Mammary intraepithelial lymphocytes and intestinal inputs shape T cell dynamics in lactogenesis

Abstract

Pregnancy brings about profound changes in the mammary gland to prepare for lactation, yet immunocyte changes that accompany this rapid remodeling are incompletely understood. We comprehensively analyzed mammary T cells, revealing a marked increase in CD4⁺ and CD8⁺ T effector cells, including an expansion of T cell receptor (TCR)αβ⁺CD8αα⁺ cells, in pregnancy and lactation. T cells were localized in the mammary epithelium, resembling intraepithelial lymphocytes (IELs) typically found in mucosal tissues. Similarity to mucosal tissues was substantiated by demonstrating partial dependence on microbial cues, T cell migration from the intestine to the mammary gland in late pregnancy and shared TCR clonotypes between intestinal and mammary tissues, including intriguing public TCR families. Putative counterparts of mammary IELs were found in human breast and milk. Mammary IELs are thus poised to manage the transition from a nonmucosal tissue to a mucosal barrier during lactogenesis.

Fig. 1 |. Late gestation and lactation lead to increased T cell populations in the mammary gland.

A, Quantification of total number of CD45+ cells normalized to mammary gland weight across stages of gestation and lactation by flow cytometry. N, nulliparous (n = 6); G12, gestation day 12 (n = 3); G17, gestation day 17 (n = 5); L3, lactation days 3–5 (n = 9); I, involution, 1 day post-weaning (n = 6). B, Representative proportions of major immune cell types in the mammary gland across stages, N (n = 5), G (n = 3), L (n = 5) and I (n = 5) quantified by flow cytometry. C, Uniform Manifold Approximation and Projection (UMAP) projection of mammary T cells. Split by stages, N, G17, L3 and I (right). The representative UMAP is from one of three independent experiments, n = 3. D, Feature plots of CD8αα T, CD4 Teff and Tn cells from C. E, Dot plot of selected highly expressed genes in T cell clusters across stages identified in C. Dot size represents the percentage of cells expressing the selected gene and color indicates expression level. F, G, Quantification by flow cytometry of cell numbers (F) and proportions (G) of T cell populations identified in C normalized to mammary gland weight. T cell populations were determined as CD4 Teff, CD4+CD44+CD62L−; CD8αα T, CD8α+CD8β−CD44+CD62L−; CD8αβ Teff, CD8α+CD8β+CD44+CD62L−; and DN, TCRβ+CD4−CD8α−. N (n = 8, n = 6 for DN); G (n = 6); L (n = 10); and I (n = 5). Two-tailed unpaired Student’s t-tests were performed on the results shown in A, F, G. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are representative of ≥3 independent experiments. Bars in plots indicate mean ± standard error of the mean (SEM).

Fig. 2 |. Mammary T cells are intraepithelial lymphocytes.

A, Summary UMAP projections of T cells from mammary gland, large intestine, spleen and small intestine of lactating mice. B, UMAP and feature plots showing the transcriptional localization of featured T cell signatures. C, Representative flow cytometry plots and quantification of CD103+Ly49+CD8αα+ T cells (gated on live CD45+TCRβ+CD8α+CD8β−) across gestation and lactation stages in the mammary gland. N (n = 7), G17 (n = 5), L3–5 (n = 6) and I, 1 day post-weaning (n = 5). D, Proportion of CD8αα+ (left) and CD8αβ+ (right) cells that express CD160, CD38, CD244 and CD103 in N (n = 5) and L (n = 6) mammary glands. E, Representative immunofluorescence images of the mammary gland at N, G17, L3 and I. Epithelial cells (Krt8+ luminal cells in magenta and Krt14+ basal cells in red) and T cells, CD8α in yellow, CD8β in cyan (top) and CD4 in cyan (bottom). Scale bars, 20 μm. Two-tailed unpaired Student’s t-tests were performed on the results shown in C, D. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are representative of ≥3 independent experiments. Bars in plots indicate mean ± SEM.

Fig. 3 |. Putative interaction networks between mIELs and epithelial cells.

A, B, Chord diagrams showing potential signaling pathways upregulated in lactation from mIEL populations to epithelial cells (A) and from epithelial cells to mIEL populations (B). Ligand–receptor pairs as summarized into functionally related signaling pathways. Outer thicker bars represent the cell population that is the source or target of the signaling pathway in the chord diagram. The inner thinner bar color is the target of the signal. The thickness of the edge represents the signaling strength (communication probability) as calculated by CellChat. C, Dot plot showing the communication probabilities of ligand–receptor pairs upregulated in lactation from mIELs to epithelial cells. Heatmap depicts the communication probability of each ligand pair for each cell pair in N and L3 mammary glands. Sender and receivers are indicated by the color bars on top. P values computed by CellChat v.2 from a one-sided permutation test. D, E, Dot plot of transcript expression levels in mIELs (D) and epithelial cells (E) in N and L3 mammary glands, depicting the percent of cells expressing and mean expression levels in each cell population. Data are generated using scRNA-seq of epithelial and mIEL populations in N (n = 2) and L3 (day 3 postpartum) (n = 2) mammary glands.

Fig. 4 |. Intestinal T cells migrate to the mammary gland during late gestation and lactation.

A, Representative flow cytometry plots and quantification of mIEL populations in control B6 (n = 6) and thymectomized (n = 6) lactating mice. mIEL populations were determined as CD4, CD4+CD44+CD62L−; CD8αα, CD8α+CD8β−CD44+CD62L−; CD8αβ, CD8α+CD8β+CD44+CD62L−; and DN, TCRβ+CD4−CD8α−. B, Experimental design for Kaede experiments. Intestines of Kaede-positive mice were photoconverted from green to red by illumination with UV light after laparotomy in mid–late pregnancy and early lactation, and migration of red cells to the mammary gland was examined after 24 h. C, Representative flow cytometry plots and quantification of Kaede red cells within T cell populations (gated on TCRβ+ followed by either CD4+, CD8β+, CD8α+CD8β− or DN) in the mammary gland of mice 24 h post-photoconversion of the intestine. Controls are non-photoconverted mice (n = 3), ‘mid’ are mice photoconverted on gestation day 10 and analyzed on gestation day 11 (n = 3) and ‘late/L’ represents mice both photoconverted on gestation day 16 and analyzed on gestation day 17 and mice photoconverted on lactation day 1 and analyzed on lactation day 2 (n = 4). Two-tailed unpaired Student’s t-tests were performed on the results shown in A and C. *P < 0.05. Data are representative of ≥3 independent experiments. Bars in plots indicate mean ± SEM. Panel B created with BioRender.com.

Fig. 5 |. mIELs share peculiar TCR repertoires with small intestinal T cells.

A, Rarefaction analysis from TCR sequencing of T cell types between nulliparous and lactating mice in the mammary gland (MG), small intestine (SI), spleen (Spl) and large intestine (LI). B, Quantification of the number of cells with repeated clonotypes between organs in N and L3 mice. Each color represents a unique clonotype. C, Chord diagrams of L3 and N stages representing clonotype sharing in different IEL populations (inner ring) between the small intestine and mammary gland (outer ring). Each line represents a TCR clonotype. D, Distance matrix between αβTCR clonotypes in iIELs. Red circle denotes ‘Newbury TCR’ and black circle denotes ‘Revere TCR’. E, CDR3 sequence of Revere and Newbury TCRs. F, Table representing instances of Revere and Newbury TCRs in CD8αα+ T cells in mammary gland and small intestine across different mice. G, Counts of Revere and Newbury TCRs in CD8αα+ T cells across multiple tissues. Data are representative of ≥3 independent experiments.

Fig. 6 |. mIEL-like cells are found in human breast and milk.

A, UMAP projection of mammary immunocytes from human breast tissue (sourced from Kumar et al.). B, Feature plots of selected genes projected on UMAP from A. C, Representative flow cytometry gating of CD8α+CD103+ T cells and CD8αα+ IEL-like cells in human milk samples. D, Quantification of CD4+, CD8αα+ and CD8αβ+ cells as percent of CD45+ cells (left) (n = 7) and cell number normalized to volume (right) (n = 8) in human milk samples. E, Proportion of human CD8αα IEL-like cells that express markers CD103 (n = 7), CD94 (n = 4) and NKG2D (n = 7) in human milk samples. Two-tailed unpaired Student’s t-tests were performed on the results shown in D and E; no significance was observed. Data represent ≥7 independent milk samples/experiments. Bars in plots indicate mean ± SEM. scRNA-seq data were from Kumar et al..

Fig. 7 |. mIEL expansion in the lactating mammary gland is partially dependent on microbes.

A, Representative hematoxylin and eosin (H&E) staining of lactating mammary glands from SPF and GF mice. Scale bars, 100 μm. B, C, Quantification of the average number (B) and average area (μm2) (C) of alveoli per x20 image fields (five images per mouse of size 900×500 μm) of SPF and GF mammary glands (n = 4). D, Quantification of the total number of CD45+ cells normalized to mammary gland weight in SPF (n = 9), GF (n = 7) and GF conventionalized mice (n = 5). E, Quantification of total cell numbers normalized to mammary gland weight, of specified mIEL populations in SPF (n = 8), GF (n = 9, n = 6 for DN) and GF conventionalized (n = 6) mice by flow cytometry. mIEL populations were determined as CD4, CD4+CD44+CD62L−; CD8αα, CD8α+CD8β−CD44+CD62L−; CD8αβ, CD8α+CD8β+CD44+CD62L−; and DN, TCRβ+CD4−CD8α−. F, UMAP projection of mIELs from lactating SPF and GF mice with feature plots of naive, CD4 effector and CD8αα T cell gene signatures. G, Dot plot of differentially expressed genes in mIEL populations between lactating SPF and GF mice from F. Two-tailed unpaired Student’s t-tests were performed on the results shown in B–E. *P < 0.05, **P < 0.01 ***P < 0.001. Data are representative of ≥3 independent experiments. Bars in plots indicate mean ± SEM.