Science Spotlight

How do antibodies evolve to kill HIV-infected cells?

From the Overbaugh and Matsen labs, Human Biology and Public Health Sciences Division

Effective antibody responses against HIV are rare both during natural infection and after vaccination. As a result, many research groups have focused their efforts on identifying and characterizing rare but potent antibodies in order to design a vaccine that would stimulate their production. One type of potent antibodies against HIV are those that can elicit antibody-dependent cellular cytotoxicity (ADCC). These antibodies bind to pathogen-derived antigens on the cell surface, where they act as markers and bridges for immune cells that recognize them and kill the infected cell. Despite their importance, the developmental pathways that give rise to HIV-specific ADCC antibodies remain elusive.

Using longitudinal antibody sequencing data coupled with computational methods to infer antibody lineages, researchers in the Overbaugh (Human Biology) and Matsen (Public Health Sciences) laboratories at Fred Hutch reconstructed the evolution of six ADCC HIV-1-specific antibodies that arose during natural HIV infection. The study, recently published in the journal eLife, was led by postdoctoral fellow Dr. Laura Doepker and Research Technician Sonja Danon. Reflecting on the contributions of this study to HIV vaccine design, Danon said: “It was fascinating to explore the developmental pathways of ADCC-mediating antibodies, specifically when they developed binding and functional capacities, as this has not been explored yet in the context of HIV. This is important because having a foundational understanding of their development sheds light on what mutations would be beneficial to elicit in a vaccine setting.”

Correlation of  binding potency with ADCC function as antibodies mature.
Image provided by Dr. Laura Doepker.

The six monoclonal antibodies with ADCC function were previously isolated from a Kenyan woman enrolled in a 1993 open-cohort study that examined the risk factors and the natural history of HIV-1 infection. To start reconstructing the probable developmental routes that gave rise to the mature ADCC antibodies, the researchers used next-generation sequencing (NGS) to obtain antibody variable region (VR) gene sequences from longitudinal blood samples collected at various time points from the time of pre-infection to 4 years post-infection. With this data and some computational inferences for unavailable VR regions, the investigators reconstructed antibodies from two lineages at the available time points and tested their ADCC capability. By day 462 post-infection, ADCC functionality had developed in both lineages.

Next, the group sought to characterize lineage intermediates that conferred HIV specificity and ADCC function. Reconstruction of intermediates was performed using a computational methodology termed Bayesian phylogenetic lineage inference, an approach specially designed to handle the sparse data that is characteristic of antibody evolution studies. With this method, the scientists identified ancestral intermediates in chronological order with high statistical confidence. Some of the computationally inferred lineages closely matched the NGS-sequences obtained from the longitudinal samples, validating the computational lineage reconstruction methodology.

Antibodies are composed of a light chain and a heavy chain, containing VRs subdivided in the complementarity determining regions (CDRs) and framework regions (FWR). With the inferred intermediates resolved, the investigators determined per-chain mutations contributing to antibody function by building antibodies with all possible (chronologically matched) combinations of inferred heavy and light chains. They first tested middle intermediates, and if ADCC function was already present, they focused on pre-middle intermediates. Alternatively, if ADCC function was not present in the middle, they concentrated on post-middle inferred intermediates to evaluate antigen-binding affinity and ADCC function.

 Across all lineages, the acquisition of antigen binding and ADCC functionality often required mutations in the CDR. Interestingly, enhancement of ADCC potency required substitutions in the FWR. Dr. Doepker explained the implications of this finding: “The reliance of several of these antibody lineages on framework region mutations for potent ADCC activity is surprising. Antibodies do not readily mutate these regions, but, then again, our study focused on samples collected many years after HIV infection, so these antibody lineages had time under chronic infection conditions to do so.  I’m very curious to know how quickly mutations in antibody framework regions can occur under both acute and chronic infection conditions. Furthermore, no one yet knows how to alter rates of mutation in antibodies, let alone by region (such as in the framework regions). The deliberate stimulation of antibody framework region mutations is an unexplored frontier in vaccine research, but it may be an important avenue for a successful HIV vaccine.”

Danon highlighted another interesting finding of the study regarding the correlation of binding affinity and ADCC function: “Binding strength alone did not necessarily predict the potency of ADCC activity. For example, despite two antibodies in the same lineage having similar antigen binding affinities, one was able to mediate ADCC while the other was not. In another instance, two antibody lineages had drastically different binding capacities, but elicited equivalent levels of ADCC activity. This suggests that other factors affecting ADCC potency are at play, and I am very interested to understand further what these factors may be.”

Doepker, L. E., Danon, S., Harkins, E., Ralph, D. K., Yaffe, Z., Garrett, M. E., Dhar, A., Wagner, C., Stumpf, M. M., Arenz, D., Williams, J. A., Jaoko, W., Mandaliya, K., Lee, K. K., Matsen, F. A., 4th, & Overbaugh, J. M. (2021). Development of antibody-dependent cell cytotoxicity function in HIV-1 antibodies. eLife10, e63444. Advance online publication. https://doi.org/10.7554/eLife.63444

This work was supported by grants from the NIH, including training grants and faculty scholar grants, and a Faculty Scholar grant from the Howard Hughes Medical Institute.

Fred Hutch/UW Cancer Consortium members Julie Overbaugh and Erick Matsen contributed to this study.