Optimizing viable leukocyte sampling from the female genital tract for clinical trials: an international multi-site study

Abstract

Background: Functional analysis of mononuclear leukocytes in the female genital mucosa is essential for understanding the immunologic effects of HIV vaccines and microbicides at the site of HIV exposure. However, the best female genital tract sampling technique is unclear.

Methods and findings: We enrolled women from four sites in Africa and the US to compare three genital leukocyte sampling methods: cervicovaginal lavages (CVL), endocervical cytobrushes, and ectocervical biopsies. Absolute yields of mononuclear leukocyte subpopulations were determined by flow cytometric bead-based cell counting. Of the non-invasive sampling types, two combined sequential cytobrushes yielded significantly more viable mononuclear leukocytes than a CVL (p<0.0001). In a subsequent comparison, two cytobrushes yielded as many leukocytes (∼ 10,000) as one biopsy, with macrophages/monocytes being more prominent in cytobrushes and T lymphocytes in biopsies. Sample yields were consistent between sites. In a subgroup analysis, we observed significant reproducibility between replicate same-day biopsies (r = 0.89, p = 0.0123). Visible red blood cells in cytobrushes increased leukocyte yields more than three-fold (p = 0.0078), but did not change their subpopulation profile, indicating that these leukocytes were still largely derived from the mucosa and not peripheral blood. We also confirmed that many CD4(+) T cells in the female genital tract express the α4β7 integrin, an HIV envelope-binding mucosal homing receptor.

Conclusions: CVL sampling recovered the lowest number of viable mononuclear leukocytes. Two cervical cytobrushes yielded comparable total numbers of viable leukocytes to one biopsy, but cytobrushes and biopsies were biased toward macrophages and T lymphocytes, respectively. Our study also established the feasibility of obtaining consistent flow cytometric analyses of isolated genital cells from four study sites in the US and Africa. These data represent an important step towards implementing mucosal cell sampling in international clinical trials of HIV prevention.

Figure 1. Comparison of immune cell yield and distribution in endocervical cytobrush (CB) and cervicovaginal lavage (CVL) samples.

(A) Representative gates for identification of CD45⁺ leukocytes from cytobrush and CVL samples. (B) Numbers of CD45⁺ leukocytes, CD4⁺ T cells, CD8⁺ T cells, CD19⁺ B cells, CD14⁺ macrophages, and CD19^neg/HLA-DQ⁺ DC enumerated from cytobrush and CVL samples. Cell numbers were log₁₀ transformed, and plotted per site (Chicago, red dots; Nairobi, blue dots; Seattle, black dots). Horizontal bars indicate median values for each site. (C) Percentage contribution of each cell subset, as well as of cells that do not fit one of the described populations (“unknown”), to the total CD45⁺ population from cytobrush and CVL samples. Data are averaged across sites. P values are listed in the text, and median and IQR in Table 1.

Figure 2. Comparison of immune cell yield and distribution in endocervical cytobrush (CB) and ectocervical biopsy samples.

(A) Representative gates for identification of CD45⁺ leukocytes from cytobrush and biopsy samples. (B) Numbers of CD45⁺ leukocytes, CD4⁺ T cells, CD8⁺ T cells, CD19⁺ B cells, CD14⁺ macrophages, and CD19^neg/HLA-DQ⁺ DC enumerated from cytobrush and biopsy samples. Cell numbers were log₁₀ transformed, and plotted per site (Chicago, red dots; Nairobi, blue dots; Seattle, black dots). Horizontal bars indicate median values for each site. (C) Percentage contribution of each cell subset, as well as of cells that do not fit one of the described populations (“unknown”), to the total CD45⁺ population from cytobrush and biopsy samples. Data are averaged across sites. P values are listed in the text, and median and IQR in Table 2.

Figure 3. Inter-site variability and sample reproducibility from female genital tract immune cell sampling techniques.

(A) Percentage contribution of each immune cell population to the total CD45⁺ population recovered from CVL, cytobrush, and biopsy samples collected at the Seattle, Chicago, and Nairobi study sites. (B) Numbers of CD45⁺ leukocytes recovered from cytobrush samples collected from all women recruited in Chicago. Filled circles and connecting lines indicate women participating in both Part 1 and Part 2 of the study (n = 12). Open circles indicate women sampled only in one part of the study. Boxes indicate median, interquartile range and total range. (C) Scatterplot of cell numbers from the twelve women sampled in both Part 1 (x axis) and Part 2 (y axis) of the study. (D) Scatterplot of cell numbers from replicate biopsies collected at the same clinic visit in seven participants at the Nairobi study site. Biopsies were taken from upper left and upper right quadrants of the ectocervix.

Figure 4. Percentages of α4β7⁺ immune cells isolated from cytobrush (CB) and biopsy samples at the Seattle (black dots) and Chicago (red dots) study sites.

Percent α4β7 expression on CD4⁺ T cells, CD8⁺ T cells, CD19⁺ B cells and CD19^neg/HLA-DQ⁺ DC. Samples are only shown if more than 50 cells were acquired for the indicated population. Horizontal bars indicate median values for each site.