Absence of Lymph Nodes in NOD Mice Treated With Lymphotoxin-β Receptor Immunoglobulin Protects From Diabetes

NOD mice were originally obtained from The Taconic Laboratory and NOD.scid mice from The Jackson Laboratory (Bar Harbor, ME). Mice were followed for the development of diabetes by bimonthly blood glucose measurement, and diabetes was defined by values >250 mg/dl on two separate occasions. All mice were housed and cared for in accordance with the guidelines of the Washington University Committee for the Humane Care of Laboratory Animals and with the National Institutes of Health guidelines on laboratory animal welfare.

Timed pregnant NOD females were injected with 100 μg i.v. LTβR-Ig (gift of J. Browning, Biogen, by way of Drs. Laura Mandik-Nayak and Paul Allen) or control human immunoglobulin (Jackson Immuno-Research) on embryonic day 11 (E11) and E14 as described (18).

The NIT-1 cell line (19) was a gift from Dr. E.H. Leiter (The Jackson Laboratory). The monoclonal antibody SF1.1.1 was used to stain H-2K^d (Pharmingen, San Diego, CA).

NIT-1 cells were incubated first with mouse sera and then with a goat anti-mouse IgG-PE (Caltag, Burlingame, CA). Fluorescence-activated cell sorter (FACS) analysis was performed on a FACScan flow cytometer, and data analysis was performed with CellQuest software (Becton Dickinson, Mountain View, CA).

Mice were killed by cervical dislocation, and pancreata were fixed in 10% formalin and then stained with hematoxylin and eosin. The spleens of control and nodeless animals were frozen and then stained with anti–CD3-FITC (fluorescein isothiocyanate) and anti–CD19-PE (Pharmingen).

Splenocytes (2 × 10⁷) from control and nodeless mice were transferred into NOD.scid recipients by intravenous tail vein injection. The mice were followed by blood glucose determination every 1–2 weeks.

Nodeless mice and 8-week-old NOD male mice were irradiated with 600 R and then received 3 × 10⁷ splenocytes from diabetic NOD donors by intravenous tail injection.

Mice from control immunoglobulin and LTβR-Ig–treated mothers were examined for the presence of lymph nodes. As previously reported, LTβR-Ig–treated mice were born without axillary, inguinal, and popliteal lymph nodes (18) and also lacked peripancreatic lymph nodes. The absence of peripancreatic lymph nodes was confirmed by en block resection and careful histological analysis of tissue containing the duodenum and head of the pancreas, including the superior mesenteric vessels. The spleens from the two groups of animals had normal cellularity and architecture (Figs. 1A and B). Specifically, there was no significant difference in the total number of cells or in the number of T-cells, B-cells, dendritic cells, and macrophages between the two groups of animals. The gross architecture of the spleens was preserved as demonstrated by normal appearing T- and B-cell zones.

The offspring of LTβR-Ig–treated mice were followed for the development of hyperglycemia by weekly blood glucose measurement. Control mice developed diabetes with kinetics and a cumulative incidence equivalent to that found in our NOD mouse colony, with an incidence of 80% (36 of 46) at 1 year of age (Fig. 2). In contrast, none of the mice (0 of 22) of LTβR-Ig–treated mothers developed diabetes at 1 year. The pancreata of mice from both control and LTβR-Ig mice were examined histologically for the presence of inflammation and β-cell death. The control mice all developed a destructive inflammatory infiltrate, with marked loss of β-cell mass; in contrast, the islets from the LTβR-Ig–treated mice were devoid of infiltrate (Fig. 3). A weak peri-insulitis was found in the pancreas of some of the nodeless mice after 40 weeks of age; however, β-cell mass was preserved in these animals.

The NOD mouse spontaneously develops antibodies that bind to β-cell surface antigens between 4 and 8 weeks of age (20). To determine the role of lymph nodes in this process, control and experimental animals were examined for the presence of anti–β-cell autoantibodies at various time points. Serum antibody titers were assessed by flow cytometric analysis of NIT-1 cells stained with mouse serum and expressed as mean fluorescence intensity. At the time points examined (6, 10, and 30 weeks of age), the development of anti–β-cell antibodies was reduced by approximately half in the LTβR-Ig–treated group when compared with control animals (P < 0.05, Table 1). Of note, LTβR-Ig–treated mice did develop low titers of β-cell autoantibodies, as demonstrated by positive staining when compared with nonautoimmune strains of mice, such as B10.BR mice.

To determine whether protected mice contained T-cells capable of inducing diabetes, splenocytes from LTβR-Ig–treated mice were transferred into NOD.scid recipients. Splenocytes from LTβR-Ig–treated mice failed to transfer diabetes into NOD.scid recipients compared with controls. Splenocytes from diabetic control mice induced diabetes in all (13 of 13) recipients by 8 weeks posttransfer; however, splenocytes from nodeless mice induced diabetes in only 1 of 13 recipients at 25 weeks posttransfer (Fig. 4A). These results taken together indicate that the presence of secondary lymphoid organs plays a key role in the expansion and/or maintenance of a diabetogenic T-cell repertoire. However, the defect in T-cell development is not global, as demonstrated in experiments where NOD and nodeless mice were immunized with the protein antigen hen-egg white lysozyme intraperitoneally and the spleen T-cells were then tested for proliferation. Cells from both control and LTβR-Ig–treated mice proliferated to hen-egg white lysozyme, indicating that T-cells could respond in an antigen-specific manner to systemic immunization. Moreover, these findings are consistent with previously published data in similar model systems (13,14,21).

However, the ability of splenocytes from diabetic NOD mice to induce diabetes upon transfer into nodeless NOD mice was no different from controls (Fig. 4B). By 8 weeks posttransfer, 60% (3 of 5) had developed disease in the two groups, indicating that the peripancreatic lymph nodes are dispensable for the adoptive transfer of diabetes induced by primed T-cells.

This study, in conjunction with others (3,6,22), reveals the critical importance of secondary lymphoid structures, most likely the peripancreatic lymph nodes, in the pathogenesis of autoimmune diabetes in the NOD mouse. Mice without lymph nodes were completely protected from the development of diabetes and had reduced anti–β-cell autoantibodies and reduced numbers of pathogenic T-cells.

The complete protection from disease enjoyed by the nodeless mice is very likely the result of the interruption or prevention of a critical step in the disease process, namely the priming and expansion of β-cell–specific pathogenic T-cells, an event that requires the local concentration of diabetogenic antigens in the milieu of a lymph node. The failure of splenocytes from nodeless animals to transfer disease (1 of 13 at 25 weeks posttransfer) supports the conclusion that β-cell–reactive T-cells were not primed in the nodeless animals or that T-cells existed at a low frequency incapable of inducing disease when transferred into NOD.scid mice.

In contrast to the surgical removal of the peripancreatic lymph nodes at 3 weeks of age, which reduced the incidence of diabetes to ∼20% at 30 weeks of age, (6) LTβR-Ig treatment provided complete protection from diabetes for over 1 year. It is possible that the critical priming event, presumably the encounter of β-cell–specific T-cells with antigen-presenting cells bearing relevant β-cell antigens, may occur before 21 days of age in a subset of animals. Interestingly, the nodeless mice produced anti–β-cell antibodies, although at half the titers compared with controls, while the mice treated with splenectomy had greatly reduced titers of anti-insulin antibodies. Taken together, these results clearly reveal the important role played by both the spleen and the peripancreatic lymph nodes in the production of a vigorous anti–β-cell humoral response.

Finally, as in the study of Gagnerault et al. (6), diabetes developed after the transfer of T-cells into the nodeless mice. This is an indication that the recognition by T-cells of diabetogenic antigens can take place in at least two sites, the local draining lymph node or the pancreas—the former is the favorable site for priming, the latter for the recruitment and development of already-activated T-cells. The peripancreatic node contains diabetogenic T-cells and is indispensable for T-cell induction, as the Gagnerault study and ours indicate (although the mice in our study have a complete absence of lymph nodes). In contrast, the islets can be infiltrated by diabetogenic T-cells independent of lymph nodes, indicating that primed T-cells do not require an intermediate migration through local nodes. The important question of whether the diabetogenic T-cell recognizes the antigen-presenting cells around the islets or within them needs to be determined.

Although the molecular identity of β-cell antigens relevant to disease pathogenesis have yet to be fully elucidated, this study, in addition to others, identifies the local lymph nodes as the critical site of T-cell priming. This knowledge of the anatomy of the disease process will hopefully facilitate the identification of disease relevant autoantigens.

M.G.L. was supported by a grant from the Howard Hughes Medical Institute to Washington University.

The authors thank Dr. Jeffrey L. Browning for the LTβR-Ig and Gina Filley for technical assistance.