Assistant Professor University of Virginia, United States
Introduction: Endometriosis presents in 10% of menstruators, yet its etiology remains unknown due to its heterogeneity. While endometriotic lesions behave differently depending on their location, they maintain common characteristics including elevated inflammation, fibrogenesis, and progesterone resistance. However, interactions between pathways are unclear. Previous work shows that Deep Infiltrating Endometriosis (DIE) preferentially resides upon stiffer substrates and displays increased myofibroblastic and inflammatory markers than other subclasses of endometriosis. Within this work, we identified relationships between increased stiffnesses, myofibroblastic phenotypes, and progesterone resistance that may promote inflammatory mediators in endometriosis lesions.
Endometriosis displays systemic changes in decidualization and survival. Decidualization involves the differentiation of stromal cells into a phenotype characterized by globular morphologies, anti-inflammatory mediators, and matrix secretion. In humans, this occurs spontaneously when progesterone peaks and guides transcription mediated by progesterone receptors and FOXO1. Eventually, progesterone dissipates and causes FOXO1 to promote proapoptotic genes, inducing menstruation to foster the cyclic uterine environment. Patients with endometriosis have reduced decidualization, known as progesterone resistance, which is known to promote infertility and prevent immune clearance of lesions. Ectopic stromal cells within lesions can also survive by evading apoptosis through increased Akt signaling compared to eutopic environments and increased use of this pathway is known to disrupt FOXO1 signaling. Therefore, we sought to determine if fibrotic environments reduce FOXO1 expression and disrupt Akt signaling. We hypothesized that environmental stiffness is sufficient to induce myofibroblastic phenotypes and inhibit decidualization via Akt signaling-mediated FOXO1 disruption.
Materials and
Methods: In this work, primary uterine fibroblasts (UFs) are cultured on gels to mimic physiological stiffnesses of the uterine wall (4kPa) and stiffer substrates (100kPa). To promote cell attachment, gels are coated with 10mg/mL collagen type I. Throughout culture, cells are exposed to a combination of cAMP and medroxyprogesterone (C+P) for 3 days with no media changes. Imaging analysis examines changes in cell morphology, FOXO1, progesterone receptors (PGR), and ser473-phosphorylated protein kinase B (pAkt) expression to determine whether fibroblasts are able to decidualize. Myofibroblastic phenotypes are confirmed through the expression of alpha-smooth muscle actin (aSMA) to determine if myofibroblastic differentiation and progesterone resistance overlap. Increases in FOXO1, pAkt, and globular cytoskeleton mark decidualization, while increases in aSMA and spindle-like cytoskeletons indicate myofibroblastic differentiation. Total levels of PGR are measured since previous work debates whether progesterone resistance stems from decreased total PGR expression. To investigate impacts of extracellular matrix (ECM) proteins, a similar experiment was run on 100kPa gel coated with 10ug/mL fibronectin. Quantification of fluorescent markers was done using deep learning models trained from whole-well images, which were applied to all remaining whole-well images. Average fluorescent intensity per nuclei was collected for FOXO1, pAkt, PGR, and aSMA, while circularity was measured by examining boundary perimeter to boundary area ratios of actin cytoskeleton objects. Statistical analysis was conducted using the Shapiro-Wilk test to assess normality before using parametric 2 way ANOVA with Tukey’s multiple comparisons post hoc tests.
Results, Conclusions, and Discussions: In response to treatment on 4kPa, cell counts decrease, FOXO1 and circularity increase, little aSMA is detected, and no significance is found with PGR nor pAkt. Treatment on 100kPa demonstrates similar cell counts and increased circularity. aSMA increases between stiffnesses and after treatment upon 100kPa. PGR and FOXO1 levels are lower compared to soft counterparts with treatment, indicating low transcription of decidual genes. No significant changes in pAkt are seen.
When comparing ECM substrates, fibronectin diminished FOXO1 expression whereas collagen resulted in no change. aSMA expression changes between collagen and fibronectin coatings depending on treatment, yet in different directions. There was no significant difference in progesterone receptor expression between stiffnesses nor treatments.
We can conclude that the 4kPa UFs are decidualized due to their increase in circularity and FOXO1 with treatment. Lack of significant findings for PGR and pAkt may be due to high standard deviations. Decreased PGR and FOXO1 levels upon 100kPa suggest an inhibition mechanism involving stiffness. UFs on 100kPa displayed significant cell count drops, yet their circularity indices, while greater than those on 4kPa, did not increase. Thus, stiffness promotes resistance to decidualization. Further work is needed to confirm whether cell count changes are due to proliferative capacities or cell death. Increased aSMA between treatments with100kPa suggests a possible mechanism between hormones and fibrogenesis. ECM substrates seem to play a role in decidualization. Changes in FOXO1 and aSMA due to fibronectin suggest significant roles in hormonal reception. Again, PGR results were statistically inconclusive, but the data does suggest a positive relationship between collagen and PGR after treatment. Previous work has identified increases in aSMA due to in vitro decidualization, however this was performed on tissue culture plastic-- a stiff substrate known to overactivate myofibroblastic properties. Our findings show that such increases are less relevant upon soft substrates, and suggest that progesterone and stiffness work together to promote myofibroblastic phenotypes.
Changes due to fibronectin exposure could be due to its relationship to decidualized phenotypes. Upon decidualization, stromal cells are known to secrete fibronectin, suggesting that the significant decrease in FOXO1 after treatment could represent a negative feedback mechanism.
