Insilico’s new program combines AI, circular mRNA and in vivo T-cell engineering to target early cellular drivers of age-related disease.
Insilico Medicine has launched a research program aimed at turning a patient’s own immune cells against cellular drivers of age-related disease, using generative AI to select targets and circular mRNA (cmRNA) to transiently engineer T cells in vivo. Called Longevity Vaccines, the initiative will initially focus on immune system rejuvenation, targeting senescent lymphocytes associated with immunosenescence, impaired vaccine responses and increasing susceptibility to infection and cancer with age.
The approach is less conventional vaccination than programmable cell therapy. Insilico proposes packaging cmRNA instructions in targeted lipid nanoparticles (LNPs), delivering them directly into the body and temporarily equipping T cells to recognize selected cellular populations; beyond senescent cells, potential targets include activated fibroblasts implicated in fibrosis and autoreactive lymphocytes involved in loss of immune tolerance.
Longevity.Technology: Preventive medicine usually conjures something rather gentler than in vivo cell engineering, but Insilico’s proposal takes prevention to its logical – and considerably more interventionist – conclusion: identify the cellular actors helping to set age-related pathology in motion, then remove them before the downstream damage becomes entrenched. The elegant part is the division of labor – transient circular mRNA would give T cells a temporary set of instructions, while AI is tasked with the altogether harder problem of deciding which cells deserve to be on the list.
That distinction matters. Oncology has already demonstrated the therapeutic power of engineered immune cells; applying related logic to otherwise healthy or presymptomatic people raises a very different safety bar, particularly when senescence is heterogeneous and not every old or damaged cell is necessarily disposable [1]. Insilico’s emphasis on self-limiting expression and highly selective targets is therefore not a technical footnote but the central wager. And while “Longevity Vaccines” is an arresting name, this is, for now, an early research program rather than a syringe waiting in the clinic.
The larger question is an intriguing one: if longevity medicine is serious about intervening before disease declares itself, prevention may increasingly cease to look passive. The trick will be ensuring that biological housekeeping does not become overenthusiastic spring cleaning.
Programming a temporary response
At the center of the proposed platform is cmRNA. Unlike linear mRNA, its closed-loop structure lacks exposed ends that are susceptible to exonuclease degradation; Insilico’s rationale is that this increased stability could provide a sustained working window while still allowing expression to fade rather than permanently modifying the cell.
That transience is deliberate. Once delivered by a targeted LNP, the genetic instructions are intended to turn endogenous T cells into temporary cellular hunters – recognizing a selected surface antigen, expanding in vivo, trafficking to the relevant tissue and eliminating antigen-expressing cells. The ambition is a single-administration treatment without the lengthy ex vivo cell manufacture associated with conventional CAR-T therapy.
“We built Insilico to treat aging and disease together, and to prove that AI can design medicines that deliver real patient impact,” said Alex Zhavoronkov, PhD, Founder and co-CEO of Insilico Medicine. “Longevity Vaccines initiative extends that same rigor to preventive medicine – delivering programmable, single-dose therapies that clear the root-cause cells of age-related disease, starting with the immune system itself [2].”
Finding the right cells
Killing a cell is only useful if the correct cell has been selected. For preventive treatment, that distinction becomes particularly unforgiving; a target must emerge sufficiently early in disease development to make intervention worthwhile, while being specific enough to avoid healthy tissues.
This is where Insilico intends its AI infrastructure to do much of the discriminating. Its Pharma.AI platform combines multi-omics target discovery through PandaOmics with generative biologics, Chemistry42 for chemistry and delivery-system design and inClinico for clinical development. Candidate surface antigens are ranked according to factors including evidence strength, expression timing, tissue specificity and deliverability, with laboratory results fed back into subsequent target and molecular-design iterations. Aging foundation models and the company’s Virtual Aging Cell are intended to add age-dependent and temporal context.
The underlying logic is to find cellular divergence early – not simply markers of damaged tissue after disease has become established.
From fibrosis to immune aging
The program extends Insilico’s existing aging-focused drug discovery strategy beyond small molecules. Its lead example is rentosertib, a TNIK inhibitor whose target was identified using AI and whose molecule was AI-designed; the drug has progressed into Phase III development for idiopathic pulmonary fibrosis following a Phase IIa study.
Longevity Vaccines moves that strategy into cell clearance. After immunosenescence, Insilico says other indications being evaluated include metabolic dysfunction and membranous nephropathy – conditions in which defined initiating cell populations might, in principle, be accessible using the same programmable chassis.
A higher bar
Much remains between a research architecture and a preventive therapy. Target specificity, delivery, durability and unintended immune effects will all require experimental evidence – particularly if the eventual population includes people without established disease.
Yet that demanding safety threshold may also make the program scientifically revealing. Preventive geroscience increasingly asks medicine to recognize pathology earlier and earlier; if cell therapy is to follow it upstream, defining exactly when an aging cell becomes a therapeutic target may prove as important as learning how to remove it.
Graphics courtesy of Insilico Medicine
[1] https://www.nejm.org/doi/full/10.1056/NEJMoa1707447
[2] https://insilico.com/news/lgv150926en-ai-driven-longevity-vaccines
