Development of monoclonal antibodies (mAbs) for oncology is increasing. An IQ DruSafe Working Group conducted an industry survey to evaluate non-human primate (NHP) use in nonclinical toxicity testing of oncology mAbs for well-characterized targets (WchT) with the aim to identify opportunities to reduce NHP use. The survey addressed sources of information used to justify WchT, weight of evidence (WoE) approaches to reduce NHP use, and design of nonclinical toxicology programs. All respondents used literature to define WchT. WoE approaches helped reduce the number and size of general toxicology studies. Most companies conducted non-Good Laboratory Practice (non-GLP) dose-range finding studies prior to GLP toxicology studies, often non-terminally, allowing for potential reuse of NHPs. For GLP studies, most companies maintained a standard design for 1-month studies (when conducted) but were more flexible with 3-month studies, often excluding recovery groups. Case studies illustrate successful regulatory acceptance of streamlined nonclinical safety packages. Engaging drug regulatory authorities (DRA) to discuss the need for additional and/or specialized studies would be beneficial to reduce NHP use. In conclusion, NHP use can be reduced in developing oncology therapeutics against WchT by optimizing nonclinical toxicology approaches with appropriate strategic planning, fit-for-purpose toxicology study designs, and discussion with DRA.

Table S5 shows normalized intensity of GSK3b, CDK4 and CDK6 in bone marrow and intestine of control rats.

Figure S2 shows A) significantly perturbed genes, B) Venn diagram of changed genes, C) hierarchical clustering of changed genes, D) volcano plot of all genes.

Supplementary Figure 5 shows the analysis of sasanlimab by SE-HPLC with UV detection at 280 nm

Supplementary Figure 11 shows sasanlimab binding to human and cynomolgus monkey Fcγ receptor and FcRn

Supplementary Figure 8 shows the solution binding affinity of sasanlimab to human and cynomolgus monkey PD-1 using the KinExA

Figure S3 shows Axin2 expression normalized to vehicle.

The first-in-patient (FIP) starting dose for oncology agents should be reasonably safe and provide potential therapeutic benefit to the patient. For late-stage oncology patients, this dose is often based on the ICH S9 guidance, which was developed primarily based on experience with cytotoxic chemotherapeutic agents using the rodent STD10 or non-rodent HNSTD and an appropriate safety factor. With the increase in molecularly targeted chemotherapeutics, it is prudent to re-evaluate how the FIP dose is derived to ensure that the appropriate balance between risk and therapeutic benefit to the patient is achieved. Blinded data on 92 small molecule oncology compounds from 12 pharmaceutical companies who are members of the IQ DruSafe consortium were gathered to investigate if a NOAEL-based starting dose without a safety factor would have been tolerated in the FIP trial and if so, estimating how many dose escalation cohorts could have been reduced. Our analysis suggests that the NOAEL-based alternative starting dose would have been tolerated in most cases evaluated, with an anticipated mean reduction of 2.3 cohorts. Of the 12 cases where the alternative approach resulted in a starting dose that would have exceeded the MTD/RP2D, none of the nonclinical toxicities in these cases were considered irreversible and would be monitorable in all but one instance. Most non-tolerated cases were within two–threefold of the MTD/RP2D, with the clinical AEs considered manageable and mitigated by dose de-escalation. No one method of FIP dose calculation will likely be appropriate for all oncology small molecules and starting dose selection should be performed using a case-by-case approach. However, the NOAEL-based method that does not utilize a safety factor should be considered when appropriate to minimize the number of patients exposed to sub-therapeutic doses of an investigational oncology agent and accelerating development to RP2D.

Supplementary Figure 15 shows no body weight changes in response to sasanlimab treatment in mice.

Supplementary Figure 12 shows the ELISA assay with the dose-dependent C1q binding to plate-adsorbed sasanlimab and the ADCC assay of sasanlimab performed with activated T cells (target cells) and stimulated PBMCs (effector cells).

Table S1 shows plasma exposure in humans or rats.

Supplementary Figure 10 shows the selectivity of sasanlimab binding to human and cynomolgus monkey PD-1 using HEK-293T transiently transfected cells and flow cytometry

Supplementary Figure 4 shows the evaluation of the charge isoform distribution of sasanlimab by imaged capillary isoelectric focusing (iCE).

Supplementary Figure 14 shows that sasanlimab blocks the binding of human PD-1 to its ligands, PD-L1 and PD-L2, by SPR analysis.

Table S2 shows IC50 values for CDK4/6 inhibitors in human and rat bone marrow.

Supplementary Figure 6 shows the rCGE profile for sasanlimab with two predominant peaks, consistent with L chain and H chain

In the fight against cancer, immunotherapies are one of the largest growing therapeutics in development. Immunotherapeutics are designed to boost or harness the power of the immune system to prevent, control, or eliminate cancer; while many immune therapies have been found to be safe others have induced severe toxicities. For instance, even compounds that target the same molecule/antigen can have dramatically differing safety profiles. Current preclinical models evaluating these therapies are underequipped to assess the safety of these compounds: in vitro assays fail to predict systemic responses and traditional animal models often fail to correlate with human responses. To better meet the needs of assessing preclinical toxicity we developed a PBMC-humanized mouse model to test a variety of therapeutics, including both monoclonal and bispecific antibodies and induce human cytokine release responses which can manifest within hours or days later resulting in tissue damage and lethality of the mice. To date we have tested a variety of therapeutics, including Blinatumomab, Rituximab, EGFRxCD3 BiTE, CAR-T, and others in our platform while evaluating the ability of the therapeutic to induce human cytokines, bodyweight loss, clinical symptom assessment, and survival in the context of toxicity alone or along with the evaluation with efficacy. We found that many of the therapeutics tested in our platform showed similarities to clinical data in humans. For example, urelumab and utomilumab are both fully humanized monoclonal antibodies against 4-1BB (CD137). However, during clinical trials, urelumab was shown to induce severe liver toxicities while utomilumab was well tolerated. In our huPBMC mouse model, we likewise showed that huPBMC mice dosed with 10 mpk of urelumab experienced body weight loss, showed liver necrosis, and met the clinical criteria for early euthanasia compared to mice treated with 10 mpk utomilumab and PBS treated controls. Serum levels of enzymes associated with liver damage: AST, ALT and GLDH were significantly higher in urelumab treated mice and terminal serum cytokine analysis revealed similarities with those found to be increased in urelumab clinical trials, including elevated IFNγ, IP-10, MIG, and MIP-1α and MIP-1β. Further, HuPBMC mice are also capable of detecting variability among donors. We have screened well over 60 human PBMC donors in huPBMC mice treated with OKT3 and αCD28 and while we always see an increase in cytokines such as IFNγ - the range of induction varies greatly among donors. Further, we see PBMC-donor variability in body weight loss and survival rate after OKT3 and αCD28 treatments. We demonstrate that the PBMC humanized mouse model shows clinical relevance. The use of these models for preclinical safety assessments has the potential to become an important part of novel immunotherapeutic development for patient safety and reducing drug development costs. Citation Format: Destanie Rose, Won Lee, Guoxiang Yang, Jiwon Yang, Mingshan Cheng, Wenqian He, Bernard Buetow, Allison Vitsky, Maggie Liu, Bart Jessen, James Keck. The use of PBMC humanized mice to test the efficacy and safety of antibody and cell-based cancer immunotherapeutics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB349.

Supplementary Figure 7 shows the binding of sasanlimab to PD-1 by ELISA

Table S3 shows A: Most significantly impacted pathways. B: Most significantly impacted upstream regulators.