Sex Differences in Correlation with Gene Expression Levels between Ifi200 Family Genes and Four Sets of Immune Disease-Relevant Genes

Background The HIN-200 family genes in humans have been linked to several autoimmune diseases—particularly to systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA). Recently, its human counterpart gene cluster, the Ifi200 family in mice, has been linked to spontaneous arthritis disease (SAD). However, many immune-mediated diseases (including RA and SLE) show gender difference. Understanding whether or not and how these genes play a role in sex difference in immune-mediated diseases is essential for diagnosis/treatment. Methods This study takes advantage of the whole genome gene expression profiles of recombinant inbred (RI) strain populations from female and male mice to analyze potential sex differences in a variety of genes in disease pathways. Expression levels and regulatory QTL of Ifi200 family genes between female and male mice were first examined in a large mouse population, including RI strains derived from C57BL/6J, DBA/2J (BXD), and classic inbred strains. Sex similarities and differences were then analyzed for correlations with gene expression levels between genes in the Ifi200 family and four selected gene sets: known immune Ifi200 pathway-related genes, lupus-relevant genes, osteoarthritis- (OA-) and RA-relevant genes, and sex hormone-related genes. Results The expression level of Ifi202b showed the most sex difference in correlation with known immune-related genes (the P value for Ifi202b is 0.0004). Ifi202b also showed gender difference in correlation with selected sex hormone genes, with a P value of 0.0243. When comparing coexpression levels between Ifi200 genes and lupus-relevant genes, Ifi203 and Ifi205 showed significant sex difference (P values: 0.0303 and 0.002, resp.). Furthermore, several key genes (e.g., Csf1r, Ifnb1, IL-20, IL-22, IL-24, Jhdm1d, Csf1r, Ifnb1, IL-20, IL-22, IL-24, and Tgfb2 that regulate sex differences in immune diseases) were discovered. Conclusions Different genes in the Ifi200 family play different roles in sex difference among dissimilar pathways of these four gene groups.


Background
Recently, Ifi200 genes, including the human HIN-200 gene cluster and its mouse counterpart, the interferon inducible-200 (Ifi200) family, have been linked to several autoimmune diseases [1][2][3][4][5]. These genes have been linked to SLE. As early as in 1994, IFI16 was recognized as a target of antinuclear antibodies in patients with SLE [6]. The Ifi200 family genes as modifiers for the SLE susceptibility have been nicely summarized in a review by Choubey [7]. A recent study suggests that anti-IFI16 antibodies hold the potential to serve as a new biomarker of disease activity in SLE [8]. IFI16 is in the same paralogue with pyrin and is an HIN domain family member 1 (PYHIN1 or IFIX), myeloid cell nuclear differentiation antigen (MNDA), and absent in melanoma 2 (AIM2) on human chromosome 1 as a human HIN-200 gene cluster. Similar function was found for other genes in this cluster [4,6]. AIM2 was found to facilitate apoptotic DNA-induced SLE via arbitrating macrophage functional maturation [9]. These genes are also linked to RA. In particular, circulating IFI16 has been found to correlate with clinical and serological features in RA [1]. In their report, Alunno et al. showed that high levels of circulating IFI16 in RA are more frequent in RF/anti-CCP-positive RA patients and significantly associated with pulmonary involvement. Most recently, a study using the IL-1RA-deficient mouse model found that decreased expression levels of Ifi genes were associated with increased resistance to SAD [10]. In addition, previously, interferon-inducible protein-10 (IP-10) was found to be linked to RA, although it is not located in the human HIN-200 cluster [11,12]. These findings suggest that it is possible that Ifi200 family genes may play an important role in the development of arthritis.
Both SLE and RA have increased prevalence in women compared to men. Whether the function of Ifi200 genes shows gender (or sex) difference is not completely understood. A few studies have been performed using animal models. Panchanathan et al. reported that cell type-and gender-dependent factors differentially regulate the expression of the AIM2 and p202 proteins-thus, suggesting opposing roles for these two proteins in innate immune responses in SLE [13]. Yang et al. reported the sexdependent differential activation of NLRP3 and AIM2 inflammasomes in SLE macrophages [14]. Regulation of Ifi202 has also been linked to sex hormones [15]. Their studies suggest that there is potentially sex difference in the function of Ifi200 genes. Therefore, understanding whether there is a sex difference and how their expression and function show sex difference would represent a significant advance in the elucidation of molecular mechanism(s) of Ifi200 genes in autoimmune diseases.
Its orthologue cluster in mouse is Ifi200 on mouse chromosome 1. In mice, the Ifi200 family genes Ifi202b, Ifi203, Ifi204, Ifi205, and Mnda are known to reside on chromosome 1. This family gene cluster is between 173,747,293 (Ifi204) bp and 174,031,810 (Ifi205) bp. The Ifi200 family genes are next to each other and are called the interferon-inducible (Ifi) gene 200 cluster. How these genes interact and exert influence on each other is not clear. Thus, elucidation of interaction mechanisms among genes in the Ifi200 cluster might enhance our understanding of relationships between these genes and autoimmune diseases, in particular the RA. However, due to the requirements for appropriate sample collections (e.g., both sexes at the same age, unified genomic background, and controlled environment), such a study using human populations has been difficult.
Animal models have been widely used to study topics that could not be easily studied using human populations. In particular, rodent models such as those in mice have contributed tremendously to our understanding of human genetics and genomics. We will examine the sex similarity and difference using data of whole genome gene expression profiles from a well-known mouse population of recombinant inbred (RI) strains derived from C57BL/6J and DBA/2J (BXD), which is the largest RI mouse population and with remarkable data on whole genome expression profiles and phenotypes [16][17][18]. The first set of 36 BXD RI strains was originally established in 1930s at The Jackson Laboratory [19]. Over the last more than a half century, the BXD RI strains have expanded into a population with almost of a hundred RI strains. Among rodent animal models, this is the largest animal RI strains in history [20]. Unlike F2 population, one RI strain needs intercrossing within the strain more than 20 generations before it established its homozygous status and survival of inbred selection. In the last decade, BXD RI strains have been used widely as the only reliable RI strain population. As of May, 2018, PubMed posted more than 500 publications that their research is based on the BXD strains. The analysis tools for these RI strains are provided by GeneNetwork [21]. These tools have been tested, applied, and approved for the last decade [22].

Mouse Gene Expression Data Sets.
For study of the sex difference and similarity of gene expression profiles, we used the whole gene expression profiles of spleen from female and male mice in a population of recombinant inbred (RI) strains from BXD (derived from C57BL/6J and DBA/2J) [18].
The data sets, UTHSC Affy MoGene 1.0 ST Spleen (Dec10), were from GeneNetwork (http://www.genenetwork. org/webqtl/main.py). There were separate sets of whole genome expression profiles from the spleen of female and male mice. In both male and female mouse sets, the whole genome gene expression profiles were all from a total of 85 strains, including 64 BXD strains, parental strain C57BL/6J, two reciprocal F1 hybrids (B6D2F1 and D2B6F1), and 18 other common inbred strains. The spleen was profiled using the Affymetrix GeneChip Mouse Gene 1.0 ST array. In most cases, two arrays were processed per strain-one for males and one for females [23].

EQTL Mapping.
We followed the standard protocol provided by GeneNetwork in conducting the eQTL mapping of female and male eQTL from spleen. eQTL maps were generated by using the command "Interval" of "Mapping Tools" at GeneNetwork. Permutation tests of 3000 was used to map the eQTLs to confirm accuracy of eQTL mapping. A simple regression method was used by GeneNetwork for mapping the expression QTL of genes with flanking markers [24]. The expression values from different strains were considered as phenotypes. Molecular markers along the chromosomes were used as genotypes or indicators of locations on chromosomes. The expression values were then compared for the probability of a specific genotype at a test location between two flanking markers. Significance of eQTL at a location was evaluated with statistical probabilities to eventually generate eQTL.
2.3. Gene Network Construction. Gene networks were constructed for genes in the Ifi200 pathway and for sex hormones. The graphic application tools in GeneNetwork were used for the gene network construction. Spring Model layout (force reduction) was selected as the graphic method for all graphic subjects. The criteria for the strong correlation, correlation, and no correlation were the absolute value of R equal to or >0.7 ( red color for positive and blue color for negative correlation), between 0.36 and 0.69 ( pink color for positive and green color for negative correlation), and between 0 and 0.35, respectively ( light pink color for positive and black color for negative correlation) [25]. In case of multiple probes, all of the probes were initially employed in the construction of gene network. The probe with the highest expression level was chosen from highly positive related probes of the same gene for the final construction of the gene network. For construction of graphic gene network, unless specified, we default to show the Pearson correlation coefficients > 0.35 or ≤0.35 between genes. The graph's canvas is 40.0 cm by 40.0 cm. The node labels and edge labels are drawn with a 16.0-point font.

Gene Categories
Analyzed for Association with Genes in the Ifi200 Family. Genes collected from four categories were used in the analysis for correlations of their expression levels to the expression levels of the Ifi200 family. We first included genes that are potentially connected to Ifi200 genes and well known for their importance in the immune system. These genes included FoxP3, Tgfb, type I IFN, and translocated promoter region (Tpr) protein. Type I IFNs belong to the class II family of α-helical cytokines, which includes type II IFN-γ, IL-17, the newly identified IFN-λs, IL-10, and several IL-10 homologs (IL-19, IL-20, IL-22, IL-24, and IL-26) [26][27][28][29]. We also included potential upstream genes such as Csf1r [30], Gata4, Nkx2.5, and Tbx5 [31].
Probes for sex hormones were searched in GeneNetwork from the whole genome expression profiles of the spleen of female and male mice by using the key words "estradiol," "progesterone," and "testosterone." The expression levels of these genes were correlated with expression levels of genes in the Ifi200 family.

Data
Organization and Comparison. The following symbols are used in the data analysis and organization: R is the correlation between each gene in the Ifi200 family and each gene from gene sets of different categories. R values were obtained from matrix analysis and graphic analysis at GeneNetwork [16,47]; R is the correlation between different R values of different comparisons; P is the result of t-test; and Raa is the average of the absolute R value, which is calculated by Raa = ∑Ri/n, where Ri is the absolute R value between a gene in a gene group and n is the total number of R values.
When analyzing the correlations between Ifi family genes and each set of other selected genes, the collective R values between each Ifi200 gene to the genes in each category are treated as a set. The four sets of R values, for example, Ifi202b, Ifi203, Ifi204, and Ifi205 in each category (known relevant immune and Ifi200 pathways, lupus, arthritis, and sex hormones) in female are compared to the set in male. The P values from t-tests and R values from correlation tests are used as criteria to judge their similarities and differences. The relevance of expression levels between a gene group and Ifi200 genes was evaluated by the average of absolute R values between genes in a group and Ifi200 genes. Raa is used to compare the strength of association of the Ifi200 family genes to genes in each category.
2.6. Statistical Analysis. Student's t-test was used for the comparison of samples. The criteria for statistical significance follow the standard values. Thus, P ≤ 0 01 and P ≤ 0 05 represent strong significant difference and difference, respectively. In the construction of network graphs, the R absolute value > 0.50 was considered as the indication of the threshold for the real connection line between two genes or probes.

The Expression Levels in Female and Male
Mice of Genes of the Ifi200 Family. The expression levels of genes of the Ifi200 family (Ifi202b, Ifi203, Ifi204, Ifi205, and Mnda) were examined. One probe for each of the Ifi202b, Ifi203, Ifi204, and Ifi205 genes was identified from female and male populations except Mnda. The probe for beta-actin was used as the control. The R value of beta-actin between female and male mice was 0.166341. The R values for Ifi202b, Ifi203, Ifi204, and Ifi205 were 0.9502, 0.6332, 0.6712, and 0.3960, respectively. The t-test resulted in a P value of 0.931. The P values between female and male for Ifi202b, Ifi203, Ifi204, and Ifi205 were 0.857239, 0.003656, 0.155603, and 0.340488, respectively. Further examination of the expression levels of Ifi203 revealed that while the expression level of Ifi203 between female and male mice in most strains was similar, a few strains showed considerable difference such as strains BXD27 and BXD58 (Figure 1(a)). eQTL mapping suggests that Ifi202b, Ifi203, and Ifi204 in both sexes were all mapped on chromosome 1 (Figure 1(b)). However, the eQTL of Ifi205 in female mice was mapped on chromosome 2, while in male mice, it was mapped on chromosome 15 (Figure 1(b)). This initial analysis suggests that there was potential sex difference among genes in the Ifi200 family.

Sex Difference in Correlation with Ifi200
Genes and the Lupus-Relevant Gene, Jhdm1d. Nineteen probes for 16 lupus-relevant genes (Abca1, Apobec3, Ccr7, Ceacam3, Ets1, Foxp3, Hdac1, Ifng, Irf9, Itgam, Jhdm1d, Mmp9, Oscar, Slc4a1, Tbx21, and Tlr7) were identified from the database of male murine spleen. Similar to that of Ifi200-relevant genes above, the correlation between the expression levels of these 17 genes and Ifi200 genes was analyzed. Overall, the correlation in expression patterns between female and male mice was similar (Figures 3(a) and 3(b)). The R values between these 16 genes and the Ifi200 gene family were 0.4921, 0.8198, 0.8788, and 0.8572, for Ifi202b, Ifi203, Ifi204, and Ifi205, respectively. In male spleen, the correlations among these 17 genes with Ifi200 genes are stronger than those for the above 18 Ifi200-relevant immune important  Grid2  Ifnb1  If10  Il17a  Il19  Il20  Il22  Il22  Il24  Nkx2−5  Sgta  Sgtb  Tbx5  Tgfb1  Tgfb2  Tgfb2  Tgfb3  Tpr  Tpr (Figures 3(a) and 3(b)) in both sexes. P values from t-test between female and male mice for the correlations between Ifi200 genes and lupus-relevant genes were 0.4783, 0.0303, 0.4147, and 0.002 for Ifi202b, Ifi203, Ifi204, and Ifi205, respectively. Therefore, more detailed information between Ifi203 and lupus-relevant genes and between Ifi205 and genes in lupus pathway were obtained. As indicated in Figure 3(c), the correlation of Ifi203 with Jhdm1d for female and male mice is different. Jhdm1d had two probes. The net sex differences of these two probes were 0.5 and 0.34. The net sex differences between Ifi205 and two probes of Jhdm1d were 0.34 and 0.35. These data suggest that among genes in the lupus pathways, Jhdm1d may regulate the sex difference. Furthermore, we examined the expression levels of Ifi200 genes and genes of other important immune-related genes (shown in Figure 2) plus Jhdm1d in female and male mice of NZB/ BlNJ which has been used for breeding of NZB/W [48]. The data shows that in the Ifi200 family, the expression levels of all four genes in female mice are higher than those in male mice. However, the expression levels of most of the immunerelated genes and Jhdm1d in female mice are lower than those in male mice (except Csf1r and Trhr) (Figure 3(e)).

The Association between Ifi200
Genes and Genes of OA and RA. Twenty-one genes with relevance to OA and RA were identified from the whole genome expression profiles of spleen from female and male mice. These genes are Akt1, C3ar1, Ccl19, Ccr1, Cd97, Cxcl1, Fyb, Ifi202b, Ifi203, Ifi204, Ifi205, Ikbke, Lyz1, Map3k13, Mapk10, Mapk10, Mapk13, Pln, Ppt1, Socs1, and Xiap. In both female and male mice, the expression level of Ifi205 showed a positive correlation with Ppt1 (Figures 4(a) and 4(b)). The expression level of Ifi203 showed positive correlation with Cd97 and Xiap (Figures 4(a) and 4(b)), which confirmed a previous report [45,46]. The probe for Pln was the only one that did not show any correlation in either male or female mice. The average absolute Raa values for Ifi202b, Ifi203, Ifi204, and Ifi205 for these genes in male mice were 0.1012, 0.2684, 0.1731, and 0.2482, respectively. In female mice, the average Raa values were 0.0847, 0.2061, 0.1418, and 0.2276. The P values between female and male mice for the R values between this set of OA-and RA-related genes and Ifi202b, Ifi203, Ifi204, and Ifi205 were 0.1769, 0.2616, 0.0676, and 0.5872, respectively. The R values for these four groups between female and male mice were 0.5392, 0.7019, 0.4282, and 0.8063, respectively.
The average absolute Raa values for Ifi202b, Ifi203, Ifi204, and Ifi205 with these genes in male mice were 0.0962, 0.1863,

Discussion
Our data have revealed that the correlation expression levels of Ifi200 genes with some immune-relevant genes have gender differences and suggest that different genes in the Ifi200 family play different roles in male and female mice among different pathways of immune-mediated diseases. First, Ifi2002b showed the most sex difference in correlation with the 18 selected immune-relevant genes among all four Ifi200 family genes. Second, Ifi202b showed gender difference in correlation with sex hormone genes [7]. Third, Ifi203 and Ifi205 showed significant gender difference in correlation with the genes in the lupus pathway. Fourth, none of the genes showed significant sex difference in correlation with genes selected for relevance to OA and RA. Our study also discovered key genes that potentially interact with Ifi200 genes that are involved in regulating gender differences in pathological pathways of inflammatory and/or immunemediated diseases.
Our initial analysis suggested that there was a potential gender difference in the expression level and regulation of Ifi200 family genes. Our analysis obtained a P value of the Ifi203 expression level of 0.0037 between female and male mice. Second, the QTL of Ifi205 in female and male mice was mapped to different chromosomes. These differences provide a foundation for their potential sex differences in regulation of different immune pathways or diseases.
Our analyses suggest that there is sex difference between genes of the Ifi200 family in their coexpression with some known coexpressed immune-related genes [28][29][30][31][32]. In particular, the coexpression of Ifi202b and these immunerelated genes between female and male mice was significantly different with a P value of 0.0004. The correlation of expression levels between Ifi202b and six well-known, immunerelated genes (Csf1r, Ifnb1, IL-20, IL-22, IL-24, and Tgfb2) is stronger in male mice than that in female mice. These genes are key components in the interleukin and Tgf pathways, which are essential for pathological processes in immune responses. In addition, both Ifi204 and Ifi205 were negatively correlated with Gata4. It has been reported that p204 (the protein product of Ifi204) is required for differentiation of P19 murine embryonal carcinoma cells to beating cardiac myocytes: its expression is activated by cardiac Gata4 and two other proteins, Nkx2.5 and Tbx5 [32].  Thra  Thrap3  Thrb  Thrsp  Trh  Trhr  Trhr2  Trip10  Trip11  Trip12  Trip13  Trip4  Trip6  Tshb  Tshr  Shbg  Lipe  Lipe   These data suggest that, in the spleen, the expression levels of Nkx2.5 and Tbx5 may be influenced by Ifi204 and Ifi205 through Gata4.
Our data indicated that the gender difference in lupus disease might be caused by the molecular pathway that is regulated through interaction between Ifi200 genes and genes in the lupus pathway. Among four genes, two of them, Ifi203 and Ifi205, showed sex difference when correlated with the expression levels of genes in the lupus pathway. The correlation of both Ifi203 and Ifi205 to Jhdm1d between female and male mice is different. Jhdm1d had two probes. Both probes showed sex difference in the correlation of expression levels between two Ifi200 genes and genes in the lupus pathway. Sex hormones have been known to moderate the susceptibility to lupus [7]. Our data strongly suggest that among genes in lupus pathways, Jhdm1d may regulate the sex difference through interaction with Ifi200 genes. Surprisingly, we did not find significant sex difference between the correlation of expression levels of Ifi200 genes and known OA-or RArelevant genes. However, these data need to be confirmed in the future studies.
Our data also established a connection of gender differences with sex hormone genes. Sex hormone genes are regarded as key genes in the gender differences in diseases. Ifi202b plays a significant role in gender difference through interaction with sex hormones. The P values between the sex hormone genes and Ifi202b were 0.0243. Correlations of several genes to Ifi202b showed sex difference. These genes were Crh, Ghitm, Thrsp, Trhr, and Tshr. However, our comparisons did not show a significant gender difference between the rest of three genes in the Ifi200 family and the sex hormone genes.

Conclusions
Four genes in the Ifi200 family play different roles in sex difference among dissimilar pathways of these four gene groups. Different genes play different roles in sex difference in different diseases. In order to understand the molecular mechanism of sex difference in different immune diseases, it is essential to study the roles of genes of the Ifi200 family.