Testosterone Influence on Gene Expression in Lacrimal Glands of Mouse Models of Sjögren Syndrome

Purpose Sjögren syndrome is an autoimmune disorder that occurs almost exclusively in women and is associated with extensive inflammation in lacrimal tissue, an immune-mediated destruction and/or dysfunction of glandular epithelial cells, and a significant decrease in aqueous tear secretion. We discovered that androgens suppress the inflammation in, and enhance the function of, lacrimal glands in female mouse models (e.g., MRL/MpJ-Tnfrsf6lpr [MRL/lpr]) of Sjögren syndrome. In contrast, others have reported that androgens induce an anomalous immunopathology in lacrimal glands of nonobese diabetic/LtJ (NOD) mice. We tested our hypothesis that these hormone actions reflect unique, strain- and tissue-specific effects, which involve significant changes in the expression of immune-related glandular genes. Methods Lacrimal glands were obtained from age-matched, adult, female MRL/lpr and NOD mice after treatment with vehicle or testosterone for up to 3 weeks. Tissues were processed for analysis of differentially expressed mRNAs using CodeLink Bioarrays and Affymetrix GeneChips. Data were analyzed with bioinformatics and statistical software. Results Testosterone significantly influenced the expression of numerous immune-related genes, ontologies, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in lacrimal glands of MRL/lpr and NOD mice. The nature of this hormone-induced immune response was dependent upon the autoimmune strain, and was not duplicated within lacrimal tissues of nonautoimmune BALB/c mice. The majority of immune-response genes regulated by testosterone were of the inflammatory type. Conclusions Our findings support our hypothesis and indicate a major role for the lacrimal gland microenvironment in mediating androgen effects on immune gene expression.

hormone action is a unique, tissue-specific effect, which is initiated through androgen binding to specific receptors in lacrimal gland epithelial cells. 7 In addition, we hypothesize that this androgen interaction then elicits the altered expression and/or activity of immune-related genes in lacrimal tissue, leading to a decrease in immunopathologic lesions and an improvement in glandular function.
To begin to test this hypothesis, we examined the nature and magnitude of testosterone's influence on immune-related gene expression in the autoimmune lacrimal tissues of female MRL/lpr mice after onset of disease. We chose the MRL/lpr strain because, like in humans, the extent of lacrimal and salivary gland inflammation in MRL/lpr mice is far greater in females compared to males, 18 and is dramatically reduced in response to androgen treatment. 7,[11][12][13][14] For comparative purposes, we also analyzed and compared the androgen impact on immune gene expression in lacrimal glands of female nonobese diabetic/LtJ (NOD) mice after onset of disease. These mice, which are an established model for type-1 insulin-dependent diabetes mellitus, 19 have been used as a model for Sjögren syndrome [20][21][22] and, like in humans, have far greater inflammation in the salivary glands of females compared to males. 18 However, unlike humans, the lacrimal glands of male NOD mice have significantly higher inflammation than those of females. 18,[23][24][25] Indeed, orchiectomy of NOD mice attenuates, whereas androgen treatment of castrated NOD males induces, lymphocyte accumulation in their lacrimal glands. 23 This anomalous hormone effect is mediated through the lacrimal microenvironment 24 and contrasts with the androgen-induced decrease in inflammation in salivary and pancreatic tissues in these mice. 26,27 Given this background, we hypothesized that androgen exposure will significantly increase the expression and/or activity of immune-related genes in the lacrimal glands of female NOD mice. We also hypothesized that these opposing actions of androgens in female MRL/lpr and NOD lacrimal tissues involve regulation of similar immune-related genes, ontologies, and pathways.

Animals and Tissue Collections
Adult female MRL/lpr and NOD mice were purchased from the Jackson Laboratories (Bar Harbor, ME, USA). Animals were maintained in constant temperature rooms with fixed light/ dark intervals 12 hours in duration. Pellets containing vehicle (cholesterol, methylcellulose, lactose) or testosterone (T; 10 mg) were implanted subcutaneously in MRL/lpr (17.1-18.1 weeks old) and NOD (21 weeks old) mice. The pellets were obtained from Innovative Research of America (Sarasota, FL, USA) and were designed for constant release of placebo (P) or physiologic amounts of androgen (for a male [11][12][13][14] for a 3-week period. After 20 to 21 days of treatment, mice (n ¼ 7-18 mice/ condition) were killed by CO 2 inhalation and exorbital lacrimal glands were removed for molecular biological procedures. Lacrimal tissue samples were prepared by combining glands from two to six mice/strain/group. Three different sample preparations were made for each treatment (i.e., 4-12 lacrimal glands/sample/treatment/strain) and then processed for analysis of gene expression.
All mouse studies were approved by the institutional animal care and use committee of the Schepens Eye Research Institute and adhered to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Molecular Biological Procedures
To determine the effect of T on lacrimal gland gene expression, total RNA was isolated from lacrimal tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and purified with RNAqueous spin columns (Ambion, Austin, TX, USA). Lacrimal gland RNA samples were treated with RNasefree DNase (Invitrogen), assessed spectrophotometrically at 260 nm to determine concentration, and examined with a RNA 6000 Nano LabChip and an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) to verify RNA integrity. The RNA samples were kept at À808C until further processing.
Gene expression was determined via two different procedures. One involved hybridization of lacrimal gland RNA samples to CodeLink (CL) UniSet Mouse 20K I Bioarrays (n2 0,000 genes/array; Amersham Biosciences/GE Healthcare, Piscataway, NJ, USA), according to reported methods. 28 cDNA was generated from RNA (2 lg) with a CL Expression Assay Reagent Kit (Amersham) and purified with a QIAquick purification kit (Qiagen, Valencia, CA, USA). Samples were dried, and cRNA was made with a CL Expression Assay Reagent Kit (Amersham), recovered with an RNeasy kit (Qiagen), and quantified with an ultraviolet spectrophotometer. Fragmented, biotin-labeled cRNA then was incubated and shaken at 300 rpm on a CL Bioarray at 378C for 18 hours. Following this time interval, the Bioarray was washed, exposed to streptavidin-Alexa 647, and scanned using ScanArray Express software and a ScanArray Express HT scanner (Packard BioScience, Meriden, CT, USA) with the laser set at 635 nm, laser power at 100%, and photomultiplier tube voltage at 60%. Scanned image files were evaluated using CL image and data analysis software (Amersham), which gave raw and normalized hybridization signal intensities for each array spot. The intensities of the approximately 20,000 spots on the Bioarray image were normalized to a median of 1. Standardized data, with signal intensities >0.50, were analyzed with bioinformatic software (Geospiza, Seattle, WA, USA). This comprehensive software also produced gene ontology, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, and z-score reports. The ontologies included those related to biological processes, molecular functions, and cellular components, and were organized according to the recommended guidelines of the Gene Ontology Consortium (available in the public domain at http://www.geneontology.org/GO.doc.html). 29 The second method to determine differential gene expression entailed hybridization of each cRNA (20 lg) sample to a GeneChip Mouse Genome 430A 2.0 Array (Affymetrix [Affy], Santa Clara, CA, USA) according to the manufacturer's protocol. Reagents for the fragmentation and hybridization steps originated from a GeneChip HT One-Cycle Target Labeling and Control Kit, and materials for the washing and staining steps were from a GeneChip HWS kit (Affy). Hybridized GeneChips were scanned with an Affy Model 700 Scanner and expression data files were generated from array images using Affy Microarray Suite 4.0 software. GeneChip data were normalized by choosing the default scaling in the Affy GeneChip operating software, which gives a trimmed mean intensity of 500 for each GeneChip microarray. Standardized data with a quality value of 1.0 then were evaluated with Geospiza GeneSifter software.
As we reported recently, 30 counts of unique mappings of probes to gene identifications in the CL and Affy arrays demonstrated that there were 15,711 and 13,265 unique genes, respectively, in these arrays. Examination of the intersection of these lists showed that there was an overlap of 11,299 genes.
Gene expression data were evaluated without log transformation and statistical analyses were conducted with Student's t-test (2-tailed, unpaired) using the GeneSifter software. Our statistical method was not tailored for multiple comparisons. Genes expressed in the same direction in comparative groups were identified using GenBank accession numbers and a Geospiza intersector program. Data used for these CL and Affy arrays are accessible for free download through the National Center for Biotechnology Information's Gene Expression Omnibus (NCBI GEO) via series accession number GSE5877.
We also compared our results to data from our studies examining the influence of sex in adult MRL/lpr and NOD mice (n ¼ 15-18/sex/strain), 30 and 2 weeks of P or T treatment of nonautoimmune, ovariectomized BALB/c mice (n ¼ 5-6 mice/ condition/experiment), 31 on lacrimal gland gene expression. The sex-and hormone-related data are available through the NCBI GEO via series accession numbers GSE5876 and GSE3995, respectively.  MRL/lpr  CL  1890  1708  3598  Affy  1120  1530  2650  NOD  CL  1474  2275  3749  Affy  1102  1150  2252 Data were evaluated without log transformation. The expression of listed genes was significantly (P < 0.05) up ()-or down ()-regulated by T treatment. Accession numbers are the sequence identities of gene fragments expressed on the CL and Affy microarrays. These sequences appear in the nucleotide database of the NCBI. Relative ratios were determined by comparing the degree of gene expression in lacrimal glands from P-and Ttreated female MRL/lpr mice. Ratios were calculated from nontransformed data.

T Influence on Gene Expression in Lacrimal Glands of Female MRL/lpr and NOD Mice
To determine the effect of androgen treatment on gene expression in lacrimal glands of autoimmune mice, tissues were obtained from female MRL/lpr and NOD mice (n ¼ 7-18 mice/strain/treatment) following 20 to 21 days of exposure to P or T. Glands were pooled according to treatment and strain (n ¼ 4-12 glands/samples/strain/treatment; n ¼ 3 samples/ treatment group), processed for isolation of total RNA, and analyzed for differentially expressed mRNAs using CL Bioarrays and Affy GeneChips. Microarray data were evaluated with Geospiza bioinformatics software.
Our results with CL and Affy microarrays showed that testosterone treatment has a significant influence on expression of thousands of genes in lacrimal glands of MRL/lpr and NOD mice (Table 1). Androgen exposure increased () the activity of genes, such as cytochrome P450, family 2, subfamily j, polypeptide 13 (Cyt), and decreased () that of pancreatic lipase-related protein 1 (PL) in both strains (Tables 2, 3). These two genes also are regulated in the same manner in lacrimal tissues of nonautoimmune female BALB/c mice (Cyt ¼ 9.9-fold ; PL ¼ 81.1-fold ; NCBI GEO GSE3995). 31 Examples of other genes upregulated in lacrimal glands of MRL/lpr mice, such as oxysterol binding protein-like 3, olfactory receptor 1086, and dopa decarboxylase (Table 2), also were very highly upregulated by 39.4-, 36.8-, and 58.6-fold amounts, respectively, in NOD lacrimal tissues. In contrast, the gene expression for cathepsin S, which is significantly elevated in the tears of Sjögren syndrome patients, 32 was significantly (P < 0.05) decreased by testosterone in female MRL/lpr lacrimal glands (CL ¼ 1.53-fold ; Affy ¼ 1.88-fold ), but increased by androgen treatment in those of female NOD mice (CL ¼ 3.87- Genes with known ontologies are listed. Relative ratios were calculated by comparing the degree of gene expression in lacrimal glands from Pand T-treated female NOD mice. Ratios were generated from nontransformed data. fold ; Affy ¼ 3.23-fold ). A similar pattern was found for moesin gene expression, which was reduced by T in female MRL/lpr lacrimal glands (Affy ¼ 3.19-fold ), but increased by androgen exposure in lacrimal tissues of female NOD mice (Affy ¼ 3.39-fold ). Other genes were regulated by T in the lacrimal tissue of only one strain (e.g., NOD, spleen tyrosine kinase [Syk]; CL ¼ 3.1-fold ).
As we 30,[33][34][35] and other investigators [36][37][38][39] have discovered, the vast majority of lacrimal gland genes in MRL/lpr and NOD female mice, which were identified as differentially expressed by the CL and Affy microarrays, were unique to each platform. Indeed, as demonstrated in Table 4, only 8.5% to 11.1% (T>P), and 7.3% to 16.8% (P>T) of the regulated genes were found by both microarrays. These data showed that there are significant differences in the ability of these platforms to detect differential gene expression.
This low concordance in gene identification appears to be due to intrinsic variations in multiple aspects of platform design, as well as to the inherent instability of lists of significantly changed genes based upon P value cutoffs. [36][37][38][39][40] The result is that CL and Affy microarrays, both of which have documented accuracy and reproducibility, seem to measure different things. 38 Most gene expression differences revealed by each platform are thought to be biologically correct, and these variations cannot be attributed to technological differences. 37,38 Comparison of gene expression between the lacrimal glands of P-treated MRL/lpr and NOD mice demonstrated that 587 genes were in common (CL). The alternate comparison (i.e., MRL/lpr, T>P; NOD, T>P) revealed 559 genes in common (CL).

T Effect on Immune-Related Ontologies in Lacrimal Glands of Female MRL/lpr, NOD and BALB/c Mice
T exerted a significant influence on the expression of a large number of immune-related gene ontologies in the lacrimal glands of female MRL/lpr and NOD mice. Many of these hormone responses were identified by CL and Affy platforms (Tables 5, 6).
As demonstrated in Table 5, androgen administration downregulated the expression of over 60 immune-associated biological process ontologies ( ‡20 genes/ontology) in lacrimal tissues of female MRL/lpr mice, including those related to immune system processes, lymphocyte activation, cytokine production, and inflammatory response. In contrast, T increased the expression of all of these same immune ontologies, as well as more, in lacrimal glands of female NOD mice (Table 6). These changes were accompanied by downand upregulation of immune-related molecular function (e.g., chemokine activity) and cellular component (e.g., MHC protein complex) ontologies ( ‡5 genes/ontology) in lacrimal tissues of MRL/lpr and NOD mice, respectively.
Some genes represented within these immune ontologies were the same (e.g., MRL/lpr and NOD : chemokine [C-X-C motif] ligand 9 [Cxcl9], IL-1b, and toll-like receptors 1 and 2 [TLR 1 and 2]), but most were not. For example, T decreased the expression of 96 immune-response genes (CL) in lacrimal glands of MRL/lpr mice (Table 7), but the majority of these genes were different than the 133 genes (CL) upregulated in NOD mouse tissues (Table 8). Despite these differences, the androgen-regulated immune-response genes were predominantly inflammatory in nature. Thus, T downregulated the expression of 41 inflammatory genes in MRL/lpr lacrimal tissues and 23 of these were the same as in Table 7. Further, androgen administration increased the expression of 52 inflammatory genes in NOD lacrimal glands and 36 of these were identical to those in Table 8.
Not all immune-related responses to T in the lacrimal glands of female MRL/lpr and NOD mice were opposite. As shown in Table 9, the expression of certain inflammatory genes was down-or upregulated in the same way in both strains.
The modulatory effect of T on immune-related gene expression in the autoimmune mouse lacrimal glands did not reflect an androgen action typically found in lacrimal tissues of a nonautoimmune strain. Indeed, the effect of T on gene ontologies in lacrimal glands of female NOD, compared to female BALB/c, mice showed significant differences. For example, 21 of 22 androgen upregulated biological process ontologies (Affy) in NOD mice (n ¼ 479 NOD > BALB/c ontologies) with the highest z-scores (z ¼ 6.85 -10.59) were all immune-related. In contrast, only two of the 161 biological process gene ontologies expressed to a greater extent in BALB/ c versus NOD mice were immune-associated. Instead, the BALB/c biological process ontologies with the highest z-scores were translation elongation (z ¼ 11.59), translation (z ¼ 9.56) and oxidation-reduction (z ¼ 6.87). In the same way, some of the top molecular function and cellular component ontologies in T-treated female NOD mice were immune-related antigen binding (z ¼ 8.72), chemokine receptor binding (z ¼ 4.63), and MHC protein complex (z ¼ 6.64), whereas they were structural constituent of ribosome (z ¼ 10.94), mitochondrion (z ¼ 12.77) and multiple ontologies related to oxidoreductase activities in androgen-treated female BALB/c mice. Data were evaluated without log transformation. Genes identified as ''unique'' were significantly (P < 0.05) increased on one, but not the other, microarray platform. The phrase ''Genes changed in the same (or opposite) direction'' means that the results were significant (P < 0.05) on both platforms. T administration led to a significant decrease in the expression of immune-related KEGG pathways in lacrimal glands of female MRL/lpr mice (Table 10). These included such pathways as chemokine signaling, cytokine-cytokine receptor interaction, and leukocyte transendothelial migration (Table 9). In contrast, T induced a significant increase in the expression of these KEGG pathways, as well as many more, in lacrimal tissues of female NOD mice (Table 11).

Comparison Between the Influence of Sex and T on Immune-Related Gene Expression in Lacrimal Glands of MRL/lpr and NOD Mice
Lacrimal glands of female MRL/lpr and male NOD mice, compared to their opposite sexes, contain a significantly greater expression of genes, ontologies, and KEGG pathways related to inflammatory responses, antigen processing, and chemokine signaling. 30 We hypothesized that many of these immune-related genes, ontologies, and pathways are analogous to those T suppresses in female MRL/lpr, and induces in female NOD mouse lacrimal tissues. To test this hypothesis, we compared the sex and T influence on immune-related gene expression in MRL/lpr and NOD mice. We also compared these findings to genes more highly expressed in inflamed (MRL/lpr female and NOD male) versus noninflamed (MRL/lpr male and NOD female) lacrimal tissues. As shown in Tables 12 to 14, many immune-related biological process ontologies (e.g., inflammatory response), immune response genes (e.g., complement component 3) and chemokine KEGG pathway genes (e.g., chemokine [C-X-C motif] ligand 9) that are influenced by sex and T in lacrimal glands of MRL/lpr and NOD mice are identical. Thus, androgen downregulates multiple immune-related genes that are highly expressed in lacrimal tissues of female MRL/lpr mice, and T upregulates the expression of these immune genes, which typically are expressed in NOD males, in female NOD lacrimal tissues. These regulated genes in Tables 12 to 14 are the same as those more highly expressed in inflamed compared to noninflamed lacrimal glands.

DISCUSSION
Our results showed that T significantly influences the expression of numerous immune-related genes, ontologies, and KEGG pathways in lacrimal glands of MRL/lpr and NOD mice. These genes are associated with processes, such as lymphocyte activation, leukocyte transendothelial migration, antigen binding, chemokine signaling, cytokine production, cytokine-cytokine receptor interaction, MHC protein complex, and the inflammatory response. The nature of this androgeninduced response depends upon the autoimmune strain and is not duplicated within lacrimal tissues of nonautoimmune BALB/c mice. The majority of immune-related genes regulated by T are of the inflammatory type. Our findings indicated the lacrimal gland microenvironment as a key mediator of androgen effects on immune gene expression and the associated immunopathology.
Our study was prompted by our earlier discovery that androgens, but not estrogens, dramatically suppress the inflammation in lacrimal tissues of the female MRL/lpr and NZB/NZW FI mouse models of SS. [11][12][13][14] We hypothesized that this androgen effect involves an alteration in the expression and/or activity of immune-related genes, because such genes are critically important in innate and adaptive immune responses. 42 These genes might also have a major role in promoting the multiple immunosuppressive actions of androgens, including those directly on T cells, monocytes, macrophages, neutrophils, and B cell precursors, and indirectly on peripheral B cells. 43,44 These androgen actions lead to regulation of the maturation, proliferation, migration, and/or function of immune cells; synthesis and secretion of antibodies, cytokines, adhesion molecules, and proto-oncogenes; and expression of autoantigens. 2,43,44 A result is that androgens are protective in SS, as well as in other autoimmune diseases, such as systemic lupus erythematosus, multiple sclerosis, and rheumatoid arthritis. 2,5,6,43 We discovered that testosterone suppresses a wide array of immune-related genes in lacrimal glands of female MRL/lpr mice. The question is whether some of these genes may be intricately involved in helping to mediate testosterone's antiinflammatory action in this tissue. Possible examples abound. For example, the androgen downregulation of complement 3, Cxcl9, moesin, IL-1b, and TLR2 genes may interfere with the early stages of SS disease development and the triggering of an adaptive immune response in the lacrimal gland. 30,[45][46][47][48][49][50] However, if these five genes are important for the androgeninduced downregulation of lacrimal gland inflammation in female MRL/lpr mice, why are these same genes upregulated by androgen treatment in lacrimal tissues of female NOD mice?
Indeed, we found that many of the immune response genes, immune-related biological process ontologies, and chemokine KEGG pathway genes that are influenced by sex and T in Biological process ( ‡20 genes/ontology), molecular function ( ‡5 genes/ontology) and cellular component ( ‡5 genes/ontology) immune ontologies were identified after the analysis of nontransformed CL and Affy data. These immune ontologies were upregulated in lacrimal gland samples from P-treated female mice, and by extension, downregulated in lacrimal gland tissues from androgen-treated mice. A z-score is a statistical rating of the relative expression of genes, and shows how greatly they are over-or underrepresented in a specific gene list. 41 Positive z-scores represent a higher number of genes meeting the criterion than is anticipated by chance, and values >2.0 are significant. CL Genes , number of genes downregulated, as calculated with a CL Bioarray; Affy Genes , number of genes downregulated, as determined with Affy GeneChips; z-score, specific score for the down-regulated genes in the CL-and Affy-related tissues.  Biological process ( ‡20 genes/ontology), molecular function ( ‡5 genes/ontology) and cellular component ( ‡5 genes/ontology) immune ontologies were identified after the evaluation of nontransformed CL and Affy data. CL Genes , number of genes up-regulated, as identified with a CL Bioarray; Affy Genes , number of genes up-regulated, as found with Affy GeneChips; z-score, specific score for the up-regulated genes in the CL and Affy related tissues. Relative ratios were calculated from CL and Affy data by comparing the degree of gene expression in lacrimal glands from P-versus T-treated female MRL/lpr mice. Listed CL genes were increased ‡2-fold.
* Genes were found to be upregulated in lacrimal glands of female NOD mice treated with T (Table 8). lacrimal glands of MRL/lpr and NOD mice are identical. Thus, androgen decreased the expression of multiple immune-related genes in lacrimal tissues of female MRL/lpr mice, and T increased the expression of these immune genes, which are typically expressed in NOD males, 30 in female NOD lacrimal tissues. We also discovered that many of these regulated genes are the same as those typically highly expressed in inflamed compared to noninflamed lacrimal glands. Are there specific genes, then, that might be responsible, at least in part, for promoting the anomalous androgen-induced inflammation in NOD lacrimal glands? Possible genes might be those encoding kallikrein 1 and its related peptidases (KLKs) b1, b4, b5, b8, b11, b24, and b26. Testosterone increased the expression of these genes by 8.4-to 216.7-fold amounts in female NOD lacrimal tissues. KLKs constitute a family of serine proteases that are stimulated by androgens in other tissues 51 and appear to have a significant role in the development and progression of autoimmune diseases. 52,53 KLK protein levels are increased in lacrimal glands in primary SS. 54,55 Further, several KLKs act as autoantigens, and may serve to elicit an autoimmune T-cell response against lacrimal tissue and to cause a decrease in aqueous tear secretion. 54,[56][57][58] However, it is unlikely that KLKs are the keys to understanding androgenimmune effects in NOD mice. The reason is that T also increases by 1.7-to 273-fold the gene expression of KLKs b1, b4, b8, b10, b11, b16, b21, b24, b26, and b27 in lacrimal glands of female MRL/lpr mice, and by 38.8-fold the KLK b24 gene activity in female nonautoimmune BALB/c mice. 31 Another gene that might be responsible for increasing the aberrant androgen-induced inflammation in NOD lacrimal glands is Syk. This tyrosine kinase is very much involved in signaling pathways in hematopoietic cells, and also functions within epithelial cells to promote inflammatory responses. 59,60 Syk inhibition has been proposed as a potential treatment for SLE and SS. 61 However, although Syk gene expression is increased in the inflamed lacrimal glands of female MRL/lpr mice (NCBI GEO series accession number GSE5876), it is not decreased by androgen treatment in this strain. Consequently, if there is a specific lacrimal gland switch that androgens turn on to induce immunopathology in NOD mice, and turn off to suppress inflammation in MRL/lpr mice, then Syk is not that switch.
What, then, is that possible on/off switch? We hypothesized that this switch, which may comprise a single or multiple genes, is triggered by an androgen-androgen receptor interaction within lacrimal gland epithelial cell nuclei. These classical androgen receptors are members of the nuclear receptor superfamily of ligand-inducible transcription factors and Relative ratios were determined from CL and Affy data by comparing the degree of gene expression in lacrimal glands from P-versus T-treated female NOD mice. Listed CL genes were increased ‡3.50-fold.
* Genes were found to be down-regulated in lacrimal glands of female MRL/lpr mice treated with T ( Table 7). Relative ratios were calculated from CL data by comparing the degree to which gene expression was significantly down-or upregulated by T treatment, relative to that of P, in lacrimal glands of female MRL/lpr and NOD mice. mediate the majority of androgen actions throughout the body. 62,63 Following androgen association with its specific receptor, the monomeric, activated androgen-receptor complex binds to androgen response elements in the regulatory region of target genes and, in combination with coactivators and enhancers, regulates gene transcription, and ultimately protein synthesis and tissue function. [62][63][64][65][66][67] We have shown that androgen receptors are located almost exclusively within acinar and ductal epithelial cell nuclei in lacrimal glands of MRL/lpr mice, and are absent within the extensive lymphocytic populations in these autoimmune tissues. 68 Moreover, we have found that androgens upregulate the expression of androgen receptor protein in MLR/lpr lacrimal gland epithelia, and this autoregulation is particularly intense in ductal epithelial cells. 68 Indeed, the highest level of androgen receptor protein in ductal nuclei 68 is elicited by those androgens that possess the greatest anti-inflammatory activity in MRL/lpr lacrimal tissue. 14 Given the role of the periductal area in promoting inflammation within the lacrimal gland, 69 it may be that an androgen-controlled on/off switch exists in ductal epithelial cells. Epithelial cells, in turn, are thought to be the primary cells involved in the initiation and perpetuation of glandular autoimmune reactivity in Sjögren syndrome. 70,71 Consistent with a regulatory role for ductal epithelial cells is the finding that infiltration of lacrimal glands in AIRE-deficient NOD mice appears to localize to ductal tissue. 72 AIRE is a transcription factor and autoimmune regulator that enforces self-tolerance; humans expressing a defective form of this gene develop multiorgan autoimmune disease. 73 Interestingly, cor-rection of ductal epithelial function also has been shown to correct acinar epithelial function. 74 This domino effect suggests that ductal cells have an essential role in the pathogenesis of lacrimal gland dysfunction and ultimately aqueous tear film deficiency Why then is there an aberrant androgen immune response in lacrimal glands of NOD mice? Could this response be related to a genetic alteration in the androgen receptor, or to changes in the hypothalamic-pituitary-adrenal (HPA) axis, or to the diabetes that is characteristic of this strain? Defects in sex steroid receptors have been linked to the onset, progression, and severity, as well as the sex-related prevalence, of a number of autoimmune disorders, including lupus, rheumatoid arthritis, and diabetes. 75 These defects often are due to gene polymorphisms or alternative splicing and may lead to marked changes in the affinity or specificity of ligand binding, nuclear translocation, receptor dimerization, DNA association, and transcriptional activation. 75 However, we found that the coding region of androgen receptors in lacrimal glands of NOD and MRL/lpr mice is not defective, but rather normal. 75 As concerns the HPA axis, we previously discovered that hypophysectomy or anterior pituitary ablation significantly interferes with androgen action on the lacrimal gland. 76 This lacrimal gland impairment appears to be tissue-specific. 77 However, although the pituitary has blunted responses in humans with SS, 78 NOD mice have a hyperactive HPA 79 and this would not inhibit androgen effects on lacrimal tissue. With regard to diabetes, insulin deficiency is known to attenuate the lacrimal gland response to androgen, 80 but there is no evidence that this condition would promote a completely Immune-related KEGG pathways that were decreased in T-, as compared to P-, treated female MRL/lpr mice are listed. Immune-related KEGG pathways that were increased in T-, as compared to P-, treated female NOD mice are listed. The number of genes (i.e., non-z-score columns) and z-scores (z) were obtained by analyzing comparative CL microarray data from lacrimal glands from female (F) versus male (M) and P-versus T-treated MRL/lpr (lpr) and NOD mice. The sex-related data originate from one of our recent publications. 30 The last two columns on the right show results obtained by comparing gene expression in inflamed (Infl) versus noninflamed lacrimal tissues, as described in the Results section. Ontologies were significantly (P < 0.05) up ()-or down ()-regulated according to the listed sex and hormone treatment. Relative ratios and P values were calculated from CL data by comparing the degree of gene expression in lacrimal glands from female versus male, P-versus T-treated, and inflamed versus noninflamed MRL/lpr and NOD mice. The categories, abbreviations, and origin of the sex-related data are described in the legend to Table 11. opposite immune response to androgens as found in NOD compared to MRL/lpr mice.
As one additional consideration, it has been proposed that a defect in male-specific, lacrimal gland-protective T regulatory cells is the cause of the lacrimal gland inflammation in NOD mice, and is driven by a T regulatory cell-extrinsic factor. 81 However, given that we were able to induce a striking increase in inflammatory gene expression in lacrimal tissue of NOD female mice, it would seem that androgen action has the key role in this T-cell effector/regulator imbalance.
The androgen-induced up-and downregulation of inflammatory gene expression in NOD and MRL/lpr mice, respectively, appears to be mediated through the lacrimal gland environment. Consistent with this hypothesis are the results of adoptive transfer experiments in NOD mice with severe combined immune deficiency (SCID). These animals lack functional T and B cells and do not suffer autoimmune disease. Transfer of splenocytes or cervical lymph node cells from a female NOD mouse to a male NOD.SCID causes massive inflammatory lesions in the lacrimal gland, whereas transfer of male NOD splenocytes or cervical lymph node cells to a female NOD.SCID does not elicit such lacrimal tissue infiltration. 24,81 Further, the lacrimal gland inflammatory response can be reduced by castration of a male NOD mouse, 23 and induced by androgen treatment of a female NOD mouse (this study).
It is possible that intracrine steroidogenic enzymes convert androgens in the NOD lacrimal gland into metabolites that act through different mechanisms than testosterone, such as may occur in the brain. 82 Such byproducts could have aberrant forms, given that unusual androgen metabolites are the key serum biomarkers for dry eye disease. 83 Alternatively, it is possible that epithelial cells in NOD lacrimal tissue, like human prostate epithelial cells, demonstrate significant plasticity in response to androgens. 84 Nevertheless, the identity of the microenvironmental switch(es) that translate androgen action into an up-or downregulation of immune-related gene expression in the lacrimal gland remains to be discovered.