TCR-mimic bispecific antibodies to target the HIV-1 reservoir

HIV-1 infection is incurable due to the persistence of the virus in a latent reservoir of resting memory CD4+ T cells. “Shock-and-kill” approaches that seek to induce HIV-1 gene expression, protein production, and subsequent targeting by the host immune system have been unsuccessful due to a lack of effective latency-reversing agents (LRAs) and kill strategies. In an effort to develop reagents that could be used to promote killing of infected cells, we constructed T cell receptor (TCR)-mimic antibodies to HIV-1 peptide-major histocompatibility complexes (pMHC). Using phage display, we panned for phages expressing antibody-like variable sequences that bound HIV-1 pMHC generated using the common HLA-A*02:01 allele. We targeted three epitopes in Gag and reverse transcriptase identified and quantified via Poisson detection mass spectrometry from cells infected in vitro with a pseudotyped HIV-1 reporter virus (NL4.3 dEnv). Sequences isolated from phages that bound these pMHC were cloned into a single-chain diabody backbone (scDb) sequence, such that one fragment is specific for an HIV-1 pMHC and the other fragment binds to CD3ε, an essential signal transduction subunit of the TCR. Thus, these antibodies utilize the sensitivity of T cell signaling as readouts for antigen processing and as agents to promote killing of infected cells. Notably, these scDbs are exquisitely sensitive and specific for the peptide portion of the pMHC. Most importantly, one scDb caused killing of infected cells presenting a naturally processed target pMHC. This work lays the foundation for a novel therapeutic killing strategy toward elimination of the HIV-1 reservoir.

Use of TCR-mimic bispecific antibodies against pMHC-II for monitoring antigen processing and redirecting cytolytic effector cells

HIV-1 infects CD4+ T cells and macrophages and is currently incurable due to a latent reservoir. Targeting infected cells using TCR-mimic antibodies to peptide-MHC (pMHC) would provide a specific, high-affinity platform for immune system activation and monitoring antigen processing. However, antibodies to pMHC are notoriously difficult to isolate. Using a phage display platform, we panned for phage expressing variable fragments against immunodominant HIV pMHC complexes. These Fab fragments were cloned into a bispecific antibody (bsAb) backbone, such that one Fab fragment is specific for an HIV pMHC and the other fragment binds to CD3. These antibodies utilize the sensitivity of T cell signaling as readouts for antigen processing and potentially as therapeutic formats. Using this approach, we have shown that bsAbs specific for the most conserved MHC-II HIV capsid epitope (Gag293) can recognize the processed epitope on human monocyte-derived dendritic cells (moDCs) fed with the whole Gag protein antigen. Recognition is exquisitely epitope-specific. Additionally, human monocyte-derived macrophages infected with HIV were able to present Gag293 using the endogenous MHC-II pathway, as read out by a Gag-specific bsAb. To our knowledge, this is the first description of endogenous class II presentation of HIV in a major cell type that supports viral replication. By developing bsAbs to detect HIV peptide presentation, we can begin to address the kinetics, pathways, and cell types contributing to the priming events in early HIV infection, which will inform improved vaccine therapies. Additionally, given the potential for therapeutic application, these bsAbs may be used to redirect cytolytic effector cells towards infected targets.

Characterizing the MHC-II immunopeptidome of HIV using a cell-free antigen processing system and peptide:MHC-specific antibodies

The breadth and specificity of CD4+T cell responses in HIV-1 infection are critical determinants of viral load and an individual’s ability to control infection. Most studies of CD4+T cell responses against HIV have relied on indirect evidence from peptide-pulsing analyses. Direct studies of antigen presentation have been hampered by the technical difficulties inherent in isolating peptide:MHC (p:MHC) complexes from HIV-infected cells. This has made it challenging to understand the HIV immunopeptidome, particularly for MHC-II epitopes. Using a cell-free antigen processing system, we have generated a near complete map of potential DR1-restricted HIV-1 epitopes. To confirm that epitopes identified using this system are actually presented by HIV-infected cells, we generated novel reagents termed single-chain diabodies (scDbs), or bispecific antibodies, which contain one Fab fragment against a p:MHC and another against CD3. These antibodies enable the use of CD8+ or CD4+T cells as readouts for antigen presentation. In proof of concept experiments, scDbs directed against immunodominant HIV MHC-I epitopes enhance cytokine production by CD8+ T cells, confirming presentation of these peptides. Preliminary experiments showed similar results with scDbs targeting MHC-II Gag epitopes on dendritic cells overexpressing the relevant peptides. We are now profiling the use of scDbs to detect antigen presentation in HIV-infected macrophages and CD4+T cells, the two target cells of infection. By establishing an immunopeptide map and tools to detect HIV peptide presentation, we can begin to address the kinetics, pathways, and cell types contributing to HIV antigen presentation on MHC-II, which will inform improved vaccine therapies.