Skip to main content
. 2020 May 26;9:e55487. doi: 10.7554/eLife.55487

Figure 2. Comparison of HIV-susceptible cells in unstimulated PBMCs and endometrium.

(A) Paired unstimulated PBMCs and endometrial cells from 4 donors were mock-treated or inoculated with F4.HSA and monitored 3 days later for levels of productive infection. Results are gated on live, singlet CD3+CD8- cells. Numbers correspond to the percentage of infected cells in each sample. (B) t-SNE plots of uninfected and infected cells from Donor #1 shown in panel A, where the PBMC and endometrial specimens were run as separate t-SNEs. Infected cells reside in unique regions of the t-SNE as a result of viral-induced remodeling. Data are representative of 4 independent donors. (C) t-SNE plots of uninfected and infected cells from Donor #1 shown in panel A, where the PBMC and endometrial specimens were run within the same t-SNE. Although the same number of HIV-infected ETs are shown here as in panel B, these cells occupy less t-SNE space since much of the space is now taken up by cells from the PBMC sample. Corresponding data for the remaining 3 donors in panel A are presented in Figure 2—figure supplement 2. (D) Schematic defining Predicted Precursor (‘PRE’) cells. A population of T cells is either mock-treated (top) or exposed to HIV (bottom). These cells include ones that are not susceptible to infection (yellow and green) and those that are (blue and purple). After 3 days, infected cells (surrounded by a black ring) are identified as those expressing the HIV reporter HSA. These cells, however, are remodeled as represented by the conversion from their original blue and purple colors to the red and pink colors, respectively. By using PP-SLIDE to identify for every infected cell the phenotypically most similar cell in the uninfected culture (upward arrows), we identify the PRE cells which harbor the predicted phenotypes of the original cells targeted for HIV infection. These cells do not have the confounder of HIV-induced phenotypic changes, and can be compared to non-susceptible cells in the uninfected culture to identify unique features associated with cells targeted for infection. (E) t-SNE plots of uninfected and PRE cells demonstrate that the dominant population of T cells targeted for infection in both blood and endometrium are memory CD4+ T (Tm) cells. Tm cells (CD4+CD45RO+CD45RA-), naïve CD4+ T cells (CD4+CD45RO-CD45RA+, abbreviated Tn cells), memory CD8+ T cells (CD8+CD45RO+CD45RA-), and naïve CD8+ T cells (CD8+CD45RO-CD45RA+) were identified by manual gating and colored as indicated. Corresponding data for the remaining 3 donors in panel A are presented in Figure 2—figure supplement 2. (F) Comparison of the proportions of Tem (CD4+CD45RO+CD45RA-CCR7-CD62L-) and Tcm (CD4+CD45RO+CD45RA-CCR7+CD62L+) among uninfected Tm and PRE cells in PBMCs and ETs reveals preferential infection of Tem in both compartments. *p<0.05, **p<0.01, ***p<0.001 as assessed using the Student’s paired t test.

Figure 2.

Figure 2—figure supplement 1. CyTOF gating strategy to identify uninfected and HIV-infected T cells.

Figure 2—figure supplement 1.

Figure 2—figure supplement 2. Comparison of HIV-susceptible cells in unstimulated PBMCs and endometrium.

Figure 2—figure supplement 2.

(A) t-SNE plots of uninfected and infected cells from PBMCs or ETs, for Donors #2, 3, and 4 in Figure 2. The PBMC and endometrial specimens were run within the same t-SNE. (B) t-SNE plots of uninfected and PRE cells from Donors #2, 3, and 4 in Figure 2 color-coded by the indicated T cell subsets. The data demonstrate that the dominant population of HIV-susceptible cells from both unstimulated PBMCs and endometrium are Tm.
Figure 2—figure supplement 3. Quantitation of viral-induced remodeling of T cells from PBMCs and ETs.

Figure 2—figure supplement 3.

SLIDE analysis (Cavrois et al., 2017; Sen et al., 2014) was conducted to quantitate the extent of HIV-induced remodeling in infected cells from endometrium and PBMCs, and reported as the SLIDE score. Scores > 1.2 indicate presence of strong viral-induced remodeling with a least 20% fold-change. Significant remodeling (p<0.0001) was observed in both PBMCs and ETs. However, the extent of remodeling was not significantly different in the two compartments (p=0.2251) as determined by two-sample t-tests on the remodeling ratios corresponding to each infected cell. See also Supplementary file 3 and Materials and Methods.
Figure 2—figure supplement 4. Comparison of uninfected and bystander cells from PBMCs and ETs.

Figure 2—figure supplement 4.

(A) t-SNE plots of uninfected and bystander cells from PBMCs or ETs. The PBMC and endometrial specimens were run within the same t-SNE. (B) SLIDE analysis revealing significantly higher remodeling in the productively-infected (HSA+) than in the bystander (HSA-) Tm cells from both blood and endometrium. *p<0.05, **p<0.01 as assessed using the Student’s paired t test. (C) SLIDE analysis revealing that endometrial Tm cells treated with conditioned media from F4.HSA-infected cultures in the presence of ART are less remodeled than productively-infected cells. Each set of line corresponds to one of two donor specimens. (D) t-SNE plots of endometrial Tm cells from two donors treated, in the presence of ART, with conditioned media from uninfected or F4.HSA-infected cultures. The Tm cells exposed to conditioned media from infected cultures reside in similar regions of the t-SNE as those exposed to conditioned media from uninfected cultures, and in different regions than that occupied by productive-infected cells.
Figure 2—figure supplement 5. Validation of susceptibility of Tm subsets to HIV infection.

Figure 2—figure supplement 5.

(A) Histograms showing the expression of CCR7 on three populations of sorted Tm cells (CD3+CD4+CD45RA-) expressing different levels of CCR7, a marker of Tcm, for two independent donors. (B) The sorted cells in panel A, along with total Tm cells, were exposed to F4.HSA and assessed for infection rates 3 days later. UI corresponds to Tm cells treated with media alone. Results are gated on live, singlet CD3+CD8- cells. The results from two donors demonstrate that infection rates inversely correlated with expression of CCR7, and validate the PP-SLIDE prediction that cells expressing high levels of CCR7 are poorly permissive to F4.HSA. Shown are the results from 2 donors representative of a total of 4. (C) The proportions of CD4+ T cells that were Tn (CD45RO-CD45RA+), and of memory CD4+ T cells (CD45RO+CD45RA-) that were Tem (CCR7-CD62L-) or Tcm (CCR7+CD62L+), were compared between uninfected endometrial and blood specimens. **p<0.01 as assessed using the Student’s paired t test. n.s.: not significant.
Figure 2—figure supplement 6. Antigens differentially expressed between PRE cells from PBMCs vs endometrium.

Figure 2—figure supplement 6.

(A) Histograms showing the expression of select antigens differentially expressed between PRE cells from PBMCs versus endometrium for one representative donor of 4 total. Results are gated on live, singlet Tm cells. (B) Box plots showing the parameters differentially expressed between PRE cells from unstimulated PBMCs vs. endometrium. Results are gated on live, singlet Tm cells and correspond to data from 4 independent donors. *p<0.05 as assessed using the Student’s paired t-test and adjusted for multiple testing using the Benjamini-Hochberg for FDR. (C) Heatmap representations of t-SNE plots showing expression levels of select antigens differentially expressed between PRE cells from PBMCs versus endometrium for 1 representative donor of 4 total. Also shown are the expression levels of the same antigens in the total population of uninfected Tm cells. The pairs of t-SNE plots at the top depict the cells as large black dots to facilitate assessment of location of the PRE cells. Results are gated on live, singlet Tm cells.
Figure 2—figure supplement 7. HIV-susceptible Tm cells from endometrium are less activated than HIV-susceptible Tm cells from stimulated PBMCs.

Figure 2—figure supplement 7.

(A) Expression levels of activation markers on uninfected Tm cells from PBMCs and ETs. *p<0.05 as assessed using the Student’s paired t test and adjusted for multiple testing using the Benjamini-Hochberg for FDR. n.s.: non-significant. (B) Proportions of Tm cells from uninfected PBMCs and ETs that expressed CCR5. ***p<0.001 as assessed using the Student’s paired t test. (C) Unstimulated ETs, stimulated PBMCs, or unstimulated PBMCs, isolated from the same donor, were mock-treated or exposed to F4.HSA and monitored 3 days later for levels of productive infection. Results are gated on live, singlet CD3+CD8- cells. Numbers reflect the percentages of HIV-infected cells. (D) t-SNE plots corresponding to the uninfected T cells and HIV-infected cells from the indicated PBMC or endometrial cultures. (E) t-SNE plots of uninfected and PRE cells from the indicated PBMC or ET cultures showing the distribution of Tm, Tn, memory CD8+ T cells, and naïve CD8+ T cells. Note that PRE cells from all 3 samples are Tm, but these Tm were phenotypically distinct from one another, as suggested by their residing in different regions of the t-SNE. (F) Heatmap showing expression levels of activation markers differentially expressed in uninfected (UI) and PRE Tm cells from the indicated ET or PBMC cultures. Compared with PRE cells from the endometrium, those from stimulated PBMCs expressed higher levels of most activation markers in the CyTOF panel. Results are gated on live, singlet Tm cells. (G) Histogram showing expression levels of CCR5 on UI and PRE Tm cells from ETs or stimulated PBMC cultures. Endometrial UI and PRE Tm cells expressed higher levels of CCR5 than the corresponding populations from stimulated PBMCs.
Figure 2—figure supplement 8. Validation of CyTOF antibodies through comparison of antigen expression on immune subsets.

Figure 2—figure supplement 8.

Tonsillar HLACs provide a good source of cells for panel validation because T cell antigens are often differentially expressed in B cells, and tonsils provide an abundant source of both T and B cells (Cavrois et al., 2017). Shown are data representative of one of 10 independent donors. (A) As demonstrated schematically in the upper left, the top population of each plot corresponds to T cells (CD3+) while the bottom population corresponds to B cells (CD3-). Each antibody in the panel was validated by standard two-dimensional plots showing expression of CD3 (y-axis) versus each indicated antibody (x-axis). The observed expression patterns are consistent with known expression patterns on T cells and B cells. Results are gated on live, singlet cells. (B) Some antigens were further validated by demonstrating differential expression in naïve (Tn) vs. memory (Tm) CD4+ T cells. As demonstrated schematically on the left, in the first row the top population of each plot corresponds to Tm cells while the bottom population corresponds to Tn cells, while in the second row the top population of each plot corresponds to Tn cells while the bottom population corresponds to Tm cells. Results s are gated on live, singlet, CD3+CD8- cells.