A Coxsackievirus B vaccine protects against virus-induced diabetes in an experimental mouse model of type 1 diabetes

Aims/hypothesis Epidemiological studies suggest a role for Coxsackievirus B (CVB) serotypes in the pathogenesis of type 1 diabetes, but their actual contribution remains elusive. In the present study, we have produced a CVB1 vaccine to test whether vaccination against CVBs can prevent virus-induced diabetes in an experimental model. Methods NOD and SOCS1-tg mice were vaccinated three times with either a formalin-fixed non-adjuvanted CVB1 vaccine or a buffer control. Serum was collected for measurement of neutralising antibodies using a virus neutralisation assay. Vaccinated and buffer-treated mice were infected with CVB1. Viraemia and viral replication in the pancreas were measured using standard plaque assay and PCR. The development of diabetes was monitored by blood glucose measurements. Histological analysis and immunostaining for viral capsid protein 1 (VP1), insulin and glucagon in formalin-fixed paraffin embedded pancreas was performed. Results The CVB1 vaccine induced strong neutralising antibody responses and protected against viraemia and the dissemination of virus to the pancreas in both NOD mice (n = 8) and SOCS1-tg mice (n = 7). Conversely, 100% of the buffer-treated NOD and SOCS1-tg mice were viraemic on day 3 post infection. Furthermore, half (3/6) of the buffer-treated SOCS1-tg mice developed diabetes upon infection with CVB1, with a loss of the insulin-positive beta cells and damage to the exocrine pancreas. In contrast, all (7/7) vaccinated SOCS1-tg mice were protected from virus-induced diabetes and showed no signs of beta cell loss or pancreas destruction (p < 0.05). Conclusions/Interpretation CVB1 vaccine can efficiently protect against both CVB1 infection and CVB1-induced diabetes. This preclinical proof of concept study provides a base for further studies aimed at developing a vaccine for use in elucidating the role of enteroviruses in human type 1 diabetes. Electronic supplementary material The online version of this article (10.1007/s00125-017-4492-z) contains peer-reviewed but unedited supplementary material, which is available to authorised users.


Animals
Animals were housed in ventilated cages in an SPF unit at the Karolinska University Hospital Huddinge, Stockholm, and provided with water and food ad libitum.
Animals were not single housed and a maximum of 5 mice were housed in the same cage.
SOCS1-tg NOD mouse generation, breeding and screening is described by Flodström et al (2002) in [1]. Briefly, SOCS1-tg mice were bred by crossing heterozygous SOCS1-tg mice with NOD mice. The offspring were genotyped to identify the transgenic (SOCS1-tg) and non-transgenic (NOD) mice (primers given in ESM Table   1 and named as hbs primer 1 and 2 as they amplify part of the hbs transgene, inserted into the human insulin promoter/SOCS1-construct used to generate the SOCS1-tg animals).
Mice were anaesthetised with isofluorane, prior to heart puncture and blood drawing, and then killed by cervical dislocation.

Virus titration and tissue homogenisation
Tissues were homogenized using sterile ceramic beads (2.8mm, QIAGEN, Hilden, Germany) by vigorous shaking with a PowerLyzer 24 Bench Top Homogenizer (MoBio Laboratories, QIAGEN) device for three 45 second repeats at 3200rpm. Lytic virus was measured in samples by standard plaque assay in GMK cells and viral titres expressed as PFU/g of wet tissue or ml of blood. Samples were analysed on two separate occasions and the mean values used.

PCR analysis
Blood samples were diluted 1:60 in Hanks solution. RNA was extracted from 140µl diluted blood sample with QIAmp RNA Blood Mini Kit (QIAGEN, Germany).
Enterovirus specific real-time PCR and product amplification was performed as described in [3] with QuantiTect Probe kit (QIAGEN, Germany) and Taqman chemistry according to the manufacturer's protocol. The primers (which amplify the 5' non coding region of all known enterovirus serotypes) and probes used are shown in ESM Table 2. The PCR reaction was run in an ABI 7900HT machine (Fisher Scientific, Vantaa, Finland). All samples were run in triplicate. CVB3 and water were used as positive and negative controls respectively.
Immunohistochemistry and antibodies VP1 was detected by immunohistochemistry using a biotinylated anti-VP1 antibody (5D8/1; DAKO, Ely, UK; biotinylated by Capra Bioscience, Ängelholm, Sweden, 0.07µg/ml final concentration, Tris EDTA pH6 antigen retrieval). The anti-VP1 antibody was biotinylated as it is raised in mouse, and biotinylation circumvents the need for a secondary antibody (which in this case would bind to mouse immunoglobulins in mouse tissue, leading to unspecific staining). The biotinylated anti-VP1 antibody was validated in infected mouse pancreas known to be VP1 positive. Insulin (1:10,000, A0564, Dako, Ely, UK) and glucagon (1:3000, A0565, Dako, Glostrup, Denmark) staining were carried out as in [1] and both antibodies were validated in mouse pancreas sections. Secondary antibody alone acted as a negative control for insulin (goat-anti guinea pig, Vector Laboratories, Burlingame, CA, USA) and glucagon (goat-anti rabbit, Dako, Glostrup, Denmark) and 2% NGS acted as the negative control for the VP1 experiments.