«913–922, 2008 doi:10.1016/j.berh.2008.09.002 available online at 9 The X in sex: how autoimmune diseases revolve ...»
Best Practice & Research Clinical Rheumatology
Vol. 22, No. 5, pp. 913–922, 2008
available online at http://www.sciencedirect.com
The X in sex: how autoimmune diseases
revolve around sex chromosomes
Carlo Selmi * MD, PhD
Department of Internal Medicine, IRCCS-Istituto Clinico Humanitas, University of Milan, Rozzano, Milan, Italy
Division of Rheumatology, Allergy, and Clinical Immunology, University of California at Davis School of Medicine, California, USA Recent estimates suggest that autoimmune diseases cumulatively affect 5–10% of the general population worldwide. Although the etiology and pathogenesis remain poorly understood in most cases, similarities between diseases outnumber differences in the initiation and perpetua- tion of the autoimmune injury. One major example is the predominance of affected women, and perhaps its most intriguing putative mechanism is related to sex chromosomes, based on the recent observation that women with autoimmune diseases manifest a higher rate of circulating leukocytes with a single X chromosome. In a complementary fashion, there have been several reports on a role of X chromosome gene dosage through inactivation or duplication in autoim- munity. It is important not to overlook men with autoimmune diseases, who might manifest a more frequent loss of the Y chromosome in circulating leukocytes. Taken together, sex chro- mosome changes might constitute the common trait of autoimmunity.
Key words: DNA methylation; female predominance; X chromosome inactivation; X monosomy.
AUTOIMMUNITY AT A GLANCEAutoimmune diseases are believed to affect 5–10% of the general population1 and are a signiﬁcant cause of morbidity and mortality worldwide. An estimate from the US National Institutes of Health suggests over 10 million cases of autoimmune dis- eases in the US2, which results in direct economic costs almost twice those of cancer.
There are currently tens of conditions in which some autoimmune features (most commonly serum autoantibodies) are encountered, although only for very few is an * Department of Internal Medicine, University of Milan, IRCCS-Istituto Clinico Humanitas, Via Manzoni 113, 20089 Rozzano, Milan, Italy. Tel.: þ39 02 5032 3088; Fax: þ39 02 5032 3089.
E-mail address: firstname.lastname@example.org 1521-6942/$ - see front matter ª 2008 Elsevier Ltd. All rights reserved.
914 C. Selmi autoimmune pathogenesis established.3 Speciﬁc age at onset, target tissue, and pat- terns of epidemiology characterize each autoimmune disease, yet several features are common throughout the majority of the autoimmunity spectrum. Among these, the recognized role for a permissive genetic background in determining individual susceptibility and the presence of a signiﬁcant sex imbalance4,5 are common to most conditions, as illustrated in Table 1, in which sex ratios and genetically identical twin concordance rates are reported.
SHOULD WE SAY ‘WHY WOMEN?’ The fact that approximately 80% of patients with autoimmune diseases are women has been known for a long time, and the causes and mechanisms for this sex imbalance are based on several hypotheses proposed over the past years.6 These include sex hormones and reproductive history, environmental factors, fetal microchimerism, a skewing in X-chromosome inactivation patterns, and major defects in the sex chromosomes.
Estrogens can inﬂuence lymphocyte maturation, activation, and synthesis of antibodies and cytokines7–10; estrogen receptor ligands modulate antigen presentation.11 These observations were used as a rationale for the comparison of sex-hormone proﬁles, parity, or the use of hormonal treatments between women with autoimmune disease and controls, but reported data were negative or inconclusive. The onset at different ages (and different times in the reproductive history) and the lack of consistent effects of estrogen replacement therapies on disease course have further weakened the sex hormone theory.
The fact that environmental factors are more commonly associated with women has also been suggested to play a role in autoimmunity initiation. Both a permissive genetic background and an environmental trigger have been suggested as being necessary (but neither on its own being sufﬁcient) to induce autoimmune diseases12, possibly through additional mechanisms such as apoptosis.13 Factors such as speciﬁc xenobiotics (as in cosmetics) or bacteria (possibly related to female-predominant urinary tract infections) are ideal candidates in the environmental model but solid evidence of a causative role is awaited. In the paradigmatic case of primary biliary cirrhosis, an organ-speciﬁc autoimmune disease, both these factors have been suggested, among others, by epidemiological or experimental data14–16, but no mechanistic relation can be assumed at the present status of knowledge, despite the most recent xenobiotic-induced murine models.17 The role of sex chromosomes in autoimmune diseases has been widely studied in the past decade, ﬁrst with the suggested model of fetal microchimerism, subsequently with X-inactivation patterns, and ﬁnally with X-chromosome monosomy and duplication. Fetal microchimerism was ﬁrst suggested based on the observation that most autoimmune diseases have their peak of incidence following menopause. Indeed, maternal and fetal cells are exchanged during pregnancy, leading to fetal cell persistence (i.e. microchimerism) in the mother. Chimeric fetal cells are often hematopoietic and can differentiate into somatic cells in multiple organs, potentially acting as targets for autoimmunity and resembling graftversus-host disease after stem-cell transplantation. Microchimeric cells were ﬁrst detected in peripheral blood mononuclear cells from patients with systemic sclerosis18, but other authors have failed to reproduce these ﬁndings.19 No signiﬁcant difference was found in the frequency of male microchimerism between women with primary biliary cirrhosis and controls.20 Cumulatively, available data on the role of fetal microchimerism in autoimmunity are weak or inconclusive, as naturally acquired fetal and maternal microchimerism is not uncommon in healthy women.21 The available data and proposed theories on sex chromosome biology and their putative role in autoimmune diseases will be discussed herein, along with the potential implications and limitations of such observations.
THE OLD AND NEW BIOLOGY OF SEX CHROMOSOMES
Female cells carry both parental X chromosomes (maternal and paternal), whereas male ones carry only the maternal X. In the majority of mammals, females are functional mosaics for X-linked genes, due to the phenomenon of X chromosome inactivation. Random inactivation patterns occur in somatic cells to achieve equivalent expression levels between sexes (i.e. dosage compensation), as an evolutionary compensatory mechanism. In nonhuman organisms, such as ﬂies and worms, dosage compensation is achieved by up- or down-regulation of X-linked 916 C. Selmi expression, respectively, while others do not manifest dosage compensation.22 A nonrandom X chromosome inactivation can be encountered as a stochastic event in nonpathological conditions, but skewing might also be secondary to X-linked mutations that cause the mutated chromosome to be predominantly inactivated. Until the appearance of recent data (discussed below), X inactivation patterns were analyzed by means of the androgen receptor gene. The methylation of this gene (which was found not to escape inactivation) is considered representative of the inactive X and is commonly studied by taking advantage of a methylation-sensitive cleavage site. Using this approach, it was reported that 16% of healthy women over 50 years of age manifest a skewed (i.e. 75:25 ratio) or extremely skewed (i.e.
90:10) X-chromosome inactivation pattern23,24, yet in the majority of cases this does not lead to a clinically relevant phenotype.25 One possible exception to this latter observation is that an extremely skewed pattern has the potential to unmask unfavorable X-linked alleles carrying mutations, thus leading to disease onset in heterozygous subjects.
The classic view of X-chromosome inactivation was recently undermined by the report of new data on additional players such as small, noncoding RNA; antisense transcription; and chromosome-wide chromatin changes. Yet DNA methylation remains the most studied phenomenon of gene silencing and X-chromosome inactivation. In particular, a recent study demonstrated that the inactive chromosome manifests the lower degree of DNA methylation26, somehow in conﬂict with the previous data on X inactivation and reversing the established relationship between overall methylation and expression potential as such enhanced methylation is concentrated in the gene body, rather than the promoter.26 More importantly, it is now established that these mechanisms do not cause an entire X chromosome to be silenced, as 10– 15% of X-linked genes escape silencing and are expressed from both X chromosomes in a variable proportion of healthy women.27 Taken altogether, these ﬁndings have radically changed our understanding of the biology of X chromosomes and allow speciﬁc genes to achieve double or a null expression in physiological conditions. The existence for some of these genes of a Y chromosome homologues makes the hypothesis ideal to explain the occurrence of autoimmune diseases also in men, as discussed below.
SEX CHROMOSOMES AND (AUTO)IMMUNITY
Clinical and experimental evidence has strengthened the link between sex chromosomes and immunity. From a clinical standpoint, we have now established that the immunodysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) and the Xlinked recessive severe combined immunodeﬁciency (XSCID) syndromes are caused by mutations of the transcription factor fork head box P3 (FoxP3) on Xp11.23,28 and by interleukin 2 receptor gamma chain on Xq13.129, respectively. Similarly, the hyperIgM syndrome is caused by defects of CD40 ligand on Xq26.30 This latter condition also provides a link between the X chromosome and autoimmunity, because it is often accompanied by autoimmune features, e.g., in Wiskott–Aldrich syndrome, an inherited X-linked, recessive disease due to mutations in the WAS gene (Xp11.
4-p11.21), sharing immune dysregulation and autoimmune manifestations.31 Furthermore, it is of particular interest that conditions characterized by major abnormalities of the X chromosome, such as Turner syndrome32 or premature ovarian failure33 are characterized by autoimmune comorbidities in a signiﬁcant proportion of cases, and might thus provide some helpful hints in female-predominant autoimmunity such as autoimmune cholangitis.34 The X chromosome and autoimmune diseases 917
X-chromosome silencing in autoimmunity
Based on the illustrations presented thus far, can we hypothesize that a skewed Xchromosome inactivation pattern causes tolerance breakdown and autoimmunity development? Can X-chromosome inactivation explain the somehow disappointing concordance rates for all autoimmune diseases in monozygotic twin sisters? As in other ﬁelds of autoimmunity, systemic lupus erythematosus (SLE) was the ﬁrst condition to be tested for de novo hypotheses. Indeed, the inactivation issue was suggested several years ago but data were not conclusive in SLE35,36 or in other autoimmune diseases37, although in the latter case not all included conditions were female predominant. The same case-control study design and androgen-receptor-based molecular approach has also been utilized in multiple sclerosis but failed to demonstrate signiﬁcant differences between patients and controls.38 In an elegant study, the X-linked CD40 ligand gene was found to be hypomethylated in women with SLE39, in contrast with previous reports of overexpression in peripheral lymphocytes.40 Through a broader approach, we note that the pharmacological inhibition of DNA methylation leads to the development of lupus-like disease in animal models41, and that DNA overall methylation is reduced in peripheral leukocytes from patients with SLE42, although the signiﬁcance of these ﬁndings in disease etiology remains uncertain.43 In recent years, four female-predominant, late-onset, often co-existing autoimmune diseases characterized by different tissue speciﬁcity have been studied for X-chromosome inactivation skewed patterns. First, it was reported that a signiﬁcant (30–49%) proportion of women with systemic sclerosis manifest an extremely skewed X inactivation44,45, yet it should be noted that the authors utilized only the androgen receptor as a marker for X inactivation, and that additional familial samples have determined that the maternal X is more frequently silenced.45 Similar ﬁndings were obtained in a Danish cohort of control women and monozygotic twins with autoimmune thyroid disease.46 Whereas a skewed X-chromosome inactivation was signiﬁcantly more common in twins than in unrelated women, we note that patterns were associated with the risk of disease in discordant twins. Similar associations were later independently recapitulated in smaller case-control studies in autoimmune thyroid disease47,48 and Sjogren syndrome49, but not in primary biliary ¨ cirrhosis.50 We should note that these negative data were obtained in a large series of patients and matched controls and utilized the androgen receptor and three additional markers to map the X chromosome inactivation for the entire chromosome.
More gene dosage in autoimmunity
Further evidence on the role of X-chromosome gene dosage in the development of autoimmunity was provided with the study of X monosomy (i.e. loss of one X chromosome) and subjects with Klinefelter syndrome (i.e. with a XXY karyotype). We recently reported (Table 2) that women with organ-speciﬁc and systemic femalepredominant, late-onset autoimmune diseases, such as primary biliary cirrhosis, systemic sclerosis, and autoimmune thyroid disease, manifest higher frequencies of peripheral blood cells with X monosomy than age-matched women.51,52 Interestingly, this observation was not replicated in SLE, as a paradigmatic disease characterized by earlier age of onset53, and was not secondary to circulating male cells (i.e. fetal microchimerism).51,52 Based on these intriguing observations, we ﬁrst proposed54,55 that the enhanced X-chromosome monosomy might cause haploinsufﬁciency of X-linked genes escaping 918 C. Selmi