Enumeration, Phenotyping, and Clinical Associations of Tissue‐Resident T Cells in the Ecto‐ and Endocervix of Women Attending a Colposcopy Clinic

ABSTRACT

Problem

Tissue‐resident memory (T_RM) cells represent important immune sentinels that mount rapid recall responses to pathogens and cancers. However, there are limited data in humans on genital tract T_RM collected by clinically feasible sampling methods, limiting a full understanding of their role in immunity and clinical disease.

Method of Study

We used flow cytometry and single cell RNA sequencing (scRNAseq) to characterize T cells isolated from ectocervical biopsies and endocervical cytobrushes collected from women attending a colposcopy clinic in Winnipeg, Canada.

Results

The ectocervix generally contained a higher frequency and abundance of immune cells and T‐cells compared to the endocervix. CD4+ and CD8+ T_RM were more approximately 5‐times more frequent and abundant in the ecto‐ compared to endocervix, even after accounting for higher T‐cell recovery from the ectocervix. Phenotypically, CD4+ T_RM showed higher Th17‐ and comparable regulatory‐associated marker expression compared to non‐T_RM in both the ecto‐ and endocervix. Cervical dysplasia and ectropion were both associated with several immune cell differences in the ecto‐ and endocervix including lower CD4+ T_RM. Single‐cell RNAseq analyses confirmed broad CD69 and core T_RM‐related gene expression and captured several heterogeneous CD4+ and CD8+ T_RM subsets with diverse gene expression and pathways associated with host immunity, homeostasis, and nonimmune cell interactions.

Conclusions

Our data suggest that T_RM are more abundant in ecto‐ versus endocervical samples, which may reflect differences in commonly used sampling methods. Location and heterogeneous expression profiles underscore the need to better understand their role in microbial interactions, inflammation, and genital infection susceptibility in women.

1. Introduction

A novel subset of long‐lived tissue resident memory T cells (T_RM) with infrequent recirculation patterns are increasingly being appreciated to play crucial roles in combating mucosal pathogens. In nonlymphoid tissues, T_RM cells mediate rapid antigen‐specific recall responses to reinfection, protect against host pathology, and mediate tissue repair during chronic infection [1, 2]. T_RM cells also perform immune surveillance in lymph nodes, where distinct subsets, derived from nonlymphoid T_RM, have been defined [3]. While often protective, persistent autoreactive T_RM have also been associated with detrimental autoimmune disease due to altered regulatory and effector T cell function [4]. T_RM cells are typically characterized by the constitutive expression of CD69 and/or co‐expression CD103. A core signature of 31 additional genes that is consistent between CD4+ and CD8+ T cells has been identified for human CD69+ T_RM cells [5] and this is generally shared across body sites, with some evidence of local tissue adaption [6]. The adhesion markers ITGAE (CD103) and ITGA1 (CD49a) are upregulated in T_RM cells, whereas sphingosine‐1‐phosphate receptor 1 (S1PR1), which prevents tissue egress, is downregulated [5, 6, 7, 8]. Given that residence appears to be a dominant mechanism used to rapidly respond to antigen reexposure, augmenting or blocking T_RM may represent attractive targets for therapeutic interventions that need to act in nonlymphoid tissues, which are major sites of disease manifestation [9, 10, 11].

While T_RM cells have been widely studied in lymphoid, intestinal, and lung tissues, studies have only recently focused on T_RM cells localized within the different anatomical regions of the human cervical mucosa such as the endometrium, endocervix, and ectocervix [12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26]. The cervicovaginal mucosa forms a crucial mucosal barrier to protect against sexually transmitted infections (STIs) such as HPV, HIV, HSV‐2, gonorrhoea, and chlamydia. The endocervical canal is covered by a single‐layered columnar epithelium and contains mucus‐producing glandular cells that can trap pathogens and prevent ascending infections. In contrast, the ectocervical mucosa is covered by an epithelium comprised of stratified squamous nonkeratinized cells. The transformation zone at the squamo‐columnar junction delineates the two distinct mucosal compartments of the female reproductive tract (FRT) with distinct physiologic functions and immune defence mechanisms. Sampling these mucosal tissues has proved to be challenging in humans and depending on the tissue sampled, the phenotype and cell recovery may differ [27]. Indeed, we have previously shown that sampling with less invasive endocervical cytobrushes resulted in similar total numbers of viable immune cells when compared to one biopsy, but with biases toward macrophages and T lymphocytes in cytobrushes and biopsies, respectively [28].

While studies of human genital T_RM cells have been limited, early studies in animal models showed that mucosal vaccination resulted in the rapid seeding of T_RM cells in the uterus, which, together with circulating memory T‐cells, were critical to clearance of genital Chlamydia trachomatis infection [29]. Furthermore, in herpes simplex virus type 2 (HSV‐2) models, T_RM in conjunction with epithelial cells were shown to clear infection in an antigen specific manner through cytokine production [30, 31]. More recently, comprehensive studies which performed both phenotypic and transcriptomic analyses, demonstrated that CD8+ T cells were more abundant than CD4+ T cells in ectocervical tissues [30]. These studies suggest that T_RM cells showed vast heterogeneity across genital tract tissues, with CD8+ T cells and Th17‐like CD103+CD69+ T_RM being most abundant in vaginal compared to cervical tissues [12]. Therefore, studies that compare in more detail the mucosal immune cell subsets at different anatomical regions within the FGT in relation to clinical outcomes could provide important insights into novel immunotherapeutic interventions for genital tract conditions.

Here, we enumerate and characterize cervicovaginal T_RM cells isolated from ectocervical biopsies and endocervical cytobrushes, two important female genital sampling techniques used in clinical studies. These techniques sample distinct anatomical locations within the cervix, potentially yielding different cell populations, and densities relevant to better understanding local immunity specifically in the context of female genital tract infections. We found important differences in the abundance and phenotypes of CD4+ and CD8+ T_RM cells at these two sites and our data support the hypothesis that T_RM cells may play important roles in genital conditions such as ectropion and cervical intraepithelial neoplasia (CIN), underscoring the need for in‐depth characterization of the genital milieu with respect to phenotype and its clinical associations to genital infections in women.

2. Results

2.1. Participant Demographics

We enrolled 50 adult women attending routine colposcopy visits at the Health Sciences Centre in Winnipeg, Canada. All participants provided informed written consent prior to study participation. The final analysis included n = 26 following exclusion of samples with insufficient yield for either flow cytometry or scRNAseq analyses (Table 1). The age range of the included participants was 18–40 years. Women had a median of 3 sex acts within the past month (IQR: 0–5, Table 1). Of those who were sexually active, 20% (5/26) reported always using condoms during sex, while 60% (15/26) reported not using condoms. Contraceptives were used by 69% (18/26) of participants, with approximately one‐third using either intrauterine devices (IUD) (38.9%, 7/18) or oral pills (27.8%, 5/18), and the remainder using other methods such as DMPA, tubal ligation and condoms.

TABLE 1.

Participant demographics.

			Sample
Characteristic	N	Total N = 26	10x 7 (26.9) [a]	Flow cytometry 20 (76.9) [a]
Age (range)	(18–40)
Number of pregnancies	26
0		13 (50.0)	3 (42.9)	10 (50.0)
1+		13 (50.0)	3 (42.9)	10 (50.0)
Number of live births	26
0		18 (69.2)	4 (57.1)	14 (70.0)
1+		8 (30.8)	2 (28.6)	6 (30.0)
Current use of contraceptives	26	18 (69.2)	6 (85.7)	12 (60.0)
Type	18
Injection, depo‐provera		1 (5.6)	0 (0.0)	1 (7.1)
IUD		7 (38.9)	1 (14.3)	6 (42.9)
Oral pill		5 (27.8)	2 (28.6)	3 (21.4)
Other (DMPA, tubal ligation, condoms)		5 (27.8)	1 (14.3)	4 (27.6)
Unusual urinary tract infection symptoms since last year	26	14 (53.8)	1 (14.3)	13 (65.0)
Number of sex acts in the past month	24	3 (0–5)	3 (2–4)	3 (0–6)
How often do you use condom?	25
Never		15 (60.0)	3 (42.9)	12 (63.2)
Sometimes		5 (20.0)	1 (14.3)	4 (21.1)
Always		5 (20.0)	2 (28.6)	3 (15.8)
Ever treated for STI?	26	12 (46.2)	4 (57.1)	8 (40.0)
Used antibiotics/Anti‐fungals in the past month	26	7 (26.9)	2 (28.6)	5 (25.0)
Vaginal discharge	25
Mucoid		13 (52.0)	4 (57.1)	9 (47.4)
Thin/Watery		12 (48.0)	2 (28.6)	10 (52.6)
Ever screened for cervical cancer	25	15 (60.0)	5 (71.4)	10 (52.6)
Ectropion	25	6 (24.0)	2 (28.6)	4 (21.1)
Dysplasia	25	16 (64.0)	5 (71.4)	11 (57.9)
CIN stage	21
Normal		3 (14.3)	1 (14.3)	2 (13.3)
Low grade lesion (CIN I, ASCUS, LSIL)		16 (76.2)	4 (57.1)	12 (80.0)
High grade lesion 9 CIN II/III)		1 (4.8)	0 (0.0)	1 (6.7)
Undetermined		1 (4.8)	1 (14.3)	0 (0.0)
Treated	19	3 (15.8)	1 (14.3)	2 (14.3)

^a n (%)

Half of participants (50%, 13/26) reported at least one prior pregnancy while 53.8% (14/26) reported urinary tract infection symptoms such as pain, bleeding, itching, presence of yeast, and urinary tract infections within a year prior to study enrolment. Almost half (46.2%, 12/26) indicated that they had been previously treated for a STI. Sixty percent (15/26) of participants reported having ever been screened for cervical cancer with approximately two‐thirds (64%, 16/25) having a history of abnormal cell growth on the cervix/dysplasia, and 24% (6/20) had cervical ectropion, a condition where the cells lining the cervical canal turn outward to cover part or all the ectocervix. Twenty‐one participants (84%, 21/26) were CIN staged with 76% (16/21) presenting with low grade lesions (CIN I) while 4.8% (1/21) had higher grade lesions (CIN II/III) and 15.8% (3/19) had or were undergoing treatment (Table 1).

2.2. Immune Cell Distribution in the Endo‐ and Ectocervix

We used an optimized flow cytometry panel to examine the distribution of immune cell subsets in matched endocervical, cytobrush, and ectocervical biopsy specimens (Figure 1). Based on the expression of CD45+, immune cells comprised a higher proportion of live gated events in ectocervix (median 16.6%, IQR: 12.1%–35.2%) compared to the endocervix (median 3.4%, IQR: 1.5%–5.3%, p < 0.001). A similar trend was observed for the absolute counts in the ecto‐ and endocervix, respectively (median 9042, IQR: 4804–15320 versus median of 3620, IQR: 825–8746, p = 0.146, paired T‐test, Figure 2A). Most immune cells in the ectocervix (median 89.1%, IQR: 83.1%–92.9%) were CD3+ T cells, compared to a median of 66.0% (IQR: 55.6%–77.0%) in the endocervix (p < 0.001). Absolute counts in both compartments showed a similar trend (median 8036, IQR: 3574–12382 versus a median of 2347, IQR: 481–5750, p = 0.061). CD4+ and CD8+ T‐cells were proportionally more frequent in the ectocervix compared to the endocervix (median 43.8%, IQR: 37.9%–49.1% versus median of 31.8%, IQR: 10.8%–41.7%, p = 0.002). Absolute counts of both CD4 (median 3778, IQR: 1636–6894 versus a median of 236, IQR: 64–1852, p = 0.002) and CD8 subsets (median 2624, 1695–5255 versus median 335, IQR: 125–1894, p = 0.003) were also higher in the ecto‐ versus endocervix (Figure 2B) consistent with previous findings [12].

FIGURE 1.

Representative plots showing gating strategy deployed to interrogate cervical T cell populations in FlowJo. (A–B) Cells were initially gated based on size, granularity, and doublets exclusion (C) live immune cells were then identified based on the expression of CD45 and negative for the amine‐reactive live‐dead stain. (D–E) T cells were defined by CD3 expression and subsequently delineated into CD4+ helper and CD8+ cytotoxic subsets. Tissue resident memory cells (T_RM) were gated based on canonical markers CD103 and CD69, with (F) CD4 T_RM being CD103+ CD69+ and or CD103‐CD69+ and (G) CD8 T_RM being CD103+ CD69+ (H) in‐depth analysis of CD4 T cells was performed to identify Th17‐like cells based on expression/ co‐expression CD161 and CCR6 in the top panel and regulatory T cells (CD25^hiCD127^lo) in the bottom panel.

FIGURE 2.

Distribution of immune cells in the endo‐ and ectocervix: (A) CD45+ (B) CD3+ (C) CD4+ and (D) CD8+ with cell frequencies in the top panel and counts in the lower panel. p value ≤ 0.05 is significant, nonparametric Wilcoxon matched pairs signed rank test.

2.3. Tissue Resident Phenotype Distribution of Cervical Immune Cells

CD4+ T_RM defined based on CD69 and CD103 co‐expression were more frequent in the ectocervix (median 11.9%, IQR: 4.7%–17.7%) compared to the endocervix (median of 2.8%, IQR: 0.8%–5.3%, p = 0.005, Figure 3A). Some literature suggests CD69+ CD103‐ CD4+ T‐cells are also tissue resident, and we observed that these cells were also increased in frequency in the ecto‐ (median 69.4%, IQR: 61.0%–75.1%) compared to the endocervix (median 47.1%, IQR: 22.3%–57.6%, p < 0.001, Figure 3B). Similar associations in proportions were observed for CD8+ T_RM (CD69+ CD103+) in ectocervix (median 51.2%, IQR: 26.8%–58.4%) compared to the endocervix (median 9.7%, IQR: 4.0%–34.4%, p < 0.001, Figure 3C). The absolute numbers of each subset were also higher in the ecto‐ compared to endocervix (p < 0.0001 for all Figure 3 lower panel).

FIGURE 3.

Distribution of CD4 and CD8 Tissue resident memory T cells in the endo‐ and ectocervix: (A) CD4+ CD69+ T_RM (B) CD4+ CD103+ CD69+ T_RM, and (C) CD8+ CD103+ CD69+ T_RM frequencies (top panel) and counts (lower panel). p value ≤ 0.05 is significant, nonparametric Wilcoxon matched pairs signed rank test.

We next profiled the T helper phenotype of CD4+ T_RM cells (CD103+ CD69+) and the potential CD4+ T_RM subset (CD103‐ CD69+) (Figure S1) and compared these to non‐T_RM cells (CD103‐ CD69‐). The co‐expression of regulatory T cell markers CD25^hiCD127^lo were similar between T_RM and non‐T_RM in both compartments (p > 0.05, Figure S1A). Expression of the Th17 marker CD161 was higher on T_RM compared to non‐T_RM both compartments (p < 0.05), while expression of or co‐expression of CCR6 was similar for T_RM and non‐T_RM in the ectocervix and endocervix (p > 0.05, Figure S1 B–D). Significant differences in CD4 T cells expressing the gut‐homing integrin complex α4β7 (fluorescence minus one gated, Figure S2) were observed in both counts and frequencies, with higher expression in the ectocervix (median 74.2%, IQR: 57.5%–96.7% in ecto‐ versus median 52.6%, IQR: 38.7%–75.2% in endocervix, p < 0.001, Figure S3). Based on CD103 expression only, a marker associated with lymphocyte retention in tissues [32], we found that the ectocervix had more CD103‐expressing T cells (both absolute counts and relative frequencies), while the endocervix had a higher proportion of these T cells (Figure S4).

2.4. Association of Cervical T Cell Subsets With Clinical Manifestations

We next compared absolute T cell subset counts and frequencies by flow cytometry in ectocervical and endocervical samples, stratified by the clinical manifestation of dysplasia and the anatomical observation of ectropion. Notably, in the endocervix, clinically observed ectropion was associated with lower CD45+ immune cell abundance (median 13.6%, IQR: 9.8%–16.4% vs. median 18.3%, IQR: 13.0%–35.7%, p = 0.032), CD4+ abundance (median 56, IQR: 19–157 versus median 381, IQR: 126–2212, p = 0.028), number of CD4+ CD103‐CD69+ cells (median 27, IQR: 2–62 versus median 151, IQR: 64–839, p = 0.014), and number of CD4+ CD161+ CCR6‐ Th17‐like cells (median 2, IQR: 2–10 versus median 111, IQR: 40–738, p = 0.017). CD4+ T_RM frequencies were also lower in those with ectropion (median 0.6%, IQR: 0.0%–1.1%) compared to those without (median 2.9%, IQR: 1.8%–5.7%, p = 0.038, Table 2). Furthermore, ectropion was associated with lower endocervical CD8+ frequencies (median 13.0%, IQR: 5.6%–20.8% versus median 39.4%, IQR: 18.6%–51.4%, p = 0.009) and counts (median 50, IQR: 28–118 versus median 765, IQR: 267–2226, p = 0.029), as well as lower ectocervical CD8+ T_RM counts (median 698, IQR: 520–825 versus median 1424, IQR: 574–3231, p = 0.007, Table 2). In contrast, dysplasia was associated with higher ectocervical CD4+ frequencies (median 46.7, IQR: 39.8%–59.0% versus median 42.4%, IQR: 33.8%–44.9%, p = 0.043) and CD4+ α4β7 T cell frequencies (median 95.0%, IQR: 74.2%–97.8% versus median 57.2%, IQR: 55.0%–67.8%, p = 0.036, Table 2). No significant associations were observed with other T cell phenotypes examined (Table 2).

TABLE 2.

Associations of cervical mucosa immune cell subsets with clinical characteristics.

T cell Phenotype	Clinical manifestation		Ectocervix		Endocervix
% CD45+	Dysplasia	No	16.6%, 13.4%–22.4%	0.854	3.6%, 2.4%–5.8%	0.367
		Yes	15.7%, 11.8%–31.4%		3.3%, 1.4%–4.9%
% CD45+	Ectropion	No	18.3%, 13.0%–35.7%	0.032	3.6%, 1.5%–5.5%	0.31
		Yes	13.6%, 9.8%–16.4%		2.7%, 1.5%–4.2%
CD45 counts	Dysplasia	No	8906, 3619–15806	0.804	4016, 1716–9636	0.617
		Yes	7768, 5682–16522		3525, 982–7926
CD45 counts	Ectropion	No	11886, 3571–16522	0.692	4317, 1346–11321	0.073
		Yes	7026, 6195–15552		1988, 256–4364
% CD3+	Dysplasia	No	87.8%, 70.2%–91.4%	0.287	69.3%, 59.7%–76.0%	0.669
		Yes	91.5%, 86.9%–93.2%		65.6%, 44.5%–74.7%
% CD3+	Ectropion	No	90.1%, 85.9%–94.7%	0.246	72.8%, 58.8%–77.3%	0.298
		Yes	72.2%, 48.8%–89.3%		57.2%, 46.0%–63.0%
CD3+ counts	Dysplasia	No	6350, 2296–14199	0.902	2895, 584–6617	0.617
		Yes	7170, 5197–12529		2753, 448–5238
CD3+ counts	Ectropion	No	8902, 3348 ‐ 12529	0.943	3052, 599–8071	0.062
		Yes	6360, 4561–10836		1052, 139–2490
% CD4+	Dysplasia	No	42.4%, 33.8%–44.9%	0.043	25.8%, 9.9%–39.2%	0.715
		Yes	46.7%, 39.8%–59.0%		35.8%, 9.0%–42.9%
% CD4+	Ectropion	No	43.0%, 37.7%–51.1%	0.586	35.8%, 9.0%–42.9%	0.881
		Yes	46.2%, 43.1%–50.6%		24.3%, 13.8%–36.7%
CD4+ counts	Dysplasia	No	2787, 520–6071	0.543	1103, 68–2282	0.944
		Yes	4461, 2302–7008		201, 50–1048
CD4+ counts	Ectropion	No	3832, 1530–7008	0.971	381, 126–2212	0.028
		Yes	3526, 2082–5844		56, 19–157
% CD8+	Dysplasia	No	51.3%, 48.2% ‐ 59.7%	0.157	31.7%, 16.0% ‐ 55.3%	0.583
		Yes	41.0%, 34.5%–49.4%		24.2%, 14.6%–41.3%
% CD8+	Ectropion	No	49.5%, 39.1%–53.5%	0.419	39.4%, 18.6%–51.4%	0.009
		Yes	38.0%, 29.8% ‐ 46.7%		13.0%, 5.6% ‐ 20.8%
CD8+ counts	Dysplasia	No	2892, 1505–4904	0.974	996, 223–1894	0.639
		Yes	2400, 2267–5396		263, 118–1574
CD8+ counts	Ectropion	No	2847, 1693–6122	0.117	765, 267–2226	0.029
		Yes	2378, 2004–2822		50, 28–118
%CD4+ CD103+ CD69+	Dysplasia	No	18.4%, 8.7%–21.9%	0.185	2.0%, 0.8%–4.3%	0.445
		Yes	8.3%, 4.7%–14.2%		2.9%, 0.8%–4.0%
%CD4+ CD103+ CD69+	Ectropion	No	11.5%, 6.3%–18.4%	0.754	2.9%, 1.8%–5.7%	0.038
		Yes	8.7%, 1.0%–19.2%		0.6%, 0.0%–1.1%
CD4+ CD103+ CD69+ counts	Dysplasia	No	264, 137–526	0.914	52, 1–120	0.217
		Yes	317, 197–738		5, 2–40
CD4+ CD103+ CD69+ counts	Ectropion	No	349, 204–804	0.223	8, 4–92	0.058
		Yes	121, 73–298		0, 0–2
%CD4+ CD103‐ CD69+	Dysplasia	No	66.0%, 48.3%–69.5%	0.211	48.5%, 25.1%–57.6%	0.591
		Yes	71.4%, 64.8%–77.7%		36.2%, 18.8%–54.0%
%CD4+ CD103‐ CD69+	Ectropion	No	71.4%, 65.3%–75.3%	0.219	47.8%, 22.8%–59.1%	0.189
		Yes	59.2%, 52.4%–65.2%		20.8%, 12.9%–33.9%
CD4+ CD103‐ CD69+ counts	Dysplasia	No	1938, 319–4010	0.432	421, 38–934	0.405
		Yes	2825, 1554–4412		87, 36–171
CD4+ CD103‐ CD69+ counts	Ectropion	No	2572, 1126–4415	0.588	151, 64–839	0.014
		Yes	2166, 1131–3428		27, 2–62
%CD8+ CD103+ CD69+	Dysplasia	No	55.6%, 26.8%–61.7%	0.568	7.7%, 2.4%–35.0%	0.712
		Yes	46.2%, 29.3%–53.4%		9.4%, 6.4%–25.1%
%CD8+ CD103+ CD69+	Ectropion	No	49.5%, 35.1%–56.8%	0.576	10.1%, 3.8%–31.8%	0.666
		Yes	33.0%, 10.4%–57.6%		5.3%, 4.0%–28.3%
CD8+ CD103+ CD69+ counts	Dysplasia	No	940, 624–3054	0.972	152, 2–578	0.183
		Yes	1365, 560–2690		47, 12–161
CD8+ CD103+ CD69+ counts	Ectropion	No	1424, 574–3231	0.007	112, 12–298	0.128
		Yes	698, 520–825		5, 1–68
% CD4+ Trm/CD127‐ CD25+	Dysplasia	No	0.9%, 0.5%–1.2%	0.392	0.1%, 0.0%–1.0%	0.248
		Yes	1.4%, 0.8%–4.5%		0.0%, 0.0%–0.0%
% CD4+ Trm/CD127‐ CD25+	Ectropion	No	1.0%, 0.6%–3.7%	0.712	0.0%, 0.0%–0.6%	0.116
		Yes	1.2%, 0.8%–3.9%		0.0%, 0.0%–0.0%
CD4+ Trm/CD127‐ CD25+ counts		No	2, 1–6	0.901	0, 0–1	0.225
		Yes	3, 2–14		0, 0–0
CD4+ Trm/CD127‐ CD25+ counts	Ectropion	No	5, 2–14	0.015	0, 0–1	0.152
		Yes	2, 1–2		0, 0–0
% CD4+ CD161+ CCR6‐	Dysplasia	No	42.2%, 34.4%–59.7%	0.314	30.3%, 0.4%–39.2%	0.91
		Yes	33.8%, 30.5%–38.2%		18.6%, 12.4%–28.2%
% CD4+ CD161+ CCR6‐	Ectropion	No	35.2%, 32.1%–52.6%	0.245	26.6%, 15.4%–34.5%	0.243
		Yes	35.4%, 29.8%–38.1%		8.3%, 5.1%–17.1%
CD4+ CD161+ CCR6‐ counts	Dysplasia	No	1430, 238–3522	0.789	373, 2–772	0.276
		Yes	1586, 876–2184		55, 15–184
CD4+ CD161+ CCR6‐ counts	Ectropion	No	1586, 497–2466	0.815	111, 40–738	0.017
		Yes	1122, 446–2232		2, 2–10
% CD4+ α4β7	Dysplasia	No	57.2%, 55.0%–67.8%	0.036	46.0%, 26.8%–53.7%	0.07
		Yes	95.0%, 74.2%–97.8%		54.7%, 45.4%–87.1%
% CD4+ α4β7	Ectropion	No	69.4%, 57.2%–95.7%	0.396	52.9%, 31.4%–69.7%	0.45
		Yes	96.9%, 84.7%–98.4%		59.6%, 44.0%–78.1%
CD4+ α4β7 counts	Dysplasia	No	1574, 475–3500	0.231	554, 50–1131	0.858
		Yes	3682, 1823–6557		110, 18–832
CD4+ α4β7 counts	Ectropion	No	2502, 982–6557	0.623	226, 53–1426	0.013
		Yes	3405, 2042–4491		42, 16–81

2.5. Single Cell Transcriptomics Analyses of Ectocervical and Endocervical Immune Cells

To better understand the cellular heterogeneity and molecular signatures of immune cells in the human cervical mucosa, scRNAseq analysis was performed. Prior to 10x genomics‐based cell capture for cDNA library generation, we aimed to enrich for CD45+ immune cells initially, using two different methods, including CD45+ cell sorting on the BD FACS Aria III Cell Sorter, then with the EasySep positive selection kit. With both methods, cell capture, and subsequent library preparation were unsuccessful for all samples tested (n = 20) due to low viable CD45+ cell counts (<10 000) post‐sort or ‐enrichment; thus, we did not proceed with library generation of these enriched samples for scRNAseq analyses. Because CD45+ cell enrichment approaches greatly compromised the cell numbers and integrity required for scRNAseq, we next used fresh single cell suspensions of subsequent specimens for capture and scRNAseq. Of the additional specimens (n = 10 participants), cell capture and sc‐cDNA library generation were successful for n = 7 participants, with one having paired successful endo‐ and ectocervical specimens. Three of the 10 participant samples were not analyzed due to their low integrity (i.e. bloody cytobrushes or low viable immune cell recovery).

On the successful samples (n = 2 endocervical cytobrushes and n = 6 ectocervical biopsies), we analyzed a total of 9001 single cells (Figure 4A) after initial quality control (see “Methods”). Almost half (n = 4006, 44.5%) of cells originated from the ectocervix, and 4995 (55.5%) cells originated from the endocervix (Figure 4C). Using high dimensionality reduction methods, these cells were classified into 28 different clusters (Figure S5A). Gene signatures in each cluster confirmed 10 putative lineage immune cell subsets predominantly, T cells (13.7%, 1230 in the ectocervix and 20.2%, 1816 in the endocervix marked by CD3δ, CD3ε, and CD3γ), monocytes (12.8%, 1149 in the ectocervix and 4.3%, 383 in the endocervix, marked by S100A9, S100A8, LYZ, and CD163), B cells (6.8 %, 609 in the ectocervix and 2.7 %, 244 in the endocervix, marked by MS4A1, CD19, CD27, and IRF4), pDC (0.5 %, 48 in the ectocervix and 2.3 %, 209 in the endocervix, marked by IRF4, IRF8, KLF4, and ZEB2), and neutrophils (1.5 %, 609 in the ectocervix, marked by CSF3R) (Figure 4E). Overall, stratified by compartments, and in agreement with the flow cytometry analyses, we found that T cells were the most abundant cell type in the endo‐ and ectocervix (20% and 13.7% respectively), while fibroblasts and endothelial cells were frequent in the endocervix (11.4% and 9% respectively) and monocytes, epithelial cells and B cells abundant in the ectocervix (Figure 4C).

FIGURE 4.

Diverse cell types in the cervical mucosa at single‐cell resolution. (A) Schematic of tissue dissociation, cell isolation, sequencing, and downstream bioinformatics analysis. (B) UMAP dimensionality reduction of all cell types identified in the ectocervix and endocervix. (C) Breakdown of cellular abundance and frequency in the ecto and endocervix. (D) Differential gene expression of all cell types identified in the cervical mucosa.

2.6. Cellular and Transcriptomic Signature Diversity of Cervical T_RM Cells

We next investigated the transcriptomic profiles of the various T cell subsets in the cervical mucosa. T cells were re‐clustered based on their gene expression profiles, which generated 22 T‐cell‐specific clusters. Based on marker expression, these T cells were further assigned to 11 subsets, notably: CD4+ Treg, Th1, Th17, CD4+ CD8+, 3 CD8+ clusters, three NKT clusters, and one naïve T cell cluster (Figure 5A). In agreement with our flow cytometry analyses, more than half (n = 1816, 59.6%) of the 3,046 CD4+ T cells derived from the ectocervix displayed high levels of CD69 (Figure 5B), the canonical marker of tissue residency, in addition to the Treg (FOXP3 and IL2RA), Th1 (IFNG, STAT4, and TNF), and Th17 (SELL, CCR5, CCR7, and TCF7) gene markers. (Figure 5C). Conversely, CD8+ T cell populations were heterogeneously distributed between compartments with CD8+pop1 and CD8+pop2 cells more abundant in the endocervix, while CD8+pop3 was evenly distributed in both the endo‐ and ectocervix. The gene ITGAE (CD103) was highly expressed in the CD8+pop1 and CD8+pop2, while GMZA and GMZB genes were enriched in Th1, CD8+pop2, and to a lesser extent CD8+pop3 cells (Figure 5C). In addition, we also found three T cell clusters mainly enriched in the endocervix with no CD4 or CD8 gene expression but highly expressed gene markers for NK cells including GZMA, GNLY, and NKG7 (Figure 5C).

FIGURE 5.

Cellular and molecular heterogeneity of T cells in the cervical mucosa. (A) UMAP shows preferential distributions of T cell types in the ectocervix and endocervix. (B) Absolute cell counts and frequencies in the ectocervix and endocervix for each T cell type. (C) Dot plots showing differential gene expression for each cell type.

We performed GO enrichment analyses by comparing significant up‐regulated gene sets from each T cell subset (Table S1). CD4+ Tregs were involved in immune tolerance and regulation, including key pathways such as Th17 cell differentiation and cytokine‐cytokine receptor interaction, which are critical in controlling inflammation and preventing autoimmunity in cervical tissues. Th1 cells showed enrichment in pathways including granzyme‐mediated programmed cell death signaling and positive regulation of lymphocyte chemotaxis, essential for responding to viral infections like HPV and controlling tumorigenesis [33, 34]. Th17 cells exhibited enriched pathways important in the regulation of T cell differentiation, B cell activation, and adaptive immune responses, which are crucial for managing inflammation in the cervical mucosa and responding to chronic HPV infection or lesions [34]. CD4+ CD8+ T cells were associated with roles in epithelial cell differentiation and morphogenesis, potentially influencing tissue repair and immune surveillance during HPV‐related dysplasia or ectropion. All three CD8+ T cell subsets are involved in acute inflammatory responses, apoptosis, and cell death, helping to control viral infections by directly killing infected cells, especially in the context of HPV persistence and dysplasia [35, 36].

Three NKT cell populations similarly demonstrated a wide range of functions critical for both immune surveillance and tissue response; NKT pop1 cells for membrane raft distribution and regulation of T cell‐mediated immunity, which is essential for controlling viral infections and preventing excessive inflammation in the cervix, NKT pop2 cells expressed collagen and extracellular matrix components, highlighting their role in tissue remodeling and immune responses to bacterial or viral stimuli, and NKT pop3 cells were associated with antibacterial humoral responses and epithelial cell morphogenesis [37]. Finally, naive T cells exhibited enriched pathways including T cell receptor gene diversification, T cell differentiation, and lymphocyte activation, which are involved in promoting effective immune responses during HPV infection and cervical dysplasia [38] (Table S1).

We next investigated the transcriptomic profile of T_RM by analyzing core transcriptional and functional gene signatures. Focusing on clusters that expressed T‐cell genes (CD3δ, CD3ϵ, and CD3γ), we found that all clusters (0,1,4,5,9,20, and 23) showed high expression of canonical tissue residency markers, including CD69, RGS1, KLF2, KLF3, DUSP6, ITGA1, ITGAE, STK38, SELL, CXCR6, and NPDC1 (Figures 6(A, B)). We validated our analysis on a publicly available dataset from Parthasarathy, et al. [18], on T cell clusters 8,10,14,16,19,21, and 29 (Figure 5B), confirming the robustness of our findings highlighting the conserved and context‐specific immune signatures in the FRT (Figures 6(A, B)).

FIGURE 6.

Transcriptomic signatures and predicted cellular interactions of tissue‐resident memory T cells (T_RM) in the cervix. (A) UMAP visualization comparing T cell clustering identified in the current study (Ssemaganda et al.) with a publicly available validation dataset (Parthasarathy et al. [18],) based on CD3δ, CD3ϵ, and CD3γ gene expression (B) Dot plot detailing the expression of core T_RM functional genes across the identified T cell clusters (C) Aggregated predicted cell‐cell communication network among cervical cell types and (D) outgoing signaling originating from T cells directed toward other immune and nonimmune stromal subsets.

To unravel factors driving tissue residency, we performed interaction analysis based on known ligand‐receptor pairs to predict biologically relevant cell‐cell communication networks in the cervical tissue [39]. Here, we found strong suggested interactions between T cells and several immune cell subsets, and with stromal nonimmune cells such as fibroblasts, epithelial cells, and keratinocytes that form the structural and functional basis of the tissue (Figures 6(C,D)).

3. Discussion

Understanding T‐cell profiles in human cervicovaginal tissues has important implications for understanding a broad range of conditions that affect women's health. By profiling immune cells isolated from biopsies and cytobrushes of women attending a colposcopy clinic, our current study offers significant insights into the immune cell compartmentalization within the FRT, specifically between the ectocervical and endocervical regions. Notably, the frequency and abundance of CD45+ immune cells were significantly higher in the ectocervix compared to the endocervix (Figure 1) while transcriptomic data (Figure 5(B)) showed a high abundance of CD45‐ nonimmune cells. Additionally, the predominance of CD3+ T cells among CD45+ immune cells further illustrate the adaptive immune potential within the ectocervix. One interpretation of these findings is that cytobrushes, which are frequently used to collect endocervical samples, may represent a more superficial specimen collection method that collects a higher frequency of non‐immune cells compared ectocervical biopsies, a more invasive sample type that nevertheless may provide important tissue immunology insights [40, 41, 42, 43]. Sampling deeper into the endocervix using methods such as endocervical curettage may address this issue but is invasive. The differential distribution of T cell populations may also be suggestive of site‐specific adaptation for heightened immune readiness in the ectocervix. Given its stratified squamous epithelium, the ectocervix serves as the first line of defense against STIs. In contrast, the endocervix is lined by columnar epithelium and is potentially more driven by the need to balance immune tolerance that prevents inflammation‐driven tissue damage, while simultaneously maintaining reproductive function. Delineating whether differences in our study are driven by sampling methods or biological variables will require additional studies that sample these tissues in a similar manner (despite the technical challenges associated with doing so).

We also observed a high density of CD4+ T cells in the ectocervix that were enriched with markers of tissue residency markers CD69 and CD103, with a high proportion of these TRM demonstrating regulatory and Th17‐like phenotypes (Figures 2, 3). We have previously shown in a cohort of female sex workers that higher regulatory T cells (Tregs) in the FRT were associated with lower inflammatory cytokines concentrations [44], suggesting that Tregs may protect the mucosal barrier and limit the ability of pathogens such as HIV to establish local infection. Additionally, Tregs have also been shown to regulate Th17 [45], further supporting the concept that the phenotype of CD4+ T_RM might indicate an immunoregulatory protective response within cervicovaginal tract.

We also observed elevated expression of the integrin α4β7 on CD4+ T cells in the ectocervix. This integrin plays a critical role in directing lymphocyte trafficking to mucosal tissues and could be suggestive of a heightened localized immune milieu that is driven by the unique microenvironment of the cervical tissue [46]. Understanding why the ectocervix harbours higher frequencies and counts of α4β7+ T cells compared to other regions of the cervical mucosa could open new avenues for targeted therapies aimed at modulating immune responses.

Our current findings align with the growing body of evidence suggesting that cervical immune cell populations are altered in conditions associated with abnormal cell growth [47, 48]. For instance, ectropion, a condition where the columnar epithelium is exposed to the ectocervical surface, often associated with hormonal influences like pregnancy and contraceptive use, has previously been associated with a reduction T cells stained from biopsies obtained from ectopic zones compared with healthy regions of the ectocervix [49, 50]. In line with this, our data also show that ectropion is correlated with a reduced immune cell abundance in the ectocervix (Table 2), which taken together could be indicative of a localized immunodeficiency to dampen immunosurveillance potentially impacting the susceptibility to infections.

Conversely, dysplasia, which represents a spectrum of precancerous changes in cervical cells often induced by HPV infection, has been shown to result in a notable increase in immune cell populations, particularly CD4+ T cells [50, 51, 52, 53, 54]. Furthermore, the increase in CD4+ α4β7 T cell frequencies (Table 1) observed in our study align with previous findings which showed that α4β7 integrin is upregulated in areas of inflammation that result from dysplasia, facilitating immune cell trafficking toward mucosal tissues [46]. This could be indicative that dysplasia may prompt recruitment of mucosa‐specific immune responses to maintain local immunological homeostasis and to counteract the risks posed by HPV infection. Although no HPV testing was performed, our current findings align with existing evidence suggesting a potential role for HPV in modulating immunosurveillance and cell‐mediated responses [55], representing a possible therapeutic intervention prior to onset of malignancies.

Our transcriptomic analyses of ecto‐ and endocervical tissues showed marked heterogeneity, with T cells being the predominant immune cell subset in both compartments. Consistent with previous findings in gut tissue, cervical T_RM had a high gene expression of CD69 (Figure 5), the most commonly used marker of tissue residency, in addition to the Treg markers FOXP3 and IL2RA [16], Th1 (IFNG, STAT4, and TNF), and Th17 (SELL, CCR5, CCR7, and TCF7) [56]. The Th17‐like phenotype that was consistent between our flow and scRNAseq data are consistent with previous findings and support the role of cervical Th17 in barrier integrity and tissue homeostasis [12]. Furthermore, our data demonstrated strong interaction of T cells with stromal nonimmune cells such as fibroblasts, epithelial cells, and keratinocytes. Epithelial cells express specific cell adhesion molecules such as Epithelial cadherin (E‐cadherin) which is a critical ligand for Integrins such as ITGAE establishing tissue residency [57, 58]. Fibroblasts on the other hand are known to secrete several chemokines, growth factors essential for the survival, maintenance, and programming of tissue‐resident cells [59]. Additionally, the stromal environment influences T‐cell interactions with other immune cells such as B cells, neutrophils, and pDCs which in turn is key to forming the long‐lived memory cells critical for a rapid response to subsequent infections in tissues [60, 61, 62].

Overall, these data suggest that the human cervical mucosa is home to diverse range of immune cell subsets with significant portion of these exhibiting T_RM ‐like characteristics, whose long‐term survival and function is dependent on strong interactions with non‐immune stromal cells. This cellular communication might be crucial for a rapid initial response to pathogens and subsequent establishment of long‐lasting immunity in the FGT.

A key limitation of our study was the cross‐sectional nature and relatively small sample size of participants recruited from a colposcopy clinic, which broadly impacts the generalizability and statistical power of the findings. In addition, our study stood the risk of selection bias, as individuals attending a colposcopy clinic may have specific gynecologic profiles not representative of the general population, potentially limiting the detection of subtle but clinically relevant differences in immune cell populations. Furthermore, our work studied premenopausal women and therefore were unable to interrogate the impact on aging and hormonal changes experienced during menopause and have been previously associated with suppression of cellular cytotoxicity and significant changes in both CD4 T cells and T_RM [22, 23]. Consequently, the clinical associations we observed may not be representative in a larger, more diverse cohort. Additionally, we experienced methodological challenges enriching for genital immune cells which limited our single cell transcriptomics dataset. This indeed limited our ability to directly compare and make inferences of our findings to previous human genital tract transcriptional studies which otherwise utilized other sample types such as hysterectomies as well as endometrial samples [18, 63]. This limitation nevertheless reflects a comparison of different sampling methods that are frequently used in the field, therefore providing important data about how sampling effects the cell populations that are obtained.

Collectively, these findings enhance our understanding of T_RM and their contribution to the genital immune milieu particularly in the context precancerous conditions emphasizing the need for routine screening and intervention. An in‐depth interrogation of the cervical immunoregulatory mechanisms could be critical in unravelling novel therapeutic interventions against these genital pathologies, ultimately improving prognosis and overall cervical health.

4. Methods

4.1. Participants, Specimen Collection and Ethics Statement

Women attending a colposcopy clinic located at the Health Sciences Centre, Winnipeg, Manitoba for their routine visit were approached to participate in this study, and no Clinical Study Registration Number is required for this study. To all consenting participants (n = 50), a questionnaire was administered to establish each participant's socio‐behavioral and reproductive characteristics. Following a full gynaecological examination, sequential sampling was then carried out starting with a vaginal swab for HPV testing followed cervical vaginal lavage performed by flushing the vaginal canal with 3 mL of sterile PBS and collecting into a 15 mL tube. Next, using a speculum, 2 consecutive endocervical cytobrushes were collected by inserting into the cervical os rotated 360° and collected into 10 mL PBS medium in separate 50 mL tubes. Lastly, 2 biopsies ∼2 mm each were collected from the upper left and right quadrant of the ectocervix and placed in 10 mL RPMI medium in 50 mL. All samples were stored on ice or in the fridge prior to transfer to the lab for processing.

4.2. Biopsy Digestion

Ectocervical biopsies were carefully placed in a sterile petri dish containing 1 mL of collagenase A (1.5 mg/mL),1% Pen Strep in RPMI medium and severed into small pieces with a sterile scalpel. Two milliliters of collagenase A solution were added and using an 18‐gauge needle and a 5 mL syringe the cut biopsy pieces were further disintegrated by plunging up and down 10 times into a 50 mL centrifuge tube. The petri dish was rinsed with 2 mL of enzyme solution and transferred into a 50 mL tube and incubated on a shaker plater at 200 rpm at 37⁰C for 45 min. Following digestion, the cell suspension was filtered through a 100 µm strainer into a new 50 mL tube and washed with 5 mL PBS 2% Fetal Bovine Serum (FBS) at 1600 rpm for 10 min ready counting and subsequent flow cytometry staining or transferred to the genomics lab for GEM preparation.

4.3. Endocervical Cytobrushes Processing

Endocervical cytobrushes were processed as previously described [29]. Briefly, the tube containing cytobrushes was vigorously vortexed to dissociate the cells and mucus from the brushes, rinsed with R10 medium (10% FBS in RPMI containing penicillin and streptomycin) and filtered through 100‐micron nylon cell strainer (Becton–Dickinson) fitted into a 50 mL tube. Cells were washed twice by centrifugation at 1600 rpm, and pellet resuspended in washing buffer (PBS 2% FBS) for counting and subsequent flow cytometry staining or GEM preparation for single cell transcriptomics analysis.

4.4. Flow Cytometry Staining

Single‐cell suspensions were washed with PBS and stained with LIVE/DEAD Fixable Dead Cell Stain (ThermoFisher) for 20 min on ice, followed by staining with an optimized antibody cocktail comprised of lineage markers CD45 (HI30, Biolegend), CD3 (UCHT1, Biolegend), CD4(OKT4, Biolegend), CD8 (RPA‐T8, Biolegend), Treg (CD25(M‐A251, Biolegend), CD127(A019D5, Biolegend), Th17(CCR6(11A9, Biolegend), CD161 (HP‐3G10, eBioscience), T_RM and integrins CD69(FN50, Biolegend), CD103(Ber‐ACT8 Biolegend), CD49d (9F10, Biolegend), β7(FIB27, Biolegend) for an additional 20 mins. Cells were then washed twice with FACs buffer (PBS containing 2% FBS), resuspended in perm wash buffer (e‐biosciences) and acquired on a BD LSR Fortessa flow cytometer (BD Biosciences). The acquired data were then analyzed using FlowJo version 10 (Becton–Dickinson Life Sciences).

4.5. EasySep CD45 Enrichment

CD45 enrichment was performed according to the manufacturer's instructions with minor modifications. Briefly, 50 µL of Selection Cocktail (STEMCELL Technologies) was added to 1 mL of cell suspension in a FACs tube and incubated at room temperature for 5 min. Releasable RapidSpheres (STEMCELL Technologies) were then added to the suspension at a concentration of 50 µL/mL and incubated for an additional 3 min at room temperature. Immunomagnetic separation was then performed by placing the tube into the magnet (STEMCELL Technologies) and incubated at RT for 5 min. In one continuous motion, the magnet and tube were inverted to discard the supernatant. The tube containing enriched suspension was removed from the magnet, release buffer (STEMCELL Technologies) added to dissociate CD45+ cells from the Releasable RapidSpheres, centrifuged at 1600 rpm for 10 min, supernatant carefully pipetted, and cells resuspended in 0.04% BSA/PBS for counting and subsequent GEM preparation for single cell transcriptomics analysis.

4.6. Single‐Cell (sc) cDNA Library Generation for scRNA‐Sequencing (scRNA‐Seq)

From each sample of ectocervical biopsies and endocervical cytobrushes, 10 000 cells were loaded onto the Chromium Next GEM Chip G (10x Genomics) according to the manufacturer's instructions. Cell lysis, single‐cell barcoding, and gel‐emulsion (GEM) generation were performed in the Chromium Controller (10x Genomics) using the Chromium Single Cell 3' v3.1 kit. Barcoded cDNA was amplified via PCR on the Veriti 96 Well Thermal Cycler and sc cDNA libraries were generated according to the manufacturer's instructions (10x Genomics). Quality control measures of the single‐cell cDNA libraries were performed on the 2100 Bioanalyzer (Agilent Technologies) and Qubit 4.0 Fluorometer (Thermo Fisher Scientific). Single‐cell cDNA libraries were sequenced (50 000 paired‐end reads, single‐indexing) on the NovaSeq 6000 at Genome Quebec.

4.7. scRNA‐Seq Data Processing and Determination of Major Cell Types

The raw read outputs were first demultiplexed, processed, and mapped to the GRCh38 human reference genome by using the Cell Ranger (version 3.0.1) [64] software. was utilized in the read alignment process. The data from all samples were read into the Seurat R package (version 4.3.0) [65] for further processing. The unique molecular identifiers (UMIs) were counted in each single cell. Cells that expressed between 200 and 5000 genes in more than 3 cells, with mitochondrial content less than 5 percent were included in the analysis. “DoubletFinder” software (version 2.0.3) [66] was used to remove doublets, identified by the top N cells with the highest pANN score for each library, where N represents the doublet rates in the 10x Genomics platform. After removing low‐quality and doublet cells, data of all samples was scaled, log‐normalized using a size factor of 10 000 molecules per cell and merged [64, 65, 66].

4.8. Dimensionality Reduction and Unsupervised Clustering

The normalized data was used to identify the top 2000 highly variable genes through the function “FindVariableGenes”. Principal component analysis (PCA) was next employed to reduce data dimensions. The best value of the number of PCs to be used in further unsupervised clustering analysis was obtained by taking a quantitative approach where the larger value between the point at which the principal components only contribute 5% of standard deviation and the principal components cumulatively contribute 90% of the standard deviation and the point where the percent change in variation between the consecutive PCs is less than 0.1%. A K‐nearest‐neighbor graph was constructed based on the Euclidean distance in the PCA space using the “FindNeighbors” function and Louvain algorithm was applied to iteratively group cells together by the “FindClusters” function with optimal cluster resolution. The “RunTSNE” and “RunUMAP” functions were applied for appropriate visualizations.

4.9. Cell Type Annotation

The algorithm‐based unbiased cell type recognition was then performed by using the SingleR R package (version 1.4.1) [67], which leverages various reference transcriptomic datasets from the “celldex” R package including “Human Primary Cell Atlas Data” for all cells and “Database Immune Cell Expression Data” for immune cells. The annotated cell clusters were checked by manually curated gene markers retrieved from the Cell Marker database [68] and published papers [16, 67]. Any discrepancies were reconciled using the strategy outlined by Clarke and colleagues [68]. The differential genes were then identified in each cell type with the following criteria: expressed in at least 25% of cells in either sample groups with log₂FoldChange > 0.25, and adjusted p value < 0.01.

4.10. Gene‐Set Enrichment Analysis

Gene‐set enrichment analyses were conducted using a well‐established and robust ShinyGO analysis pipeline [68]. Differential gene sets in each T cell cluster were used to compute enriched gene Ontology (GO) biological processes and results were tabulated to display the 50 differentially expressed genes for the T cell clusters.

4.11. Receptor‐Ligand Interaction Analysis

CellChat R package was utilized to construct the intercellular communication network in the cervix from scRNA‐seq data [37]. CellChat infers cell‐cell interactions via a ligand‐receptor pair database, subsequently calculating the signaling probability and strength between cell populations and generating a visualization of the overall cellular communication landscape.

4.12. Statistical Analysis

Statistical analysis and data visualization in the present study was performed by using the R software (version 4.2.2, R Foundation for Statistical Computing, Vienna, Austria; http://www.r‐project.org) and Prism version 10 (GraphPad Software, LLC). Unless specifically stated, p or FDR values < 0.05 were considered as statistically significant.

Supporting information

Supplemental Table 1: Differentially expressed genes (DEGs) and associated enriched GO biological process pathways of T_RM‐associated markers in T cell lineage clusters, identified through single‐cell transcriptomic analysis of endo‐ and ectocervical cells. GSEA outputs were generated using ShinyGO 0.82 (https://bioinformatics.sdstate.edu/go/) ⁷¹

Supplemental Figure 1: CD4 T_RM phenotypes in the endo‐ and ectocervix (A) CD4+ Tregs (B) CD4+ CD161+ CCR6+ (C) CD4+ CD161+ CCR6‐ (D) CD4+ CD161‐ CCR6+ with frequencies in the top panel and absolute counts in the lower Panel.

Supplemental Figure 2: Representative fluorescence minus one (FMO) gating showing gut homing expression of integrin (A) CD49d (α4) and (B) β7.

Supplemental Figure 3: T_RM homing based on the expression of α4β7. (A) CD4+ CD69+ T_RM (B) CD4+ CD103+ CD69+ T_RM and (C) CD8+ CD103+ CD69+ T_RM. For each of these subsets, the frequencies are presented in the top panels, while the absolute cell counts are shown in the bottom panels.

Supplemental Figure 4: CD103+ expression on CD4 top Panels and CD8 bottom panel in the endo‐ and ectocervix with (A) showing Frequencies, (B) Relative abundance and (C) absolute counts.

Supplemental Figure 5: Dimensionality reduction and unsupervised UMAP clustering of immune cell clusters based on expression of putative T cell genes CD3γ, CD3δ, and CD3ε for (A) our current study Ssemaganda et al., and (B) a publicly available dataset Parthasarathy et al., 2023.¹⁸

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.