Recombinant human IL-2 (rhIL-2) is now rarely used to treat patients with cancer because it too often causes severe toxicities. In this issue, Nirschl and colleagues report the development and preclinical characterization of an engineered IL-2 prodrug called WTX-124 that activates in the tumor microenvironment and has minimal systemic toxicity. It will be intriguing to watch the translation of this approach to the clinic.
See related article by Nirschl et al., p. 581 (5).
The remarkable anticancer activity of recombinant human IL-2 (rhIL-2; ref. 1) is often outweighed by its highly toxic side effects, leading to its discontinued use in the clinic. Advances in cell and structural biology have identified two main reasons for rhIL-2 toxicity. One is its preferential activation of off-target CD25+ cells, and the other is its short blood half-life. Such findings led to a resurgence of interest in engineering next-generation IL-2 cancer immunotherapies that do not interact with CD25, or at least have significantly reduced interaction, and have improved blood half-life (2). The available data for such efforts indicate that reducing or abrogating the interaction with CD25 has an overall beneficial effect. Nevertheless, it also suggests that systemic toxicity might still occur at high(er) doses, which can be problematic because the prevailing notion is that the primary factor limiting the therapeutic efficacy of IL-2–like molecules is their MTD.
Further improving the therapeutic index of IL-2–like molecules will require new, more sophisticated approaches, such as localization of the drug into the tumor microenvironment (3). However, targeting IL-2 by simple fusion to binding domains is challenging because unintended off-target localization easily happens despite the strength/affinity of the fused targeting domain used. In the end, it is just a matter of an overwhelming quantity of cells in the body containing IL-2 receptors, compared with the relatively tiny amounts of genuine sites available for binding by the targeting domain (4).
In this issue, to achieve effective IL-2 targeting, Nirschl and colleagues go one step beyond simple IL-2 localization (5). Using a multipronged protein-engineering approach, they have built an IL-2 prodrug, INDUKINE WTX-124, that harmlessly/inactively diffuses in blood for extended periods but becomes activated once in the tumor microenvironment. To achieve this feat, the authors combined natural human IL-2 with three protein engineering elements: first, they introduced an IL-2 (steric) inactivation domain near the binding interface to block the interaction of WTX-124 with the IL-2 receptor(s); second, they added a fusion with an human serum albumin (HSA)-binding domain to extend the blood half-life of the inactive molecule; and third, they introduced an efficient tumor-associated protease-cleavable linker that holds the inhibitory and life-extension elements in place. Although none of the components used to create WTX-124 is new, the novelty of WTX-124 is that the authors have successfully combined and optimized the three elements to maximize their beneficial effects at once. Thus, when WTX-124 finally encounters cells secreting tumor-associated proteases, the “linker” element becomes proteolyzed, unleashing two synergistic effects: first, on-the-spot activation of the IL-2 domain, and second, the (now) activated IL-2 molecule loses its half-life extension property, therefore limiting its potential systemic toxicity.
Altogether, the clever approach of Nirschl and colleagues restricts the amount of active IL-2 available systemically while allowing for a high local concentration of IL-2 in the tumor, so long as the required tumor-associated proteases are present. The experimental data presented for WTX-124, including multiple cell lines and preclinical mice cancer models, support the success of their approach. However, it is to be seen whether it will translate into in-human clinical trials. Despite this, it seems likely that this elegant protein-engineering approach could, in principle, extrapolate to other promising protein drugs riddled by inherent systemic toxicities.
D.-A. Silva reports personal fees from Neoleukin Therapeutics within the last 18 months; in addition, D.-A. Silva has a patent 20220056095 issued and licensed, a patent 20220017588 issued and licensed, a patent 11117944 issued and licensed, a patent 10844105 issued and licensed, a patent 20200347109 issued and licensed, a patent 20200339648 issued and licensed, a patent 20200317744 issued and licensed, a patent 10766929 issued and licensed, a patent 10703791 issued and licensed, and a patent 20200002398 issued and licensed. No other disclosures were reported.