Sex Matters: Hormonal and Chromosomal Determinants of Autoimmunity and Anti-Cancer Immunity Across the Lifespan.

While differences in sex hormone levels underlie many sex differences in immunity, differences in sex chromosome complement also drive sexual dimorphism in the immune response. Most often, sex chromosomes consist of XX for females and XY for males, and this distinction has multiple downstream consequences. For instance, despite mechanisms in place to equalize dosage of X chromosomes between males and females, multiple sex chromosome‐linked genes are differentially expressed in female vs. male immune cells. Here we will review recent findings on how sex chromosomes contribute to sexual dimorphisms observed in anti‐cancer immunity and autoimmunity. These mechanisms include (1) the expression of Y chromosome genes in male cells; (2) loss of Y chromosome (LOY) that occurs in male cells; (3) expression of genes from the inactive X (Xi) chromosome in female cells, and (4) the expression of XIST by female cells. These mechanisms will be discussed in more detail below. The structure of the Y chromosome is quite complex and full of base pair repeats. As a consequence, the Y chromosome has been difficult to fully sequence and assemble, resulting in it being the last human chromosome to be completed. This recent achievement has contributed to our overall understanding of the Y chromosome and the genes it possesses [ 115 ]. While the Y chromosome is much smaller and contains fewer genes than the X chromosome, it contains 17 homologs to X chromosome genes (Figure  3B ). These homologs (e.g., KDM5D, KDM6C, DDX3Y ) encode for proteins with critical functions, and their X‐linked counterparts are often expressed from both the active X chromosome (Xa) and inactive X chromosome (Xi) in females (i.e., escape inactivation) [ 116 ]. Genes expressed from both the Xi and Xa (XCI ‘escapees’) in female cells and homologous X and Y gene pairs. (A) Xi and Xa chromosomes with known immune‐related and Y homolog genes that escape XCI labeled. (B) Representative X and Y chromosomes with homologs labeled that are known to escape XCI in females. The Y‐linked gene Kdm5d , for instance, encodes for an epigenetic regulator, and its X‐linked counterpart ( Kdm5c ) is expressed from both the Xa and Xi in females. Recently, Kdm5d was found to be a key driver of colorectal carcinoma (CRC) outcome and metastasis [ 117 ]. Deletion of Kdm5d improved tumor outcomes in mouse models, in part through indirect effects on the CD8 + T cell compartment. CD8 + T cells exposed to Kdm5d deficient cancer cells showed more efficient killing. Furthermore, constitutive expression of Kdm5d in cancer cells resulted in decreased CD8 + T cell infiltration. These findings suggest an immune mechanism by which expression of Y‐linked Kdm5d in CRC cells can contribute to increased metastases and mortality in males. Male‐specific Y linked genes [e.g., sex determining region y (Sry)], which do not have X‐linked counterparts, may also underlie sex differences in anti‐cancer immunity. For instance, transcription factor Sry has been implicated in sex differences in cancer recurrence. Primary liver cancers occur more often in males after surgery, in both patients and mice [ 118 ]. Male mice had fewer CD8 + T cells and increased polymorphonuclear‐myeloid derived suppressor cells (PMN‐MDSCs) consistent with an unfavorable anti‐tumor immune response. Moreover, hepatocyte‐specific Sry overexpression in mice was found to be associated with postsurgical liver cancer recurrence whereas hepatocyte‐specific Sry deletion was protective. These findings implicate Sry in suppressing anti‐cancer immunity and unfavorable cancer outcomes. How Sry functions in hepatocytes to mediate these immune effects, however, will require further study. Clonal mosaicism can occur in cells over time and accumulate in individuals with aging. The most common form of clonal mosaicism is loss of Y (LOY) chromosome in male cells (Figure  4 ), which is detectable in 20% of male peripheral leukocyte samples in the UK Biobank and 40% of samples in males over the age of 70 years [ 119 ]. Mosaic LOY (mLOY) in circulating leukocytes has been associated with multiple disease states including increased mortality, cardiovascular events, and other age‐related conditions [ 120 ]. While these associations are well established, a recent study has also demonstrated that mLOY is causal in promoting cardiac fibrosis with aging [ 121 ]. In a mLOY mouse model, LOY macrophages showed aberrant differentiation and increased pro‐fibrotic transforming growth factor (TGF)‐β production in the myocardium. Thus, LOY immune cells have a pathogenic capacity in age‐associated disease. Consequences of LOY. LOY in immune cells can lead to dysfunction or exhaustion as seen in T cell subsets, overall partial or complete loss of important regulatory genes such as Kdm5d and Uty , and the further propagation of LOY. Together, these result in an increase in cancer incidence and poor prognosis for various malignancies. Emerging evidence also suggests that LOY may confer pathogenic capacity to immune cells beyond myeloid cells. A recent analysis of scRNA‐seq datasets from liver, lung, and colorectal cancer patients revealed that regulatory T cells (Tregs) have a marked increase of LOY and comprise a higher proportion of CD4 + T cells compared to other T cell subtypes [ 122 ]. LOY Tregs exhibit alterations in the expression of multiple genes, with many genes having known immune functions. How LOY may affect Treg suppressive function and the mechanisms by which these changes can alter cancer susceptibility remain to be determined. While these findings suggest that LOY in immune cells may have effects on their function and phenotype, recent work has also highlighted an indirect mechanism by which LOY in cancer cells impacts immune cell function [ 123 ]. In bladder cancer, LOY in cancer cells was associated with increased tumor burden in immune‐competent mice. Remarkably, this increased tumor burden was not due to increased growth by LOY cancer cells, since LOY cancer lines did not proliferate more than control cancer lines. Moreover, faster growth was also not seen when LOY cancer cells were implanted into immunodeficient Rag2 −/− Il2rg −/− mice, suggesting that increased tumor burden was due to increased efficiency of LOY tumors in evading anti‐tumor immunity [ 123 ]. Y chromosome genes Kdm5d and Uty were found to be crucial to tumor growth, as single knockouts of each in cancer cells significantly increased tumor growth while overexpression diminished tumor size. LOY tumor cells displayed more dysfunctional and exhausted CD8 + T cells (i.e., higher levels of TOX expression), suggesting that LOY in cancer cells alters T cell function in the surrounding tumor microenvironment. More recently, LOY in cancer cells has been proposed to be linked to LOY in neighboring T cells, and this concurrent LOY impacts cancer outcomes [ 124 ]. A comprehensive analysis of bulk‐ and single‐cell RNA sequencing datasets from 29 different human cancers revealed LOY not only in malignant cancer cells but also in tumor stromal and immune cells [ 124 ]. LOY had the greatest impact on CD8 + and CD4 + T cell transcriptomes, increasing immunosuppressive and exhaustion signatures while decreasing cytotoxicity signatures. While more studies are needed to fully understand the mechanisms by which LOY spreading might occur, concurrent LOY in benign and malignant cells was associated with poor survival in cancer patients. In addition to anti‐cancer immune responses, LOY in circulating immune cells has also been implicated in predisposition to autoimmunity. Previous studies have noted associations between autoimmune thyroiditis in male patients and LOY in peripheral blood [ 125 ]. However, the mechanism by which LOY may enhance autoimmunity is not currently clear. Interestingly, polymorphisms in Y chromosome genes have been reported to alter autoimmunity predisposition in multiple mouse models of autoimmunity and to modify sexual dimorphism seen in these models, suggesting possible mechanisms by which LOY may enhance autoimmune capacity [ 126 ]. Females undergo a process known as X chromosome inactivation (XCI) to achieve X‐linked gene dosage compensation. This process is driven by the long non‐coding RNA (lncRNA) Xist, which initiates the process of XCI by coating the randomly selected future Xi and propagating epigenetic and transcriptional changes that result in a silenced X chromosome [ 127 ]. Importantly, however, 20%–30% of X chromosome genes in humans are still expressed at varying degrees from the Xi [ 128 ]. As a consequence, genes that are expressed from both the Xa and Xi in females (also known as XCI “escapees”) have sexually dimorphic expression, with higher expression in XX females compared to XY males (Figure  3A ). Although to a lesser extent, biallelic expression of X‐linked genes also occurs in mouse cells, with about 3%–7% of genes exhibiting expression from the Xi [ 129 ]. Thus, biallelic expression of X‐linked genes in females is a conserved mechanism underlying sexual dimorphism across species. Among the XCI “escapees” in humans and mice are epigenetic regulators that play critical roles in immune cell function and may underlie sexually dimorphic responses in autoimmunity. For instance, the histone 3 lysine 27 (H3K27) demethylase Kdm6a (which encodes UTX) has been implicated as an XCI escapee whose higher expression level in female CD4 + T cells and microglia is responsible for female bias in the incidence of MS [ 130 , 131 ]. Interestingly, deletion of UTX in microglia protected female mice in a mouse model of but did not have an impact on male mice. In addition to MS, UTX has also been implicated in regulating immune cell function in other female‐biased conditions. In SLE, for example, pharmacological inhibition of UTX (with GSKJ4) in SLE patient monocytes resulted in decreased interferon‐stimulated gene (ISG) expression [ 132 ]. Similarly, GSKJ4 treatment in a resiquimod (R848) mouse model of SLE resulted in decreased production of autoantibodies as well as lower ISG expression. While these findings suggest a potential role for UTX, it is important to note that GSKJ4 is not a specific inhibitor of UTX but instead is an inhibitor of the histone demethylase activity of both UTX and JMJD3 (encoded by Kdm6b ). Thus, more specific approaches for inhibiting UTX are needed to attribute these findings to UTX. Additionally, in a mouse model of colitis, T cell specific deletion of UTX protected mice from the development of autoimmune colitis [ 133 ]. Together, these recent findings support the idea that higher levels of UTX in females can lead to female‐biased incidence of autoimmune diseases. At the same time, a recent report suggests an opposing role for the X‐linked histone 3 lysine 4 (H3K4) demethylase Kdm5c (which encodes Jarid1c) [ 134 ]. Similar to UTX, Jarid1c is expressed from both the Xa and Xi in humans and mice. Doss et al. utilized a male‐biased adoptive transfer model of EAE to show that overexpression of Jarid1c in Th17 cells protected against disease transfer [ 134 ]. These findings suggest that Jarid1c negatively regulates pathogenic Th17 cells in this model and is therefore protective in Th17‐mediated autoimmune conditions. Indeed, although the incidence of MS is higher in females, males with MS have a more severe disease course. Thus, these findings implicate lower Jarid1c expression in male Th17 cells as a mechanism underlying increased MS severity in males. In addition to autoimmunity, accumulating evidence suggests a key role for XCI escapees in anti‐cancer immunity. Dysregulation of XCI, inferred using an X‐Reactivation (X‐Ra), in circulating monocytes, peripheral blood, and other cell types has been correlated with triple‐negative breast cancer and worse survival [ 135 ]. Additionally, specific XCI escapees have been implicated in the control of anti‐cancer immunity. For example, UTX has been implicated in the sex‐biased immune response against glioblastoma [ 136 ]. Female mice had higher survival in a mouse model of glioblastoma due to a more robust T cell response. Specifically, females had fewer progenitor exhausted T cells (CD8 + CD44 + PD1 + TCF1 + TIM3 − ) and increased effector T cells (CD8 + CD44 + TCF1 − TIM3 − ). Treatment of mouse CD8 + T cells with the pharmacological inhibitor GSKJ4 led to decreased expression of IFNγ and increased expression of the exhaustion marker TIM3, suggesting that higher UTX expression may underlie the more robust anti‐cancer T cell response in females. As noted above, however, GSKJ4 inhibits the histone demethylase activity of both UTX and JMJD3, so a more specific approach to UTX inhibition is needed to define UTX's role in the female‐biased immune response against glioblastoma. In support of a role for UTX in promoting anti‐cancer immunity, T cell‐specific UTX deletion has been reported to decrease the immune response against colon cancer [ 137 ]. Additionally, multiple studies have now reported that UTX mutations in cancer cells can alter tumor‐infiltrating immune cell populations. In bladder cancer, for instance, low UTX expression due to UTX mutations in cancer cells was correlated with decreased immune cell infiltration in the tumor [ 138 ]. Moreover, UTX‐deficient bladder cancer cells have also been reported to increase production of cytokines important in the polarization of macrophages to the M2 lineage [ 139 ]. Thus, UTX may alter anti‐cancer immunity both through its direct effects in T cells and through its indirect effects in cancer cells. As discussed above, the long noncoding RNA (lncRNA) Xist directs XCI in female cells. Consistent with this role, perturbation of Xist in B cells released X‐linked genes (i.e., TLR7 ) from inhibition. As a consequence, Xist deletion in B cells was associated with increased formation of pathogenic CD11c + atypical B cells and a higher incidence of SLE and RA [ 140 ]. In addition, mild global Xist deficiency was induced in mice due to a mutation in Ftx , an X inactivation center gene. These Xist‐deficient mice also reactivated gene (i.e., TLR7) expression from the Xi in immune cells, with subsequent dysregulation in B cells, dendritic cells, and monocytes [ 141 ]. Mild Xist‐deficiency in mice led to the spontaneous development of inflammatory signals and symptoms consistent with SLE. These studies highlight the importance of Xist‐mediated TLR7 repression in maintaining tolerance and preventing systemic autoimmune conditions such as SLE. In T cells, Xist has been shown to dynamically regulate XCI with stimulation [ 142 ]. In unstimulated T cells, the Xi was found to be transcriptionally silent with an enrichment of repressive histone marks. After stimulation, T cells acquired epigenetic modifications on the Xi that poised certain genes for expression. Xist localization was found to be dependent on NF‐κB signaling, as both treatment with NF‐κB inhibitors and a genetic murine T cell‐specific knockout of NF‐κB kinases resulted in decreased Xist cloud formation. Interestingly, this appears to be lymphocyte specific, as similar reductions in Xist cloud formation were not observed in a murine embryonic fibroblast line. Taken together, these findings demonstrate the importance of Xist in repressing X‐linked genes in immune cells and define the molecular underpinning of Xist‐mediated repression of X‐linked genes. Notably, Xist has recently been shown to have roles beyond XCI. For instance, Xist has been shown to dampen gene expression not only on the X chromosome but also on autosomes [ 143 ]. Xist RNA antisense purification (RAP‐seq) was used to identify autosomal targets of Xist, and the expression of these autosomal gene targets was lower in females than males. Because these findings were made in human pluripotent stem cells (hPSCs) and murine PSCs, it remains unclear whether Xist also regulates autosomal gene expression in immune cells. If shown empirically to occur, this non‐canonical Xist function may further explain sex differences in immunity through alterations in autosomal gene expression. In addition to autosomal gene regulation, other “non‐canonical” Xist functions have now been delineated which may contribute to immune activation (Figure  5 ). First, Xist RNA was found to be a TLR7 ligand that can stimulate pDCs in patients with SLE [ 144 ]. Xist contains a motif (UU‐containing sequence) that TLR7 preferentially binds to, leading to its induction. Transfection of pDCs with the sequence of Xist containing the TLR7 agonist motif led to robust IFNα production. Depletion of Xist significantly reduced the amount of TLR7 ligands present in cells. Furthermore, Xist expression was shown to be higher in SLE patient PBMCs compared to healthy controls. Canonical and noncanonical functions of Xist. (A) Canonical Xist function of coating the Xi and enacting the process of XCI. Absence of Xist has been shown to lead to increased autoimmunity through increased TLR7 expression from the Xi. (B) Noncanonical functions of Xist. Xist has been shown to dampen autosomal gene expression and function as a TLR7 agonist to promote incidence of autoimmunity. Xist can also form RNP complexes (with proteins such as Spen, PRC1, DHX9, and SSB) that are antigenic and are targeted by autoantibodies in systemic autoimmunity. Second, Xist complexes with multiple ribonucleoproteins (RNPs), and these RNPs serve as autoantibody targets to promote autoimmunity in a SLE mouse model [ 145 ]. In SJL/J male mice, inducible expression of Xist was associated with higher autoantibody production and more severe lupus‐like disease. Furthermore, the T and B cell compartments of male mice expressing Xist were reprogrammed to have decreased immune modulatory programs and increased gene signatures associated with female mice.

This work was supported by the National Institute of Allergy and Infectious Diseases (Grants 5R01AI174519‐03, 5R01AI174519‐03S1, 5U19AI181729‐02).

Much progress has been made in recent years in understanding the biological underpinnings of sex differences in immunity. A major development has been an appreciation for sexually dimorphic changes that occur across the lifespan. Understanding the effects on immune responses of hormonal changes that occur over time, for instance, is an important step toward optimizing immune health in individuals undergoing menopause or andropause. Additionally, a recognition that somatic mosaicism accumulates with age and plays a role in altering the immune response is also an important step toward improving the immune response in cancer and tolerance toward self‐antigens with aging. Finally, with the expanding number of immunotherapies being utilized in the treatment of cancer and autoimmunity, how chromosomal complement and hormonal levels interact with these immunotherapies is a subject of much interest. A deeper understanding of these interactions will allow the maximization of clinical benefit while minimizing adverse outcomes in individuals eligible for these therapies.

The immune response to cancer and autoimmunity are closely linked. This is illustrated by the high rates of autoimmune side effects that occur with treatments that boost anti‐cancer immunity [ 1 ]. Approximately 40% of cancer patients treated with immune checkpoint immunotherapies developed immune related adverse events (IRAEs). Moreover, better cancer outcomes were reported in patients who developed IRAEs and better outcomes with reduced incidence of melanoma in individuals who developed vitiligo [ 2 , 3 , 4 ]. In both anti‐cancer immunity and autoimmunity, sex plays a critical role as a biologic variable in determining incidence and outcomes. Here we will review recent progress in understanding hormonal and chromosomal differences that underlie sex differences in autoimmunity and anti‐cancer immunity. Because hormone levels change and chromosomal changes accumulate throughout the lifespan, the impact of age on sex differences is also discussed in this review. Understanding these differences will pave the path to developing new approaches for modulating the immune response to benefit patients with cancer, autoimmunity, and both cancer and autoimmunity.

The authors declare no conflicts of interest.