Tumor Necrosis Factor Alpha-Induced Interleukin-1 Alpha Synthesis and Cell Death Is Increased in Mouse Epithelial Cells Infected With Chlamydia muridarum

Abstract

Chlamydia trachomatis-genital infection in women can be modeled in mice using Chlamydia muridarum. Using this model, it has been shown that the cytokines tumor necrosis factor (TNF)α and interleukin (IL)-1α lead to irreversible tissue damage in the oviducts. In this study, we investigated the contribution of TNFα on IL-1α synthesis in infected epithelial cells. We show that C muridarum infection enhanced TNFα-induced IL-1α expression and release in a mouse epithelial cell line. In addition to IL-1α, several TNFα-induced inflammatory genes were also highly induced, and infection enhanced TNF-induced cell death. In the mouse model of genital infection, oviducts from mice lacking the TNFα receptor displayed minimal staining for IL-1α compared with wild-type oviducts. Our results suggest TNFα and IL-1α enhance each other’s downstream effects resulting in a hyperinflammatory response to chlamydial infection. We propose that biologics targeting TNF-induced IL-1α synthesis could be used to mitigate tissue damage during chlamydial infection.

Chlamydia trachomatis infection has increased by 60% in the last decade in the United States. Chlamydial infection can cause persistent inflammation leading to pelvic inflammatory disease in approximately 10% of cis-gender women and other individuals assigned female sex at birth [1]. Infection of genital tract epithelial cells initiates a robust inflammatory response [2, 3]. Using the female mouse model of genital Chlamydia infection, it has been shown that the inflammatory response in the genital tract can persist even after the infection is cleared [4]. Therefore, besides antibiotic treatment to clear infection, there is a critical need to develop therapies that prevent upper genital tract sequelae.

Chlamydia muridarum infection in the mouse genital tract serves as a model system to study oviduct inflammation and tissue damage. Using this model, we and others have shown that the interleukin-1 receptor (IL-1R) and tumor necrosis factor-α receptor (TNFR) signaling significantly contribute to oviduct pathology [5–7]. The cytokine TNFα, which signals via TNFR, is also shown to be essential for oviduct tissue damage [8, 9]. Using mice deficient for the cytokines IL-1α or IL-1β, which bind to IL-1R, we have shown that IL-1α plays a larger role in IL-1R-mediated pathology [9]. Interleukin-1α is an alarmin released from dying cells under inflammatory conditions, recruiting immune cells to the damaged tissue site (reviewed in [10]). Interleukin-1α is expressed predominantly in infected genital tract epithelial cells and contributed to neutrophil recruitment and long-term fibrosis in the genital tract. It is notable that low levels of IL-1α persisted in the tissue after the infection was cleared, suggesting low-grade inflammation in the genital tract [9]. These findings demonstrate that during chlamydial infection, tissue-damaging inflammation is largely driven by TNFα and IL-1α. Increased levels of TNFR-induced and IL-1R-induced cytokines have also been detected in cervical secretions of women with endometrial infection [11], supporting their pathologic role in human chlamydial infection.

The finding that mice lacking either TNFα or IL-1α are protected from oviduct disease during infection suggests an interdependence between the 2 cytokine synthesis or activity. Tumor necrosis factor-α has been shown to induce IL-1α in epithelial cells, which in turn can function in an inducible autocrine manner to enhance TNF-mediated inflammatory effects [12]. Therefore, we hypothesize that the expression and/or inflammatory effect of IL-1α and TNFα are interdependent and amplified in infected cells. To test this hypothesis, we examined the effect of TNFα on IL-1α synthesis in Chlamydia-infected mouse epithelial cells. Relative to uninfected cells exposed to TNFα, exposing infected cells to TNFα increased IL-1α synthesis and release. Several inflammatory cytokines were also increased in parallel, and increased cell death was observed in TNFα-exposed infected cells. Moreover, in the mouse model of genital C muridarum infection, mice deficient for TNFR had reduced IL-1α synthesis in oviducts with minimal destruction to the epithelial layer. These studies suggest that blocking TNFR signaling to reduce IL-1α synthesis could protect the oviducts during Chlamydia infection.

MATERIAL AND METHODS

Chlamydial Strain, Cell Lines, and Reagents

Mouse chlamydial strain C muridarum (AR-Nigg) has been cultured and titrated in our laboratory [13]. Mouse oviduct epithelial cell line Bm1.11 was cultured as described [14]. Interleukin-1α antibody was purchased from Abcam (ab9724), and recombinant mouse interferon (IFN)γ and TNFα were from PreproTech.

Interferon-γ Pretreatment and Tumor Necrosis Factor-α Exposure

Bm1.11 oviduct epithelial cells (20 000/well in a 24-well plate) were pretreated with IFNγ at 20 ng/mL for 6 hours, infected with C muridarum at 1 multiplicity of infection by centrifuging the plate at 3000 rpm for 60 minutes at 37°C. Cells were washed and exposed to TNFα (50 ng/mL), and 18 hours after infection, supernatant and cell lysates were collected, or cells were fixed for immunostaining.

Interleukin-1α Messenger Ribonucleic Acid, Enzyme-Linked Immunosorbent Assay, Bioassay, Immunostaining, and Cell Death Assay

Interleukin-1α protein in the supernatant was determined by enzyme-linked immunosorbent assay (ELISA) using mouse IL-1α Duo set (BioLegend). Total ribonucleic acid (RNA) was extracted using QIAGEN RNeasy kit and IL-1α messenger RNA (mRNA) expression in cells determined by quantitative reverse-transcription polymerase chain reaction as described earlier [13] using primers GCCACCAAAGAACAAAGT and ATTACAGAGACTCAGGACAT, and normalized to actin gene expression. For IL-1α Bioassay, a HEK-IL-1R reporter cell line (HEK-IL-1Ra; InvivoGen) expressing IL-1R1 linked to secretory alkaline phosphatase (SEAP) as a reporter was used. Cells were exposed to supernatant from Bm1.11 cells for 16 hours, or recombinant mouse IL-1α, and SEAP in the supernatant was quantified using Quantiblue substrate (InvivoGen). Immunostaining was done as described earlier [15]. To measure the rate of apoptosis and necrosis in cells exposed to TNFα, Bm1.11 cells were plated in 96-well plate and run in triplicate for each condition. RealTime-Glo Annexin V Apoptosis and Necrosis assay (Promega) was used, and luminescence and fluorescence were measured periodically on a Biotek Synergy plate reader.

Mice Strains and Infection

Age-matched (8–10 weeks) TNFR1 knockout (KO) mice (catalog ID003242) and wild-type (WT) C57BL/6J (catalog ID 000664) female mice were purchased from Jackson Laboratory. Infections were performed with 3 × 10⁵ inclusion forming units of C muridarum, as described earlier [9]. Mice were sacrificed at day 7 postinfection, and genital tracts were processed for paraffin embedding. All studies were approved by the Institutional Animal Care and Use Committee at the University of North Carolina.

Immunohistochemistry and Immunofluorescence Staining of Tissue Sections

Immunohistochemistry staining for IL-1α in mouse tissue is described in detail [9]. For immunofluorescence (IF) staining of chlamydial inlcusions, 4-μm sections were de-paraffinized and rehydrated, and heat-induced epitope retrieval was performed using pH 6.0 buffer (no. TA-135-HBL; ThermoFisher). Sections were blocked in 10% normal goat serum for 1 hour and incubated overnight at 4°C in antichlamydial mouse serum at 1:100. Sections were then incubated in Alexa Fluor 568 goat antimouse immunoglobulin G (1:2000; no. A11031; Life Technologies) for 2 hours at room temperature. Sections were coverslipped using Fluorogel II with 4’,6-diamidino-2-phenylindole (no. 17985-50; EMS Bio).

Ribonucleic Acid Library Preparation, Sequencing, Data Processing, and Differential Expression Analysis

Bm1.11 cells were IFNγ-primed, and total RNA was collected from 4 conditions for RNAseq: uninfected, C muridarum-infected, uninfected cell exposed to TNFα, and infected cells exposed to TNFα. Total RNA (250 ng) from each sample was then used in KAPA RNA HyperPrep Kit with RiboErase (HMR) and dual indexed with a KAPA Dual-Indexed Adapter Kits. Libraries were sequenced using a Nextseq High-Output Kit utilizing 76-base pair paired end reads. Raw RNA-seq reads were demultiplexed and quality trimmed using the standard Illumina bcl2fastq software. Reads were aligned to a custom mouse ribosomal RNA sequence index and then to the mouse genome (GRCm38) with GENCODE annotation (vM16) using STAR v2.5.2b in quantmode [16]. Genes with more than 30 raw read counts in at least 2 biological samples were included in expression comparisons. The raw gene read count data were normalized using DESeq2’s regularized logarithm (rlog) transformation [17]. Differentially expressed genes were defined as those with an absolute log₂-fold change greater than 1.

Statistical Analysis

Comparison between groups in in vitro experiments was done using one-way analysis of variance. All statistical analyses were performed using GraphPad’s PRISM software.

RESULTS

Tumor Necrosis Factor-α Exacerbates Chlamydia-Induced Interleukin-1α Expression and Release in Mouse Epithelial Cells

Interleukin-1α is detectable at >2 ng/mL by ELISA in genital secretions of C muridarum-infected mice but not in uninfected mice [9]. However, when mouse epithelial cells were infected with C muridarum, IL-1α protein was detected only at low levels (<2 pg/mL) at 32 hours postinfection in the supernatant and cell lysates (data not shown). The low levels of IL-1α in infected cells indicated an important second stimulus that was present in vivo is absent in this in vitro infection model. We hypothesized that TNFα could be the second stimulus, because TNFα (1) has been shown to induce IL-1α in epithelial cells and (2) functions in an inducible autocrine manner [12]. To test this hypothesis, infected epithelial cells were exposed to TNFα for 16–18 hours, and cell supernatants were analyzed for IL-1α. Cells were briefly pretreated with IFNγ before TNFα exposure, because it removes the inhibitory effect of STAT-1 on the TNFR signaling complex [18]. Cells pre-exposed to IFNγ or exposed to TNFα alone were used in parallel. Tumor necrosis factor-α exposure of IFNγ-pretreated cells induced IL-1α in the supernatant, but this effect was significantly increased by >5-fold in infected cells treated similarly (Figure 1A). To determine whether protein levels in supernatant were a result of an increase in IL-1α transcription, we examined the total RNA from cells for IL-1α transcripts. Relative to uninfected cells, IL-1α mRNA was upregulated in infected cells exposed to TNFα, with or without IFNγ prestimulation, by >1000-fold (Figure 1B). Comparison of levels of IL-1α protein in supernatant and RNA in cells suggests that prestimulation of cells with IFNγ before TNFα exposure and infection significantly increases IL-1α protein.

Figure 1.

Tumor necrosis factor (TNF)α exacerbates Chlamydia-induced interleukin (IL)-1α gene expression. Bm1.11 oviduct epithelial cells (20 000 cells/well) were prestimulated with interferon IFNγ (pIFNγ) (20 ng/mL) for 6 hours. Interleukin-γ-exposed cells were washed and infected with Chlamydia muridarum at 1 multiplicity of infection (MOI) and exposed to TNFα (50 ng/mL). Interleukin-1α protein in the supernatant was determined by enzyme-linked immunosorbent assay (A), and messenger ribonucleic acid expression in cells was determined by quantitative reverse-transcription polymerase chain reaction (B) at 16–18 hours postinfection. In a separate experiment, cells were exposed to a variable concentration of TNFα, and cells infected with 1 MOI of C muridarum or equivalent amounts of ultraviolet (UV)-killed C muridarum. Cell supernatants (C) or cell lysates (D) were added to HEK293T cells expressing IL-1 receptor and nuclear factor κB-driven secreted alkaline phosphatase (SEAP). The SEAP activity was assayed in cells after 18 hours as an index of bioactive IL-1α added. Significance was determined by one-way analysis of variance. (*P < .05, **P < .01, and ****P < .0001). A and B are representative of 4 experiments, and C and D are representative of 2 experiments. UT (untreated).

Interleukin (IL)-1α Released From Tumor Necrosis Factor-Exposed Infected Cells Is Bioactive, and IL-1α Induction Requires Infection With Viable Chlamydiae

To determine whether the IL-1α released in the supernatant from infected cells is functional, we used a bioassay using reporter HEK293T cells expressing IL-1R, in which IL-1α binding to IL-1R leads to expression of SEAP. Bm1.1 cells pre-exposed to IFNγ were treated with variable doses of TNFα and infected with C muridarum as described earlier. Cells were mock-infected with UV-killed C muridarum to determine the effect of chlamydial growth on IL-1α synthesis and release. Both supernatants and cell lysates were collected and tested in the bioassay to compare intracellular and released IL-1α. Tumor necrosis factor-α activation at as low as 5 ng/mL led to IL-1α synthesis, and maximal stimulation was observed at 25 ng/mL (Figure 1C). Ultraviolet-killed chlamydiae failed to induce IL-1α expression, which suggests that infection with viable chlamydiae is essential for upregulating TNF-induced IL-1α expression.

Interleukin-1α Is Predominantly in Its Proform in Infected Cells

Interleukin-1α is synthesized in its proform, which translocates to the nucleus because it contains a nuclear localization signal [19]. Pro-IL-1α can be cleaved by enzymes such as calpain, triggered by cell stress, to form mature IL-1α, which is cytoplasmic [20]. However, both pro- and mature IL-1α are biologically active [21]. To determine the localization of IL-1α in infected cells, we performed IL-1α IF staining of infected cells exposed to TNFα. Interleukin-1α staining was more pronounced in TNFα-exposed infected cells primed with IFNγ, relative to unprimed cells. Staining for IL-1α protein was predominantly restricted to cell nuclei, suggesting that it is the uncleaved pro-IL-1α (Figure 2). Chlamydial staining revealed that cells with chlamydial inclusion stained bright for IL-1α, although some cells in the infected culture with no inclusions also stained for IL-1α.

Figure 2.

Interleukin (IL)-1α was predominantly found in the cell nucleus in Chlamydia-infected cells. Bm1.11 cell primed interferon (pIFNγ) or not primed with IFNγ were infected with Chlamydia muridarum (1 multiplicity of infection) and exposed to tumor necrosis factor (TNFα) or untreated (UT). Cells were fixed and stained for chlamydial inclusion (green) and IL-1α (red) 24 hours postinfection and mounted with a mounting solution containing 4’,6-diamidino-2-phenylindole ([DAPI] blue), which stains for the nucleus. Inclusions are indicated by asterix, and nuclear IL-1α are indicated with an arrow. Data are representative of 3 experiments.

Tumor Necrosis Factor-α Exposure of Infected Cells Increased Cell Death and Induced an Hyperinflammatory State

Because IL-1α was predominantly nuclear in the cells, its detection in the supernatant was likely a result of cell death. Therefore, we tested whether TNFα exposure of infected cells increased cell death using a real-time cell death assay to measure apoptosis and necrosis. In infected cells primed with IFNγ, we observed that TNFα enhanced apoptotic and necrotic cell death, compared with uninfected cells primed with IFNγ and exposed to TNFα (Figure 3A). Apoptotic cell death was observed starting from 5 hours and reached a maximum at 12 hours after TNFα exposure and was marginally but significantly increased in infected cells compared with uninfected cells exposed to TNFα. Necrosis followed apoptosis and was observed 12 hours after TNFα exposure, and it was also significantly increased in infected cells versus uninfected cells exposed to TNFα. Furthermore, because TNFR signaling could increase the expression of several inflammatory genes in addition to IL-1α, we performed RNA sequence (RNAseq) from RNA obtained from IFNγ-primed infected and uninfected cells exposed to TNFα. The RNAseq revealed that infected epithelial cells exposed to TNFα highly upregulated several inflammatory genes besides IL-1α, including CXCL9, NOS2, CXCL11, CXCL10, CXCL1, CCL2, IL-6, and TNF, relative to uninfected cells exposed to TNFα or infected cells (Figure 3B). Therefore, TNFα-induced cell death and gene expression was enhanced in infected cells, relative to uninfected cells exposed to TNF or unexposed infected cells.

Figure 3.

Increased cell death and inflammatory gene expression in infected cells exposed to tumor necrosis factor (TNF)α. Bm1.11 cells primed with interferon (IFN)γ were exposed to TNFα with or without infection, and the rate of apoptosis and necrosis was determined using the real-time glo-Annexin necrosis assay. Each sample was run in triplicate, and data were analyzed for significance by one-way analysis of variance. Luminescence and fluorescence were measured as an index for apoptosis and necrosis, respectively (A). Gene expression in infected cells with or without TNFα exposure is shown relative to uninfected cells, where genes differentially expressed in at least 1 of the 3 treatment groups are shown (B). Arrows indicate genes whose overexpression was confirmed by quantitative reverse-transcription polymerase chain reaction.

Interleukin-1α Expression Is Reduced in Chlamydia-Infected TNFR1 Knockout Mice

To determine whether TNF signaling is essential for increased IL-1α expression in the mouse genital tract, mice deficient for TNFR1 (TNFR1-KO mice) were infected in parallel with WT mice and genital tract excised at day 7, at a time point where we observed increased IL-1α expression in epithelial cells in WT mice [9]. Interleukin-1α staining was predominantly on the epithelial cells as reported earlier [9], and the level and pattern of staining were similar between WT and TNFR1 KO mice uterus (data not shown). However, in the oviducts, IL-1α staining was observed in WT oviducts but was very weak in a TNFR1-KO oviduct (Figure 4A). Although the data are not quantitative, reduced IL-1α staining was observed in all 5 TNFR1-KO oviducts compared with WT mice. Tissue architecture of the oviduct showed significant destruction of the epithelial barrier in the WT oviducts, whereas they were retained in the TNFR1-KO mice. Neutrophils were observed in the oviduct lumen of both strains. When adjacent tissue sections were stained for chlamydiae, chlamydial inclusions were observed in the epithelial cells of both WT and TNFR1-KO oviducts (Figure 4B). These results suggest that TNFR signaling contributes to IL-1α expression in the oviduct. Furthermore, in the absence of TNFR signaling, the epithelial structure is retained despite increased infection and neutrophil recruitment.

Figure 4.

Interleukin (IL)-1α staining is reduced in oviducts of tumor necrosis factor-α receptor-knockout (TNFR-KO) mice infected with Chlamydia muridarum. Wild-type (WT) and TNFR1-KO (N = 5) mice were infected with 3 × 10⁵ inclusion forming units/mouse, and mice were sacrificed at day 7 postinfection. Genital tracts were excised, fixed in formalin, and paraffin-embedded section stained for IL-1α by immunohistochemistry (IHC) (A) or for Chlamydia by immunofluorescence (B). Brown staining indicates IL-1α in IHC (A), and red staining indicates chlamydial inclusions (B) in a representative mouse in each group.

Discussion

Tumor necrosis factor-α and IL-1 receptor signaling are essential for oviduct pathology in a mouse model of genital C muridarum infection [5, 6]. Tumor necrosis factor-α signaling via TNFR is suggested to contribute to CD8 T cell-mediated oviduct pathology during chlamydial infection [8]. Interleukin-1α contributes to increased neutrophil recruitment to the infection site and tissue fibrosis, but it is dispensable for chlamydial clearance [9]. Our long-term goal is to identify the molecular and cellular interactions of TNFR and IL-1R signaling during chlamydial infection, to therapeutically target ongoing inflammation and mitigate oviduct disease. The primary objective of this study was to determine what drives IL-1α production in Chlamydia-infected cells. Using an in vitro cell culture model, we show that TNFR signaling is enhanced in infected cells and increases IL-1α synthesis, release, and cell death. The effect of increased TNFα signaling also leads to overexpression of several inflammatory genes. Our study suggests that blocking TNF-dependent IL-1α synthesis can protect from oviduct pathology, without affecting chlamydial clearance.

IL-1a is a nuclear factor (NF)κB inducible gene [12], and NFκB activation is observed during C muridarum infection [22, 23]. However, infection alone did not lead to robust IL-1α expression in cell culture, although it was observed in infected mice. This led us to hypothesize that TNFα may be essential for IL-1α synthesis. Tumor necrosis factor-α-mediated IL-1α expression is a 2-step activation process [12], requiring first rapid release of preformed mature transforming growth factor α, which binds to epidermal growth factor receptor (EGFR) promoting MEK/ERK activation. The combined effect of EGFR signaling coupled with TNFR signaling is essential for IL-1α expression [12]. When infected cells were exposed to TNFα in vitro, we found that IL-1α was induced several fold. In addition, chlamydial replication was essential for augmented TNFα signaling and IL-1α expression. Therefore, it is likely that chlamydial effectors contribute to enhanced TNFα signaling in infected cells. In fact, it has been shown that chlamydial infection activates the ERK pathway for its survival [24], and activation of EGFR is required for C trachomatis development [25], both of which could be potential mechanisms for enhancing the downstream effects of TNF-signaling. As a result, differential gene expression is not limited to IL-1α and was also observed for downstream inflammatory genes such as IL-6.

In addition to increased inflammatory gene expression, infected cells showed increased cell death upon TNFα exposure. There is conflicting literature on the effect of TNF on Chlamydia-infected cells. Waguia Kontchou et al [26] showed that in infected cells, TNF-induced NFκB activation was not affected, whereas internalization of TNF-TNFR was blocked, preventing TNF-induced apoptosis. Sixt et al [27] showed that although cells are protected from TNF-induced apoptosis during the early and mid-stage of infection, they are susceptible to TNF-induced cell death late in infection and die by a mechanism that resembles necrosis. Our results are more in line with the latter, showing that chlamydial antiapoptotic activity is insufficient to protect its intracellular environment under inflammatory conditions. Furthermore, we show that TNF-induced cell death is, in fact, enhanced in infected cells. A key difference between our study and others is that unlike previous studies that used cycloheximide-treated cells to study TNF-signaling, we have prestimulated the cells with IFNγ, to remove the natural inhibition on TNFR. Pre-stimulation with IFNγ is similar to what is observed in an infected tissue where cells are exposed to IFNγ from natural killer or T cells, and TNFα from multiple inflammatory cells.

Our strongest evidence for the TNFR and IL-1α synergistic loop comes from in vivo studies in the mouse model. Mice lacking TNFR showed significantly reduced IL-1α staining in their oviduct at day 7 postinfection, compared with WT mice. It is notable that this difference was observed only in the oviducts and not in the uterine horns, which suggests that TNFR signaling-mediated IL-1α is not the sole mechanism of IL-1α synthesis and could be a tissue-specific phenotype, such as differential expression of TNFR. Tumor necrosis factor-α receptor KO oviducts were infected at similar levels in WT oviducts and had several neutrophils in the lumen, which surprisingly did not have a damaging effect on the epithelial cells. These data support our in vitro results that the absence of TNFR signaling and IL-1α reduce the “prodeath” effect of TNFα, protecting the tissue from long-term damage. The presence of a large number of neutrophils that do not cause damage in the TNFR KO oviducts also suggests a contribution of TNFR signaling in neutrophil activation, in this model.

Conclusions

Although it is useful, the mouse model of genital infection has a few differences from human infection, such as requirement for depo-provera, high dose of infection, rapid ascension to oviducts, and significant immunity to reinfection (reviewed in [28]). Nevertheless, the contribution of TNFα and IL-1α during inflammation is described in several inflammatory disease in humans [29, 30], which suggests that observation from mice in this study could be valid in humans. In infected women, antibiotic treatment clears infection but does not prevent reinfection, ongoing inflammation, or risk of reproductive sequelae. Although IL-1α antibody was able to protect infected WT mice from oviduct disease [9], a potential extension of the current study would be to test specific inhibitors (anticytokine/receptor/drugs) of TNFα/IL-1α signaling as adjunctive therapies with antibiotics to mitigate upper reproductive tract tissue damage.

Notes

We thank Jesse Thomas and Allie Coor (Centers for Disease Control and Prevention [CDC]) for assistance during the CDC-Pelvic Inflammatory Disease meeting, Dr. Manoj Tripathy for laboratory assistance, and Bryan Mcqueen for editorial review.

Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Financial support. This study was funded by National Institutes of Health (NIH) (Grant AI067678), University of North Carolina (UNC) NCTracs Institute CTSA (Grant UL1TR002489), and UNC-Boost Award, all to UNC, and NIH (Grant HL130126 to S. N.). This study used UNC Translational Pathology Laboratory, which is supported in part by grants from the National Cancer Institute (5P30CA016086-42), NIH (U54-CA156733), National Institute of Environmental Health Sciences (5 P30 ES010126-17), University Cancer Research Fund, and Neurobehavioral Core Test Battery (2015-IDG-1007).

Supplement sponsorship. This supplement is sponsored by the Centers for Disease Control and Prevention.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

Presented in part: AAI 2019 Meeting, 9–13 May 2019, Austin, TX; Pelvic Inflammatory Disease, Centers for Disease Control and Prevention, 5–9 November 2019, Atlanta, GA.