Circulating MAIT cells frequency before (n = 69) and at 3, 6, and 12 months after surgery (n = 35, 34, and 35, respectively). (Control individuals, n = 23.) *P = 0.01, ***P < 0.002, † P < 0.0001. Circulating MAIT cell frequency was significantly lower in obese patients at each time point compared to control individuals (P < 0.05). (F) Cytokine production after PMA-ionomycin stimulation of MAIT cells from healthy individuals (n = 20) and obese patients before surgery (n = 39) and 3, 6, and 12 months after surgery (n = 38, 33, and 31, respectively).

The nonobese diabetic (NOD) mouse was established as an inbred strain in 1980 and proposed as a model of type I diabetes mellitus (1). By 6-8 wk of age, mononuclear cells start infiltrating the periphery of pancreatic Langerhans' islets of both males and females. Progressive invasion inside the islets occurs later and is correlated with selective destruction of insulin-producing ß cells and with the onset of clinically overt diabetes . Diabetes is first observed at 12 wk of age and strongly predominates in females. By 30 wk of age,^-70% of females have become diabetic, while <20% of males develop overt disease (2). Several lines of evidence suggest that diabetes in the NOD mouse is an autoimmune disease mediated by T cells: First, Thy-1,2+ cells predominate in the cellular islet infiltration (3); second, neonatal thymectomy prevents the disease (4) ; third, NOD nu/nu mice do not develop diabetes (5).Further identification of immune cells involved in the destruction of insulinproducing cells has been hindered by lack of suitable in vivo transfer models . The NOD mouse has a particular MHC haplotype due to unique I-A specificity (6) that prevents the inoculation of NOD lymphoid cells into MHC-compatible strains. Attempts to derive lines of nondiabetic mice from the original NOD nucleus have also been unsuccessful . Recently, it has been shown (7) that the transfer of spleen cells from diabetic mice into diabetes-prone NOD adults greatly promoted the onset of overt diabetes, provided that the recipients had been sublethally irradiated. In addition, adoptive transfer required recipients older than 6-8 wk who, presumably, had already begun to self-damage their pancreatic islets as inferred from histological studies (2). This latter condition limits the validity of the model to account for the whole history of ß cell destruction, particularly in its initial stages .In this study we show that diabetes can be adoptively transferred to NOD neonates by spleen cells from diabetic NOD donors . Overt diabetes, which is correlated with >90% of ß cell destruction (8, 9) occurred as early as 21 d of age, at a time when pancreases of noninjected mice were still free of histological changes. The susceptibility of the recipients to the transfer was limited in time and declined after 3 wk of age. We also show that the neonatal model of transfer A. Bendelac was supported by a fellowship from the Foundation pour la Recherche Médicale. This work was supported by funds from INSERM and Centre National de la Recherche Scientifique. J. Exp. MED.

Maturity-onset diabetes of the young type 5 encompasses a wide clinical spectrum. Analysis for mutations of HNF-1beta is warranted, even without a family history of diabetes, in nonobese patients with diabetes and slowly progressive nondiabetic nephropathy, particularly when pancreatic atrophy or genital abnormalities are present.