|Home | About | Journals | Submit | Contact Us | Français|
Scurfy (Foxp3Sf/Y), Il2-/- and Il2rα-/- mice are deficient in CD4+Foxp3+ regulatory T cells (Treg) but only the latter two develop inflammation in the submandibular gland (SMG), a critical target of Sjögren's syndrome. Here, we investigated the reason why SMG of Sf, Sf.Il2-/-, Sf.Il2rα-/-, and the long-lived Sf.Faslpr/lpr mice remained free of inflammation, even though their lymph node cells induced SMG inflammation in Rag1-/- recipients. A strong correlation was observed between the development of the granular convoluted tubules (GCT) of the SMG in these mice and SMG resistance to inflammation. Moreover, GCT development in Sf.Rag1-/- mice was not impeded, indicating a role of adaptive immunity. In the Sf.Faslpr/lpr mice, this block was linked to atrophy and inflammation in the accessory reproductive organs. Testosterone treatment restored GCT expression but did not induce SMG inflammation, indicating GCT is not required for inflammation and additional mechanisms were controlling SMG inflammation. Conversely, oral application of LPS induced SMG inflammation but not GCT expression. LPS treatment induced up-regulation of several chemokines in SMG with little effect on the chemokine receptors on CD4+ T cells in Sf mice. Our study demonstrates that Sf mutation affects SMG development through adaptive immunity against accessory reproductive organs and the manifestation of SMG inflammation in Sf mice is critically controlled through innate immunity.
Sjögren's syndrome is an autoimmune disease affecting the salivary and lacrimal glands, causing dry eyes and dry mouth (1). The mechanism that triggers leukocyte infiltration into these organs remains largely unknown. We observed Sjögren's syndrome-like disease in Il2rα-/- mice and Il2-/- mice; both are partially deficient in the naturally occurring CD4+Foxp3+ regulatory T cells (Treg)3 (2-4). These mice display inflammation in the submandibular gland (SMG) with impaired salivation function (2). Paradoxically, the SMG of Scurfy (Sf) mice which carry the Foxp3sf mutant gene and are totally devoid of Treg is not inflamed (2). Added to the complexity are the observations that transfer of Sf CD4+ T cells into Rag1-/- recipients induced strong inflammation in SMG and that oral application of LPS induced SMG inflammation in Sf mice (2).
Mouse SMG contains three major components, acini, striated ducts and granular convoluted tubules (GCT). The early stages of SMG development occur soon after birth and are dominated by the formation of acini whereas GCT development occurs at 2 wks later and continues to adulthood (5). GCT development is sexually dimorphic, i.e., dominantly expressed in male soon after GCT development begins. Dominant GCT expression in female or castrated male can be induced by testosterone (6). Because Foxp3sf mutation effect is X-linked (7), we wish to address whether the SMG development in Sf mice is blocked as a result of reproductive organ damage and what is the consequence of the SMG growth arrest with respect to inflammation attack.
Here, we report an unusual regulation of organ-specific development and T cell-mediated inflammation uniquely associated with not only the SMG of Sf mice but also the long-lived Sf.Il2-/-, Sf.Il2rα-/- and Sf.Faslpr/lpr mice. We observed that the SMG development based on GCT expression was severely blocked in these mice. Importantly, GCT expression in male Il2-/- and Il2rα-/- mice was moderately inhibited but introducing Foxp3sf gene severely blocked their development and rendered their SMG resistant to inflammation whereas SMG in Sf.Rag1-/- mice developed normally and became inflamed upon transfer of Sf lymph node cells. Atrophy and inflammation were observed in the accessory reproductive organs in long-lived Sf.Faslpr/lpr mice. We could restore GCT dominance by treating Sf.Faslpr/lpr mice with testosterone but their SMG remained free of inflammation. Conversely, daily oral application of LPS to Sf.Faslpr/lpr mice induced SMG chemokine production and inflammation but the GCT growth arrest remained. Thus, the X-linked Foxp3sf gene dominantly inhibits both SMG development and inflammation but by different mechanisms through adaptive and innate immunity, respectively. The significance of the study to the etiology, mechanism of action, and the developmental process of SMG inflammation are discussed.
C57BL/6 (B6), B6.Faslpr/lpr, B6.Il2+/-, B6.Il2rα+/-, B6.Cg-Foxp3sf/+/J, and B6.129S7-Rag1tm/Mom/J (Rag1-/-) mice were obtained from the Jackson Laboratories, Bar Harbor, ME. B6.Il2-/-, B6.Il2rα-/-, Sf mice (Foxp3sf/Y), and Sf.Il2-/- mice were generated as previously described (8, 9). Mice (Sf.Il2rα-/-) carrying both Il2rα-/- and Foxp3sf/Y genes were generated by breeding B6.Il2rα+/- males with B6.Il2rα+/-Foxp3sf/+ females. Mice carrying both of the Foxp3sf and Rag1-/- genes (Sf.Rag1-/-) were generated by breeding Rag1-/- male with B6.Cg-Foxp3sf/+/J mice, followed by breeding Rag1+/-Foxp3sf/+ female progeny with Rag1-/- male. Presence of the Il2-/-, Il2rα-/-, Foxp3sf, Faslpr/lpr and Rag1-/- mutation was confirmed by PCR as detailed in the Jackson Laboratory's website. Mice were examined twice weekly for clinical signs of disease including skin inflammation, body weight loss, wasting etc. All animal experiments were approved by the Animal Care and Use Committee of the University of Virginia.
Adoptive transfer experiments were carried out by intravenous injection of 15×106 lymph node cells into adult Rag1-/- male recipients. Various organs and tissues were harvested at 4-12 wks after transfer. Testosterone treatment was conducted on 4-wk old Sf.Faslpr/lpr mice by twice weekly subcutaneous injection of 25 μl of testosterone oenanthate (Sigma-Aldrich) solution (2 mg/ml in corn oil) or vehicle control. Various organs and tissues were harvested after 4 wks of treatment. Sf.Faslpr/lpr mice and control littermates (4 wks old) were also treated by daily oral feeding of LPS or poly-I:C solution (10 μl of 1 mg/ml, one in the morning and one in the evening) (Sigma-Aldrich). Various organs and tissues were harvested 3 wks later, H&E stained and examined under microscope.
Tissues/Organs were fixed with 10% neutral buffered formalin (Fisher Scientific). Sections of paraffin-embedded samples were stained with H&E and examined under microscope.
The extent of leukocyte infiltration in tissue sections were examined and categorized into five groups, from 0 to 4, to indicate respectively the degree of inflammation as none (0 infiltration lesion), mild (1 lesion), moderate (2 lesions), strong (3-4 lesions), and severe (>4 lesions). The overall inflammation index of samples of a group of mice was calculated as mean± standard deviation. For quantification of GCT, images of H&E stained SMG sections were acquired using Olympus BX51 microscope (Olympus America Inc.) equipped with a digital camera at 100 × magnifications. Images of two random optical fields from at least three different mice of each strain were captured. The images were then analyzed using Kodak 1D image analysis software (Eastman Kodak Company) and the area occupied by the GCT was marked using the “Region of Interest” tool. The total area occupied by the GCT in each image was expressed as “% area occupied by GCT”.
SMG were removed from Sf mice treated either with LPS or PBS and collected in RNA later reagent (Qiagen Inc, Valencia, CA). Pieces of SMG were suspended in RLT buffer and homogenized using TissueLyser system (Qiagen). Total RNA from homogenate was purified using RNAeasy mini-kit, followed by first strand cDNA synthesis using the QuantiTect Reverse Transcription kit (Qiagen). The expression levels of CCL2, CCL3, CCL4, CCL5, CCL9, CXCL2, CXCL10 and CXCL12 were determined by Real-Time PCR on MyiQ machine (Bio-Rad laboratories, Hercules, CA) using the Taqman gene expression assays (Applied Biosystems, Foster City, CA). A RNA sample from the SMG of a normal C57BL/6 mouse was used as a calibration control. This sample was chosen because its cytokine expression pattern closely represents the average pattern obtained from 5 normal samples. The data are represented as fold change in gene expression over the B6 calibrator. Statistical significance of variance was determined by the non-parametric Mann-Whitney test using GraphPad Prism software. A p<0.05 was considered statistically significant.
Flow cytometric staining for various chemokine receptors was conducted on the draining lymph nodes of SMG in Sf mice treated with PBS and LPS. Age-matched, male B6 mice were included for comparison. Lymphocytes were stained with anti-CD4 mAb along with antibodies specific to CCR3 (TG14/CCR3), CCR5 (HM-CCR5), CXCR2 (TG11/CXCR2), and CXCR3 (CXCR3-173) (Biolegend, San Diego, CA). The expression of chemokine receptors on gated CD4+ T cells was presented.
II2-/- and II2rα-/- but not Sf mice developed spontaneous inflammation in SMG even though Sf mice contained T cells capable of inducing SMG inflammation in Rag1-/- recipients (2). To determine if the resistance is inherently associated with Foxp3sf/Y mice, we introduced Foxp3sf gene into male II2-/- and II2rα-/- mice to generate Sf.II2-/- and Sf.II2rα-/- mice, respectively. In all cases, these mice lived significantly longer than Sf mice but their SMG remained free from inflammation (8). By contrast, inflammation in colon normally observed in Il2-/- and Il2rα-/- mice remained (Fig. 1A).
We transferred lymph node cells from Sf, and Sf.II2-/- mice to determine if they contained competent cells capable of inducing SMG inflammation in Rag1-/- recipients. In addition, we also tested Sf.Faslpr/lpr mice because these mice lived more than 15 wks and their SMG remained free of inflammation (see Fig. 3A below). All induced moderate/strong inflammation in the SMG of the recipients (Fig. 1B). Taken together, these results indicate that the SMG resistance to inflammation in Sf, Sf.Il2-/-, and Sf.Faslpr/lpr mice occurs even in the presence of competent inflammation-inducing T cells and this resistance is inherently and dominantly associated with Foxp3sf in male mice in an organ-specific manner, i.e., inflammation is inhibited in SMG but not colon.
We compared the SMG expression level between male and female Il2-/- and Il2rα-/- mice that were 8 wks old using the semi-quantification method. B6 mice were used as control. As shown in Fig. 2A, a strong age-dependent expression of GCT was observed in B6 male as opposed to the weak expression of GCT with less granule content in B6 female. The SMG of Il2-/- and Il2rα-/- mice also displayed sexual dimorphism but the GCT in male Il2-/- and Il2rα-/- mice was less than age-matched B6 male. This could be caused by SMG inflammation that induces GCT atrophy (Fig. 2B) (2). By contrast, the SMG of Sf.Il2-/- and Sf.Il2rα-/- mice were severely under-developed with fewer and smaller GCT than male Il2-/- and Il2rα-/- mice (Fig. 2B). Indeed, highly statistically significant differences in GCT expression levels between Il2-/- and Sf.Il2-/- mice and between Il2rα-/- and Sf.Il2rα-/- mice were observed (Fig. 2C). Unlike Il2-/- and Il2rα-/- mice, the SMG growth arrest in Sf.Il2-/- and Sf.Il2rα-/- mice occurred in the absence of SMG inflammation, indicating that the inhibition of GCT growth in this case is not caused by local inflammation, but rather the consequence of the early systemic inflammatory response. In support of this, we observed SMG sexual dimorphism between Sf.Rag1-/- male and Sf.Rag1-/- female, both lacked a functional adaptive immune response (Fig. 2D). Moreover, transfer of Sf lymph node cells into male Sf.Rag1-/- recipients readily induced SMG inflammation (Fig. 2E). These results indicate that Foxp3sf/Y indirectly controls SMG development and SMG inflammation through the adaptive immune system.
Foxp3 expression has been demonstrated in the epithelial cells of mammary and prostate glands (10). Foxp3 expression in the SMG has not been determined. The facts that SMG of Sf.Rag1-/- mice displays sexual dimorphism for GCT expression and becomes inflamed upon transfer of Sf lymph node cells indicate that Foxp3, if expressed in SMG per se, does not play a role in the regulation of SMG development and the SMG resistance to inflammation-inducing T cells.
Whether GCT growth arrest was due to a response of SMG to the systemic inflammation or lack of testosterone resulting from the inflammation of the reproductive organs was investigated. The latter possibility was considered because Sf.Faslpr/lpr, despite living beyond 15 wks, remained reproductive incompetent. We examined the gross structure and histological features of the reproductive organ of 6-8 wks old Sf.Faslpr/lpr mice with age-matched B6 male (Fig. 3). In gross examination, we observed tremendous atrophy in the various accessory reproductive organs of the Sf.Faslpr/lpr mice but not of the B6 male. These included coagulation glands/seminal vesicle (black arrows, top panel), preputial glands (red arrowheads), epididymis (black arrowheads) and prostate (row 3, Fig. 3). Histological examination confirmed the atrophy of these organs, displayed as shrunken glands and empty lumens. This was accompanied with a prominent presence of leukocytic infiltrates in the peri-glandular regions (green arrowheads). Leukocytic infiltrates could also be observed in the interstitial regions of the testis and in the regions containing the interstitial cells of the Leydig between the seminiferous tubules (green arrowheads). The data suggest that the inflammation-induced atrophy of the reproductive organs was responsible for the arrest of GCT development. As testosterone is the major hormone responsible for the development of GCT and reproductive organs, we treated Sf.Faslpr/lpr mice with the long-lasting testosterone oenanthate as described in Materials and Methods (6, 11). Testosterone successfully restored the growth of the accessory reproductive organs of Sf.Faslpr/lpr mice (Fig. 3, right column) but the leukocytic infiltrates in the reproductive organs were still present. Thus, the potential anti-inflammatory effect of testosterone is unlikely the reason for the restoration of growth and development of inflamed organs. The data suggest that atrophy of the reproductive organs was caused by the lack of testosterone resulted from inflammation in the reproductive organs. However, restoration of organ development by the injected testosterone could not reverse the inflammation.
Testosterone treatment also fully restored the dominant GCT expression in the SMG of Sf.Faslpr/lpr mice (Fig. 4A). Both the number of GCT and the granule content were restored to the level of normal B6 male. The observation supports the interpretation that the lack of testosterone from the atrophic reproductive organs was responsible for the GCT growth-arrest. Despite a fully restored SMG, the SMG remained free of inflammation, suggesting GCT is not required for SMG inflammation. The treatment did not inhibit inflammation in other organs normally observed in Sf.Faslpr/lpr mice such as skin, lungs, and the accessory reproductive organs (8, Fig. 3).
In the second approach, we treated 4 wks old Sf.Faslpr/lpr mice with LPS that has been shown to induce SMG inflammation in Sf mice (2). The latter experiment treated one wk old Sf mice and SMG inflammation was determined 2 wks later, i.e., at 3 wks old when normal GCT expression was not obvious. In the present experiment, SMG inflammation was determined at 8 wks old. In addition, we also treated these mice with a different TLR agonist poly-I:C. A strong inflammation was induced by both treatments but the GCT expression of the treated mice remained inhibited (Fig. 4B). The two approaches provided reciprocal evidence that dissociate GCT growth inhibition from resistance to inflammation attack.
Several chemokine receptors/integrins are strongly up-regulated on the CD4+ T cells in Sf mice (12). Yet these T cells failed to enter the SMG. We hypothesize that LPS can directly act on SMG by increasing their chemokine production. This change sensitizes the SMG to the inflammation-inducing T cells already present in the Sf mice. We analyzed the chemokine expression pattern of the SMG of control Sf mice and those treated with LPS using quantitative RT-PCR. We chose eight chemokines including those that are known to be up-regulated upon LPS stimulation (13). Four chemokines, CCL3, CCL5, CXCL2, and CXCL10 were significantly increased in the LPS-treated Sf mice as compared with PBS-treated control (Fig. 4C). Although the expression levels of CCL2, CCL4, CCL9, and CXCL12 were not significantly up-regulated, additional samples are needed to firmly determine if this is really the case with CCL4. These data demonstrated that LPS induced a direct change in the SMG of Sf mice and the under-developed SMG in Sf mice is capable of responding to LPS by producing chemokines.
We then analyzed CD4+ T cells of the SMG draining lymph nodes of these mice for the expression of chemokine receptors specific to the up-regulated chemokines. The expression of chemokine receptors targeted by CCL3, CCL5, CXCL2, and CXCL10 showed considerable variability and the differences between PBS-treated and LPS-treated Sf mice were not significant (Fig. 4D). Although a strong trend of increase in their expression was observed when compared with B6 controls, the difference was not significance except for CXCR2. Apparently, the overall mild increase in these trafficking/retention receptors on CD4+ T cells of Sf mice was not sufficient to induce inflammation unless increase in chemokine production in SMG was achieved by LPS treatment. These observations provided evidence that LPS acted on SMG chemokine production and induced inflammation without affecting GCT development.
Our study provides evidence for the following scheme for SMG development and inflammation response. The genetic defect in Sf mice results in the inhibition of two critical components of the process: one required for GCT expression and one for the susceptibility to inflammation, both depended on a functional immune system. A consequence of the severe and early systemic inflammatory response in Sf mice is the inhibition of GCT-inducing agents such as testosterone by the adaptive immunity. Testosterone treatment induced GCT over-expression but not the component controlling susceptibility to inflammation-inducing T cells. The reciprocal result, obtained with LPS and poly-I:C treatments, demonstrated that TLR agonist induced SMG inflammation but not GCT development in Sf mice. Evidence provided strongly suggests that the inflammation in the under-developed SMG is limited by its innate immune response which can be overcome by TLR agonists through chemokine induction.
We have previously shown that daily LPS application induces SMG inflammation in Sf mice (2). In the present study, daily LPS application to the long-lived Sf.Faslpr/lpr mice was also able to induce SMG inflammation, indicating that although the key defect in the induction of SMG inflammation is long-lasting, the ability to break it by LPS remained as well. In addition, an agonist for a different TLR also induced SMG inflammation in these mice. Although multiple intraperitoneal injections of these agonists into adult (NZWxNZW)F1 mice activate innate immunity in the SMG (14), we chose oral applications to neonates for ease of application and less injury. It has been shown that LPS directly affects various organs through circulation (15). We showed that LPS treatment up-regulated the production of several chemokines in the SMG of Sf mice, thereby indicating that oral applications of LPS can act on the SMG of Sf mice and induce biological changes. In contrast, the corresponding chemokine receptors on CD4+ T cells were not up-regulated upon LPS treatment. These observations strongly suggest that the neonatal and under-developed SMG in Sf mice is limited in its ability to produce sufficient chemokines required for the attraction of inflammation-inducing CD4+ T cells to cause SMG pathology and salivation dysfunction.
As compared with normal B6 male, the SMG of the sex- and age-matched Il2-/- and Il2rα-/- mice are under-developed but not severely growth-arrested as the SMG of Sf.Il2-/- and Sf.Il2rα-/- mice. Unlike Sf or Sf.Faslpr/lpr mice, Il2-/- and Il2rα-/- mice still contain a significant level of Treg. In addition, they lack IL2 and IL-2Rα, respectively. Thus, the inflammation process in Il2-/- and Il2rα-/- mice is reduced and delayed as compared with the corresponding Sf.Il2-/- and Sf.Il2rα-/- mice. Like Sf.Il2-/- and Sf.Il2rα-/- mice, Il2-/- and Il2rα-/- male are reproductive incompetent with light body weight and probably lack normal level of testosterone but they live significantly longer than Sf.Il2-/- and Sf.Il2rα-/- mice. These factors could account for the weak inhibition of SMG development in male Il2-/- and Il2rα-/- mice. In this regard, the relatively more advanced SMG development in these mice with appropriate environmental changes may allow sufficient chemokine production to attract inflammation-inducing CD4+ T cells to SMG and induce inflammation as the mice lived beyond weaning to adult age.
Several hormones can induce dominant GCT over-expression in female or castrated male (5, 6). We examined the long-lived Sf.Faslpr/lpr mice and observed dramatic atrophy and significant inflammation in many of the accessory reproductive organs. Specifically, severe atrophy was observed in the seminal vesicle, prostate and epididymis with varying degree of inflammation. Spontaneous inflammation in multiple male accessory reproductive organs is rare and our study showed it for the first time that it does indeed develop in Sf.Faslpr/lpr mice that lived beyond weaning and reach adult age. This finding strongly suggests that testosterone is the main reason that Sf mice fail to display a normal male dominant presentation of the GCT. Indeed, our testosterone treatment fully restored GCT development but the SMG remained free of inflammation. Conversely, testosterone treatment also restored the development of the accessory reproductive organs but the inflammation remained. This result dissociates the testosterone-dependent organ development from inflammation induced by Treg-deficiency in both directions, i.e., inflammation in accessory reproductive organs or lack of inflammation in SMG was maintained regardless of testosterone-induced organ development. This result also suggests that lacking GCT is not the reason for the inability of Sf mice to develop SMG inflammation.
Despite totally lacking Treg, many of the organs/tissues such as central nervous system, joints, endocrine and mucosal organs are free of inflammation in Sf mice despite the fact they have a large functional inflammatory repertoire (16, 17). In this regard, the present study has implication toward the genetic control of regulatory mechanisms involved in the multi-organ inflammation in Sf mice. In contrast to the Foxp3Sf/y -dependent regulation of SMG inflammation in Sf.Il2-/- mice, the Il2-/- defect conversely inhibited the skin and lung inflammation in Sf mice but not the liver inflammation and colitis associated with Il2-/- mice, i.e., Sf.Il2-/- mice do not develop skin and lung inflammation (12). This apparent “organ-specific” control of inflammation is in fact mediated by different mechanisms. It turns out that IL-2 is required for the expression of a panel of receptors involved in the trafficking/homing and retention of the pathogenic CD4+ T cells to skin and lungs. Among the receptors that are highly up-regulated are cysteinyl leukotriene receptor 1, leukotriene B4 receptor 1, CCR8, CXCR6, IL-1R-like1, and CD103 of the αEβ7 (12). We have extensively characterized the role of IL-2 in the up-regulation of CD103 of the αEβ7 integrin on CD4+ T cells and their retention in the skin and lungs. The fact that these pathogenic CD4+ T cells failed to infiltrate SMG of Sf mice suggests either TLR agonists induce the ligands for these receptors for the infiltration of the T cells or that a different set of chemokine receptors and/or integrins are involved. The result of LPS treatment suggests that the latter case is likely a possible mechanism. However, whether LPS induces the specific ligands for each of these up-regulated receptors has not been exhaustively determined. Additional study is needed to establish the exact mechanism by which SMG inflammation is induced by LPS treatment.
We thank Mr. C. Abaya and Ms. Angela Ju for their technical assistance and Dr. K. Tung for his critical review of the manuscript.
1Supported by in part by NIH grants AR-051203 (STJ), DE-017579 (STJ), AR-045222 (SMF), AR-047988 (SMF), and AR-049449 (SMF), AR051391 (USD) and a grant from Sjogren's Syndrome Foundation (USD).
3Abbreviations: GCT, granular convoluted tubules; Sf, Scurfy; SMG, submandibular gland.