CD4⁺ T cell protection against murine Salmonella infection is female-specific and estrogen-dependent

Abstract

CD4⁺ T cells are necessary to control intracellular bacterial infections, but the role of sex in these infections is poorly understood. Here we show that upon challenge with Salmonella, a model intracellular bacterium, more male mice succumb to infection than females, despite having more Salmonella specific CD4⁺ T cells at later infection stages. Upon infecting mice lacking CD4⁺ T cells, survival differences reverse, which suggests that female and male CD4⁺ T cells play opposing roles during infection. Transfer of purified CD4⁺ T cells from either sex into CD4 deficient mice restores survival in females but decreases survival in males. Decreasing the hormone 17β-estradiol in females induces a more severe male-like phenotype. Exogenous 17β-estradiol induces both inflammatory and suppressive cytokines in CD4⁺ T cells from infected female mice and restores the protective function of T cells during infection. These findings reveal a sex difference in CD4⁺ T cell behavior during intracellular bacterial infection.

CD4⁺ T cells play critical roles in the immune response to intracellular bacterial infection. Here the authors suggest that CD4⁺ T cells can protect in a murine model of Salmonella infection but that the protection occurs in females and is oestrogen dependent.

Introduction

Infections caused by intracellular bacteria account for significant morbidity worldwide, and bacteria-specific helper CD4⁺ T cells are important to prevent bacterial replication and spreading¹. We and others have shown that mice lacking CD4⁺ T cells are unable to control infection with the intracellular bacterium Salmonella^2–5, along with other intracellular bacteria such as Yersinia sp., Francisella tularensis, and Mycobacterium tuberculosis^6–8. During these infections, CD4⁺ T cells produce potent T helper 1 (Th1)-related proinflammatory cytokines such as TNF and IFNγ that are essential for controlling intracellular bacterial replication⁹. This Th1-driven cytokine response activates a variety of innate immune cells, in particular macrophages, that ultimately control infection¹.

Historically, most immunological experiments in mice were conducted in females, or the sex of the mice was not specified, so whether CD4⁺ T cells serve the same roles in both sexes is not known. This highlights a gap between experimental and clinical studies, given the clear sex bias during intracellular bacterial infection in humans. During many infections, women initiate a stronger adaptive immune response than men, leading to lower female susceptibility^10,11. Sex differences in mortality to intracellular bacterial infection have been documented as far back as the Black Death, caused by the intracellular bacterium Yersinia pestis, in the 1300s¹². Even today, tuberculosis, caused by the intracellular pathogen Mycobacterium tuberculosis, is significantly more prevalent and deadly in males¹³. While there is no sex difference in the infection rate for Salmonella, men are 1.6 times more likely to develop invasive non-typhoidal Salmonella enterica and constitute the majority of its dangerous sequela, typhoid intestinal perforation^14–16.

In the absence of infection, women have slightly higher, although not always significantly different, numbers of circulating CD4⁺ T cells that display increased activation and proliferation in vitro in comparison with male CD4⁺ T cells^11,17–23. Virus-related immunization and infection studies have implicated 17β-estradiol (E2), a sex hormone found in higher concentrations in females, in mobilizing a potent CD4⁺ antiviral response¹⁰. In a bacterial pneumonia model, exogenous E2 promotes regulatory T cell (Treg) expansion and lower lung injury, highlighting the important role E2 plays on CD4⁺ T cell regulation of tissue damage²⁴. To be activated by E2, CD4⁺ T cells express multiple estrogen receptors, including estrogen receptor α (ERα) and β (ERβ) as well as the G protein-coupled estrogen receptor (GPER); however, ERα is the most highly expressed estrogen receptor on CD4⁺ T cells²⁵. Despite previous studies, little is known about the sex-based behavior of CD4⁺ T cells during infection with intracellular bacteria.

Based on prior work and clinical data, we hypothesized that female CD4⁺ T cells are functionally different from male T cells, leading to more efficient clearance of intracellular infection, but that T cells in both sexes would ultimately be important for containing bacterial spread and subsequent disease. To test this, we took advantage of a well-characterized infection model with the intracellular bacterium Salmonella enterica, serovar Typhimurium (hereafter S. Tm). We initially found that females exhibited increased survival when compared with males, despite males exhibiting higher numbers of S. Tm -specific CD4⁺ T cells at later time points. Resistance to S. Tm in females was lost in CD4 knockout mice (hereafter CD4KO mice), supporting previous studies showing a dependence on CD4⁺ T cells^2–5. Unexpectedly, CD4KO males were protected from infection, while conversely, adoptive transfer of CD4⁺ T cells reduced survival. CD4⁺ T cells transferred prior to infection in males or females adopted the protective fate of the recipient sex, demonstrating that female CD4⁺ T cells were not intrinsically more resistant to S. Tm infection and could not retain their protective phenotype when removed from the female. We also found that CD4⁺ T cell protection in females required E2 acting specifically on ERα expressed in CD4⁺ T cells, providing a cellular mechanism for these results. Together, our findings advance our fundamental understanding of the role of CD4⁺ T cells in intracellular infections based on biological sex and have therapeutic implications for treating these infections.

Results

Female mice displayed increased survival and infection resolution compared with males

To compare sex differences seen in clinical studies with experimental outcomes, we infected mice of both sexes with an attenuated strain of S. Tm, at a dose expected to cause disease but not death^4,26–30. Survival in males and females was similar until 10 days post-infection (dpi) when males began to succumb to infection. By 15 dpi, female survival was significantly higher than males (Fig. 1a). Because the spleen and liver are important blood filtering organs vital for S. Tm clearance, we assessed bacterial burdens in these organs over time. Initial bacterial burdens were similar but were higher at 5 dpi in females compared with males in both the spleen (Fig. 1b, c) and liver (Fig. 1f, g). By 10 dpi, bacterial clearance had shifted in favor of females (Fig. 1d and h). By 15 dpi, splenic bacterial burdens were similar in both sexes (Fig. 1e); however, surviving males had significantly greater liver burdens compared to females (Fig. 1i).

Fig. 1. Female mice have increased survival and resolution against S. Tm infections and increased clearance of bacteria in the liver.

a Mice were intravenously infected with $1\times {10}^5$ CFUs of S. Tm retro-orbitally. Survival of females (purple) and males (green) was measured 15 days post-infection. Bacterial burdens were assessed at 1, 5, 10, and 15 dpi in (b–e) spleens and (f–i) livers. Data are represented as probability of survival (a) or geometric mean +/− geometric SD (b–i). Statistical analyses were determined using a Log-rank Mantel–Cox tests (a) or an unpaired, two-sided t-test (b–i). For (a) *, P = 0.0303; for (c) ***, P = 0.0004; for (g) *, P = 0.0313; for (i) **, P = 0.0019. Represents 4 independent experiments combined for (a) (n = 20 total mice for each sex) and 2 independent experiments combined for (b and f) (n = 8 mice for each sex in total); (c) (n = 9 females and 7 males in total); (g) (n = 8 mice for each sex in total); (h) (n = 10 females and 8 males in total) and (e and i) (n = 9 females and 6 males in total) and 3 independent experiments combined for (d) (n = 15 females and 13 males in total). Source data are provided as a Source Data file.

Infected female mice produced a more balanced immune response compared with male mice

Although female mice showed higher bacterial burdens at 5 dpi, they also displayed increased serum levels of the chemokines CCL4 and CXCL10 (Fig. 2a, b). Both are crucial for attracting macrophages and T cells to infected tissues following intracellular bacterial infection^31,32. Higher serum levels of the T cell chemotactic chemokine CCL2 were also found at 5 dpi (Fig. 2c)³³. Females had higher serum levels of the antimicrobial cytokines IFNγ and TNF at 5 dpi, but by 10 dpi, both sexes had similar levels of both cytokines (Fig. 2d, e). Females also showed higher serum levels of the immunosuppressive cytokine IL-10 at 5 dpi (Fig. 2f). Because male survival begins to drop around 7 dpi when the adaptive CD4⁺ T cell response is normally peaking, we next assessed the S. Tm-specific CD4⁺ T cell response over time. The strain of S. Tm used was engineered to express the CD4⁺ T cell epitope 2W1S²⁶, allowing assessment of the expansion of endogenous S. Tm-specific CD4⁺ T cells in response to infection using tetramer-based flow cytometry (Fig. S1a)^26,34,35. Male and female mice were infected, and spleens and livers were harvested at 5, 7, 10, and 15 dpi. While S. Tm-specific T cell numbers were higher in females at 7 dpi, males showed significantly higher numbers of S. Tm-specific CD4⁺ T cells in the liver by 15 dpi (Fig. 3a–d).

Fig. 2. Females produce higher circulating levels of immune cell recruiting chemokines and cytokines involved in host defense against S. Tm early in infection.

Levels of serum a CCL4, b CXCL10, and c CCL2 at 5 dpi are shown. Levels of serum d IFNγ, e TNF, and f IL-10 at 5, 10, and 15 dpi are shown. Data are represented as mean +/− SEM. Statistical analyses were determined using a Mann–Whitney test (a), an unpaired, two-sided t-test (b, c), and a two-way ANOVA with Šídák’s multiple comparisons test (d–f). For (a) ***, P = 0.0006; for (b) **, P = 0.0027; for (c) *, P = 0.0221; for (d) ***, P = 0.0003; for (e) *, P = 0.0277; for (f) *, P = 0.0461. Represents 2 independent experiments combined with total n = 7 mice for each sex for (a); n = 8 mice for each sex for (b, c); n = 6 females and 3 males for day 0, n = 6 for each sex for day 1, n = 8 for each sex for day 5, n = 10 females and 8 males for day 10, and n = 8 females and 7 males for day 15 for each sex for (d); n = 6 females and 3 males for day 0, n = 3 for each sex for day 1, n = 13 females and 12 males for day 5, n = 10 females and 9 males for day 10, and n = 8 females and 6 males for day 15 for (e); n = 4 females and 3 males for day 0, n = 6 for each sex for day 1, n = 13 for each sex for day 5, n = 10 females and 9 males for day 10, and n = 8 females and 7 males for day 15 for (f). Source data are provided as a Source Data file.

Fig. 3. Female S. Tm-specific CD4⁺ T cells are increased at 7 dpi and produce more IL-10 but less TNF compared with male T cells.

S. Tm-specific T cells were quantified using flow cytometry and the total number of S. Tm-specific CD4⁺ T cells are shown in the spleen and liver at 5, 7, 10, and 15 dpi (a–d, respectively). The percentage of S. Tm-specific CD4⁺/CD44^hi T cells stained intracellularly for e TNF and f IL-10 are shown at 7 dpi. g Numbers of CD25⁺/FoxP3⁺ Tregs are shown at 7 dpi in the spleen. Data are represented as mean +/− SEM. Statistical analyses were determined using an unpaired, two-sided t-test. For (b) *, P = 0.0359; **, P = 0.0090; for (d) *, P = 0.0170; for (e) *, P = 0.0406; for (f) **, P = 0.0072; for (g) **, P = 0.0014. Represents 2 independent experiments combined for (a) (n = 8 mice for each sex in total) & (b) (n = 9 females and 8 males in total) and (e) (n = 8 mice for each sex in total) & (f) (n = 8 females and 7 males in total) & (g) (n = 9 mice for each sex in total) and 3 independent experiments combined for (c) (n = 12 females for spleen and liver, 11 males for spleen and 10 males for liver in total) & (d) (n = 14 females and 10 males in total). Source data are provided as a Source Data file.

To examine the possibility that CD4⁺ T cells were functionally different between infected males and females, we assessed S. Tm-specific CD4⁺ T cell cytokine production. Total CD4⁺ T cells were isolated from male and female spleens at 7 dpi and restimulated with 2W1S peptide to characterize S. Tm-specific T cell cytokine responses midway through infection prior to the stark drop in male survival at 10 dpi. Interestingly, female CD4⁺ T cells produced less TNF but higher levels of IL-10 (Figs. 3e, f and S1b). The increased IL-10 led us to examine whether females display an increased regulatory T cell (Treg) response during infection. Staining for the canonical Treg transcription factor Foxp3 revealed that females exhibited significantly higher numbers of S. Tm-specific Tregs in the spleen at 7 dpi (Fig. 3g).

Infected female mice showed reduced inflammation, liver necrosis, and splenomegaly compared with male mice

In disseminated multi-organ bacterial infections, IL-10 plays a key role in downregulating inflammation and tissue damage, while high TNF can enhance tissue injury, so we next explored whether there were differences in inflammatory pathology between males and females^4,36–38. No inherent sex-based differences in histology were found in the absence of infection between the female and male livers (Fig. 4a). Naïve livers showed normal architecture, including hepatocytes, portal triads, and central veins with no significant inflammation or necrosis. Hepatocytes were arranged in cords with patent sinusoids and evenly spaced portal triads and central veins. Single bile ductules were present alongside most portal veins. Contrastingly, by 10 dpi, males showed disruption of normal liver architecture, hepatocyte swelling, and chronic inflammatory infiltrate surrounding portal triads and central veins. Focal necrosis with associated mixed inflammation was shown (Fig. 4b). Livers from male mice also displayed consistent and significantly increased levels of biliary inflammation and focal and geographic tissue necrosis when compared with females (Fig. 4c, d). In spleens, while there were no significant histological differences seen between females and males (Fig. 4e–g), males displayed significantly increased splenomegaly compared with females, even when accounting for overall body weight differences (Fig. 4h). This data highlights that infected male mice are significantly more prone to severe tissue damage.

Fig. 4. Females have decreased inflammation and tissue necrosis in livers and decreased splenomegaly at 10 dpi in comparison to males.

Female and male livers were H&E-stained, imaged, and scored blinded by a pathologist. Representative images from a naïve and b 10 dpi livers are shown at 4× magnification with a 20× magnification inlay. c Liver inflammation scores and d tissue necrosis scores are shown at 10 dpi. Female and male livers were H&E stained, imaged, and scored by a pathologist. Representative images from e naïve and f 10 dpi spleens are shown at 4× magnification with a 10× magnification inlay. Degree of g pathologic change scores and h spleen to body weight ratio of spleens are shown at 10 dpi. Data are represented as mean +/− SEM. Statistical analyses were determined using an unpaired, two-sided t-test (c, d, and h). For (c) *, P = 0.0101; for (d) *, P = 0.0493; for (h) *, P = 0.0359. Data represents 2 independent experiments combined (n = 6 mice for each sex in (c, d, g), n = 9 females and 10 males for each sex in (h) in total). Each individual data point in (c, d, g, and h) represents an individual mouse/ biological replicate. Source data are provided as a Source Data file.

CD4⁺ T cells conferred protection against infection in females but were detrimental in males

The observed decreased CD4⁺ T cell production of TNF but increased IL-10 from female mice led us to speculate that female CD4⁺ T cells possessed a greater intrinsic immune balance compared with males that would ultimately protect against infection. To test this, we infected mice genetically deficient in CD4⁺ T cells (CD4KO mice). As anticipated, S. Tm infected female CD4KO mice showed a significant drop in survival compared with wild-type females; however, unexpectedly, the total loss of CD4⁺ T cells completely restored survival in males (Fig. 5a). To assess whether adding CD4⁺ T cells back could restore survival in the female CD4KO mice, we adoptively transferred CD4⁺ T cells isolated from naïve wild-type females and infected those mice. Survival between female CD4KO mice receiving CD4⁺ T cells compared with PBS-transferred control CD4KO mice was significantly different with probability of survival increasing from 25 to 75% with the addition of female CD4⁺ T cells (Fig. 5b). In contrast, transferring CD4⁺ T cells isolated from WT male mice into male CD4KO mice caused a significant drop in survival compared with PBS-transferred controls (Fig. 5c), recapitulating the data observed in WT mice and indicating that CD4⁺ T cells in male mice consistently displayed an adverse effect on survival.

Fig. 5. Adoptively transferred CD4⁺ T cells confer protection against S. Tm infection in female mice but are detrimental in males.

a Survival of female and male CD4KO mice infected intravenously with $1\times {10}^5$ CFUs of S. Tm retro-orbitally is shown. b Survival of female CD4KO mice adoptively transferred with approximately $2-3\times {10}^7$ CD4⁺ T cells isolated from WT female mice (dashed pink line) or given a PBS vehicle control (solid purple) 24 h before infection is shown. c Survival of male CD4KO mice adoptively transferred with approximately $2-3\times {10}^7$ CD4⁺ T cells isolated from WT male mice (dashed light green) or given a PBS vehicle control (solid green) 24 h before infection is shown. d Survival of male CD4KO mice adoptively transferred with approximately $2-3\times {10}^7$ CD4⁺ T cells isolated from WT female mice (dashed blue) or given a PBS vehicle control (solid green) 24 h before infection is shown. e Survival of female CD4KO mice adoptively transferred with approximately $2-3\times {10}^7$ CD4⁺ T cells isolated from WT male mice (dashed blue) or given a PBS vehicle control (solid purple) 24 h before infection is shown. Data are represented as probability of survival. Statistical analyses were determined using a Log-rank Mantel–Cox test (a–e). For (a) *, P = 0.0116; for (b) *, P = 0.0346; for (c) **, P = 0.0014; for (d) **, P = 0.0051; for (e) **, P = 0.0057. a represents 2 independent experiments combined (n = 8 female CD4KO and 10 male CD4KO mice in total), b represents 3 independent experiments combined (n = 7 female CD4KO mice transferred with female CD4⁺ T cells and 12 female CD4KO mice transferred with PBS control in total), c represents 2 independent experiments combined (n = 10 male CD4KO mice transferred with male CD4KO T cells and 9 male CD4KO mice transferred with PBS control in total), d represents 3 independent experiments combined (n = 12 male CD4KO mice transferred with female T CD4⁺ T cells and 21 male CD4KO mice transferred with PBS control in total), and e represents 2 independent experiments combined (n = 10 female CD4KO mice transferred with male CD4⁺ T cells and 8 female CD4KO mice transferred with PBS control in total). Source data are provided as a Source Data file.

The previous finding led us to surmise that female, but not male CD4⁺ T cells were intrinsically protective, so we transferred female CD4⁺ T cells into male CD4KO mice, anticipating protection. Instead, we found that neither male nor female CD4⁺ T cells improved survival in male CD4KO mice (Fig. 5c, d). We next transferred CD4⁺ T cells isolated from males into female CD4KO mice and found that recipient females displayed increased survival compared with PBS-transferred control female mice (Fig. 5e).

Ovariectomy decreased female survival during infection that was rescued by exogenous 17β-estradiol

This restoration of survival in female CD4KO recipient mice transferred with CD4⁺ T cells regardless of donor sex led us to consider that recipient extrinsic factors played a substantial role in protection against S. Tm infection. Because E2 is found at differing levels in male and female host environments, we sought to better understand the role of estrogen in female protection. Prepubescent female mice were ovariectomized (OVX) and allowed to recover, ensuring that endogenous ovarian hormones were depleted. OVX and ovary-intact sham-operated control mice were infected with S. Tm and assessed for weight loss and survival. OVX females succumbed more than their sham counterparts (Fig. 6a) and displayed increased inflammatory pathology and inflammation in the livers (Fig. S2). OVX females also had significantly higher bacterial burdens in the spleen and liver by 15 dpi (Fig. 6b). While both groups generated similar numbers of S. Tm-specific CD4⁺ T cells in the spleen, OVX females had significantly higher numbers of S. Tm-specific CD4⁺ T cells in the liver (Fig. 6c). To further investigate whether the ovarian-induced hormone environment induced CD4⁺ T cell mediated protection in females, we adoptively transferred female CD4⁺ T cells into OVX CD4KO females and found that, unlike transfer into ovary-intact CD4KO mice, transferred cells were not protective in OVX recipient mice (Fig. 6d), supporting our other findings that the hormonal environment directly affects the CD4⁺ T cell response in females to provide protection not present in male mice.

Fig. 6. Loss of ovarian hormones decreases survival during S. Tm infection, while exogenous E2 addback restores protection.

a–c Female mice were ovariectomized or underwent sham surgery and infected with S. Tm. a Survival of sham females (solid purple line) and OVX females (dotted line) is shown. b Bacterial burdens in the spleen and liver are shown at 15 dpi. c S. Tm-specific T cells were quantified using flow cytometry and total number of S. Tm-specific CD4⁺ T cells are shown in the spleen or liver at 15 dpi. d OVX or sham CD4KO mice were adoptively transferred with female WT CD4⁺ T cells and 24 h later, mice were infected with S. Tm and survival monitored. e–g Female mice were ovariectomized or underwent sham surgery, implanted with extended-release placebo or E2 pellet, then infected with S. Tm. Bacterial burdens in the spleen or liver are shown at e 5 and f 15 dpi. g S. Tm-specific T cells were quantified, and total number of S. Tm-specific CD4⁺ T cells are shown in the spleen and liver at 15 dpi. h OVX and sham WT mice were implanted with a placebo or E2 pellet then infected with S. Tm and survival was monitored. Data are represented as probability of survival (a, d, h), mean +/− SEM (c, g), or geometric mean +/− geometric SD (b, e, f). Statistical analyses were determined using an unpaired, two-sided t-test or a Log-rank Mantel–Cox test. For (a) *, P = 0.0256; for (b) *, P = 0.0157 (spleen) and *, P = 0.0401 (liver); for (c) *, P = 0.0450; for (d) *, P = 0.0368; for (f) *, P = 0.0474; for (g) *, P = 0.0204; for (h) *, P = 0.0173. a represents 2 independent experiments combined (n = 11 mice receiving sham surgery and 13 mice receiving OVX in total). b represents 2 independent experiments combined (n = 7 mice receiving sham surgery and 9 mice receiving OVX in total); c represents 2 independent experiments combined (n = 11 mice for each group in total); d represents 3 independent experiments combined (n = 9 female CD4KO mice receiving sham surgery & female CD4⁺ T cells and 12 female CD4KO mice receiving OVX surgery & female CD4⁺ T cells in total); e represents 3 independent experiments combined (n = 6 mice total per group); f represents 3 independent experiments combined (n = 7 female mice receiving OVX surgery & E2 addback and 8 female mice receiving OVX surgery & placebo total); g represents 2 independent experiments combined (n = 7 female mice receiving both OVX surgery & E2 addback and OVX surgery & placebo total for spleen, n = 6 female mice receiving OVX surgery & E2 addback and 9 female mice receiving OVX surgery & placebo total for liver); h represents 3–4 independent experiments combined (n = 13 female mice receiving OVX surgery & E2 addback and 18 female mice receiving OVX surgery & placebo total). Each individual data point in (b, c, e–g) represents an individual mouse/ biological replicate. Source data are provided as a Source Data file.

The loss of protection in OVX mice led us to hypothesize that 17β-estradiol (E2), a sex hormone found at higher levels in females than males, was playing a protective role in the ovary-intact females. To test this, we implanted OVX females subcutaneously with an extended-release matrix-driven delivery pellet releasing a physiologically relevant dosage of E2 and infected with S.Tm one week after implantation³⁹. E2 levels and FRT weights returned to those found in mice that underwent sham surgery demonstrating that E2 was functioning as expected (Fig. S3). Despite similar bacterial burdens at 5 dpi, bacterial burdens in the OVX female mice implanted with the E2 pellet were significantly lower in the spleen, but not liver, in comparison with placebo-implanted OVX females at 15 dpi, supporting our hypothesis (Fig. 6e, f). E2 also decreased expansion of S. Tm-specific CD4⁺ T cells in the spleen, suggesting that E2 regulates expansion of CD4⁺ T cells during infection (Fig. 6g). Restoration of E2 to OVX mice was also sufficient to rescue mice from mortality (Fig. 6h).

17β-estradiol directly impacts T cell cytokine production and fails to protect mice lacking T cell ERα

Since E2 was required for protection against infection in females, we sought to determine the cellular mechanism for this protection and posited that estrogen receptor signaling in the T cells themselves was required. Initially, we assessed whether E2 directly affected cytokine production in mice lacking gonadal hormones. OVX female mice were supplemented as above with E2 or implanted with a placebo pellet, and levels of serum TNF and IL-10 were assessed five days post-infection. While TNF was not significantly affected by the loss of E2, IL-10 was completely diminished (Fig. 7a, b). When these same cytokines were analyzed 15 days post-infection, IL-10 remained diminished in the absence of E2 (Fig. 7c) while TNF was somewhat, but not significantly, lower when E2 wasn’t present (Fig. 7d). We next assessed whether E2 could directly affect cytokine production by CD4⁺ T cells from infected mice. CD4⁺ T cells from female mice 15 days post-infection were purified and restimulated in vitro with CD3/CD28 beads in the presence or absence of exogenous E2 using charcoal-treated FBS and phenol red-free media to mitigate potential estrogenic effects from the media. We found that E2 enhanced the production of both TNF and IL-10 in these T cells (Fig. 7e, f). We next assessed whether E2 has a direct effect on T cell protection in vivo. GPER is expressed by CD4⁺ T cells and potentially plays a role in Treg function, while ERα is highly expressed by CD4⁺ T cells, leading us to assess the function of these two receptors^40,41. Naïve CD4⁺ T cells purified from GPER or ERα knockout mice were transferred into separate groups of CD4KO mice one day before infection. There was no difference in survival between female mice transferred with GPER KO or WT T cells, indicating that GPER is dispensable for T cell-mediated protection (Fig. S4). In contrast, transfer of ERα KO T cells failed to protect mice from infection, demonstrating a significant decrease in survival compared with mice receiving WT T cells (Fig. 7g). Combined, this data shows that E2 is an essential soluble mediator of protection against infection in female mice that activates CD4⁺ T cells, and that these cells require ERα-expression to maintain their protective effect against infection.

Fig. 7. E2 induces T cell cytokines and acts on T cell ERα receptor to restore control of infection.

Levels of serum a, c TNF and b, d IL-10, at a, b 5 and c, d 15 dpi are shown from OVX mice supplemented with E2 or placebo. CD4⁺ T cells from female spleens 15 days post-infection were restimulated with CD3/CD28 beads in the presence or absence of E2 and e TNF and f IL-10 in the supernatant was measured 48 h later; g female CD4KO mice were adoptively transferred with WT or ERα KO CD4⁺ T cells, then infected with S. Tm, and survival was monitored. Data are represented as mean +/− SEM (a–f) or probability of survival (g). Statistical analyses were determined using a Mann–Whitney test (a–d), an unpaired, two-sided t-test (e, f), or a Log-rank Mantel–Cox test (g). For (b) *, P = 0.0152; for (d) *, P = 0.0210; for (e) *, P = 0.0111; for (f) *, P = 0.0192; for (g) *, P = 0.0250. a–d represent 2 independent experiments combined (n = 6 mice per group for (a and c), n = 6 mice for OVX + E2 and 5 mice for OVX + placebo for (b and d)); e, f represents 6 samples from three mice pooled and (g) represents 4 independent experiments combined (n = 9 female CD4KO mice receiving WT CD4⁺ T cells and 16 female CD4KO mice receiving ERα KO CD4⁺ T cells in total). Source data are provided as a Source Data file.

Discussion

The most unanticipated finding from the current study was that CD4⁺ T cells adopted completely opposing roles depending on whether they resided in males or females. This is one of clearest cases for the sex difference in a T cell response that we are aware of. One possibility for why this outcome has not been previously revealed is prior experimental design and reporting. Historically, animal studies are biased towards gathering data from only a single sex. Indeed, immunological studies favor the use of female mice and, in over 50% of studies, the sex of the mice is not listed at all^42,43. Thus, prior studies that found an essential role for CD4⁺ T cells during Salmonella infection are consistent with the female data in the results presented here. Some studies show sex differences in CD4⁺ T cell numbers^11,17–23, but these small differences could not entirely explain the contrasting outcomes in the CD4⁺ T cell roles observed here. Th1 T cells producing TNF are important for combatting many intracellular bacterial infections, including Salmonella, so our findings support induction of this protective cell type⁹. This antimicrobial phenotype was balanced by an associated increase in the anti-inflammatory cytokine IL-10 produced by S. Tm-specific T cells and concomitant Treg expansion in female mice. We posit that female effector T cells aided in bacterial control, while increased Treg expansion dampened tissue inflammation and pathology, a protective effect not seen in males. This balance has been observed in other infections such as those caused by Leishmania, Schistosoma, and Helicobacter pylori⁴⁴. Treg depletion a week post S. Tm infection abrogates Th1 differentiation, suggesting Tregs are essential to selectively shape the immune response to S. Tm⁴⁵.

We found that T cells from mice lacking ovarian hormones appeared to be more male-like and lost protective capacity. This was further reflected in tissue damage that was more severe in ovariectomized mice compared with ovary-intact females. T cells transferred into ovariectomized CD4KO mice were unable to rescue survival, indicating that the extrinsic hormonal environment of the recipient mice directly influenced the transferred T cells. Exogenous supplementation of E2 in OVX mice increased survival and reduced infectious burdens. Sex differences in clinical patient outcomes for many bacterial infections are well-established; following the acute phase of bacteremia, surviving patients often display impairments in lymphocyte functionality, including the loss of CD4⁺ helper T cells, while addition of E2 is protective^46–48. While men and women acquire infections caused by intracellular bacteria at similar rates, men are more likely to exhibit severe disease and succumb to infection than women, even when accounting for external factors, such as smoking and socialization⁴⁹. We carefully selected the dose of E2 to mimic physiological levels since these levels promote Th1-type responses. In contrast, supraphysiological concentrations promote Th2-type humoral responses that are less likely to protect against intracellular infections^50,51. Supraphysiological doses of estrogens also inhibit T cell proliferation, reduce the T cell activating cytokine IL-2, and increase susceptibility to infection with Listeria monocytogenes^52,53. In mice, ERα is expressed by 85% of circulating CD4⁺ T cells, and signaling through ERα is generally proinflammatory, while ERβ signaling is anti-inflammatory, but expression is below 1%⁵⁴. GPER is expressed by both naïve CD4⁺ T cells and Foxp3⁺ Treg cells and induces both IL-10 and Foxp3 expression in CD4⁺ T cell populations⁴¹. While GPER was dispensable for the protective capacity of CD4⁺ T cells in this study, ERα on the T cells themselves was necessary to confer protection. It is conceivable that E2 binding to ERα on bacteria-specific CD4⁺ T cells directly induced a protective Th1 phenotype, balanced by Treg induction, whereas lower E2 levels in males were not sufficient to induce this balance. Indeed, we found that E2 directly enhances both TNF and IL-10 in CD4⁺ T cells from infected female mice, supporting its role in regulating T cell function. Loss of E2 via OVX led to a significant decrease in IL-10 in the serum of infected mice, while TNF was unaffected. We speculate this loss of immunosuppression leads to the increased liver pathology in OVX mice, akin to what is observed in the males. E2 supplementation rescued IL-10 production, possibly preventing the increased tissue damage.

While many intracellular bacterial infections rely on CD4⁺ T cell immunity, it is likely that other immune cells, not explored here, play key roles that might also influence the outcome of these infections. Many of these bacteria co-opt host macrophages to serve as an intracellular survival niche, and the interplay between CD4⁺ T cells and macrophages dictates protection and survival¹. The increase in females in macrophage recruiting chemokines in serum in our study also supports the likelihood that the female immune response overall supports control of Salmonella infection. Macrophages are also influenced by sex hormones, which likely impact their control of intracellular bacteria^24,55. E2 shifts macrophages towards more anti-inflammatory, reparative profiles in tissue injury models^56,57. While both sexes produce both estrogens and androgens, androgens such as testosterone circulate at higher levels in males and may play a role in the loss of male survival we observed in this study. An immunosuppressive role for androgens has been suggested, where androgens suppress the Th1 response to limit the protective capacity of CD4⁺ T cells⁵⁸. This feedback loop, whereby lack of immunological bacterial control allows for greater bacterial growth, ultimately may lead to the observed pathology and likely loss of organ function and death. It is worth noting both ERα KO female and male mice display high levels of testosterone that could potentially affect development of the ERα KO CD4⁺ T cells used for adoptive transfer⁵⁹.

While this study reveals differences between male and female CD4⁺ T cells during Salmonella infection, some study limitations should be highlighted. Salmonella is inherently a fecal-oral pathogen. We chose to infect intravenously with attenuated bacteria for consistency of dosing, but it is possible that oral or colitis infection models might show different sex biases. Further, C57BL/6 mice are susceptible to Salmonella, so future studies in resistant models (e.g. 129 Sv) would be informative. Notably, CD4KO mice display higher levels of MHC class II-restricted CD8⁺ T cells⁶⁰. It is not thought that CD8⁺ T cells play significant roles in Salmonella control during primary infection^3,5,61; however, we cannot rule out that these cells play a more prominent function in our study. We also recognize survivor bias may play a role in the observed differences in T cells and bacterial burdens between male and female mice on day 15 post-infection, but contend that the adoptive transfer experiments address this limitation to some degree.

These findings have broader implications for our understanding of the impact of sex on immunity to infection. Both clinical and animal studies indicate that females are protected against a variety of infections, including those caused by intracellular bacteria^10,25, often due to a more robust and proinflammatory immune response. This enhanced protection may have evolved to better protect females of childbearing age to ensure reproductive fitness and preservation at a species level^62–64. Our T cell data supports an enhanced immune response in females that was dependent on the host rather than the originating sex of the cell. E2 levels are constantly in flux due to menstrual cycles, pregnancy, and menopause. Our findings would suggest that lower E2 levels in post-menopausal women may contribute to more severe infections. Indeed, this is confirmed by clinical studies showing increased susceptibility to these infections^65,66. Conversely, providing exogenous E2 to males might confer protection in the mouse model used here or perhaps during infection in human populations. Adjusting E2 levels in this way might offer a therapeutic mechanism for generating better immunity against infections caused by intracellular bacteria, regardless of the sex of the infected person. Ultimately, our findings reveal that CD4⁺ T cells contribute to completely different outcomes during intracellular bacterial infection depending on the sex of the host. Future studies will attempt to further delineate the roles for these cells and if these outcomes extend to other infections and species.

Methods

Mice

All experimental procedures were approved and performed in compliance with the guidelines established by Tulane University School of Medicine’s Institutional Animal Care and Use Committee (protocol numbers 210, 842, and 1708). C57BL/6J and CD4KO (B6.129S2-Cd4^tm1Mak/J) female and male mice were purchased from Jackson Laboratory (Bar Harbor, ME) and were maintained under specific-pathogen-free conditions in the vivarium at Tulane University School of Medicine. Mice aged 8–12 weeks were used. Female mice were age-matched with male mice for experiments, and both were harvested at specified time intervals after infection. Animal breeding of CD4KO mice was conducted in accordance with recommendations from the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Spleens harvested from ERα KO mice were received from Troy Roepke (Rutgers University, New Brunswick, NJ, USA) and spleens from GPER mice were received from Sarah Lindsey (Tulane University, New Orleans, LA, USA).

Surgeries and hormone replacement

At 4–5 weeks of age, mice were anesthetized with isoflurane and bilaterally ovariectomized (OVX) as described, with surgeries on control mice performed in a similar fashion without removal of the ovaries (Sham)^67,68. Wounds were closed using absorbable sutures. All mice were given 4 weeks of recovery. For exogenous estradiol addback experiments, ovariectomized (OVX) mice were assigned to receive subcutaneous implants of placebo (−E2) or 0.01 mg 17β-Estradiol (+E2) 60-day release pellets (Innovative Research of America) 7–10 days prior to S. Tm inoculation. Uterine horns were removed at all time points upon euthanasia of mice, and wet weight was quantified as a bioassay ensuring ablation of ovarian hormones.

Bacteria growth conditions and preparation

Salmonella enterica subsp. enterica serovar Typhimurium (S. Tm) strain SL3261 (an aroA attenuated derivative of S. Tm strain SL1344) was tagged chromosomally with the 2W1S peptide (EAWGALANWAVDSA) and kanamycin resistance using the lambda red recombinase system²⁶. To do this, primers were designed with extension arms homologous to the 3’ portion of the OmpC gene, deleting the stop codon but extending downstream from it. A single FLAG sequence was inserted to assess insertion-positive bacteria. The stop codon was inserted with kanamycin resistance introduced downstream of the newly incorporated stop codon. PCR products were generated from a template plasmid encoding the 2W peptide and FLAG epitope (pJM1). PCR products were used directly for electrotransformation into S. Tm containing the temperature-sensitive pKD46 plasmid, carrying arabinose-inducible bacteriophage lambda Red genes^69,70. Bacterial suspensions in 10% glycerol were mixed with 0.5–1 μg of PCR product and incubated on ice for 30 min before transferring to a chilled 0.2-cm cuvette. Cuvettes were subjected to a single pulse of 12.5 kV/cm. After recovering for 1 h at 37 °C in Super Optimal broth with Catabolite repression medium, bacteria were plated on LB agar plates with 50 μg/ml kanamycin. DNA sequencing was used to verify recombination. Tagged bacteria were grown overnight at 37 °C in a 5% CO₂ incubator on Luria-Bertani (LB) agar plates (Miller) with added 50 μg/ml kanamycin. A single colony was then suspended in LB Broth (Thermo Fisher Scientific) with added 50 μg/ml kanamycin and incubated in a static 37 °C water bath for 5–6 h. For infections, bacteria were washed once in sterile phosphate-buffered saline (PBS) then diluted in sterile PBS (Invitrogen) to an infection dose of 1 × 105 CFUs of S. Tm in 100 μL of PBS.

Infections and bacterial burdens

Mice were infected with 1 × 10⁵ CFUs of S. Tm intravenously via the retroorbital route. Mice were weighed daily post-infection. At designated times post-infection, mice were euthanized by CO₂ asphyxiation, and spleens and livers were harvested for determination of bacterial burdens. If mice reached 20% loss of their starting weight, they were humanely euthanized and considered to not have survived the challenge. Organs were homogenized in sterile 0.1% Triton ×-100 and plated on LB agar containing 50 μg/ml kanamycin for enumeration of viable CFU using serial dilutions. The data were expressed as the geometric mean CFU per tissue.

Cytokine and chemokine quantification

Serum was harvested at all time points via cardiac puncture immediately after euthanasia. Serum was used to measure GM-CSF, IFNγ, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-12p70, IL-13, IL-18, TNFα, IL-9, IL-10, IL-17A (CTLA-8), IL-22, IL-23, IL-27, Eotaxin (CCL11), GROα (CXCL1), IP-10 (CXCL10), MCP-1 (CCL2), MCP-3 (CCL7), MIP-1α (CCL3), MIP-1β (CCL4), MIP-2, and RANTES (CCL5) using the Cytokine & Chemokine Convenience 26-Plex Mouse ProcartaPlex™ Panel 1 kit (Invitrogen). Following 1 h incubation at room temperature, blood was spun in a BD Microtainer tube at 10,000 × g for 10 min. Serum fraction was collected. Antigen standard was prepared by 4× dilution in diluent buffer. Fifty μL of capture beads were added to each well on the plate and plate was washed with 150 μL of 1× Wash Buffer using a Hand-Held Magnetic Plate Washer. Twenty-five μL of 1× universal assay buffer (UAB) was added to each well followed by 25 μL of prepared standards or serum samples. Twenty-five μL of 1× UAB was used for negative control wells. Plate was sealed and covered and shaken at 600 rpm for 2 h at room temperature. Plate was subsequently washed three times with 150 μL of 1× Wash Buffer using the Hand-Held Magnetic Plate Washer and Biotinylated detection Antibody Mix (25 μL/well) was added. The plate was sealed and covered and shaken at 600 rpm for 30 min at room temperature. The plate was subsequently washed with 150 μL of 1× Wash Buffer using the Hand-Held Magnetic Plate Washer. Streptavidin-PE (50 μL/well) was added, and the plate was sealed and covered and shaken at 600 rpm for 30 min at room temperature. The plate was subsequently washed with 150 μL of 1× Wash Buffer using the Hand-Held Magnetic Plate Washer. Reading buffer was added to the plate at 120 μL/well, and the plate was sealed, covered, and shaken at 600 rpm for 5 min at room temperature. The plate was run on a Luminex100.

Antibodies

Antibodies used throughout the study are listed specifically in each subsection but overall are shown here: BD Horizon Viability stain (Cat no 564406), Biolegend anti-CD4-BV605, clone RM4-5 (Cat no 100547); Biolegend anti-CD44-BV785, clone IM7 (Cat no 103041); Biolegend anti-CD44-PE, clone IM7 (Cat no 103008); Tonbo anti-CD11c-eFluor450, clone N418 (Cat no 75-0114-U100); Tonbo anti-CD11b-eFluor450, clone M1/70 (Cat no 75-0112-U100); Tonbo anti-CD19-eFluor450, clone 1D3/CD19 (Cat no 75-0193-U100); Tonbo anti-F4/80-eFluor450, clone BM8 (Cat no 75-4801-U100); Biolegend anti-CD11b-FITC, clone M1/70 (Cat no 101206); Biolegend anti-CD11c-FITC, clone N418 (Cat no 117306); Biolegend anti-F4/80-FITC, clone BM8 (Cat no 123108); Biolegend anti-CD19-FITC, clone 1D3/CD19 (Cat no 152404); Biolegend anti-CD3-PerCP-Cy5.5, clone 17A2 (Cat no 100218); Biolegend anti-CD4-PE, clone RM4-4 (Cat no 116016); Biolegend anti-CD25-PE-Cy7, clone A18246A (Cat no 113711); Biolegend anti-CD8-APC-Cy7, clone 53–6.7 (Cat no 100714); Biolegend anti-CD4- eFluor450, clone RM4-4 (Cat no 116018); Biolegend anti-IL-17-BV605, clone TC11-18H10.1 (Cat no 506927); Biolegend anti-CD8-FITC, clone 53–6.7 (Cat no 100706); ThermoFisher anti-CD8-eFluor450, clone 53–6.7 (Cat no 48-0081-82); Biolegend anti-IFN-gamma-PerCP-Cy5.5, clone XMG1.2 (Cat no 505822); Biolegend anti-TNF-PE, clone MP6-XT22 (Cat no 506306); Biolegend anti-CD4-PE-Cy5, clone GK1.5 (Cat no 100410); Biolegend anti-CD3-PE-Cy7, clone IM7 (Cat no 103030); Biolegend anti-IL-10-APC-Cy7, clone JES5-16E3 (Cat no 505036); ThermoFisher anti-Foxp3-FITC, clone FJK-16s (Cat no 11-5773-82).

Generation of single cell preparations and 2W1S-specific T cell enrichment for flow cytometry

Spleens were made in single-cell suspensions by homogenizing the organs over a 100-μm nylon mesh filter in cold sorter buffer (1× PBS, 2% newborn calf serum, and 0.1% sodium azide). Livers were placed into cRPMI, minced into 0.5 cm pieces, and added to conical tubes with pre-warmed cRPMI containing 1 mg/mL Collagenase IV. Samples were shaken for 30 min at 200 rpm at 37 °C. Following digestion, samples were gently mashed through a 70 μm cell strainer and subsequently washed. Hepatocytes were removed from livers by sedimenting the suspension at 50 × g for 2 min. Supernatant was collected and spun at 500 × g at 25 °C for five minutes to pellet the remaining cells. Cells were then underlaid with a 15% OptiPrep gradient (Sigma-Aldrich) prepared with Hanks’ balanced salt solution (HBSS) and spun at 750 × g for 20 min at room temperature. Interfaces containing lymphocytes were harvested. Samples were spun at 500 × g at 4 °C for five minutes (wash step). Single-cell preparations were resuspended in 200 μL with Fc Block. Cells were enriched by staining with 10 μM 2W1S:I-A^b MHC-II tetramer, conjugated to allophycocyanin (APC), incubated at room temperature in the dark for 1 h, then washed, stained with anti-APC magnetic beads (Miltenyi), and passed over an LS column on a quadroMACS magnet⁷¹. Eluted cells were stained with the following anti-mouse antibodies/stains: Viability (BV510), CD45 (BV605), CD44 (BV786), CD11b (FITC), CD11c (FITC), F4/80 (FITC), CD19 (FITC), CD4 (PE), CD25 (PE-Cy7), and CD8 (APC-Cy7). 2W1S-specific cells were enumerated using AccuCheck counting beads (Invitrogen) that were prepared at a known concentration. The number of 2W1S-specific T cells was calculated by multiplying the total cell number by the percentage of tetramer-positive cells in that population^72,73. Cells were collected on an LSRFortessa instrument or a FACSCelesta™ Cell Analyzer instrument (Beckton-Dickson). Data were analyzed using FlowJo software version 10 (TreeStar, Ashland, OR).

In vitro intracellular cytokine staining

Isolated CD4⁺ T cell suspensions from spleens were cultured with irradiated splenocytes to act as antigen-presenting cells in a 1:1 ratio and stimulated in vitro with 2W1S peptide (EAWGALANWAVDSA) at 100 ng/ml in complete RPMI at 37 °C. After 2 h of stimulation, 10 μg/ml brefeldin A (Sigma-Aldrich) was added for an additional 5 h. The plate was spun and washed at 500 × g at 4 °C for five minutes, and stained for viability for thirty minutes at 4 °C. The plate was washed as described and surface stains were added with incubation at 4 °C for 30 min. Cells were then fixed and permeabilized using the BD Cytofix/Cytoperm kit by adding 100 μL of Fixation/ Permeabilization solution and incubating for twenty minutes at 4 °C. The plate was washed twice 200 μL of 1× BD Perm/Wash buffer, spinning at 500 × g at 4 °C for five minutes in between washes. The plate was then stained overnight for intracellular cytokines. Cytokine-positive gates were set on naive, CD44^low T cells, which do not produce cytokines, from each organ, as other studies have shown previously³⁴. Cells were stained with the following anti-mouse antibodies/stains: CD4 (ef450), Viability (BV510), IL-17 (BV605), CD11c (FITC), CD11b (FITC), CD8 (FITC), F4/80 (FITC), CD19 (FITC), IFNγ (PerCP-Cy5.5), TNF (PE), CD4 (PE-Cy5), CD44 (PE-Cy7), IL-10 (APC-Cy7).

Transcription factor staining

Spleens and livers were harvested from infected mice, and cells were made into single-cell preparations, as described above. Cells were stained with 2W1S:I-A^b-APC and enriched for antigen-specific cells, as described above. Cells were washed and blocked with Fc Block and subsequently stained with viability dye (BV510) for 30 min at 4 °C, washed with 100 μL of sorter buffer by spinning at 500 × g at 4 °C, and stained with anti-mouse surface antibodies. Following an incubation for 30 min at 4 °C, cells were washed by spinning at 500 × g at 4 °C for five minutes and fixed by adding 1 mL off Foxp3 Fixation/Permeabilization working solution (eBioscience), pulse vortexed, and incubated for 30 min at 4 °C protected from light. Following incubation, a wash of 2 mL of 1× Permeabilization Buffer was added to each tube and samples were centrifuged at 500 × g at 4 °C for five minutes, repeated twice. Pellet was resuspended in residual buffer and nuclear transcription factors stains were added and incubated overnight at 4 °C. The next morning, samples were washed twice with 1× permeabilization wash buffer by spinning at 500 × g at 4 °C for five minutes and resuspended in sorter buffer for fluorescence-activated cell sorter (FACS) analysis. Samples were stained with the following anti-mouse antibodies/ stains: Gata3 (BV421), Viability (BV510), T-bet (BV605), CD44 (BV785), Foxp3 (FITC), RorgT (PE), CD4 (PE-Cy7), and CD25 (APC-Cy7).

Hematoxylin and Eosin Staining and pathology scoring

After the mice tissues were harvested, the tissues were fixed in formalin. Tissue from livers and spleens of infected and control mice was paraffin-embedded, sections cut at 5 μm intervals, fixed to negatively charged glass slides, and then stained by traditional hematoxylin and eosin staining. The slides were scored by a pathologist (RC) who was blinded to their treatment arm. Slides were examined on a Nikon Eclipse 50i using Nikon objectives, including 4, 10, 20, and 40×. Images were produced utilizing an Infinity 5 camera from Teledyne Luminera and their Infinity Analyze 7 software package.

Slides from all subjects were subjected to evaluation and scoring based on several categories, including inflammation, cell types, necrosis, cell swelling, pigment (bile) deposition, or disruption of normal architecture. Inflammation scoring followed a mild (+), moderate (++), or severe (+++) rating system considering amount, cell types, and location of inflammatory infiltrate. Normal hepatic architecture is considered cords/plates of hepatocytes with patent sinusoids, a lobular distribution with evenly spaced central veins and hepatic triads (hepatic artery, portal vein, and bile ductules), and essentially no inflammatory infiltrate. Normal splenic architecture is considered mild variation in follicle size and even distribution with narrow, open sinusoids, and minimal trilineage hematopoiesis.

CD4⁺ T cell isolation and adoptive transfer

Spleens and mesenteric lymph nodes were harvested from donor C57BL/6J mice. Spleens harvested from ERα KO mice were received from Troy Roepke (Rutgers University, New Brunswick, NJ, USA) and spleens from GPER mice were received from Sarah Lindsey (Tulane University, New Orleans, LA, USA). Cells were brought to single cell suspension at a concentration of 1 × 10⁸ cells/mL and CD4⁺ T cells were isolated using the EasySep™ Mouse CD4⁺ T Cell Isolation Kit (Stemcell). Samples were incubated with 50 μL rat serum per mL of sample. Fifty μL Isolation cocktail per mL of sample was added, and samples were incubated at room temperature for 10 min. Following incubation, 75 μL RapidSpheres were added per each mL of sample. Samples were mixed and incubated for 2.5 min and topped up to a total volume of 2.5 mL. Samples were placed in StemCell EasyEights™ magnets and incubated at room temperature for five minutes. Following this, isolated CD4⁺ T cells were collected in the supernatant. CD4⁺ T cell purity for each experiment was 91–96%. Absolute cell numbers of CD4⁺ T cells were determined, and cells were transferred into anesthetized mice via retroorbital injection in 50 μL PBS into the left eye. Mice were infected 24 h later into the right eye as described above. All protocols were conducted using sterile technique in a laminar flow hood.

CD4⁺ T cell culture and cytokine quantification

Using sterile technique in a laminar flow hood, CD4⁺ T cells were isolated as described above. Two hundred thousand cells per well were cultured with Iscove’s Modified Dulbecco’s Medium with added L-Glutamine and 25 mM HEPES but lacking Phenol Red (Gibco), supplemented with 2% charcoal-stripped FBS (Sigma) and Dynabeads CD3/CD28 activator beads (Thermo Fisher) with or without the addition of 10⁻⁹ M water-soluble β-Estradiol (Millipore Sigma). Cultures were incubated for 48 h at 37 °C in a 5% CO₂ incubator. Following 48 h, plates were spun at 500 × g for five minutes at 4 °C to pellet the cells and culture supernatant was collected. Cytokines TNF and IL-10 in cultures were identified through Mouse TNF alpha Uncoated ELISA Kit (Thermo Fisher) and Mouse IL-10 Uncoated ELISA Kit (Thermo Fisher), respectively. Plates were coated with 100 μL per well of capture antibody (anti-TNF alpha or anti-IL-10, respectively) and incubated overnight at 4 °C. Wells were washed three times with 250 μL of Wash Buffer per well. Plates were then blocked with 200 μL of 1× ELISA Diluent at room temperature for 1 h. Standards were prepared by twofold dilutions for a total of eight points. Following a wash as described above, prepared standards and samples were added at 100 μL per well to appropriate wells, and 100 μL of 1× ELISA Diluent was added to the blank wells as a negative control. The plate was sealed and incubated overnight at 4 °C. Plates were washed three times and 100 μL of diluted Detection Antibody was added to the plate. The plate was then incubated at room temperature for 1 h. Diluted Streptavidin-HRP at 100 μL per well was added to the plate and incubated at room temperature for 30 min. Following incubation, the plate was washed as described above five times. One hundred μL of TMB solution per well was added across the plate, and the plate was incubated for 15 min. Following incubation, 100 μL per well of stop solution was added to each well, and results were read at 450 nm and 570 nm on a plate reader.

Estradiol quantification

Estradiol in serum was quantified through Estradiol ELISA (Calbiotech). To do this, 25 μL of standards and samples were dispensed into the streptavidin-coated plate with 50 μL Estradiol Biotin Reagent per well. Estradiol Biotin Reagent alone was added as a negative control. The plate was incubated for 45 min at room temperature. One hundred μL of E2 enzyme conjugate was added on the top of the reaction mixture further incubated for an additional 45 min. Liquid was removed from all wells, and the wells were washed thrice with 300 μL of Wash Buffer. Next, 100 μL of TMB Reagent was added and incubated at room temperature for 20 min. The development was stopped with the addition of 50 μL of stop solution, and the absorbance measured spectrophotometrically at 450 nm on a plate reader.

Statistics and reproducibility

Statistical analysis was performed using an unpaired two-sided t-test (normal), Mann–Whitney test (non-normal), one-way ANOVA, two-way ANOVA, or Mantel–Cox test as stated in the figure legends (Prism version 10; GraphPad Software, Inc.). Significance is displayed as exact p-values listed in the figure legends for each panel of each figure if significant. Outliers were removed using the GraphPad Prism ROUT outlier test, and test was applied equally to all data sets. Prior research from our group and others was used to determine sample size; no statistical method was used to predetermine sample size. Experimental repeatability is listed in each figure legend for each figure. The Investigators were generally not blinded to allocation during experiments and outcome assessment, except for histological scoring in Fig. 4, which was single-blind with respect to the scoring pathologist.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Conceptualization: S.J.D., R.M.H., J.R.K., J.B.M., Methodology: S.J.D., R.M.H., S.H.L., J.A.M., F.M.J., J.B.M., Investigation: S.J.D., R.M.H., L.B., V.E.I., M.S.G., C.M.H., V.M.L., D.L.B, A.J.P., A.M.B., R.C., S.A.B., H.Lee, H.Lu, S.H.L., A.Y., V.A., T.A.R., J.B.M., Funding acquisition: J.B.M., Project administration: J.B.M., Supervision: J.B.M., Writing: S.J.D., R.M.H., J.B.M., Writing–eview & editing: S.J.D., R.M.H., F.M.J., S.H.L., H.L., J.A.M., L.A.M., J.B.M.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

Source data are provided with this paper. All data are available in the main text or the supplementary materials and as a zipped source file containing Excel spreadsheets encompassing the raw source data used in the manuscript. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71805-1.