobrushes (n = 2), or TearFlo paper strips (n = 1) were used to sample the cervicovaginal environment. Nine women experienced discordant HIV shedding during the study period (shedders) and 22 did not (non-shedders). Non-shedders were older than shedders at study enrollment (non-shedder median age = 34 years, shedder median age = 28 years; Mann–Whitney test, p = 0.047) (Supplementary Fig. 1A). Shedders and non-shedders had similar CD4 counts at enrollment and similar CD4 recoveries when values from enrollment and after 2 years of ART were compared (Mann–Whitney test; adjusted p-value = 0.94 at enrollment, adjusted p-value = 0.94 at final visit, adjusted p-value = 0.94 for change from enrollment to final visit) (Supplementary Fig. 1B). Using metagenomic NGS from centrifuged specimen pellets to characterize the bacteriome, we obtained a median of 15.6 × 106 (IQR 13.7 × 106–17.8 × 106) raw read pairs and 1.95 × 106 (IQR 8.7 × 105–4.7 × 106) quality-filtered, deduplicated individual reads per sample. Most samples clustered into groups dominated by L. crispatus (n = 29), L. iners (n = 21), Gardnerella vaginalis (n = 48), or a mixed community (n = 21), with a few samples dominated by Lactobacillus jensenii (n = 2) or Escherichia coli (n = 4) (Fig. 1B, Supplementary Data 2). There were no significant differences in bacterial alpha diversity, microbiome composition, or cluster assignment between different sample types (CVL, vaginal swabs, cytobrushes, or TearFlo strips) (Supplementary Fig. 1C–F and Supplementary Data 3).
A Timeline of samples analyzed from each individual. Non-shedder samples are shown as blue circles, shedder samples from concordant timepoints are shown as red circles, and shedder samples from discordant shedding timepoints are shown as black triangles. ART antiretroviral therapy. B Samples grouped by microbiome community cluster. Samples are shown as vertical bars colored by species relative abundance. Metadata is shown as rows of colored squares beneath the samples. By multinomial logistic regression, community clusters were not associated with ART duration or shedder/non-shedder status, except for cluster 3, which was associated with non-shedder status (two-sided, adjusted p-values in Supplementary Data 3). Cluster 3 consists of two samples from the same non-shedder individual.
Intrapersonal microbiome variation (beta diversity) was lower than interpersonal variation, suggesting intraindividual stability over time (Fig. 2A; Bray–Curtis dissimilarity; two-sided permutation test, Z = 10.24 and adjusted p-value < 0.0001 for non-shedders, Z = 3.77 and adjusted p-value = 0.00051 for shedders). Shedders had higher intrapersonal microbiome beta diversity than non-shedders (Fig. 2A; Bray–Curtis dissimilarity, two-sided permutation test, Z = −2.92 and adjusted p-value = 0.0070). We tested the hypothesis that perturbation was most pronounced during discordant shedding by comparing diversity between discordant and concordant timepoints within individuals (Fig. 2B, left). Beta diversity was significantly higher between samples when one timepoint was discordant and the other was concordant, compared to pairs of concordant timepoints (Bray–Curtis dissimilarity; two-sided permutation test, Z = 2.67 and adjusted p-value = 0.011), suggesting that higher intrapersonal variation in shedders reflects dysbiosis during discordant shedding rather than prolonged fluctuations. When we compared the equivalent timepoints from matched non-shedder individuals, we did not observe this difference (Bray–Curtis dissimilarity; two-sided permutation test, Z = −0.24 and adjusted p-value = 0.81) (Fig. 2B, right). Principal coordinates analysis (PCoA) using Bray–Curtis dissimilarity showed no association between microbiome composition and ART duration (PERMANOVA, F(1) = 0.69, p = 0.33, R2 = 0.0056) or shedder/non-shedder status (PERMANOVA, F(1) = 1.5, p = 0.34, R2 = 0.012) (Fig. 2C), although discriminant analysis with MaAsLin251 showed higher Mycoplasma hominis relative abundance in shedders compared to non-shedders (Fig. 2D, adjusted p-value = 0.048). PCoA of shedder samples showed no difference between discordant and concordant timepoints (Fig. 2E; PERMANOVA, F(1) = 0.35, p = 0.85, R2 = 0.0099). Microbiome community clusters were not associated with shedder/non-shedder status, ART duration, or discordant shedding timepoints (Fig. 1B and Supplementary Data 3). These results suggest that no specific microbiome composition is associated with shedder/non-shedder status, ART duration, or discordant shedding events, implying that higher variation during discordant shedding reflects individual-specific, not shared, dysbioses. To determine whether this could be explained by common functional characteristics encoded by bacterial genomes that might not be reflected in taxonomic profiles, we performed functional profiling with HUMAnN352 (Supplementary Fig. 2A). We observed similar results for bacterial metabolic pathway and gene family beta diversity as for taxonomic beta diversity (Supplementary Fig. 2B–I). One hundred seventy-two metabolic pathways were associated with microbiome community clusters but not with shedder/non-shedder status, discordant shedding events, or ART duration (MaAsLin2, Supplementary Data 4). Results were similar for gene families (Supplementary Data 5). This suggests that functional profiles mirror taxonomic profiles, with similar beta diversity dynamics during discordant HIV shedding.
A Bacterial species Bray–Curtis dissimilarity between pairs of samples from the same individual (within-individual comparisons; n = 134 pairwise comparisons for non-shedder samples and n = 58 pairwise comparisons for shedder samples) or from different individuals (between-individual comparisons; n = 3694 pairwise comparisons for non-shedder samples and n = 608 pairwise comparisons for shedder samples). In this and panel B, statistical significance was assessed using a two-sided, two-sample permutation test, and p-values were adjusted for multiple comparisons using the Benjamini–Hochberg method. In this and B, D, center lines show the median, box bounds show the interquartile range, and whiskers show the minimum and maximum. B Bacterial species Bray–Curtis dissimilarity between pairs of concordant samples within the same shedder individual, or sample pairs where one sample was from a discordant shedding timepoint and the other was from a concordant timepoint within the same individual. The corresponding timepoints from matched non-shedder individuals serve as a control. C PCoA of shedder and non-shedder samples, using Bray–Curtis dissimilarity calculated from bacterial species relative abundance. Samples are colored by duration of ART (antiretroviral therapy). Non-shedder samples are shown as circles, and shedder samples are shown as triangles. Statistical significance was determined by PERMANOVA, p = 0.33 for ART duration and p = 0.34 for shedder/non-shedder status. D Mycoplasma hominis relative abundance in non-shedders and shedders. Statistical significance was determined by MaAsLin2, adjusted two-sided p-value = 0.048. E PCoA of shedder samples, using Bray–Curtis dissimilarity calculated from bacterial species relative abundance. Discordant shedding timepoints are colored brown, and concordant timepoints are colored green. Statistical significance was determined by PERMANOVA, p = 0.85.
Non-shedders’ viromes are less stable than shedders’ viromes, with anellovirus and papillomavirus relative abundance changing over time
Given the temporal dynamics of the bacteriome, we examined the virome for similar patterns. After enrichment for virus-like particles (VLPs), we visualized VLPs in specimens using epifluorescence microscopy to corroborate VLP enrichment (Supplementary Fig. 3A). VLP concentrations ranged from 2.5 × 105 to 2.2 × 108 particles per mL of VLP-enriched supernatant, with a median of 1.1 × 107 particles per mL (Supplementary Fig. 3B). Using metagenomic NGS of nucleic acids extracted from VLP-enriched supernatants and amplified by multiple displacement amplification (MDA), we obtained a median of 1.6 × 107 (IQR 4.5 × 106–2.2 × 107) raw read pairs and 2.4 × 106 (IQR 1.1 × 106–3.5 × 106) quality-filtered, deduplicated individual reads per sample. 12,574 unique contigs ≥ 500 nt were assembled from the samples and sequencing controls, and 1596 were identified as viral. A median of 89.6% of quality-filtered reads per sample could be mapped back to the viral contigs (IQR 72.6–92.5%). One hundred sixteen samples had >1000 viral reads and were retained for further analysis. There were no significant differences in virome diversity metrics or overall virome composition between specimen types (CVL, cytobrush, vaginal swab, or TearFlo strip) (Supplementary Fig. 4A–D). Samples clustered into groups dominated by Anelloviridae (n = 46), Papillomaviridae (n = 54), or bacteriophages including Microviridae and Caudoviricetes (n = 16) (Fig. 3A and Supplementary Data 6–8). Virome alpha diversity was inversely associated with ART duration (linear mixed modeling (LME), t(84) = −6.9 and p = 7.7 × 10−10 for viral contig richness, t(84) = −5.9 and p = 8.3 × 10−8 for Shannon index), with no significant difference between shedders and non-shedders (LME, t(29) = 1.3 and p = 0.19 for richness, t(29) = 0.12 and p = 0.90 for Shannon index) (Fig. 3B). Because MDA only amplifies DNA, we performed sequence-independent amplification (SIA) on the viral nucleic acids to amplify RNA and DNA. We obtained a median of 1.6 × 107 raw read pairs per sample by SIA (IQR 2.5 × 106–2.0 × 107 read pairs), and a median of 5.2 × 105 clean, deduplicated individual reads (IQR 1.6 × 105–1.4 × 106 individual reads). A pilot investigation of contigs assembled from SIA sequencing data in 39 samples found mostly DNA viruses and a few plant-infecting RNA viruses. Therefore, we employed a more sensitive read-based search against the NCBI RefSeq database. We found plant-infecting RNA viruses in a few samples, but no RNA viruses known to infect humans or bacteria, and most samples had no RNA viruses identified (Supplementary Fig. 5A). Therefore, we focused our analyses on the DNA virome using the MDA data.
A Samples grouped by DNA virome community cluster. Samples are shown as vertical bars colored by family relative abundance. Metadata is shown as rows of colored squares beneath the samples. ART antiretroviral therapy. B DNA virome alpha diversity (viral contig richness and Shannon index) by duration of ART. Shedder and non-shedder samples are plotted together. Purple lines indicate the linear regression lines, and 95% confidence bands are shown in gray. Statistical significance was assessed by LME, p = 7.7 × 10−10 for richness and p = 8.3 × 10−8 for Shannon index. C Viral Bray–Curtis dissimilarity (weighted) between pairs of samples from the same individual (within-individual comparisons; n = 116 pairwise comparisons for non-shedder samples and n = 49 pairwise comparisons for shedder samples) or from different individuals (between-individual comparisons; n = 3205 pairwise comparisons for non-shedder samples and n = 512 pairwise comparisons for shedder samples). In this and D statistical significance was assessed using a two-sided, two-sample permutation test and p-values were adjusted for multiple comparisons using the Benjamini–Hochberg method. In this and D center lines show the median, box bounds show the interquartile range and whiskers show the minimum and maximum. D Viral Sorensen dissimilarity (unweighted) between pairs of samples from the same individual (within-individual comparisons; n = 116 pairwise comparisons for non-shedder samples and n = 49 pairwise comparisons for shedder samples) or from different individuals (between-individual comparisons; n = 3205 pairwise comparisons for non-shedder samples and n = 512 pairwise comparisons for shedder samples). E PCoA of shedder and non-shedder samples, using Bray–Curtis dissimilarity calculated from virus contig relative abundance. Samples are colored by duration of ART. Non-shedder samples are shown as circles, and shedder samples are shown as triangles. Statistical significance was determined by PERMANOVA, p = 0.001 for ART duration and p = 0.016 for shedder/non-shedder status. F Anelloviridae relative abundance (left) and Papillomaviridae relative abundance (right) by ART duration. Shedder and non-shedder samples are plotted together. Purple lines indicate the linear regression lines, and 95% confidence bands are shown in gray. Statistical significance was determined using MaAsLin2, adjusted two-sided p-value = 8.8 × 10−5 for Anelloviridae and adjusted p-value = 0.037 for Papillomaviridae.
Like bacteriome variation, intrapersonal virome variation (beta diversity) was significantly lower than interpersonal variation (Fig. 3C, Bray–Curtis dissimilarity (weighted); two-sided permutation test, Z = 23.21 and adjusted p-value < 0.0001 for non-shedders, Z = 13.66 and adjusted p-value < 0.0001 for shedders; Fig. 3D, Sorensen dissimilarity (unweighted); two-sided permutation test, Z = 7.89 and adjusted p-value < 0.0001 for non-shedders, Z = 4.17 and adjusted p-value < 0.0001 for shedders). However, in contrast to the bacteriome, weighted intrapersonal virome variation was not significantly different between shedders and non-shedders (Fig. 3C, Bray–Curtis dissimilarity; two-sided permutation test, Z = −0.35 and adjusted p-value = 0.73). Further, unweighted intrapersonal virome variation was significantly higher in non-shedders than in shedders (Sorensen dissimilarity; two-sided permutation test, Z = 3.02 and adjusted p-value = 0.0024) (Fig. 3D). In weighted (Bray–Curtis) PCoA, virome composition was associated with ART duration (PERMANOVA, F(1) = 1.9, p = 0.001, R2 = 0.016) and shedder/non-shedder status (PERMANOVA, F(1) = 1.6, p = 0.016, R2 = 0.014) (Fig. 3E). Because ART duration trended toward interaction with shedder/non-shedder status (PERMANOVA, p = 0.069), we carried out PCoA on shedders and non-shedders separately (Supplementary Fig. 6A). Virome composition was associated with ART duration in non-shedders (PERMANOVA, p = 0.001), and trended toward association with ART duration in shedders but was not significant (PERMANOVA, p = 0.072). In unweighted (Sorensen) PCoA of combined shedders and non-shedders, virus presence-absence was associated with ART duration (PERMANOVA, F(1) = 5.8, p = 0.001, R2 = 0.048) but not shedder/non-shedder status (PERMANOVA, F(1) = 2.6, p = 0.63, R2 = 0.021), with no interaction between ART duration and shedder/non-shedder status (PERMANOVA, p = 0.15) (Supplementary Fig. 6B). These results suggest that, unlike the bacteriome, virome composition changes in a specific trajectory over time in WLHIV taking ART, with more pronounced changes among non-shedders.
To identify viruses responsible for changes in virome beta diversity over time, we used differential abundance analysis (MaAsLin251). Anelloviridae relative abundance was inversely associated with ART duration (MaAsLin2, adjusted p-value = 8.8 × 10−5), while Papillomaviridae relative abundance increased over time (MaAsLin2, adjusted p-value = 0.037) (Fig. 3F). ART duration trended toward interaction with shedder/non-shedder status for several viral families, but these trends were not significant (Supplementary Data 9). When non-shedders were analyzed separately, Anelloviridae relative abundance decreased over time after ART initiation, and Papillomaviridae relative abundance increased (Supplementary Fig. 6C). No taxa changed significantly over time in shedders. These results suggest that changes in virome composition after ART initiation are driven by Anelloviridae and Papillomaviridae, especially among non-shedders.
Anellovirus communities change over time, with decreased alphatorquevirus abundance and increased betatorquevirus proportions
Because the Anelloviridae family contributed significantly to the temporal virome signature (Fig. 3), we performed phylogenetic analysis of the anellovirus contig ORF1 gene sequences to better define anellovirus community dynamics. We obtained predicted ORF1 sequences from 751/799 anellovirus contigs. 21 contigs had more than one predicted ORF1 sequence, for a total of 774 sequences. At the genus level, most anelloviruses were classified as alphatorqueviruses (n = 339), betatorqueviruses (n = 339), or gammatorqueviruses (n = 85), with a few hetorqueviruses (n = 4) and sequences falling outside known anellovirus genera (n = 7) (Fig. 4A). Compared to all viral genera, relative abundance of alphatorqueviruses and incomplete anellovirus genomes (i.e., without ORF1) decreased over time after ART initiation (MaAsLin2, adjusted p-value = 0.0019 and adjusted p-value = 0.0019) (Supplementary Fig. 7A). There were no significant changes in the relative abundance of other anellovirus genera. This suggests that decreased Anelloviridae relative abundance reflects decreased Alphatorquevirus relative abundance.
A Anellovirus ORF1 amino acid phylogeny. Colored tips (red, blue, orange, green, and pink) represent sequences generated in this study. Reference sequences are shown in black. B Weighted anellovirus UniFrac distance between pairs of samples from the same individual (within-individual comparisons; n = 108 pairwise comparisons for non-shedder samples and n = 49 pairwise comparisons for shedder samples) or from different individuals (between-individual comparisons; n = 2973 pairwise comparisons for non-shedder samples and n = 512 pairwise comparisons for shedder samples). Statistical significance was assessed using a two-sided, two-sample permutation test, and p-values were adjusted for multiple comparisons using the Benjamini–Hochberg method. The center lines show the median, box bounds show the interquartile range, and whiskers show the minimum and maximum. C PCoA of shedder and non-shedder samples, using weighted anellovirus UniFrac distance. Samples are colored by duration of ART (antiretroviral therapy). Non-shedder samples are shown as circles, and shedder samples are shown as triangles. Statistical significance was assessed by PERMANOVA, p = 0.005 for ART duration and p = 0.81 for shedder/non-shedder status. D Cervicovaginal Anelloviridae NGS relative abundance vs plasma CD4 counts. Shedder and non-shedder samples are plotted together. The purple line indicates the linear regression line, and 95% confidence bands are shown in gray. Statistical significance was determined by LME, p = 2.1 × 10−4. E Cervicovaginal alphatorquevirus qPCR copy numbers vs plasma CD4 counts. Copy numbers are expressed in copies/μL of extracted total nucleic acid (TNA). Shedder and non-shedder samples are plotted together. The purple line indicates the linear regression line, and 95% confidence bands are shown in gray. The dotted horizontal line separates samples with detectable and undetectable alphatorquevirus by qPCR. Statistical significance was assessed by LME, p = 0.04. To facilitate visualization on a log scale, a pseudocount of 0.01 copies/μL was added to all alphatorquevirus qPCR copy numbers, but LME analyses were based on the raw copy numbers.
We used the anellovirus phylogeny to calculate UniFrac distance, a beta diversity metric that accounts for phylogenetic distance (Supplementary Data 10 and 11). In shedders, intrapersonal anellovirus beta diversity was significantly lower than interpersonal diversity (weighted UniFrac, two-sided permutation test, Z = 3.4 and adjusted p-value = 0.00080), while for non-shedders, there was no significant difference between intrapersonal and interpersonal anellovirus beta diversity (weighted UniFrac, two-sided permutation test, Z = 1.3 and adjusted p-value = 0.21) (Fig. 4B). Intrapersonal variation was also significantly lower in shedders than non-shedders (Fig. 4B; two-sided permutation test, Z = 4.1 and adjusted p-value = 0.0001). Results were similar for unweighted UniFrac distance (Supplementary Fig. 7B). These results suggest that anellovirus communities are less stable over time in non-shedders compared to shedders. In PCoA using weighted anellovirus UniFrac distance, anellovirus community compositions changed significantly by ART duration (PERMANOVA, F(1) = 3.5, p = 0.005, R2 = 0.030), with no significant difference between shedders and non-shedders (PERMANOVA, F(1) = 2.4, p = 0.81, R2 = 0.020) (Fig. 4C). Because ART duration trended toward interaction with shedder/non-shedder status (PERMANOVA, p = 0.078), we carried out separate PCoA analyses on non-shedder and shedder samples. In these separate analyses, ART duration was associated with anellovirus community structure in non-shedders (PERMANOVA, F(1) = 3.6, p = 0.006, R2 = 0.044), but not in shedders (PERMANOVA, F(1) = 1.2, p = 0.12, R2 = 0.035) (Supplementary Fig. 7C). In unweighted anellovirus UniFrac PCoA of combined shedders and non-shedders, anellovirus presence-absence was associated with ART duration (PERMANOVA, F(1) = 6.0, p = 0.001, R2 = 0.051) but not with shedder/non-shedder status (PERMANOVA, F(1) = 1.8, p = 0.59, R2 = 0.015), with no interaction between ART duration and shedder/non-shedder status (PERMANOVA, p = 0.31) (Supplementary Fig. 7D). We performed differential abundance analyses (MaAsLin2) to identify specific anelloviruses responsible for changing anellovirus diversity over time. When shedders and non-shedders were analyzed together, Betatorquevirus relative abundance increased, and relative abundance of partial anellovirus genomes decreased relative to other anellovirus genera (Supplementary Fig. 7E; MaAsLin2, adjusted p-value = 0.036 and adjusted p-value = 0.036). When non-shedders were considered separately, Betatorquevirus relative abundance increased (MaAsLin2, adjusted p-value = 0.028) (Supplementary Fig. 7F) and Alphatorquevirus relative abundance tended to decrease but did not reach significance (MaAsLin2, adjusted p-value = 0.060). No anellovirus genus changed significantly over time in shedders. Together with the whole-virome analyses, where Alphatorquevirus relative abundance decreased over time in the virome as a whole, these results suggest that betatorqueviruses consequently make up a larger proportion of the remaining anellovirus community. While these temporal changes are more pronounced among non-shedders, there is no specific anellovirus community profile or temporal signature associated with shedders.
We corroborated the NGS data with quantitative PCR (qPCR) specific to the untranslated region of the alphatorquevirus genome (Supplementary Fig. 8Α). Alphatorquevirus genome copy numbers were correlated with Alphatorquevirus NGS relative abundance (linear regression, t(114) = 2.1, p = 0.041, multiple R2 = 0.036) and trended toward association with Anelloviridae relative abundance (linear regression, t(114) = 1.7, p = 0.085, multiple R2 = 0.026), but not Betatorquevirus or Gammatorquevirus relative abundance (linear regression, t(114) = 0.046, p = 0.96, multiple R2 = 1.8 × 10−5 and t(114) = −0.20, p = 0.84, multiple R2 = 3.6 × 10−4, respectively) (Supplementary Fig. 8Β). Alphatorquevirus copy numbers tended toward inverse association with ART duration, but this was not significant (LME, t(84) = −1.7, p = 0.090) (Supplementary Fig. 8C), and there was no significant difference between shedders and non-shedders (LME, t(29) = −0.67, p = 0.51).
Because anellovirus abundance has previously been associated with immune status45,53,54,55,56,57,58,59,60, we considered whether increasing CD4 counts could explain decreasing anellovirus relative abundance. Seventy-eight specimens (67%) had CD4 count data available for analysis. CD4 counts were inversely associated with Anelloviridae relative abundance (LME, t(46) = −4.2, p = 1.1 × 10−4) (Fig. 4D), Alphatorquevirus relative abundance (Supplementary Fig. 8D, LME, t(46) = −4.1, p = 1.7 × 10−4) and alphatorquevirus qPCR copy numbers (Fig. 4E; LME, t(46) = −2.1, p = 0.04). There was no association between CD4 counts and Betatorquevirus or Gammatorquevirus relative abundance (LME, t(46) = 0.30 and p = 0.77, t(46) = −0.92 and p = 0.36, respectively) (Supplementary Fig. 8D). When we analyzed non-shedders separately, Anelloviridae NGS relative abundance, Alphatorquevirus NGS relative abundance and alphatorquevirus qPCR copy numbers were inversely associated with CD4 counts (LME; t(35) = −4.1 and p = 2.1 × 10−4, t(35) = −4.29 and p = 1.3 × 10−4, t(35) = −2.1 and p = 0.04, respectively), but there were no significant associations in shedders (Supplementary Fig. 8E). These results suggest a relationship between immune status and anellovirus abundance, specifically alphatorqueviruses, in the FGT.
Papillomavirus community diversity is stable over time in shedders and non-shedders
Because the Papillomaviridae family also contributed to the temporal virome signature (Fig. 3), we conducted phylogenetic analysis of the concatenated papillomavirus contig E1, E2, L2, and L1 genes to characterize papillomavirus diversity over time. We obtained E1–E2–L2–L1 sequences from 169/190 papillomavirus contigs. Most were alphapapillomaviruses (n = 128), with some betapapillomaviruses (n = 7) and gammapapillomaviruses (n = 34) (Fig. 5A). Alphapapillomavirus relative abundance increased over time (Supplementary Fig. 9A, MaAsLin2, adjusted p-value = 0.044), while other papillomavirus genera did not change significantly.
A Papillomavirus E1–E2–L2–L1 nucleotide phylogeny. Colored tips (red, blue, and orange) represent sequences generated in this study. Reference sequences are shown in black. B Weighted papillomavirus UniFrac distance between pairs of samples from the same individual (within-individual comparisons; n = 108 pairwise comparisons for non-shedder samples and n = 49 pairwise comparisons for shedder samples) or from different individuals (between-individual comparisons; n = 2973 pairwise comparisons for non-shedder samples and n = 512 pairwise comparisons for shedder samples). Statistical significance was assessed using a two-sided, two-sample permutation test, and p-values were adjusted for multiple comparisons using the Benjamini–Hochberg method. The center lines show the median, box bounds show the interquartile range, and whiskers show the minimum and maximum. C PCoA of shedder and non-shedder samples, using weighted UniFrac distance calculated from papillomavirus phylogeny and relative abundance. Samples are colored by duration of ART (antiretroviral therapy). Non-shedder samples are shown as circles and shedder samples are shown as triangles. Statistical significance was determined by PERMANOVA, p = 0.33 for ART duration and p = 0.28 for shedder/non-shedder status.
Unlike anellovirus beta diversity, both non-shedders and shedders had lower intrapersonal papillomavirus beta diversity compared to interpersonal diversity, with no significant difference between shedders and non-shedders (Fig. 5B and Supplementary Data 12; weighted papillomavirus UniFrac distance, two-sided permutation test, Z = 6.7 and adjusted p-value < 0.0001, Z = 2.9 and adjusted p-value = 0.0051, Z = 0.51 and adjusted p-value = 0.61, respectively). Unweighted papillomavirus beta diversity results were similar (Supplementary Fig. 9B and Supplementary Data 13). In PCoA using papillomavirus UniFrac distance, there were no significant differences by ART duration or shedder/non-shedder status (Fig. 5C and Supplementary Fig. 9C; PERMANOVA; weighted, F(1) = 0.84, p = 0.33, R2 = 0.0074 and F(1) = 2.2, p = 0.28, R2 = 0.020, respectively; unweighted, F(1) = 1.1, p = 0.18, R2 = 0.0095 and F(1) = 1.6, p = 0.10, R2 = 0.014, respectively). These results suggest that papillomavirus diversity is stable over time in shedders and non-shedders.
Discussion
We identified intraindividual and longitudinal patterns in the cervicovaginal bacteriome and virome of WLHIV with and without discordant HIV shedding. Specifically, the bacteriome was less stable over time in shedders. Virome composition changed significantly in association with ART duration, with more pronounced changes among non-shedders.
Higher intraindividual bacterial beta diversity in shedders suggests more frequent changes in microbiome composition compared to non-shedders. This was driven by shedding timepoints, suggesting that microbiome alterations may be associated with shedding events. By contrast, lower intraindividual diversity in non-shedders suggests stability over time. No specific microbiome composition was associated with shedders or shedding events, suggesting that stability over time is more important than specific taxa in the context of discordant shedding. This is interesting because previous studies have often focused on specific microbiome community states or bacterial taxa as markers of FGT health11. Previous analysis of this cohort linked discordant shedding with proliferation of infected cells, rather than HIV replication4. Other studies have identified associations between cervicovaginal microbiota, inflammation, and T cell populations in the FGT14,61. Thus, the microbiome may contribute to discordant shedding by influencing the number of infected cells in the FGT, which may proliferate in response to disruption of the FGT. Alternatively, the microbiome may be influenced by other factors (trauma through intercourse, inflammation, etc.) that also contribute to discordant shedding. Because specimens were collected every three months, episodes of discordant shedding could have occurred between study visits without being detected, including in women in the non-shedder group. Importantly, changes in bacterial beta diversity were directly linked to discordant shedding timepoints (Fig. 2B) and not solely to shedder/non-shedder status, making it less likely that undetected shedding timepoints significantly influenced the results.
Unlike the bacteriome, which was stable over time in non-shedders, the virome was less stable over time, particularly in non-shedders. The temporal instability quantified by overall virome beta diversity metrics was consistent with longitudinal changes we identified in specific viral taxa. We found that Anelloviridae and Alphatorquevirus relative abundance decreased, and overall anellovirus community diversity changed during two years of ART. Anellovirus relative abundance and alphatorquevirus absolute abundance measured by qPCR were inversely associated with plasma CD4 counts. Anelloviruses are prevalent in healthy adult humans and are not known to be pathogenic50. Interestingly, numerous studies have associated anellovirus loads with immune suppression, including in people living with HIV45,53,54,55,56,57,58,59 and in organ transplant recipients50,60, suggesting that anellovirus replication may be controlled by host immunity, although the specific mechanism is poorly understood50. Thus, the anellovirus dynamics that we observed may reflect participants’ immune recovery during two years of ART. The fact that the associations were more pronounced among non-shedders may be a consequence of sampling size, or may suggest underlying immune dysregulation in shedders.
While Papillomaviridae relative abundance increased over time, papillomavirus diversity did not change significantly. Because relative abundance data is compositional, papillomavirus relative abundance may have been inflated when anellovirus relative abundance decreased, with no underlying change in papillomavirus viral load. Although Phi29 polymerase may preferentially amplify small circular single-stranded DNA62, our alphatorquevirus qPCR absolute copy numbers were consistent with NGS relative abundance, suggesting that our results represent an actual biological signature. In an effort to define absolute viral loads in the specimens, we used epifluorescence microscopy to measure concentrations of VLPs. Because several CVL specimens had measured VLP concentrations lower than the negative controls (Supplementary Fig. 3B), we chose to treat this data conservatively and did not incorporate VLP concentrations into any of the statistical models. Besides plant virus reads identified in a few samples, we did not find any RNA viruses by SIA, including HIV. Data from the sequencing positive control (rhinovirus B14 RNA) indicated that the protocol itself was effective (Supplementary Fig. 5). Therefore, limited detection of RNA viruses in the samples suggests that the quantity or integrity of viral RNA in the samples was limited, either because the initial quantities were low or because of RNA degradation during storage. Other than the timepoints at the beginning of the study (before ART treatment), HIV RNA copy numbers measured at shedding timepoints were low (30–721 copies/mL of CVL, limit of detection = 30 copies/mL). Further, specimens were not DNase-treated before qRT-PCR4. Therefore, HIV DNA from host cells could have been detected by the qRT-PCR assay, but is less likely to have been detected by our sequencing protocol. Finally, RNA can degrade over time even when stored at –80 °C63. The aliquots used for bacteriome and virome analysis were stored for 10–14 years at −80 °C before VLP preparation and TNA extraction. Therefore, some degree of RNA degradation may have occurred and contributed to limited detection of RNA viruses.
Our cohort consisted of ART-naïve women from a clinical setting who were initiating government-sponsored ART in Lima, Peru. Because all the study participants received ART, we cannot directly compare our results with longitudinal data from WLHIV not receiving ART. While this comparison would be biologically informative, such a study would be incompatible with current treatment standards. For our cohort, data were not available on sexual activity in the days prior to sample collection, which can affect the FGT microbiome64,65,66. To our knowledge, only a few studies have analyzed the cervicovaginal bacteriome or virome in women from Peru67,68,69. Therefore, our data contributes valuable information about the FGT bacteriome and virome in women from this region. Prevalence of vaginal microbiome community states varies between cohorts, with some studies reporting differences between women from different geographic regions or different racial and ethnic backgrounds11,12,13,14,15,20,70,71,72. Therefore, studying diverse groups of WLHIV worldwide would indicate whether our results are generalizable to other populations.
This longitudinal analysis of the cervicovaginal virome and bacteriome in WLHIV identifies intrapersonal and temporal patterns associated with discordant HIV shedding and ART duration. These findings support further investigation of the role of bacterial communities in discordant HIV shedding in women with suppressed plasma viral loads, as well as the role of anelloviruses during immune suppression and recovery.
Methods
Specimens
This study was approved by the Institutional Review Boards of the Hospital Dos de Mayo (Lima, Peru), Seattle Children’s Hospital (Seattle, Washington, USA), and Arizona State University (Tempe, Arizona, USA). Written informed consent was provided by all participants. Cervicovaginal specimens were collected as part of an 18–24-month study evaluating low-level viremias (aka “blips”) and discordant genital shedding in ART-naïve men and women with HIV-1 from Lima, Peru, between July 2007 and February 20094,73. Participants were eligible if they would be receiving care at Hospital Dos de Mayo for the next two years and were willing to provide cervicovaginal lavage specimens quarterly. First-line, non-nucleoside reverse transcriptase inhibitor-based ART was prescribed at enrollment (stavudine (d4T) or zidovudine (ZDV), lamivudine (3TC), and nevirapine (NVP) or efavirenz (EFV)). Two participants (one shedder and one non-shedder) were prescribed atazanavir (ATV) instead of NVP or EFV. Specimens were collected at enrollment (i.e., before participants began ART) and every three months over a two-year period for each participant, i.e., the subsequent study duration was equivalent to the ART duration. Specimen collection was timed to avoid days when participants were menstruating, with specimens generally collected during the follicular phase of the menstrual cycle. Plasma CD4 counts were measured every six months. Data on sexual activity and antibiotic usage prior to sample collection were not recorded as part of this study. CVL were collected by washing the uterine cervix with sterile 1× phosphate-buffered saline (PBS) and collecting the 1× PBS from the vaginal fornix as described in ref. 74. Plasma and CVL HIV RNA were quantified by qRT-PCR as described in ref. 75. Discordant shedding was assessed in participants with ART suppression, defined as median plasma HIV RNA below the limit-of-quantification (1.48 log10 copies(c)/mL), without virologic failure (HIV RNA > 3.0 log10c/mL), who had HIV RNA detected in the CVL (1.48 log10 copies(c)/mL) for the women. All other timepoints were considered to be concordant, including initial timepoints with HIV detectable in both plasma and CVL; later timepoints with no HIV detectable in plasma or CVL; and later timepoints with intermittent low-level plasma viremia (1.48–3.0 log10 copies/mL; 3 women, 4 timepoints). Nine women experienced discordant HIV shedding (shedders), and 22 women did not experience discordant shedding (non-shedders). Participant median age at enrollment was 28 years for shedders and 34 years for non-shedders. To select timepoints for bacteriome and virome analysis, individual shedders were matched with individual non-shedders based on average CD4 counts over the duration of ART (2–4 non-shedders for each shedder), so that equivalent timepoints could be analyzed for shedders and non-shedders. We analyzed the bacteriome and virome at enrollment (month 3 for one shedder who did not have an enrollment specimen available); for shedders, discordant shedding timepoints and immediate pre- or post- shedding timepoints, as available; the corresponding timepoints for matched non-shedders; and at 24 months (month 21 for two shedders who did not have month 24 specimens available) (Fig. 1A). When CVL specimens were not available, we analyzed other available specimen types: vaginal swabs (n = 2), cytobrushes (n = 2), or TearFlo filter paper strips (HUB Pharmaceuticals; n = 1). All specimen types were stored at −80 °C until processing, with no additional media added to swabs, cytobrushes, or TearFlo strips during storage.
Sample processing
Specimens were randomized into batches for processing using a random number generator. Specimens were thawed on ice and mixed by pipetting. Up to 800 μL of CVL were centrifuged for 10 min at 7000×g at 4 °C. The supernatants were used immediately for virus-like particle enrichment and virome sequencing, and the pellets were frozen at -80 °C until used for bacteriome sequencing. PBS negative controls (800 μL) were processed alongside the samples through bacteriome and virome sequencing to assess contamination. In the case of vaginal swabs, cytobrushes, or TearFlo strips, 1 mL of 1× PBS was added to the specimen and vortexed for 30 s, and 800 μL was then processed in the same way as the CVLs.
Bacteriome sequencing
Frozen bacterial pellets (CVL samples and 1× PBS controls) were thawed on ice. DNA was extracted with the DNeasy PowerSoil Pro kit (Qiagen; Germantown, MD), following the manufacturer’s instructions modified to include 775 μL of Solution CD1, 20 min bead beating in a TissueLyser II, 3 min of centrifugation at 16,000×g to dry the spin column membrane, and elution in 60 μL of Solution C6. DNA concentrations were measured on a NanoDrop spectrophotometer, and 5–25 μL of DNA (approximately 100–500 ng DNA) per sample was used as input for the Illumina DNA Prep library kit, following the manufacturer’s protocol with six cycles for the indexing PCR. Two microliters of ZymoBIOMICS Microbial Community DNA Standard were used to build a positive control library. Libraries were pooled, quantified with a Qubit fluorometer, and sequenced (paired-end, 2 × 150 bp) on a NextSeq 1000/2000 sequencer (Illumina; San Diego, CA). Depending on the number of reads obtained per library, some libraries were re-sequenced and the results concatenated to obtain a minimum of 10 million read pairs per sample.
Virus-like particle enrichment and virome sequencing
CVL supernatants and PBS controls were passed through a 0.2-μm pore filter to exclude bacterial and eukaryotic cells. Filtrates were treated with benzonase and Baseline-ZERO DNase to digest non-encapsidated nucleic acids. Total nucleic acids (TNA) were then extracted from 500 μL of the treated filtrates on the bioMérieux eMAG, with PBS added as needed if filtrates were below the required volume. Viral DNA was amplified by MDA using Phi29 DNA polymerase (GenomiPhi V2 (Cytiva; Marlborough, MA)), with the kit’s lambda phage DNA for a positive control. The amplified DNA was variably diluted (up to 1:10 depending on band brightness after gel electrophoresis) and used as input for the Illumina DNA Prep library kit, with six cycles for the indexing PCR. Libraries were individually quantified with a Qubit fluorometer, pooled, and sequenced on an Illumina NextSeq 2000 sequencer (paired-end, 2 × 150 bp). Separately, we performed SIA of viral RNA and DNA as previously described76. First-strand cDNA was synthesized with SuperScript IV reverse transcriptase (ThermoFisher; Waltham, MA) and primers 5’-GTTTCCCAGTCACGATCNNNNNNNNN-3’ and 5’-GTTTCCCAGTCACGATC-3’, second strand DNA was synthesized with DNA Polymerase I Klenow fragment (New England Biolabs; Ipswich, MA), and amplification was performed with AccuPrime Taq high fidelity polymerase (ThermoFisher; Waltham, MA) and primer 5’-GTTTCCCAGTCACGATC-3’. We used cell-culture-derived rhinovirus B14 RNA as a positive control. The SIA products were used undiluted as input for the DNA Prep library kit (Illumina; San Diego, CA) and sequenced in the same way as the MDA products.
Virus-like particle enumeration by fluorescence microscopy
Unextracted, nuclease-treated filtrates remaining from the virus-like particle (VLP) enrichment were frozen at −80 °C until processed for VLP enumeration by fluorescence microscopy, using an adaptation of previously described methods76,77. Fifteen-100 μL of CVL filtrate and PBS negative controls were mixed with 10 mL sterile SM buffer and filtered onto a 0.02-μm pore, 25 mm diameter Anodisc aluminum oxide filter (Cytiva; Marlborough, MA). De-identified, remnant stool specimens (ASU IRB record STUDY00011967) (resuspended in SM buffer, 0.45-μm and 0.2-μm filtered, and nuclease treated) and myxoma virus with a known titer were used as positive controls. Filters were dried, stained with 25× SYBR Gold (ThermoFisher; Waltham, MA), and mounted on slides in 15 μL of glycerol/PBS buffer with 0.1% ascorbic acid (working stock: 1 mL glycerol, 980 μL PBS, and 20 μL 10% ascorbic acid). Slides were visualized at 60× with oil immersion on an AX R confocal microscope (Nikon; Tokyo, Japan) with a 488-nm wavelength laser. Five fields were randomly selected per sample, and NIS-Elements software (Nikon; Tokyo, Japan) was used to count particles between 50 nm and 500 nm in size. VLP counts for each sample were averaged across the number of fields viewed and used to calculate VLP concentrations in the CVL filtrates and controls.
Bacteriome analysis
Bacterial sequencing reads were quality-filtered, sequencing adapters, PhiX, and human sequences removed, reads deduplicated, and paired reads merged using BBTools (v. 38.38)78. Both merged and unmerged reads were retained for downstream analysis. Taxonomy was assigned to the clean reads with KrakenUniq (v. 0.7.3)79 using a custom database (RefSeq archaeal, bacterial, viral, plasmid, and fungal genomes, plus the GRCh38 human genome; downloaded in January 2022). To reduce false positives, we masked taxa that were assigned fewer than 1800 unique kmers per million reads. For downstream analyses, we parsed the read counts at the species level, i.e., we included reads that could be classified at the species level or lower and excluded reads that could only be classified at the genus level or higher. We excluded probable false positive taxa (Cyanobacteria/Melainabacteria and Rickettsiales), as well as eukaryotes, archaea, viruses, and plasmids. Contaminant bacterial taxa were identified using decontam (R package v. 1.18.0)80 and removed at a threshold of 0.1 (Acidovorax sp. KKS102, Streptococcus equi, and Cutibacterium acnes). To normalize for different sequencing depths between samples, species read counts were expressed as the number of reads per 100,000 clean reads in a sample, or as relative abundance where appropriate.
Bacterial alpha diversity was measured by species richness, calculated as the number of taxa present in each sample, and Shannon index, calculated using vegan (R package, v. 2.6-4)81. Bacterial beta diversity was measured by Bray–Curtis dissimilarity (weighted) and Sorensen dissimilarity (unweighted), calculated with vegan using species relative abundance and species presence-absence, respectively, with statistical significance measured by a permutation test. PCoA analyses were conducted using phyloseq (R package, v. 1.42.0)82 with statistical significance assessed by PERMANOVA. Microbiome community groups were defined by k-means clustering using the R stats package (v. 4.2.2)83 on species relative abundance with k = 6, and associations between cluster assignment and sample metadata were tested by multinomial logistic regression. To identify species differentially associated either with ART duration or shedder/non-shedder status, MaAsLin251 was used on species relative abundance with a minimum prevalence threshold of 0.1 (i.e., only species present in at least 10% of samples were included in the analysis), and no additional normalization or transformation. For MaAsLin2 analyses, participant ID was considered a random effect, ART duration and shedder/non-shedder status were considered fixed effects, and q < 0.05 was considered statistically significant.
For functional profiling, we assigned reads to bacterial gene families and metabolic pathways using HUMAnN352. Contaminant identification and beta diversity analyses were performed as described above, using normalized gene family and metabolic pathway read counts or relative abundance for input as appropriate. We used MaAsLin2 to identify gene families and pathways associated with sample metadata and microbiome community clusters, with the same parameters as in the taxonomic analysis.
Virome analysis
MDA and SIA virome data were analyzed separately. For the MDA data, sequencing adapters were trimmed with Cutadapt (v. 4.0)84. Reads were quality-filtered, PhiX and human reads removed, reads deduplicated, and paired reads merged with BBTools. Both merged and unmerged reads were retained for downstream analysis. Contigs were assembled with metaSPAdes (v. 3.15.4)85 and deduplicated with cd-hit-est (v. 4.8.1)86 at 95% identity across 95% of contig length. We retained contigs ≥ 500 bases in length for further analysis.
Identification of viral contigs
We used Cenote-Taker2 (v. 2.1.5)87 with the standard database and VirSorter2 (v. 2.2.4)88 to identify candidate viral contigs. We queried the candidate viral contigs against the NCBI NT database (downloaded June 2020) with megablast (BLAST+ v. 2.13.0)89 and removed contigs with ≥95% nucleotide identity and query coverage to human sequences. We used CheckV (v. 1.0.1)90 to check for false positive, non-viral sequences. We considered contigs to be viral if they were classified as complete, high quality, or medium quality by CheckV; or if they were low-quality with at least three viral genes and no more than one host gene; or if they were classified as proviruses. For contigs classified as proviruses, we kept only the viral regions and discarded flanking host sequences. We discarded contigs flagged by CheckV as having high kmer frequency, or that had been classified as bacteria by Cenote-Taker2 during candidate contig identification. To assign taxonomy, we queried the viral contigs against a custom viral protein database (NCBI RefSeq and neighbor viral sequences, downloaded January 2023) using blastx (e-value 1 × 10−3) and used the taxonomy of the best hit in the database. We discarded contigs classified as Mimiviridae as likely false positives. We used Bowtie 2 (v. 2.5.1)91 to map the quality-filtered sample reads against the viral contigs. Contaminant contigs were identified using decontam (threshold = 0.1), considering both spiked positive and unspiked negative controls. 6 contigs identified as statistically significant contaminants in the positive sequencing controls were removed. Total quality-filtered read counts in the negative controls ranged from 42 to 293 reads. None of the negative controls had any mapped viral reads except for one control, which had 11 reads mapped to 5 anellovirus contigs; these contigs were found more frequently in the CVL specimens (74 samples with one or more reads mapped) and did not meet the criteria for removal. For downstream analyses, we kept samples that had at least 1000 viral reads (116 of 125 samples (93%)).
Ecological analyses
Virome alpha (Shannon index and richness) and beta diversity metrics (Bray–Curtis and Sorensen dissimilarity) were calculated, and PCoA was conducted as described for the bacteriome, using viral contig relative abundance for weighted analyses and presence-absence for unweighted analyses. K-means clustering was conducted on virus family relative abundance in R as described for the bacteriome, with k = 3. MaAsLin2 was used at different taxonomic levels (family, genus, and contig) to identify viral taxa associated with ART duration or shedder/non-shedder status, using the same parameters as described for the bacteriome. LME was used to assess statistical significance for alpha diversity over time, permutation tests were used to test statistical significance for beta diversity within and between individuals, and PERMANOVA was used to assess differences in overall virome composition.
Anellovirus phylogenetic analysis
We used Geneious Prime (v. 2023.0.4, http://www.geneious.com) to predict open reading frames (ORFs) from the anellovirus contigs called by blastx during taxonomy assignment. ORFs ≥ 1000 nucleotides in length were extracted, manually curated, translated to amino acids, and aligned with reference Alphatorquevirus, Betatorquevirus, Gammatorquevirus, Hetorquevirus, Omegatorquevirus, Epsilontorquevirus, and Zetatorquevirus sequences retrieved from GenBank (accessions in Supplementary Data 14) using MAFFT (v. 7.505)92 with auto algorithm selection. The alignment was trimmed with trimAl (v. 1.4.rev15)93 using the -gappyout option, and a maximum likelihood phylogeny was constructed with IQ-TREE (v. 2.0.3)94 with the model LG + F + G4 and 1000 bootstrap replicates. The phylogeny was visualized and rooted in FigTree (v. 1.4.4)95 using Epsilontorquevirus/Zetatorquevirus as the outgroup, and contigs were assigned to genera based on clade location. Phylogeny-based genus assignments, rather than blastx-based genus assignments, were used for anellovirus contigs in genus-level virome analyses (e.g., MaAsLin2 analyses). For anellovirus UniFrac distance, a separate phylogeny was built with the contig ORF1 sequences and a single Zetatorquevirus reference sequence, using the method described above. This phylogeny was re-rooted in R using ape (v. 5.6-2)96, the reference sequence was pruned using phyloseq (v. 1.42.0)82, and weighted and unweighted UniFrac distance were calculated with phyloseq using the pruned phylogeny and anellovirus contig relative abundance. PCoA, PERMANOVA, and permutation analyses were performed using weighted and unweighted UniFrac distance based on the pruned phylogeny and anellovirus contig relative abundance or presence-absence.
Papillomavirus phylogenetic analysis
We used Geneious Prime to predict ORFs from the papillomavirus contigs called by blastx during taxonomy assignment. We mapped contig ORFs ≥ 1000 nucleotides in length against reference Alphapapillomavirus, Betapapillomavirus, Gammapapillomavirus, Mupapillomavirus, and Nupapillomavirus genomes and annotated the ORFs based on mapping to the reference genes. After manual curation, concatenated contig E1–E2–L2–L1 nucleotide sequences and reference Alphapapillomavirus, Betapapillomavirus, Gammapapillomavirus, Mupapillomavirus, and Nupapillomavirus sequences retrieved from RefSeq (accessions in Supplementary Data 15) were aligned with MAFFT with auto algorithm selection, trimmed with trimAl with the -gappyout option, and a maximum likelihood phylogeny constructed with IQ-TREE with the model GTR + G and 1000 bootstrap replicates. The phylogeny was visualized in FigTree and rooted using Mupapillomavirus/Nupapillomavirus as the outgroup. Contigs were assigned to genera based on their clade, and phylogeny-based genus assignments were used for papillomavirus contigs in genus-level virome analyses. Papillomavirus beta diversity analyses were performed in the same way as for anelloviruses, with a separate phylogeny built from the contig E1–E2–L2–L1 sequences and a single Mupapillomavirus reference.
SIA virome analysis
SIA virome sequencing reads were processed through the same quality control pipeline as described for the MDA virome reads. SIA reads were then queried against the viral RefSeq and neighbor sequences database using blastx (e-value 1 × 10−3) and parsed with MEGAN (community edition, v. 6.24.21)97.
Alphatorquevirus qPCR assay
Forward primer 5’-GTGCCGIAGGTGAGTTTA-3’, reverse primer 5’-AGCCCGGCCAGTCC-3’, and probe 5’-[6-FAM]-TCAAGGGGCAATTCGGGCT-[TAMRA]-3’ were used with Applied Biosystems 2× TaqMan Fast Universal PCR Master Mix and 5 μL of sample TNA per 25-μL reaction. Reactions were performed on an Applied Biosystems QuantStudio 7 Flex Real-Time PCR system (ThermoFisher; Waltham, MA), with reaction conditions of 95 °C for 20 s and 40 cycles of 95 °C for 1 s and 60 °C for 20 s.
Statistical analysis (bacteriome and virome)
All statistical analyses were conducted in R (v. 4.2.2)83. p-values < 0.05 were considered statistically significant. Where appropriate, corrections were made for multiple comparisons using the Benjamini–Hochberg method98. We corrected for repeated sampling using participant ID as a random effect in linear mixed models, multinomial logistic regression, and MaAsLin2 analyses where appropriate. PERMANOVA was performed using the adonis2 function in the vegan package (v. 2.6-4)81 and we corrected for repeated sampling by restricting permutations within individuals. Multinomial logistic regression was done with the mblogit function in the mclogit package (v. 0.9.6)99. Differential abundance analysis was conducted using MaAsLin2 (v. 1.12.0)51, with a minimum prevalence threshold of 0.1 and no additional normalization or transformation, and adjusted p-values < 0.05 were considered statistically significant. LME was done using the nlme package (v. 3.1-160)100. For MaAsLin2 and LME, participant ID was included as a random effect, ART duration and shedder/non-shedder status were included as fixed effects, and taxa relative abundance was the dependent variable. For linear mixed models of anellovirus abundance and CD4 counts, 78 samples had CD4 count data available and were included in the analysis. CD4 counts were included as a fixed effect, anellovirus relative abundance or alphatorquevirus qPCR copy number was included as the dependent variable, and participant ID was included as a random effect. Participant age was not included as a variable in the bacteriome and virome analyses. Details of the code used in statistical analyses can be found in the GitHub code repository for this study. Plots were generated using R, GraphPad Prism (v. 9.3.1, www.graphpad.com), and FigTree.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The de-identified sequencing reads generated in this study, with human sequences removed, have been deposited in the NCBI Sequence Read Archive under accession code PRJNA1077994. Assembled viral contig sequences are available upon request from the corresponding author, E.S.L.
Code availability
Code used in this study can be found at https://doi.org/10.5281/zenodo.15724178.
References
Ghosn, J., Taiwo, B., Seedat, S., Autran, B. & Katlama, C. HIV. Lancet 392, 685–697 (2018).
Menéndez-Arias, L. & Delgado, R. Update and latest advances in antiretroviral therapy. Trends Pharm. Sci. 43, 16–29 (2022).
Joint United Nations Programme on HIV and AIDS. HIV Estimates with Uncertainty Bounds 1990–Present (UNAIDS, 2023).
Bull, M. et al. Genital shedding of human immunodeficiency virus type-1 (HIV) when antiretroviral therapy suppresses HIV replication in the plasma. J. Infect. Dis. 222, 777–786 (2020).
Homans, J. et al. Permissive and protective factors associated with presence, level, and longitudinal pattern of cervicovaginal HIV shedding. J. Acquir Immune Defic. Syndr. 60, 99–110 (2012).
Uvin, S. C. & Caliendo, A. M. Cervicovaginal human immunodeficiency virus secretion and plasma viral load in human immunodeficiency virus-seropositive women. Obstet. Gynecol. 90, 739–743 (1997).
Hart, C. E. et al. Correlation of human immunodeficiency virus type 1 RNA levels in blood and the female genital tract. J. Infect. Dis. 179, 871–882 (1999).
Cu-Uvin, S. et al. Effect of highly active antiretroviral therapy on cervicovaginal HIV-1 RNA. Aids 14, 415–421 (2000).
Kovacs, A. et al. Determinants of HIV-1 shedding in the genital tract of women. Lancet 358, 1593–1601 (2001).
King, C. C. et al. Prevalence, magnitude, and correlates of HIV-1 genital shedding in women on antiretroviral therapy. J. Infect. Dis. 216, 1534–1540 (2017).
France, M., Alizadeh, M., Brown, S., Ma, B. & Ravel, J. Towards a deeper understanding of the vaginal microbiota. Nat. Microbiol. 7, 367–378 (2022).
Ravel, J. et al. Vaginal microbiome of reproductive-age women. Proc. Natl. Acad. Sci. USA 108, 4680–4687 (2011).
Gajer, P. et al. Temporal dynamics of the human vaginal microbiota. Sci. Transl. Med. 4, 132ra152 (2012).
Gosmann, C. et al. Lactobacillus-deficient cervicovaginal bacterial communities are associated with increased HIV acquisition in young South African women. Immunity 46, 29–37 (2017).
Liu, F. et al. Comparative metagenomic analysis of the vaginal microbiome in healthy women. Synth. Syst. Biotechnol. 6, 77–84 (2021).
Anahtar, M. N. et al. Cervicovaginal bacteria are a major modulator of host inflammatory responses in the female genital tract. Immunity 42, 965–976 (2015).
Gudnadottir, U. et al. The vaginal microbiome and the risk of preterm birth: a systematic review and network meta-analysis. Sci. Rep. 12, 7926 (2022).
Borgdorff, H. et al. Cervicovaginal microbiome dysbiosis is associated with proteome changes related to alterations of the cervicovaginal mucosal barrier. Mucosal Immunol. 9, 621–633 (2016).
Zevin, A. S. et al. Microbiome composition and function drives wound-healing impairment in the female genital tract. PLoS Pathog. 12, e1005889 (2016).
Łaniewski, P. et al. Linking cervicovaginal immune signatures, HPV and microbiota composition in cervical carcinogenesis in non-Hispanic and Hispanic women. Sci. Rep. 8, 7593 (2018).
Norenhag, J. et al. The vaginal microbiota, human papillomavirus and cervical dysplasia: a systematic review and network meta-analysis. BJOG 127, 171–180 (2020).
Brotman, R. M. et al. Bacterial vaginosis assessed by gram stain and diminished colonization resistance to incident gonococcal, chlamydial, and trichomonal genital infection. J. Infect. Dis. 202, 1907–1915 (2010).
Borgdorff, H. et al. Lactobacillus-dominated cervicovaginal microbiota associated with reduced HIV/STI prevalence and genital HIV viral load in African women. ISME J. 8, 1781–1793 (2014).
Sha, B. E. et al. Female genital-tract HIV load correlates inversely with Lactobacillus species but positively with bacterial vaginosis and Mycoplasma hominis. J. Infect. Dis. 191, 25–32 (2005).
Thurman, A. R. et al. vaginal microbiota and mucosal pharmacokinetics of tenofovir in healthy women using a 90-day tenofovir/levonorgestrel vaginal ring. Front. Cell Infect. Microbiol. 12, 799501 (2022).
Dabee, S. et al. Genital microbiota of women using a 90 day tenofovir or tenofovir and levonorgestrel intravaginal ring in a placebo controlled randomized safety trial in Kenya. Sci. Rep. 12, 12040 (2022).
Mazibuko-Motau, N. et al. Vaginal microbial shifts are unaffected by oral pre-exposure prophylaxis in South African women. Sci. Rep. 12, 16187 (2022).
Klatt, N. R. et al. Vaginal bacteria modify HIV tenofovir microbicide efficacy in African women. Science 356, 938–945 (2017).
Donahue Carlson, R. et al. The female genital tract microbiome is associated with vaginal antiretroviral drug concentrations in human immunodeficiency virus-infected women on antiretroviral therapy. J. Infect. Dis. 216, 990–999 (2017).
Taneva, E. et al. Vaginal microbiome modulates topical antiretroviral drug pharmacokinetics. JCI Insight. https://doi.org/10.1172/jci.insight.99545 (2018).
Nicol, M. R. et al. Depot medroxyprogesterone acetate and the vaginal microbiome as modifiers of tenofovir diphosphate and lamivudine triphosphate concentrations in the female genital tract of Ugandan women: implications for tenofovir disoproxil fumarate/lamivudine in preexposure prophylaxis. Clin. Infect. Dis. 70, 1717–1724 (2020).
Cheu, R. K. et al. Impact of vaginal microbiome communities on HIV antiretroviral-based pre-exposure prophylaxis (PrEP) drug metabolism. PLoS Pathog. 16, e1009024 (2020).
Lantz, A. M. et al. Vaginal microbiome, antiretroviral concentrations, and HIV genital shedding in the setting of hormonal contraception initiation in Malawi. Aids 37, 2185–2190 (2023).
Liu, C. M. et al. The effect of antiretroviral therapy initiation on the vaginal microbiome in HIV-infected women. Open Forum Infect. Dis. 6, ofz328 (2019).
Keita, A. et al. Evidence of HIV-1 genital shedding after one year of antiretroviral therapy in females recently diagnosed in Bamako, Mali. Microorganisms. https://doi.org/10.3390/microorganisms9102164 (2021).
Marques, M. P. C. et al. Protease inhibitor effects on prevalence of bacterial vaginosis and human papillomavirus-related lesions. J. Obstet. Gynaecol. Res. 46, 899–906 (2020).
Politch, J. A. et al. Highly active antiretroviral therapy does not completely suppress HIV in semen of sexually active HIV-infected men who have sex with men. AIDS 26, 1535–1543 (2012).
Gianella, S. et al. Shedding of HIV and human herpesviruses in the semen of effectively treated HIV-1-infected men who have sex with men. Clin. Infect. Dis. 57, 441–447 (2013).
Lawn, S. D. et al. Correlation between human immunodeficiency virus type 1 RNA levels in the female genital tract and immune activation associated with ulceration of the cervix. J. Infect. Dis. 181, 1950–1956 (2000).
Wylie, K. M. et al. Metagenomic analysis of double-stranded DNA viruses in healthy adults. BMC Biol. 12, 71 (2014).
Jakobsen, R. R. et al. Characterization of the vaginal DNA virome in health and dysbiosis. Viruses. https://doi.org/10.3390/v12101143 (2020).
Kaelin, E. A. et al. Cervicovaginal DNA virome alterations are associated with genital inflammation and microbiota composition. mSystems 7, e0006422 (2022).
Madere, F. S. et al. Transkingdom analysis of the female reproductive tract reveals bacteriophages form communities. Viruses. https://doi.org/10.3390/v14020430 (2022).
Li, Y. et al. Altered vaginal eukaryotic virome is associated with different cervical disease status. Virol. Sin. 38, 184–197 (2023).
Siqueira, J. D. et al. Composite analysis of the virome and bacteriome of HIV/HPV co-infected women reveals proxies for immunodeficiency. Viruses. https://doi.org/10.3390/v11050422 (2019).
Wylie, K. M. et al. The vaginal eukaryotic DNA virome and preterm birth. Am. J. Obstet. Gynecol. 219, 189.e1–e12 (2018).
Happel, A. U. et al. Cervicovaginal human papillomavirus genomes, microbiota composition and cytokine concentrations in South African adolescents. Viruses. https://doi.org/10.3390/v15030758 (2023).
Happel, A. U. et al. Presence and persistence of putative lytic and temperate bacteriophages in vaginal metagenomes from South African adolescents. Viruses 13, 2341 (2021).
Sung, H. et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 71, 209–249 (2021).
Taylor, L. J., Keeler, E. L., Bushman, F. D. & Collman, R. G. The enigmatic roles of Anelloviridae and Redondoviridae in humans. Curr. Opin. Virol. 55, 101248 (2022).
Mallick, H. et al. Multivariable association discovery in population-scale meta-omics studies. PLOS Comput. Biol. 17, e1009442 (2021).
Beghini, F. et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife 10, e65088 (2021).
Christensen, J. K. et al. Prevalence and prognostic significance of infection with TT virus in patients infected with human immunodeficiency virus. J. Infect. Dis. 181, 1796–1799 (2000).
Shibayama, T. et al. Inverse relationship between the titre of TT virus DNA and the CD4 cell count in patients infected with HIV. Aids 15, 563–570 (2001).
Thom, K. & Petrik, J. Progression towards AIDS leads to increased Torque teno virus and Torque teno minivirus titers in tissues of HIV infected individuals. J Med Virol. 79, 1–7 https://doi.org/10.1002/jmv.20756 (2007).
Schmidt, L. et al. Torque Teno Virus plasma level as novel biomarker of retained immunocompetence in HIV-infected patients. Infection 49, 501–509 (2021).
Honorato, L. et al. The Torque Teno Virus Titer in saliva reflects the level of circulating CD4(+) T lymphocytes and HIV in individuals undergoing antiretroviral maintenance therapy. Front. Med. 8, 809312 (2021).
Li, L. et al. AIDS alters the commensal plasma virome. J. Virol. 87, 10912–10915 (2013).
Liu, K. et al. HIV-1 infection alters the viral composition of plasma in men who have sex with men. mSphere. https://doi.org/10.1128/mSphere.00081-21 (2021).
van Rijn, A. L., Roos, R., Dekker, F. W., Rotmans, J. I. & Feltkamp, M. Torque teno virus load as marker of rejection and infection in solid organ transplantation—a systematic review and meta-analysis. Rev. Med. Virol. 33, e2393 (2023).
Arnold, K. B. et al. Increased levels of inflammatory cytokines in the female reproductive tract are associated with altered expression of proteases, mucosal barrier proteins, and an influx of HIV-susceptible target cells. Mucosal Immunol. 9, 194–205 (2016).
Shkoporov, A. N. & Hill, C. Bacteriophages of the human gut: the “Known Unknown” of the microbiome. Cell Host Microbe 25, 195–209 (2019).
Olivieri, E. H. et al. Biobanking practice: RNA storage at low concentration affects integrity. Biopreserv Biobank 12, 46–52 (2014).
Lebeer, S. et al. A citizen-science-enabled catalogue of the vaginal microbiome and associated factors. Nat. Microbiol 8, 2183–2195 (2023).
Vodstrcil, L. A. et al. The influence of sexual activity on the vaginal microbiota and Gardnerella vaginalis clade diversity in young women. PLoS One 12, e0171856 (2017).
Plummer, E. L. et al. Sexual practices have a significant impact on the vaginal microbiota of women who have sex with women. Sci. Rep. 9, 19749 (2019).
Blostein, F., Gelaye, B., Sanchez, S. E., Williams, M. A. & Foxman, B. Vaginal microbiome diversity and preterm birth: results of a nested case-control study in Peru. Ann. Epidemiol. 41, 28–34 (2020).
Carrillo-Ng, H. et al. Variations in cervico-vaginal microbiota among HPV-positive and HPV-negative asymptomatic women in Peru. BMC Res. Notes 14, 4 (2021).
Holliday, M. et al. Insights into the genital microbiota of women who experienced fetal death in utero. Microorganisms 11, 1877 (2023).
Borgdorff, H. et al. The association between ethnicity and vaginal microbiota composition in Amsterdam, the Netherlands. PLoS One 12, e0181135 (2017).
Marconi, C. et al. Characterization of the vaginal microbiome in women of reproductive age from 5 regions in Brazil. Sex. Transm. Dis. 47, 562–569 (2020).
Dabee, S., Passmore, J. S., Heffron, R. & Jaspan, H. B. The complex link between the female genital microbiota, genital infections, and inflammation. Infect. Immun. https://doi.org/10.1128/iai.00487-20 (2021).
Soria, J. et al. Transmitted HIV resistance to first-line antiretroviral therapy in Lima, Peru. AIDS Res. Hum. Retroviruses 28, 333–338 (2012).
Reichelderfer, P. S. et al. Effect of menstrual cycle on HIV-1 levels in the peripheral blood and genital tract. WHS 001 Study Team. Aids 14, 2101–2107 (2000).
Bull, M. E. et al. Monotypic human immunodeficiency virus type 1 genotypes across the uterine cervix and in blood suggest proliferation of cells with provirus. J. Virol. 83, 6020–6028 (2009).
Liang, G. et al. The stepwise assembly of the neonatal virome is modulated by breastfeeding. Nature 581, 470–474 (2020).
Patel, A. et al. Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR green I. Nat. Protoc. 2, 269-276 (2007).
Bushnell B. BBMap https://sourceforge.net/projects/bbmap/ (2019).
Breitwieser, F. P., Baker, D. N. & Salzberg, S. L. KrakenUniq: confident and fast metagenomics classification using unique k-mer counts. Genome Biol. 19, 198 (2018).
Davis, N. M., Proctor, D. M., Holmes, S. P., Relman, D. A. & Callahan, B. J. Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome 6, 226 (2018).
Oksanen, J. et al. vegan: Community Ecology Package. R Package Version 2.6-4 https://CRAN.R-project.org/package=vegan (2022).
McMurdie, P. J. & Holmes, S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8, e61217 (2013).
R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, 2022); https://www.R-project.org/.
Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 3. https://doi.org/10.14806/ej.17.1.200 (2011).
Nurk, S., Meleshko, D., Korobeynikov, A. & Pevzner, P. A. metaSPAdes: a new versatile metagenomic assembler. Genome Res. 27, 824–834 (2017).
Li, W. & Godzik, A. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22, 1658–1659 (2006).
Tisza, M. J., Belford, A. K., Domínguez-Huerta, G., Bolduc, B. & Buck, C. B. Cenote-Taker 2 democratizes virus discovery and sequence annotation. Virus Evol. 7, veaa100 (2021).
Guo, J. et al. VirSorter2: a multi-classifier, expert-guided approach to detect diverse DNA and RNA viruses. Microbiome 9, 37 (2021).
Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinforma. 10, 421 (2009).
Nayfach, S. et al. CheckV assesses the quality and completeness of metagenome-assembled viral genomes. Nat. Biotechnol. 39, 578–585 (2021).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Capella-Gutiérrez, S., Silla-Martínez, J. M. & Gabaldón, T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25, 1972–1973 (2009).
Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534 (2020).
Rambaut, A. FigTree v1.4.4. Github https://github.com/rambaut/figtree (2025).
Paradis, E. & Schliep, K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics 35, 526–528 (2019).
Huson, D. H. et al. MEGAN Community edition—interactive exploration and analysis of large-scale microbiome sequencing data. PLoS Comput. Biol. 12, e1004957 (2016).
Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 57, 289–300 (1995).
Elff, M. mclogit: Multinomial Logit Models, with or without Random Effects or Overdispersion. R Package Version 0.9.6 https://CRAN.R-project.org/package=mclogit (2022).
Pinheiro, J., Bates, D. & R Core Team. nlme: Linear and Nonlinear Mixed Effects Models. R Package Version 3.1-160. https://CRAN.R-project.org/package=nlme (2022).
Acknowledgements
The authors would like to acknowledge the men and women who participated in this study and the work of the clinical study personnel (Carmela Ganoza, Marcela Rodriguez, Angie Roldan) who made this study possible. Additionally, we would like to acknowledge the technical support from Kelli Kraft, Jillian Legard, Jennifer McKernan-Mullin, and LaRinda Holland. Mara Fredenberg performed the VLP microscopy. We thank Dr. Bertram Jacobs (ASU) for providing the rhinovirus B14 stock sample and Dr. Masmudur Rahman (ASU) for providing the myxoma virus stock sample. This work was supported by the National Institute of Allergy and Infectious Disease at the National Institutes of Health: R01 AI071212 and R01 AI091550 awarded to L.M.F.; UM1-AI-106701 awarded to R.W.C.; UW/FHCRC CFAR P30-AI-027757; and R00DK107923 awarded to E.S.L.
Author information
Authors and Affiliations
Contributions
E.A.K.: data curation, investigation, formal analysis, writing—original draft, and writing—review and editing. C.M.: clinical set-up for the trial, medical expertise as an OB-GYN. J.S.: clinical set-up for the trial, medical expertise with experience in this particular community. A.L.R.: clinical set-up for the trial, medical expertise with experience in this particular community. E.T.: clinical set-up for the trial, medical expertise with experience in this particular community. R.W.C.: CLIA-certified quantification of HIV in the plasma and CVL. L.M.F.: design and funding of the parent clinical trial, input on the current study design. M.E.B.: conceptualization, resources, data curation, and writing—review and editing. E.S.L.: conceptualization, funding, supervision, writing—original draft, and writing—review and editing.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethical approval
This study was approved by a local institutional review board and included local researchers in study design, implementation, and authorship. Roles and responsibilities of collaborators were agreed on prior to the research. The work was locally relevant, and available regional research has been referenced. The research does not result in stigmatization or other harm to the participants, and appropriate biosafety precautions were implemented to minimize risk to the researchers.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Kaelin, E.A., Mitchell, C., Soria, J. et al. Longitudinal cervicovaginal bacteriome and virome alterations associate with discordant shedding and ART duration in women living with HIV in Peru. Nat Commun 16, 7904 (2025). https://doi.org/10.1038/s41467-025-63158-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-025-63158-y
