Insilico Medicine is known for using AI for new drug discovery and also works with nutraceutical companies like SRW Laboratories in developing new products.
Speaking to NutraIngredients-Asia on the sidelines of Founder Longevity Forum held in Singapore on February 27 to 28, the company talked about common queries from nutraceutical companies when it comes to the use of AI.
One, companies would like to find supporting evidence on how their nutraceutical products work.
This can help them better explain the way their products work to consumers.
“There is a growing interest for nutraceutical companies to use AI to find evidence to support the rationale of their existing products.
“Traditionally, we relied on literature findings to back up claims and understand how certain ingredients work.
“Now, with advanced AI models, we can better interpret and analyse the vast amount of available scientific data,” Frank Pun, Head of Insilico Hong Kong, Insilico Medicine, told us.
However, before finding the supporting evidence, the task is to first identify the active ingredients used in the nutraceuticals at the molecular level.
This also presents a challenge as active ingredients such as botanical extracts are complex mixtures of molecules.
However, once the molecule is identified, AI can be used to find out how the genes the molecule interacts with and influences.
“I think the primary challenge for nutraceuticals lies in identifying active ingredients at the molecular level. Once we characterise these specific molecules, we can then determine which genes they interact with and influence,” he explained.
Following which, companies will still need to conduct further studies, such as cell line studies, animal studies, or proceed to human clinical trials to validate the scientific evidence found using AI.
Some companies may also want to refine their existing formulations after getting to know how their products work.
“It really depends on the company’s goals. For many companies, the primary purpose is to explain how their current products work, followed by identifying better formulations and combinations.
“Compared to pharmaceuticals, nutraceuticals historically lacked strong scientific validation.
“Therefore, companies are now using AI to find more comprehensive evidence supporting their products' health claims,” he said.
For Insilico Medicine, one of the company’s latest research showed that a particular medication for dry eye can potentially address endometriosis.
The medication, known as Lifitegrast, is a prescription eye drop used to treat the signs and symptoms of dry eye disease.
By using AI, the company found out that Lifitegrast targets a specific gene - integrin subunit beta 2 (ITGB2) - by inhibiting its gene expression.
This particular gene is also upregulated in human endometriotic specimens. Therefore, inhibiting the gene expression may potentially address endometriosis.
The company then went on to validate the potential benefits of Lifitegrast for endometriosis by conducting a mice study.
The findings, published in Advanced Science last December, Lifitegrast has effectively suppressed lesion growth. This shows that the dry eye medication could potentially be repurposed for endometriosis treatment.
“Using a similar approach, if we can identify the active ingredients within nutraceutical products at the molecular level, then we can find their therapeutic gene targets and conduct comprehensive pathway analyses, similar to pharmaceutical drug repurposing strategies,” said Pun.
Pun also presented the above research findings when presenting on the topic “AI for Drug Discovery” during the second day of the forum.
Hallmarks of Ageing research
Insilico Medicine is also aiming to find out the genes that play a role in the hallmarks of ageing.
Twelve hallmarks of ageing have been proposed by scientists so far, including chronic inflammation, mitochondrial dysfunction and cellular senescence. The idea is that ageing can be slowed down by targeting these hallmarks.
Insilico Medicine is aiming to find out the genes that affect these hallmarks of ageing, as well as active ingredients at the molecular level that influence these genes.
For example, if a gene is involved in the cell senescence pathway, it could either be that the gene is pro or anti-cellular senescence - working in either of these two directions.
Once the gene target is identified, companies can then design a molecule that tackle this particular gene.