Altered Macrophage and Dendritic Cell Response in Mif-/- Mice Reveals a Role of Mif for Inflammatory-Th1 Response in Type 1 Diabetes.

Macrophage migration inhibitory factor (Mif) is highly expressed in type 1 diabetes mellitus (T1DM). However, there is limited information about how Mif influences the activation of macrophages (Mφ) and dendritic cells (DC) in T1DM. To address this issue, we induced T1DM by administering multiple low doses of streptozotocin (STZ) to Mif-/- or wild-type (Wt) BALB/c mice. We found that Mif-/- mice treated with STZ (Mif-/-STZ) developed lower levels of hyperglycemia, inflammatory cytokines, and specific pancreatic islet antigen- (PIAg-) IgG and displayed reduced cellular infiltration into the pancreatic islets compared to Wt mice treated with STZ (WtSTZ). Moreover, Mφ and DC from Mif-/-STZ displayed lower expression of MHC-II, costimulatory molecules CD80, CD86, and CD40, Toll-like receptor- (TLR-) 2, and TLR-4 than WtSTZ. These changes were associated with a reduced capacity of Mφ and DC from Mif-/-STZ to induce proliferation in ovalbumin-specific T cells. All the deficiencies observed in Mif-/-STZ were recovered by exogenous administration of recombinant Mif. These findings suggest that Mif plays a role in the molecular mechanisms of Mφ and DC activation and drives T cell responses involved in the pathology of T1DM. Therefore, Mif is a potential therapeutic target to reduce the pathology of T1DM.


Introduction
T1DM is an autoimmune-mediated disease characterized by selective destruction of insulin-producing pancreaticcells, resulting in the need for lifelong administration of exogenous insulin for patient survival [1], and represents 5-10% of all cases of diabetes [2,3]. The pathological condition begins with an autoimmune inflammatory process known as insulitis, which leads to leucocyte infiltration into the pancreatic islets of Langerhans, resulting in the autoimmune destruction of pancreatic -cells [4]. Thus, the onset of T1DM is the result of the recognition of self-antigens due to molecular mimicry against infection with various viruses such as Coxsackie virus, -cell cytotoxicity, or -cell cytolysis [5][6][7]. Consequently, molecules such as preinsulin, insulin, glutamate decarboxylase-65 (GAD- 65), and islet antigen 2 become recognized as self-antigens [8].
Mif is a pleiotropic cytokine produced during the immune response by activated T cells, M , DC, and a variety of nonimmune cells and plays a pivotal role in the systemic inflammatory immune response by promoting the production of proinflammatory cytokines including TNF-and IL-6, which are involved in inflammatory and autoimmune diseases such as septic shock [25], cancer [26], inflammatory bowel disease [27,28], rheumatoid arthritis [29,30], obesity [31,32], and diabetes [33][34][35]. Moreover, Mif has recently been proposed as a diagnostic biomarker for autoimmune diseases [36] such as arthritis [37,38], type 2 diabetes [35], and ulcerative colitis [39] in both animals and humans.
The pathogenic contribution of Mif to T1DM was demonstrated by showing that Mif mRNA expression was upregulated in splenic lymphocytes during the development of spontaneously diabetic nonobese diabetic (NOD) mice, as well as cyclophosphamide-treated NOD mice. Diabetes incidence was increased to 86% in NOD mice treated with recombinant Mif (rMif) protein, compared with the 55% incidence observed in untreated control NOD mice [20]. Furthermore, studies performed using Mif−/− mice rendered diabetic by administering multiple low doses of streptozotocin (STZ) have shown that the absence of Mif affected several aspects of experimental T1DM, including initial immunopathological events and the production of proinflammatory and cytotoxic mediators, thereby interfering with both inflammation and tissue destruction [22].
All the results described above provide evidence that Mif plays a critical role in the pathogenesis of T1DM. However, the precise mechanism by which Mif promotes insulitis and the mechanism underlying its proinflammatory role remain unclear. The activities of Mif may reside at the levels of both the inductive and effector phases of the inflammatory response attributed to antigen-presenting cells.
Here, we analyzed the influence of Mif on M and DC activation using an autoimmune diabetes model in which multiple low doses of STZ were administered to Mif−/− and wild-type (Wt) mice (Mif−/−STZ and WtSTZ, resp.) in the BALB/c genetic background. As previously demonstrated Mif−/−STZ developed less severe hyperglycemia, reduced levels of IFN-and TNF-, a smaller amount of pancreatic islet antigen-(PIAg-) specific IgG, and decreased cell infiltration into the pancreatic islets compared to WtSTZ. Interestingly, we found for the first time that M and DC from Mif−/−STZ displayed decreased expression of costimulatory molecules CD80, CD86, and CD40, as well as Toll-like receptor-(TLR-) 2, TLR-4, and major histocompatibility complex-(MHC-) II. Importantly, we demonstrated that due to diminished upregulation of costimulatory molecules, these cells exhibited a reduced capacity to induce proliferation and cytokine expression in cocultures with allogeneic ovalbumin-(OVA-) specific T cells. All deficiencies observed in Mif−/−STZ were reversed by exogenous rMif protein administration.

Ethics Statement.
All experiments in this study were performed according to the guidelines in the Mexican Regulations on Animal Care (NOM-062-ZOO-1999, 2001) and were approved by the local Institutional Animal Care and Use Committee. All efforts were made to minimize animal suffering over the course of these studies.

Animals.
Six-to 8-week-old male BALB/c mice were purchased from Harlan Laboratories (Mexico City, Mex.) and were maintained as a breeding colony in a pathogenfree environment at our animal facilities in accordance with institutional guidelines. Mif−/− mice were kindly provided by Dr. Abhay R. Satoskar (The Ohio State University, USA) and were maintained as breeding colonies for more than 10 generations in the BALB/c genetic background on the Transgenic Mouse Core Facility at our institution. Genotyping of Mif−/− mice was routinely performed on DNA isolated from tail snips using a PCR procedure [40]. The PCR were performed using the following primers: Mif: forward 5 -AGACCACGTGCTTAGCTGAG-3 and reverse 5 -GCATCGCTACCGGTGGATAA-3 ; Neomycin (Neo): forward 5 -ATTGAACAAGATGGATTGCAC-3 and reverse 5 -CGTCCAGATCATCCTGATC-3 . PCR for the amplification of Mif and NEO was performed by adding 100 ng of the extracted DNA to 25 L of a reaction mixture that contained 18.4 L of distilled water, 2.5 L of 10x PCR buffer, 0.4 L of dNTPs (10 mM), 1.5 L of MgCl 2 (25 mM), 1 L of the forward and reverse primers, and 0.2 L (2.5 units) of Taq DNA polymerase (Ampliqon, Bioreagents and Molecular Diagnostics). The amplification protocol consisted of 5 min at 95 ∘ C, 35 cycles of 95 ∘ C for 30 sec, 58 ∘ C for 40 sec, and 72 ∘ C for 30 sec and a final extension at 72 ∘ C for 5 min. All PCR experiments were conducted with positive and negative controls. A PCR fragment of 200 bp, corresponding to Mif, or 500 bp, corresponding to NEO, was visualized to identify Wt or Mif−/− mice, respectively. The PCR products were analyzed by electrophoresis on a 1.5% agarose gel and were viewed under UV light (Bio-Rad, USA).

Induction of T1DM
. Mif−/− and Wt mice were deprived of food for 8 h before induction of diabetes via intraperitoneal (i.p.) injection of STZ at doses of 40 mg/kg of body weight, daily for five consecutive days (days 0-4) (Sigma-Aldrich, St. Louis, MO, USA). STZ was diluted in cold 0.01 M citrate buffer (pH 4.5) and was used within 5 min of preparation, in accordance with a previously reported protocol [41]. Healthy mice from each group received i.p. injections of an equivalent volume of vehicle (citrate buffer) as negative controls.

Analysis of Blood Glucose, Serum Insulin, and Cytokine
Levels. Blood samples were collected by tail snipping from Wt and Mif−/− mice that had been fasting for 6 h. Samples were obtained once before STZ injection and 2, 4, and 8 weeks after STZ injection. Blood glucose levels were measured with a portable glucometer (Accu-Chek Sensor glucometer; Roche Diagnostics, Indianapolis, IN, USA). Mice with a glucose concentration exceeding 300 mg/dL were considered to have T1DM. Blood was collected and centrifuged at 1300 ×g, and the serum levels of Mif (Neobiolab, USA), IL-12, IFN-, IL-17, IL-4, IL-13 (PeproTech, Mex), and insulin (Lincoln, St. Charles, MO, USA) were determined via ELISA according to the manufacturer's instructions.

Pancreatic Islets Antigen (PIAg) Isolation.
PIAg islets were obtained from healthy Wt mice as described below. After isolation, the islets were lysed by five freeze-thaw cycles followed by sonication (six 60-Hz cycles for 1 min each) on ice. After centrifugation (10 000 g, 15 min, 4 ∘ C), supernatants were collected and filtered through a 0.2 m membrane (Corning, Cambridge, MA, USA). The protein concentration was determined using the Lowry method [42], and PIAg aliquots were stored at −70 ∘ C until use.
2.6. Analysis of IgG Antibody Production. Serum samples were analyzed for the levels of pancreatic islet-specific Th1associated IgG2a and Th2-associated IgG1 antibodies by ELISA. Briefly, 96-well ELISA plates (Costar) were coated with 100 L/well soluble PIAg in Tris buffer, pH 7.8. The plates were incubated overnight at 4 ∘ C. Then, the wells were washed thoroughly with phosphate-buffered saline (PBS) containing 0.05% Tween-20 (PBS-Tween; Merck, France) and were blocked with PBS supplemented with 1% bovine serum albumin (PBS-BSA; Sigma-Aldrich) for one hour at room temperature (RT). The serum samples were diluted 1 : 100, followed by serial dilution of each sample (in PBS-BSA) from healthy or STZ-treated Wt or Mif−/− mice. The plates were then incubated at 4 ∘ C overnight. After extensive washing with PBS-Tween, the samples were incubated for 45 min at RT with isotype-specific peroxidase-labeled goat anti-mouse antibodies (anti-IgG1 and anti-IgG2a at 1/1000 dilutions; Zymed, San Francisco, CA, USA). Then, the plates were washed, and immunoreactivity was detected with ABTS solution (Zymed). The results were expressed as endpoint titers based on optical density.

Cell and Pancreatic Islet
Isolation. Spleen and pancreatic islet cells from WtSTZ and Mif−/−STZ were collected after 0, 2, 4, and 8 weeks of injection with STZ and stained for flow cytometry analysis. Briefly, the spleen was removed under sterile conditions, and spleen cells were obtained by mincing and filtering the tissue, followed by washing and suspension in DMEM culture medium supplemented with 10% fetal bovine serum, 100 units of penicillin/streptomycin, 2 mM glutamine, and 1% nonessential amino acids (all from GIBCO, BRL, Grand Island, NY, USA). Spleen cells were suspended at 5 × 10 6 cells/mL in the same medium. The pancreas was also removed under sterile conditions, and pancreatic islets were isolated using the collagenase method as previously described [44]. Briefly, the pancreas was removed and cut into small pieces (approximately 3 mm in size). The tissue was subsequently incubated in collagenase (0.3 mg/mL; Roche Diagnostics Corp., Indianapolis, IN, USA) for 10 minutes at 37 ∘ C in a total of 1 mL of digestion solution under constant shaking and intermittent vortexing. Islets were subsequently washed several times in HBSS containing BSA (5 mg/mL) and were hand-picked under a dissecting microscope. Islets were dispersed into single cells by suspension in trypsin-EDTA (GIBCO, BRL) and passage through a siliconized Pasteur pipette. Then, the cells were incubated at 37 ∘ C.
2.9. Flow Cytometry Analysis. Cells from the spleen or pancreas obtained as described above were then used for flow cytometry analysis. In brief, cells were washed in flow cytometry wash solution (Dulbecco's PBS containing 1% FCS and 0.05% sodium azide), followed by incubation with allophycocyanin-(APC-) conjugated anti-F4/80 and anti-CD11c antibodies for differentiation of M and DC, respectively. Then, the selected cells were incubated in 3% BSA-PBS containing a phycoerythrin-(PE-) labeled anti-CD80, anti-CCR5, or anti-TLR-4 antibody or a fluorescein isothiocyanate-(FITC-) labeled anti-CD86, anti-CD40, anti-MHC-II, or anti-TLR-2 antibody (all antibodies from Biolegend, San Diego, CA, USA) at 4 ∘ C for 30 min. After incubation, the cells were washed several times in buffer, fixed in 1% paraformaldehyde (Sigma-Aldrich), and stored at 4 ∘ C in the dark, followed by analysis using a FACSCalibur flow cytometer and CellQuest software (Becton Dickinson, Franklin Lakes, NJ, USA).

Coculture of Macrophages with Spleen Cells.
Coculture of M with naive spleen cells was performed as follows. Adherent M among peritoneal exudate cells (PECs) from healthy or 8 weeks after STZ-treatment Wt or Mif−/− mice were obtained. Briefly, the M density was adjusted to 5 × 10 6 cells/mL, and the cells were plated (100 L) in 96-well flat-bottom plates (Costar, Cambridge, MA, USA). Three hours later, the PECs were washed three times with warm sterile PBS to remove nonadherent cells, and 10 g of OVA (Worthington, USA) in 100 L of DMEM supplemented media was added. Three hours later, adherent M were washed three times with warm sterile DMEM to remove excess OVA that had not been phagocytosed. Spleen cells from OVA-transgenic mice were obtained as previously described [45], suspended at 1 × 10 6 cells/mL, and added (100 L) to PECs at a ratio of 1 M : 5 spleen cells. The cocultures were maintained at 37 ∘ C in 5% CO 2 for 5 days. Then, [3H]-thymidine (185 GBb/mmol activity, Amersham, UK) was added at 0.5 Ci/well, and the cells were incubated for a further 18 h. The cells were harvested using a 96-well harvester (Tomtec, Toku, Finland) and then counted using a microplate counter (Trilux, Toku, Finland). The values are presented as counts per minute (CPM) from triplicate wells.

Statistical Analysis. Comparisons between Wt and
Mif−/− mice that were healthy or treated with STZ were performed using either Student's unpaired -test or ANOVA followed by Turkey's multiple comparisons test for data that displayed a normal distribution. values less than 0.05 were considered significant and were designated as * < 0.05, * * < 0.01, or * * * < 0.001. All data were analyzed using GraphPad Prism 6 software (San Diego, CA, USA).

Mif−/− Mice Developed Less Severe Hyperglycemia Than
Wt Mice after STZ Administration. We first investigated the clinical effects of STZ on Mif−/− and Wt mice. Our data demonstrated that Wt mice rapidly developed hyperglycemia after STZ administration. These mice sustained high blood glucose levels from week 2 (281.7 ± 16 mg/dL) until week 8 of STZ administration (470.5 ± 24 mg/dL) (Figure 1 WtSTZ showed a peak blood insulin level at 2 weeks after STZ treatment, and after the sixth and eighth weeks of treatment, their blood insulin levels were significantly lower than those of healthy Wt mice (Figure 1  The serum levels of the anti-inflammatory cytokine IL-4 were significantly lower in WtSTZ (squares) than in  Mif−/−STZ (inverted triangles) (Figure 2(e)). No differences in the serum IL-13 levels were found between WtSTZ and Mif−/−STZ (Figure 2(f)).  [41].

Mif−/− Mice Showed Reduced Pancreatic Islet
To assess the damage to cells in a model of T1DM induced by multiple low doses of STZ, we sacrificed the mice at 8 weeks after T1DM induction and removed the pancreas for H&E staining and histological analysis. We assessed the number and size of normal and infiltrated pancreatic islets in each slide for the experimental and healthy mice. Thirty pancreatic islets were quantified per experimental group.
Islets from healthy Wt and Mif−/− mice had a round morphology and well-defined borders without cellular infiltration (Figures 3(a) and 3(c)). As expected, the pancreas from WtSTZ showed fewer and smaller islets than those from healthy Wt mice. Additionally, evident cellular infiltration led to the breakdown of islet morphology in the pancreas of WtSTZ (Figure 3(b)). In contrast, partial islet damage (Figure 3(d)) and no significant reduction in islet number or size were observed in Mif−/−STZ compared to WtSTZ ( Table 1).
As shown in Figure 3(e), Mif−/− mice at 8 weeks after STZ administration displayed marked reductions in both invasive insulitis (insulitis grades 3 and 4) and mild periinsulitis (insulitis grade 2) in pancreatic islets compared to WtSTZ. Thus, pancreatic islets from Mif−/−STZ had reduced damage associated with reduced leucocyte infiltration compared to pancreatic islets from WtSTZ. These results demonstrated that STZ administration triggers -cell destruction followed by insulin deficiency and hyperglycemia in Wt mice. Although STZ reached the pancreatic islets in Mif−/− mice, Mif−/−STZ exhibited minor damage compared to WtSTZ.

Mif−/− Mice Showed Lower Levels of Specific Pancreatic
Islet Antibodies Than Wt Mice after T1DM Induction. In T1DM, the immune system targets self-antigens within pancreatic islets and destroys the inhabiting insulin-secretingcells. Therefore, autoantibody detection serves as a predictive factor for the onset of diabetes in both humans and mice [46,47].
To investigate how Mif contributes to autoantibody production in T1DM, the serum levels of pancreatic islet-specific IgG2a and IgG1 antibodies were determined in Wt and Mif−/− mice after 2, 4, and 8 weeks after STZ administration. WtSTZ produced high levels of IgG2a at all time points analyzed (Figure 4(a)) and IgG1 only at week 8 (Figure 4(b)). Importantly, Mif−/−STZ mice displayed significantly lower levels of both IgG2a and IgG1 than WtSTZ (Figures 4(a)  and 4(b), resp.), suggesting suppressed development of an adaptive immune response. These findings suggest that Mif modulates the incidence and severity of diabetes by favoring the development of autoantibodies in T1DM.    Figure 4: Kinetics of antibody responses in type 1 diabetes mellitus models induced by low-dose STZ. Specific production of IgG2a (a) and IgG1 (b) antipancreatic islet antigen antibodies. Sera from Wt (◼) and Mif−/− mice ( ) at weeks 0, 2, 4, and 8 after STZ administration were processed by ELISA, using pancreatic islets antigen (PIAg) as the source of specific antigen as indicated in Section 2. Data are expressed as the mean reciprocal endpoint titer ± SEM. Five to six animals were analyzed in each group. The data shown are representative of one out of three identical experiments with similar results. Mann-Whitney U test. * * < 0.01, * * * < 0.001, GraphPad Prism software 6.

F4/80 + Macrophages and CD11b + Monocytes Obtained from Mif−/−STZ Show Impaired Capacity to Induce Spleen
Cell Proliferation. The ability of M to activate T cells was investigated using OVA-transgenic T cell cocultures. M were collected from Wt or Mif−/− mice treated or untreated for 8 weeks with STZ. As shown in Figure 9(a), after priming the cells with OVA in vitro, M from Mif−/−STZ mice induced less T cell proliferation in response to OVA than M from WtSTZ. A similar trend was observed in cocultures of CD11b + cells from Mif −/−STZ and OVA-transgenic T cells (Figure 9(b)). These data suggest a role of Mif in promoting M activation, which in turn induces specific T cell proliferation, particularly in this experimental T1DM model.  (Figure 10(a)).

Mif Reconstitution in STZ-
Comparable serum levels of the cytokines IL-6 and IL-12 were observed between Mif−/−STZ+rMif and WtSTZ during the first 6 weeks after STZ treatment. However, the levels of these inflammatory cytokines significantly decreased in Mif −/−STZ+rMif mice at 8 weeks after STZ treatment compared to WtSTZ (Figures 10(b) and 10(d)). Interestingly, the serum levels of TNF-from Mif−/−STZ+rMif were higher than those of WtSTZ at all time points analyzed (Figure 10(c)). These results confirm that Mif acts as a powerful inducer of proinflammatory cytokines involved in the development of experimental T1DM.

Discussion
Currently, there is no doubt that Mif is a key molecule that promotes proinflammatory immune responses [48]. This proinflammatory property of Mif contributes to developing protective inflammatory-Th1 immune response in different models of parasitic diseases [49]. In contrast the same proinflammatory property of Mif participates in the pathogenesis of many inflammatory diseases [50]. In line with the last one, recently it has been established that high blood levels of Mif are associated with human T1DM, similar to the findings in experimental mouse models of T1DM [20,21,24,51]. Studies in NOD mice or mice treated with multiple low doses of STZ (despite the pathogenic differences between these models) have shown that pancreatic cell destruction results from the toxic effect of free radicals (O 2 − , H 2 O 2 , and nitric oxide) and inflammatory cytokines released by activated M and T cells [15,41,52]. Therefore, both models have been widely used to dissect the role of Mif in the pathogenesis of T1DM using anti-Mif monoclonal antibody treatment or using Mif −/− mice. In all cases, the lack of Mif resulted in diminished manifestation of the disease, decreased glucose blood levels, and reduced production inflammatory cytokines associated with the development of T1DM, including TNF-, IL-1 , IFN-, IL-12, and IL-23 [21,22,24]. Our study validates and extends these findings by demonstrating an important role for Mif in promoting costimulatory molecule expression in M and DC during T1DM development. We further demonstrate that, in addition to regulating M and DC activation in T1DM, M isolated from T1DM Mif−/− mice exhibit reduced T cell activation.
Here, we observed that Wt mice treated with STZ exhibited high blood glucose levels greater than 400 mg/dL but Mif−/−STZ mice developed lower glucose levels of approximately 200 mg/dL, in association with lower serum levels of proinflammatory cytokines. After reconstituting Mif using exogenous rMif, Mif−/−STZ showed blood glucose levels and IL-6 and IL-12 levels similar to those in WtSTZ from 2 to 6 weeks after STZ treatment. Interestingly, TNFserum levels from Mif−/−STZ+rMif were higher than those of WtSTZ at all time points analyzed. The reduction levels of blood glucose, IL-6, and IL-12 observed in Mif−/−STZ mice, on week 8, probably were because the animals did not receive rMif injections on week 8, as mentioned in Section 2, so the residual effect of rMif from week 1 to week 7 was insufficient to induce IL-6 and IL-12 levels at 8 week similar to that observed in the WtSTZ mice in this point, at least for these two cytokines. These observations confirm that Mif exogenous acts as a powerful inducer of IL-6 and IL-12, but only if it is present in steady high concentration.
In addition, we found that Mif−/−STZ did not produce detectable levels of islet autoantibodies, in contrast to WtSTZ, which produced high levels of islet autoantibodies. It is well known that the early presence of islet autoantibodies is decisive in the development of diabetes by NOD mice as well as humans [46,47,53]. Our results confirm that Mif is essential for the development of hyperglycemia and suggest a role for Mif not only in the innate immune response but also in the adaptive immune response in T1DM.
Recently, it has been reported that Mif is produced by pancreatic -cells and that Mif is released by insulin granules in an autocrine fashion [54,55]. The chemical destruction of -pancreatic islets by STZ in Wt mice damaged -pancreatic islets; this condition could reduce one major source of Mif in this experimental T1DM model. However, we did not observe a reduction in Mif levels in this model; in contrast, Wt mice produced high serum levels of Mif after STZ administration. In line with this finding, it is known that the pancreatic islets remaining after treatment with STZ produce high levels of Mif [21] and that an elevation of Mif secretion precedes pancreatic islet death induced by IFN-, TNF-, and IL-1 [24]. This evidence establishes that STZ did not influence Mif production/release by pancreatic islets or other cellular sources, such as T cells, DC, and M infiltrating the pancreas.
The loss of insulin production in T1DM is related to pancreatic -cell destruction due to insulitis [56]. We observed that WtSTZ developed high serum levels of Mif and low insulin levels compared to Mif−/−STZ, which expressed insulin levels comparable to those in healthy mice. The histological analysis of pancreatic islets showed that WtSTZ displayed 100% insulitis, compared to the 33% insulitis observed in Mif −/−STZ mice. These observations confirm  that Mif deficiency resulted in pancreatic islet protection, probably by controlling the functional activity and modulating the secretory capacity of proinflammatory cytokines produced by T cells, DC, and M that reach the pancreatic cells.
Mif has been recognized as a molecule that not only promotes proinflammatory cytokine production but also acts as a chemokine. For example, Mif plays a crucial role in leukocyte recruitment and arrest during atherosclerosis development [57]. Therefore, Mif could participate in the process of insulitis to promote the production of proinflammatory cytokines, but Mif could also promote leukocyte recruitment to pancreatic -cells. For this reason, Mif −/−STZ exhibited reduced insulitis. Antigen-presenting cells, M and DC, are key mediators of the development of T1DM [4]. Moreover, it has been proposed that DC orchestrate the autoimmune response in T1DM via TLR-2 and TLR-4 [58]. Previous studies by us and others have shown that Mif induces the expression of costimulatory molecules on M and DC in some pathological infections [59][60][61]. Here, we identified the expression of costimulatory molecules and TLR-2 and TLR-4 on M and DC, as well as the ability of M to activate T lymphocytes.
Our results demonstrated that both M and DC from the spleen and pancreas of Mif−/−STZ mice expressed lower levels of CD80, CD86, MHC-II, TLR-2, and TLR-4 than those of WtSTZ. These results demonstrate a role of Mif in the activation of M and DC to promote costimulatory molecule expression, which might drive pancreas-specific T cell activation and effector Th1 subset differentiation, processes that have been associated with subsequent pancreatic injury in T1DM.  Figure 10: After Mif reconstitution, proinflammatory cytokine production in Mif−/− mice was increased.Mif reconstitution was initiated at the same time of STZ administration to Wt (◼) or Mif−/− mice ( ) until 7 weeks after STZ treatment. The glucose levels (a) and the levels of the inflammatory cytokines IL-6 (b), TNF-a (c), and IL-12 (d) were measured. Control mice receiving vehicle injection did not display significant changes on cytokine production compared to week 0 (data do not shown). The values are expressed as the means ± SEM of three independent experiments ( = 8). * < 0.05, * * < 0.01, or * * * < 0.001, GraphPad Prism software.
In line with the results described above, we observed that both M (F4/80) and monocytes (CD11b + ) from WtSTZ were more reactive and had greater ability to induce T lymphocytespecific proliferation in response to OVA than those from Mif−/−STZ mice. This finding is consistent with the evidence that proliferation of antigen-specific T cells from TCRtransgenic mice is highly dependent on CD28/CD86 costimulation. Disrupting this interaction dramatically reduces cell proliferation [62]. This conclusion agrees with the result that blocking CD86 prevents the development of diabetes in NOD mice [63]. Moreover, mutations in the MHC-II molecule lead to development of autoimmune diabetes [64].
By another hand, the differentiation state of M is an important determinant for T cell response in T1DM. Two major populations have been defined, the classically activated (CA) M secreting proinflammatory cytokines such as TNF-, IL-6, and IL-1 and reactive oxygen species; and alternatively activated (AA) M which secrete antiinflammatory factors including TGF-and IL-10. In T1DM it has been established that CAM trigger inflammatory responses which initiates insulitis and pancreatic cell death, whereas the AAM decreases hyperglycemia, insulitis, and inflammation in the pancreas [65].
As Mif is a regulator of many proinflammatory cytokines that are characteristic for the CAM , Mif has been proposed as CAM macrophage-polarizing factor [66]. However, there are few and contrary experimental evidences about how Mif might participate in the polarization to AAM or CAM . In a mouse model of nonalcoholic fatty liver disease M in liver from Mif−/− mice were skewed toward AAM [67]. By contrary, in the melanoma mouse model M -derived Mif participates in AAM polarization [68]. Moreover, murine Mif and filarial nematode parasite (Brugia) Mif protein induced proinflammatory cytokines release. However, Mif also induce upregulation of IL-4 on bone marrow-derived mouse M , which when treated in vitro with Mif and IL-4 induce AAM [69]. Here, we show that M from Mif−/−STZ mice display reduced proinflammatory cytokine production and exhibit reduced ability to induce T lymphocyte proliferation in response to OVA. It is possible that Mif deficiency influences on AAM polarization in this model; however, more experiments are necessary to establish this.

Conclusions
We show for the first time a role of Mif in promoting costimulatory molecule expression in M and DC in T1DM. These results reveal Mif as a key regulator of proinflammatory function and M and DC activation in T1DM. Although more specific experiments are required, there is no doubt that Mif represents a potential target for anti-Mif therapy, which might attenuate the autoimmune process in T1DM.