The tobacco variant Nicotiana benthamiana has recently emerged as a versatile host for the manufacturing of protein therapeutics, but the fidelity of many recombinant proteins generated in this system is compromised by inadvertent proteolysis. Previous studies have revealed that the anti‐HIV‐1 antibodies 2F5 and PG9 as well as the protease inhibitor α1‐antitrypsin (A1AT) are particularly susceptible to N. benthamiana proteases. Here, we identify two subtilisin‐like serine proteases (NbSBT1 and NbSBT2) whose combined action is sufficient to account for all major cleavage events observed upon expression of 2F5, PG9 and A1AT in N. benthamiana. We propose that downregulation of NbSBT1 and NbSBT2 activities could constitute a powerful means to optimize the performance of this promising platform for the production of biopharmaceuticals. Databases NbSBT sequence data are available in the DDBJ/EMBL/GenBank databases under the accession numbers MN534996 to MN535005.

Broadly neutralizing antibodies (bNAbs) to human immunodeficiency virus type 1 (HIV‐1) hold great promise for immunoprophylaxis and the suppression of viremia in HIV‐positive individuals. Several studies have demonstrated that plants as Nicotiana benthamiana are suitable hosts for the generation of protective anti‐HIV‐1 antibodies. However, the production of the anti‐HIV‐1 bNAbs 2F5 and PG9 in N. benthamiana is associated with their processing by apoplastic proteases in the complementarity‐determining‐region (CDR) H3 loops of the heavy chains. Here, it is shown that apoplastic proteases can also cleave the CDR H3 loop of the bNAb 2G12 when the unusual domain exchange between its heavy chains is prevented by the replacement of Ile19 with Arg. It is demonstrated that CDR H3 proteolysis leads to a strong reduction of the antigen‐binding potencies of 2F5, PG9, and 2G12‐I19R. Inhibitor profiling experiments indicate that different subtilisin‐like serine proteases account for bNAb fragmentation in the apoplast. Differential scanning calorimetry experiments corroborate that the antigen‐binding domains of wild‐type 2G12 and 4E10 are more compact than those of proteolysis‐sensitive antibodies, thus shielding their CDR H3 regions from proteolytic attack. This suggests that the extent of proteolytic inactivation of bNAbs in plants is primarily dictated by the steric accessibility of their CDR H3 loops.

Background: Macrophage Migration Inhibitory Factor (MIF) has been described as a pleiotropic cytokine known to exacerbate tumor growth (Chesney and Mitchell, 2015). Due to its ubiquitous nature, MIF can be considered as an inappropriate target for therapeutic intervention. However, we discovered oxidized MIF (oxMIF) as the disease-related isoform of MIF, which is specifically present in tumor tissue (Schinagl A et al., 2016). A 1st generation anti-oxMIF monoclonal antibody (mAb) demonstrated an acceptable safety profile and efficacy signal in a phase 1 clinical trial (Mahalingam D et al., 2020). Two bioengineered 2nd generation anti-oxMIF mAbs with highly improved biophysicochemical and biological properties were generated and compared to the 1st generation anti-oxMIF mAb. Both bioengineered mAbs share identical variable domains, but distinct heavy-chain constant regions, to increase effector functions and efficacy of the therapeutic mAb ON203 and to reduce interactions with FcγRs for the radio diagnostic mAb ON102. Methods: Hydrophobicity and stability were determined by HIC and SEC. ADCC and ADCP activity were investigated by reporter and PBMC-mediated cell killing assays. Tumor penetration was assessed using IRDye 800CW or Zr89-labeled mAb in tumor-bearing Balb/c or Balb/c nude mice. PK and bioavailability were assessed in Balb/c nude mice. Efficacy was determined in PC3 xenograft models in NMRI nude mice. Results: Bioengineering significantly reduced hydrophobicity of ON203 and ON102, leading to improved stability and strongly reduced aggregation but retaining the low nM affinity for oxMIF. This further resulted in a two-fold improved bioavailability and tumor accumulation after 24-48h and a three-fold enhanced retention on day 7 in mice harboring solid tumors of the colon, when compared to the 1st generation anti-oxMIF mAb. In contrast to previously described anti-oxMIF reference, ON203 and ON102 did not mediate any unspecific release of MCP-1, IL-6, or TNF-α from PBMCs. Reporter and PBMC-mediated cell killing assays proved ADCC activity of ON203 with EC50 values of 0.1-0.8 nM and ADCP activity with EC50 values of 2.3 nM, whereas the 1st generation anti-oxMIF mAb was at least 10-fold less potent. Our findings for ON203 translated into abolishment of tumor growth in NMRI nude mice harboring human PC3 prostate cancer cells, showing superior efficacy compared to the 1st generation anti-oxMIF mAb. Conclusion: ON203 has a high potential to significantly improve efficacy compared to the 1st generation anti-oxMIF mAb, with Zr89-ON102 as companion diagnostic for patient stratification. We aim to develop these anti-oxMIF mAbs together for clinical use to create new treatment options for patients with solid tumors, with a clear rationale to combine ON203 with other immunotherapies or checkpoint inhibitors (Noe and Mitchell, 2020). Citation Format: Alexander Schinagl, Michael Thiele, Irina Mirkina, Gregor Rossmueller, Alejandro A. Puchol Tarazona, Randolf J. Kerschbaumer. Novel bioengineered monoclonal antibodies targeting oxidized macrophage migration inhibitory factor as anti-cancer therapeutics and companion diagnostics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 313.