Adoptive transfer of CD3+ T cells and CD4+ CD44high memory T cells induces autoimmune pancreatitis in MRL/MpJ mice

Abstract The immunopathogenesis of autoimmune pancreatitis (AIP) is poorly understood. Here, we have used MRL/MpJ mice, a model of spontaneous AIP, to address the role of cellular autoimmune processes in the initiation and progression of the disease. Therefore, different T cell subpopulations were adoptively transferred from sick to still healthy (but susceptible) MRL/MpJ mice. Unpurified splenocytes and CD3+ T cells both efficiently induced AIP, while CD4+ and CD8+ T cells alone, as well as splenocytes from healthy mice, were insufficient to trigger the disease. Strikingly, CD4+ CD44high memory T cells, although transferred at lower numbers than other T cells, also induced AIP in recipient mice. Employing a modified experimental design, we also evaluated the effects of regulatory T cells (T regs) on the progression of AIP in already diseased mice. Under the given experimental conditions, there was no significant suppressive effect of adoptively transferred T regs on pancreatic histopathology. The results of our studies suggest a key role of T cell‐mediated processes in murine AIP. The effects of CD4+ CD44high memory T cells are in accordance with genetic studies of our group, which had previously implicated this cell type into the pathogenesis of AIP. In follow‐up studies, we will focus on the interplay of cellular and humoral autoimmunity in the context of AIP.

Less is known about cellular immune processes in AIP, although T cells are the prevalent type of infiltrating immune cells in affected tissue. 8,9 Both, CD4 + and CD8 + T lymphocytes are present in pancreatic parenchyma of AIP patients, 3 but CD4 + T cells predominate in tissue infiltrates. Increased production of interferon (IFN)-c by CD4 + T helper -cells type 1 (Th1) has been proposed to promote AIP, 8 and treatment with IFN-c strongly aggravated AIP in mice. 10 On the other hand, cytokines promoting the development of Th2 have also been described in the context of AIP. 11 The MRL/MpJ mouse model has proven useful to study the pathogenesis of AIP. These mice spontaneously develop an AIP at the age of about 6 months. 12,13 The murine AIP histopathologically resembles the human AIP type 1. The incidence and the severity of the disease are higher for female than for male mice and can be triggered using polyinosinic:polycytidylic acid (poly I:C). 14 Using this mouse model in previous studies, we gained genetic evidence for an involvement of CD4 + CD44 high memory T cells in the pathogenesis of AIP. 15 Likewise, our studies implicated regulatory T cells (T regs ) in the mediation of the therapeutic effects of rapamycin through a suppression of the effector T cell response. 9 One characteristic of autoimmune diseases such as AIP is the possibility to adoptively transfer the disease from sick to healthy individuals in in vivo experiments. In 1992, Kanno

| Isolation of splenocytes and cell culture conditions
Splenocytes were obtained from female MRL/MpJ mice. Young and adult mice (8.5 AE 0.2 and 41 AE 0.6 weeks old, respectively) served as spleen donors. The animals were sacrificed by an overdose of ketamine/xylazine, serum was collected using clot activator containing tubes (Sarstedt, N€ umbrecht, Germany), and the pancreas, liver and kidneys were harvested and cryo-or paraffin-embedded for further analyses. The spleen was kept on ice in splenocyte culture medium RPMI-1640 (Biochrom, Berlin, Germany) supplemented with 10% foetal calf serum (FCS; Biochrom), 1% penicillin/streptomycin (Biochrom) and 50 lmol/L b-mercaptoethanol (Merck, Darmstadt, Germany). Subsequently, the organ was forced through a 70 lm cell strainer (Greiner bio-one, Kremsm€ unster, Austria) and incubated in 10 mL splenocyte culture medium containing 50 lg/mL DNase I (Roche Applied Science, Mannheim, Germany) for 10 minutes. The cell suspension was centrifuged (3009 g, 10 minutes, 4°C), resuspended in 1 mL phosphate-buffered saline (PBS) and incubated for 3 minutes on ice with 4 mL of ice-cold 0.25 M NH 4 Cl solution to lyse erythrocytes. Lysis was stopped by adding 10 mL of culture medium and centrifugation (3009 g, 10 minutes, 4°C). The cells were sowed at a density of 1 9 10 6 cells/mL in cell culture flasks.
Based on the protocol of Kanno et al, 12 25 lg/mL phytohaemagglutinin (PHA; Merck) was added to each flask. Cells were cultured for 3 days at 37°C in a 5% CO 2 humidified atmosphere.
2.3 | Adoptive transfer of unpurified splenocytes, CD3 + T cells, CD4 + T helper -cells, CD8 + cytotoxic T cells and CD4 + CD44 high memory T cells   Six weeks after cell injection, the mice were sacrificed, serum was collected, and the pancreas, liver and kidneys were cryo-and paraffin-embedded for further analyses. The workflow for the adoptive transfer is illustrated in Figure 1. female MRL/MpJ mice (pretreated with cyclophosphamide as described above; for workflow, see Figure 1). Mice (29 AE 0.2 weeks old) that were not treated with cells served as controls. End-point analyses for both groups were carried out 6 weeks after cell injection.

| Histology and immunohistochemistry
Paraffin-embedded pancreatic sections were stained with haematoxylin and eosin (H&E), and pathological changes were graded on a semi-quantitative scale from 0 (none) to 4 (severe) as previously described (H&E-score). 9 Table 1) for splenic cell isolation. The severity of the AIP was evaluated by scoring H&E stained pancreatic tissue ( Figure 2) and CD3 stained sections (data not shown) in a semi-quantitative manner employing scores that ranged from 0 to 4. 9,10,13 The average AIPscore for all groups of adult donors was approximately 3 (Table 1),

| Statistical evaluations
representing a severe inflammation with parenchymal destruction.
Adult mice without a pancreatic phenotype were disregarded as donors of lymphocytes. For comparison, we also employed one group of young healthy female donors for splenocyte isolation (Table 1).

| CD3 + T cells effectively transfer murine AIP
For the adoptive transfer of splenocytes, the cells acquired from donor MRL/MpJ mice were cultured for 3 days before CD3 + , CD4 + or CD8 + T cells were purified. All isolations lead to highly pure cell populations ( Figure S1). Either one of the isolated subpopulations or unpurified splenocytes were then transferred into young and still healthy female MRL/MpJ recipient mice (Table 1). Additionally, unpurified splenocytes were also transferred into male recipient mice. Young female MRL/MpJ mice that were not treated with cells served as a control group. Six weeks after the cell injection, the AIPscore of the recipient mice was evaluated ( Figure 3).  were not significantly higher than the score of untreated controls.

| Composition of inflammatory foci
Pancreatic inflammatory foci of MRL/MpJ mice with spontaneous AIP mainly consist of CD3 + T cells, with CD4 + T cells being predominant over CD8 + T cells ( Figure 4A). 10,12,13,16 Autoimmune foci that were induced by the transfer of unpurified splenocytes ( Figure 4B), CD3 + T cells ( Figure 4C), CD4 + T cells ( Figure 4D) and CD8 + T cells ( Figure 4E) also largely consisted of CD3 + T cells, and higher numbers were observed for CD4 + T cells than for the CD8 + counterpart.
Thus, our immunohistochemical investigation revealed no differences in the composition of the inflammatory foci of MRL/MpJ recipient mice, independent of the kind of transferred cells.

| Other organ involvements
AIP of MRL/MpJ mice is occasionally accompanied by autoimmune lesions in liver and kidney (Ref. [9] and Figure 5). Investigating control mice and recipients of unpurified splenocytes and CD3 + T cells from adult mice with AIP, we observed autoimmune foci in these organs in 10%-30% of the animals ( Figure 5 and Figure S2). There Splenocytes were obtained from adult (sick) and young (healthy) female MRL/MpJ donor animals as indicated, and subpopulations (CD3 + , CD4 + or CD8 + T cells) were isolated 3 days later. Either unpurified splenocytes or the indicated subpopulations were transferred into young recipient mice. Age-matched MRL/MpJ mice without cell transfer served as controls (darker bar). Six weeks after cell injection, the mice were sacrificed, the pancreata were CD3 and H&E stained and subjected to semi-quantitative evaluation (scores from 0 to 4). Data are presented as mean AE SEM (n = 6-10 per group; Table 1), *P < .05 vs untreated controls were no significant differences between treated mice and controls.
Therefore, under the given experimental conditions, the injected cells did not have an impact on other vulnerable organs than the pancreas.

| CD4 + CD44 high memory T cells transfer AIP with high efficiency
To gain further mechanistic insights, we employed CD4 + CD44 high memory T cells for additional cell transfer experiments. Splenocytes were isolated from adult female MRL/MpJ mice (average AIP-score: 3.0 AE 0.2; Table 1) and cultured for 3 days, prior to the isolation of CD4 + CD44 high T cells (for cell purity, please refer to Figure S1). To . There were no significant differences among the groups F I G U R E 6 Pancreatic AIP-score of MRL/MpJ mice treated with CD4 + CD44 high memory T cells. Splenocytes were obtained from adult female MRL/MpJ donor animals and CD4 + CD44 high T cells were isolated 3 days later. Subsequently, 2 9 10 6 cells (unpurified splenocytes or CD4 + CD44 high T cells) were transferred into young, female recipient mice. Age-matched MRL/MpJ mice without cell transfer served as controls (darker bar; same controls as in Figure 3). Data are presented as mean AE SEM (n = 9-10 per group), *P < .05 vs untreated controls F I G U R E 7 Pancreatic AIP-score of MRL/MpJ mice treated with regulatory T cells. CD4 + CD25 + T regs were isolated from adult female MRL/MpJ mice (27 AE 0.1 weeks old) and expanded for 2 weeks. Afterwards 2.5 9 10 6 cells were transferred into female MRL/MpJ mice at an advanced age of 30 AE 0.2 weeks, where spontaneous AIP occurs at a high frequency. 12 Age-matched control mice did not receive cells. Six weeks after the injection, pancreatic tissue was CD3 and H&E stained and scored in a semi-quantitative manner (scores from 0 to 4). Data are presented as mean AE SEM (n = 8-10 per group). There was no significant difference between the two groups induce AIP in adoptive transfer experiments. 17 Accordingly, T cells play a key role in cellular immune processes in different experimental models of AIP. In the human situation, the presence of T cells within pancreatic parenchyma has also been described. 8,18,19 Moreover, CD4 + CD44 high memory T cells were sufficient to transfer murine AIP. Strikingly, the average AIP-score for recipient mice treated with memory T cells was as high as the AIP-score for recipient mice treated with CD3 + T cells, although the number of transferred cells was less than the half. The studies on the role of CD4 + CD44 high T cells were encouraged by the results of previous investigations of our group: Genotype-phenotype correlations studies in AIP-susceptible mice lead to the mapping of quantitative trait loci (QTLs) containing putative susceptibility genes for AIP. 16 Interestingly, the relative frequency of CD4 + CD44 + splenocytes and the development of AIP were found to be controlled by overlapping QTLs. 15,16 Additionally, the relative frequency of CD4 + CD44 high memory T cells in the spleen correlated with the severity of AIP. 15 The unique role of CD4 + CD44 high memory T cells in murine AIP is underscored by the fact that they were the only type of immune cell that fulfilled both criteria (overlapping QTLs with AIP and correlation with disease severity).
CD4 + CD44 high CD62L low effector memory T cells (T EM ) are, in contrast to na€ ıve T cells, antigen-primed and thus deliver T cell memory. 20 T EM are located in lymphoid and non-lymphoid tissue and provide immediate local immune response. 20,21 With the presented work, we demonstrate a functional role of CD4 + CD44 high memory T cells in the pathogenesis of autoimmune-related experimental pancreatitis.
The role of CD4 + and CD8 + T cells in AIP has been discussed before. 3,8,9,16 Interestingly, neither CD4 + nor CD8 + T lymphocytes alone were sufficient to induce AIP in healthy (but susceptible) recipients. Together, these findings suggest a contribution of both types of T cells to the exhibition of the disease phenotype.
Regulatory T cells are immune cells with largely inhibitory characteristics. 11 The specific role of these cells in AIP, however, remains to be elucidated. In MRL/MpJ mice, the therapeutic effect of rapamycin could be linked to the activation of T regs , which suppress the activity of effector T cells. 9 This suppressive effect of The MRL/MpJ model of spontaneous AIP that was used in this study mimics the human disease, specifically type 1, in some important regards. Thus, inflammation starts from the exocrine tissue, involves destruction of acinar architecture and increased deposition of extracellular matrix, and leaves the islets largely unaffected. Moreover, dense infiltrates of immune cells, such as T cells and activated B-cells/plasma cells, are characteristic of both the human and the murine disease. 3,13,16,22 As a major limitation of any mouse model of AIP, the lack of IgG4 has to be taken into account. The relevance of our findings for the human situation therefore needs to be studied further.
While this study was dedicated to the analysis of the role of different T cell populations, previous experimental studies by us and others have implicated macrophages (particularly, of the M1 subtype) and neutrophils into the pathogenesis of the disease. 9,10,12,13,16 Specifically, it has been proposed that activated macrophages may induce destruction of pancreatic parenchyma directly or via antibody-dependent cellular cytotoxicity. 12 Neutrophils may be activated by endogenous danger signals, such as damage-associated molecular patterns and autoantibodies, and trigger activation of plasmacytoid dendritic cells through neutrophil-derived structures termed neutrophil extracellular traps (NETs). 22 Induction and progression of AIP in mice and men, therefore, depend on a complex interplay between innate and adaptive immune responses that remains to be further elucidated.
In summary, the results of this study provide new insights into the immunopathogenesis of experimental AIP by showing that both CD3 + and CD4 + CD44 high T cells effectively transfer AIP from sick MRL/MpJ mice to healthy, but AIP-susceptible individuals. While our data clearly support a key role of T cells in experimental AIP, they do not address the involvement of autoantibodies to disease induction and progression. As both cellular and humoral immune reactions are likely to contribute to the pathogenesis of human AIP, the interplay between these processes is subject of follow-up studies of our laboratory.

CONFLI CT OF INTEREST
The authors confirm that there are no conflict of interests.