2022.02.18
In under 40 days since announcing the partnership, Insilco Medicine and Fosun have selected preclinical candidate and initiated IND-enabling studies with an AI-discovered inhibitor of QPCTL target
Following recent success in identifying a novel pan-fibrotic target and promising drug candidates in Idiopathic Pulmonary Fibrosis (IPF) and Kidney Fibrosis for Insilico Medicine's internal programs, the company's end-to-end AI platform has delivered a speed record-setting milestone in cancer immunotherapy for our external industry partner -- Fosun Pharmaceuticals (stock code: 600196.SH, 02196.HK), a leading innovation-driven international healthcare group in China. In under 40 days since the strategic collaboration announcement, Insilico Medicine and Fosun Pharmaceuticals have achieved the first preclinical candidate ISM004-1057D, -- a potentialfirst-in-class small molecule inhibitor of QPCTL targeting CD47-SIRPα pathway, providing not only high tumor penetration but also reduced hematologic side effects and avoiding antigen sink. Insilico Medicine and Fosun Pharmaceuticals are now working together on the IND-enabling studies to advance this program to clinical trials.
"The world needs innovative, effective first-in-class therapeutics, and in this collaboration with Fosun Pharma, we are setting records in speed and novelty with artificial intelligence. I am delighted that in less than 40 days since our collaboration announcement, we managed to achieve preclinical candidate and start IND-enabling studies for the QPCTL program"
Alex Zhavoronkov
PhD, CEO of Insilico Medicine.
A rebel of cells
Cancer is a disease in which some of the body's cells go rogue, spreading uncontrollably to other parts of the body and leading to functional disruption of normal processes in tissues and organs, and eventually to death. In 2020, an estimated 19.2 million people worldwide were diagnosed with cancer, and almost 10 million people died the same year from this devastating cause -- around 1 of every 6 deaths worldwide.

In 1957, Burnet and Thomas proposed the "immuno-monitoring" hypothesis, suggesting our body is constantly on the alert -- searching and killing any "rebel" cells of our own to maintain a homeostatic environment. Our immune system reacts to various tumor-specific as well as non-tumor-specific antigens expressed by malignant or damaged cells and kills such cells. The anti-tumor immunity is enabled by an army of professional bodily "guards" -- macrophages, T-cells, natural killer cells and other important effector cells. However, cancer cells are particularly "smart" at escaping the fierceness of our immune system by utilizing a variety of mechanisms.

One key mechanism cancerous cells use to deceive our immune system is by overexpressing CD47, a signaling molecule which is widely regarded as a "don't eat me" message for the macrophages and other immune cells. Under normal physiological conditions, our healthy cells use this protein molecule routinely as a means of maintaining peace with an otherwise highly aggressive immune system. CD47 protein interacts with another important player in this process -- signal regulatory protein α (SIRPα), a membrane glycoprotein present on the surface of some immune cells (myeloid cells) -- leading to a negative control of the innate immune cells.
Removing the blindfold from the immune system
The knowledge about how cancer cells use CD47-SIRPα interaction to hide from the immune system opens doors for promising new strategies to cure cancer. The idea behind immunotherapy is to "remove the blindfold" from the immune cells and allow them to recognize and kill tumor cells, and one of the promising ways of achieving this goal is by inhibiting CD47 signaling in cancer cells and disrupting the very mechanism of escape from phagocytosis.

A number of companies and research groups have attempted to inhibit CD47 glycoprotein using anti-CD47 antibodies. The problem with this approach -- leading to discontinuation of several clinical trials in the past years -- is that CD47 signaling proteins are present not only on cancerous cells, but also ubiquitous among healthy cells, for instance, those of the hematopoietic system (normal red blood cells, senescent red blood cells, and platelets). This often leads to serious off-target effects, leading to anemia and other complications.

The so-called "antigen sink" effect could also pose a problem in the development of anti-CD47 antibody treatments. CD47 is expressed ubiquitously among malignant and healthy cells, meaning that large initiation doses and/or frequent administrations may be required for a drug to achieve an effective therapeutic effect. The alternative route is to therapeutically target SIRPα molecule, which is less distributed vs. CD47. SIRPα is highly expressed, however, on myeloid cells and central and peripheral nervous system cells, so the potential for neurological side-effects should be considered when using therapeutics that target SIRPα.
A speed record-setting milestone achieved using Artificial Intelligence (AI)
An alternative approach to stripping the cancerous cells of CD47-induced protection from attacks by the immune system has been the focus of our collaboration with Fosun Pharmaceuticals (stock code: 600196.SH, 02196.HK), a leading innovation-driven international healthcare group in China.
Today we are pleased to announce the first major milestone in this collaboration -- a nomination of the first preclinical drug candidate ISM004-1057D, a potential first-in-class small molecule inhibitor that targets glutaminyl-peptide cyclotransferase-like protein (QPCTL). QPCTL is responsible for CD47 pyroglutamate formation, the modification essential for the interaction of CD47 with SIRPα. The small molecule inhibitors of QPCTL mimic the therapeutic effect of CD47 antibodies to attack tumor cells by blocking this signal.
The preclinical candidate compound (PCC) has been nominated by our AI platform and R&D team in under 40 days since the collaboration announcement -- what we believe an industry record for speed -- and showed in vivo anti-tumor efficacy in both liquid and solid tumors. Meanwhile, the compound demonstrated favorable pharmacokinetics and ideal safety profiles in vivo preclinical studies, with auspicious value in further clinical development. Moreover, ISM004-1057D as a small molecule inhibitor of QPCTL could potentially provide not only high tumor penetration but also reduce hematologic side effects and avoid antigen sink, which is commonly seen upon treatment by most of the current CD47 antibodies owing to wide expression of CD47 beyond tumor cells.

"It has been a phenomenal experience to work together with Fosun Pharma to achieve the PCC nomination milestone for the QPCTL program in record time"
Feng Ren
Ph.D., CSO of Insilico Medicine

The discovery has become possible due to our AI system PandaOmics, which was used to develop the target hypothesis in oncology. The other artificial intelligence component, Chemistry42, was used to generate compounds with good drug-like properties. After synthesizing and testing 71 compounds, we managed to deliver preclinical candidate compounds within 9 months from project initiation.

This new success of our AI platform comes as a continuation of a series of proof-of-concept milestones in various therapeutic areas – including Idiopathic Pulmonary Fibrosis and Kidney Fibrosis – for our internal programs. Having invested years of research and development into our cutting-edge AI platform, we are now reaping the rewards of a truly industrialized drug discovery process which is fast, cost efficient, and highly innovative.
Stay tuned, follow us on social media!