The involvement of peritoneal GATA6+ macrophages in the pathogenesis of endometriosis

Endometriosis is a chronic inflammatory disease and affects approximately 10% of reproductive-aged women, representing ~190 million women worldwide ( 1 , 2 ). It is associated with debilitating chronic pelvic pain, dysmenorrhea, and dyspareunia that dramatically reduces the quality of life of women ( 3 – 6 ). Endometriosis is commonly classified into four stages according to the revised criteria from the American Society of Reproductive Medicine (rASRM) or American Fertility Society (AFS) based on lesion size, location, and the extent of adhesions ( 6 , 7 ). However, endometriosis-associated pain does not correlate with the staging system ( 6 , 8 , 9 ). Patients with stage I disease can have severe pain, while stage IV patients can be asymptomatic ( 2 , 10 ), indicating that several other factors contribute to disease symptoms. Rigorous prior research suggests that aberrant inflammation contributes to the onset and progression of endometriosis ( 11 – 21 ). Macrophages are considered to be key players in promoting disease progression ( 14 , 20 , 21 ), as abundant macrophages are present in ectopic lesions ( 22 ) and elevated in the peritoneal cavity ( 20 , 23 , 24 ). Transcriptionally and functionally dysregulated macrophages can establish an inflammatory environment by secreting cytokines and chemokines, which encourage lesion growth and progression ( 13 , 17 , 20 , 21 , 24 , 25 ) and contribute to endometriosis-associated pain ( 24 , 26 , 27 ). Peritoneal macrophages also contribute to the inflammatory condition by releasing cytokines and growth factors that stimulate lesion infiltration, growth, and vascularization ( 20 , 24 , 28 , 29 ). Indeed, depletion of peritoneal macrophages by clodronate liposomes reduced lesion progression ( 24 ). Macrophages are highly heterogeneous immune cells that play various roles in homeostasis, innate immunity, and tissue development and repair depending on the signals from their local environment ( 30 ). In the peritoneum, two major subsets of macrophages have been distinguished based on their ontogeny and differential expression of F4/80 and MHCII ( 31 ). Large peritoneal macrophages (LPM) expressing high F4/80 and low MHCII, F4/80 hi MHCII lo , are derived from embryonic progenitors in the yolk sac and are maintained in the peritoneum by self-renewal ( 32 ), while some replenishment from bone marrow (BM)-derived monocytes continuously occurs with aging ( 33 , 34 ). Small peritoneal macrophages stemmed from hematopoiesis in the BM, expressing low F4/80 and high MHCII, F4/80 lo MHCII hi ( 31 ). When inflammatory stimuli are induced in the peritoneum, large numbers of BM-derived monocytes are recruited and differentiate into monocyte-derived macrophages that secrete proinflammatory cytokines as an immune response ( 35 ). Monocyte-derived macrophages are transcriptionally and functionally distinct from tissue-resident macrophages even though they serve to replenish resident macrophages ( 36 , 37 ). Replenishment or reprogramming of monocyte-derived inflammatory macrophages to resident macrophages depends on the degree of inflammatory stimuli ( 37 ). Recent evidence ( 15 ) indicates that monocyte-derived macrophages may protect the peritoneum against lesion establishment, whereas endometrial macrophages are pro-endometriosis, suggesting further diverse functions of macrophage contribution in endometriosis pathophysiology. LPM represent most of the macrophage population in the peritoneum, and GATA6 is selectively expressed in LPM ( 38 , 39 ). See ( 21 ) https://kanakohayashilab.org/hayashi/en/mouse/peritoneal.immune.cells/ A zinc finger transcription factor, GATA6, is known to transcriptionally regulate residential LPM proliferation, differentiation, metabolism, and survival ( 38 – 40 ). LPM are essential as immunosurveillance patrollers in maintaining peritoneal homeostasis, as summarized in ( 41 , 42 ). In response to inflammatory stimuli such as infection or sterile inflammation, LPM undergo the macrophage disappearance reaction (MDR) at the early phase of the inflammatory reaction, in which LPM disappear from the peritoneum due to LPM death and/or adherence to the mesothelium ( 41 , 42 ). Fate-mapping studies exhibit that MDR, due to the death of large numbers of GATA6 +  LPM, induces significant replenishment with BM-derived monocytes ( 33 , 34 ). GATA6 + LPM also adhere to the mesothelium to repair sterile injury on the surface of visceral organs, such as the liver or intestine ( 43 , 44 ). They generally cannot interact with mesothelium under normal homeostasis as they are free-floating cells ( 45 ). In contrast, multiple studies have demonstrated that peritoneal macrophages contribute to disease advancement, such as ovarian cancer progression and metastasis ( 46 – 49 ), as deleting peritoneal macrophages reduces tumor progression and peritoneal metastasis ( 46 , 47 ). Gastric cancer patients bearing more peritoneal macrophages show poor prognosis ( 50 ). GATA6 + LPM invade liver metastases directly from the peritoneum and contribute to metastatic tumor growth and recurrence ( 51 ). Thus, GATA6 + LPM can also become adapted and act as tumor-associated macrophages ( 41 , 42 ). In endometriosis, CD14 hi peritoneal macrophages, which are the predominant resident population in the peritoneum in humans ( 52 ), are increased in women with endometriosis ( 9 ). However, CD14 hi peritoneal macrophages negatively correlate with the severity of pelvic pain, which is independent of disease outcome ( 9 ). We have previously reported that GATA6 + LPM were involved in the establishment of the aberrant inflammatory environment in the peritoneum, which was associated with increased levels of cytokines and chemokines in the peritoneum and lesion vascularization and innervation ( 20 ), indicating dysregulation of GATA6 + LPM are involved in disease pathophysiology. In the present study, we thus examined the role of peritoneal GATA6 + LPM following endometriotic lesion induction on the recruitment and differentiation of peritoneal macrophages and how they affect disease progression, inflammation, and endometriosis-associated hyperalgesia using Lyz2 cre Gata6 f/f mice, selectively deficient for Gata6 only in the macrophage lineage to reduce GATA6 + LPM ( 38 – 40 ).

Before studying the impact of ELL induction on macrophage populations, we sought to establish the baseline parameters for immune cell distribution, cytokine levels, and pain sensitivity in mice lacking GATA6 + macrophages. In agreement with reported values in three significant papers in 2014 characterizing GATA6 + LPM ( 38 – 40 ), we observed an approximately 50% reduction of CD11b + cell populations (macrophages) in the peritoneum in Mac Gata6 KO mice ( Figures 1A, B ). The percentages of CD11b + cell populations in Gata6 f/f mice (36.5 ± 3.2%) and Mac Gata6 KO mice (19.7 ± 4.0 %) were shown in Figure 1B ( P < 0.05). GATA6 + CD11b + cells ( Figures 1A, B ; Gata6 f/f , 91.2 ± 2.0 % vs Mac Gata6 KO, 6.7 ± 1.5 %, P < 0.01 in Figures 1B ), or a selective population of F4/80 hi MHCII lo LPM ( Supplementary Figure S1A ; Gata6 f/f , 87.9 ± 0.8 % vs Mac Gata6 KO, 3.8 ± 1.7 %, P < 0.05) were greatly reduced in the peritoneum, confirming that the Lyz2 Cre successfully deleted Gata6 , leading to the loss of GATA6 + LPM expressing high F4/80 in Mac Gata6 KO mice ( 38 , 40 ). In addition, GATA6-regulated genes were also down-regulated in the peritoneal exudate cells of mice lacking GATA6, confirming the loss of GATA6-driven transcriptional factor action ( Supplementary Figure S1B ). However, GATA6 deficiency did not affect CD3 + and CD19 + cells in the peritoneum ( Figures 1A, B ). When MDR occurs by sterile stimuli, inflammation leads to the recruitment of BM-derived monocytes to replace partially long-lived LPM ( 32 , 34 , 57 , 58 ), although remaining LPM proliferate substantially ( 59 ). In the persistent loss of GATA6 + F4/80 hi LPM, an increase in F4/80 lo MHCII hi macrophages has been reported ( 39 ). We have also observed an increased F4/80 lo MHCII hi ( Supplementary Figure S1A ; Gata6 f/f , 7.8 ± 1.0 % vs Mac Gata6 KO, 38.8 ± 2.0 %, P < 0.05) and Ly6C + cells in Mac Gata6 KO mice ( Figure 1C ; Gata6 f/f , 4.6 ± 0.5 % vs Mac Gata6 KO, 9.2 ± 0.9 %, P < 0.01). Furthermore, F4/80 int cells had high levels of MHCII in Mac Gata6 KO mice (F4/80 int MHCII int , 54.0 ± 3.4 %), which was lowly expressed in GATA6 f/f mice (3.7 ± 0.8 %) ( Supplementary Figure S1A ; P < 0.05), indicating that GATA6 deficiency induces persistent replenishment from BM in the peritoneum. TIM4 + residential macrophages were reduced due to the loss of GATA6 + LPM ( Figure 1C ; Gata6 f/f , 80.3 ± 1.7 % vs Mac Gata6 KO, 51.5 ± 3.1 %, P < 0.01). As expected, F4/80 + cells in Mac Gata6 KO mice were extremely low in either TIM4 + macrophage populations ( Figure 1C ; Gata6 f/f , 97.2 ± 0.6 % vs Mac Gata6 KO, 8.2 ± 2.9 %, P < 0.01), or TIM4- macrophage populations ( Figure 1C ; Gata6 f/f , 10.6 ± 3.0 % vs Mac Gata6 KO, 0.2 ± 0.1 %, P < 0.01). In support of persistent replacement from BM-derived recruited inflammatory macrophages, a high level of MHCII in TIM4 + macrophages was observed in Mac Gata6 KO mice ( Figure 1C ; Gata6 f/f , 3.3 ± 0.6 % vs Mac Gata6 KO, 25.5 ± 9.4 %, P < 0.01). While a higher basal proliferation rate (SG 2 M phases) of peritoneal macrophages was reported in Mac Gata6 KO mice ( 39 ), Ki67 + cells in either TIM4 + or – macrophages were not altered by loss of GATA6 + LPM in the steady state ( Figure 1C ). Peritoneal cytokine levels, such as TNFα, IL-1β, and IL-6 ( Figure 1D ), as well as abdominal and paw hypersensitivity, were not significantly different between Gata6 f/f and Mac Gata6 KO mice ( Figure 1E ). We next induced ELL in Gata6 f/f and Mac Gata6 KO mice to examine how the loss of GATA6 + LPM affects BM-derived and peritoneal residential macrophages and their resolutions from the stimuli of lesion induction ( Figures 2 , 3 ). While the CD11b + population was reduced in Mac Gata6 KO mice prior to ELL induction, within 42 days after ELL induction, there were no significant differences in the CD11b + population between Gata6 f/f and Mac Gata6 KO mice after ELL induction ( Figure 2B ). As shown in Figures 2C, D , Supplementary Figure S2A , ELL induction in GATA6 f/f mice induced the disappearance of TIM4 hi MHCII lo residential macrophages 3 days after ELL induction (day -1, 81.3 ± 2.2 % vs. day 3, 26.2 ± 9.1 %, P < 0.0001 in Supplementary Figure S2A ). Concurrently, the influx of increased Ly6C + monocytes was observed 3 days after ELL induction ( Figure 2E ; day -1, 3.8 ± 0.5 % vs. day 3, 36.1 ± 3.6 %, P < 0.0001). At the same time, TIM4 lo MHCII hi macrophages, as shown in Figures 2C, D and Supplementary Figure S2A , were significantly elevated (day -1, 11.8 ± 1.4 % vs. day 3, 25.2 ± 2.5 %, P < 0.001 in Supplementary Figure S2A ) following an increase of Ly6C + and TIM4 lo MHCII lo cells as the recruitment of inflammatory macrophages ( Figures 2C–E , Supplementary Figure S2A ) ( 15 , 21 , 60 ). By day 7, MHCII hi inflammatory macrophages exhibited TIM4 int MHCII hi phenotypes consistent with inflammatory macrophages. From days 21-42, TIM4 hi MHCII lo residential macrophages recovered and repopulated to the baseline distribution ( Figures 2C, D , Supplementary Figure S2A ). In contrast, a more significant disappearance of TIM4 hi MHCII hi residential macrophages occurred in Mac Gata6 KO mice within 3 days ( Figures 2F, G, I , Supplementary Figure S2B ). The percentages of TIM4 hi MHCII hi macrophages were 49.5 ± 2.5 % at day -1 and 2.3 ± 0.9 % at day 3 as shown in Supplementary Figure S2B ( P < 0.0001). Enhanced recruitment of Ly6C + monocytes ( Figures 2H ; day -1, 9.2 ± 0.9 % vs. day 3, 32.6 ± 8.8 %, P < 0.0001), or TIM4 lo MHCII hi inflammatory macrophages ( Figures 2F, G ; day -1, 28.5 ± 1.2 % vs. day 3, 56.5 ± 3.2 %, P < 0.01 shown in Supplementary Figure S2B ) was also observed by day 3. Interestingly, the recruitment of MHCII hi inflammatory macrophages was extensive in Mac Gata6 KO mice ( Figures 2F, G, I ), probably due to the severe disappearance of TIM4 hi MHCII hi residential macrophages. While repopulation of TIM4 hi MHCII hi macrophages continuously occurred in Mac Gata6 KO, the replenishment of TIM4 hi MHCII hi residential macrophages was slower than in Gata6 f/f mice ( Figures 2F, G, I , Supplementary Figure S2B ). As shown in Supplementary Figure S2AB , TIM4 hi matured residential macrophages recovered to the baseline by day 21 in Gata6 f/f mice ( Supplementary Figure S2A ; day -1, 81.3 ± 2.2 % vs. day 21, 61.3 ± 4.0 %, N.S.), while replenishment of TIM4 hi macrophages was still ongoing at day 42 in Mac Gata6 KO mice ( Supplementary Figure S2B ; day -1, 49.5 ± 2.5 % vs. day 21, 27.8 ± 2.6 %, P < 0.001). Indeed, TIM4 int MHCII hi macrophages (unmatured) were elevated at only day 7 in Gata6 f/f mice ( Figures 2C, D , Supplementary Figure S2A ), whereas TIM4 int MHCII hi macrophages in Mac Gata6 KO mice disappeared within 3 days and then repopulated but remained persistent at day 42 during the resolution phase of resident macrophages ( Figures 2F, G, I , Supplementary Figure S2B ). Thus, GATA6 deficiency in LPM limits the ability of TIM4 hi residential macrophage repopulation following MDR within 6 weeks due to probably loss of proliferative ability with transcriptional signatures ( 38 , 39 ) and persistently recruits MHCII hi inflammatory macrophages but fails to acquire a mature resident phenotype. Subsequently, Ki67 expression confirmed GATA6-dependent proliferative ability, showing different proliferative phenotypes of TIM4 + residential macrophages ( Figure 3 ). After the initial disappearance of TIM4 + residential macrophages by ELL induction in Gata6 f/f mice, the remaining TIM4 + residential macrophages actively proliferated by self-renewal within day 3 ( Figures 3A, B ). As shown in Figure 3B , the percentages of TIM4 + Ki67 + macrophages were significantly increased on day 3 (10.15 ± 4.0 %) compared to day -1 (2.4 ± 0.6 %). This active proliferation probably contributed to the recovery of the TIM4 + residential macrophage population by day 21 ( Supplementary Figure S2A ). In Mac Gata6 KO mice, TIM4 + residential macrophages almost entirely disappeared and lost their proliferative ability, although they gradually gained proliferative activity to the baseline ( Figures 3A, B ). In contrast, TIM4 - macrophages possess a higher proliferative state in Mac Gata6 KO mice, possibly due to the continuous recruitment of inflammatory macrophages. Analysis of TIM4 + Ki67 + and TIM4 - Ki67 + cells in CD11b + subsets after ELL induction in Gata6 f/f and Mac Gata6 KO mice. (A) Representative flow cytometric plots of CD11b + cells gated with TIM4 and Ki67. (B) Bar graphs showing the TIM4 + Ki67 + (top panel) and TIM4 - Ki67 + (bottom panel) cell percentages in Gata6 f/f in blue, or Mac Gata6 KO in red, or together. Data are presented as the mean ± SEM (n=5). One-way ANOVA, followed by Dunnett’s multiple comparison test, was used to analyze the time-dependent differences in each group. Two-way ANOVA was performed to determine the main effects of ELL induction (depicted as ELL), GATA6 status (depicted as GATA6), and the interaction effect between the two factors (depicted as Int.). The differences between Gata6 f/f and Mac Gata6 KO mice at each time point were compared by the Multiple Mann-Whitney test. * P < 0.05, *** P < 0.001. While abundant cytokines and chemokines have been observed in the pelvic cavity of endometriosis patients ( 13 , 16 ), TNFα, IL-1β, and IL-6 are considered the key factors involved in maintaining the aberrant peritoneal inflammatory environment, promoting lesion growth, and mediating peripheral sensitization ( 61 – 63 ). Elevated TNFα, IL-1β, and IL-6 levels have been reported in the blood, peritoneal fluids, and/or eutopic and ectopic endometrial tissues of women with endometriosis ( 13 , 64 – 66 ). Specifically, the levels of TNFα, IL-1β, and IL-6 are increased in peritoneal macrophages isolated from endometriosis patients ( 67 ). Thus, we next analyzed the secretion of proinflammatory cytokines, TNFα, IL-1β, and IL-6, in the peritoneal fluids to assess how dysregulated peritoneal macrophage resolutions in the absence of GATA6 affect cytokine secretions ( Figure 4 ). ELL induction time-dependently affected cytokine secretions ( Figure 4A ). Peritoneal TNFα, IL-1β, and IL-6 were elevated immediately after ELL induction on day 3 and remained high on day 7 but returned to the baseline by day 21 in both Gata6 f/f and Mac Gata6 KO mice ( Figure 4A ). When cytokine levels were compared between Gata6 f/f and Mac Gata6 KO mice at each time point, TNFα at day 21 in Gata6 f/f mice was higher than in Mac Gata6 KO mice ( Figure 4B ; Gata6 f/f , 412.1 ± 60.7 pg/ml vs. Mac Gata6 KO, 186.1 ± 8.9 pg/ml, P < 0.05). Although MHCII hi macrophages were continuously recruited 3-6 weeks after ELL induction in Mac Gata6 KO mice ( Figures 2F, G, I ), the persistent recruitment of inflammatory macrophages did not accelerate the cytokine levels in the peritoneum. Proinflammatory cytokine levels (TNFα, IL-1β, and IL-6) after ELL induction in Gata6 f/f and Mac Gata6 KO mice. Cytokine levels were shown in a single group (A) or grouped plots (B) . Data are presented as the mean ± SEM (n=5). (A) Data were analyzed with one-way ANOVA followed by Dunnett’s multiple comparisons test. ** P < 0.01, *** P < 0.001, and **** P < 0.0001. (B) Two-way ANOVA was performed to analyze the main effects of ELL induction (depicted as ELL), GATA6 status (depicted as GATA6), and the interaction effect between the two factors (depicted as Int.). The difference between Gata6 f/f and Mac Gata6 KO mice at each time point was determined by the Multiple Mann-Whitney test. * P < 0.05. Inflammatory factors in the peritoneum from endometriosis patients can be inflammatory mediators that stimulate neurite outgrowth and endometriosis-associated hyperalgesia ( 20 , 26 , 68 ). We next performed the von Frey test to examine the abdominal and hind paw retraction threshold to determine whether loss of GATA6 + LPM affects endometriosis-associated hyperalgesia ( Figure 5 ). ELL induction affected time-dependently both abdominal and hind paw sensitivities ( Figure 5A ). Gata6 f/f and Mac Gata6 KO mice withdrew abdominal retraction thresholds with significantly lighter stimuli during the entire study period ( Figure 5A ), even though cytokines levels were only elevated by day 7 ( Figure 4 ). Gata6 f/f mice tended to show higher sensitivity after day 21 when the abdominal retraction was compared between Gata6 f/f and Mac Gata6 KO mice at each time point ( Figure 5B ; Gata6 f/f , 0.04 ± 0.01 g vs. Mac Gata6 KO, 0.09 ± 0.02 g at day 21, P = 0.18; Gata6 f/f , 0.04 ± 0.01 g vs. Mac Gata6 KO, 0.08 ± 0.01 g at day 42, P = 0.08). The hind paw retraction thresholds were sensitive in Gata6 f/f mice until day 21, but after day 7 in Mac Gata6 KO mice, sensitivity returned to the preinduction threshold ( Figure 5A ), whereas no different sensitivity was observed between the groups ( Figure 5B ). The results suggested that Gata6 f/f mice tended to sustain higher endometriosis-associated hyperalgesia for extended periods. Evaluation of endometriosis-associated hyperalgesia followed by ELL induction in Gata6 f/f and Mac Gata6 KO mice. Abdominal and hind paw withdrawal thresholds were assessed by the von Frey test in Gata6 f/f and Mac Gata6 KO mice. Data are shown by individual group (A) and grouped graphs (B) as mean ± SEM (n = 10). (A) Date were analyzed with one-way ANOVA followed by Dunnett’s multiple comparison test. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. (B) Two-way ANOVA was performed to analyze the main effects of ELL induction (depicted as ELL), GATA6 status (depicted as GATA6), and the interaction effect between the two factors (depicted as Int.). The difference between Gata6 f/f and Mac Gata6 KO mice at each time point was determined by the Multiple Mann-Whitney test. Endometriosis-associated hyperalgesia, especially local abdominal sensitivity, should be correlated with the stimuli in the peripheral nerve terminals of nociceptor neurons ( 69 ), such as an increase in the expression of transient receptor potential channels, TRPV1. Peripheral nerves can be activated with the increased release of neurotransmitters, such as SP and CGRP. We thus examined TRPV1, SP, and CGRP in the L4-6 DRG, the primary spinal ganglia receiving sensory input from pelvic organs ( Figure 6 ). ELL induction increased TRPV1, SP, and CGRP expression time-dependently ( Figures 6A, B ). More TRPV1 + DRG were observed until day 21 in Gata6 f/f mice and until day 7 in Mac Gata6 KO mice ( Figures 6A, B ). Higher TRPV1 + neurons were detected in Gata6 f/f mice at day 21 compared with those in Mac Gata6 KO mice ( Figure 6B ; Gata6 f/f , 14.1 ± 1.2 % vs. Mac Gata6 KO, 7.1 ± 1.3 % at day 21, P = 0.039). SP + DRG were significantly high until day 21 compared with that in the preinduction level in both mouse groups ( Figures 6A, B ). CGRP + DRG were elevated until day 42 in Gata6 f/f mice and until day 21 in Mac Gata6 KO mice ( Figures 6A, B ). Thus, the results support behavior analysis, in which Gata6 f/f mice potentially sustained higher endometriosis-associated hyperalgesia than Mac Gata6 KO mice at later periods. Expression of TRPV1, SP, and CGRP in DRG from Gata6 f/f and Mac Gata6 KO mice after ELL induction. (A) Representative images showing DRG sections double stained with TRPV1, SP, or CGRP (green), and PGP9.5 (red), as a marker of DRG cell body. (B) Quantification of TRPV1 + , SP + , or CGRP + cells in PGP9.5 + cells. Data are shown as the mean ± SEM (n=5). Two-way ANOVA was performed to analyze the main effects of ELL induction (depicted as ELL), GATA6 status (depicted as GATA6), and the interaction effect between the two factors (depicted as Int.). Following two-way ANOVA, Dunnett’s multiple comparison test was used to analyze time-dependent differences compared with day -1 within a group labeled blue ( Gata6 f/f ) or red (Mac Gata6 KO). The difference between Gata6 f/f and Mac Gata6 KO mice at each time point was determined by the Multiple Mann-Whitney test. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Lastly, we assessed how GATA6 deficiency affects the progression and growth of endometriosis lesions. Loss of GATA6 + LPM did not alter lesion number during the study period ( Figure 7A ). Because we previously reported that peritoneal macrophages can infiltrate ELL ( 20 ), we addressed macrophage infiltration in the lesion staining with CD68, a macrophage marker. We only examined macrophage infiltration in the lesion on days 21 and 42, as lesion number, endometriosis-associate hyperalgesia, cytokine levels in the peritoneal cavity, and TRPV1, SP, and CGRP expression in the DRG were not altered by GATA6 deficiency until day 7 in Gata6 f/f and Mac Gata6 KO mice. CD68 + macrophages were comparable in both mouse groups at days 21 and 42 ( Figure 7B ). Thus, the absence of GATA6 in the LPM did not affect lesion progression and macrophage infiltration. Analysis of ELL in Gata6 f/f and Mac Gata6 KO mice. (A) Quantification of lesion numbers in Gata6 f/f and Mac Gata6 KO mice after ELL induction (n=10). (B) Representative immunohistochemical images (left) and quantification (right) of CD68 + macrophages in the lesions of Gata6 f/f and Mac Gata6 KO mice on days 21 and 42 (n=5). Data are shown as the mean ± SEM. Two-way ANOVA was performed to analyze the main effects of ELL induction (depicted as ELL), GATA6 status (depicted as GATA6), and the interaction effect between the two factors (depicted as Int.). The difference between Gata6 f/f and Mac Gata6 KO mice at each time point was determined by the Multiple Mann-Whitney test. No significant difference was detected.

Peritoneal macrophages are known to contribute to endometriosis pathophysiology, as dysregulated macrophages increase the secretion of proinflammatory cytokines and reduce phagocytic functions in women with endometriosis ( 23 , 25 , 70 , 71 ). Bacci et al. ( 24 ) demonstrate that peritoneal macrophages are required for lesion establishment, growth, and vascularization in a mouse model of endometriosis. However, peritoneal macrophages are highly diverse ( 21 , 72 ), differ in their ontogeny ( 31 ), and have transcriptionally and functionally divergent features depending on the signals of the local environment ( 30 ). Hogg et al. ( 15 ) have reported different characteristics of macrophages depending on their ontogeny in the pathogenesis of endometriosis. They found that monocytopenic mice, which limit monocyte recruitment and subsequently reduce LPM and SPM replenishment, develop increased lesions ( 15 ). Furthermore, fewer lesions are developed in mice when embryo-derived LPM are replaced by monocyte-derived LPM, indicating that monocyte-derived LPM protect against endometriosis development ( 15 ). In the present study, we focused on the roles of embryo-derived GATA6 + LPM, which represent the majority of peritoneal residential macrophages, during ELL establishment and progression. When menses-like tissues are introduced in the peritoneum, it causes acute inflammatory responses. GATA6 + LPM are important for the initial uptake where they adhere to the mesothelium to cover organs ( 41 , 42 ) or die via pyroptosis to release proinflammatory cytokines, such as IL-1β ( 73 ), called MDR. If LPM die/disappear, they appear to be replaced by BM-derived macrophages ( 34 ). However, when LPM adhere to the walls or mesothelium, they eventually detach and return to the residential LPM in the peritoneum ( 34 ). In the present study of Gata6 f/f mice, peritoneal residential LPM (TIM4 + ) disappeared after ELL induction, BM-derived monocytes (Ly6C + ) were recruited/increased, and MHCII hi macrophages were elevated in the peritoneum. This likely happens immediately after ELL induction, as sterile inflammation induced by intraperitoneal injection with a low dose of zymosan (10 µg) causes MDR and recruited monocytes and inflammatory macrophages in the peritoneum within 4 hours ( 37 ). MDR from ELL induction was still observed on day 3 in our study. The significant recruitment of Ly6C + monocytes and MHCII hi macrophages was present until day 21, indicating that LPM pyroptosis with replenishment from BM-derived macrophages also occurs. Supporting the MDR via pyroptosis ( 34 ), proinflammatory cytokines, such as TNFα, IL-1β, and IL-6, were significantly elevated in the first week after ELL induction. In contrast, because resident macrophages did not wholly disappear, the remaining macrophages rebounded in number by day 21. In support of this, proliferation in residential LPM was active in the first 3 days and tended to be high even on day 7. Based on the results of increased Ly6C + cells, cytokine levels, and proliferation, “acute inflammatory reactions” continue until 7 days after ELL induction, which is longer than that is induced by a low dose of zymosan ( 37 ). During days 21-42, repopulated TIM4 hi residential macrophages and BM-derived TIM4 lo MHCII hi and TIM4 int MHCII hi macrophages returned to the baseline level. LPM in the peritoneum appear to be reestablished by replenishment from BM-derived macrophages and proliferation of remaining residential LPM. Thus, Days 21-42 are likely to be a resolution phase from the inflammatory stimuli (ELL induction). However, we are uncertain whether LPM are fully functional, as monocyte-derived LPM are functionally distinct from embryo-derived residential LPM ( 37 ). Deficiency of GATA6 in the peritoneum induces loss of F4/80 hi MHCII lo macrophages (LPM), as previously reported ( 38 , 40 ). We also observed a significantly reduced macrophage population of highly expressed F4/80. Loss of GATA6 reduced TIM4 + resident macrophages as well. Due to the persistent reduction of LPM or resident macrophages, consistent recruitment from BM-derived monocytes and macrophages was seen in our study, as Ly6C + monocytes and MHCII hi macrophages were elevated. In contrast, proinflammatory cytokines were not elevated in the peritoneum of Mac Gata6 KO mice, suggesting that genetically engineered deletion of GATA6 is different from MDR via pyroptosis to release proinflammatory cytokines. When ELL was induced in Mac Gata6 KO mice, severe MDR occurred, as TIM4 + resident macrophages almost completely disappeared in addition to the loss of F4/80 hi macrophages on day 3. At the same time, MHCII hi macrophages were continuously recruited and maintained at high levels, even on day 42. GATA6 deficiency disrupts peritoneal macrophage proliferation, as GATA6 controls peritoneal macrophage proliferation via its target genes ( 38 , 39 ). After ELL induction, the remaining residential macrophages are likely less proliferative due to the severe disappearance of TIM4 + residential macrophages in Mac Gata6 KO mice on day 3. Instead, recruited TIM4 lo MHCII hi macrophages presented higher proliferative activity. However, TIM4 + resident macrophages did not fully recover by day 42, indicating that BM-derived TIM4 + macrophages probably have a less proliferative ability due to GATA6 deficiency; they are probably distinct from embryo-derived LPM. Severe MDR also elevated proinflammatory cytokine levels in the peritoneum of Mac Gata6 KO mice, although we only evaluated TNFα, IL-1β, and IL-6, probably via pyroptosis, similar to those in Gata6 f/f mice. Interestingly, TNFα was higher on day 21 in Gata6 f/f mice than in Mac Gata6 KO mice. GATA6 + macrophages use an inhibitory pathway to suppress IL-1β processing, a potent activator of the inflammatory cascade in response to inflammatory stimuli, to solve early stages of inflammation ( 74 ). It is unclear whether and how GATA6 + macrophages are still involved in cytokine secretion after the acute phase of inflammation, and it remains to be examined. However, newly recruited monocyte-derived LPM can protect against lesion development ( 15 ). Since GATA6 deficiency enhanced the recruitment of BM-derived macrophages to replenish resident macrophages, they might be involved in the reduction of cytokine levels in the peritoneal cavity. Alternatively, the persistent recruitment of BM-derived macrophages is not directly correlated with cytokine secretion in the peritoneal cavity. On the other hand, GATA6 + macrophages can become tumor-associated macrophages to contribute to tumor progression ( 41 , 42 ), as deletion of peritoneal macrophages reduces tumor progression and peritoneal metastasis in ovarian cancer ( 46 , 47 ). Peritoneal macrophages expressing higher amounts of a proangiogenic factor, VEGF, are associated with peritoneal metastasis in gastric cancer ( 50 , 75 ). GATA6 + LPM can invade liver malignant tumors and contribute to metastatic tumor growth and recurrence ( 51 ). In the mouse model of endometriosis, >25% of macrophages in the lesions are LPM in the peritoneum, indicating that significant trafficking of LPM into the endometriotic lesion might be favorable to maintaining chronic inflammation and disease progression ( 15 ). We showed that the GATA6-deficient peritoneal macrophages did not affect abdominal and paw hyperalgesia in a steady state. When ELL was induced, both Gata6 f/f and Mac Gata6 KO mice immediately exhibited increased endometriosis-associated hypersensitivity in the abdomen on day 3 and sustained until day 42. Increased expression of pain-related mediators supports the result of endometriosis-associated hyperalgesia that was sustained for nearly 6 weeks. However, the sensitivity in the abdomen tended to be severe in Gata6 f/f mice on days 21-42, when most peritoneal macrophage replenishment was resolved. In support of local abdominal sensitivity in Gata6 f/f mice, TRPV1 expression was higher on day 21. Systemic paw sensitivity was significantly high in Gata6 f/f mice until day 21, whereas it was no longer significant on day 21 in Mac Gata6 KO mice. The results suggest that ELL induce acute inflammation in the peritoneum, immediately sensitizing the pain-related neurons. While the resolution of the immune cell population from the stimuli of ELL induction is relatively fast, endometriosis-associated hyperalgesia is sustained longer because the stimuli likely further sensitize pain-related neurons. Mac Gata6 KO mice showed slightly less hyperalgesia, which can be supported by TNFα level in the peritoneum and TRPV1 expression in the DRG. Hogg et al. ( 15 ) demonstrate that monocyte-derived peritoneal macrophages are anti-endometriosis to protect the cavity from lesion development. Mac Gata6 KO mice constantly recruited SPM (monocyte-derived macrophages) due to the deficiency of GATA6 and further enhanced the replenishment of residential macrophages after ELL induction. Thus, large numbers of BM-origin monocyte-derived macrophages exist in the peritoneum. Indeed, TIM4 int macrophages differentiated from monocytes are significantly higher in Mac Gata6 KO mice on days 21-42. Therefore, although its mechanism is unclear, monocyte-derived macrophages might also protect against inflammatory stimuli during the resolution phase of inflammation. However, increased TIM4 int macrophages were observed in both Gata6 f/f and Mac Gata6 KO mice, suggesting that GATA6 + macrophages do not affect earlier lesion development, including establishment. As a supporting result, the lesion numbers were not altered at any stage. Gibson et al. ( 9 ) have reported a human study about endometriosis and peritoneal macrophages. The group shows that CD14 hi peritoneal macrophages considered resident peritoneal macrophages in humans ( 52 ) are significantly higher than CD14 lo peritoneal macrophages in women with endometriosis ( 9 ). CD14 hi peritoneal macrophages negatively correlate with pain score, whereas CD14 lo peritoneal macrophages are positively correlated, independent of endometriosis diagnostic outcome at laparoscopy ( 9 ). They assume CD14 hi and CD14 lo in humans to be equivalent to LPM and SPM in mice ( 76 ). However, the authors have noted that CD14 hi macrophages likely account for BM-derived macrophages, as monocyte-derived LPM (CD14 hi macrophages) are not distinguished. The group concludes that pain symptoms do not correlate with disease extent, as pain scores do not differ among patients at the rASRM stages, indicating inflammatory or immune cell profiles might be a better predictor of endometriosis pain than lesion extent identified during laparoscopy ( 9 ). Evolutionarily, retained body cavities protect internal organs from the external toxic environment, and cavity macrophages contribute to various immune responses ( 42 ). LPM maintain peritoneal cavity homeostasis, providing the first line of defense mechanisms ( 41 , 42 ). Rigorous studies present the vital roles of LPM in repairing injuries and controlling infections by GATA6-driven transcriptional program ( 41 , 42 ). In the present study, we used a mouse model of endometriosis to study the role of GATA6 resident macrophages following the resolution of inflammatory stimuli (lesion induction). The results suggest the involvement of peritoneal GATA6 + macrophages in the peritoneal inflammatory environment, including the reprogramming of monocyte-derived macrophages. Because GATA6 deficiency induced the extensive recruitment of monocyte-derived macrophages and enhanced the replenishment of residential macrophages after ELL induction, they might protect against inflammatory stimuli, especially during the resolution phase. Therefore, GATA6 + LPM might be favorable toward sustaining local inflammation in the peritoneum and sensitivities in the neurons, reflecting potentially endometriosis-associated pain during the process of resolution of inflammatory stimuli. Retrograde menstruation causes massive inflammatory responses in the peritoneum. However, menstrual cycles repeatedly occur in women; each retrograde menstruation induces composite inflammation in the pelvic cavity, and unsolved inflammation is expected to worsen to develop chronic conditions further. It might be necessary to induce multiple inflammatory stimuli (multiple ELL induction) to understand the role of residential macrophages and their resolution process in the peritoneum associated with endometriosis-associated hyperalgesia.

Gata6 -floxed mice ( Gata6 f/f , Gata6 tm2.1Sad/ , JAX #008196) and Lysozyme M (Lyz2) Cre mice ( Lyz2 Cr e , Lyz2 tm1(cre)Ifo/J , JAX #004781) were purchased from the Jackson Laboratory and Lyz2 Cre/+ Gata6 f/f (denoted as Mac Gata6 KO) and Lyz2 +/+ Gata6 f/f (denoted as Gata6 f/f ) mice were generated. All animal experiments were performed at Washington State University according to the NIH guidelines for the care and use of laboratory animals (protocol #6751). Characterizations of peritoneal macrophages in Gata6 f/f and Mac Gata6 KO mice are shown in Figure 1 . Comparison of peritoneal immune cell profiles between Gata6 f/f and Mac Gata6 KO mice. (A) Representative flow plots illustrating the composition of CD11b + , GATA6 + , CD3 + , and CD19 + cells in Gata6 f/f and Mac Gata6 KO mice. (B) Proportions of CD11b + , GATA6 + , CD3 + , and CD19 + in Gata6 f/f and Mac Gata6 KO mice. (C) Flow cytometric analysis of CD11b + subsets. Bottom left: Proportions of Ly6C + and TIM4 + cells from CD11b + subsets. Right panel: TIM4 + (top) and TIM4 - (bottom) cells were further gated with F4/80, MHCII, and Ki67. Results were shown as representative overlaid histogram plots ( Gata6 f/f in blue and Mac Gata6 KO in red) with summarized bar plots below. (D) Proinflammatory cytokine levels (TNFα, IL-1β, and IL-6) in the peritoneal fluid were analyzed by IQELISA. (E) Abdominal and hind paw hyperalgesia were analyzed using the von Frey test. Data were analyzed by the Mann-Whitney test and shown mean ± SEM (n=5). * P < 0.05, ** P < 0.01. An experimental mouse model of endometriosis was employed by adopting procedures we have used and described previously with minor modifications ( 20 , 21 , 53 ). Briefly, a ‘menses-like’ event was induced in ovariectomized estradiol-17β- and progesterone-primed donor mice following an established protocol ( 54 ). Then, mouse menses-like endometrium scraped from myometrium and cut into fragments (1-2 mm per side) were introduced as the source of syngeneic mouse endometrium (donor: Gata6 f/f mice) via injection (in 0.2 mL PBS) into the peritoneal cavity in untreated naïve mice (recipient: Gata6 f/f or Mac Gata6 KO mice) under anesthesia via inhaled isoflurane. Mac Gata6 KO mice served only as recipient mice, as we have already confirmed that peritoneal macrophages do not arise from donor tissue in this model ( 20 ). Endometriosis-like lesions (ELL) were induced in the recipient, Gata6 f/f or Mac Gata6 KO mice, as described above ( Figure 2A ). On Day -1, 3, 7, 21, and 42 (Day 0 = ELL induction), a behavioral test was performed, and mice were then euthanized for sample collections: peritoneal fluid (PF) was recovered by lavage (4 mL x 2 of ice-cold PBS with 3% FBS), ELL, and bilateral lumbar (L4-6) dorsal root ganglia (DRG) were collected for further analysis. Characterization of peritoneal macrophage profiling following lesion induction in Gata6 f/f and Mac Gata6 KO mice. (A) Experimental study design as described in Materials and Methods. (B) Graph summarizing flow cytometry data for the proportion of CD11b + cells on Days 3, 7, 21, and 42 after endometriosis-like lesions (ELL) induction (n=5). Two-way ANOVA was performed to determine the main effects of ELL induction (depicted as ELL), GATA6 status (depicted as GATA6), and the interaction effect between the two factors (depicted as Int.). The differences between Gata6 f/f and Mac Gata6 KO mice at each time point were compared by the Multiple Mann-Whitney test. (C, F) Representative flow cytometric plots showing the dynamics of macrophages following ELL induction in Gata6 f/f (C) and Mac Gata6 KO mice (F) . CD11b + cells were further gated by TIM4 and MHCII (top) or Ly6C (bottom) (n=5). Statistical analysis of TIM4 high(hi) MHCII low(lo) (pink), TIM4 inter(int) MHCII hi (yellow), and TIM4 lo MHCII hi (green) cells in Gata6 f/f mice (D) , or TIM4 hi MHCII hi (pink), TIM4 int MHCII hi (yellow), and TIM4 lo MHCII hi (green) cells in Mac Gata6 KO mice (G) was performed with two-way ANOVA to determine the time-dependent difference after ELL induction (depicted as ELL), the mean difference among cell populations (depicted as Group), and the interaction effect between the two factors (depicted as Int.). Time course analysis for Ly6C + cells in (E, H) was conducted by one-way ANOVA with Dunnett’s multiple comparisons test. * P < 0.05, and **** P < 0.0001. (I) Time-dependent proportions of TIM4 hi (pink), TIM4 int (yellow), TIM4 lo (green) cells, and (J) Ly6C + cells in Gata6 f/f and Mac Gata6 KO mice (n=5). Data were analyzed by two-way ANOVA and Multiple Mann-Whitney tests as described in (B) . * P < 0.05. Data are presented as the mean ± SEM. Single-cell suspensions of peritoneal exudate cells were used for analyzing immune cell profiles by flow cytometry as described previously ( 20 , 21 ). Briefly, peritoneal exudate cells were lysed using Red Blood Cell Lysis Buffer (BioLegend) and incubated at room temperature for 20 min with Zombie Aqua™ Fixable Viability dye (Bio-Legend). The cells were blocked on ice for 20 min with Fc Block anti-CD16/CD32 (ThermoFisher) and stained with fluorochrome-conjugated monoclonal antibodies for 1 hour ( Supplementary Table S1 ). For GATA6 and Ki67 staining, the cells were fixed and permeabilized with Foxp3/Transcription Factor Staining Buffer Set and Permeabilization Buffer (ThermoFisher). Samples (n=5/group) were acquired with the Attune NxT Acoustic Focusing Cytometer using Attune NxT software (ThermoFisher), and data were analyzed with FlowJo v10.4 software (FLOWJO). A standard behavioral (mechanical sensitivity) test was performed before sample collection, as we have performed and others described previously ( 26 , 55 , 56 ). Mice were allowed to acclimate in the testing room for 30 min, and then the von Frey test was performed using von Frey filaments (BIO-VF-M, Bioseb). Filaments were applied 10 times to the skin perpendicular to the lower abdomen and bilateral hind paws. The force in grams (g) of the filament evoking a withdrawal response (50% response count as sensitive) was recorded. Three behaviors were considered positive responses to filament stimulation: 1) sharp retraction of the abdomen, 2) immediate licking and/or scratching of the area of filament stimulation, or 3) jumping. All behavioral tests were performed blindly without describing the identity and details of treatment groups to investigators assessing pain. These data were then analyzed by another independent investigator. Protein yield from peritoneal fluid was quantitated by BCA assay (Pierce), and TNFα (IQM-TNFA-1), IL-1β (IQM-IL1b-1), and IL-6 (IQM-IL6-1) were further quantified by IQELISA kits (Ray Biotech) according to the manufacturer’s instructions (n=5/group). Immunostaining of TRPV1, SP, CGRP, PGP 9.5, and CD68 was performed with cross-sections (5 µm) of paraffin-embedded tissues using specific primary antibodies ( Supplementary Table S1 ) and AlexaFluor 488 or 568-conjugated F(ab’) secondary antibody (Molecular Probe) or VECTASTAIN ABC kit (Vector lab). Immunostaining images were acquired by Leica DM4 B microscopy. Cell-specific CD68 positive and total cell numbers were counted by Image J in the area of 0.07244 mm 2 , and the percentage of CD68 + cells was shown. PGP9.5 was used as a pan-neuronal marker and was co-stained with TRPV1, SP, or CGRP. TRPV1, SP, or CGRP positive DRG neurons in the section were counted by Image J in the area of 0.07244 mm 2 , and the percentages of TRPV1, SP, or CGRP positive cells per PGP9.5-positive DRG were shown (n=5/group). Total RNA was isolated from peritoneal exudate cells collected from Gata6 f/f and Mac Gata6 KO mice using TRIzol reagent (Sigma). cDNA templates were synthesized from 1 μg of purified RNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher). Relative gene expression was examined by SYBR Green Master Mix (Thermo Fisher) using a CFX Opus 96 Real-time PCR system (BioRad) as described previously ( 20 , 21 ). Rpl19 was used as the reference gene to normalize mRNA expression levels. Data were analyzed using the 2 -ΔΔCt method, and shown as the relative mRNA expression (n=5/group). Specific primer sets were designed using NCBI’s design suite and are shown in Supplementary Table S2 . Statistical analyses were performed using GraphPad Prism (version 9.5). Mann-Whitney test was used to compare the differences between Gata6 f/f and Mac Gata6 KO mice. One-way ANOVA followed by Dunnett’s multiple comparison tests was used to analyze time-dependent differences within a single group. For grouped data, two-way ANOVA was used to analyze the main effects of ELL induction (depicted as ELL on graphs) and group factor (i.e., GATA6 status, depicted as GATA6; or cell populations, depicted as Group). The interaction effect between the two factors was then calculated (depicted as Int. on graphs). Following two-way ANOVA, to determine the group differences at each time point, multiple Mann-Whitney test with Holm-Šídák correction was used. A P value less than 0.05 was considered to be statistically significant.