Chlamydial YAP activation in host endocervical epithelial cells mediates pro-fibrotic paracrine stimulation of fibroblasts

ABSTRACT Infection of the female genital tract by Chlamydia trachomatis can produce severe fibrotic sequelae, including tubal factor infertility and ectopic pregnancy. While infection demonstrably mediates a pro-fibrotic response in host cells, it remains unclear if intrinsic properties of the upper genital tract exacerbate chlamydial fibrosis. The relatively sterile environment of the upper genital tract is primed for a pro-inflammatory response to infection, potentially enhancing fibrosis; however, subclinical C. trachomatis infections still develop fibrosis-related sequelae. Here, we compare infection-associated and steady-state gene expression of primary human cervical and vaginal epithelial cells. In the former, we observe enhanced baseline expression and infection-mediated induction of fibrosis-associated signal factors (e.g., TGFA, IL6, IL8, and IL20), implying predisposition to Chlamydia-associated pro-fibrotic signaling. Transcription factor enrichment analysis identified regulatory targets of YAP, a transcriptional cofactor induced by infection of cervical epithelial cells, but not vaginal epithelial cells. YAP target genes induced by infection include secreted fibroblast-activating signal factors; therefore, we developed an in vitro model involving coculture of infected endocervical epithelial cells with uninfected fibroblasts. Coculture enhanced fibroblast expression of type I collagen, as well as prompting reproducible (albeit statistically insignificant) induction of α-smooth muscle actin. Fibroblast collagen induction was sensitive to siRNA-mediated YAP knockdown in infected epithelial cells, implicating chlamydial YAP activation in this effect. Collectively, our results present a novel mechanism of fibrosis initiated by Chlamydia, wherein infection-mediated induction of host YAP facilitates pro-fibrotic intercellular communication. Chlamydial YAP activation in cervical epithelial cells is thus a determinant of this tissue’s susceptibility to fibrosis. IMPORTANCE Chronic or repeated infection of the female upper genital tract by C. trachomatis can lead to severe fibrotic sequelae, including tubal factor infertility and ectopic pregnancy. However, the molecular mechanisms underlying this effect are unclear. In this report, we define a transcriptional program specific to C. trachomatis infection of the upper genital tract, identifying tissue-specific induction of host YAP—a pro-fibrotic transcriptional cofactor—as a potential driver of infection-mediated fibrotic gene expression. Furthermore, we show that infected endocervical epithelial cells stimulate collagen production by fibroblasts and implicate chlamydial induction of YAP in this effect. Our results define a mechanism by which infection mediates tissue-level fibrotic pathology via paracrine signaling and identify YAP as a potential therapeutic target for the prevention of Chlamydia-associated scarring of the female genital tract.

To determine if infection of a specific cell type is associated with pro-fibrotic gene expression, we have examined the host response to C. trachomatis in primary human cervical epithelial cells (HCECs) and primary human vaginal epithelial cells (HVEs).We observe both steady-state and infection-associated induction of pro-inflammatory and pro-fibrotic signaling in the former, including increased expression of TGFA, IL6, IL8, and IL20.Transcription factor enrichment analysis of this gene set implicated a broad portfolio of known regulatory targets of YAP, a pro-fibrotic transcriptional cofactor.Infection promoted YAP activation in cervical epithelial cells, but not vaginal epithelial cells, suggesting that YAP may contribute to pro-fibrotic gene expression specific to the upper genital tract.To assess the downstream effects of pro-fibrotic/ pro-inflammatory signaling by infected cells, we have developed an in vitro model of infection wherein Chlamydia-infected, immortalized endocervical epithelial cells (End1s) are cocultured with uninfected uterine fibroblasts (KCO2s).We show that fibroblasts cocultured with infected epithelial cells exhibit significantly increased expression of collagen I, a phenotype sensitive to siRNA-mediated knockdown of YAP in host epithelial cells.Taken together, our results define a novel means by which infection promotes fibrotic gene and protein expression, dependent upon host YAP activity and intercellular communication.

Cervical epithelial cells exhibit intrinsic pro-inflammatory and pro-EMT gene expression relative to vaginal epithelial cells
To identify gene expression potentially driving fibrotic pathologies in the epithelial tissues of the upper genital tract, we compared the transcriptomes of mock and Chlamydia-infected cervical epithelial cells to that of infected vaginal epithelial cells.Primary HCECs and HVEs were either mock-infected or infected with the anogenital C. trachomatis serovar L2, then harvested for polyadenylated RNA for bulk RNA sequenc ing at 24 hours post-infection (hpi).Comparison of the Chlamydia-infected HCEC and HVE transcriptomes identified 7,354 genes differentially expressed (false discovery rate P-value [FDRP] ≤ 0.05) in infected cervical cells relative to infected vaginal cells (Fig. 1A; Table S1).To avoid the introduction of selection bias by solely examining genes functionally associated with pro-fibrotic and pro-EMT signaling, further characteriza tion of this gene set was performed using the ranked functional enrichment analysis functions of the STRING database (31,32).The differentially expressed gene set exhibited disproportionately high representation of induced genes associated with pro-inflammatory signaling, with GO-BP (gene ontology, biological process) terms associated with a type-I interferon response, leukocyte chemotaxis, and complement activation showing significant enrichment (Table 1).Intriguingly, the infected HCEC transcriptome also exhibited enrichment of GO-BP terms associated with keratinocyte and epidermal differentiation (Table 1).Recent reports have also shown Chlamydia can induce an EMT in endocervical epithelial cells, a pro-fibrotic transdifferentiation event associated with loss of cytokeratin expression (16,33,34).Accordingly, we observe significantly increased expression of EMT-regulating transcription factors such as SNAIL/SNAI1 (FC, 7.35) and TWIST1 (FC, 11.37) in infected HCECs relative to the infected HVE transcriptome.These data collectively suggest that cervical epithelial cells may be more susceptible than vaginal epithelial cells to infection-associated EMT.
The differences we report between vaginal and cervical epithelial cell gene expres sion during infection could be attributed to HVEs being less permissive to chlamydial infection; should HVEs exhibit a reduction in bacterial burden at 24 hpi relative to HCECs, this discrepancy would account for the diminished host response observed in the former.To assess this possibility, we examined chlamydial replication in HCECs and HVEs via the quantification of chlamydial genomes over the course of a 36-hour infection.Surpris ingly, pathogen replication was substantially increased in HVEs over HCECs (Fig. S1A), comparable to infection of immortalized endocervical epithelial cells (End1/E6E7) previously used by our laboratory and others to model chlamydial infection (12,35).Consistent with this result, we observe significant differences in production of infectious chlamydial organisms after a 36-hour infection (Fig. S1B), with HCECs exhibiting a greater than 10-fold reduction in recovered inclusion-forming units (IFUs) relative to HVEs and End1s.Taken together, these results indicate that reduced chlamydial fitness in HVEs cannot account for their diminished pro-inflammatory and pro-EMT response to infection.
Alternatively, differences in pro-inflammatory and pro-EMT gene expression between infected HCECs and HVEs may be the product of cell type-intrinsic differences in basal gene expression.Thus, we next compared the transcriptome of mock-infected control HCECs to that of mock-infected HVEs, identifying 4,720 genes differentially expressed in an HCEC-intrinsic fashion (Fig. 1B).Functional characterization of this gene set via STRING revealed a similar pattern of enriched terms as in the comparison of Chlamydia-infected cells, with relative induction of innate immunity as well as repression of keratinization  and epidermal differentiation (Table 2).Expression of specific interleukins and CXCL-fam ily cytokines followed a similar trend, with infected HCECs exhibiting strikingly increased expression of CXCL10 (FC, 480.76),IL6 (FC, 1423.03), and IL8 (FC, 186.51), as well as modest induction of IL1B (FC, 3.73) and IL24 (FC, 6.52) (Table S1).However, expression of other interleukins, such as IL7, IL15, and IL20, differed only between Chlamydia-infected HCECs and HVEs (Fig. 1C).Importantly, both IL-15 and IL-20 have been implicated in fibrosis (36,37), suggesting that Chlamydia-infected endocervical epithelial cells may produce a wider portfolio of fibromodulatory cytokines relative to infection of other cell types.In contrast cytokeratin repression in HCECs was far more consistent between uninfected and Chlamydia-infected transcriptomes, as was relative induction of SNAI1 and TWIST1 (Fig. 1C).These data suggest HCEC-specific expression of EMT-associated genes may be the product of an intrinsic predisposition toward EMT induction, rather than an altered response to infection.

Cervical epithelial cells do not exhibit collagen or EMT induction during midcycle infection by C. trachomatis
While the effect of C. trachomatis infection on gene expression of host cervical epithelial cells has been extensively reported (20,21,38), prior study has focused on human papillomavirus (HPV)-transformed cell lines (e.g., HeLa, HEp-2).Given the potential for HPV transformation to dysregulate gene expression (39,40), we next defined a more physiologically relevant host transcriptome by assessing differential expression of Chlamydia-infected primary HCECs.Relative to mock-infected control HCECs, infection with C. trachomatis serovar L2 induced differential expression (FDRP ≤ 0.05) of 8,241 genes at 24 hpi (Table S1).In accordance with prior reports describing the extensive pro-inflammatory response to infection, functional characterization of this gene set via STRING exhibited enrichment of immune and inflammation-associated GO-BP terms, including both positive and negative regulation of type I/II interferon signaling (Table 3).
Critically, this included infection-dependent induction of specific interleukins (e.g., IL17, IL15, and IL20) and other genes associated with a type-I IFN response previously noted to exhibit HCEC-specific expression (Fig. 1B, Fig. 2A), such as IFITM1 (lFC, 7.56), BST2 (logFC, 6.63), IFI44L (FC, 7.00), and CXCL10 (FC, 112.38).Intriguingly, the infection-associated transcriptome exhibited repression of ECM components, with GO-CP (gene ontology, cellular component) terms associated with collagens and the ECM exhibiting modest enrichment among the downregulated gene set .Indeed, infected HCECs exhibited reduced expression of fibrillar (e.g., COL1A1 and  ), fibril-associated (e.g., COL16A1 and COL21A1), and network-forming (COL4A1) collagens relative to mock-infected HCECs, despite collagen expression of infected HCECs generally exceeding that of infected HVEs (Fig. 2B).The mesenchymal marker vimentin/VIM was repressed by infection (FC, −2.22), whereas common markers of EMT (e.g., CDH1/2, SNAI1/2, and ZEB1/2) were not differentially expressed (Fig. 2B), suggesting that infection does not induce EMT in endocervical epithelial cells at 24 hpi.To confirm this hypothesis, we next examined expression of EMT-associated markers at this time point during C. trachomatis serovar L2 infection of immortalized endocervical epithelial cells (End1/E6E7), which have previously been demonstrated to undergo EMT in a chronic/persistent model of C. trachomatis infection (35).Critically, infected End1s did not exhibit induction of mesenchymal differentiation markers (e.g., CDH2 and VIM) or repression of epithelial differentiation markers (e.g., CDH1 and KRT13) (Fig. 2C).In similar fashion, transcription factors associated with EMT induction did not exhibit induction at 24 hpi, with the notable exception of SNAI1/SNAIL (Fig. 2D).Collectively, these data imply that chlamydial induction of EMT in cervical epithelial cells does not occur by 24 hpi, despite an apparent predisposition toward mesenchymal differentiation relative to vaginal epithelial cells.Thus, pro-fibrotic gene expression observed at this time point likely occurs via an EMT-independent mechanism.

Induction of the pro-fibrotic transcriptional cofactor YAP is specific to infection of cervical epithelial cells
To identify alternative mechanisms of pro-fibrotic gene expression potentially modulated by Chlamydia, we next defined a common signature of pro-fibrotic genes exhibiting distinct expression in the upper genital tract as well as differential expression during C. trachomatis infection.Importantly, only a subset of genes differentially expressed in an HCEC-intrinsic fashion (mock HCEC vs mock HVE expression) exhibited comparable induction/repression in an infection-mediated fashion (mock HCEC vs Ct L2-infected HCEC expression).Upon comparing the HCEC-intrinsic (4,720 genes) and infection-medi ated (8,241 genes) differentially expressed gene sets, only 2,391 genes were common to both (Fig. 3A).Of these, 1,158 (49.4%) were equivalently induced or repressed in both comparisons, suggesting their likely relevance to infection-associated fibrosis of the upper genital tract (Fig. 3A).More than half (629 genes, or 54.3%) of this gene set mapped to the Comparative Toxicogenomics Database of pro-fibrotic genes (Fig. 3B), including numerous pro-fibrotic signal factors (e.g., IL6, IL24, and TGFA) and ECM-remod eling enzymes (e.g., ADAMTS9, MMP13, and TIMP2/3) (Table S2).Critically, expression of this gene set in infected HCECs significantly correlated with our previously repor ted bulk RNA sequencing of Ct serovar L2-infected End1/E6E7 immortalized endocer vical epithelial cells (Fig. 3B).These data suggest that modulation of this putatively fibrotic gene set is conserved between in vitro models of chlamydial pathogenesis in the upper genital tract, additionally recommending use of End1/E6E7 cells for subse quent investigation of the mechanisms underlying infection-associated pro-fibrotic gene expression.
One potential means by which infection may facilitate host expression of pro-fibrotic gene expression is via modulation of host transcription factors.Indeed, past work by our laboratory and others have defined mechanisms by which C. trachomatis and related species manipulate host transcription factor activity (12,41,42).Thus, we next performed transcription factor enrichment analysis of this putatively fibrotic gene set via ChIP-X Enrichment Analysis 3 (ChEA3) (43), identifying 56 transcription factors exhibiting significant enrichment (Table S3).This included known mediators of pro-fibrotic and pro-inflammatory signaling, such as RELA, IRF1/8, SMAD3, and EGR1.Strikingly, nearly half (25) of the enriched transcription factors are known regulatory targets of YAP (Table 4), a pro-fibrotic transcriptional cofactor of which we recently reported Chlamy dia-directed induction (12).Given the potential for YAP activation to drive expression of pro-fibrotic gene expression specific to the upper genital tract, we next assessed the degree to which C. trachomatis infection facilitates YAP nuclear translocation in vaginal epithelial cells.After infecting confluent monolayers of each cell type with C. trachomatis serovar L2, YAP nuclear translocation was quantified as a ratio of nuclear to cytosolic immunofluorescence.Critically, infected HVEs did not exhibit YAP nuclear translocation relative to mock-infected control monolayers, in stark contrast to the significant increase in nuclear YAP observed during infection of either primary (HCEC) or immortalized (End1) endocervical epithelial cells (Fig. 3C and D).Collectively, these data suggest that chlamydial YAP activation may mediate pro-fibrotic gene expression specific to infection of the upper genital tract.

Coculture with infected endocervical epithelial cells alters the putative myofibroblast transcriptome
Given our data indicating that infected endocervical epithelial cells exhibit induction of pro-fibrotic signal factors, one potential mechanism through which C. trachomatis may induce fibrosis is via signaling to stromal fibroblasts.Indeed, differentiation of tissue-resident fibroblasts into ECM-remodeling myofibroblasts has been consistently associated with other forms of fibrotic disease (26); however, the contribution of fibroblasts to Chlamydia-associated scarring has not been assessed.Importantly, the lack of clinical evidence of Chlamydia-infected fibroblasts in vivo necessitates a model of infection wherein uninfected fibroblasts are cocultured with infected endocervical epithelial cells.Past study of the kinetics of C. trachomatis infection epithelial cells indicates that pathogen-directed uptake of the bacterium is rapid, with a majority of invading bacteria internalized as soon as 10 minutes post-infection (64,65).Critically, this event requires interaction of Chlamydia with the host cell plasma membrane in two distinct stages: transient adhesion relying on electrostatic interactions with membrane glycosaminoglycans (66,67), followed by irreversible attachment due to binding with host cell receptors and/or engagement of the pathogen's type III secretion system (68).Thus, we infected immortalized End1s seeded at 35% of confluence with the genetically tractable C. trachomatis serovar L2, adding immortalized KCO2-44D uterine fibroblasts at 1 hpi.After 23 hours of incubation, cocultures prepared via this method exhibited a consistent morphology characterized by islands of epithelial cells interspersed with fibroblasts, of which the latter were readily distinguished by immunofluorescence staining of actin stress fibers (Fig. 4A).Importantly, fibroblasts did not exhibit visible chlamydial inclusions containing DAPI-positive organisms, in contrast to the readily apparent inclusions of infected epithelial cells (Fig. 4A).
To both confirm the absence of Chlamydia-infected fibroblasts and examine the transcriptome of uninfected fibroblasts, we then harvested cocultures at 24 hpi via trypsinization for library preparation and single-cell RNA sequencing.End1/KCO2 cell populations were readily distinguished on the basis of epithelial and mesenchymal marker expression, with End1s exhibiting heightened expression of cytokeratin 19 (KRT19) and KCO2s exhibiting heightened expression of vimentin (VIM) (Fig. 4B).To identify Chlamydia-infected cells, sequencing results were aligned to the hg18 refer ence genome appended with C. trachomatis reference sequences for highly expressed, polyadenylated RNA transcripts most likely to be captured by library preparation and sequencing, including the chlamydial outer membrane protein ompA, the polymorphic membrane protein pmpC, the nucleoside phosphate kinase ndk, and the pgp8 antisense sRNA-2.Expression of chlamydial reference sequences identified an epithelial cell cluster distinct from the mesenchymal population-consistent with our previous findings in bulk RNA-sequencing of both HCECs and End1s, this cluster exhibited induction of genes associated with a type I interferon response (e.g., IFI6, IFITM2, IFITM3, and HLA-A/C) (Table S4).Importantly, chlamydial marker genes were not detectable in uninfected cocultures (Fig. S2)-taken together, these data demonstrate the viability of this approach as a means to render cocultured fibroblasts inaccessible to Chlamydia.
To determine if coculture with infected epithelial cells altered the transcriptome of "bystander" fibroblasts, we performed clustering analysis on the mesenchymal cell population via Seurat.We identified four mesenchymal subclusters distinguished by marker gene expression, designated clusters F1-F4 (Fig. 4C, top), with the mesenchy mal population of Chlamydia-infected cocultures exhibiting a modest increase in the proportion of cluster F3 relative to a mock-infected control (24.3% and 17.4%, respec tively).Given that cluster F3 also exhibited enhanced expression of the myofibroblast differentiation marker smooth muscle actin (α-SMA/ACTA2), as well as tissue inhibitor of matrix metallopeptidases 1 (TIMP1), this result suggests that coculture with infec ted-epithelial cells may enhance myofibroblast differentiation (Fig. 4C, bottom).Further comparison of cluster F3 gene expression in Chlamydiaand mock-infected cocultures revealed differential expression of a variety of genes associated with fibrosis (Fig. 4D).For example, cluster F3 fibroblasts from Chlamydia-infected cocultures exhibited significant repression of HES1, a bHLH-family transcription factor known to inhibit the regulator of myofibroblast differentiation MyoD (69,70).Coculture with infected epithelial cells additionally inhibited expression of chromobox protein homolog 3 (CBX3), which was recently shown to inhibit proliferation and collagen production of vascular smooth muscle cells (71)."Bystander" cluster F3 fibroblasts also exhibited increased expres sion of pro-inflammatory genes, including C1QBP and ISG15, suggesting that cocul tured fibroblasts may respond to infection-associated induction of pro-inflammatory signaling by host endocervical epithelial cells.Collectively, these data imply that infected epithelial cells may signal to cocultured fibroblasts, thereby stimulating pro-fibrotic gene expression.

Chlamydia-infected endocervical epithelial cells induce fibroblast collagen I expression in a YAP-dependent fashion
A key functional consequence of myofibroblast differentiation is the capacity to remodel the extracellular matrix via the production of collagens (72,73).In healthy epithelial tissues, the ECM of the basement membrane is predominantly composed of type IV network-forming collagen, laminins, and the collagen/laminin crosslinking protein nidogen (74,75).In contrast, fibrotic tissues exhibit heightened proportionality of fibrillar collagens (e.g., types I, II, III, and V) in the basement membrane, which has been attributed to deposition by myofibroblasts in idiopathic pulmonary fibrosis and other forms of scarring disease (72,74,76,77).In similar fashion, it has been shown that Chlamydia-infected epithelial cells exhibit increased production of fibrillar collagen due to chlamydial induction of EMT and consequent host cell transdifferentiation into a myofibroblast-like phenotype (17).
In light of these results, one potential consequence of the altered fibroblast transcriptome in Chlamydia-infected cocultures may be increased production of fibrillar collagens.Thus, we prepared an alternative in vitro model of chlamydial infection, wherein End1 epithelial cells were seeded/infected on the semipermeable membrane of a transwell insert, washed with heparan sulfate at one hpi, then transferred to a plate containing seeded KCO2 fibroblasts (Fig. 5A).This approach provided a means to perform population-level measurement on cocultured fibroblasts, given that the 0.4-µm pore size of the transwell membrane would necessarily restrict passage of cells, while still allowing for cell-cell communication via the action of diffusible signal factors.Critically, KCO2s cocultured with Ct serovar L2-infected End1s in transwells for 23 hours (24 hpi) exhibited a significant increase in production of type I collagen relative to an uninfected control, as measured by immunofluorescence (Fig. 5B and C).Induction of collagen I expression occurred in a fashion consistent with mock-infected cocultures treated with TGF-β1 (a known inducer of myofibroblast differentiation and collagen deposition) (78)(79)(80).Consistent with this result, we observed a significant increase in expression of COL1A1 in cocultured fibroblasts as measured by reverse transcription quantitative PCR (RT-qPCR); importantly, this phenotype was observed in both infections with C. trachomatis serovar L2 and serovar D (Fig. 5D), suggesting that infection-associated induction of fibroblast activity is conserved between anogenital and urogenital serovars of Chlamydia.
We have previously demonstrated that C. trachomatis infection of endocervical epithelial cells induces expression of the fibroblast-activating signal factors CTGF and INHBA, in a fashion dependent upon pathogen-directed induction of the host transcrip tional cofactor YAP (12).Furthermore, here, we observe infection-mediated YAP activa tion specific to cervical epithelial cells, suggesting a role for chlamydial YAP activation in the fibrotic pathology associated with chlamydial infection of the upper genital tract.Thus, to assess the role of chlamydial YAP activation in stimulation of fibroblast collagen production, we next opted to knock down YAP expression in epithelial cells via siRNA at 24 hours prior to infection.After 23 hours of coculture (24 hpi), fibroblast coculture expression was assessed via RT-qPCR.Knockdown of YAP in Chlamydia-infected End1s was approximately 40% efficient, as measured by Western blotting of total YAP protein levels, in accordance with other reports of siRNA-mediated knockdown efficiency in End1s (12, 81) (Fig. S3).Fibroblast expression of collagen I was significantly increased in infected cocultures transfected with non-targeting control siRNA, consistent with our prior observations via immunofluorescence.However, transfection of infected End1s with YAP-targeting siRNA significantly attenuated this phenotype, with cocultured fibroblasts exhibiting collagen I (COL1A1) expression indistinguishable from mockinfected cocultures (Fig. 5D).Intriguingly, expression of the myofibroblast differentiation marker α-SMA (ACTA2) exhibited a similar, albeit statistically insignificant trend (Fig. 5E).Taken together, these data confirm that Chlamydia-infected epithelial cells stimulate fibroblast production of collagen, as well as implicate chlamydial induction of host cell YAP activity in this effect.

DISCUSSION
To elucidate the mechanisms underlying Chlamydia-mediated fibrosis of the upper genital tract, we have presented comparative analysis of both steady-state and infectionassociated gene transcription in primary HCECs and primary HVEs.Both mock-and Chlamydia-infected HCECs exhibited enhanced expression of pro-inflammatory and pro-fibrotic cytokines (e.g., IL6 and IL8), induction of the pro-EMT transcription fac tors SNAI1/SNAIL and TWIST1, and relative repression of multiple cytokeratins.How ever, expression of specific fibromodulatory interleukins (e.g., IL7, IL15, and IL20) was enhanced only in Chlamydia-infected HCECs, suggesting a cell type-specific response to infection.Transcription factor enrichment analysis of genes equivalently modulated by infection and HCEC-intrinsic expression relative to HVEs revealed a broad portfolio of known YAP regulatory targets.YAP nuclear translocation of YAP was enhanced during infection of HCECs, but not HVEs, suggesting a role for this transcriptional cofactor in mediating the HCEC-specific induction of pro-fibrotic signaling.To investi gate this possibility, we examined the response of uninfected fibroblasts to coculture with Chlamydia-infected endocervical epithelial cells, observing a modest expansion of α-SMA/TIMP1-expressing putative myofibroblasts.Further iteration on this coculture approach via use of transwells enabled population-level measurement of cocultured fibroblast gene expression.We find that cocultured fibroblasts exhibit heightened expression of type I collagen-critically, this phenotype was conserved between anogenital (L2) and urogenital (D) serovars of Chlamydia and sensitive to siRNA-mediated knockdown of YAP in infected endocervical epithelial cells.Taken together, our results present a novel mechanism by which Chlamydia mediates fibrotic tissue remodeling, wherein infection-mediated YAP activation in host endocervical epithelial cells promotes fibroblast collagen expression via the induction of pro-fibrotic signaling.
The apparent repression of collagen expression in infected HCECs is surprising, given past reports of chlamydial EMT leading to production of type I collagen in infected primary cells of the murine oviduct (17).Deposition of fibrillar collagen is a hallmark of pro-fibrotic remodeling of the ECM in other forms of scarring disease (74)(75)(76); in light of this, the role of Chlamydia-infected epithelial cells in collagen synthesis is critically understudied.These data suggest two possibilities: either previous reports of fibrillar collagen production by infected cells are unique to infection of murine oviduct epithelial cells, or infection inhibits collagen expression in host cells prior to induction of EMT.Importantly, prior study of chlamydial EMT induction indicates this phenotype occurs in late-cycle or persistent models of infection, with murine oviduct epithelial cells and End1 cells expressing mesenchymal markers at 72 hours or 7 days post-infection, respectively.Consequently, infection-mediated repression at 24 hpi may be a transient effect, superseded by pathogen-directed host transdifferentiation into a myofibroblast-like and collagen-producing phenotype.Indeed, prior reports indicating collagen expression is not enhanced in HeLa cells at 24 hpi would seem to support this hypothesis (20).Further study is required to elucidate the impact of chlamydial EMT induction on basement membrane architecture and thereby determine the physiolog ical relevance of this phenotype to the development of infection-associated fibrosis.Importantly, the infection-mediated induction of fibroblast collagen expression we observe may constitute a means by which Chlamydia may stimulate ECM deposition in an EMT-independent fashion.
Past study of C. trachomatis in murine models of infection has shown that the pathogen induces EMT in host epithelial cells of the murine oviduct, as well as host production of collagen I and TGF-β1 (a known inducer of myofibroblast differentiation) (13,16,17).However, fibrotic sequelae arising from C. trachomatis infection of the murine female genital tract has not been reported, suggesting that animal models of infection may not faithfully recapitulate the molecular events underlying Chlamydiaassociated scarring in humans.A likely explanation for this discrepancy is the repeated observation that murine intravaginal infections with C. trachomatis fail to ascend into the upper genital tract and are readily cleared by the innate immune system (82,83).
That being said, recent work has demonstrated transcervical infections with C. trachoma tis serovar D severely impact fertility in the mouse model (84,85), emphasizing the importance of infection of cervical epithelial cells to the development of chlamydial sequelae.Intriguingly, the immune response to transcervical C. trachomatis infections includes robust recruitment of CD4+ T-cells and secretion of IL-17 (85, 86)-features consistent with the cell type-intrinsic induction of IL-17 expression and CD4+ T-cell survival and chemotactic factors (e.g., IL-6 and IL-8) we observe in both uninfected and infected HCECs.In light of this, use of immortalized human cell lines in an in vitro model of infection thus has the potential to provide greater insight into how Chlamydia-infected cervical epithelial cells drive tissue-wide fibrotic pathologies.Here, we report that infected endocervical epithelial cells alter gene expression and induce collagen production in cocultured, uninfected fibroblasts-demonstrating a means by which C. trachomatis may coordinate pro-fibrotic activity of cell types beyond the simple columnar cervical epithelium the pathogen demonstrably infects in vivo and complementary to the immune cell recruitment that has been repeatedly associated with clinical study of Chlamydia-associated infertility.
The observation that collagen I expression in fibroblasts was sensitive to siRNA-medi ated knockdown of YAP in epithelial cells is in accordance with our prior report of infection-and YAP-dependent expression of pro-fibrotic signal factors (12).However, it is unclear to what extent specific signaling factors are relevant to or dispensable for induction of this phenotype.Given consistent role in induction of fibroblast differentiation and ECM remodeling (25,(87)(88)(89), this signal factor constitutes a likely candidate for YAP-and infection-dependent induction of collagen synthesis in fibroblasts in cocultured fibroblasts.That being said, YAP has also been implicated in the expression of several other pro-fibrotic signaling factors.A recent report implicated YAP in inducing expres sion of the pro-fibrotic interleukin IL-6 in the context of endometrial cancer (90).YAP has also been shown to enhance TGF-β expression and signal transduction via binding to and enhancing the activity of SMAD2/3 (48).It has also been implicated in the expression of the myofibroblast-activating interleukin IL-33, with Cre/Lox-mediated YAP knockout significantly attenuating IL-33 expression and scarring in a murine model of post-infarction cardiac fibrosis (91).YAP has been additionally shown to modulate the Wnt/β-catenin signaling pathway, via enhancement of both β-catenin expression and transcriptional activity (92).Given the demonstrable role of this signaling pathway in fibrosis generally and regulation of TGF-β expression specifically (93,94), these data collectively demonstrate multiple, complementary means by which chlamydial induction of YAP may facilitate pro-fibrotic signaling.Ultimately, further study of the role of YAP must begin with clearly defining the YAP regulon in Chlamydia-infected endocervical epithelial cells.This may necessitate the development of a YAP-knockout cell line, given the relative inefficiency of siRNA-mediated knockdown we and others observe in End1/ E6E7 cells (81).
It is important to note that chlamydial fibrosis is clinically associated with chronic or repeated infections, with the latter demonstrably increasing the risk of scarring infertility and/or ectopic pregnancy (6).Critically, the lack of a viable and tractable animal model of C. trachomatis-associated scarring has heretofore precluded more longitudinal study of the pro-fibrotic effects of chronic or repeated C. trachomatis infection.Given our observation that Chlamydia-infected cells stimulate cocultured fibroblast expression of collagen, these data suggest intriguing possibilities: that infected epithelial cells sensitize affected tissues to other pro-fibrotic stimuli or that repeated infections at the same site have an aggregate effect in stimulating scar-associated tissue remodeling.In a similar fashion, collagen deposition facilitated by infected epithelial cells may itself act as a stimulus of myofibroblast activation.Ultimately, the capacity of coculturestimulated fibroblasts to return to a naïve, collagen-repressed state must be assessed, as must the potential for successive rounds of coculture to enhance fibroblast activa tion and collagen synthesis.Further refinement of the in vitro model of chlamydial pathogenesis presented here may provide critical insight into how the pro-fibrotic stimulus of infected epithelial cells may effectively persist after host cell death and pathogen clearance, thereby facilitating the tissue-level scarring associated with chronic or persistent infection.

Chlamydial infections
Chlamydia trachomatis serovars L2 (434/Bu) and D were originally obtained from Dr. Ted Hackstadt (Rocky Mountain National Laboratory, NIAID).Chlamydial EBs were isolated from infected, cycloheximide-treated McCoy cells at 36-40 hpi and purified by den sity gradient centrifugation as previously described (95).For infection of 6-well plates (Greiner Bio-One 657160 and 662160) or 12-well plates (Costar 3513), HCECs or HVEs seeded at 5,000 cells/cm 2 and grown to confluence (6-7 days) or End1s seeded at 125% of confluence and incubated for 16 hours were washed with pre-warmed Hanks buffered saline solution (HBSS) prior to inoculation with Chlamydia-containing CECBM, VECBM, or KSFM (respectively) at the indicated multiplicity of infection (MOI).To synchronize infection of cultured monolayers, tissue culture plates were centrifuged at 4°C and 500 rcf (Eppendorf 5810 R tabletop centrifuge, A-4-81 rotor) for 15 minute to permit transient electrostatic adherence of C. trachomatis to the host cell surface.Inoculum was then aspirated, and cells were washed with chilled HBSS to remove any non-adherent EBs.After aspiration of the wash, chlamydial invasion was initiated via the addition of CECBM, VECBM, or KSFM (HCECs, HVEs, or End1s respectively) pre-warmed to 37°C.Infected cultures were then incubated for 24 hours at 37°C and 5% CO 2 for subsequent harvesting of total RNA (see below).
For infection of End1/KCO2 same-well co-cultures, End1s were seeded at 35% of confluence on six-well tissue culture plates (Greiner Bio-One 657160), incubated for 16 hours, then infected at an MOI of 5 as described above.Infected End1 cultures were returned to a tissue culture incubator until 1 hpi, at which time cultures were washed three times with pre-warmed HBSS containing 100 µg/mL heparin (Sigma-Aldrich H3393) to remove any transiently adhered, non-internalized EBs (96).After aspiration of the final wash, KCO2s in complete DMEM were seeded in each well at 75% of confluence.Tissue culture plates were subsequently incubated until 24 hpi prior to harvest for single-cell RNA library preparation (see below).
For transwell coculture experiments, End1s seeded at 100% of confluence in six-well-sized transwell inserts (Costar 3450, Fisher Scientific) and incubated for 16 hours (or 40 hours, in the case of experiments involving siRNA transfection) were infected at an MOI of 5 as above.At 1 hpi, cells were washed with HBSS containing 100 µg/mL heparin (Millipore-Sigma H3393) to remove transiently adhered chlamydial organisms.After replacement of wash media with DMEM, transwells were subsequently transferred to six-well plates containing KCO2s seeded at 60% of confluence immediately after infection.Cocultures were then incubated for 24 hours prior to fixation or RNA harvest (see below).

Bulk RNA sequencing and analysis
HVEs infected at an MOI of 2 as described above were harvested for RNA at 24 hpi using TRIzol (Thermo Fisher Scientific 15596026) and the DNA-free DNA removal kit (Thermo Fisher Scientific AM1906), according to manufacturers' protocols.Total RNA samples were subsequently assayed for fragmentation using an Agilent Bioanalyzer 2100.Polyadenylated transcript enrichment and cDNA library preparation of intact samples were performed using the NuGEN Universal mRNA-Seq Library Preparation kit, and sequencing of cDNA libraries was performed using the NextSeq 550 system (Illumina).HCEC and End1 libraries were prepared as described previously (12).
Read alignment and downstream analysis was performed using CLC Genomics Workbench (Qiagen); each treatment group was comprised of libraries from three biological replicates, each with a minimum of 30 million reads (unstranded single read, mean length 150 bp); and genes with an FDRP ≤ 0.05 were considered differentially expressed.Functional characterization of each differentially expressed gene set was performed using the STRING database of protein-protein association networks, using the ranked protein list functional enrichment analysis function (31,32).GO/Wikipath way overrepresentation analysis was also performed in R, using the Bioconductor package clusterProfiler (97).Pearson's correlation coefficients between HCEC, End1, and HeLa expressions of the fibrosis-associated gene set identified by clusterProfiler were subsequently calculated in R, using log 2 -transformed fold changes of all genes differentially expressed in at least one data set.Sequencing data for mock-infected and Ct L2-infected HVEs are available in the Gene Expression Omnibus, accession GSE228774; sequencing data for the equivalent HCEC and End1 infections are also available in GEO as part of our recent publication (accession no.GSE180784) (12).

Single-cell library preparation and sequencing
Infected cocultures were harvested as single-cell suspensions via the 10× Genom ics Sample Preparation Demonstrated Protocol (Revision B).Briefly, cocultures were disassociated in 0.25% Trypsin-EDTA and then harvested in complete DMEM.After centrifugation and resuspension in complete DMEM to eliminate residual trypsin, cell suspensions were filtered over 30-µm cell strainers (Miltenyi Biotec, 130-098-458) to eliminate clustered cells, washed twice in 1× phosphate-buffered saline (PBS) containing 0.04% bovine serum albumin (BSA), then filtered again using Flowmi 40 µM cell strainers (Bel-Art H13680-0040).After staining, aliquots of each suspension with trypan blue were used to validate a minimum of 95% cell viability for each sample as well as to determine suspension density for controlling cell library size.Suspensions were stored on ice for library preparation according to the 10× single-cell protocol (10× Genomics, Chromium Single Cell 3' v2 Chemistry).cDNA libraries were submitted to the UNMC Genomics Core facility for molarity quantification via qPCR, followed by sequencing using the NextSeq 550 platform and two 150-cycle High-Output flow cells.Sequencing data for this experiment are also available on GEO, accession no.GSE228647.

Single-cell RNA sequencing analysis
Sequencing data were demultiplexed and aligned to the hg38 reference genome via CellRanger (10× Genomics, version 6.0.1), using an expected cell count of 1,000 informed by hemacytometer counts performed on an aliquot of the single-cell suspension.After aggregating the mock-and Ct L2-infected alignment results for each replicate (n = 2), initial cell type determination was performed in Loupe Cell Browser (10× Genomics, version 3.1.1).Epithelial cell (End1) and fibroblast (KCO2) populations were identified on the basis of keratin 19 and vimentin expression, respectively.Subclustering analysis of the epithelial/fibroblast populations was performed in the Seurat R package (98).After associating per-cell expression data with their respective treatment (mock or Ct L2 infected) and cell type (epithelial cell or fibroblast) metadata, per-cell expression data were separated into separate End1/KCO2 Seurat objects.
Subsequent quality control and dimensionality reduction were performed analo gously to the Seurat guided vignette on single-cell expression analysis.After trim ming cells with extreme unique feature counts (200 < n < 8,000), outlier features were identified using the vst method (local polynomial regression of log-transformed mean/variance, selecting top 2,000 features).The dimensionality of each data set was determined via the Seurat JackStraw method; for both the End1 and KCO2 cell popula tions, the top 15 principal components encompassed the majority of variance in the data.Dimensionality reduction via uniform manifold approximation and projection was then performed on each data set, using their respective top 15 principal components, followed by cluster identification via Seurat's shared nearest neighbor algorithm (using a clustering resolution of 0.5).Per-cluster differential expression for each cluster was then determined via the FindMarkers function, using a cutoff of genes expressed by at least 25% of cells in each compared population.Per-cell and per-cluster gene expression was visualized using the Seurat DimPlot, FeaturePlot, and VlnPlot functions, splitting plots across treatment groups or clusters where applicable.

Immunofluorescence microscopy
For assessment of chlamydial titer in infected HCECs, HVEs, and End1s, confluent monolayers of each cell type were infected with C. trachomatis serovar L2 in 12-well plates as described above, then harvested at 36 hpi via scraping.Host cells were then lysed via centrifugation (20,000 rcf for 30 min at 4°C); pelleted chlamydial organisms and host cell debris were resuspended in 100-µL SPG buffer (219 mM sucrose, 10 mM Na 2 HPO 4 , 3.8 mM KH 2 PO 4 , 5 mM glutamic acid) for storage at −80°C.Aliquots of each lysate were diluted serially 1:100 and 1:1,000 in HBSS for infection of McCoy cells seeded in triplicate on glass cover slips and then infected via centrifugation as described above.
At least five fields of each cover slip were imaged using an Axio Observer inver ted fluorescent microscope (Zeiss, chlamydial titers) or CSU-W1 Spinning-Disk Confocal Microscope (Nikon, all other experiments).For titration of chlamydial IFUs, the number of inclusions (defined here as nuclei-adjacent structures exhibiting DAPI and anti-Chlamy dia staining) was counted per field and averaged across technical triplicates to determine a mean count of IFUs/field for each lysate, proportional to total IFUs/well by a ratio of areas (1:947.94).The total number of IFUs per lysate was then calculated by dividing the estimated total IFUs by volume of inoculum added to each well (500 µL), the multiplying this result by the dilution factor of the original sample lysate (1:100 or 1:1,000) and the total volume of each lysate (100 µL).Statistical analysis was performed in R, using pairwise Student's t-tests and Bonferroni's correction for multiple comparisons; P-values less than 0.05 were considered statistically significant.
For measurement of YAP nuclear translocation, 10 nuclei were selected at random per field using only the DAPI channel, manually masking the nuclear area, and record ing the mean YAP fluorescence intensity per nucleus.Blinded image quantification was performed by assigning image filenames randomized number codes.To account for variation in total YAP between cells, staining efficiency between cover slips, and compression of the nuclear/cytosolic compartments by the chlamydial inclusion, 10 cytosolic regions not occluded by an inclusion body and adjacent to measured nuclei were selected per field, with the mean YAP fluorescence intensity of these regions averaged to produce a per-field measurement of mean cytosolic YAP fluorescence intensity; nuclear translocation of YAP was thereby expressed as a ratio of mean nuclear fluorescence to mean cytosolic fluorescence.
For measurement of KCO2 collagen I expression, a selection mask of all cell area per field was defined via fluorescence thresholding of the phalloidin counterstain in ImageJ; per-field measurements of mean collagen I fluorescence intensity were subsequently recorded using this selection mask.Blinded image quantification was performed by assigning image filenames randomized number codes.Statistical analysis of YAP nuclear translocation and fibroblast collagen expression was performed in R, using a Kruskal-Wallis test to first verify a statistically significant (P-value < 0.05) difference between treatment groups.Subsequent pairwise comparisons were performed using a Wilcoxon rank sum test and Bonferroni's correction for multiple comparisons, with P-values less than 0.05 being considered statistically significant.

siRNA transfection
End1s were transfected with either an ON-TARGETplus non-targeting siRNA pool (Horizon Discovery D-001810-10-05) or an ON-TARGETplus YAP1-targeting siRNA SMARTpool (Horizon Discovery L-012200-00-0005) using Lipofectamine 3000 (Thermo Fisher Scientific L3000008), per manufacturer's instructions, at an empirically determined optimal concentration of 10 nM.At 16 hours post-seeding of End1s at 125% of confluence on six-well transwell inserts as described above, siRNA was combined in Opti-MEM (Thermo Fisher Scientific 31985062) with the Lipofectamine 3000 reagent, incubated for 5 minutes at room temperature to allow for liposome formation, then added to wells dropwise with mixing.Transfected End1s were then incubated for 24 hours prior to infection with Chlamydia as described above.

SDS-PAGE and Western blotting
To minimize activity of the chlamydial protease CPAF, End1s seeded on six-well plates, transfected with siRNA, and then infected at an MOI of 5 as described above were subsequently lysed in 1% SDS buffer heated to 95°C, as previously described (99).After treatment with Pierce Universal Nuclease (Thermo Fisher Scientific, 1:1,000 dilution) for 5 minutes at room temperature, lysates were combined with 4× Laemmli Sample Buffer (Bio-Rad 1610747) for loading on a 10% acrylamide/bis-acrylamide gel for SDS-PAGE (1.5 hour, 100 V).Gels were then transferred to PVDF membranes (Bio-Rad 1620177) using a semi-dry transfer method (50 minutes, 20 V).After blocking in 5% BSA in PBST (PBS containing 0.1% Tween-20) for 1 hour at room temperature, membranes were incubated overnight at 4°C with primary antibodies in 5% BSA-PBST: rabbit anti-YAP (CST 4912, 1:1,000 dilution) and rabbit anti-GAPDH (CST 2118, 1:1,000 dilution).Membranes were subsequently washed in PBST and then incubated with a goat anti-rabbit HRPconjugated secondary antibody (Dako P0448, 1:2,000 dilution in 5% BSA-PBST) for 2 hours at room temperature.After additional washing in PBST, membranes were imaged using Immobilon HRP Substrate (Millipore Sigma WBKLS0500) or an Azure Biosystems c600.Images were analyzed using the ImageJ gel analysis tool to quantify the fluorescence density of total YAP total protein relative to the GAPDH loading control.

Reverse transcription quantitative PCR
For assessment of host cell permissiveness to chlamydial infection, End1s, HCECs, and HVEs were seeded or grown to confluence on 12-well plates and infected with C. trachomatis serovar L2 at an MOI of 2, as described above.Wells were subsequently harvested for isolation of host and chlamydial genomic DNA (gDNA) at 0, 8, 16, 24, and 36 hpi, using the DNeasy Blood & Tissue kit (QIAGEN 69506), per manufacturer's instructions for adherent cells.After confirming recovered gDNA concentration and purity using a NanoDrop One spectrophotometer (Thermo Fisher Scientific), the amount of chlamydial gDNA in each sample was determined via quantitative PCR.Briefly, 20-ng aliquots of each sample were tested in triplicate against a standard curve of 2 pg-20 ng C. trachomatis serovar L2 gDNA, prepared via 1:10 serial dilution (R 2 ≥0.995).Quantitative PCR of the C. trachomatis serovar L2 gene euo was performed on a QuantStudio 3 (Thermo Fisher Scientific) using PowerUp SYBR Green (Thermo Fisher Scientific A25742) and the following primers: 5′-GCTGTTCCTGTTACTTCGCAAA-3′ (euo_F), 5′-AACATAGATAG CCTGACGAGTCACA-3′ (euo_R).

FIG 1
FIG 1 Cervical epithelial cells exhibit intrinsic pro-inflammatory and pro-EMT gene expression relative to vaginal epithelial cells.(A and B) Volcano plot of differential gene expression of Chlamydia trachomatis serovar L2-infected (A) or mock-infected (B) primary HCECs relative to equivalent HVEs at 24 hpi (see also Table S1).n = 3, with a minimum of 3 × 10 7 unstranded single reads per replicate with a mean length of 150 bp.All fold changes are relative to the mock-infected control; red dots: FDRP ≤ 0.05, labels: top 20 genes with lowest FDRP values.(C) Heatmap of cytokine and EMT-associated gene expression in mock-infected (left columns) and C. trachomatis serovar L2-infected (right columns) HCECs.All fold changes are relative to an equivalent infection in HVEs; only genes differentially expressed (FDRP ≤ 0.05) in at least one comparison are shown.

FIG 2
FIG 2 Cervical epithelial cells do not exhibit collagen or EMT induction during midcycle infection by C. trachomatis.(A and B) Heatmap of cytokine (A) and collagen/EMT-associated (B) gene expression in mock-infected HCECs relative to equivalent HVEs (left columns), Ct L2-infected HCECs relative to equivalent HVEs (center columns), and Ct L2-infected HCECs relative to mock-infected HCECs (right columns).Only genes differentially expressed (FDRP ≤ 0.05) in at least one comparison are shown.(C and D) Expression of epithelial/mesenchymal differentiation markers (C) and EMT-associated transcription factors (D) at 24 hpi in mock-and Ct L2-infected End1/E6E7 immortalized endocervical epithelial cells, as measured by reverse transcription-quantitative PCR.n = 3 biological replicates; fold changes are relative to mean expression of the mock-infected and untreated control.Whiskers: minimum to maximum; asterisks: P-values ≤ 0.05, using pairwise Student's t-tests and

FIG 3
FIG 3 Induction of the pro-fibrotic transcriptional cofactor YAP is specific to infection of cervical epithelial cells.(A) Venn diagrams of genes differentially expressed (top, FDRP ≤ 0.05), induced (middle, FDRP ≤ 0.05, log 2 FC > 0), or repressed (bottom, FDRP ≤ 0.05, log 2 FC < 0) in an HCEC-intrinsic (red, mock-infected HCECs vs mock-infected HVEs) or infection-mediated fashion (green, Ct L2-infected HCECs vs mock-infected HCECs).(B) Scatter plot of gene expression Ct L2 infection of HCECs (x-axis) and End1s (y-axis) at 24 hpi of the subset of genes equivalently modulated by steady-state HCEC expression relative to HVEs (see also Table S2).All fold changes are relative to each cell type's respective mock-infected control; blue line: linear regression model of correlation; gray shading: 95% confidence interval.R 2 and P-values calculated using Pearson's correlation.(C) Representative micrographs of YAP (green) translocation into the nuclei (blue) of confluent mock-and Ct L2-infected HVEs, HCECs, and End1s at 24 hpi.Scale bar: 20 µm.(D) Quantification of YAP nuclear translocation in C as a ratio of nuclear to cytosolic YAP fluorescence normalized to each cell type's respective mock-infected control.n = 3 biological replicates, 50 cells measured per sample.Black bars: group means; asterisks: P-values ≤ 0.05, using pairwise Wilcoxon rank sum tests and Bonferroni's correction for multiple comparisons.

FIG 5
FIG 5 Chlamydia-infected endocervical epithelial cells induce fibroblast expression of collagen I in a YAP-dependent fashion.(A) Schematic representation of the transwell-mediated approach of coculturing C. trachomatis-infected endocervical epithelial cells (top compartment) with uninfected uterine fibroblasts (bottom compartment).(B) Representative micrographs of collagen I (green) produced by KCO2 fibroblasts cocultured with mock-infected, Ct L2-infected, or TGF-β1-treated (20 ng/mL) End1 cells for 24 hours, starting at 1 hpi.Scale bar: 20 µm.(C) Quantification of mean fluorescence intensity of collagen I in B. n = 3 biological replicates, five fields measured per sample.Black bars: group means; asterisks: P-values ≤ 0.05, using pairwise Wilcoxon rank sum tests and Bonferroni's correction for multiple comparisons.(D) COL1A1 expression as measured by RT-qPCR in KCO2 fibroblasts cocultured for 24 hours with mock-, Ct L2-, or Ct D-infected End1 cells (coculture starting at one hpi).n = 4 biological replicates; fold changes are relative to mean expression of the mock-infected control.Whiskers: minimum to maximum; asterisks: P-values ≤ 0.05, using Student's t-tests and Bonferroni's correction for multiple comparisons.(E) COL1A1 and ACTA2 expression as measured by RT-qPCR in KCO2 fibroblasts cocultured for 24 hours with siRNA-transfected (10 nM of YAP1-targeting or non-targeting siRNA, for 24 hours prior to infection), Ct L2-infected End1 cells (coculture starting at 1 hpi).n = 3 biological replicates; fold changes are relative to mean expression of the equivalent mock-infected control.Whiskers: minimum to maximum; asterisks: P-values ≤ 0.05, using Student's t-tests and Bonferroni's correction for multiple comparisons.

TABLE 1
GO biological process terms significantly enriched (FDR ≤ 0.05) in comparison of C. trachomatis serovar L2-infected primary HCECs to primary HVEs at 24 hours post-infection via bulk RNA sequencing a

GO-BP term description Genes mapped Enrichment score Modulation False discovery rate
a Induction/repression of each term is relative to the HVE infection.

TABLE 2
GO biological process terms significantly enriched (FDR ≤ 0.05) in comparison of steady-state expression in primary HCECs and primary HVEs via bulk RNA sequencing a

TABLE 3 GO biological process terms significantly enriched (FDR ≤ 0.05) in comparison of C. trachomatis serovar L2-infected HCECs to mock-infected control HCECs via bulk RNA sequencing a GO-BP term description Genes mapped Enrichment score Modulation False discovery rate
a Induction/repression of each term is relative to the mock infection.

TABLE 4
YAP-associated transcription factors exhibiting significant enrichment (FDR ≤ 0.05) in the set of genes equivalently modulated both by Ct L2 infection of HCECs and steady-state HCEC expression relative to HVEs