Antibodies target intracellular oncodrivers through PIGR-mediated transcytosis

MAIN TEXT

Monoclonal antibodies have revolutionized the landscape of cancer and autoimmunity patient care and are at the forefront of contemporary drug development. However, antibodies have been traditionally considered ineffective for targeting intracellular antigens, due to their large size, which hinders effective penetration into the cytoplasm. However, it has been known for a long time that IgG crosses the placenta through binding to FcRn and endosomal intracellular trafficking, a phenomenon known as transcytosis. In addition, dimeric IgA and pentameric IgM also undergo directional transcytosis through mucosal cells in a Polymeric Immunoglobulin Receptor (PIGR)-dependent manner. This intricate process not only enables the transport of these immunoglobulins to the luminal side of mucosal locations, but neutralizes viruses that have already infected epithelial cells¹.

We previously demonstrated the quasi-universal expression of PIGR in human ovarian² and endometrial cancer cells³, which drives bona fide transcytosis through cancer cells of dimeric IgA (dIgA), which heavily accumulates in the microenvironment of these tumors. In a recent publication in Immunity⁴, we demonstrated that non-small cell lung cancer (both adenocarcinomas and squamous cell carcinomas) also express PIGR and abundant IgA at the protein level, whereas virtually all epithelial cancers in TCGA datasets appear to express PIGR mRNA. We therefore surmised that, similar to intracellular viruses, transcytosing dIgA, engineered to specifically target cytosolic proteins, could neutralize commonly mutated oncodrivers, as well as other intracellular proteins. In support of this proposition, we demonstrated that dimeric IgA specific for KRAS^G12D mutations, but not an IgG4 using the same VH/VL sequences, neutralizes mutant (but not wild-type) KRAS inside the cytosol of cancer cells. Furthermore, dIgA, through directional transcytosis in early and late endosomes, transports the mutated oncogene outside the tumor cell, where it is expelled (Figure 1).

Figure 1. Antigen-specific dIgA and intracellular target interact in early and late endosomes inside tumor cells.

Our results demonstrate that KRAS^G12D-specific dIgA changes the traffic of KRAS^G12D from recycling endosomes to the same early and late endosomes in which dIgA transcytoses. A mechanism of heterotypic endosomal fusion near the cell membrane is suggested.

Neutralization of KRAS^G12D was not restricted to in vitro experiments, because KRAS^G12D-specifc dimeric IgA abrogated the progression of multiple carcinomas spontaneously or ectopically carrying the mutation in vivo, both in immunodeficient and immunocompetent mice. Equally important, transcytosing dIgA can also target cytosolic proteins that are not located close to the cell membrane, where interactions between early and recycling endosomes could be less frequent. Thus, we demonstrated that IDH1^R132H-specific dIgA can also abrogate the growth of IDH1^R132H-mutant tumors, in a mutation- and PIGR-dependent manner. Our study therefore opens new avenues for targeting virtually any cytosolic protein using antibodies that are internalized through PIGR-mediated transcytosis, including hotspots in other oncodrivers or other tumor-promoting molecules, such as the cytosolic fraction of survivin or Myc. In addition, the range of proteins that could be intracellularly neutralized could be extended to other diseases. For instance, one could envision the cGAS–STING pathway as a therapeutic target in inflammatory mucosal diseases.

Ongoing experiments and other studies in the field should clarify multiple emerging questions. For instance, how exactly dIgA interacts with cognate antigens through endosomal membranes inside the cytoplasm? We demonstrated that dIgA transcytosis abrogates the recycling of mutant KRAS, which is then found in the same early and late endosomes through which dIgA transcytoses. Therefore, it is possible that heterotypic fusion occurs between dIgA-carrying endosomes and recycling endosomes that bring-back mutant KRAS to prevent lysosomal degradation (Figure 1). Second, although we did not find nuclear localization of antigen-specific dIgA, modifications of the antibody with nuclear translocation motifs, or binding to nuclear proteins that spontaneously translocate to the nucleus, could bring dIgA in contact with nuclear proteins, where these antibodies could neutralize previously undruggable transcription factors. Third, although we did not find evidence of IgG transcytosis through epithelial cells, we found expression of FcRn, the receptor that mediates IgG transcytosis through the placenta, in many tumor and healthy epithelial cells. It is theoretically possible that IgG could transcytose through a subset of tumors, although we could not capture this in our experiments. Most importantly, what improvements could be made to enhance the effectiveness of these interventions? For instance, the half-life of IgA in primates is ~6 days. However, the half-life of therapeutic IgG antibodies is more than 20 days. Therefore, more persistent neutralization could be achieved using IgG. In addition, the yield of recombinant monomeric IgG is greater than that for dimeric IgA. Therefore, we are currently engineering IgGs to include PIGR-binding motifs, which could trigger transcytosis while maintaining the superior half-life of this isotype.

It should also be noted that PIGR-binding antibodies are perfectly compatible with emerging small molecule inhibitors. This is important because, after 40 years of relentless efforts aimed at targeting KRAS mutations, KRAS^G12C-specific small molecule inhibitors have been recently approved by the FDA. In addition, a KRAS^G12D-specific inhibitor, MRTX1133, is currently undergoing clinical testing under accelerated approval. However, mechanisms of resistance have already emerged for KRAS^G12C inhibitors⁵, and will likely occur for MRTX1133. We are currently testing how our antibody-based approaches could be effective against tumors that have become resistant to these drugs.

Furthermore, future studies should determine how intracellular oncodriver-targeting antibodies synergize with other immunotherapies. For instance, our previous studies demonstrated that dIgA transcytosis, in a non-antigen-specific, MHC-I-independent manner, sensitizes tumor cells to T cell-mediated killing². This likely has to do with the poorly understood effects of dIgA (and perhaps pentameric IgM) transcytosis through tumor cells, which we found elicits genome-wide transcriptional changes that lead to pro-apoptotic pathways (such as CHOP), as well as T cell-sensitizing NKG2D ligands (such as ULBP1)³. Therefore, it is likely that the effectiveness of PIGR-binding antibodies is enhanced by immune checkpoint inhibitors. Thus, our studies in immunocompetent KRAS^G12D tumor-bearing mice showed that therapeutic effectiveness is dependent on CD8 T cells⁴. Because these studies were conducted using human antibodies, it is possible that they were neutralized after sensitization, thus limiting their effectiveness.

In summary, our study demonstrates the feasibility of specifically targeting, inside the cytoplasm of epithelial cells, recurrently mutated hotspots within oncogenic loci that drive malignant progression across frequent and aggressive human carcinomas. This can be achieved through antigen-specific dIgA, or other antibodies engineered to trigger PIGR-mediated transcytosis across all epithelial cells. The possibility of using antibodies to effectively target intracellular molecules therefore opens many avenues for new immunotherapies, and novel experimental tools.