Type 1 Diabetes in NOD Mice Unaffected by Mast Cell Deficiency

Mast cells are innate immune cells that are the main effectors in IgE-mediated allergic inflammation. In response to cross linking of the high-affinity IgE receptor FcεRI, mast cells release preformed molecules stored in granules, such as histamine, proteoglycans, and proteases, that can contribute to allergic inflammation and to anaphylactic shock (1). In addition to this well-established role in IgE-mediated allergic diseases, mast cells, due to the plethora of factors that they can produce, are believed to take part in, and modulate, many immune responses, including autoimmunity. Based on either correlative evidence, for instance, the presence of mast cells in inflamed tissues, or through the use of mast cell–deficient Kit mutant mice and mast cell inhibitors, mast cells have been suggested to play roles in several autoimmune diseases that include thyroid eye disease (2), bullous pemphigoid (3), pemphigus vulgaris (4), rheumatoid arthritis (5), multiple sclerosis (6), systemic sclerosis (7,8), Guillain-Barre syndrome (9), and, notably, type 1 diabetes (T1D) (10–13; reviewed in 14,15). However, recent experiments in newly developed Kit-independent mast cell–deficient mice did not corroborate the proposed roles for mast cells, at least in models of rheumatoid arthritis and multiple sclerosis (16). The observed discrepancies call for a reevaluation of mast cell functions in models more specific than the traditional Kit mutants (reviewed in 17,18).

T1D, which affects ∼37 of every 100,000 children aged 14 years or younger in countries with the highest incidence (19), results from an autoimmune attack on insulin-producing β-cells by both cellular and humoral components of the immune system. In humans, the susceptibility for T1D is linked to the histocompatibility leukocyte antigen locus (20). Akin to humans, in the universally used nonobese diabetic (NOD) mouse model of T1D, the highest genetic contributor to disease susceptibility maps to the major histocompatibility complex (21,22). In view of the T-cell dependency of this disease and the putative roles of mast cells in the control of T-cell responses, which may occur via direct antigen presentation to either CD4⁺ or CD8⁺ T cells (23–25), by induction of T-cell migration to lymph nodes (26,27), by control of dendritic cell activation and migration to lymph nodes (28–31), or by control of the Th1 versus Th2 skewing (32–35), it has been postulated that mast cells are likely important players in the development of T1D (14,15,36).

Earlier studies have reported an association between mast cells and T1D by showing upregulation of mast cell genes in the pancreatic lymph nodes (PLNs) of the BioBreeding (BB) DR^lyp/lyp rat (13) or by delaying disease onset in the BB rat following disodium cromoglycate–mediated mast cell stabilization (10). In contrast, a third study found that activation, rather than inhibition, of mast cells with anti-FcεRI antibody injections in NOD mice led to delayed onset of T1D (11). Thus the existing data are either correlative or conflicting, and hence the potential role of mast cells in the incidence and progression of T1D remains largely unknown. Here we report the analyses of two different mast cell–deficient mouse models (NOD.Cpa3^Cre/+ and NOD.Kit^W-sh/W-sh), which were independently generated at three different institutions (German Cancer Research Center, Harvard Medical School, and VIB) to evaluate the effects of mast cell deficiency on the incidence and progression of T1D. Our data show that the progression or incidence of T1D in NOD mice are unaffected by mast cell deficiency.

Carboxypeptidase A3 (Cpa3)^Cre/+ mice have been previously described (16). These mice are wild type for Kit, lack mast cells in all tissues analyzed, and have no other known defect in the hematopoietic system, with the exception of a reduction in basophils. To obtain NOD.Cpa3^Cre/+ mice, Cpa3^Cre/+ mice were backcrossed into the NOD/ShiLtJ background (Jax stock number 001976) for at least 12 generations. Kit^W-sh/W-sh (“sash”) mice (37) were backcrossed into the NOD/ShiLtJ background for 11 or 12 generations to generate two independent lines of NOD.Kit^W-sh/W-sh mutants.

Male and female NOD.Cpa3^Cre/+, their NOD.Cpa3^+/+ littermate controls, and NOD/ShiLtJ controls, all housed in the same animal facility (German Cancer Research Center, Heidelberg, Germany), were monitored for the development of diabetes starting from 10 weeks of age by weekly assessing blood glucose levels via tail-vein bleeding using a OneTouch Ultra glucometer (LifeScan, Inc., Milpitas, CA). Female NOD.Kit^W-sh/W-sh, NOD.Kit^W-sh/+, and NOD.Kit^+/+ controls were monitored for the development of diabetes starting at 10 weeks of age. Two independent studies were performed with NOD.Kit^W-sh/W-sh, one at the VIB in Leuven, Belgium, and one at Harvard Medical School, Boston, MA. In all studies, mice with glucose concentrations >250 mg/dL on 2 consecutive days were considered diabetic and were immediately killed by CO₂ asphyxiation.

Pancreata from 15-week-old (Heidelberg) or 10-week-old (Boston) mice were isolated and fixed in 10% formalin (Sigma-Aldrich, St. Louis, MO) overnight. Subsequently, tissues were paraffin embedded and cut into 5 μm sections. Serial paraffin sections were stained by hematoxylin and eosin (H&E) coupled with aldehyde fuchsin as previously described (38). Images of islets were taken using a Zeiss Axioplan light microscope coupled with an AxioCam ICc3 color camera at 200× or 630× magnifications. For the NOD.Cpa3^Cre/+ mice (Heidelberg), the scoring was performed using the following scale in a blinded manner (Supplementary Fig. 1): A) 0, no insulitis; B) 1, peri-islet insulitis; C) 2, intermediate insulitis; D) 3, intraislet insulitis; and E) 4, complete islet insulitis. For the NOD.Kit^W-sh/W-sh mice (Boston), scoring was divided into three categories: insulitis, peri-insulitis, and no insulitis. Pancreas and ear paraffin sections from NOD.Cpa3^Cre/+, NOD.Cpa3^+/+, NOD.Kit^W-sh/W-sh, and NOD.Kit^+/+ mice were stained with a solution of 0.1% (w/m) toluidine blue (Sigma-Aldrich) and analyzed by light microscopy for the presence of metachromatic mast cells.

PLNs from 15-week-old mice were isolated, finely minced, and mechanically sieved through a 40 μm mesh to isolate immune cells. Isolated single cells were stained for flow cytometry as previously described (39). The following antibodies were used: CD45 Alexa Fluor 700 (30-F11; eBioscience), CD117 (Kit) allophycocyanin (2B8; BD Pharmingen), IgE fluorescein isothiocyanate (R35–72; BD Pharmingen), CD19 PerCP-Cy5.5 (1D3; BD Pharmingen), CD3 PerCP-Cy5.5 (145–2C11; BioLegend), CD4 allophycocyanin (RM4.5; eBioscience), CD8 PE-Cy7 (53–6.7; eBioscience), CD44 fluorescein isothiocyanate (IM7; BD Pharmingen), CD62L Vio605 (MEL-14; BioLegend), and Foxp3 PE (FJK-16S; eBioscience). SYTOX blue (Life Technologies) was used for dead cell exclusion. All samples were measured using a BD LSRFortessa flow cytometer (BD Biosciences, Heidelberg) and analyzed using FlowJo X software.

Progression to diabetes and diabetes incidence were calculated using the product limit method (Kaplan-Meier analysis). Statistical significance between a pair of curves was determined by the log-rank (Mantel-Cox) test using Prism 5.0 software. Flow cytometry data were analyzed with two-tailed Student t test. P values ≤0.05 were considered statistically significant.

As mast cell–deficient models, we used Cpa3^Cre/+ mice, which lack mast cells and are wild type for Kit, and Kit^W-sh/W-sh mice, which are double deficient for mast cells and Kit. Cpa3^Cre/+ mice are targeted knock-in mice that bear Cre recombinase in the Cpa3 locus. Cre expression under the control of the endogenous Cpa3 gene leads to genotoxic ablation of mast cells through a Trp53-dependent mechanism (16,40,41). Cpa3^Cre/+ mice are devoid of mucosal and connective tissue mast cells under steady state and remain so under various challenges (16,40,41). In Kit^W-sh/W-sh mice, an ∼3.1 Mb genomic inversion upstream of the Kit locus disrupts the physiological regulation of Kit transcription that results in mast cell deficiency (42,43). In addition, Kit^W-sh/W-sh have further Kit-related and Kit-unrelated defects (43). Cpa3^Cre/+ or Kit^W-sh/W-sh mice were backcrossed onto the NOD background to obtain mast cell–deficient NOD mouse lines (see 2Research Design and Methods).

To confirm that NOD.Cpa3^Cre/+ mice remained mast cell–deficient during the progression of insulitis and diabetes, we examined NOD.Cpa3^Cre/+ and NOD.Cpa3^+/+ mice for the presence of mast cells in islets, ears, PLNs, and the peritoneal cavity (Fig. 1A–I). Histological analysis of pancreatic islets from 15-week-old female mice, a time point when autoimmune pathology is progressing, showed complete lack of mast cells in the islet-immune infiltrates of both NOD.Cpa3^+/+ and NOD.Cpa3^Cre/+ mice, indicating that even in mast cell–proficient mice (NOD.Cpa3^+/+), mast cells do not infiltrate the pancreatic islets (Fig. 1A and B). This was substantiated by an independently performed flow cytometric analysis of islets from 10-week-old NOD/ShiLtJ mice (Supplementary Fig. 2). To substantiate the mast cell–deficient phenotype by histological analysis, paraffin-embedded ear sections were stained with toluidine blue; these showed a large number of mast cells delineating the epidermis (larger magnification image depicts mast cells) in the NOD.Cpa3^+/+ mice (Fig. 1C and D) and complete absence of mast cells in the sections of NOD.Cpa3^Cre/+ mice (Fig. 1E). Flow cytometric analysis of PLN cells showed the presence of mast cells in NOD.Cpa3^+/+ mice, supporting previous reports (Fig. 1F) (10,13). However, these cells were absent in NOD.Cpa3^Cre/+ mice (Fig. 1G). Moreover, flow cytometric analysis of peritoneal lavage cells revealed that ∼4% of all CD45⁺ cells in the peritoneal cavity of NOD.Cpa3^+/+ mice were mast cells (Fig. 1H), while mast cells in the peritoneal cavity lavage of NOD.Cpa3^Cre/+ mice were undetectable (Fig. 1I). Toluidine blue staining of ear paraffin sections confirmed mast cell deficiency in NOD.Kit^W-sh/W-sh mice (Supplementary Fig. 3). Collectively, in NOD.Cpa3^+/+, but not in NOD.Cpa3^Cre/+ mice, mast cells are present in the skin, in the peritoneal cavity, and in draining PLNs, but mast cells were not found in the islet-immune infiltrates.

NOD.Cpa3^Cre/+ mice, their NOD.Cpa3^+/+ littermate controls, and the original NOD/ShiLtJ mice, all housed in the same animal facility, were monitored for diabetes starting at 10 weeks of age. In female mice, there was no statistically significant difference in the incidence (∼70%) or median onset time (21 vs. 19 weeks) between NOD.Cpa3^Cre/+ mice and their NOD.Cpa3^+/+ littermates. A similar incidence and onset time (20 weeks) was also observed for the NOD/ShiLtJ female mice (Fig. 2A). A parallel study was conducted for a cohort of male mice, and again, no differences were found between NOD.Cpa3^Cre/+, NOD.Cpa3^+/+, and NOD/ShiLtJ mice (Fig. 2B). Comparing male and female mice, the sex differences (44) in median onset of disease were similar for NOD.Cpa3^Cre/+ (21 vs. ∼35 weeks; P = 0.0015), NOD.Cpa3^+/+ (19 vs. 36 weeks; P = 0.0005), and NOD/ShiLtJ mice (20 vs. 35 weeks; P = 0.0149). In keeping with analyses of NOD.Cpa3^Cre/+ mice, female NOD mice on the Kit mutant background generated in Leuven (Fig. 2C) and in Boston (Fig. 2D) showed no significant differences in diabetes incidence or onset time comparing NOD.Kit^W-sh/W-sh and NOD.Kit^+/+ littermates.

Histological analyses were performed on serial paraffin sections of pancreata stained by H&E together with aldehyde fuchsin to assess the level of insulitis in 15-week-old female and male NOD.Cpa3^Cre/+, NOD.Cpa3^+/+, and NOD/ShiLtJ mice. Numbers of islets were counted on five serial sections, each 5 μm in thickness and with 25 μm distance between each section. Relative numbers of islets per mouse did not differ between genotypes (Supplementary Fig. 1F). Quantification of insulitis showed no differences in scores comparing female (Fig. 3A) and male (Fig. 3B) NOD.Cpa3^Cre/+, NOD.Cpa3^+/+, and NOD/ShiLtJ mice. Moreover, the degree of islet insulitis was comparable in female NOD.Kit^W-sh/W-sh and NOD.Kit^+/+littermates (Supplementary Fig. 4).

In summary, these experiments establish that the incidence, progression, and histopathological degree of insulitis during diabetes development are unaffected by the presence or absence of mast cells in the NOD mouse.

PLNs from NOD.Cpa3^+/+ and NOD.Cpa3^Cre/+ mice were collected, and subsequently, cells were counted and analyzed by flow cytometry. Absolute overall cell numbers (2.8 × 10⁶ ± 1.0 [mean ± SD] for NOD.Cpa3^+/+ [n = 4] vs. 2.6 × 10⁶ ± 0.5 for NOD.Cpa3^Cre/+ [n = 3] mice), absolute numbers of CD45⁺ cells (1.8 × 10⁶ ± 1.1 for NOD.Cpa3^+/+ vs. 2.0 × 10⁶ ± 0.5 for NOD.Cpa3^Cre/+ mice), percent of CD3⁺ T cells (78 ± 11 for NOD.Cpa3^+/+ vs. 76 ± 6 for NOD.Cpa3^Cre/+ mice), and absolute numbers of CD3⁺ T cells (1.8 × 10⁶ ± 1.1 for NOD.Cpa3^+/+ vs. 2.0 × 10⁶ ± 0.5 for NOD.Cpa3^Cre/+ mice) were comparable in mast cell–sufficient and mast cell–deficient NOD mice (Supplementary Fig. 5). In a larger cohort of NOD.Cpa3^+/+ (n = 8) and NOD.Cpa3^Cre/+ (n = 9) mice, we determined in the PLNs the relative proportions of T cells, B cells, and dendritic cells (Fig. 4A–C), all of which were similar in both genotypes. In accordance with earlier analysis (16), proportions of CD4⁺ (Fig. 4D) and CD8⁺ (Fig. 4E) cells among total T cells and the fraction of T regulatory cells per total CD4⁺ T cells (Fig. 4F) were all not affected by mast cell deficiency.

Immune-driven effector cell differentiation of conventional T cells was assessed by expression analysis for CD62L and CD44, and T cells were classified into three categories: naïve cells (CD62L⁺CD44^–), central memory cells (CD62L⁺CD44⁺), and effector memory cells (CD62L^–CD44⁺) (reviewed in 45). In the CD4 T-cell population (Fig. 4G), percentages of naïve, central memory, and effector memory subsets were comparable in NOD.Cpa3^Cre/+ and NOD.Cpa3^+/+ mice.

Collectively, all analyzed immunological parameters were similar comparing mast cell–deficient and mast cell–proficient NOD mice. This is in full agreement with the lack of evidence for a role of mast cells on incidence, progression, or degree of insulitis during diabetes development in the NOD strain.

Many years of work by several groups leading to the current study have postulated that mast cells are important players in the initiation and progression of autoimmune diseases (14,15,36,46,47). In most cases, these studies have been performed using nonspecific mast cell “stabilizers” (48) or Kit mutant mice that, although mast cell–deficient, have other immune (18) and relevant nonimmune abnormalities, which include, notably, the recent discovery of the role of Kit in pancreatic β-cell function (49). Taking into account the pleiotropic abnormalities in Kit mutant mice, it is conceivable that several roles that have been attributed to mast cells are not mast cell–specific, but instead caused by the Kit mutations. To circumvent this problem, our group (16) and others (50,51) have generated Kit-independent mast cell–deficient mice; experiments in these mice have already challenged the role of mast cells in antibody-induced autoimmune arthritis and experimental autoimmune encephalomyelitis (16). The role of mast cells in the autoimmune destruction of β-cells, or T1D, is not well understood. In fact, although highly speculated (14,15,36), the effect of mast cell deficiency on T1D had never been directly assessed genetically in vivo, with the exception of an abstract suggesting that NOD.Kit^W/Wv mice failed to develop T1D (52). Indirectly, one study tested the effect of mast cell inhibition using the mast cell stabilizer disodium cromoglycate and found that this significantly delayed the onset of T1D (10); however, the specificity and function of this drug is controversial (48). In contrast, a second study activated mast cells and basophils via anti-FcεRI antibody treatment and reported that this treatment delayed the onset of T1D (11). It is difficult to interpret these opposing results, and therefore the role of mast cells on T1D remained, at best, controversial.

In the current study, we assessed the potential effects of mast cell deficiency on T1D, using both a traditional Kit mutant (Kit^W-sh/W-sh) and the more recent Kit-independent mast cell–deficient Cpa3^Cre/+ strain. Both strains were backcrossed to the NOD background until the new lines developed diabetes with similar onset times and rates as the NOD/ShiLtJ mice housed under identical conditions. We present data from independent studies conducted in three different institutions, all of which showed that lack of mast cells did not affect the incidence and progression of T1D in mast cell–deficient NOD strains. It was critical to show that NOD.Cpa3^Cre/+ mice remained mast cell–deficient during initiation and progression of the autoimmune attack. Both in the peritoneal cavity and in the relevant PLNs, there was a complete absence of mast cells in these NOD mice. Of note, the absence of mast cells in the islets of NOD/ShiLtJ mice or NOD.Cpa3^+/+ mice at stages when they were infiltrated by other immune cells strongly suggests that mast cells cannot be directly involved in islet immunopathology. However, the presence of mast cells in secondary lymphoid organs, including the PLNs, in mast cell–proficient NOD mice may reflect an active role of mast cells in the immunomodulation of the autoimmune response leading to T1D. After establishing mast cell deficiency in Cpa3^Cre/+ mice on the NOD background, we assessed the onset and severity of diabetes and found that the presence or absence of mast cells had no impact on the development of T1D. Furthermore, evaluations of T1D in NOD.Kit^W-sh/W-sh mice in Leuven and in Boston showed a similar outcome regardless of the Kit genotype (wild type or mutant) and, hence, irrespective of the presence of mast cells. Moreover, we compared immunological parameters associated with the autoimmune response in NOD.Cpa3^Cre/+ mice and their littermate controls and found comparable cellularity of lymphocyte and dendritic cell subsets and in T helper cell activation.

Collectively, at least two certain conclusions can be drawn from this study. Firstly, the T1D autoimmune response in NOD mice is mast cell–independent, as well as independent from defects associated with hypomorphic mutations in Kit. Secondly, our study, along with earlier studies (16), put into question the suggested ability of mast cells to control or modulate autoimmunity-promoting adaptive immune response (14,46). Mast cell deficiency did not affect the levels of conventional or regulatory T cells, B cells, or dendritic cells in the PLNs; moreover, T-cell activation, much of which had previously been suggested to be associated with mast cell function, was also unaffected by mast cell deficiency (23–25,28–35). Obviously, we refrain from extrapolating negative data obtained in experiments on arthritis, experimental autoimmune encephalomyelitis (16), and T1D (this study) to the potential roles of mast cells in other autoimmune responses or adaptive immunity in general. However, as discussed recently (17,53), it remains important to reevaluate mast cell functions, which had previously been suggested based on in vitro experiments, on the use of inhibitors or mast cell stabilizers or on the use of Kit mutants by turning to more conclusive in vivo mouse models for mast cell deficiency.

Although our results clearly show that mast cell deficiency does not affect T1D development in NOD mice, our results suggest, but do not prove, that mast cells are unimportant in rat and human T1D. In the BB DR^lyp/lyp rat, upregulation of mast cell genes in the PLNs suggest that mast cell numbers or their activity increase during the disease (13), and in this model, disease onset was delayed by mast cell stabilization (10). A link between T1D progression and mast cell activation has also more generally been invoked in parasite infections or during asthma, which might be negatively correlated to T1D development (11,36,54). Further experiments should test these ideas by examining the prevalence of T1D in patients with mastocytosis or by studying the dependency on mast cells of helminth-mediated T1D inhibition. As it stands, mast cells are likely to play important evolutionarily conserved roles in immunity beyond their notorious role in allergic diseases. While beneficial mast cell roles have emerged in venom degradation (55,56) and in IgE-mediated protection from lethal doses of venom (57,58), the possible involvement of mast cells in the regulation of innate and adaptive immune responses and their roles as effector cells remain enigmatic.

Acknowledgments. We thank Sathya Muralidhar and Tabea Arnsperger (Division of Cellular Immunology, German Cancer Research Center, Heidelberg, Germany) for technical assistance in this project.

Funding. D.A.G. was a recipient of a postdoctoral fellowship from the German Cancer Research Center. H.-R.R. was supported by European Research Council advanced grant 233074 and Deutsche Forschungsgemeinschaft-Sonderforschungsbereiche 938-project L. A.L. was supported by a Career Development Award from JDRF. D.M. received funding from JDRF.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. D.A.G. designed and performed experiments, analyzed the data, and wrote the manuscript. W.F. and S.S. designed and performed experiments. T.B.F. contributed to experiment design, assisted with experiments, and edited the manuscript. A.O.-L. and Y.L. assisted with experiments. A.L., D.M., and H.-R.R. conceived the studies, contributed to experiment design, and edited the manuscript. H.-R.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.