To adjust for the genetic imbalance on the X chromosome between males (XY) and females (XX), one of the X chromosomes is silenced and inactivated in mammalian female diploid cells. This is accomplished by paternal X chromosome imprinting in some lower mammals, which ensures inactivation of the X chromosome from the father side at all times. X chromosome inactivation in higher mammals is accomplished by the process of random X inactivation rather than imprinting. In this process, X inactivation is random, so that either the maternal or the paternal X chromosome is inactivated in a given diploid cell. In contrast to paternal imprinting, random X inactivation has the advantage of minimizing the effects of mutations, if present, on either the paternal or maternal X chromosomes, since diploid cells are mosaic with either paternal or maternal X inactivation [21
]. Therefore, this is more compatible with the natural selection process and is thought to be a logical evolutional alternative to inactivation by paternal imprinting [22
The process of X chromosome inactivation is complex, and remains incompletely understood. The process of X chromosome inactivation starts by counting the number of X chromosomes and allowing X inactivation only if more than one X chromosome is present per diploid cell [23
]. This is followed by selecting which X chromosome will be inactivated. The counting and selection process are regulated by a locus called the X-inactivation center (Xic
), and appears to occur during transient pairing of the two X-inactivation centers in both X chromosomes [24
]. The Xist
genetic locus, located within the Xic
is critical for the X inactivation process. The X chromosome that is selected for inactivation expresses Xist
RNA that accumulates over the chromosome from which it is transcribed [5
RNA expression is followed by recruitment of polycomb group proteins to the Xist
RNA coated X chromosome [22
]. Polycomb group proteins are necessary to maintain X chromosome inactivation throughout cell division. Indeed, the presence of polycomb group proteins such as polycomb repressive complex 1 (PRC1) and polycomb repressive complex 2 (PRC2) triggers a number of epigenetic modification in the X chromosome that result in a heterochromatin that is inaccessible for transcription [22
]. Most notably, the inactive X chromosome is marked by histone H3 methylation at lysine 9 and lysine 27, and hypoacetylation of histones H3 and H4 [26
]. In addition, promoter sequences in the X chromosome are heavily methylated resulting in transcriptional silencing [27
]. Both histone modifications and promoter DNA methylation that characterize the inactive X chromosome are heritable changes that are carried over and maintained throughout cell division.
This DNA methylation suppresses gene expression by a number of mechanisms. These include stearic hindrance of transcription factor binding, and recruitment of methylcytosine binding proteins resulting in an increase in chromatin density [28
]. Methyl CpG binding protein 2 (MECP2) binds methylated DNA and recruits histone deacetylases, which by deacetylating histone tails, increase the charge attraction between methylated DNA and histones, therefore increasing chromatin density resulting in chromatin that is transcriptionally inaccessible.
Therefore, DNA methylation is thought to play a central role in maintaining and stabilizing X chromosome inactivation. Indeed, treating proliferating cells with the DNA methylation inhibitor 5-azacytidine, results in reversal of X inactivation and gene expression from the previously inactivated X chromosome [29
]. These data imply that if DNA methylation is defective, then a potential exists for reactivating the inactive X chromosome, resulting in duplication of gene expression levels of X chromosome genes in females but not in males who only have one X chromosome.
Can a gene dose effect on the X chromosome, therefore, explain the predominance of lupus in females? Indeed, we believe the answer is yes. In fact, DNA methylation is defective in CD4+ T cells from lupus patients [30
]. Do T cells from female lupus patients over-express X chromosome genes compared to male lupus patients and compared to normal females? Before we discuss this, we would like to summarize the literature that supports a central role for defective T cell DNA methylation in the pathogenesis of lupus.
DNA methylation is an enzymatic reaction that refers to the addition of a methyl group to the 5th carbon in the cytosine ring within CG dinucleotide pairs. This process is mediated by a group of enzymes, known as DNA methyltransferases. DNA methyltransferases 3a and 3b are responsible for establishing the pattern of DNA methylation in utero
and are known as de novo
DNA methyltransferases [31
]. On the other hand, the DNA methyltransferase 1 maintains DNA methylation patterns during mitotic cell divisions [32
]. In lupus CD4+ T cells, DNA is globally hypomethylated, resulting from reduced expression of DNA methyltransferase 1 [35
]. Indeed, T cells from active lupus patients over-express a number of methylation sensitive genes such as ITGAL
(perforin), and TNFSF7
]. Promoter sequences of the aforementioned genes are hypomethylated in lupus T cells, and become hypomethylated to the same extent in normal T cells treated with DNA methylation inhibitors such as 5-azacytidine [30
]. Indeed, normal human T cells treated with 5-azacytidine over-express the same genes (CD11a, CD70, and perforin), similar to T cells from active lupus patients [30
], and become autoreactive in vitro
]. T cells treated with 5-azacytidine are capable of killing autologous or syngeneic macrophages [38
], and induce immunoglobulin production in co-culture assays with autologous B cells [40
]. Further, T cells treated with 5-azacytidine results in autoimmunity when adoptively transferred into syngeneic mice [41
]. These mice develop a lupus-like disease characterized by glomerulonephritis, alveolitis, CNS disease, and anti-histone and anti-dsDNA autoantibody production [41
]. Studies from the MRL/lpr lupus-prone mouse also demonstrate defective DNA methylation and reduced Dnmt1 expression in CD4+ T cells, and over-expression and hypomethylation of CD70 gene [42
]. These data powerfully support an important role for defective T cell DNA methylation in the pathogenesis of lupus.
CD40L is a T cell co-stimulatory molecule that plays an important role in T cell-B cell interaction. Indeed, the interaction between CD40L on T cells and CD40 on B cells is important for immunoglobulin class switching [43
]. Further, CD40L-CD40 ligation is involved in B cell proliferation, germinal center formation, and increasing the sensitivity of B cells to cytokine stimulation. CD40L is over-expressed on both T cell surface and in the serum in lupus patients [44
]. Serum CD40L levels correlate with both anti-dsDNA production and disease severity in lupus patients [46
]. The gene encoding for CD40L is located on the X chromosome. Gene dose balance of CD40L expression in normal males (46,XY) and females (46,XX) is maintained by random X chromosome inactivation in female cells. This is maintained by epigenetic changes including heavy DNA methylation of the inactive X chromosome as discussed above. In fact, the CD40L promoter sequence is demethylated in normal male CD4+ T cells, while females have one methylated and one demethylated alleles, reflecting random X chromosome inactivation (as reflected by heavy methylation) in one of the two X chromosomes [45
]. When normal CD4+ T cells are treated with the DNA methylation inhibitor 5-azacytidine, the normally methylated CD40L allele demethylated and the expression of CD40L is doubled in T cells from females compared to males [45
The normal physiology described for CD40L is disrupted in patients with SLE, however. In CD4+ T cells from active lupus patients, both CD40L alleles are demethylated and CD40L is expressed as twice as much as compared to male lupus patients, or normal female CD4+ T cells [45
]. This suggests an explanation for the gene dose effect observed on the X chromosome in lupus patients; and, therefore, may be part of the explanation for the higher prevalence of lupus in patients with Klinefelter’s syndrome that is discussed earlier. Since DNA methylation, and therefore X inactivation is defective in lupus patients, male patients with Klinefelter’s syndrome (47,XXY) have two CD40L alleles that are available for transcription, similar to female patients (46,XX). By this mechanism, males with Klinefelter’s syndrome (47, XXY) may be at a risk of SLE similar to females, and considerably higher than in karotypically normal males (46,XY).
Another example for an X chromosome gene copy number excess resulting in autoimmunity is the duplication of TLR7 in mice. In BxSB mice, the presence of the Yaa
genetic locus on the Y chromosome is implicated in accelerated autoimmunity in male mice [47
]. Recently, it was discovered that the Yaa
locus is a 4-megabase genetic region at the telomeric end of X chromosome that is translocated onto the Y chromosome in Yaa
bearing mice [48
]. Therefore, BxSB mice have two copies available for transcription of this genetic element – one on the X chromosome and the other on the Y chromosome; whereas, females have one copy since one X chromosome is randomly inactivated. The gene in this locus that is responsible for the autoimmune phenotype acceleration in male BxSB mice was confirmed to be TLR7. Indeed, the expression of TLR7 was increased by twofolds in CD19 B cells from Yaa
strains compared with B6 controls [49
]. To confirm that TLR7 duplication is responsible for the Yaa
effect, transgenic mice that over-express TLR7 were generated and showed that increased TLR7 dosage was sufficient to induce autoimmunity [50
In human lupus, however, TLR7 duplication has so far not been evident in a study of 99 unrelated SLE patients [51
]. We have studied 22 SLE men from 11 families in which at least 2 men have SLE. So, these men all share a Y chromosome with one other man with SLE. We did not find a translocation of a segment of the X chromosome containing the TLR7 gene on to the Y chromosome in these men with familial male SLE. Thus, a yaa
gene equivalent among men with is either not present at all, or is very uncommon.