Peripheral immune characteristics and subset disorder in reproductive females with endometriosis

Abstract

Pathogenesis of endometriosis (EN) is still unknown, but growing evidence suggests that immune regulation may be important, and the pattern of peripheral immune changes in reproductive women with EN has yet to be fully explored. In this study, we conducted a comprehensive and systematic analysis of immune cell subsets within T cells, B cells, NK cells, and γδ T cells in peripheral blood (PB) samples from women with EN, women with uterine fibroids (UF) but without EN (UF-alone), and healthy controls using multi-parameter flow cytometry. Our findings revealed that UF, a common comorbidity of EN, exhibited similar peripheral immune features to EN, particularly in T cell and B cell immunity. Compared to healthy controls, we constructed the peripheral immune profile of EN. This profile highlighted that the immunopathogenic factors in EN predominantly relate to the immune disorder of B cells and their subsets, as well as the functional abnormalities within immune cell subsets of CD4⁺ T cells, CD8⁺ T cells, and γδ T cells. Moreover, using the random forest (RF) machine-learning method, we developed a diagnostic model that can effectively identify the patients with EN from healthy controls. The immune factors identified within this model could be pivotal for unraveling the immune pathogenic mechanisms of EN. Our study is the first to present a comprehensive depiction of the circulating immune features in EN, although the detailed roles and underlying mechanisms of these immune factors in the context of EN require further investigation.

Keywords: diagnosis; endometriosis; flow cytometry; immune cell subset; peripheral immune profile.

Figure 1

Gating strategies of flow cytometry and significances of immune indexes between patients with EN patients and healthy controls. (A) The representative dot diagrams showed the gating strategy of T cells, B cells, NKT cells, and NK cells; (B) statistical results indicated these subsets in the EN patients and the healthy controls. (C) The representative dot diagrams showed the gating strategy of different T cell subsets; (D) statistical results indicated these subsets in the EN patients and the healthy controls. (E) The representative dot diagrams showed the gating strategy of different NK cell subsets; (F) statistical results indicated these subsets in the EN patients and the healthy controls. (G) The representative dot diagrams showed the gating strategy of different B cell subsets; (H) statistical results indicated these subsets in the EN patients and the healthy controls.

Figure 2

Gating strategies of flow cytometry and significances of CD4⁺ and CD8⁺ T cell subsets between EN patients and healthy controls. (A) The representative dot diagrams showed the gating strategy of different CD4⁺ and CD8⁺ T subsets representing functional state of the T cells; (B–D) comparison of the percentages of the above subsets of CD4⁺ and CD8⁺ T cells in the EN patients and the healthy controls. (C) The representative dot diagrams showed the gating strategy of different CD4⁺ and CD8⁺ T subsets representing distinct effector T subsets; (F–H) comparison of the percentages of the above subsets of the CD4⁺ and CD8⁺ T cell subsets in the EN patients and the healthy controls.

Figure 3

Gating strategies of flow cytometry and significances of γδ T cell subsets between patients with EN and healthy controls. (A) The representative dot diagrams and histograms showed the gating strategy of γδ T cells; (B–E) comparison of the percentages of γδ1⁺ and γδ2⁺ T cells, and the frequencies of NKG2D, PD-1, NKP30, and NKP46 in the EN patients and the healthy controls.

Figure 4

Peripheral immune indexes detected by multiparameter flow cytometry. (A) The overview of the significantly changed immune indexes between the patients with EN and the healthy controls (HC), and between the different groups from EN and non-EN patients. Up, the percentages of immune indexes were increased in the front group; down, the percentages of immune indexes were decreased in the front group. The significantly changed immune subsets among the EN patients, UF-alone patients, and HC (B), among EN-alone patients, UF-alone patients, and HC (C), among EN-alone patients, EN patients with UF (EN + UF) and HC (D). *P<0.05, **P<0.01, ***P<0.001,****P<0.0001.

Figure 5

The diagnostic models of distinguishing the patients with EN from healthy controls. (A) The ROC curve of the efficiency of the diagnostic model constructed by one of random samplings using 70 immune indexes. (B) The AUC distribution of ROC curves from 1000 random samplings of the efficiency of the diagnostic model with 70 immune features. (C) The importance ranking calculated by the random forest model of the 70 immune cell subsets. (D) The average AUC distribution of ROC curves of the efficiencies of diagnostic models constructed by top important immune features. (E) The AUC distribution of ROC curves from 1000 random samplings of the efficiency of the diagnostic model with top 16 important immune features. (F) The importance ranking calculated by the random forest model for the top 16 important immune features.

Figure 6

The relationship between the percentages of immune indexes and chronological age of healthy controls or the patients with EN. (A, B) The immune indexes that with significant correlations between their percentages and chronological age in healthy cohorts and in patients with EN. P < 0.05, Spearman correlation analysis; red, positive correlation; blue, negative correlation. (C, D) Comparison of the percentages of the immune indexes that with significant correlations between the younger and older individuals in the healthy cohorts and in the patients with EN, Mann-Whitney U tests.

Figure 7

Analysis of the relationship between peripheral immunity in EN patients and their clinical stage and symptoms. The significantly changed immune indexes between EN patients with I + II stages (Early EN) and III + IV stages (Advanced EN) (A), between EN patients with and without pain symptoms (B), and between EN patients with and without abnormal uterine bleeding (C).