Reversal of Diabetes in Non-Obese Diabetic Mice Without Spleen Cell-Derived ß Cell Regeneration

See allHide authors and affiliations

Science  24 Mar 2006:
Vol. 311, Issue 5768, pp. 1774-1775
DOI: 10.1126/science.1123510


Autoimmune destruction of β cells is the predominant cause of type 1 diabetes mellitus (T1DM) in humans and is modeled in non-obese diabetic (NOD) mice. Many therapeutic interventions prevent the development of T1DM in NOD mice, but few can induce its reversal once established. Intervention with Freund's complete adjuvant, semi-allogeneic splenocytes, and temporary islet transplantation has been reported to cure NOD mice of established T1DM. Using the same approach, we report here that this treatment cured 32% of NOD mice of established diabetes (>340 milligrams per deciliter blood glucose), although β cells in these mice were not derived from donor splenocytes.

Current therapy for T1DM, based on the replacement of insulin or on islet transplantation and immune suppression, has substantial limitations (1). An alternative approach to curing T1DM is through the simultaneous reversal of preexisting autoreactivity and restoration of endogenous β cell function. Recent studies by Kodama et al. (2) suggested that treatment with Freund's complete adjuvant (FCA), semi-allogeneic spleen cells, and temporary islet transplantation leads to the control of autoimmunity and alloreactivity and to a marked restoration in β cell mass through the in vivo differentiation of stem cells from the spleen. We attempted to independently replicate these findings using spleen cells from a transgenic mouse strain in which the mouse insulin promoter drives the expression of green fluorescent protein (MIP-GFP); thus, GFP is expressed only in their β cells and not in their spleen cells (3).

Severely diabetic NOD mice (blood glucose levels of >300 mg/dl for 2 consecutive days; table S1) were injected with FCA and live semi-allogeneic spleen cells (CByB6F1) from MIP-GFP mice. To initially control blood glucose levels, 400 to 500 islets from NOD-scid mice were implanted under a single kidney capsule of NOD recipients. Of the 22 diabetic NOD mice that were transplanted and treated with FCA and spleen cells, 14 (63.6%) redeveloped diabetes (blood glucose >200 mg/dl) within 40 days, and one within 80 days, of initiation of treatment (Fig. 1A and table S1). To confirm that the redevelopment of diabetes was due to a recurrence of autoimmunity, a second syngeneic islet transplant (200 islets per mouse) was performed in six mice (with >200 mg/dl blood glucose levels for 2 to 30 days). Four of the six mice became diabetic again within 1 week of retransplantation, consistent with recurrent autoimmunity. Two of the six remained normoglycemic, one for >52 days and the other for >100 days, suggesting that the initial recurrence of diabetes was due not to the inability of the treatment to control autoimmunity but to insufficient β cell mass. Thus, treatment with FCA and spleen cells reversed autoimmunity in 41% (9 out of 22) of the diabetic NOD mice. Seven of the nine mice remained normoglycemic for ≥90 days; the remaining two mice with normal glycemia were observed for only 52 and 53 days. Six of the seven mice were successfully nephrectomized, after which all remained normoglycemic for >29 days (Fig. 1B and table S1), indicating that normoglycemia could be maintained by endogenous β cells. All four of the successfully treated (Rx-NOD) mice that were subject to intraperitoneal glucose tolerance testing displayed almost normal responses (fig. S1).

Fig. 1.

Reversal of established diabetes in NOD mice. (A) Kaplan-Meier plot of the percentage of nondiabetic mice after the first islet transplantation (n = 22). (B) The blood glucose (BG) levels of three NOD mice successfully treated with FCA, biweekly infusions for 6 weeks of semi-allogeneic spleen cells, and a subrenal transplant of syngeneic islets. The bar indicates the period of normoglycemia after the subrenal islet grafts were removed. (C to E) Immunohistochemistry of β cells. Pancreatic sections from age-matched NOD-scid (C), successfully treated NOD mice (D), and age-matched (≥45 weeks old) NOD mice that did not spontaneously develop diabetes (E). (F) The maximal diameter (empty bars) and average volume (shaded bars; Embedded Image, where r is the radius) of the islets ± SEM were calculated for all the islets observed in successfully treated (Rx) and control (Ctrl) NOD mice, and all insulin-positive islets (>80 per section) from two randomly selected sections per NOD-scid (Scid) mouse (n = 3 per group). The total volume ± SEM (filled circles) of β cells was estimated by multiplying the average volume by the total number of islets (NOD-scid mice were assumed to have 2000 islets). (G to H) Visualization of β cells from control MIP-GFP (G) or successfully treated NOD mice (H). Occasional green fluorescence observed in the pancreas of successfully treated NOD mice was due to autofluorescence and was not associated with β cells.

Histological examination of the entire pancreas indicated that five of the six Rx-NOD mice had very few hyperplastic islets (Fig. 1, D to F) and that one had few small islets that stained very weakly or did not stain for insulin, similar to the observations by Suri et al. (4). The mean number of hyperplastic islets (±SEM) in the pancreas of the five Rx-NOD mice examined was 12.2 ± 2.2. This number is considerably lower than that observed in age- and sex-matched NOD-scid mice or young prediabetic NOD mice [which are estimated to have ≥2000 islets per NOD mouse (5)]. The average diameter of the islets from Rx-NOD mice was 5.2 ± 0.1 times as large as that of age-matched NOD-scid mice (Fig. 1, C, D, and F; P < 0.05). This increase in islet size was due not to hypertrophy of individual β cells but to increased numbers of β cells and not of glucagon-expressing α cells (fig. S2). We estimated that the total β cell volume in the Rx-NOD mice was 22.5 ± 4.1% of that of age-matched NOD-scid mice. Similarly low numbers (12.2 ± 1.7) of significantly enlarged islets (P < 0.05) were observed in age-matched NOD mice that failed to develop spontaneous diabetes (Fig. 1, D to F).

To determine whether the β cells in the Rx-NOD were of host origin or derived from donor spleen cells, we examined all the Rx-NOD islets for expression of GFP. No β cells expressing GFP were detected in any of the islets examined (Fig. 1, G and H) or in the organs outside of the pancreas. Thus, our data do not support a conclusion of β cell regeneration from spleen-derived stem cells. All the hyperplastic islets from the five cured Rx-NOD mice were associated with circumferentially distributed lymphocytes (peri-insulitis) of donor origin (fig. S3) and devoid of invasive insulitis. The majority of these peri-islet lymphocytes were CD4+CD25+ cells, with a small portion expressing the transcriptional factor FoxP3+ (fig. S3) (6). These observations suggest that regulatory T cells may function to control autoimmunity locally where β cells are situated. The islets that stained very weakly or did not stain for insulin in the one Rx-NOD mouse without hyperplasia did not have a notable peri-islet lymphocytic infiltrate.

The finding that autoimmune diabetes in NOD mice can be reversed by a complex protocol combining islet transplantation, FCA treatment, and transfer of semi-allogeneic splenocytes suggests a potential for novel β cell replacement therapies for human T1DM (2, 7). We confirmed these observations and demonstrated restoration of diabetic NOD mice to normoglycemia with this therapeutic protocol. We did not observe β cell reconstitution from the infused spleen cells, and these studies were not designed to determine whether the observed host-derived β cells were the result of replication of preexisting β cells or differentiation from stem-cell precursors. Our studies indicate that normoglycemia in Rx-NOD mice can be maintained by a very low number of hyperplastic islets or, in one case, by weakly insulin-positive or insulin-negative islets surviving the autoimmunity. The consistent association between islet hyperplasia and peri-insulitis (8, 9) suggests a hypothesis that autoreactivity, when restrained by regulatory T cells, may facilitate the development of islet hyperplasia. The limited β cell mass required to maintain normoglycemia in this model contrasts with the large numbers of transplanted islets required to maintain normal glycemia (1, 10). Our studies confirm that autoimmune diabetes can be reversed and that sufficient endogenous β cell mass can be restored to cure diabetic NOD mice with the treatment protocol developed by Faustman and colleagues (2, 7). Translating these findings into therapies useful for curing T1DM in humans remains a challenge.

Supporting Online Material

Materials and Methods

Figs. S1 to S3

Table S1


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article