Immunological Reversal of Autoimmune Diabetes Without Hematopoietic Replacement of ß Cells

See allHide authors and affiliations

Science  24 Mar 2006:
Vol. 311, Issue 5768, pp. 1778-1780
DOI: 10.1126/science.1123500


Type 1 diabetes mellitus results from the autoimmune destruction of the β cells of the pancreatic islets of Langerhans and is recapitulated in the nonobese diabetic strain of mice. In an attempt to rescue islet loss, diabetic mice were made normoglycemic by islet transplantation and immunization with Freund's complete adjuvant along with multiple injections of allogeneic male splenocytes. This treatment allowed for survival of transplanted islets and recovery of endogenous β cell function in a proportion of mice, but with no evidence for allogeneic splenocyte–derived differentiation of new islet β cells. Control of the autoimmune disease at a crucial time in diabetogenesis can result in recovery of β cell function.

It is now generally accepted that two issues must be addressed for the successful treatment of type 1 diabetes mellitus (T1DM), an autoimmune disease in which T cells kill the β cells responsible for insulin production (1). The persistent autoreactivity to β cell antigens that characterize the disease needs to be controlled and the β cell mass, which is extensively reduced at the time of onset of clinical disease, must be restored. A treatment protocol was recently developed in nonobese diabetic (NOD) mice, an extensively used model of T1DM, that resulted in cessation of autoimmunity and reversal of diabetes through the generation of new β cells from splenic cells (2, 3). Because of the importance of these conclusions in offering novel clinical strategies for disease intervention, we repeated the same experimental protocol as a prelude to further analysis of the process.

The major protocol from the paper of Kodama et al. involved three experimental manipulations in female diabetic NOD mice made between 7 and 20 days after the development of hyperglycemia (2, 3). First, a single subcutaneous injection of Freund's complete adjuvant (FCA) was administered, which is known to stop the autoimmune process in diabetic NOD mice (46). Second, the mice were given a series of intravenous injections of spleen cells derived from major histocompatibility complex (MHC)–mismatched donor F1 (CByB6F1) male mice. Finally, to control the hyperglycemia, syngeneic islets were transplanted under the capsule of one kidney. The mice were followed for a period of at least 120 days after transplant, at which time nephrectomy was performed to remove the islet grafts. Most mice remained normoglycemic, which indicates that β cell function in the pancreas was restored during the time of immunological control (2). Moreover, the islets were reported to be derived from the injected spleen cells because they contained the male Y-chromosome marker of the injected male cells (2).

In our experiments, female NOD mice were maintained on insulin for 7 to 20 days before the three treatments after the first indications of hyperglycemia, to ensure that the islet β cell mass was depleted (7). Mice that remained persistently normoglycemic (22 out of 53) were followed for at least 120 days after treatments. (The 31 transplanted mice that developed periods of hyperglycemia before the 120-day observation period were eliminated.) A nephrectomy, which also removed the transplanted islets placed under the capsule, was performed between days 120 and 146 after transplantation. Examples of experimental mice that reverted back to diabetes, or that maintained normoglycemia after nephrectomy, are shown in Fig. 1. The large majority of the mice (82%) subsequently reverted to the diabetic state (Table 1), which suggests that endogenous β cell function had not been restored in these individuals, and histological analysis confirmed the presence of very few small islets made up entirely of glucagon-positive cells (8). Thus, no insulin-positive cells were detected either in the few remaining small islets or in the ducts of revertant mice (8). In other experiments, the same manipulations were performed with FCA injections and islet transplants (5) but without the inocula of allogeneic cells. In this case, 20 of 29 mice became normoglycemic, some for as long as >100 days (8), an indication that the FCA injection was sufficient for control of the autoimmune process in these mice.

Fig. 1.

Serum glucose levels of two representative mice treated according to the major protocol of Kodama et al. (2, 7). Transplanted mice were followed for at least 120 days, after which the islet transplant under the kidney capsule was removed by nephrectomy. The day of nephrectomy for each animal is indicated in superscript. Of the two mice, one became diabetic immediately after nephrectomy, whereas the other remained normoglycemic.

Table 1.

Compiled results of 22 diabetic female NOD mice after nephrectomy. Mice marked with an asterisk (*) were transplanted with 300 islets; all others received a transplant of 600 to 650 islets. The kidney containing the islet graft was removed between days 120 and 146 after transplant, after which most mice reverted to a diabetic state (“yes” in third column) and few remained normoglycemic (“no” in third column). Many transplanted mice were analyzed for the presence of chimerism (i.e., presence of CByB6F1 cells in spleen and peripheral blood); the presence of alloantibodies (Allo-Abs) to CByB6F1 cells in serum; the ability to mount a mixed lymphocyte reaction (MLR) against CByB6F1 cells; and the ability of their spleen cells to transfer diabetes into NOD.scid recipient mice. n.d., not determined.

MouseNephrectomy (day)DiabeticChimerismAllo-AbsMLRTransfer of diabetes into NOD.scid recipients
2* 126 yes no +++ +++ 5/5 (100%)
35D* 133 no no +++ +++ 5/5 (100%)
38* 133 yes no +++ +++ 3/3 (100%)
61* 136 yes no +++ +++ 5/5 (100%)
66 122 yes no +++ n.d. 3/3 (100%)
79 146 yes no +++ +++ 4/5 (80%)
122 130 yes no +++ - n.d.
124 130 yes no +++ +++ 4/5 (80%)
136D 124 no no +++ +++ 1/3 (33%)
137 124 yes no +++ +++ 3/5 (60%)
138 123 no no +++ n.d. n.d.
140 138 no no +++ n.d. n.d.
147 124 yes no +++ n.d. n.d.
152 120 yes no +++ n.d. n.d.
157 139 yes n.d. n.d. n.d. n.d.
158 134 yes n.d. n.d. n.d. n.d.
164 134 yes n.d. n.d. n.d. n.d.
159 132 yes n.d. n.d. n.d. n.d.
160 132 yes n.d. n.d. n.d. n.d.
166 128 yes n.d. n.d. n.d. n.d.
174 121 yes n.d. n.d. n.d. n.d.
175 125 yes n.d. n.d. n.d. n.d.

In the four remaining mice of the original experiment (Table 1) that maintained normoglycemia after nephrectomy, few islets were found, all of which contained about 75% or more glucagon- and somatostatin-positive cells (8). However, some islets also contained insulin-positive cells and were also positive for the β cell–specific transcription factor PDX-1 (fig. S1, A and B) (7, 9). A few cells were negative for insulin but positive for PDX-1, which suggests that such β cells had undergone extensive insulin degranulation (fig. S1B) (7). Fluorescence in situ hybridization (FISH) analysis failed to detect the presence of Y chromosome–positive cells in any of the islets examined (Fig. 2A and fig. S2) (7).

Fig. 2.

(A) FISH analysis of islets from mouse 138 to detect the presence of Y chromosome (18). Nuclei are stained with 4′,6′-diamidino-2-phenylindole (DAPI). Positive control is indicated. Islets are indicated by arrows. (B) Experiment searching for GFP-positive cells in islets. Immunoblot analysis of pancreatic extracts from F1-GFP, NOD or five experimental mice of protocol 2 (15 to 19) searching for GFP. Glyceraldehyde phosphate dehydrogenase (GAPDH) was used as an internal loading control. Although the control was positive, none of the extracts from the experimental mice were reactive with GFP antibodies.

No evidence of chimerism of the F1 male cells in the NOD female mice was detected with flow cytometry analysis (Table 1 and fig. S3A) (7). Moreover, the spleen cells exhibited a mixed lymphocyte reaction when tested with irradiated F1 cells (Table 1 and fig. S3B) (7), and the sera contained alloantibodies against the F1 cells (Table 1 and fig. S3C) (7). Taken together, these demonstrate a clear host immune response against the injected allogeneic male F1 cells. In addition, the spleen cells from the treated mice transferred diabetes when injected into NOD.scid recipients (Table 1 and fig. S3D) (7), which demonstrates that diabetogenic T cells had persisted, most likely in a quiescent state, in mice treated with FCA and F1 cells (5, 10).

In the second of their protocols, Kodama et al. treated prediabetic NOD female mice with FCA and green fluorescent protein (GFP)–expressing male (CByB6F1) spleen cells (2). These mice failed to develop diabetes and expressed GFP in their islets. Repeating this second protocol, 12-week-old prediabetic NOD female mice received a single injection of FCA, along with four injections of GFP-expressing CByB6F1spleen cells (5 × 105 per injection, administered over 2 weeks). Included was a control group of 12-week-old prediabetic NOD female mice that received a single injection of FCA alone. In both cases, all mice were protected from diabetes (fig. S4) (7). Moreover, intact islets, some with evidence of insulitis, were present in both groups of protected mice. However, we failed to detect GFP+ cells by immunohistochemistry or Western blot analyses (Fig. 2B). In addition, spleen cells from GFP-treated mice retained the capacity to transfer diabetes into third-party NOD.scid recipients (fig. S5) (7).

In our experiments, no evidence of replacement of islets by the allogeneic spleen cells was observed. Moreover, immunological reaction to the injected cells was seen, as would be anticipated. Finally, the presence of diabetogenic T cells, albeit in a quiescent state, indicated that the abnormal autoimmune process persisted in these treated mice. It is likely that the normoglycemia observed in the four mice from the original protocol resulted from a few β cells that had survived the initial T cell attack and subsequently expanded to maintain blood glucose levels. Potentially, these could derive from precursor cells found in ducts or from surviving preexisting β cells, which is consistent with a similar conclusion found in two recent studies (11, 12). We also examined the pancreas of 14 diabetic mice maintained on insulin for 2 to 3 weeks, that is, under the same initial situation as the mice in the original protocol. We observed in this experiment a large variation in the number of surviving islets and β cells, with 3 of the 14 mice examined containing about 6% of the normal content of β cells and the rest having less (fig. S6A) (7). Most of the islets were abundant in glucagon-positive cells, although some cells were PDX-1 positive but negative for insulin (fig. S6B) (7). Thus, even in the presence of marked hyperglycemia, an indication of advanced diabetes, some β cells can survive for at least a period of time. It is possible that, as their insulin content decreases, the surviving β cells become less susceptible to attack by the insulin-specific T cells that dominate the immunological reaction (1315). It is noteworthy that at the time of the first reading of hyperglycemia, the number of β cells was decreased by 71% in five mice examined. Between the time of the first evidence of hyperglycemia and the time the protocol was started [that is, 7 to 20 days (2, 3)], there was a precipitous drop in islet β cell numbers. Along these lines, other studies demonstrated the successful reversion of disease with antibodies to CD3 or CD4 molecules, but only in newly diabetic NOD mice (16, 17). In sum, the present findings in the 20% of mice displaying restored islet function when there is immunological regulation indicate that some β cells survived immunologic attack. Finding means of preserving these cells and stimulating their expansion could represent an exciting new approach to the prevention and treatment of T1DM.

Supporting Online Material

Materials and Methods

Figs. S1 to S6


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article