Marc Horwitz

Autoimmune disorders like type 1 diabetes (T1D) are complex diseases caused by numerous factors including both genetic variance and environmental factors. Two such factors, enterovirus infection and disruption of the intestinal microbiota, have been independently associated with T1D onset in both humans and animal models. Existing research has established a gut-pancreas interaction that links dysbiosis and inflammatory events within the gastro-intestinal (GI) environment to pancreatic immune homeostasis. Thus, I hypothesized that modulation of gut homeostasis due to infection with a T1D-associated virus, coxsackievirus B4 (CVB4) can shape autoreactive potential and pathology within the islet microenvironment. To determine the link between virus infection and the microbiome, non-obese diabetic (NOD) mice were infected with CVB4 and changes in the microbiome and GI homeostasis were observed. Infection accelerated diabetes autoimmunity and caused restructuring of intestinal microbial community composition as well as alteration of intestinal physiology exhibited by eroded mucosal barriers, increased intestinal permeability, bacterial translocation, and modulation of anti-commensal bacteria antibody production. These findings support an overlap in known environmental risk factors of T1D and suggest that microbiome disruption and impaired intestinal homeostasis contribute to CVB-enhanced autoreactivity and T1D. Commensal microbes influence host immunity through production of small molecule metabolites including short-chain fatty acids (SCFAs) to skew host immunity. Targeted metabolomics and supplementation of mice with exogenous SCFAs prior to CVB4 infection determined both how virus infection modifies the functional activity of the microbiome as well as the ability of bacterial metabolites to affect virus-accelerated diabetes. Collectively, these results suggest bacteria-produced metabolites may protect from virus-induced diabetes autoimmunity. The events within pancreatic tissues leading to T1D are not entirely understood and disruption of immune homeostasis within the pancreas is detrimental for T1D development. Flow cytometry was used to characterize how infection with CVB4 can alter immune profiles in the islets to promote inflammation and destruction of the insulin-secreting beta cells leading to T1D onset. Ultimately, CVB4 infection promoted autoreactive tissue-resident memory T (TRM) cells to invade the islet microenvironment, suggesting this cell population represents a particularly pathogenic T cell subset that contributes to diabetes autoimmunity in NOD mice.

View record

Multiple sclerosis (MS) and rheumatoid arthritis (RA) are two of the most common autoimmune diseases, but the etiologies are not well understood. Both genetic and environmental factors contribute to disease susceptibility. In particular, a history of Epstein-Barr virus (EBV) infection, a common gammaherpesvirus, is epidemiologically and clinically associated with the development of MS and RA, though the immunological mechanisms(s) that mediate this relationship are not well understood. We hypothesize that latent EBV infection contributes to MS and RA via a common immunologic mechanism. Here, I present evidence towards a novel model by which gammaherpesvirus infection contributes to arthritis. Mice with latent gammaherpesvirus infection prior to the induction of collagen-induced arthritis display exacerbated joint swelling and a more inflammatory immune response compared to non-infected mice. This builds off of our prior work that demonstrated that latent gammaherpesvirus infection exacerbates disease in a murine model of MS. These two in vivo models were studied, along with samples from adults with MS, to investigate immunological mechanisms of the viral contribution to autoimmunity and age-associated B cells (ABCs). ABCs are a recently identified B cell population that increases in number during both viral infection and autoimmunity. Here, I performed a side-by-side comparison of the ABC population during viral infection and autoimmunity. I show that the circulating ABC population is expanded in people with MS and that EBV infection and MS status differentially impact the ABC phenotype. Mouse models of infection and disease are studied to characterize ABC abundance and phenotype and I found that ABCs are required for viral exacerbation of disease in models of both RA and MS. In the final chapter, ABCs were characterized throughout acute and latent gammaherpesvirus infection. I demonstrated that during infection, there was a female sex bias resulting towards increased frequency of ABCs and that a portion of ABCs were infected with gammaherpesvirus. Finally, while ABCs are dispensable for the development and maintenance of latency in steady-state conditions, I showed that ABCs are critical for controlling gammaherpesvirus latency during challenge. Here, the ABC population is identified as a critical mediator of latent gammaherpesvirus infection and autoimmunity.

View record

Multiple Sclerosis (MS) is caused by a combination of genetic and environmental factors. It is believed that previous infection with Epstein Barr Virus (EBV) plays an important role in the development of MS. To examine this, our lab developed a murine model where latent infection with gamma herpesvirus 68 (γHV-68), the murine homolog to EBV, enhanced the symptoms of experimental autoimmune encephalomyelitis (EAE), resulting in disease that more closely resembles MS development in humans. Latency of γHV-68 primarily occurs in memory B cells. Here, we explored which conditions were necessary for EAE enhancement. We showed that latently infected memory B cells were capable of enhancing EAE symptoms when transferred from mice infected with γHV-68 into uninfected mice. We observed a prevention of enhancement when B cells were depleted before infection. However, depletion after the establishment of latency only partially reduced EAE. This indicated the existence of a mechanism where B cells likely play an important role as antigen presenting cells (APCs) prior to EAE induction for the priming of Th1 cells, and that these signals persist in the mouse even after B cell depletion. Using RNAseq analysis we identified 22 genes that differ between B cells from infected versus uninfected mice. Upon the establishment of latency, CD11b+CD11c+ cells show upregulation of CD40 and pSTAT1. This suggests a contribution by type I IFNs in EAE enhancement. To test this, we used IFNARko mice, however, there was no difference in EAE enhancement, suggesting that while type I IFNS are relevant for maintenance of latency and antiviral responses, they are not contributing to EAE exacerbation. Finally, while infection of mice at a young age did not result in EAE enhancement, infection in adult mice left a long-lasting immune impact that directed enhanced autoimmunity even after 5 months of latent infection. Based on these results we propose a mechanism for how gammaherpesviruses contribute to the development of autoimmunity and suggest that cells infected with EBV can be the target of new therapeutic treatments that would be highly effective and less damaging to the immune system of MS patients.

View record

Over 2 million people are infected with HIV each year. The majority of these infections occur in women residing in low-income countries, where their access to and control over preventative measures is often limited. This suggests that female-controlled prevention options for HIV-1 are urgently needed. Microbicides, which can be topically applied to the vaginal tract in advance of sexual activity to protect from HIV-1 infection, represent a promising female-controlled prevention option. We have investigated the development of an HIV-1 specific microbicide using the non-pathogenic, freshwater bacterium Caulobacter crescentus. C. crescentus contains a Surface or S-layer that is easily modified for high-density display of recombinant proteins. We have developed 18 recombinant C. crescentus that display anti-HIV proteins, including decoy receptors and ligands, anti-viral lectins and fusion inhibitors, that can prevent various steps of the HIV-1 attachment and entry process. In vitro testing with these recombinant C. crescentus indicated that 15 were able to provide substantial protection from HIV-1 infection. Studies with immune-competent mice demonstrated that application of C. crescentus to the vaginal tract does not induce the production of inflammatory cytokines or recruitment of immune cells. To test for protection against HIV-1 in vivo we have combined the implantation of human fetal liver and thymus tissue with the intravenous injection of autologous CD34+ cells into NOD-scid IL2Rγnull mice to create humanized Bone Marrow-Liver-Thymus (BLT) mice and we have demonstrated that the peripheral blood of these mice contains human CD45+ cells, including CD4+ and CD8+ T cells, B cells, myeloid cells, and NK cells. Our data indicates that vaginal application of recombinant C. crescentus at the time of HIV-1JR-CSF infection provides protection from HIV-1 infection. Seven of the recombinant C. crescentus were predicted to also prevent infection with herpes simplex virus 2 (HSV-2). HSV-2 is a major co-morbidity for HIV-1 infection, contributing to a 2-4 fold increase in acquisition. Four recombinant C. crescentus provided significant protection from HSV-2 infection in vivo. Taken together this data suggests that a C. crescentus based microbicide could be a safe and effective method to prevent infection with HIV-1 and HSV-2, having considerable impact on public health.

View record

Type 1 diabetes (T1D) is a debilitating disease involving the autoimmune destruction of insulin-producing pancreatic β-cells. The personal and economic burden of this disease is enormous, therefore simpler and more cost effective therapeutic approaches than those currently available must be explored.In children at risk for T1D, a unique type 1 interferon (IFN-I) transcriptional signature precedes islet autoimmunity. Recent onset of T1D strongly associates with infection by RNA viruses like coxsackievirus that induce IFN-I. Importantly, genetic variants in the T1D risk locus IFIH1 are linked to protection from T1D and result in reduced expression of the RNA virus sensor melanoma differentiation-associated protein 5 (MDA5), which is also a critical component in establishing the IFN-I signature. In chapter 2 we describe a novel model where we have translated the reduced MDA5 expression phenotype observed in patients onto the non-obese diabetic (NOD) mouse and established its importance in T1D. We describe the first observations that a reduction in MDA5 in the NOD mouse protects from spontaneous and coxsackievirus B4 (CB4)-induced T1D. We also establish the importance of a specific IFN-I signature in the development of T1D as a result of reduced (not eliminated) MDA5 sensing of CB4 that allows for regulatory T cells (Tregs) at the site of autoimmunity and protects from CB4 induced T1D. In chapter 3 we show that this unique IFN-I signature is limited to MDA5 and disease pathogenesis is linked to the specific IFN-I response induced by the virus as a strain of CB3 failed to modify the IFN-I signature associated with disease. Further RNA sequencing discussed in Chapter 4 demonstrates unique tissue-specific differential gene profiles associated with a reduction in MDA5 following CB4 infection. Our results support our hypothesis that there is a direct correlation between the IFN-I signature induced following environmental challenge with the induction of a strong effector T cell and a matched Treg response. This work demonstrates the essential role of MDA5 signaling in regulating the IFN-I signature, implicates MDA5 in T1D susceptibility and in protection against IFN-I and T1D-inducing agents like CB4 and suggests restricting MDA5 function as a potential T1D therapeutic.

View record

The development of autoimmune diseases is thought to involve both genetic and environmental factors. Epstein-Barr virus (EBV) has been associated with the development of both multiple sclerosis (MS) and systemic lupus erythematosus (SLE). My research projects aimed at identifying the mechanism that this virus is exploiting to cause autoimmunity.In the first project, the role of EBV infection of the blood brain barrier (BBB) as a trigger of MS was investigated. EBV was found to be able to infect human primary endothelial cells isolated from the BBB. EBV infected brain endothelial cells upregulated pro-inflammatory mediators and supported increased immune cell adhesion. These results suggest that EBV has the ability of increasing the BBB permeability. EBV latency and reactivation in endothelial cells could lead to initial inflammation and infiltration of the first wave of autoreactive immune cells during MS initiation. The second and third research project were aimed at developing a mouse model to study the interactions between latent gammaherpesvirus infection and the host’s immune system that may lead to autoimmunity. The role of murine gamma herpesvirus 68 (γHV-68), the murine equivalent to EBV, was analyzed in the experimental autoimmune encephalomyelitis (EAE) model, an experimentally induced model to study MS, and in the New Zealand Black and White (NZBW) model, a spontaneous SLE mouse model. Mice latently infected with γHV-68 developed more severe EAE that mirrored human MS more closely than EAE in uninfected mice. γHV-68 EAE mice developed lesions composed of CD4 and CD8 T cells, loss of myelin in the brain parenchyma and spinal cord. Further, T cells from the CNS of γHV-68 EAE mice were primarily Th1, producing heightened levels of IFN-γ and T-bet accompanied by IL-17 suppression and decreased regulatory T cells frequencies. γHV-68 NZBW mice exhibited a similar Th1 skewed response and they produced different types of autoantibodies, if compared to uninfected NZBW mice. Clearly, gammaherpesvirus latency polarizes the adaptive immune response in both mouse models, directs a heightened brain pathology following EAE induction reminiscent of human MS and portrays a novel mechanism by which EBV likely influences MS and other autoimmune diseases.

View record

Myocarditis-induced dilated cardiomyopathy is a major cause of heart disease and sudden death in young adults. Development of myocarditis is thought to involve both genetic and environmental factors such as pathogen infection, with the most commonly associated pathogen being the enterovirus coxsackievirus. Herein is a summary of three research projects aimed at identifying genes associated with coxsackievirus-induced myocarditis. In the first project, the role of IL-6 in coxsackievirus-induced myocarditis was investigated. IL-6 was found to have a protective role in disease development as IL-6 deficient mice developed increased chronic disease pathology following viral infection. Recombinant IL-6 treatment in these mice decreased the disease severity, suggesting that IL-6 production during the initiation of the disease regulates myocarditis severity. The second research project was aimed at identifying genetic loci that confer susceptibility to coxsackievirus-induced myocarditis. Using chromosome substitution mouse strains (CSS) and congenic mice generated from the CSS mice, three loci on chromosome 17 were shown to confer susceptibility to chronic myocarditis. Two of the loci, Vam1 and Vam2, do not contain any genes previously associated with myocarditis development. Real-time PCR analysis identified Igf2r and Cacna1h to be strong candidates for the susceptibility genes in the Vam1 and Vam2 loci, respectively. In the third research project, the immune response following coxsackievirus infection was monitored in four inbred mouse strains in order to identify immune factors involved in disease development. Three of the strains, A/J, NOD and BALB/c, were susceptible to disease while the fourth strain, C57BL/6, was resistant to disease. Two interesting responses were observed. The first was increased TNFα levels in the resistant mice and the second was increased PD-L1 expression in the susceptible mice. Administered recombinant TNFα to the susceptible A/J mice was sufficient to increase their survival, indicating a disease protective role for TNFα in virus-induced myocarditis. All together, these data indicate that a network of many genes, both immune and non-immune, is involved in myocarditis development following virus infection. Identifying and elucidating the role of these genes in myocarditis induction could suggest ways to limit disease progression by developing therapies to target these disease modifiers.

View record

Susceptibility to autoimmune diseases is dictated by the interplay of geneticdeterminants and environmental factors including diet, toxins and infections. Viral infectionshave long been suspected to play a role in the etiology of several autoimmune disorders. Inparticular, coxsackieviruses are common human pathogens that have been linked toautoimmune myocarditis and type 1 diabetes (T1D). Evidence suggests that interactionsbetween a pathogen and components of the innate immune system may influence thegeneration of a dysregulated adaptive response ultimately resulting in autoimmune diseasedevelopment. Early recognition of viral infection is mediated by pattern recognition receptors(PRRs) expressed by a variety of cells including antigen presenting cells (APCs). PRR mediatedrecognition of an invading pathogen results in wide ranging functionalconsequences that serve to trigger innate antiviral mechanisms as well as the maturation ofAPCs and the activation of adaptive immune responses. As such, innate interactions betweenviruses and APCs likely represent a potential risk factor for the development ofautoimmunity following infection.Here, I demonstrate that early protection from coxsackievirus infection is criticallydependent on Toll-like receptor (TLR) 3 signaling on CD11b+CD11c- APCs. Interestingly,my work demonstrates that this same subset of APCs is central to the acceleration of T1Dand that manipulation of the maturation and inflammatory status of CD11b+CD11c- APCs issufficient to protect from coxsackievirus-induced autoimmune myocarditis and T1D.Protection from T1D is dependent on the reduction of costimulatory molecule expression,particularly CD40, on the surface of CD11b+CD11c- APCs which in turn increases thecapacity of these APCs to induce protective regulatory T cells (Tregs) in the pancreas.Protection from autoimmune myocarditis is not dependent on Tregs and can be circumventedby activation of the TLR4 signaling pathway.Taken together, this work illustrates an important role for a particular subset of APCsthat is critical for both early protection of the host as well as the induction of autoimmunityfollowing infection with coxsackieviruses. This strongly suggests that CD11b+CD11c- APCsrepresent a potential therapeutic target for the prevention of viral-induced autoimmunity.

View record