Cytomegalovirus (CMV) is a ubiquitous infection, infecting approximately 70% of adults in the US. CMV infects in childhood and establishes lifelong latent infection, usually without symptoms. Despite its limited morbidity, immunity to CMV occupies a surprisingly high proportion of the immune system throughout life: around 4% of CD4+ and 10% of CD8+ T cells in chronically infected people are specific for CMV, probably more than for any other virus.

This large number of T cells is thought to be maintained by recurrent virus activity, which nevertheless is exceedingly difficult to detect. However, if immunity is compromised, CMV rapidly reactivates and is a major cause of morbidity in immunocompromized individuals. It thus appears that the chronic equilibrium between CMV and its host involves a lot of activity from both sides. The long-term goal of my lab is to understand this virus-host equilibrium, and to explore the possible use of this powerful immune response in cancer immunotherapy.

Understanding how memory CD8 T cell inflation occurs after chronic herpesvirus infection

A model diagram showing the source and fate of inflationary CD8 T cells.

The CD8 T cell response to many infections follows a predictable pattern of rapid expansion, rapid contraction, then stable long-term memory. In contrast, murine CMV (MCMV) drives memory ‘inflation,’ where the CD8 T cell response increases over time. These CD8 T cells have a highly-differentiated phenotype (CD62Llo, CD127-, KLRG1+), similar to human CMV-specific CD8 T cells. We are using a combination of mouse and viral genetics to explore which infected cells stimulate these CD8 T cells during latency. Specifically, we use an MCMV expressing the CD8 T cell epitope SIINFEKL, behind a lox-stop-lox sequence, together with mice that express cre recombinase only in a specific tissue. These experiments are allowing us to define the host cell type(s) capable of driving CD8 T cell inflation.

Analysis of fibroblast-tropic HCMV for evidence of unconventional CD8 T cell responses

Groundbreaking work at the OHSU Vaccine and Gene Therapy Institute has shown that rhesus CMV (RhCMV), expressing Simian Immunodeficiency Virus (SIV) antigens, primes CD8 T cells capable of clearing highly pathogenic SIV infection. These CD8 T cells break multiple immunological paradigms. For example, many of these CD8 T cells recognize peptide presented by MHC II, rather than MHC I, while the remainder recognize peptides presented by the non-classical MHC molecule HLA-E, with virtually no CD8 T cells responding to classical MHC I. We are evaluating whether similar phenomena occur in human subjects vaccinated with a human CMV vaccine. 

Recombinant CMV as a cancer vaccine

A multiplexed immunohistochemistry section of a tumor, with T cell subsets identified by CD3, CD4 and CD8 staining.

Murine CMV, like human CMV, elicits robust T cell and antibody responses. We previously found that MCMV recombinantly expressing Trp2, a tumor-associated antigen from melanoma, causes rejection of highly immunosuppressive B16F10 melanoma. We are extending this work into mammary tumor models, using recombinant MCMV expressing the proto-oncogene Neu (a.k.a. HER2 in humans). We are testing MCMV-Neu for efficacy in a transplantable tumor model, and also in a Genetically Engineered Mouse (GEM) model of mammary tumors through a collaboration with Rosie Sears’ lab. Ongoing work is focused on determining how these vaccines work, including the role of both T cells and antibodies.

Evaluation of pMHCI-IgGs as a novel immunotherapy and as a research tool

A model of how bifunctional pMHCI-IgG molecules bind specifically to tumor cells and target them for CD8 T cell killing.

Through a collaboration with Roche Pharmaceuticals, we are performing pre-clinical testing of a novel immunotherapy fusion protein for efficacy against tumors in mice. The fusion protein (pMHCI-IgG) consists of a monoclonal antibody (mAb), specific for a tumor antigen, genetically attached to an MHC I molecule containing a CD8 T cell epitope from MCMV (see Figure). Binding of the mAb to a tumor cell leads to activation of virus-specific CD8 T cells adjacent to the tumor cell, turning fully functional bystander CMV-specific CD8 T cells into tumor-specific CD8 T cells. The utility of these pMHCI-IgG fusion proteins is being explored in models of melanoma, colon cancer, lymphoma and leukemia.

In addition, we are using these pMHCI-IgGs as a research tool. It is widely recognized that the tumor microenvironment (TME) is very hostile to CD8 T cells, leading to exhaustion and other forms of dysfunction. The pMHCI-IgGs allow us to follow the fate of functional CD8 T cells activated within the TME in an effort to identify novel immunotherapy targets.