Considerable epidemiologic, immunohistochemical, and virological evidence has accumulated supporting the hypothesis that type 1 diabetes (T1D) is associated with enteroviral infection. However, the majority of immunohistochemical evidence is based on the detection of the enteroviral capsid antigen VP1, using a commercially available monoclonal antibody (clone 5D8/1). Hence, there is a pressing need to confirm these data using additional methods. Accordingly, a coordinated blinded analysis of enteroviral protein and RNA was undertaken in multiple laboratories using different techniques in tissues recovered from controls, T1D patients and those at risk of T1D, from the nPOD collection.
Materials and Methods
Serial pancreas sections from a total of 30 controls, type 1 diabetes and autoantibody positive cases were prepared by the nPOD Pathology Core and distributed blindly to nPOD-V consortium laboratories. Samples were immunostained (IHC) for enteroviral capsid protein and class I MHC (UK, Finland) and probed by in situ hybridisation for enteroviral genome (Finland). Each case was assigned a score, ranging from 1-4, based on the number of independent positive indicators of viral infection (IHC+ UK; IHC+ Finland; ISH+ Finland; Hyper expression of class I MHC, UK).
Unblinding revealed that IHC results were concordant between laboratories in 29/30 cases (96%) while the concordance rate between ISH and IHC signals was 69% (20/29 cases). In patients with residual insulin-containing islets (ICIs), 11 of the 15 cases examined had more than 3 positive indicators of viral infection. In contrast, no viral indicators were found in 6/7 control samples. Analysis of T1D cases containing only insulin-deficient islets (IDIs) revealed a total lack of viral indicators. This agrees with previous results demonstrating that once beta cells have been destroyed, evidence of viral infection is lost.
The results support the conclusion that a bona fide enteroviral infection occurs at much higher frequency in T1D pancreas than in controls.