The Role Of Antibody in Enabling Cell-mediated Control of HCMV Infection

Human cytomegalovirus (HCMV) is a serious clinical problem. If a mother catches it during pregnancy and it passes to the foetus, the foetus may die, or it may suffer from lifelong disabilities including deafness, blindness or intellectual disability. In the UK over 5,000 babies contract CMV every year, while in America, over 44,000 babies will catch it. It also causes major problems in patients with weakened immune systems, such as transplant recipients or HIV sufferers. The virus is a major cause of suffering, and places a high financial burden on the health system; caring for a single CMV infected baby costs >$330,000 per year. Current treatment are antivirals, however these can be very toxic, and the virus can rapidly become resistant to the drugs.

For many viruses, and in the study of cancer, great success has been achieved by giving patients antibodies that can bind to cells, and activate immune cells to kill infected or cancerous cells. HCMV antibodies can control infection in a similar way, yet we have very little understanding of which HCMV proteins these antibodies are binding, or which immune cells are responsible for killing the infected cells. Answering these questions is important to generating optimal vaccines, novel therapies, and for understanding how the immune system controls disease in patients.

We have identified all the HCMV proteins that are present on the surface of infected cells, and shown that antibodies can bind to these proteins, enabling them to be 'seen' by immune cells. We will now exploit these observations to determine which virus proteins are best at activating immune cells, and which immune cells are best at controlling the infection. We will also investigate the ways that the virus might try and prevent antibodies from binding to infected cells, so that we can circumvent these 'evasion' mechanisms.

Taken together, this data will enable us to design effective antibodies that are optimised for controlling HCMV disease in human patients, and will enable the rational design of more effective vaccine strategies.

HCMV treatment relies on antivirals, yet these have significant drawbacks including toxicity and resistance. Many clinical efforts are focussed on the development of neutralising antibodies. However, although antibody titres correlate with control of disease, their neutralising capacity is not critical; in a trial of neutralising human IgG, no reduction in disease was observed. Consistent with this observation, we have shown that dissemination by clinical HCMV is resistant to neutralising antibodies.

In contrast to results with neutralising antibodies, in two recent vaccine trials, 50% protection was achieved when the induced antibodies acted through non-neutralising mechanisms. Non-neutralising mechanisms including antibody-dependent cellular cytotoxicity (ADCC), complement dependent cytolysis (CDC) and antibody-dependent cellular phagocytosis (ADCP). Both in vivo and in vitro findings strongly suggest that ADCC is critical for virus control. We have examined the cell surface proteome to define, for the first time, the viral antigens that activate ADCC rapidly after infection. We will determine the relative effectiveness of each viral antigen at activating ADCC in isolation, in the context of the wildtype virus and across a range of clinically-relevant cell types. We will then investigate the relative ability of these viral antigens to activate each non-neutralising effector mechanism (ADCC, CDC, ADCP). On the basis of these findings, humanised monoclonal antibodies will be generated to the strongest candidates, and the relative ability of ADCC, ADCP and CDC to control infection assessed.

Finally, we will enable optimisation of therapeutic antibody efficacy by assessing the role of HCMV-encoded Fc-binding proteins in sequestering therapeutic antibodies, before crystallising the viral Fc-binding proteins in complex with Fc. This will guide the development of optimised antibodies that activate human, but not viral, Fc-receptors.

Knowledge of which viral antigens are optimal at activating ADCC, ADCP, and CDC, and how efficiently targeting these antigens enables the virus to be controlled, is crucial to the design of future vaccines. Similarly, a comprehensive comparison of which antigens enable which effector mechanisms to be activated, and which effector mechanisms are the most effective at controlling infection, is also a critical parameter in vaccine design.

These parameters are also crucial to the design of monoclonal antibodies that could be infused directly therapeutically to kill infected cells within an individual, and we have submitted a patent covering this use. Not only would these antibodies be targeted against an optimal antigen, and activating an optimal cellular immune response, but our work would also enable them to activate these effector mechanisms more strongly than 'natural' antibodies, and to avoid being bound by viral Fc binding proteins. As a result they would have the potential to be much more effective than any other current antibody-based therapeutic; furthermore, the lead antigens we have identified are different from those being pursued as neutralising antibodies. Although neutralising antibody targets have the potential to be present on the cell surface, they are not present until much later in infection, and are thus significantly less likely to be effective at controlling HCMV through activating cell-mediated immunity.

The proposed antibodies would offer a number of advantages over existing therapies: they would be a defined, antigen specific product. They can be grown to almost limitless amounts, with minimal lot-to-lot variation. Monoclonal antibody products often provide a rapid route to a clinical proof of concept and have high success rate in regulatory approval in comparison to small molecule drugs. Reasons include clear understanding of their mechanism of action, efficient production platforms, good toleration and high specificity. The risk of unexpected safety issues in human clinical trials of humanised antibody products is lower than with many other types of therapeutic products, thus they are unlikely to suffer from the issues of toxicity that are cause-for-concern with antivirals.