CXCR3 escapes X chromosome inactivation in T cells during infection: potential implications for sex differences in immune responses

Abstract

CXCR3, an X-linked gene, is subject to X-chromosome inactivation (XCI), but it is unclear whether CXCR3 escapes XCI in immune cells. We determined whether CXCR3 escapes XCI in vivo, evaluated the contribution of allelic CXCR3 expression to the phenotypic properties of T cells during experimental infection with Leishmania, and examined the potential implications to sex differences in immune responses. We used a novel bicistronic CXCR3 dual reporter mouse, with each CXCR3 allele linked to a green or red fluorescent reporter without affecting endogenous CXCR3 expression. Our results show that CXCR3 escapes XCI, bi-allelic CXCR3 expressing T cells produce more CXCR3 protein than mono-allelic CXCR3 expressing cells, and bi-allelic CXCR3 expressing T cells produce more IFN-γ, IL-2 and CD69 compared to T cells which express CXCR3 from one allele during L. mexicana infection. These results demonstrate that XCI-escape by CXCR3 potentially contributes to the sex associated bias observed during infection.

INTRODUCTION

Sex disparities in autoimmune and inflammatory diseases have been extensively studied and are well documented. This disproportionality has been shown to be attributable to an over-active immune response in females (1), which is generally associated with increased T cell activation and higher production of Th1 cytokines such as IFN-γ (2). Similarly, sex biased resistance to bacterial, viral and parasitic infections has been described in epidemiological and experimental studies (3, 4). While the underlying causes of this sex disparity in immune responses are still incompletely understood, sex hormones and genetic factors appear to play major roles.

Genetic and epigenetic factors which contribute to sexual dimorphism are largely centered around the X chromosome. Females have two X chromosomes (XX), therefore to equalize gene expression between sexes, they undergo a process called X chromosome inactivation (XCI). During XCI, one X chromosome is randomly chosen to be transcriptionally silenced (5). However, it is known that a number of X-linked genes escape XCI and display bi-allelic gene expression (6). Recent studies demonstrate that genes which escape XCI vary, depending on the cell type, tissue, individual, age and even the methods used (7–9). This highlights the need for a deeper understanding of X-linked genes that escape XCI and the potential contribution of gene specific XCI escape on sexual dimorphism in autoimmune and infectious diseases.

Immune-associated genes that are mapped to the X-chromosome are of particular interest because they could affect the nature and intensity of immune responses that exist between sexes, depending on whether or not they escape XCI. An important X-linked immune associated gene is CXCR3, an inducible chemokine receptor and significant regulator of Th1 immune responses (10). CXCR3 is expressed by activated CD4+ T cells, CD8+ T cells, macrophages, dendritic cells and other immune cells (11). In T cells, CXCR3 mRNA has been shown to be overexpressed in females relative to males (12). Regulation of CXCR3 expression is critical in immune cells so as to prevent uncontrolled amplification of the immune response. CXCR3 has been identified as a crucial mediator of inflammatory autoimmune diseases, which predominantly affect females (12, 13). CXCR3 and its ligands have also been shown to play a critical role in resistance to intracellular infectious diseases such as leishmaniasis (14). Interestingly, males are more susceptible to leishmaniasis than females, despite similar pathogenic exposure (15, 16). These studies demonstrate the need to further explore the contribution made by allelic CXCR3 expression in determining sex mediated immunity to Th1 associated infectious and autoimmune diseases.

Unlike all other chemokine receptors, CXCR3 is mapped to the X-chromosome (17), and is therefore subject to XCI. However, whether CXCR3 escapes XCI, and what cells and tissues are involved, remains unclear. A number of studies examining gene-wide XCI and escape from XCI does not reveal CXCR3 as an XCI escapee. This may allude to the lack of sensitivity of these methods or the absence of single nucleotide polymorphisms (SNPs) that could facilitate determination of allelic usage using next generation sequencing technologies (18). Although a recent study identified CXCR3 as a possible escapee of XCI in humans (19), a more suitable method is needed to clarify whether CXCR3 escapes XCI, what cells and tissues may be involved in XCI escape, and the possible biological implications during inflammatory disease conditions in vivo.

Current methods such as RNA-FISH and sequencing technologies have improved the identification of genes which escape XCI (20, 21). However, the ability to visualize allelic gene expression in single cells in vivo, as well as determine the phenotypic and biological consequences of bi-allelic gene expression in these cells is highly improbable, since these methods inevitably result in cell death. A method which is able to visualize gene allelic usage in live cells in vivo, as well as determine their phenotypic properties, will allow for a more in-depth investigation of the consequences of escape from XCI. Here we describe a novel experimental mouse model that permits visualization of allelic usage of the X-linked gene CXCR3 using imaging and flow cytometry methods. We also report the identification of a hyper-inflammatory subset of T cells which express CXCR3 from both X chromosomes and likely contribute to the enhanced development of Th1 immune response in females. Further, we describe the phenotype of these bi-allelic CXCR3 expressing T cells in the context of localized inflammation.

MATERIALS AND METHODS

Mice

All mice were maintained in a pathogen-free animal facility in accordance with NIH and institutional guidelines and approved by The Ohio State University Animal Care and Use Committee (Protocol #2008A0105).

Generation of Cxcr3^RFP and Cxcr3^GFP/RFP bicistronic reporter mice

The Cxcr3^RFP targeting vector consisted of a 13kb mouse genomic fragment which contained CXCR3 exons 1 and 2, flanking sequences of 5.8kb and 3.4kb at the 5’ and 3’ ends of the coding region respectively. The vector was further modified by introducing an IRES element, tdTomato, and a bovine growth hormone polyA signal. This vector was linearized, then electroporated into C57BL/6 ES cells, followed by selection of correctly targeted clones. Positive ES cell clones were microinjected into albino C57BL/6 blastocysts and implanted into pseudopregnant female mice. The resulting chimeras were bred to wild type albino C57BL/6 females. The targeted mutation was confirmed by Southern blot and PCR genotyping. To obtain Cxcr3^GFP/RFP bicistronic dual reporter mice, homozygous Cxcr3^RFP females were bred to Cxcr3^GFP males.

Southern blot and PCR screening

Southern blot screening was performed as previously described (22). Briefly, genomic DNA was prepared from ES cell clones or mouse tail snips. DNA was digested using BamHI, and hybridized to a 500bp digoxigenin (DIG) labelled probe. Primers for PCR genotyping of Cxcr3^RFP mice were similar for Cxcr3^GFP mice as described previously (22). Expected PCR product sizes for WT and Cxcr3^RFP mice were 448bp and 230bp respectively.

T cell activation

Isolated cells were incubated with plate bound anti-CD3 (3μg/ml, clone 145–2C11) and anti-CD28 (4μg/ml, clone 37.51) antibodies (Biolegend, SanDiego CA USA). Splenocytes or lymph node cells from Leishmania infected mice were restimulated in vitro with 20μg/ml of freeze-thawed L. mexicana antigen for 72 h, or PMA, ionomycin and brefeldin for 4 h at 37°C. Following activation, cells were either rested in conditioned media and analyzed by flow cytometry, or collected for real time PCR analysis.

Leishmania infection

Leishmania mexicana (MNYC/BZ/62/m379) parasites were maintained by serial passage in murine BALB/c footpads. Eight week old female WT, Cxcr3^GFP, Cxcr3^RFP or Cxcr3^GFP/RFP mice were infected with 2.0 × 10⁶ L. mexicana promastigotes in the hind-left footpad or 1 × 10⁴ promastigotes in the left ear and monitored for eight weeks. The ear lesion thickness was measured at 8 weeks post infection and thickness of the non-infected ear was subtracted from the thickness of the infected ear for each mouse. At the time of harvest, ear lesions and draining lymph nodes were isolated, homogenized and used for parasite quantification.

Statistical analysis

Student’s unpaired t test was used to determine statistical significance of values obtained. P values less than 0.01 were considered statistically significant.

RESULTS AND DISCUSSION

CXCR3 bicistronic dual reporter mice faithfully report allelic CXCR3 expression

We designed an in vivo mouse model that allowed us to determine allelic expression of the X-linked gene CXCR3. To do this, we generated and characterized a CXCR3 bicistronic dual reporter mouse Cxcr3^GFP/RFP. In female Cxcr3^GFP/RFP mice, each CXCR3 allele is linked to a green or red fluorescent reporter via an internal ribosome entry site (IRES) element, which permits expression of the downstream cistron (EGFP or td-Tomato RFP) without interfering with CXCR3 expression (Figure 1A). We have previously reported on the generation and characterization of a CXCR3 bicistronic EGFP reporter mouse (Cxcr3^GFP) (22, 23). Using the same approach, we generated a CXCR3 bicistronic RFP reporter mice (Cxcr3^RFP mice), which were screened for successful integration into the CXCR3 locus (Figure 1B and 1C). These mice were viable and healthy with no phenotypic defects.

Figure 1: Generation, characterization and validation of CXCR3^RFP reporter and CXCR3^GFP/RFP dual reporter mouse.

(A) Targeting vector for generation of CXCR3^RFP reporter mouse and strategy for homologous integration onto the CXCR3 gene locus. Southern blot strategy for detection of wild type and mutated gene in electroporated ES cells. A similar strategy was used for confirming germ line transmission in F1 mice. p – probe; b – BamH1 site; tk – Thymidine kinase gene cassette; E1 and E2 – exons 1 and 2. (B) Southern blot screening of genomic DNA extracted from embryonic stem cell clones transfected with CXCR3^RFP targeting vector. (C) Genotyping of wild type, homozygous and heterozygous CXCR3^RFP mice by PCR is also shown. (D) Assessment of CXCR3 production in CXCR3^GFP, CXCR3^RFP and CXCR3^WT mice by flow cytometry. Splenoctyes were activated in vitro with anti CD3 and CD28 antibodies according to a protocol previously described (22). CXCR3 was detected by staining with PE-Cy7 conjugated anti CXCR3 antibodies (CXCR3^GFP, CXCR3^RFP and CXCR3^WT mice), GFP fluorescence (CXCR3^GFP mice), and RFP fluorescence (CXCR3^RFP mice). (E) Generation of female CXCR3^GFP/RFP dual reporter mice by breeding CXCR3^GFP and CXCR3^RFP mice. In these female mice, one CXCR3 allele is tagged with a GFP reporter while the other allele is tagged with an RFP reporter.

Next, we analyzed CXCR3 expression in Cxcr3^RFP mice. Similar to Cxcr3^GFP mice, Cxcr3^RFP mice were designed to report CXCR3 expression by RFP fluorescence, without interfering with endogenous expression of CXCR3. As expected, CXCR3 RNA and protein expression in Cxcr3^RFP mice was similar to that of WT C57BL/6 mice. We also determined whether CXCR3 expression in Cxcr3^RFP mice correlated with tdTomato RFP production (24). Our results showed that, similar to Cxcr3^GFP mice, expression of tdTomato RFP was directly proportionate to surface expression of the CXCR3 protein in these cells (Figure 1D). These data demonstrate that tdTomato RFP is a reliable reporter system for CXCR3 expression in cells of Cxcr3^RFP mice.

Our generated Cxcr3^RFP and Cxcr3^GFP mice provide a novel and highly useful model to study allelic usage of CXCR3 in females in vivo at the cellular level. F1 female progeny derived from the cross of male hemizygous Cxcr3^RFP mice and female homozygous Cxcr3^GFP mice, or male hemizygous Cxcr3^GFP mice and female homozygous Cxcr3^RFP mice are CXCR3 dual reporter mice with each CXCR3 allele tagged to a GFP or RFP reporter (Figure 1E). We are therefore able to determine allelic usage of CXCR3 in individual cells, and examine the phenotypic properties of cells that display mono- or bi- allelic usage of CXCR3. This tool will also enable the in vivo assessment of CXCR3 allelic usage in various cells, tissues and organ systems, as well as in different infectious autoimmune and neoplastic disease contexts. Female Cxcr3^GFP/RFP dual reporter mice are viable and healthy, and display no observable phenotypic differences compared to WT, Cxcr3^GFP and Cxcr3^GFP/RFP mice.

CXCR3 gene escapes XCI in activated T cells

We first determined whether immune cells from Cxcr3^GFP/RFP mice are capable of expressing CXCR3 from both alleles. Since CXCR3 is not typically constitutively expressed in naïve T cells, we induced CXCR3 expression in vitro as described previously (24). We determined allelic expression of CXCR3 by analyzing GFP and RFP expression in individual cells by flow cytometry. Our results indicate that a small percentage of activated T cells (about 6.6%) express CXCR3 from both alleles (Figure 2A). Although extreme care was taken to eliminate doublet cells in the flow cytometric analysis, we confirmed bi-allelic CXCR3 expression in single cells by imaging flow cytometry (25). Imaging flow cytometry of activated cells further confirmed bi-allelic expression of CXCR3 in activated T cells (Figure 2B).

Figure 2: CXCR3 gene escapes X-chromosome inactivation in T cells.

(A) T cells were isolated from spleens of CXCR3^GFP/RFP mice and were activated in vitro. Allelic expression of CXCR3 was determined by GFP and RFP fluorescence using flow cytometry. WT, CXCR3^GFP and CXCR3^RFP mice were used as controls. Representative plots of bi-allelic expression of CXCR3 in CXCR3^GFP/RFP mice are shown in the CD3 gated double positive (upper right) quadrant of CXCR3^GFP/RFP mice. Graph showing quantification of GFP, RFP, and GFP+RFP expressing cells over three independent experiments. (B) Determination of bi-allelic CXCR3 gene expression in CXCR3^GFP/RFP mice by imaging flow cytometry. Single cells were captured and visualized for CXCR3-GFP, CXCR3-RFP and CD3 expression. Dot plots and images of selected mono-allelic and bi-allelic CXCR3 expressing T cells are shown. Yellow arrows indicate images of bi-allelic CXCR3 expressing CD3⁺ T cells. (C) Mono-allelic, and bi-allelic CXCR3 expressing cells in female CXCR3^GFP/RFP mice were activated and stained for CXCR3 protein expression using PE-Cy7 conjugated anti-CXCR3 antibodies. A dot plot of gating used to determine the levels of CXCR3 expression represented by mean fluorescence intensities (MFI) of the respective T cell populations, are shown on the left. MFI are shown as a bar graph. The negative control T cell population was obtained from CXCR3 knock out mice. *p value < 0.05. Error bars represents SEM.

Next, we examined the effect of bi-allelic CXCR3 expression on total CXCR3 protein levels on individual T cells. To do this, we stained activated Cxcr3^GFP/RFP T cells with PECy7 conjugated anti-CXCR3 antibody, and analyzed CXCR3 protein expression in GFP-only, RFP-only, and GFP+RFP cells. As expected, cells showing bi-allelic CXCR3 expression displayed higher CXCR3 protein on their cells compared to mono-allelic CXCR3 expressing T cells (Figure 2C). Similar results were found in bi-allelic CXCR3 expressing T cells from spleens and lymph nodes of infected Cxcr3^GFP/RFP mice compared to mono-allelic CXCR3 expressing cells (Supplemental Figure 1). Given that CXCR3 is involved in immune cell migration, differentiation and activation, the increased protein expression observed in bi-allelic CXCR3 expressing cells can potentially result in enhanced immune response, leading to significant biological consequences in autoimmune and infectious diseases.

Bi-allelic CXCR3 expressing T cells display a more activated phenotype

In addition to our ability to determine allelic expression of CXCR3 in individual cells in vivo, our novel Cxcr3^GFP/RFP mice allows us to further examine the phenotypic properties of individual cells that exhibit bi-allelic CXCR3 gene expression. Therefore, we next explored the biological implications of bi-allelic CXCR3 expression using an experimental in vivo infection model in Cxcr3^GFP/RFP mice. Our hypothesis was that a stimulatory environment (as is the case with infection) will result in increased bi-allelic expression of CXCR3 in activated T cells in vivo. We used the well-established Leishmania infection model, which requires a Th1 immune response for protection against disease. Since CXCR3 is predominantly expressed in T helper type 1 (Th1) cells, and is involved in the generation and maintenance of a Th1 immune response, this was an ideal model to test our hypothesis. We first validated our model using an intradermal ear infection of Cxcr3^WT, Cxcr3^+/− and Cxcr3^−/− mice with L. mexicana promastigotes. At 8 weeks, we observed that lesion incidence and subsequent lesion ulcerations (which was indicative of degree of infection) directly correlated with allelic expression of CXCR3 in these mice (Figure 3A). Specifically, 100% of Cxcr3^−/− mice developed the lesion, while 85.7% of Cxcr3^+/− and 60% of Cxcr3^WT mice showed lesions. Similarly, lesion ulcerations were observed in 60%, 42.9% and 20% of Cxcr3^−/−, Cxcr3^+/− and Cxcr3^WT mice respectively. Furthermore, thickness and circumference of lesions were greater in Cxcr3^−/− mice compared to Cxcr3^WT mice, while parasitic burdens were higher in Cxcr3^−/− mice compared to Cxcr3^+/− (Figure 3B). Interestingly parasitic burdens in the draining lymph nodes were similar in all 3 groups, which is consistent previous results on the role of CXCR3 in immune response to L. major at the site of infection (14). Taken together, our results demonstrate the importance of CXCR3 copy number in controlling L. mexicana infection.

Figure 3: Bi-allelic CXCR3 expressing T cells are increased during in vivo infection with Leishmania mexicana.

(A) Representative ear lesions in of CXCR3^−/−, CXCR3^+/− and CXCR3^WT mice infected with L. mexicana. (B) Lesion thickness, circumference and parasitic burdens of CXCR3^−/−, CXCR3^+/− and CXCR3^WT mice infected with L. mexicana. (C-E) Flow cytometric analysis of allelic CXCR3 expression in CD3+ gated cells of the (C) draining popliteal lymph node, (D) spleen and (E) ears of L. mexicana infected and uninfected CXCR3^GFP/RFP dual reporter mouse. Graphs showing percentages of mono-allelic and bi-allelic CXCR3 expressing T cells are also shown. *p value < 0.05, **p value < 0.01. Error bars represents SEM. All experiments were performed at 8 weeks post infection.

Our analysis of spleens and lymph nodes of L. mexicana infected and uninfected Cxcr3^GFP/RFP mice using a high dose footpad infection model, showed a significant increase in T cells displaying bi-allelic expression of CXCR3 in infected mice compared to uninfected mice (Figure 3C and 3D). We did not see any significant increase in overall cell numbers during infection, suggesting that increased bi-allelic expression of CXCR3 are representative of the increased proportion of bi-allelic cells. Next, we determined the contribution of bi-allelic CXCR3 expression to T cell immune responses at the infection site, using a low dose L. mexicana model, which mimics human infection and allows us to characterize T cell infiltration to the skin (26). We observed bi-allelic CXCR3 expressing cells migrating to the skin (Figure 3E). Interestingly, although the percentages of bi-allelic CXCR3 T cells were lower than our previous high dose footpad model, proportions of double positive CXCR3 expressing T cells were greater at the skin than the regional lymph nodes of ear infected mice. These data suggest that a stimulatory micro-environment leads to XCI escape and bi-allelic expression of CXCR3 in T cells.

In order to fully explore the phenotypic consequences of bi-allelic CXCR3 expression in our experimental in vivo infection model, we compared gene expression profiles between mono-allelic and bi-allelic CXCR3 expressing T cells in Cxcr3^GFP/RFP mice infected with L. mexicana. It should be noted that sex associated differences in resistance to L. mexicana in mice have been well documented (27, 28). Female mice are more resistant to infection, and this process is mediated by an increased IFN-γ production relative to male mice (27). Using our CXCR3^GFP reporter mice, we confirmed that female mice are more resistant than males to L. mexicana footpad infection (Figure 4A). Although levels of IFN-γ production in re-stimulated T cells were low, likely due to the suboptimal stimulation conditions, we did not see changes in the overall cell numbers (Figure 4B). Given that CXCR3 is involved in the activation and recruitment of IFN-γ producing Th1 cells, we hypothesized that T cells which display bi-allelic CXCR3 expression during L. mexicana infection produce more IFN-γ compared to mono-allelic CXCR3 expressing T cells, which could account for the increased IFN-γ production and resulting resistance to L. mexicana infection observed in female mice. To test this, mono-allelic and bi-allelic CXCR3 expressing T cells were sorted from spleens and lymph nodes of Cxcr3^GFP/RFP infected mice, and transcripts were measured by real time PCR. Interestingly, we observed increased gene expression levels of Il-2, Ifn-g and Cd69 in bi-allelic CXCR3 expressing T cells compared to mono-allelic CXCR3 expressing T cells (Figure 4C). These genes are associated with increased proliferation, differentiation and activation of Th1 cells. Levels of Il-4, were slightly lower in bi-allelic CXCR3 expressing T cells. T-bet, a Th1 master regulator, known to regulate the expression of CXCR3 in activated CD4+ T cells (11), was significantly increased in CXCR3 positive T cells compared to CXCR3 negative T cells, but we did not observe any difference between mono-allelic and bi-allelic CXCR3 expressing T cells (Figure 4C). Gata-3 and Irf-1 were not affected by the expression of CXCR3 in T cells of Leishmania infected Cxcr3^GFP/RFP mice (Figure 4C). Further, IFN-γ production in re-stimulated bi-allelic CXCR3+ T cells of L. mexicana infected mice were significantly higher than in mono-allelic CXCR3+ T cells (Figure 4D). Finally, expression of activation marker CD44 was increased, while CD62L (secondary lymphoid organ homing receptor) was decreased in bi-allelic CXCR3+ T cells compared to mono-allelic CXCR3+ T cells in infected Cxcr3^GFP/RFP mice (Figure 4E and F). Taken together, our data suggest that XCI escape and bi-allelic expression of CXCR3 in female mice potentially contributes to enhanced Th1 immunity against Leishmania infection. This can explain, in part, why female mice are generally more resistant to Leishmania infection than male mice (15, 16, 27).

Figure 4: Bi-allelic CXCR3 expressing T cells in female mice display a more activated phenotype upon activation during in vivo infection with Leishmania mexicana.

(A) parasitic loads in footpads of L. mexicana infected homozygous CXCR3^GFP female and hemizygous CXCR3^GFP male reporter mice (n=3 per group). (B) IFN-γ production and total cell count in sub-optimally re-stimulated CD3⁺ T cells from lymph nodes of L mexicana infected male and female CXCR3^GFP reporter mice and negative control as determined by flow cytometry (n=3 per group). CD3+ T cells were gated. (C) Real-time PCR analysis of IFN-γ, IL-2. IL-4, CD69, CD25, T-bet, Gata3, and IRF-1 in mono-allelic, bi-allelic and non-CXCR3 expressing T cells in spleens of L. mexicana infected CXCR3^GFP/RFP dual reporter mice. GFP^–/RFP ^–, GFP⁺/RFP^– GFP^–/RFP⁺ and GFP⁺ /RFP⁺ T cells were sorted from infected CXCR3^GFP/RFP mice for gene expression analysis. *p value < 0.05. Error bars represents SEM. (D) Percentage of GFP⁺/RFP^– GFP^–/RFP⁺ and GFP+/RFP+ splenic T cells producing IFN-γ after restimulation. (E-F) Expression of CD62L (E) and CD44 (F) in the spleen of L. mexicana infected CXCR3^GFP/RFP dual reporter mice. *p value < 0.05. Error bars represents SEM.

The results reported here using our Cxcr3^GFP/RFP mice demonstrate that stimulatory microenvironments can affect the allelic expression of CXCR3 in female T cells in vivo. This observation supports the notion that escape from XCI is dynamic, and in addition to genetic elements, environmental factors impact the epigenetic regulatory mechanisms that lead to bi-allelic gene expression. Further research into the specific activation mechanisms that contribute to XCI escape of CXCR3 in different cells, tissues, individuals and disease conditions will provide more insights on sex-associated differences attributable to X-linked genes. Our novel Cxcr3^GFP/RFP mouse will provide an ideal model system to study these mechanisms in vivo. The notion that multiple mechanisms could contribute to sex biased susceptibility to autoimmune and infectious diseases, including the contribution made by XCI escape in X-linked genes is becoming increasingly recognized (29, 30). As demonstrated by our data, even a low frequency of XCI escape can be a potential biologically relevant determinant of disease outcome.

In conclusion, our newly generated CXCR3 bicistronic dual reporter mouse model demonstrates that CXCR3 escapes XCI in T cells when activated in vitro as well as during in vivo infection with L. mexicana. This novel approach to the study of XCI escape of CXCR3 will deepen our understanding of the contribution that bi-allelic CXCR3 expression makes in sex-associated predisposition to autoimmune disease and resistance to infection.

KEY POINTS.

A bicistronic dual reporter mouse enables visualization of allelic CXCR3 expression

CXCR3 escapes X chromosome inactivation (XCI) in T cells in vivo during infection

XCI escape by CXCR3 potentially contributes to enhanced Th1 responses in females