Opinion: AI Is Pushing the I&I Pendulum From Biologics to Small Molecules – BioSpace

This post was originally published on this site.

Biologics have been the dominant treatment modality for chronic inflammation and immunology diseases for three decades. Yet even as biologics continue to generate excitement and draw investment in this space, we perceive the pendulum is beginning to swing toward small molecules, thanks in large part to artificial intelligence.

Although the application of AI to small molecule discovery is an industry-wide trend, for inflammatory and immunology diseases it is still relatively new—and extremely exciting, given how biologics have dominated the space to an extent not found in other disease areas. With AI, research is opening new types of chemistry and identifying small molecules with target selectivities, safety profiles and pharmacological properties that have been difficult to achieve, enabling a new generation of I&I therapeutics.

Our company, Montai Therapeutics, is one of several recognizing this pendulum swing and seeking to capitalize on the opportunity to bring breakthrough oral I&I medications to market. Here, we explore the convergence of small molecule and AI-driven discovery in the I&I space.

Biologics Are Not Unassailable

Biologics have several advantages that explain their dominance in the I&I space: high selectivity for their targets, good safety profiles and predictable pharmacology. In addition, it’s possible to identify a biologic for virtually any extracellular target with a high degree of success—a startling new concept when the first biologics were developed 30 years ago.

By contrast, small molecules, which have long been the therapeutic mainstays for many diseases, have been prone to failure in I&I. This is because the treatment of chronic autoimmune diseases imposes very high safety standards, yet small molecules often have suboptimal pharmacological properties and lack high target specificity, leading to off-target toxicities. As a result, fewer than 8% of small molecules and 11% of autoimmune drugs in development succeed in reaching the market, and they face post-approval setbacks, up to and including market withdrawals.

Biologics also have their limitations, however. Many of these therapies are approved for relatively narrow patient populations, and the side effects of their on-target immunosuppressive action further reduce the number of patients who can use them. Typically, biologics must be administered intravenously or subcutaneously in a clinical setting, and patients may be reluctant to receive these therapies because of the invasive nature or inconvenience of the procedures. Notably, the therapies are not available to patients in some parts of the world due to cost, transportation and storage issues.

With the help of AI, small molecule therapies for I&I diseases have the potential to address these limitations of biologics. First, due to their size, small molecules can engage intracellular targets such as transcription factors that are beyond the reach of biologics. This opens up the possibility of directly interfering with powerful intracellular pathways in a way large biologics can’t.

Second, AI will be instrumental in elucidating the complex pathological mechanisms underlying I&I diseases, which should lead to better-targeted small-molecule therapies. It’s also easier to combine two or more small molecules than to combine two or more large biologics, and it’s these combos that might prove to be breakthroughs in patient care.

Finally, small molecule therapies are more readily accessible to patients globally because they are, in general, less expensive to make and more readily transported.

While AI is also useful for discovering new biologics, it is especially well-suited to overcoming the historical limitations of small molecules and discovering new ones that capture all of the natural advantages small molecules have for the treatment of I&I diseases.

AI Is Amping Up Small Molecule Discovery for I&I Diseases

The first wave of AI platforms, across the biopharma industry and various disease areas, improved on known chemistry. While attractive and exciting at the time, some of these early AI companies may have over-hyped the potential of their platforms to transform drug discovery, generating skepticism about the companies’ abilities to deliver on those promises.

The new wave of AI platforms is better poised to find solutions to the problems of small molecule development by opening completely new chemical spaces, increasing the probability of success and reducing cycle times and costs. This is now possible because major technological advances have enabled the generation of vast amounts of data on which models can be trained, and advances in AI and computational power have led to better models that are capable of discovering molecular structures that have not previously been seen or considered as drug leads—and predict how they will work. AI thus allows the industry to imagine a two- or threefold better success rate for small molecules than was previously possible.

AI is changing the landscape of small molecule drug development across the board, but perhaps nowhere more so than in I&I diseases.

AI is just starting to make major inroads into the I&I disease space, but small molecules are fast becoming as developable as biologics were two or three decades ago. Although the Inflation Reduction Act (IRA) gives small molecules a shorter period of protection against price negotiation than biologics—creating a so-called “pill penalty”—we think the efficiency gains afforded by AI-based small molecule development will offset this disincentive, preserving the enormous value of small molecules to patients and the healthcare system overall.

The Companies Betting on Small Molecules in I&I

Building on these advantages, Montai and a handful of other companies are applying AI in various ways to the development of small molecule I&I therapies.

Nimbus and Atomwise are taking structure-based approaches that focus on modeling drug-target interactions. Psivant and Relay also have structure-based approaches, but they model moving proteins. Odyssey is using AI and other computational approaches across its target and drug discovery efforts.

At Montai, we take a biology-first approach, exploring diverse chemistry from nature’s bioactive chemical space. This space is far more diverse than conventional libraries of synthetic small molecules and is a known source of drugs. Yet it has been largely inaccessible to traditional discovery efforts, which have required significant resources and relied largely on chance and serendipity for success. Only with AI can we systematically mine this space for therapeutic solutions that evolution has already developed.

Investment in AI and Small Molecules for I&I

In the last four years, there has been a tremendous upsurge of investment both in using AI to develop small molecules across a range of disease areas, and in small molecules for I&I diseases (see graphic).

These are distinct but overlapping groups, as shown by deals involving Relay, Nimbus and Atomwise, each of which uses AI for small molecule development and has at least one I&I disease program. As interest in AI, small molecules and I&I diseases continues to grow, we expect to see greater convergence between the two groups and greater investment in companies that are using AI to discover small molecules for I&I diseases.

AI is changing the landscape of small molecule drug development across the board, but perhaps nowhere more so than in I&I diseases. We believe the technology will drive the major therapeutic breakthroughs for chronic I&I diseases in the coming decade by enabling the discovery of small molecule drugs that could provide attractive alternatives to biologics and be available to more patients globally.