Sulfatide (3 ' sulfogalactosylceramide) is a glycosphingolipid present with
in the nervous system and in the islets of Langerhans. Anti-sulfatide antib
odies have been observed in both prediabetic and newly diagnosed type I dia
betic patients. The aim of this study was to test in vivo, the therapeutic
effect of sulfatide on the development of diabetes in the NOD mouse. In fou
r separate experiments diabetogenic splenocytes from newly diabetic NOD mic
e were injected iv into 7-8 week old irradiated (700R) female NOD mice (4-1
0 million cells/mouse). Each experiment consisted of four treatment groups
to which the mice were randomly divided: 1) sulfatide; 2) galactosylceramid
e (the precursor to sulfatide without sulfate); 3) GM I, a glycosphingolipi
d negatively charged as sulfatide but with a different sugar composition; a
nd 4) phosphate buffered saline (PBS). The mice received 100 tg glycosphing
olipid iv on the day of cell transfer and 1-3 times thereafter at four day
intervals, and were screened for diabetes three times a week the next 52 da
ys. Among all the 35 sulfatide-treated mice 54% became diabetic compared to
93 % of 43 PBS-treated animals (p < 0.00001). Correspondingly, galactosylc
eramide reduced diabetes incidence to 52% (25 mice, p < 0.00001). On the ot
her hand, 86% of GM1-treated mice (n=28) became diabetic indicating that no
effect was obtained by this glycosphingolipid. In two experiments in which
less spleen cells were transferred (4-5 mill.) and glycosphingolipids were
given 4 times, 35% of the sulfatide-treated animals (n = 17) developed dia
betes compared to 85% of PBS-treated mice (n = 20, p <less than> 0.001). A
robust proliferative response to sulfatide, but none to GM1, was observed w
hen spleen cells were rechallenged with glycosphingolipid in vitro. Thus, l
ike insulin and GAD, sulfatide is able to prevent diabetes in NOD mice.