Pathogenic NKT cells attenuate urogenital chlamydial clearance and enhance infertility

Abstract

Urogenital chlamydial infections continue to increase with over 127 million people affected annually, causing significant economic and public health pressures. While the role of traditional MHCI and II peptide presentation is well defined in chlamydial infections, the role of lipid antigens in immunity remains unclear. Natural killer (NK) T cells are important effector cells that recognize and respond to lipid antigens during infections. Chlamydial infection of antigen‐presenting cells facilitates presentation of lipid on the MHCI‐like protein, CD1d, which stimulates NKT cells to respond. During urogenital chlamydial infection, wild‐type (WT) female mice had significantly greater chlamydial burden than CD1d^−/− (NKT‐deficient) mice, and had significantly greater incidence and severity of immunopathology in both primary and secondary infections. WT mice had similar vaginal lymphocytic infiltrate, but 59% more oviduct occlusion compared to CD1d^−/− mice. Transcriptional array analysis of oviducts day 6 post‐infection revealed WT mice had elevated levels of Ifnγ (6‐fold), Tnfα (38‐fold), Il6 (2.5‐fold), Il1β (3‐fold) and Il17a (6‐fold) mRNA compared to CD1d^−/− mice. In infected females, oviduct tissues had an elevated infiltration of CD4⁺‐invariant NKT (iNKT) cells, however, iNKT‐deficient Jα18^−/− mice had no significant differences in hydrosalpinx severity or incidence compared to WT controls. Lipid mass spectrometry of surface‐cleaved CD1d in infected macrophages revealed an enhancement of presented lipids and cellular sequestration of sphingomyelin. Taken together, these data suggest an immunopathogenic role for non‐invariant NKT cells in urogenital chlamydial infections, facilitated by lipid presentation via CD1d via infected antigen‐presenting cells.

1. INTRODUCTION

Urogenital Chlamydia trachomatis infections are the world's leading sexually transmitted bacterial infection with an estimated 127 million new cases per annum. ¹ Despite widespread public awareness and education, infections continue to rise in both the developed and developing world, doubling in the United States over the last two decades with almost two‐thirds of all reported cases occurring among persons aged 15‐24 years ² and costing an estimated US$3 million per annum. ³ Chlamydial infection of the urogenital tract is also often asymptomatic in males and females (85%‐90%), allowing continued dissemination without antibiotic treatment. ⁴ Chlamydial infection initially infects the vagina and cervix causing localized inflammation, and in as many as 30% of cases, it can ascend to infect the uterine and fallopian tube epithelia of the upper female reproductive tract (FRT). Infection and inflammation of the upper FRT causing pelvic inflammatory disease (PID) can lead to scarring, occlusion and tubal infertility in 10%‐20% of cases. ⁵ As chlamydial infections continue to rise globally, it is widely agreed in the scientific community that there is an urgent requirement for a vaccine.

Chlamydia spp. are obligate intracellular bacteria with a dimorphic developmental cycle consisting of an infectious extracellular elementary body (EB) phase and an intracellular replicating reticulate body (RB) phase. Following infection, the EB rapidly dissociates from the host endocytic pathway preventing lysosomal degradation, and then traffics to the host Golgi apparatus via the microtubule network where it then replicates within a parasitophorous vesicle termed the inclusion. ⁶ As the chlamydial infection proceeds inclusion (Inc), membrane proteins facing the host cell cytoplasm and type III secretion system proteins interact with host proteins, metabolites and lipids hijacking host cell trafficking pathways, signal transduction and immune responses. ⁶ Within the inclusion, replicating Chlamydia not only is sequestered from the host cell cytoplasm and many innate defences but also parasitizes host cell glycerolphospholipids, sphingolipids, cholesterol and other metabolites. ⁶ Chlamydia spp. have a condensed genome (~1 Mbp) relative to other Gram‐negative bacteria (E. coli = 4.6 Mbp, Neisseria gonorrhoea > 2 Mbp), ⁷ and have evolutionarily lost the enzymatic genes necessary for lipid synthesis and are thus reliant on host‐derived lipids trafficked to the inclusion for replication and survival. ⁸ Chlamydia spp. generally infect epithelial cells lining the mucosa of the reproductive tract, but infection of immune cells such as macrophages, dendritic cells and neutrophils has been widely reported. ⁹ Infection of professional antigen‐presenting cells (APCs) facilitates cross‐presentation of peptides, metabolites and lipid antigens to lymphocytes.

Natural Killer T (NKT) cells are a group of innate‐like T lymphocytes that share properties of both NK cells and T cells. Like conventional T cells, NKT cells recognize self and foreign antigens through the αβT cell receptor (TCR) and upon activation can produce a wide variety of T helper (Th) 1, Th2, Th10 and Th17 cytokines. ¹⁰ , ¹¹ However, NKT cell development and function are independent of peptide antigen presented via major histocompatibility complex (MHC), but rather by lipid antigen presented via CD1d. ¹¹ CD1d is an MHC class I‐like glycoprotein that is stabilized by forming a heterodimer with β‐2‐microglobulin (β2m) and is presented together with endosomally processed lipids or glycolipids on the cell surface of APCs including macrophages and dendritic cells, but also by interferon‐γ (IFNγ)‐activated epithelial cells. ¹² CD1d‐restricted lipid presentation to the αβ TCR of NKT cells leads to rapid production and secretion of cytokines influencing localized tissue‐specific responses dependent on the phenotype of the activated NKT cell.

There are 2 main subsets of CD1d‐restricted NKT cells termed types I and II NKT cells. Semi‐invariant NKT (iNKT) cells are type I ‘classical’ NKT cells that have a conserved Vα24‐Jα18 TCR in humans and Vα14‐Jα18 in mice, which can be used to identify them due to their ability to bind and be activated by α‐galactosylceramide (αGalCer)‐loaded CD1d tetramers. ¹¹ Conversely, type II NKT cells express a different and more diverse αβ TCR repertoire (albeit less than conventional αβ TCRs) and are ‘non‐classical’ NKT cells. Type II NKT cells also fail to recognize αGalCer limiting their identification as a non‐iNKT (CD3⁺ NK1.1⁺ CD1dtet⁻) population instead recognizing other types of lipid antigens such as sulfatides. There is also a third smaller subset of ‘atypical’ NKT cells which appear to share some unique features like the ability to uniquely bind with αGalCer‐CD1d tetramers in mice and humans, and the ability to recognize the mycobacterial lipid glucuronosyldiacylglycerol. ¹³ , ¹⁴ As replicating chlamydiae parasitize host cell lipids and highjack host‐derived lipids, the role of lipid presentation by infected CD1d⁺ cells to NKT cells is unclear. We sought to determine if there was a role in urogenital chlamydial infections and pathogenesis.

Due to their potential to secrete a broad repertoire of cytokines, it is not surprising that NKT cells have been implicated in a variety of immune responses, both protective and pathogenic. Mucosal infection of mice with Streptococcus pneumoniae results in rapid production of IFNγ by NKT cells and blocking NKT activation in the context of CD1d greatly increasing bacterial and viral burden. ¹⁵ , ¹⁶ , ¹⁷ NKT cells have also been implicated in protection against cancers including melanoma and fibrosarcoma cells ¹⁸ , ¹⁹ but not osteosarcoma cells, ²⁰ regulating several different types of autoimmune diseases in mice. ²¹ , ²² Lipid presentation by APCs such as DCs, monocytes and macrophages to NKT cells can also enhance marginal zone B cell responses which express high levels of CD1d, and can present lipids to NKT cells, ²³ , ²⁴ , ²⁵ which can then modulate NK cell activity. ²⁶ , ²⁷ With regard to Chlamydia infections, NKT cells have been implicated in the induction of Th1 protective immunity against C. pneumoniae lung infection, via modulation of dendritic cell function, primarily via increased IL‐12 production by CD8α⁺ DCs. ²⁸ , ²⁹ Because of their rapid activation by both exogenous and some endogenous lipids, and subsequent secretion of multiple cytokines, NKT cells serve as a key cell lineage linking the arms of the innate and adaptive immune responses.

As NKT cells have been shown to influence the outcome of chlamydial lung infection, we sought to investigate how NKT cells affect the outcome of chlamydial urogenital infection and demonstrate an immunopathogenic role for non‐invariant NKT cells in chlamydial infections, mediated by CD1d‐restricted chlamydial lipid presentation by infected antigen‐presenting cells.

2. METHODS

2.1. Ethics statement

This study was approved by the Queensland University of Technology Animal Ethics Committee (QUT UAEC # 1500000029) and was carried out in accordance with any recommendations. Mice were euthanized by intraperitoneal injection with sodium pentobarbital (200 mg kg⁻¹).

2.2. Animals

Mice (BALB/c and C57BL/6) were sourced from the Animal Resource Centre (ARC; Canningvale, WA, Australia) at 6 weeks of age and given food and water ad libitum. Animal experiments were performed at QUT Medical Engineering Research Facility (MERF). Female BALB/c CD1d^−/− mice were provided by Professor Mark Smyth (Queensland Institute of Medical Research Berghofer [QIMR‐B], Brisbane, QLD, Australia) and were bred at the QUT MERF. Female C57BL/6 Jα18^−/− mice were generously provided by Dr Stephen Mattarollo (University of Queensland, QLD, Australia). ³⁰ Ten animals were used in each group, as determined to be statistically powerful a priori unless stated otherwise, in 1 experiment.

2.3. Cell and Chlamydia culture

Chlamydia muridarum (Weiss strain) a generous gift from Dr Catherine O'Connell (University of North Carolina – Chapel Hill, NC, USA) was cultured from McCoy cells and purified as previously described. ³¹ Mouse McCoy fibroblasts (Cat # CRL‐1696) and mouse RAW 264.7 macrophages (Cat # TIB‐71) were purchased from the American Type Cell Culture (ATCC; Manassas, VA, USA). Cells were cultured in RPMI 1640 supplemented with 10% v/v heat‐inactivated fetal calf serum, 100 μg ml⁻¹ streptomycin sulphate, 50 μg ml⁻¹ gentamycin and 1× Glutamax (Thermo Fisher Scientific, Brisbane, QLD, AUS).

2.4. Intravaginal infections

Female mice (6‐8 weeks) were hormonally synchronized with 2.5 mg depot medroxyprogesterone (Depo Provera, Pfizer, New York, USA) subcutaneously 1 week prior to intravaginal infection with 5 × 10⁴ inclusion forming units (IFUs) of C. muridarum in 10 μl volume.

2.5. Vaginal shedding and gross oviduct pathology

Chlamydial vaginal shedding was determined by swab culture every 3rd day, and the oviduct diameter was measured on day 35 post‐infection to assess the incidence and severity of pathology as previously described. ³² The incidence of hydrosalpinx was the total number of occluded oviducts.

2.6. Th17 response RT2‐gene array and qRT‐PCR

Following 6 days of infection, pooled oviducts (n = 6) from mock‐infected or infected BALB/c female mice were collected and RNA was extracted using TRIzol (Thermo Fisher) and 10 μg of glycogen carrier protein (Cat No. AM9510, Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. Complementary DNA (cDNA) was synthesized using High‐Capacity Reverse Transcriptase Kit (Applied Biosystems) (Cat. No. 4368814) as per the manufacturer's instructions. Mouse Th17 Response RT² Profiler PCR Array (PAMM‐073Z) (Qiagen, Clayton, VIC, AUS) was performed as per the manufacturer's instructions on an ABI 7900 HT plate PCR (Applied Biosystems).

2.7. Flow cytometry

At indicated time points, post‐infection animals were euthanized and reproductive tract organs were collected. Each region (uterine horns and oviducts) was dissected and placed into separate tubes for tissue digestion. Tissue digestion media consisted of RPMI media containing 5% v/v fetal bovine serum (FBS), 50 μg ml⁻¹ gentamicin, 4 mmol L⁻¹ L‐glutamine, 500 U ml⁻¹ collagenase A and 0.057 KU μl⁻¹ DNase I (Sigma‐Aldrich, North Ryde, Australia). Tissues were cut into 1 mm² pieces and placed in 15 ml tubes containing digestion media. Tubes were incubated at 37°C in a shaking incubator at 200 rpm for 1.5 hours to allow cells to dissociate. Cells were isolated into single‐cell suspensions as previously described. ³³

After single cells had been isolated, each sample was halved to allow for antibody staining panels to be performed and aliquoted into 96‐well plates at a density of 10⁶ cells/well. Cells were blocked in 5% v/v FBS in PBS for 20 minutes and 5% v/v normal human serum to block FcγR for 20 minutes. Cells were washed twice in PBS via centrifugation to remove traces of serum. Cells were labelled with Alexa Fluor 700 Fixable Viability Stain (1/1000; BD Biosciences, San Jose, USA) in PBS for 7 minutes at 37°C, then quenched with 5% v/v FBS in PBS. Each sample was stained with CD3e‐APC (1/100; Clone 145‐2C11, BD Biosciences), CD4‐V450 (1/1000; Clone RM4.5, eBiosciences, San Diego, USA), CD8‐PECy7 (1/2000; Clone 53–6.7, eBiosciences) and CD1d‐tetramer αGal‐Cer‐PE (1/400, kindly provided by Professor Dale Godfrey, University of Melbourne, which is loaded with PBS‐44 α‐GalCer glycolipid, kindly provided by Professor Paul Savage, Utah, USA). Cells were incubated with antibodies for 20 minutes in the dark, washed and fixed in 4% v/v paraformaldehyde in PBS to preserve the RNA. Samples were stored at 4°C in the dark until flow cytometry. Samples were analysed and sorted using a FACSAria III flow cytometer (BD Biosciences) with cells being collected straight into 300 μl of RLT buffer (Qiagen). Cells were vortexed and then stored at −80°C until processing. Flow cytometry gating strategy is displayed in Figure S3.

2.8. RNA extraction and gene expression

Because the oviduct is the major site of inflammatory pathology leading to oviduct occlusion and infertility, oviduct cells in RLT buffer were defrosted, vortexed and had 590 μl Ultrapure H₂O (Life Technologies, Mulgrave, VIC, AUS) and 10 μl Proteinase K (200 μg ml⁻¹ final concentration) (Roche, Millers Point, NSW, AUS) added. The samples were then vortexed, incubated at 56°C for 15 minutes, mixing by inversion every 3 minutes, followed by incubation at 85°C for 15 minutes and mixing again by inversion every 3 minutes. Samples were then cooled on ice for 3 minutes. RNA was extracted using RNA extraction kits, including the genomic DNA removal step according to the manufacturer's instructions (Analytik Jena, Jena, Germany). Samples were eluted into 25 μl of ultrapure H₂O and quantified using the Qubit RNA HS Assay Kit (Qubit 3.0, Life Technologies) and stored at −80°C until use. Reverse transcription was performed on 103.1 pg of RNA for each sample using iScript reverse transcription Supermix for RT‐qPCR (Bio‐Rad, Gladesville, Australia) according to manufacturer's instructions to generate cDNA.

To quantify the level of expression of Il17a, Il17f, Il17ra and Il17rc mRNA, the Bio‐Rad digital droplet automated droplet generator (AutoDG, Bio‐Rad) PCR system was utilized. All procedures were performed in strict accordance with the manufacturer's instructions using the 4‐plex probe protocol. Briefly, a master mix was made (10 μl 2× ddPCR Supermix for Probes [No dUTP], 1 or 0.5 μl 20× target primer/probe [Life Technologies, Waltham, USA; Table 1] and 5 μl ultrapure H₂O per sample) and 15 μl was aliquoted into each well of 96‐well Eppendorf plates. Each sample had 2 μl of cDNA added to the appropriate well, with no template controls included in each plate run. Plates were sealed, centrifuged for 2 minutes at 500 g and placed in the Bio‐Rad AutoDG for droplet generation using Auto DG Oil for Probes (Bio‐Rad). Droplet samples then underwent PCR using the following cycling conditions: (i) 95°C for 10 minutes, (ii) 94°C 30 seconds, (iii) 60°C 1 minute, × 50 cycles of steps (ii‐iii), (iv) 98°C 10 minutes. Samples were analysed using the Bio‐Rad QX200 Droplet Reader and QuantaSoft™ Analysis Pro (version 1.0.596) (Bio‐Rad).

TABLE 1.

TaqMan® gene expression assay primer/probes.

Gene	Assay ID	Concentration (nmol L−1)	Reporter probe
Il17a	Mm00439618_m1	125	VIC
Il17f	Mm00521423_m1	125	FAM
Il17ra	Mm00434214_m1	250	FAM
Il17rc	Mm00506606_m1	250	VIC

2.9. Immunohistochemistry (IHC)

Paraffin‐embedded reproductive tract tissues were sectioned into 5 μm sections, dewaxed and rehydrated into PBS. Haematoxylin and eosin were performed as previously described. ³³ For IHC, endogenous peroxide was blocked using 0.5% v/v hydrogen peroxide for 15 minutes. Sections were then blocked with 5% v/v FCS in PBS for 1 hour. Sections were washed in PBS, and incubated with primary antibody for 1 hour. Rat anti‐mouse CD1d‐biotinylated IgG (10 μg ml⁻¹) (BioLegend), rat anti‐mouse F4/80 IgG (10 μg ml⁻¹) (clone BM8; Abcam, Melbourne, VIC, AUS), hamster anti‐mouse CD11c IgG (10 μg ml⁻¹) (clone N418; eBioscience), rabbit anti‐mouse CD3 IgG (2 μg ml⁻¹) (Dako, Mulgrave, VIC, AUS), rat anti‐mouse CD4 (2 μg ml⁻¹) (clone: A15.1.17. BD Pharmingen) and rat anti‐mouse CD8 IgG (clone 53–6.7) (2 μg ml⁻¹) (BD Pharmingen) were used as indicated. Slides were washed 3 times with PBS. Primary antibodies were detected with biotinylated mouse anti‐hamster IgG (1:1000) (Jackson Immunresearch, Waterford, QLD, AUS), biotinylated donkey anti‐rat IgG (1:1000) (Jackson Immunresearch) and biotinylated donkey anti‐mouse IgG (1:1000) (Jackson Immunresearch) as appropriate to the primary antibody used. Slides were probed for 1 hour with streptavidin‐HRP (1:1000) (Southern Biotechnology, Melbourne, VIC AUS). Slides were washed 3 times with PBS and incubated with DAB for 5 minutes. Slides were then counterstained with haematoxylin or methyl green for 1 minute, washed in PBS, dehydrated in ethanol and coverslips mounted with DPX (Sigma Aldrich). Slides were imaged using the Leica Aperio AT Turbo slide scanner and visualized and positive staining was quantified in Aperio ImageScope (v12.4.3).

2.10. Transfection of pCD1d

The full‐length mRNA sequence for mouse (Mus musculus) CD1d mRNA sequence (NM_007639.3) was obtained from the NCBI GenBank Nucleotide database. The domains including the signal peptides (1‐21), extracellular (22‐305 bp), transmembrane (306‐326 bp) and cytoplasmic tail (332‐335 bp) domains were identified, and a sequence containing a hexahistidine (6xHis) sequence followed by a Tobacco Etch Virus (TEV) protease site between the extracellular and transmembrane domain of CD1d. A BamHI and EcoRI restriction enzyme site was included at the N′ and C′ of the full‐length CD1d gene respectively (Figure S4). The modified pCD1d was synthesized and inserted into the pCDNA3.1 expression vector by GeneArt (Thermo Fisher). Vectored pCD1d was transfected into RAW264.7 cells using Lipofectamine 2000 (Thermo Fisher) as per the manufacturer's instructions. Three days after transfection, cells containing the plasmid were selected by incubating in growth media supplemented with 500 μg ml⁻¹ of G418 antibiotic (Thermo Fisher). Media containing G418 were replaced every 3 days for 3 weeks, then cells were maintained in G418‐free growth media and screened for surface CD1d expression by flow cytometry using rat anti‐CD1d‐IgG‐biotin (1:500) (Cat: 5538421, BD Biosciences) and detected using V450‐conjugated streptavidin (1:2000) (BD Biosciences). Positive transfectants were frozen in liquid nitrogen until required.

2.11. Lipid mass spectrometry of infected macrophages

pCD1d‐transfected RAW cells were grown to confluence in T‐75 flasks and were then incubated in media (control) or infected with 10⁷ IFUs of C. muridarum (multiplicity of infection = 1) for 24 hour at 37°C + 5% CO₂. Adherent cells were lifted using a cell scraper and pelleted by centrifugation at 300 g for 5 minutes. Cells were resuspended in PBS and pelleted by centrifugation at 300 g for 5 minutes. The wash step was repeated 2 more times and then cells were resuspended in 1 ml of PBS containing sodium azide (0.5%) and 100 μg of TEV Protease (Sigma Aldrich) and transferred to a 1.7‐ml microfuge tube. Cells were incubated overnight on a rotating wheel at 4°C. Cells were then pelleted by centrifugation at 300 g for 5 minutes. Supernatant was collected and clarified by centrifugation at 12 000 g for 15 minutes at 4°C. Clarified supernatants (1 ml) and cell lysates (resuspended in 1 ml PBS) were transferred into sterile glass vials. Lipids were extracted as previously described. ³⁴ Briefly, 1.15 ml of ice‐cold MeOH was added to 1 ml samples and vortexed for 3 minutes. Following mixing, 3.85 ml of methyl‐tert‐butyl ether (MTBE) (MTBE: MeOH 10:3) was added and briefly vortexed. Samples were then incubated with rotation for 1 hour at 4°C. Following mixing, 0.95 ml of 0.15 mol L⁻¹ ammonium acetate (aqueous) was added. Samples were then centrifuged at 2000 g for 5 minutes at 4°C. The supernatant was collected. The aqueous layer was then washed with 1.5 ml of MTBE:MeOH:NH₄OAc (20:6:5). Samples were dried under a stream of liquid N₂ and resuspended in MeOH:CHCl₃(2:1) in 0.01% butylated hydroxytoluene (BHT) and 5 mmol L⁻¹ NH₄OAc for mass spectrometry analysis.

The lipid extracts were analysed using an automated lipidomics profile described previously. ³⁵ Briefly, an autosampler (Shimadzu LC‐20A, West Eagle Farm, QLD, AUS) was used to deliver 100 μl of sample via loop injection into a 20 μl min⁻¹ flow of MeOH containing 5 mmol L⁻¹ NH₄OAc. The autosampler was directly connected via a PEEKsil restriction (500 mm length, 50 μm I.D) to the electrospray ionization source of a QTRAP 6500 hybrid triple quadrupole mass spectrometer (Sciex, Concord, ON, Canada) operating in positive‐ion mode. Spray voltage was set to +5 kV, and both source gasses (GS1 and GS2) were set to 15 (arb. Units). Lipid classes were profiled using either precursor ion scans or neutral loss scans, with the detected m/z value being indicative of the lipid sum composition.

The resulting data were processed using LipidView (version 1.3 beta; Sciex, Concord, ON, Canada). Data tables were exported from LipidView as comma‐separated variable files and processed using Python, Pandas and Microsoft Excel. For each sample, lipid abundances were normalized to total summed abundance of the phosphatidylcholine lipid class for qualitative comparison between samples.

2.12. Power calculations and statistical analysis

Power calculations were performed as previously described. ³² Data were tested for normality and then appropriate parametric tests were applied. Statistical analysis of graphical data was performed using unpaired t‐tests (multiple comparisons for kinetics and hydrosalpinx severity) and Chi‐squared (for hydrosalpinx incidence) tests on GraphPad Prism version 7 (GraphPad, La Jolla, CA, USA) as indicated. Error bars represent the standard error of mean.

3. RESULTS

3.1. NKT cells delay chlamydial clearance and enhance immunopathology

To determine the expression of CD1d in the female reproductive tract, the whole female reproductive tract was removed from progesterone‐primed mice and stained for CD1d expression. CD1d⁺ cells were found to be localized to the lamina propria of the cervicovagina, uterine horns and oviducts with the highest expression in the cervicovagina (Figure 1A). In the ovary, small numbers of CD1d⁺ cells were observed surrounding the thecal layer of developing follicles and in larger numbers within the antrum of atretic follicles. To determine the presence of macrophages and dendritic cells in the reproductive tract, FRT tissues were also stained for F4/80 and CD11c expression respectively. Macrophages (Figure 1B) and dendritic cells (Figure 1C) were found in the lamina propria of the cervicovagina, uterine horns and oviducts, indicating APCs are present throughout the FRT prior to chlamydial challenge.

FIGURE 1.

NKT cells attenuate vaginal chlamydial clearance and exacerbate the severity and incidence of infertility. Female BALB/c mice were vaginally infected with Chlamydia muridarum and shedding and hydrosalpinx were determined following primary and secondary infection. (A) CD1d, (B) F4/80 and (C) CD11c expression in progesterone‐primed BALB/c FRT tissues prior to challenge (day 0) was determined by IHC. Vaginal shedding kinetics (D) and total burden (E) were determined by swab culture (n = 8 or 10 mice/group, in technical duplicate). The severity and incidence of hydrosalpinx formation were determined following 5 wk of primary (F) (n = 20 oviducts from 10 mice/group), or 5 wk of secondary (G) (n = 18 or 19 oviducts from 10 mice/group) C. muridarum infection. Scale = 200 μm. Data are mean ± SEM. Unpaired t‐tests were performed on clearance and area under the curve, and chi‐squared for hydrosalpinx data, *P < .05, in GraphPad Prism (v7).

Vaginal chlamydial infection leads to an ascending infection that causes scarring and occlusion of the oviducts. Vaginal infection of WT and CD1d^−/− mice led to a peak of infection on day 6, which was significantly greater (P < .05) in WT mice (Figure 1D). Secondary infection of both WT and CD1d^−/− mice led to rapid chlamydia clearance, indicating a limited role for NKT cells in recall immunity generated from primary infections. When the total burden of the primary infection was quantified (area under the curve), there was a significantly greater (more than double; P = .006) burden in WT mice compared to CD1d‐deficient mice, indicating NKT cells play a detrimental role in chlamydial clearance (Figure 1E). When the severity and incidence of hydrosalpinx were quantified, there was a significant enhancement of oviduct occlusion (infertility) in WT mice in both primary (Figure 1F) and secondary (Figure 1G) infections, indicating CD1d presentation of lipid antigen/s to NKT cells significantly contributes to immunopathology during chlamydial infection of the FRT.

To determine how CD1d expression significantly alters how the immune system responds to infection and enhances immunopathology, immunohistochemistry of FRT tissues from WT and CD1d^−/− mice was performed. H&E staining of FRT tissues following primary (d35) and secondary (d80) chlamydial infection (Figure 2A) demonstrated that WT mice had significantly more leukocyte infiltrate and inflammation in the cervicovaginal tissues shortly following primary infection compared to CD1d^−/− mice (Figure 2B; P < .05). At 80 days, when the infection had cleared, the difference in leukocyte numbers between WT and CD1d^−/− tissues was not significant, although there was still a trend towards greater numbers in the WT mice (Figure 2B). This reduced leukocyte infiltration following primary infection in CD1d^−/− mice is consistent with the reduction in hydrosalpinx in these mice. Both WT and CD1d^−/− mice had minimal inflammation or infiltration following secondary infection, with oviducts remaining occluded despite the egress of inflammatory cells.

FIGURE 2.

Uterine T cell infiltrate in the presence and absence of CD1d. Female BALB/c mice were vaginally infected with Chlamydia muridarum for 5 wk. Uterine tissues were fixed and stained by (A) H&E and infiltrating leukocyte numbers were counted (B, Aperio Image Scope). Tissues were also stained by IHC for expression of (C) CD3ε, (D) CD4; (E) CD8α and quantified in (F). Mock‐infected uterine tissues and spleens were used as controls. N = 5 mice per group, from 1 experiment. Scale = 200 μm. Data are mean ± SD. Unpaired t‐tests were performed on cell counts. Graphs produced in GraphPad Prism (v7), *P < .05.

T cells play a crucial role in adaptive immunity with CD4⁺ T cells being necessary to resolve chlamydial infection. ³⁶ , ³⁷ Staining of lower reproductive tract tissues from chlamydial‐infected mice demonstrated comparable levels of T cell (CD3⁺) infiltration in the cervicovaginal lamina propria of WT and CD1d^−/− mice, despite WT mice having more leukocyte infiltration (Figure 2C). T cells were commonly observed in aggregates, in close proximity to the basolateral side of the simple mucosal epithelia lining the cervicovagina. When the cervicovaginal tissue was stained more specifically for CD4⁺ and CD8⁺ expressions, both WT and CD1d^−/− mice were elevated over uninfected controls, (Figure 2D,E) but no difference was found between the infected groups (Figure 2F). Conversely, within the upper reproductive tract (ovaries), IHC staining for T cells in the ovaries demonstrated enhanced T cell infiltration in WT mice compared to CD1d^−/− mice (Figure 3A) for CD4⁺ cells, and significantly for CD3⁺ and CD8⁺ cells (Figure 3B). There was no significant difference in PMN numbers between WT and CD1d^−/− mice (Figure 3B). However, systemic and local production of chlamydial antigen‐specific IgA and IgG, or Th1 bias (IgG_2a > IgG₁) was not significantly affected by the presence of CD1d expression although CD1d^−/− mice had quantitatively more IgG_2a and IgG₁, suggesting NKT cells play a minimal role in B cell differentiation and antibody secretion during chlamydial infection of the FRT and may warrant further investigation in future studies (Figure S1).

FIGURE 3.

Ovarian T cell infiltrate is minimized in CD1d‐deficent mice. Female BALB/c mice were vaginally infected with Chlamydia muridarum for 5 wk. Ovarian tissues were (A), fixed and stained by H&E for quantitation of PMNs or IHC for expression of CD3ε, CD4 and CD8α, then (B), those cell markers were quantified. Scale = 200 μm. N = 3–5 mice per group, from 1 experiment, graphs show mean ± SD and were produced in GraphPad Prism (v7), *P < .05 from unpaired t‐tests.

Taken together, these data demonstrate that CD1d is expressed by APCs in the lamina propria of the lower and upper FRT, and suggest that CD1d (likely CD1d‐loaded lipid antigen presentation to NKT cells) leads to attenuated chlamydial clearance of primary infections, that NKT cells play a negligible role in immunity to secondary infection and importantly, that CD1d presence potentially enhances the severity and incidence of oviduct occlusion and infertility. Furthermore, CD1d expression during chlamydial infection leads to enhanced T cell infiltration in the ovaries.

3.2. CD1d expression enhances pro‐inflammatory transcription during early chlamydial infection

Prior to the T cell infiltration and dominance, chlamydial infection of the mouse FRT peaks at day 6 and declines thereafter. ³² It is also the point of no return for the onset of pathology, as azithromycin treatment after day 6 is able to clear infection but does not prevent hydrosalpinx formation. ³⁸ Thus, cytokine and chemokine signalling in the oviducts early during infection is likely innate or innate‐like cell mediated. An innate source of IL17A has also been implicated in hydrosalpinx development, ³⁹ thus to determine the role of NKT cells in influencing pro‐inflammatory signalling, a cytokine gene array was performed on oviducts of WT and CD1d^−/− mice infected with Chlamydia for 6 days (Figure 4).

FIGURE 4.

Pro‐inflammatory signalling is dampened in CD1d‐deficent mice. Female BALB/c mice (n = 5/group, in 1 experiment) were vaginally infected with Chlamydia muridarum for 6 d and oviducts were collected and pooled. RNA was extracted and qRT‐PCR was performed in duplicate using a Mouse Th17 response RT2‐PCR array. Results were normalized to uninfected oviducts. Fold change expression of inflammatory cytokines (A, B), chemokines (C), surface markers (D) and signalling pathways (E) was determined and represented as a heatmap.

The transcription of pro‐inflammatory cytokines Ifnγ (1054‐fold), Il1β (130‐fold), Il6 (261‐fold), Il12b (97‐fold) and Tnf (534‐fold) was highly elevated in infected WT mice oviducts compared to CD1d^−/− mice oviducts (Ifnγ 177‐fold, Il1β 42‐fold, Il6 105‐fold, Il12b‐10‐fold and Tnf 14‐fold) (Figure 4A). In all experiments, the fold increase was compared to non‐infected oviducts of the same strain. There was no significant difference in transcriptomes of non‐infected WT and CD1d^−/− mice (data not shown). Moderate upregulation of Il10 (20‐fold), Il17a (6‐fold), Il17c (4‐fold), Il22 (13‐fold) and Il23a (2‐fold) was observed in WT mice, with only less upregulation of Il10 (10‐fold), Il17a (no change), Il17c (2‐fold) and Il22 (no change) and downregulation of Il23a (8‐fold) in CD1d^−/− mice (Figure 4B). Upregulation of Il27 was similar between WT (61‐fold) and CD1d^−/− (53‐fold) mice. Upregulation of chemokines Ccl2 (200‐fold vs 83‐fold), Ccl7 (108‐fold vs 66‐fold), Cxcl2 (140‐fold vs 52‐fold) and Cxcl5 (198‐fold vs 118‐fold) was observed between WT and CD1d^−/− mice respectively (Figure 4C). Increased T cell activation markers Cd40lg (11‐fold vs 6‐fold) and Icos (75‐fold vs 5‐fold) were also observed between WT and CD1d^−/− mice respectively (Figure 4D). When determining the transcriptional markers, similar levels of expression were observed for most genes, however, WT mice expressed more Socs 1 (13‐fold vs 6‐fold), Socs 3 (9‐fold vs 5‐fold) and Tbx21 (Tbet) (23‐fold vs 6‐fold) than CD1d^−/− mice, indicative of a Th1 phenotype (Figure 4E). As a control for pathology, BALB/c female mice were also infected with the infectious but non‐pathology‐causing plasmid‐cured strain of C. muridarum, ⁴⁰ and showed minimal upregulation of pro‐inflammatory genes in either WT or CD1d^−/− mice when contrasted with plasmid‐containing C. muridarum (Figure S2A–E). Taken together, WT mice had upregulated more pro‐inflammatory cytokines (Ifnγ, Il1β, Il6, Tnf, Il17a, Il17c, Il22 and Il23a), inflammatory chemokines (Ccl2, Ccl7, Cxcl2 and Cxcl5), T cell activation markers (Cd40lg and Icos) and Th1 phenotype transcriptional markers (Tbet/Tbx21) when compared to CD1d^−/− mice, suggesting NKT cells play a pivotal role in the enhancement of immunopathology during chlamydial infections.

3.3. Invariant NKT cells play no significant role in immunopathology during chlamydial infection

Next, to determine the subclass of NKT cells orchestrating chlamydial immunopathology, we examined the role of the most widely studied CD1d‐restricted NKT cells, which express an invariant T cell receptor α chain with a restricted repertoire and can respond rapidly during a pro‐inflammatory event. ¹¹ The iNKT conserved TCR is made of a Vα14‐Jα18 α chain paired with a limited set of β chains (Vβ8.2, Vβ7 and Vβ2). ¹⁰ Mice deficient in the Jα18 gene have a significant loss in iNKT cells as the Vα14‐Jα18 TCR cannot be generated. These cells have been shown to be important in clearance of respiratory chlamydial infection and protection from pathology of the respiratory tract. ²⁸ We therefore vaginally infected WT‐ and iNKT (Jα18^−/−)‐deficient mice with C. muridarum. Over the course of a vaginal infection, there was a significantly lower chlamydial burden on day 3 in Jα18^−/− mice, with no further differences throughout the duration of infection (Figure 5A). There was no difference in total burden (Figure 5B), demonstrating iNKT cells play a minor role in chlamydial clearance from the FRT. Interestingly, neither hydrosalpinx severity (Figure 5C) nor incidence (Figure 5D) significantly differed between WT and Jα18^−/− mice. These data show that iNKT cells play no essential role in the development of oviduct occlusion and infertility.

FIGURE 5.

iNKT cells do not contribute to immunopathology. Female C57BL/6 WT or Jα‐18^−/− mice were vaginally infected with Chlamydia muridarum for 35 d. Vaginal shedding kinetics (A, via swab in duplicate), total burden (B), hydrosalpinx severity (C, dashed line represents normal diameter) and incidence (D) were determined (n = 8 or 10 mice/group, in 1 experiment). Data are mean ± SEM. Unpaired t‐tests were performed on clearance and area under the curve, and chi‐squared for hydrosalpinx data, *P < .05, in GraphPad Prism (v7).

In WT female mice, chlamydial infection did cause a small increase in iNKT cells in the uterus and oviducts, although these cells constituted <5% of recruited CD3⁺ T cells (data not shown). iNKT cells isolated from the FRT of infected mice on day 6 were enriched by sorting CD4⁺ T cells labelled with fluorescent CD1d‐αGalCer‐tetramers, and were found to have enhanced expression of Il17a and the Il17 receptors Il17ra and Il17rc mRNA (Figure S4). Thus, although iNKT cells appear to play no role in the duration and pathology associated with genital chlamydial infection, they do express IL17A and IL17 receptors, an inflammatory cytokine signalling pathway that has previously been associated with pathology induced by C. muridarum infection. ³⁹ , ⁴¹

3.4. CD1d presentation of glycolipids is enhanced in Chlamydia‐infected APCs

As NKT cells recognize CD1d‐mediated host and foreign glycolipid antigen presentation, we next sought to explore the spectrum of lipid presentation of chlamydial‐infected immune cells. As the cell membrane is mainly composed of lipids to differentiate CD1d‐loaded lipids from contaminating cell membrane lipids, a chimeric CD1d protein with a cleavable CD1d domain was developed. To isolate surface presented CD1d alone (as opposed to intracellular CD1d‐loaded lipid that is not able to present to T cells), a hexahistidine tag and TEV protease site were included between the extracellular and transmembrane domains of CD1d, allowing only cell surface presented CD1d to be isolated, and then specific cleavage of the extracellular CD1d domain (Figure S5). The mouse macrophage cell line RAW264.7 was transfected with a plasmid containing the chimeric CD1d, and the surface expression of CD1d was determined by flow cytometry (Figure S6). LPS‐activated RAW cells endogenously expressed CD1d, but in RAW cells transfected with pCD1d, surface expression of CD1d was enhanced by 8.8%. To determine the lipid profile of CD1d presentation in Chlamydia‐infected macrophages, pCD1d‐transfected RAW cells were infected with C. muridarum for 24 hours, followed by cleavage of CD1d by TEV protease, lipid purification of supernatants and analysis by lipid mass spectrometry.

Lipid mass spectrometry of supernatants and cell lysates from mock and Chlamydia‐infected macrophages showed lipid profiles across major membrane classes – phosphatidylethanolamine, phosphatidylglycerol, phosphatidylcholine, phosphatidylserine (PE, PG, PC and PS), sphingomyelins (SM) and cholesteryl ester (CE). For each sample, lipid abundances were normalized to total summed abundance of the phosphatidylcholine lipid class (PC) for qualitative comparison between samples (Figure S7). Figure 6 shows the above lipid classes in supernatants (Figure 6A) and CD1d‐cleaved whole‐cell lysates with each lipid class normalized to total PC. As shown in Figure 6A, in supernatant samples lipid classes PG, PE and CE showed significant differences in relative intensity of infected and uninfected samples. The CD1d‐cleaved cell lysates (Figure 6B) also show all same classes as seen in the supernatant samples and show a significant difference in PE (increased) and SM (sequestered) in infected and uninfected cells. An increase in the levels of PE is observed across both the infected samples.

FIGURE 6.

CD1d‐presented lipids during chlamydial infection. Surface CD1d cleaved supernatants (A) and CD1d cleaved whole‐cell lysates (B) of uninfected and Cmu‐infected macrophages were analysed for lipid content by lipid mass spectrometry. For each sample, lipid abundances were normalized to total summed abundance of the phosphatidylcholine (PC) lipid class for qualitative comparison between samples. Lipid classes include cholesteryl ester (CE), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylserine (PS) and sphingomyelins (SM). Graphs and statistical analysis (multiple comparisons unpaired t‐tests) are generated in GraphPad Prism (v7), *P < .05, ***P < .001, data are mean ± SD, N = 3 replicates, from 1 experiment.

4. DISCUSSION

Vaginal infection of common mouse strains with C. muridarum, following prior administration of progesterone to induce dioestrus, elicits an infection that usually resolves in 30 days. However, these infections induce pathology in most strains of mice. This is characterized by the formation of hydrosalpinx, which occludes the oviduct, resulting in infertility. Primary infection protects against a secondary infection challenge, but hydrosalpinx persists and may be increased by repeat infections. CD1d⁺ cells were found scattered throughout the murine reproductive tract (Figure 1A), with numbers highest in the cervicovagina, the site of initial infection. CD1d knockout mice demonstrated a significantly reduced infectious burden (Figure 1E), incidence of hydrosalpinx (Figure 1F) and inflammation, as determined by H&E staining (Figure 2A,B), compared to WT mice following primary Cmu infection. Differences in infiltrating leukocyte numbers were significant with 10 fields of view, even though H&E staining was only performed on 5 animals per group at day 35. The infectious burden was equivalent after secondary infection of both WT and CD1d^−/− mice. CD3⁺ T cell numbers were equivalent in uterine horns of both WT and CD1d^−/− mice (Figure 2C), although CD3⁺ T cell numbers in the WT ovary were higher than in the CD1d^−/− ovary (Figure 3). In infected WT mice reproductive tracts, transcripts encoding multiple inflammatory cytokines (Ifnγ, Il1β, Il6, Tnf, Il17a, Il17c, Il22 and Il23a), chemokines (Ccl2, Ccl7, Cxcl2 and Cxcl5) and T cell activation/Th1 (Cd40l, Icos, Socs1/3 and Tbet) markers were all increased following infection, compared to CD1d^−/− mice (Figure 4). While these experiments required pooling of tissues due to size and amounts of mRNA generated and should be viewed with caution, pooling tissues minimize any effects of outliers, and our results showed a clear enhancement of differences in gene expression observed between WT and CD1d^−/− mice. Future studies should involve PCR validation of key analytes and/or single‐cell RNA sequencing (scRNA‐Seq) of FACS‐purified NKT cell subsets to confirm our findings. In Jα18^−/− mice, however, there was no significant difference in burden or hydrosalpinx (Figure 5), suggesting that invariant NKT cells (iNKT, type 1) were not responsible for the observed effects in CD1d^−/− mice, however, Jα18^−/− are not only deficient in iNKT cells but also have a reduced TCR repertoire, ⁴² so results must be interpreted with caution, and knockout mice, while inbred and widely used, are on different backgrounds. Despite there being no difference in burden or hydrosalpinx between WT and Jα18^−/− mice, FACS‐purified iNKT cells from Chlamydia‐infected mice did express increased mRNA for Il17a and Il17ra and Il17rc. Thus, the role of iNKT in pathology requires further investigation as IL17 signalling plays a key role in chlamydial pathology. ³² , ³⁹ , ⁴¹

Understanding the role of NKT cells in animal models of chlamydial infection is complicated by the use of different infection routes (genital, pulmonary and intra‐articular), different species of Chlamydia (C. trachomatis (Ctr), C. pneumoniae (Cpn) and C. muridarum (Cmu)) and different mouse strains (C57BL/6 and BALB/c). For example, NKT‐deficient mice showed exacerbated susceptibility to Cpn lung infection, however, both CD1d^−/− and Jα18^−/− were more resistant to Cmu lung infection, suggesting a detrimental role of both type I iNKT and type 2 NKT cells in Cmu lung infections. ²⁸ Pretreatment of BALB/c mice with α‐GalCer, prior to genital Cmu infection, led to reduced bacterial burden and pathological changes, suggesting a protective role for iNKT cells. ⁴³ Activation of iNKT cells by α‐GalCer injection also enhanced protection against intranasal Cpn, intra‐articular Ctr and intravaginal Cmu infection. ²⁸ , ⁴³ , ⁴⁴ In a lung infection model, using WT and Jα18^−/− mice on a C57BL/6 background, iNKT cells promoted protective immunity through their actions on lung dendritic cells (LDC), particularly CD8α⁺ DCs. ²⁹ , ⁴⁵ iNKT cells from Cpn‐infected WT mice induced expansion of LDC, enhanced cytokine and co‐stimulatory molecule expression by LDC that promoted the development of protective Th1/Tc1 T cell responses. ²⁹ LDC from Cpn‐infected Jα18^−/− mice elicited non‐protective Th2/Tc2 T cell responses. Interestingly, while a‐GalCer is the archetypal antigen used to stimulate iNKT cells GLXA, a glycolipid exoantigen ⁴⁶ from Cmu infection stimulates cytokine production in WT mouse sera, but not serum of Jα18^−/− mice ⁴⁷ providing evidence of chlamydial glycolipids that can stimulate iNKT cells. NKT cells may also influence the outcomes of chlamydial infections through their modulating effects on NK cell functions. Following Cmu infection of Jα18^−/− mice, there was a reduced expansion of IFNγ‐secreting NK cells but an increase in CD107a‐degranulating NK cells, ²⁶ suggesting an interaction between these 2 unconventional T cell populations. ²⁷ Overall, the findings that both CD1d^−/− and Jα18^−/− mice showed increased resistance to Cmu lung infection, but increased susceptibility to Cpn lung infection, suggest that the protective versus pathogenic roles may be determined by the chlamydial pathogen used to infect mice. ²⁹ NKT cells, both type 1 and 2, can be rapidly activated to secrete various combinations of cytokines similar to Th1, Th2 and Th17 CD4⁺ T cells, which enable these cells to regulate the activation of other cell types including DCs, neutrophils, NK cells and B cells by either cytokine‐mediated (indirect) or cell‐cell contact dependent (direct) mechanisms, ⁴⁸ ultimately regulating infection outcomes. As a chlamydial glycolipid antigen (GLXA) has already been isolated from Cmu and shown to stimulate iNKT cytokine production, ⁴⁷ it is not surprising that different chlamydial species will express different glycolipid antigen profiles that associate with CD1d, and therefore, partially determine the outcome of infection pathology.

Because our data showed that NKT cells attenuated vaginal chlamydial clearance and exacerbated the severity and incidence of Infertility (Figure 1), we investigated the effects of Cmu infection of macrophages on expression of CD1d‐associated chlamydial lipids. The RAW cells were transfected with a chimeric CD1d protein that could be cleaved in order to isolate cell surface–expressed CD1d‐associated lipids. Chlamydia are obligate intracellular parasites that require host cell lipids, including glyerophospholipids, sphingolipids and cholesterol, to maintain the membrane‐bound inclusion in which replication occurs ⁴⁹ and as such have developed multiple means of modulating host cell lipid pathways, including both vesicular and non‐vesicular pathways to maintain this intracellular niche. Important pathways include acquisition of sphingomyelin (SM) and cholesterol from the Golgi apparatus by vesicular stacking, acquisition of ceramide and SM by the cytosolic lipid transporter CERT, acquisition of SM following Golgi fragmentation, acquisition of lipids by Rab GTPases and interactions with lipid droplets (reviewed in Elwell et al ⁴⁹ ). Lipid mass spectrometry analysis of cleaved CD1d‐containing microfiltered supernatants from infected and non‐infected macrophages revealed significant differences in the spectrum of CE, PE and PG lipid families. Surface presented CE, PE and PG lipid families were the significantly different CD1d‐loaded lipid families presented in Cmu‐infected macrophages, suggesting these families may be associated with the dramatic differences observed in chlamydial shedding and immunopathology. However, in surface CD1d‐cleaved cell lysates, PE levels remained equivocal to surface membrane levels, suggesting a limited role of CD1d‐loaded PE, and despite the quantitatively low‐abundance CE and PG lipids, they may have a significant role in infection and immunopathology. Unsurprisingly, Cmu infection of macrophages revealed sequestration of cellular but not cell surface SM lipids following infection. These data are consistent with the role of host cell–produced SM in maintaining the chlamydial inclusion integrity. Perhaps more importantly, these data demonstrate that CD1d lipid presentation is affected by chlamydial infection, and it is not unlikely that various chlamydial species affect lipid production and CD1d presentation of lipids differently, thereby eliciting different pro‐ or anti‐inflammatory phenotypes in resident NKT cells. A unique chlamydial glycolipid antigen (GLXA) has already been isolated from Cmu and shown to stimulate iNKT cytokine production, ⁴⁷ thus it might be possible to identify other glycolipid antigens from Chlamydia‐infected cells that could be used to induce protective immunity. The use of such glycolipid antigens has not been explored in terms of anti‐chlamydial immunity, however, as both infection and hydrosalpinx were reduced in the absence of NKT cells, this would have to be approached with caution, but the use of α‐GalCer to promote anti‐tumour immunity ⁵⁰ suggests this may be a possibility. An alternative approach may be to induce anergy in local NKT cell populations in the reproductive tract to reduce the pro‐inflammatory role of NKT cells. Anergy of NKT cell responses occurs naturally in many tumour models, ⁵¹ often involving PD‐1:PD‐L1/2 interactions, ⁵² suggesting a potential method to achieve this.

T cells are essential and sufficient for clearance of genital Cmu infections. ³⁶ , ³⁷ While there was no difference in CD3⁺, CD4⁺ or CD8⁺ (Figure 2F), T cell numbers in the cervicovagina of infected CD1d^−/− mice vs WT mice, the upper tract had significant T cell infiltration (Figure 3B). Systemic and local Chlamydia‐specific antibody production was not significantly affected by the absence of CD1d (Figure S1), suggesting NKT cells do not affect antibody production, but may suggest a role in B cell maturation, phenotype and antibody class switching that encourages further investigation. Thus, our data suggest that CD1d presentation of glycolipids enhances the severity and incidence of oviduct occlusion, increases the duration and magnitude of infection, possibly by compromising the activation of protective T cells.

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest to declare.

Supporting information

Figures S1–S7

Armitage CW, Carey AJ, Bryan ER, Kollipara A, Trim LK, Beagley KW. Pathogenic NKT cells attenuate urogenital chlamydial clearance and enhance infertility. Scand J Immunol. 2023;97:e13263. doi: 10.1111/sji.13263

DATA AVAILABILITY STATEMENT

The data that supports the findings of this study are available in the supplementary material of this article.