Novartis Terminates Development of STING Agonist ADU-S100 Following Disappointing Clinical Trial Results

News Event

Today, Novartis announced that it will terminate the development of ADU-S100, a STING agonist acquired from Aduro Biotech. The discontinued programs primarily included combination therapies with Novartis’ own PD-1 antibody spartalizumab and with Bristol Myers Squibb’s Yervoy. However, Aduro plans to continue developing ADU-S100 in combination with Keytruda for head and neck cancer in Phase II clinical trials. The ADU-S100/spartalizumab combination yielded only a 9% response rate in Phase I trials, while the ADU-S100/Yervoy combination achieved just a 7% response rate in PD-1-refractory malignant melanoma. Aduro’s stock fell another 10% today to $1 per share, a stark contrast to its peak price of $40.

Drug Source Analysis

Aduro is one of the pioneers in immunotherapy. Its Listeria-based platform (LADD) yielded CRX-207, a mesothelin-based pancreatic cancer tumor vaccine, which failed in Phase III clinical trials in 2017. ADU-S100 is arguably the earliest human STING agonist. Another STING agonist, DMXXA, demonstrated promising efficacy in animal studies but activates only murine STING and exhibits low binding affinity to human STING. In 2015, during the bubble period for immuno-oncology (IO) and biotechnology, Novartis acquired this asset with an upfront payment of $200 million and milestone payments totaling $500 million. At that time, Novartis also acquired a 2.7% equity stake in Aduro for $25 million and invested an additional $25 million in the development of this product.

STING is an intracellular danger-sensing mechanism discovered approximately a decade ago. Any double-stranded DNA (dsDNA) can activate the enzyme cGAS, leading to the production of cGAMP, a natural STING agonist, which subsequently activates the immune system to eliminate hazardous cells. Under normal conditions, dsDNA is confined to the nucleus; however, in tumor and damaged cells, DNA may leak into the cytoplasm. Along with the cell surface Toll-like receptor (TLR) family, STING constitutes a major component of the innate immune system. Similar to TLR agonists, STING agonists exhibit relatively high toxicity, necessitating intratumoral administration for agents such as ADU-S100. Even with intratumoral delivery, the two STING agonists currently in clinical trials have demonstrated modest efficacy, showing little monotherapy activity; promising signals have been observed only in combination with PD-1 antibodies. Although many patients received lower doses (as early-phase clinical trials aim to determine the maximum tolerated dose), these data indicate that first-generation STING agonists have a narrow therapeutic window. In recent years, GlaxoSmithKline identified a class of STING agonists amenable to systemic administration; however, no clinical data have been reported to date.

The company that has benefited the most from STING is IFM, which sold STING agonists and antagonists at high prices to Bristol Myers Squibb and Novartis, respectively. Currently, STING remains one of the important targets of interest in the pharmaceutical industry, with at least 10 companies developing different types of STING agonists. However, much exploration is still needed to determine how to best leverage such a critical cellular component as STING. Intratumoral injection has, to some extent, addressed the issue of selectivity. Moreover, other mechanistic drugs administered via this route have already demonstrated certain early-stage efficacy; for example, last month, both the monotherapy and combination therapy of the intratumorally administered TLR9 agonist CMP-001 showed favorable responses in malignant melanoma. Combination regimens may also represent a viable strategy, but it remains debatable whether PD-1 inhibitors, which likewise lack selectivity, are the optimal combination partners. Low-level activation of STING may render tumor cells more sensitive to certain highly targeted therapies. CRISPR-based knockout of different genes could help identify which gene deficiencies make cells intolerant to low-level STING activation, thereby providing direction for the development of highly selective STING combination therapies. Of course, various targeted delivery technologies exist, but these may require the drug itself to possess sufficient potency. In any case, the term “broad-spectrum anticancer drug” has become somewhat self-contradictory. For STING-targeted agents to become successful therapeutics, their selectivity must be improved.