An engineered IL-2 reprogrammed for anti-tumor therapy using a semi-synthetic organism

The implementation of applied engineering principles to create synthetic biological systems promises to revolutionize medicine, but application of fundamentally redesigned organisms has thus far not impacted practical drug development. Here we utilize an engineered microbial organism with a six-letter semi-synthetic DNA code to generate a library of site-specific, click chemistry compatible amino acid substitutions in the human cytokine IL-2. Targeted covalent modification of IL-2 variants with PEG polymers and screening identifies compounds with distinct IL-2 receptor specificities and improved pharmacological properties. One variant, termed THOR-707, selectively engages the IL-2 receptor beta/gamma complex without engagement of the IL-2 receptor alpha. In mice, administration of THOR-707 results in large-scale activation and amplification of CD8+ T cells and NK cells, without Treg expansion characteristic of IL-2. In syngeneic B16-F10 tumor-bearing mice, THOR-707 enhances drug accumulation in the tumor tissue, stimulates tumor-infiltrating CD8+ T and NK cells, and leads to a dose-dependent reduction of tumor growth. These results support further characterization of the immune modulatory, anti-tumor properties of THOR-707 and represent a fundamental advance in the application of synthetic biology to medicine, leveraging engineered semi-synthetic organisms as cellular factories to facilitate discovery and production of differentiated classes of chemically modified biologics.


Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences
Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative. Sample sizes were not statistically pre-determined. Appropriate sample sizes were selected for each experiment to allow meaningful statistical confidence where required For initial screening studies using the Discoverx assay, each compound was tested in duplicate to ensure reproducibility to identify potential hit compounds and discern from pharmacologically inappropriate variants. Duplicate measurements were considered sufficient for screening as the density of sampling across doses ensured reasonable accuracy of curve fits sufficient to identify large differences in potency. For pSTAT5 flow studies, sample size were selected to minimize natural donor variability effects and allow mean and SEM calculations for potency data. For in vivo studies, sample size was selected to balance animal use with statistical power, and allow temporal sampling density required to observed the effects in cases where terminal collections were required (PK/PD studies). For PK/PD studies, at least three animals per time point per compound were selected as appropriate to minimize animal to animal variability and statistical measure of the standard error of the mean effect. For tumor studies, sample sizes were selected to balance animal use and statistical power, and to minimize animal to animal variability effects on the results.
In the SPR binding study, rhIL-2 concentrations of 5-10uM and above produced aberrant signals that were not able to be fit. These signals saturated the surface and were excluded from the analysis to avoid artifacts in the fitting.
For other studies, data were excluded when a sample failed due to a technical issue, and exclusion criteria were established before data were collected. All such exclusions are listed in the Source Data file and summarized below. * exclusion based on low event count in flow cytometry analysis (<100 events in target cell population). Criteria for event counts were >100 events in target population. All such exclusions are highlighted in the Source Data file All data produced was based on biologically independent samples with the exception of the Discoverx screen, in which technical replicates were employed to ensure technical reproducibility for screening. No technical replicates were used for statistical measurements. In vitro studies were performed with multiple donors where applicable and described in methods section. All attempts at replication were successful.

Antibodies
Antibodies used absence of Treg expansion were confirmed in >2 additional studies. PK/PD study in tumor bearing mice was performed in singlicate using 4-7 animals per time point. B16-F10 tumor efficacy study was performed in singlicate using 15 animals per condition. PBMC dose response for pSTAT5 study (Supplementary Fig 5) was performed in singlicate with 3 independent donor samples. Supplementary Fig 11-IL-5 levels after rhIL-2 or THOR-707 administration was performed in singlicate in 3 independent animals per time point Samples/animals in all experiments were randomized to each group.
In PK/PD study in naive mice, animals were assigned to each group at random. In B16F10 tumor PK/PD study, mice were randomly assigned to groups for treatment. Randomization was performed based on average tumor volume (mm3) at~60-80mm3 for each group on day -1 (a day prior to dosing). In B16F10 efficacy study, mice were randomized at tumor volumes of 50 mm3 for the treatment. All animals were randomly allocated to the different study groups.
For PK/PD studies in mouse, mice were allocated to study groups at random.
All studies reported herein report quantitative data measured without subjective scoring. The investigators were not blinded during experiments All antibodies used in the studies reported are described in detail in the supplemental materials section of the manuscript and are listed below. Note that full information on the approval of the study protocol must also be provided in the manuscript.

Human research participants
Policy information about studies involving human research participants