Background
In the development of IL-2 based immunotherapy for cancer, it was generally accepted that interaction with the IL-2Rα (CD25) should be eliminated to avoid unwanted toxicity. However, we have been developing a novel long-acting IL-2 applying the opposite strategy of incorporating CD25 engagement. Recent studies have demonstrated the critical role of CD25 binding property in terms of safety and antitumor activity, supporting our development strategy. Here, we explored how the CD25 engagement functions to mitigate systemic toxicity in IL-2 based immunotherapies.
Methods
HM16390, which strongly binds to CD122 and possesses optimal CD25 binding characteristics, and its non-alpha variant were administered subcutaneously (SC) in mice. After weekly doses of them, survival rate, immune cell profile, and toxicity were evaluated. Cynomolgus monkeys were given SC injections of both molecules once every three weeks in a stepwise dose-escalation regimen, and toxicological finding and peripheral immune cell profiles were monitored.
Results
SC administration of HM16390 was generally tolerated in mice across various dose ranges. However, non-alpha variant showed significant toxicity, including body weight loss, increased lung weight, elevated ALT and AST levels, subsequent mortality at a dose 12 times lower than HM16390, despite similar CD8+ T cells expansion in lymphoid organ. In B16F10 mice, while maintaining a significantly increased CD8+ T/Treg ratio, peripheral Tregs also increased in line with the elevation of CD8+ T cells, suggesting a potential peripheral role for Tregs. Consistent with the results in mouse models, HM16390 was well tolerated in monkeys, whereas non-alpha variant showed signs of intolerance during the dose escalation phase. HM16390 did not induce any severe toxicity reactions up to the highest dose. In contrast, clinical observations pointed to more pronounced adverse effects with non-alpha variant, and found dead at the highest dose with severe weight loss (-21% vs baseline). The peripheral immune cell profiles in monkey also highlighted that HM16390 slowly and steadily increased effector cells in a dose-dependent manner through the beneficial buffering effect of optimal Tregs expansion. In contrast, non-alpha variant induced a steep and uncontrolled increase in peripheral effector cells in the absence of Tregs expansion, ultimately leading to severe toxicity and mortality.
Conclusions
The CD25 binding characteristics within HM16390 was finely tuned to mitigate unwanted toxicity derived from uncontrolled immune cell expansion. The crucial role of CD25 in terms of safety has been demonstrated in rodents and non-human primates, and these findings support HM16390 as a safe and effective immuno-oncology agent.