Susceptibility to immune-mediated diabetes (IMD) in humans and NOD mice involves their inherently defective T cell immunoregulatory abilities. We have followed natural killer (NK) T cell numbers in patients with IMD, both by flow cytometry using mAbs to the characteristic junctions found in the T cell receptors of this cell subtype, and by semiquantitative RT-PCR for the corresponding transcripts. Both before and after clinical onset, the representation of these cells in patients' PBMCs is reduced. We also report low numbers of resting CD4(+) CD25(+) T cells in IMD patients, a subset of T cells shown to have important immunoregulatory functions in abrogating autoimmunities in 3-day thymectomized experimental mice. Whereas a biased Th1 to Th2 cytokine profile has been suggested to underlie the pathogenesis of IMD in both species, we found defective production of IFN-gamma in our patients after in vitro stimulation of their PBMCs by phorbol-myristate acetate and ionomycin and both IFN-gamma and IL-4 deficiencies in V(alpha)24(+) NK T-enriched cells. These data suggest that multiple immunoregulatory T (Treg) cell defects underlie islet cell autoimmunity leading to IMD in humans and that these lesions may be part of a broad T cell defect.

Bloom syndrome is a rare autosomal disorder characterized by predisposition to cancer and genomic instability. BLM, the structural gene mutated in individuals with the disorder, encodes a DNA helicase belonging to the RecQ family of helicases. These helicases have been established to serve roles in both promoting and preventing recombination. Mounting evidence has implicated a function for BLM during DNA replication; specifically, BLM might be involved in rescuing stalled or collapsed replication forks by a recombination-based mechanism. We have tested this idea by examining the binding and melting activity of BLM on oligonucleotide substrates containing D-loops, DNA structures that model the presumed initial intermediate formed during homologous recombination. We find that BLM preferentially melts those D-loops that are formed more favorably by the strand exchange protein Rad51, but whose polarity could be less favorable for enabling restoration of an active replication fork. We propose a model in which BLM selectively dissociates recombination intermediates likely to be unfavorable for recombination-promoted replication.

Bloom syndrome (BS) is caused by homozygous or compound heterozygous mutations in the RecQ DNA helicase gene BLM. Since the molecular isolation of BLM, characterization of BS-causing mutations has been carried out systematically using samples stored in the Bloom's Syndrome Registry. In a survey of 134 persons with BS from the Registry, 64 different mutations were identified in 125 of them, 54 that cause premature protein-translation termination and 10 missense mutations. In 102 of the 125 persons in whom at least one BLM mutation was identified, the mutation was recurrent, that is, it was shared by two or more persons with BS; 19 of the 64 different mutations were recurrent. Ethnic affiliations of the persons who carry recurrent mutations indicate that the majority of such persons inherit their BLM mutation identical-by-descent from a recent common ancestor, a founder. The presence of widespread founder mutations in persons with BS points to population genetic processes that repeatedly and pervasively generate mutations that recur in unrelated persons.