Sex differences exist in adult heart group 2 innate lymphoid cells

To investigate the differences in heart ILC2s between male and female mice, we collected Percoll-enriched heart lymphocytes from 3- and 8-week-old male and female mouse hearts. The gating strategy for heart ILC2s is shown in Fig. 1A. At 3 weeks of age, the percentage and number of heart ILC2s were indistinguishable between male and female mice (Fig. 1B). At 8 weeks of age, both the percentage and the total number of heart ILC2s were higher in female mice than in age-matched male mouse hearts (Fig. 1C).

Based on Klrg1 expression, a recent study reported that the numbers of Klrg1⁻ ILC2s in the bone marrow and lungs are downregulated by androgen [12]. Therefore, we further investigated the sex differences in Klrg1⁻ and Klrg1⁺ ILC2s. The percentages and numbers of heart Klrg1⁻ ILC2s were higher in female mice than in male mice at the age of both 3 and 8 weeks when CD45⁺Lin⁻CD127⁺CD90.2⁺ST2⁺ ILC2s were assessed. For heart Klrg1⁺ ILC2s, the percentage was decreased, but the numbers were increased in 8-week-old mice, but a decreased percentage and a decreasing trend in the number of heart Klrg1⁺ ILC2s were found at 3 weeks of age (Fig. 1D–E).

In addition, we also measured the protein levels of Gata3 in heart ILC2s and found that CD45⁺Lin⁻CD127⁺CD90.2⁺ST2⁺ cells showed over 90% Gata3 expression, and the percentage and numbers of heart CD45⁺Lin⁻CD127⁺CD90.2⁺ST2⁺Gata3⁺ cells [9] were also higher than those in male mice (Additional file 1: Fig. S1A). To further illustrate heart ILC2s exist sex difference, we also gated CD45⁺Lin⁻CD90.2⁺CD127⁺Gata3⁺ for ILC2 as done in another published article [13], and the data showed that the percentage of ILC2s among CD45⁺ cells were higher in female mice than male mice (Additional file 1: Fig. S1B). The above data suggest that sex hormones may play an extrinsic role in determining the number of heart ILC2s, especially the numbers of Klrg1⁻ ILC2s and Klrg1⁺ ILC2s in the heart.

We next investigated the sex differences in murine heart ILC2 phenotypes, including surface markers, transcription factors, and proliferation. In total heart ILC2s, the geometric mean fluorescence intensity (gMFI) of Klrg1 and Sca-1 was lower, while CD90.2 (Thy1.2) was higher in 8-week-old female mice than in age-matched male mice (Fig. 2A). There were no sex differences in the gMFIs of other surface markers, including Icos, CD25 (IL-2Rα), CD127 (IL-7R) and ST2 (IL-33R), or in that of the transcription factor Gata3 (Fig. 2A). In addition, female mouse heart ILC2s had a stronger proliferation ability than male mice, which was reflected by increased Ki67⁺ cells (Fig. 2B). These results demonstrated that mouse heart total ILC2s had distinct sex differences in terms of the surface expression of Klrg1, Sca-1, CD90.2 and proliferation ability.

Because our results showed that the numbers of Klrg1⁻ ILC2s and Klrg1⁺ ILC2s in the heart exhibited sex differences, we next determined the sex differences in heart Klrg1⁻ or Klrg1⁺ ILC2 phenotypes. Our results showed that the gMFI of CD90.2 was increased, and the gMFI of Sca-1 was decreased in both Klrg1⁻ and Klrg1⁺ ILC2s in the heart of female mice, but the gMFI of Klrg1 was also decreased in Klrg1⁺ ILC2s in female mouse hearts (Fig. 2A). The number of Ki67-positive cells was higher in both Klrg1⁻ ILC2s and Klrg1⁺ ILC2s in female mouse hearts (Fig. 2B). There were no sex differences in CD25, Icos or Gata3 among heart Klrg1⁻ ILC2s and Klrg1⁺ ILC2s (Fig. 2A).

ILC2s are known to produce the main type 2 cytokines, IL-4 and IL-5, after stimulation with IL-25, IL-33 and TSLP. To investigate the expression of these cytokines in male and female mouse heart ILC2s, we stimulated heart lymphocytes with 50 ng/ml IL-33 for 4 h and then determined the production of IL-4 and IL-5 by the ILC2s through flow cytometry. Our results showed that heart ILC2s showed a lower percentage of IL-4 in female mice than in age-matched male mice (Fig. 3A), but the total number of IL-4⁺ ILC2s and the gMFI of IL-4 exhibited no significant difference between male and female mouse heart ILC2s (Fig. 3A). There was no significant difference in IL-5 production, regarding the percentage, total number and gMFI, in heart ILC2s between male and female mice (Fig. 3B). The above data suggest that IL-4 and IL-5 secreted by heart ILC2s are not affected by sex hormones.

ILC2s are thought to be regulated by Treg cells that also express ST2 [10, 11], and sex hormones are known to affect Treg cells [6]. As such, we also determined the percentage and number of heart ILC2s in male and female Rag2^−/− mice, which lack functional T and B cells, at the age of 8 weeks old. The data showed a higher percentage and number of total heart ILC2s in female mice than in male mice (Fig. 4A). The percentages and numbers of heart Klrg1⁻ and Klrg1⁺ ILC2s between male and female Rag2^−/− mice showed trends similar to those observed for WT mice (Fig. 4B). Interestingly, the number of total heart ILC2s in Rag2^−/− mice was almost threefold higher than that in WT mice (male mice: 258.89 ± 157.84 vs. 1000.6 ± 220.18; female mice: 570.11 ± 310.08 vs. 1612.00 ± 321.78). The increased number of heart ILC2s in Rag2^−/− mice was mainly contributed by Klrg1⁻ ILC2s. This difference might be attributed to Treg cells, partly because Treg cells repress ILC2 functions directly or compete for IL-33 in heart tissue [10].

Because our results showed that the phenotypes of heart ILC2s and Klrg1⁻ ILC2s and Klrg1⁺ ILC2s, including Klrg1 and CD90.2, were different in male and female WT mice, we also determined the sex difference of heart Klrg1⁻ or Klrg1⁺ ILC2s in phenotypes. Our results showed that the gMFI levels of Klrg1 were lower but CD90.2 was higher in female Rag2^−/− mice (Fig. 4C), which was similar in WT mice. Moreover, the gMFI level of CD90.2 were increased in both subsets of female mice heart, but the gMFI level of Klrg1 were decreased in female mouse heart Klrg1⁺ILC2s, and the gMFI level of CD127 was higher in only female mouse heart Klrg1⁻ ILC2s (Fig. 4C). There was no sex difference in ST2 among heart Klrg1⁻ ILC2s and Klrg1⁺ ILC2s (Fig. 4C).

We further explored whether the sex differences in heart ILC2s are dependent on cell intrinsic sex hormone receptor expression levels or extrinsic cytokines from local heart tissues. Previous studies have found that lung ILC2s showed higher expression levels of androgen receptor (Ar), lower expression of estrogen receptor 1 (Esr1) and no expression of Esr2 [14], and sex hormones, such as androgen, play an extrinsic role in determining the numbers of lung ILC2s and ILC2 progenitors [15, 16]. To this end, we measured the mRNA expression of Ar, Esr1 and Esr2 in both male and female mouse heart ILC2s and found that the mRNA levels of Ar, Esr1 and Esr2 were undetectable in both male and female mouse heart ILC2s (Fig. 5A). Consistently, the percentages and surface marker expression of heart ILC2s from both male and female mice were not changed after stimulation with different concentrations of 17β-E2 (estrogen) or testosterone (androgen) for 12 h (Additional file 1: Fig. S2A-S2D). Interestingly, there was no significant difference between Klrg1, CD90.2 and Sca-1 expression levels in heart ILC2s between female and male mice after in vitro coculture. This suggests a cell-extrinsic factor existed in the heart tissue may contribute to the sex differences of heart ILC2s.

To further explore the extrinsic factors, such as cytokines, that respond to sex differences in heart ILC2s, we determined the mRNA levels of IL-33, which is reported to maintain ILC2 homeostasis and expansion in the heart tissues of both male and female mice in a homeostatic state [9, 17, 18]. As IL-33 is mainly expressed by nonhematopoietic cells, such as fibroblasts, epithelial cells, and endothelial cells [19], we used whole heart tissue for RNA preparation. The data revealed relatively higher mRNA levels of IL-33 in female mouse heart tissue than in male mouse heart tissue (Fig. 5B). Collectively, these data suggested that the sex difference in heart ILC2s is not dependent on sex hormone receptor expression but might be associated with the higher levels of IL-33 within the whole heart tissue of female mice.

To further explore the sex difference of heart ILC2s at the single-cell level, we reanalyzed the single-cell RNA sequencing (scRNA-seq) data by using two published datasets together, one containing 2 [20] and one containing 4 cardiac nonmyocytes samples [21], both of which were pooled from one paired WT female and male heart tissue, for scRNA-seq. After initial quality control checks, a total of 5793 CD45⁺ cells were acquired and subdivided into 8 main clusters based on marker gene expression across all cells by uniform manifold approximation and projection (UMAP) analysis (Additional file 1: Fig. S3A). Cell plots with red circles were attributed to ILC2s, which was confirmed by the high expression of Gata3 and Il7r (Additional file 1: Fig. S3A-S3B). Because these scRNA-seq samples were mixtures of cells derived from both male and female mice, we isolated male and female CD45⁺ cells based on the expression of the female-specific gene Xist (X-inactive specific transcript) [21] (Additional file 1: Fig. S3C). The data showed relatively higher percentages of ILC2s among CD45⁺ cells in female mice than in male mice (Fig. 6A), which is consistent with our flow cytometric results. Due to the limited cell numbers of heart ILC2s acquired from the published data, we only identified Xist and Tyrosylprotein sulfotransferase 2 (Tpst2) as differentially expressed genes (DEGs) in heart ILC2s from male and female mice (adjust P < 0.05) (Fig. 6B and Additional file 2: Table S1). Although without statistical significance, DEAD-Box helicase 3 Y-linked (Ddx3y) (30.8% of total male ILC2), Killer cell lectin like receptor K1 (Klrk1) (25.6% of total male ILC2s), and Il18 receptor 1 (Il18r1) (25.6% of total male ILC2s) were only detected in male ILC2s, while Fragile X mental retardation autosomal homolog 1 (Fxr1) (27.4% of total female ILC2s) and ADP ribosylation factor like GTPase 5b (Arl5b) (33.9% of total female ILC2s) were mainly detected in female ILC2s (Additional file 2: Table S1). Gene set variation analysis (GSVA) revealed that the glycine, serine and threonine metabolism pathway was upregulated, while other signaling pathways, such as the B-cell receptor signaling pathway and alpha-linolenic acid metabolism, were downregulated in female heart ILC2s compared with male heart ILC2s (Fig. 6C).

To explore the heterogeneity of heart ILC2s in both female and male mice, heart ILC2s were further subdivided into 2 clusters, Cluster 0 and 1 (Fig. 6D). Due to the limited cell numbers of heart ILC2s acquired from the published data, only 3 DEGs were found between the two clusters after adjusting P values, including Semaphorin 4a (Sema4a), Thioredoxin interacting protein (Txnip), and Hypoxia-inducible lipid droplet-associated (Hilpda) (adjust P < 0.05) (Additional file 2: Table S2). Cluster 1 showed relatively lower expression of Sema4a and Txnip and higher expression of Hilpda than Cluster 0 (Fig. 6E). In addition, Cluster 1 also showed relatively higher expression of Gata3, Areg, Icos, and Kit but lower expression of Klrg1 than Cluster 0, although without statistical significance (Fig. 6E). The percentages of Clusters 0 and 1 were 35.71% and 64.29%, respectively, in male heart ILC2s and 55.56% and 44.44%, respectively, in female mouse heart ILC2s (Fig. 6F left panel). After calculating the percentage of Cluster 0 and 1 heart ILC2s among CD45⁺ cells, we found that the percentage of Cluster 0 heart ILC2s among CD45⁺ cells showed no difference between female and male mice, whereas the percentage of Cluster 1 heart ILC2s among CD45⁺ cells was obviously higher in female mice than in male mice (Fig. 6F right panel). Due to limited cell numbers of heart ILC2s in this dataset, we did not find any DEGs after adjusting P values between male and female mice in cluster 0 and 1 except Xist (Additional file 2: Table S3-S4). Although without statistical significance, 149 genes were mainly detected in cluster 0 of female, including C–C motif chemokine ligand 3 (Ccl3) (25% of female in cluster 0) (Additional file 2: Table S3) that has been reported to have sex difference in the hippocampus of rats [22]. And 26 genes were only detected in cluster 1 of female, such as Peroxisome proliferator activated receptor gamma (Pparg) (22% of female in cluster 1) (Additional file 2: Table S4), which has been found that Pparg-regulated genes were upregulated in female adipose tissue compared with male adipose [23]. These findings suggest that the higher percentage and greater numbers of ILC2s in female heart tissue are mainly due to more Cluster 1 cells.

Male and female wild-type (WT) C57BL/6 mice were maintained in the Hunan Children’s Research Institute pathogen-free animal facility of Hunan Children’s Hospital (Changsha, Hunan, China). Mice were housed in cages (4–5 mice maximum per cage) at 22–25 °C and 50 ± 10% relative humidity with a 12-h light/dark cycle, periodic air changes, and free access to water and food. Congenic Rag2^−/− mice on the C57BL/6 background were obtained from Changzhou Cavens Laboratory Animal Co., Ltd. (Changzhou, Jiangsu, China). All animal procedures and protocols were approved by the Animal Ethics Committee of Hunan Children’s Hospital and followed the guidelines of the Institutional Animal Care and Use Committees of Hunan Children’s Hospital (Changsha, Hunan, China).

A single-cell suspension was prepared as described previously [3]. Mice were anesthetized with 2% pentobarbital sodium, and the heart was slowly perfused with cold phosphate-buffered saline (PBS) administered via the left ventricle with a 5-ml syringe to remove peripheral blood cells. Then, heart tissues were cut into approximately 1-cm² pieces and digested for 45 min at 37 °C in Hank’s solution containing 10% fetal bovine serum (FBS) (Biological Industry, Kibbutz Beit Hemek, Israel), 0.5 mg/ml collagenase I (Sigma‒Aldrich, St. Louis, MO, United States), and 0.5 mg/ml collagenase II (Gibco, Waltham, MA, United States). After digestion, the cells were resuspended in 20% Percoll (GE Healthcare, Pittsburgh, PA, United States) in RPMI 1640 medium (Biological Industry, Kibbutz Beit Haemek, Israel) containing 5% FBS, and a single mononuclear cell suspension was collected after centrifugation (2000 rpm, room temperature, 5 min).

The antibodies used for flow cytometry were commercially purchased and are listed in Table 1. For surface markers, single-cell suspensions derived from heart tissues were stained by incubating the cells with antibodies in staining buffer (PBS containing 2% mouse serum, 2% horse serum and anti-CD16/CD32 blocking antibodies (eBioscience, San Diego, CA, United States)) for 15 min at room temperature in the dark. For live-dead staining, cells were incubated with 7-AAD in apoptosis staining buffer (BioLegend, San Diego, San Diego, CA, United States) for 15 min at 4 °C after surface marker staining. Gata3 and Ki67 were stained as recommended by the manufacturer using the Foxp3/Transcription Factor Staining Buffer Set Kit (eBioscience, San Diego, CA, United States). The lineage (Lin) markers included CD3ε and CD19. Isotype-matched control antibodies were used at the same concentration as the corresponding test antibody. All flow cytometry experiments were carried out on a BD LSRFortessa (BD Biosciences, San Diego, CA, United States). Data were analyzed with FlowJo software (version 10.0; FlowJo LLC, Ashland, OR United States).

For analysis of the sex hormone receptor of heart ILC2s, single mononuclear cell suspensions were isolated from male and female mouse heart tissues according to the above description, and heart ILC2s were sorted by fluorescence-activated cell sorting (FACS) using a FACSAria III cell sorter (BD Biosciences, San Jose, CA, USA) after gating on CD45⁺Lin⁻CD127⁺CD90.2⁺ST2⁺ cells. The purity of heart ILC2s was measured with the gating strategy (Additional file 1: Fig. S4). Then, the sorted heart ILC2s with purity > 90% were used for RNA extraction with an RNAprep Pure Micro Kit (Tiangen Biotech, Beijing, China). To analyze the expression of IL-33 in heart tissue, total RNA was extracted from whole heart tissue from male and female mice using TRIzol (Invitrogen, Waltham, MA, United States), as described previously [3].

Total RNA was then reverse transcribed into cDNA using Evo M-MLV RT Premix (AG Biotechnology (Hunan), Changsha, China). Real-time qPCR was performed using a SYBR® Green Premix Pro Tag HS qPCR Kit (AG Biotechnology (Hunan), Changsha, China) with a Roche LightCycler® 480 II. Primer sequences used for qRT‒PCR were obtained from reported literature or designed by Pubmed Primer-BLAST. Primer sequences used for qRT‒PCR were designed by PubMed Primer-BLAST, including Ar forward, 5′-CAGGAGGTAATCTCCGAAGGC-3′; Ar reverse, 3′-ACAGACACTGCTTTACACAACTC-5′; Esr1 forward, 5′-CCCGCCTTCTACAGGTCTAAT-3′; Esr1 reverse, 3′-CTTTCTCGTTACTGCTGGACAG-5′; Esr2 forward, 5′-CTGTGATGAACTACAGTGTTCCC-3′; Esr2 reverse, 3′-CACATTTGGGCTTGCAGTCTG -5′. Primer sequences used for qRT‒PCR were obtained from the reported literature, including IL-33 forward, 5′-CCCTGGTCCCGCCTTGCAAAA-3'; IL-33 reverse, 3′-AGTTCTCTTCATGCTTGGTACCCGA-5′ [3]; GAPDH forward, 5′-AGGTCGGTGTGAACGGATTTG-3′; GAPDH reverse, 3’ TGTAGACCATGTAGTTGAGGTCA-5′.

For the analysis of the heart ILC2 response to sex hormones, single mononuclear cell suspensions isolated from heart tissues were stimulated with 0, 0.1, 1 and 10 nM 17β-estradiol (17β-E2) and testosterone (Medchem Express, Monmouth Junction, NJ, United States) supplemented with 10 ng/mL IL-2 and IL-7 (BioLegend, San Diego, San Diego, CA, United States) to survival for 12 h [9] and then stained for surface markers. To rule out sex hormone effects from the culture medium and FBS, we used charcoal dextran stripped FBS (Serana Europe GmbH, Brandenburg, Germany) and phenol red free culture medium (Boster Biological Technology, Wuhan, China) in this experiment. All flow cytometry experiments were carried out on a BD LSRFortessa. Data were analyzed with FlowJo software.

For intracellular IL-5 and IL-4 staining, single-cell suspensions isolated from heart tissues were stimulated with 50 ng/ml IL-33 (BioLegend, San Diego, San Diego, CA, United States) plus BD Golgi Plug protein transport inhibitor (BD Biosciences, San Diego, CA, United States) for 4 h and then stained for surface markers. After washing, the cells were fixed with the Fixation/Permeabilization Solution Kit (BD Biosciences, San Diego, CA, United States) following the manufacturer’s instructions and stained with anti-IL-4 or anti-IL-5 antibodies. Isotype-matched control antibodies were used at the same concentration as the corresponding test antibody. All flow cytometry experiments were carried out on a BD LSRFortessa. Data were analyzed with FlowJo software.

A total of 6 cardiac nonmyocyte samples pooled from one paired WT female and male mouse heart tissue were downloaded from the ArraryExpress database [20, 21] for downstream analysis. Cell Ranger version (version 6.0.1) was used to process raw sequencing data. The Seurat R package (version 4.0.5) was applied in our study to convert the scRNA-seq data as a Seurat object [43]. Cells that expressed fewer than 300 genes or more than 5000 genes or more than 20% mitochondrial genes were removed at the quality control step. Data were then normalized by the SCT transform R package (version 0.3.2) [44]. Next, we used the “RunPCA” function to reduce the dimensionality of the scRNA-seq data. Subsequently, we used the “RunPCA” function to conduct UMAP analysis. We also used the “Find Clusters” and “Find All Markers” functions to conduct cell clustering analysis and detect gene expression markers. Afterward, we used the Single R package and Cell Marker dataset to annotate the cell types in our study [45]. The “subset” function was also applied to extract the subcluster for downstream analysis and then annotated and analyzed as described above. Cells with Xist gene expression ≥ 0.1 were identified from female mice. To analyze the DEGs between female and male ILC2s, the value of the logFC threshold was set to 0.25 to filter the DEGs (Adjust p value < 0.05), and the heatmap was produced by the Complex Heatmap R package (version 2.11.1). A total of 186 KEGG pathway were download from the Molecular Signatures Database (MSigDB, http://​www.​gsea-msigdb.​org/​gsea/​msigdb/​index.​jsp) by the package “msigdbr”, with the permission from Kyoto Encyclopedia of Genes and Genomes (KEGG, https://​www.​kegg.​jp/​) website [46‐49]. GSVA analysis was performed to assess these KEGG pathway activation in the GSVA R package [50]. The parameters of msigdbr were set as follows: species = “Mus musculus”, category = “C2”, subcategory = “KEGG”.

All data are expressed as the mean ± SD, and statistical analyses were performed with SPSS software for Windows (Version 23, SPSS Inc., Chicago, IL, United States). Statistical analysis was performed with an unpaired Student’s t test for comparisons of two independent experimental groups. If the litter effect was very obvious among independent experiments, two-way ANOVA (sex and litter) followed by Dunnett’s test was used [51]. In each analysis, there were n = 5–12 replicates per group, and the results are representative of at least two independent experiments. Statistical significance was defined as p < 0.05. The number of mice used in each group is indicated in the figure legends. All graphs were produced by GraphPad Prism 8.0 for Windows software (GraphPad Software Inc., La Jolla, CA, United States).