All investigations involving human subjects were approved by the Dartmouth College Committee for the Protection of Human Subjects, Dartmouth-Hitchcock Medical Center (DHMC), and the Dartmouth Health Institutional Review Board. Informed written consent was obtained from all patients prior to surgery and all research was conducted in accordance with the policies of the Declaration of Helsinki. Endometrial (EM), endocervical (CX), and ectocervical (ECX) tissues were obtained from patients undergoing surgical hysterectomy at DHMC (Lebanon, NH, USA) for benign indications. Patients with evidence of reproductive tract infection, reproductive tract cancer, who were taking hormonal birth control, or had a recent history of exogenous hormone therapy were excluded from this study. All tissue received was distal to any sites of pathology. Menopausal status and menstrual cycle stage were determined by trained clinical pathologists via histological assessment with post-menopausal status defined as patients with atrophic EM. Pre-menopausal status was assigned as those patients with histological evidence of normal menstrual cycling. The mean age of pre-menopausal women was 37.8 years (SD = 8.6; range: 22-51) and the mean age of post-menopausal women was 66.5 years (SD = 7.7; range: 52-80).

After surgery, fresh tissues were promptly processed, sectioned, and assessed for inclusion criteria by the staff of the DHMC Department of Pathology and Laboratory Medicine prior to transfer to our laboratory. Mucosal tissue sections were then weighed with mean tissue mass received measured as 3.0 g (95% CI: 2.3 to 3.7 g) for pre-menopausal EM tissue, 1.8 g (95% CI: 1.4 to 2.1 g) for post-menopausal EM tissue, 1.6 g (95% CI: 1.2 to 2.0 g) for pre-menopausal CX tissue, 1.4 g (95% CI: 1.1 to 1.7 g) for post-menopausal CX tissue, 2.5 g (95% CI: 1.8 to 3.3 g) for pre-menopausal ECX tissue and 1.4 g (95% CI: 1.0 to 1.7 g) for post-menopausal ECX tissue.

As described previously (25, 37), samples were then minced under aseptic conditions into 1-2 mm fragments using sterile scalpels (Feather, Osaka, Japan) before being enzymatically digested in a HBSS solution containing 0.05% Type 1V collagenase (Sigma-Aldrich, St, Louis, MO, USA) and 0.01% DNase (Worthington Biochemical, Lakewood, MO, USA) 37°C. After 1 h incubation, digested tissues were passed through 200 μm followed by 35 μm nylon mesh filters (LBA, Miami, FL, USA) to remove debris and epithelial cells sheets. Red blood cells were then removed via osmotic lyses, dead cells were removed using a commercial dead cell removal kit (Miltenyi Biotec, Auburn, CA, USA), after which the flow through was again filtered using a 30 μm nylon filter (Miltenyi Biotec). The resulting mixed cell suspensions contained immune cells, stromal fibroblasts, and residual single epithelial cells. Total numbers of live cells were determined by hemacytometer and trypan blue (VWR-Amresco, Solon, OH, USA) exclusion.

A previously described pNL-GFPm.6ATRi-BaL.ecto reporter virus (38) was used for all infections. This CCR5 trophic virus, referred to as “BaL-GFP” throughout this paper, was produced through 293T cell transfection and only expresses GFP in productively infected HIV-1-susceptible cells. The TZM-bl cell line was used to determine the infectious titer of all virus stocks prior to experimental use as previously described (39).

Unstimulated mixed cell solutions were resuspended in X-VIVO 15 media (Lonza, Walkersville, MD, USA) supplemented with 10% charcoal dextran stripped human serum (Valley Biomedical, Winchester, VA, USA) and plated out in triplicate for each sample at a density of 300,000 cells/well in a round bottom ultra-low attachment 96-well plate (Corning, Corning, NY, USA). In one well for each tissue sample, 10 μM of the reverse transcriptase inhibitor Azidothymidine (AZT; NIH HIV Reagent Program) was added as a viral infection control and cells were incubated for 2 h at 37°C. Cells were then infected at an MOI of 1 (based on TZM-bl titer) with HIV-1 BaL-GFP with uninfected control wells receiving medium without virus. After 2 h of virus exposure, all cells were washed twice with media to remove residual virus and cell cultures kept at 37°C for 6 days with AZT at a concentration of 10 μM added to replication control wells. On days 3 and 5 post infection, half the media was removed from each well and replaced with fresh media and with AZT at a concentration of 10 μM again added to replication control wells.

After 6 days in culture, mixed cell suspensions were stained with combinations of the following anti-human antibodies: CXCR4-PE (Clone 12G5; BD), CX3CR1-PE/Dazzle (Clone 2A9-1; BioLegend), CCR5-PerCPCy5.5 (Clone J418F1; BioLegend), α4β7-AF647 (R&D Systems), CD45-AF700 (Clone 2D1; BioLegend), CD3-APC/E780 (Clone UCHT1; BioLegend), CD8-VF450 (Clone SK1; Tonbo Biosciences), CD69-VioGreen (Clone REA824, Miltenyi Biotec), CCR10-PE (Clone 6588-5; BioLegend), CCR6-PE/Dazzle (Clone G034E3; BioLegend), CXCR3-PE/Cy5 (Clone G025H7; BioLegend), CCR4-PE/Cy7 (Clone L291H4; BioLegend), CD127-APC (Clone A019D5; BioLegend), CD25-APC/Cy7 (Clone BC96; BioLegend), CD3-E450 (Clone SK7; Invitrogen), CD8-BV510 (Clone SK1; BioLegend), CD45-PC5.5 (Clone J33; IOTest), and CD14-PE/Vio770 (Clone TÜK4; Miltenyi Biotec). Following staining, cells were fixed with 2% methanol free formaldehyde (Polysciences Inc., Warrington, PA, USA) and data acquired using a 10-color Gallios Flow Cytometer (Beckman Coulter Life Sciences, Indianapolis, IN, USA) running Kaluza software. Data analysis was performed using FlowJo v10.10 Software (Tree Star, Inc., Ashland, OR, USA).

For select experiments tissue-specific mixed cell suspensions were generated from pre- and post-menopausal patients, with a portion of post-menopausal cells incubated for 2 h with 17β-estradiol (E₂; Sigma-Aldrich, Burlington, MA, USA) at a concentration of 5 x 10^-8 M prior to HIV-1 BaL-GFP infection. E₂ exposure at a concentration of 5 x 10^-8 M was maintained throughout the 6-day post-infection period with E₂ prepared as described previously (40).

In this study, CD4⁺ T cells were defined as CD45⁺ CD3⁺ CD8^- single cells; CD14⁺ macrophages were defined as CD45⁺ CD14⁺ single cells exhibiting high side-scatter ( Figure 1A ). Reported infection frequencies represent the % GFP⁺ cells from HIV-1 infected samples minus the % GFP⁺ cells from the corresponding HIV-1 + AZT treated control to ensure that reported GFP expression is solely the result of productive HIV infection.

Discrete measures were analyzed by Mann-Whitney U test, Kruskal-Wallis test followed by Dunn’s post-test for pairwise multiple comparisons, Mann-Whitney U test with Holm-Šídák correction for multiple comparisons or 2-way ANOVA with Tukey’s post-test for multiple comparisons. Statistical analyses were performed using GraphPad Prism 10.2.3 software and mean differences were considered significant at two-sided P < 0.05.

The quantity of HIV target cells at the site of mucosal exposure has been proposed to modify HIV infection susceptibility (41). To determine whether menopause alters the density of HIV-1 target cells in the FRT, tissue specific single cell suspensions were generated from mucosal tissues of the EM, CX, and ECX as described in the methods. CD4⁺ T cell and CD14⁺ cell frequencies were subsequently determined by flow cytometry ( Figure 1A ) and used to calculate cell density normalized to cell number per gram of tissue.

Among the three tissue types, mean CD4⁺ T cell density was significantly greater in the EM (1.62 x 10⁵ cells/g) relative to CX (0.65 x 10⁵ cells/g) and ECX (0.64 x 10⁵ cells/g) with density not differing between the latter two tissues ( Figure 1B ). Following stratification by menopausal status, the density of EM (2.58 x 10⁵ cells/g vs 0.92 x 10⁵ cells/g; Figure 1C ) CD4⁺ T cells was significantly lower in post-menopausal relative to pre-menopausal women. In contrast, menopausal status did not alter CD4⁺ T cell density in the CX ( Figure 1D ) or ECX ( Figure 1E ).

Unlike CD4⁺ T cells, mean CD14⁺ cell density did not significantly differ between the EM (1.23 x 10⁴ cells/g), CX (0.95 x 10⁴ cells/g), or ECX (0.70 x 10⁴ cells/g; Figure 1F ), though CD14⁺ cell density was significantly lower than CD4⁺ T cell density for all three tissues ( Figure 1A vs. Figure 1E ; approximately 10-fold; P > 0.0001). Following stratification by menopausal status, a significant decrease in CD14⁺ cell density was observed in post-menopausal versus pre-menopausal women in the EM only (1.99 x 10⁴ cells/g vs 0.51 x 10⁴ cells/g; Figure 1G ). In contrast, no differences in CD14⁺ cell density were observed with menopausal status in the CX ( Figure 1H ) or ECX ( Figure 1I ). Taken together these data indicate that the density of HIV-1 target cells decreases in the mucosa of the EM, but not the CX or ECX, following menopause. Moreover, these data suggest that CD4⁺ T cells are the most abundant putative HIV-1 target cell type throughout the mucosa of the FRT.

We next assessed whether menopausal status alters the susceptibility of CD4⁺ T cells and CD14⁺ cells to HIV-1 infection in the FRT mucosa. As before, tissue specific mixed cell suspensions were generated from the EM, CX, and ECX prior to either mock infection or infection at a MOI of 1 with a HIV-1 BaL reporter virus that expresses GFP following productive viral infection. After 6 days of viral exposure, flow cytometry was used to determine CD4⁺ T cell and CD14⁺ cell infection frequency in the mixed cell cultures. As seen in ( Figure 2A ), while no GFP expression was detected in mock infected controls, GFP expression in virally infected cells could be readily detected by flow cytometry. Moreover, GFP expression was essentially absent in cells exposed to HIV-1 BaL-GFP in the presence of the reverse transcriptase inhibitor AZT, thus, demonstrating the specificity of detected GFP expression as a marker for de novo infection of both CD14⁺ cells and CD4⁺ T cells.

Following six days of viral exposure, mean CD4⁺ T cell HIV-1 infection frequency was significantly greater in the EM (4.8%) relative to the CX (2.0%) or ECX (1.6%; Figure 2B ), with no differences observed between the CX and ECX. When stratified by menopausal status, mean CD4⁺ T cell infection frequency was significantly greater in the EM of post-menopausal women (6.7%) relative to pre-menopausal women (2.3%; Figure 2C ), with pre-menopausal EM infection frequencies consistent with those observed for the CX and ECX for both pre- and post-menopausal donors ( Figure 2B ). Thus, despite the changes observed in the EM, menopause did not alter the HIV-1 infection susceptibility of CD4⁺ T cells isolated from the CX ( Figure 2D ) nor ECX ( Figure 2E ) as measured by relative and absolute infection frequency.

In contrast to CD4⁺ T cells, no differences in mean CD14⁺ cell infection frequency were observed between the three tissue types ( Figure 2F ). Likewise, menopausal status did not alter mean CD14⁺ cell infection frequency in the EM ( Figure 2G ), CX ( Figure 2H ) nor ECX ( Figure 2I ). These data suggest that greater CD4⁺ T cell infection frequency following menopause is driven by tissue specific changes unique to the EM.

Previous work by our group and others suggest that greater CCR5 expression may increase CD4⁺ T cell susceptibility to HIV infection post-menopause (25, 35). To test this hypothesis, we first measured the expression of CCR5 and other select surface markers of CD4⁺ T cell HIV-1 infection susceptibility on cells isolated from the FRT. These included the expression of the alternate coreceptors CXCR4, CX3CR1, and integrin α4β7 in addition to the early activation marker CD69. Consistent with previous observations, post-menopausal status was associated with significant increases in the frequency of CCR5 expression on CD4⁺ T cells isolated from the EM ( Figure 3A ), CX ( Figure 3B ), and ECX ( Figure 3C ). Increased α4β7 expression frequency was also observed on CD4⁺ T cells of the EM ( Figure 3A ) and CX ( Figure 3B ), but not the ECX ( Figure 3C ) following menopause, while CCR5 + α4β7 co-expression was observed to be greater only in the EM ( Figure 3A ). In contrast, no changes in CD4⁺ T cell expression of CXCR4, CX3CR1, or CD69 were observed for any tissue ( Figures 3A–C ).

We next evaluated whether observed changes in surface marker expression ( Figure 3A ) correlated with increased infection susceptibility in the EM ( Figure 3D ) by comparing the BaL-GFP infection frequency of specific receptor expressing cells between pre- and post-menopausal women. After controlling for changes in surface marker expression frequency, the infection frequency of CCR5, α4β7, CCR5 + α4β7, CXCR4, and CD69 expressing CD4⁺ T cells remained greater in the EM of women following menopause ( Figure 3D ). HIV-1 infection frequency of CX3CR1⁺ CD4⁺ T cells likewise trended greater in the EM of post-menopausal women but did not reach statistical significance ( Figure 3D ). In contrast to the EM, no changes in infection frequency were observed with menopause in the CX ( Figure 3E ) nor ECX ( Figure 3F ) for any of the markers measured. Taken together this data indicates that despite broad increases in the expression of previously described determinates of CD4⁺ T cell HIV-1 infection susceptibility across all three tissue types following menopause, the increased expression of CCR5 and α4β7 is only associated with increases in EM CD4⁺ T cell infection susceptibility.

Functional polarization has been shown to further influence CD4⁺ T cell HIV infection susceptibility with CCR6 expressing Th17 and Th22 cells identified as preferential targets of HIV infection in the FRT (25, 26, 41). As prior work by our group suggested that the frequency of CCR6 expressing CD4⁺ T cells increases after menopause in the EM, but not the CX and ECX (25), we next sought to determine whether shifts in the relative composition of T helper cell subsets could be driving increased infection frequency in the EM. To test this hypothesis, the relative frequencies and densities of Th1 cells, Th1/17 cells, Th2 cells, Th9 cells, Th17 cells, Th22 cells, and regulatory T cells (Tregs) were quantified by surface maker expression as described previously (42, 43) using the gating strategy shown in ( Figure 4A ).

Th1 cells were the predominant CD4⁺ T helper cell subset in the mucosa of the EM (mean frequency of 26.0%), with reduced, but roughly equal, mean frequencies of Th1/17 cells (8.2%), Th2 cells (6.8%), Th9 cells (8.1%), Th17 cells (5.2%), and Tregs (5.7%; Figure 4B ). Th22 cells were the least common T helper subset in the EM with a mean frequency of 2.8%. T helper subset compositions similar to that of the EM were observed in the CX ( Supplementary Figure 1A ) and ECX ( Supplementary Figure 1B ), with a greater mean frequency of Th1/17 cells observed in the CX (14.1%) relative to the EM (8.2%) and ECX (8.8%). Following stratification by menopausal status, the mean frequency of Th1 cells was observed to be greater following menopause in the EM (17.5% in pre-menopausal patients versus 33.5% in post-menopausal patients; Figure 4C ), which was offset by decreases in the mean frequency of Th17 (6.4% vs 4.0%) and Tregs (7.3% vs 4.1%). No changes in Th1/17, Th2, Th9, or Th22 cell frequency were observed following menopause in the EM. Unlike the EM, no changes in the frequency of any phenotype was observed in the CX ( Figure 4D ) or ECX ( Figure 4E ) though a non-significant (p = 0.09) increase in Th1 cell frequency was observed in the CX following menopause.

To further evaluate the effects of menopause on T helper subset composition in the female reproductive tract, we next compared the mean densities of Th1 cells, Th1/17 cells, Th2 cells, Th9 cells, Th17 cells, Th22 cells, and Tregs in EM, CX, and ECX between pre- and post-menopausal women. Following stratification by menopausal status, mean densities of Th2 (2.44 x 10⁴ cells/g vs 0.67 x 10⁴ cells/g), Th17 (1.94 x 10⁴ cells/g vs 0.44 x 10⁴ cells/g), Th22 (0.67 x 10⁴ cells/g vs 0.29 x 10⁴ cells/g), and Tregs (1.79 x 10⁴ cells/g vs 0.47 x 10⁴ cells/g) were all greater in the EM ( Supplementary Figure 2A ) of pre- relative to post-menopausal women. In contrast, no changes in the mean density of any T helper subset were observed with menopausal status in the CX ( Supplementary Figure 2B ) and ECX ( Supplementary Figure 2C ).

To determine whether observed changes in T helper subset composition ( Figure 4C ) correlated with increased infection susceptibility in the EM ( Figure 4F ) we next compared the BaL-GFP infection frequency of specific T helps subsets between pre- and post-menopausal women. After controlling for changes in T helper cells subset composition, the mean frequency of infection remained significantly greater for Th1 (4.3% vs 1.2%), Th1/17 (8.1% vs 2.7%), Th2 (3.5% vs 1.2%), Th9 (10.7% vs 2.0%), Th17 (9.7% vs 3.3%), Th22 (9.5% vs 2.9%), and Tregs (10.3% vs 3.0%) in the EM of post-menopausal women relative to pre-menopausal women ( Figure 4F ). In contrast, no changes in infection frequency were observed for any specific T helper subset in the CX ( Figure 4G ) nor ECX ( Figure 4H ) with menopausal status. Thus, phenotypic changes in T helper subset composition are unlikely to explain increases in HIV-1 infection susceptibility in CD4⁺ T cells isolated from the EM.

Menopause is associated with dramatic declines in the concentrations of both progesterone and estradiol (E₂) (29). Because E₂ is known to reduce CD4⁺ T cell HIV-1 infection susceptibility in vitro (40, 44), we next assessed whether E₂ treatment could reverse the observed increases in endometrial CD4⁺ T cell infection susceptibility following menopause as described in the methods.

In the absence of E₂ treatment, mean CD4⁺ T cell infection frequency was again significantly greater in the EM of post-menopausal versus pre-menopausal women (6.3% versus 1.7%; Figure 5A ). In contrast, E₂ treatment of post-menopausal cells significantly reduced the mean infection frequency by 52.4% (from 6.3% to 3.0%) in CD4⁺ T cells isolated from post-menopausal patients. Unlike with the EM, no significant differences were observed with either menopause or E₂ treatment in CD4⁺ T cell infection frequency in the CX ( Supplementary Figure 3A ) or ECX ( Supplementary Figure 3B ).

To evaluate how E₂ could be reducing EM CD4⁺ T cell HIV-1-infection frequency following menopause, the expression of CCR5, integrin α4β7, CXCR4, CX3CR1, and CD69 on these cells were measured by multicolor flow cytometry. While the expression frequencies of both CCR5 ( Figure 5B ) and α4β7 ( Figure 5C ) were greater on endometrial CD4⁺ T cells following menopause, unexpectedly, we found that their expression was not altered with E₂ treatment. As before, the expression of CXCR4 ( Figure 5D ), CX3CR1 ( Figure 5E ), and CD69 ( Figure 5F ) did not differ between pre- and post-menopausal women and their expression frequencies in post-menopausal women were likewise not altered following E₂ treatment. Taken together, E₂ treatment can reduce menopausal associated increases in EM CD4⁺ T cell HIV-1 infection susceptibility in vitro. This reduction in infection frequency was independent of changes in the expression of CCR5, α4β7, and other previously described surface determinates of HIV-1 infection susceptibility.