One engine, promising drugs for many diseases
We tested tens of thousands of drugs for their ability to induce therapeutic immune cell behaviors and validated them in follow-up experiments – surfacing promising therapies for multiple age-related diseases.
March 15, 2022 • 10 minute read

Age-related diseases cause a tremendous amount of suffering. They are many, they are awful, and they don’t discriminate. Scientists work tirelessly to find therapies, but they need new tools – tools that give them the superpowers such a fight demands. At Spring, we’ve dedicated ourselves to building this technology.

We built an engine to discover therapies for the diseases of immune aging – an engine that maps changes in our immune system, searches thousands and thousands of potential treatments, and gives scientists the tools to understand them.

This opens up a door to finding novel drugs for many different diseases. But delivering on such an ambitious opportunity requires:

  1. New technology to profile and perturb a wide array of immune biology.
  2. A non-traditional approach to drug discovery, placing engineered biological tools in the hands of scientists to help steer towards the most promising hypotheses.
  3. Validation across many diseases to prove the engine’s broad therapeutic relevance.

Running a multi-disease therapeutic engine

Starting with one example of the engine’s output: drugs to reduce the precancerous skin lesions that form after a lifetime of sun exposure.

How mice reacted after...
7 days
Untreated mice
4.1 skin lesions per mouse
Mice treated with Drug 601
2.7 skin lesions per mouse
View mice
Evaluating the impact of drug treatment on the formation of precancerous skin lesions. Mice were exposed to UV light over a sixteen-week period, followed by an eight-week treatment window. Drug 601 was identified by measuring therapeutic immune cell behaviors via Spring's multi-disease engine.

In a recent Spring study (manuscript in preparation), multiple drugs evaluated by our novel screening technology and surfaced via MegaMap helped cure mice of these sorts of precancerous skin lesions associated with sun exposure. The treatment group experienced drastically fewer and smaller lesions, along with improved skin health overall.

This study is an example of the kind of follow-up validation we run at Spring – orthogonal, established protocols used to test the outputs of our engine. We run such validation experiments for many different diseases (more below).

But even more importantly, the hypotheses tested in this study weren't generated through a traditional approach to drug discovery.

Instead, they came from putting engineered biological tools in the hands of scientists – tools designed to help them wield enormous biological data, answer hard questions quickly, and surface the most promising hypotheses. Equipping scientists with such superpowers creates repeatable conditions for promising drug discovery – an engine capable of addressing multiple diseases.

An engine should run again and again and again

Typical experiments in drug discovery measure very narrow ways in which drugs may or may not be helpful for affecting a specific disease target. This reductivist approach works really well, when it works – when we know the right targets and have strong conviction in their clinical relevance. But these aren’t easy conditions to recreate again and again – and certainly not for complicated diseases.

Meanwhile, aging and immune function are two incredibly complex, multi-faceted biological processes. Any singular measurement focused on an individual target or disease mechanism will fail to capture anything close to a complete picture of the biology.

This makes it especially challenging to apply the reductivist approach – and presents an opportunity to use new tools to search for clinically relevant therapies. Building these new tools for the field of immune aging has opened up a powerful therapeutic search process.

A landscape of novel immune therapies, using known drugs as waypoints

At Spring, we start by profiling a multitude of high-level immune cell behaviors that point towards therapeutic applications for multiple disease pathways – creating a rich, high-dimensional landscape of relevant biology for scientists to explore.

We then profile thousands of therapies within that landscape of relevant biology – placing multiple waypoints along the way using known compounds and cell perturbations.

Identifying novel drugs based on known waypoints
Vaccine controls
Activation controls
Anti-inflammation controls
Novel Drug 207
Suppressed cell death
in follow-up assay.
Novel Drug 453
Reduced tumor size
in animal study.
Novel Drug 672
Enhanced vaccine response
in follow-up assay.
Selecting the most promising drugs for validation in follow-up experiments based on their similarity to known reference profiles. Drug 207, Drug 453, and Drug 672 were surfaced by our platform and validated for therapeutic efficacy in follow-up experiments.

These reference profiles help guide our search towards areas of therapeutic interest – for example, therapies that activate the immune system, curb inflammation, or bolster vaccine response. By searching for unknown drugs that ‘look like’ these known experimental waypoints, we discover both novel drugs and novel mechanisms.

Combined with the phenotypic detail surfaced by MegaMap, our approach gives scientists a compass to navigate potential therapies’ impact on complex immune function – uncovering drugs that induce the immune responses most relevant to their clinical interest, whether it be precancerous lesion formation, tumor growth, vaccine response, or another age-related indication.

Validating efficacy in follow-up experiments for multiple diseases

The repeatability of this engine has allowed scientists to find promising drugs for multiple diseases. But validating their efficacy requires gold-standard experimental data – testing the outputs of our engine using independent, established protocols.

The design of these experiments varies from disease to disease: one, as seen above, involved exposing mice to UV to accelerate skin aging; another, seen below, involved using model vaccines to assess effects on vaccine efficacy.

In each disease case, we validated efficacy for our top drugs -- each of which was surfaced by the same, repeatable immune aging engine.

Tumor Size Reduction - Immuno-oncology
Tumor sizes decrease with drug treatment
Testing the anti-tumor effects of drugs that look similar to immuno-oncology "positive controls" known to drive anti-tumor immunity.
Skin Lesion Reduction - Dermatology
Lesion counts decrease with drug treatment
Mean lesion count (SEM)
Testing the effect of drugs which look similar to immune activation "positive controls" on age-related skin lesions.
Vaccine Response - Immunology
Production of protective antiviral antibodies increases with drug treatment
Antiviral Antibodies
Testing the vaccine enhancing effects of drugs that look similar to vaccine adjuvant "positive controls".
Immune Maturation
Cell surface levels of maturation marker CD86 increase with drug treatment
CD86 Expression
Testing the immune maturation effects of drugs that look similar to immune activation "positive controls".
Anti-Inflammatory Activity
Inflammatory cell death (pyroptosis) decreases with drug treatment
Maximal pyroptosis (%)
Testing the anti-inflammatory effects of drugs that look similar to inflammasome inhibitor "positive controls".

From decreasing tumor size to reducing skin lesions to improving vaccine response to maturing immune cells to decreasing proinflammatory cell death, this suite of clinically-relevant validation results have strengthened our belief in the multi-disease approach of our immune aging engine.

More fuel in the engine

We’re now ramping up the number of therapeutic programs running on this engine both internally and in in collaboration with partners – while simultaneously making non-stop investments in the fundamental technology behind it and moving its most promising outputs towards clinical trials.

We strongly believe that scientists who are working on treatments for age-related diseases are one of our greatest forces for good. And that equipping them with world-leading tools is one of the best possible uses of technology.

We look forward to sharing more as we do just that.

Charlie Marsh, Ben Komalo, Brandon White, Ben Kamens, Matt Booty, Christian Elabd

Thanks to the following for feedback on drafts of this post: Sarah Headland, Lauren Nicolaisen, Kelly Reynolds, and Tim Sullivan.

Thanks to Jeffrey Chan for the disease model illustrations.
PartnersSpring's technology has been used in many different therapeutic areas, and we're open to more. We're excited to discuss strategic collaborations with biotechs, pharma companies, and academic research groups looking for novel biological insights or tools to accelerate their searches. Contact us.
© 2022 Spring Discovery. All rights reserved.