When Emily Weiss and Adair Borges quit their jobs last May, they weren't thinking small. Weiss, a microbiologist who'd spent years at Illumina and DuPont, and Borges had a thesis that bordered on the unsettling: parasitic worms—creatures that have perfected the art of dampening human immune responses over millions of years without killing their hosts—might just hold the secret to resetting autoimmune diseases.
Less than a year later, their San Francisco startup, Ditto Biosciences, had screened more than a million parasite, virus, and tick proteins through an AI platform, won a spot in Y Combinator's Winter 2026 batch, and reported early binding candidates hitting single-digit nanomolar affinity against key autoimmune targets. Now the three-person team is building a tissue biobank to map how human immune systems remember encounters with parasite proteins, with an eye toward engineering nature's own immunomodulators into treatments for Crohn's disease, ulcerative colitis, and multiple sclerosis.
It's an audacious plan, and Ditto isn't alone in making it. The company's emergence reflects something of an inflection point in immunology—a field where the old playbook of chronic immune suppression is giving way to a conviction, shared by scientists and pharma executives alike, that lasting remission demands immune tolerance, not indefinite dampening. More than 15 million Americans—roughly 4.6% of the population—live with at least one diagnosed autoimmune disease, according to electronic health record data published last December. The question, increasingly, isn't whether to pursue "reset" therapies, but how fast they can reach patients. And which biological toolkits will prove most tractable.
When Blockbusters Face Erosion
Immunology remains one of the largest therapy areas by spending, with autoimmune indications alone accounting for 91% of immunology sales in 2023, IQVIA data show. Dupixent, the IL-4/IL-13 blocker from Sanofi and Regeneron, posted $17.8 billion in global sales for 2025—a 26% jump year-over-year—proving that targeted biologics for allergic and Type 2 inflammatory conditions still command enormous commercial appetite.
But the first wave of anti-TNF blockbusters is now facing biosimilar substitution at scale, a shift that's upending economics across the sector. Adalimumab biosimilars gained multiple FDA interchangeability designations in April and May of 2025. By mid-2024, they'd already captured roughly 8% of U.S. units, and uptake has accelerated since. Meanwhile, Medicare's Inflation Reduction Act brought negotiated pricing directly to immunology: Stelara, widely used in psoriasis, Crohn's, and ulcerative colitis, saw its "maximum fair price" take effect on January 1, 2026, with CMS projecting $6 billion in first-year savings across the initial ten drugs subject to negotiation.
Into this shifting landscape, Big Pharma has poured billions—sometimes frantically—into next-generation approaches. Merck acquired Prometheus Biosciences and its TL1A program for $10.8 billion in 2023. Roche paid $7.1 billion upfront, plus $150 million in near-term milestones, for Telavant's anti-TL1A antibody that October. More tellingly, perhaps, Sanofi acquired Dren Bio's deep B-cell depletion candidate DR-0201 for $600 million upfront—potentially $1.9 billion total—this past March, explicitly citing the potential for "sustained treatment-free remission" in refractory autoimmune diseases. Weeks later, Sanofi bid up to $9.5 billion for Blueprint Medicines to diversify beyond Dupixent's dominance. A feature in Nature Reviews Drug Discovery this February framed the moment as a "race to reset autoimmune diseases," underscoring the industry's pivot from symptom control toward something closer to curative intent.
Evolution's Back Catalog
The scientific rationale for mining parasite biology rests on evolutionary immunology—a discipline that's less glamorous than it sounds but rich with potential. Helminths, the parasitic worms that infect billions of humans globally, have co-evolved with mammalian immune systems for hundreds of millions of years. They secrete proteins that induce regulatory T-cell phenotypes, dampen inflammatory pathways like JAK/STAT and NF-κB, and modulate dendritic cell function to ensure both host survival and parasite persistence. A 2024 review in the Annual Review of Immunology described how gastrointestinal nematodes shape systemic immune states far beyond the gut—an observation with potentially sweeping therapeutic implications.
Molecules like FhHDM-1, a small peptide from the liver fluke Fasciola hepatica, have shown NLRP3 inflammasome inhibition and protective effects in preclinical arthritis models dating to 2016. Heligmosomoides polygyrus secretes Hp-TGM, a TGF-β mimic that triggers regulatory signaling without any sequence homology to mammalian TGF-β—a biological trick with obvious therapeutic appeal. ES-62, a glycoprotein from the filarial nematode Acanthocheilonema viteae, has demonstrated anti-inflammatory activity in some models, though results have been inconsistent.
Yet translating epidemiologic observations—populations with endemic helminth infections often show lower autoimmune disease rates—into actual drugs has proven treacherous. Coronado Biosciences' Phase 2 trial of Trichuris suis ova (live pig whipworm eggs) in Crohn's disease failed in October 2013, a setback that underscored the gap between hygiene hypothesis correlations and controlled therapeutics. The field pivoted. Rather than dosing patients with whole organisms, researchers began isolating, characterizing, and engineering parasite proteins with known targets and mechanisms.
This is where artificial intelligence enters—and complicates—the picture. Ditto's founders claim that over 98% of viral and parasite proteins remain uncharacterized, a figure broadly consistent with the literature on the "dark proteome" in non-model organisms. Advances in protein structure prediction—AlphaFold 3, released in May 2024, broadened to multi-molecule complexes—and generative design models now allow researchers to screen massive sequence databases, predict binding partners, and engineer scaffolds for developability, all before synthesizing a single molecule. Generate:Biomedicines' billion-dollar-plus deal with Novartis in 2024 and its March 2026 IPO illustrate sustained capital allocation to AI-first biologics platforms. Ditto's founders claim to have screened over a million proteins and identified candidates binding autoimmune-relevant targets at 1–2 nanomolar affinity, with lab data reportedly generated by partner Adaptyv, according to LinkedIn posts from February and March 2026.
Regulatory and payer dynamics, meanwhile, create whitespace. With Humira biosimilars eroding legacy revenue and IRA negotiation pressuring incumbent brands, insurers and health systems are newly receptive to therapies that promise durable benefit and reduced long-term cost—provided, of course, they deliver on efficacy and safety. The FDA finalized guidance on "Nonclinical Evaluation of the Immunotoxic Potential of Pharmaceuticals" in June 2023, tightening expectations for immunomodulators. For parasite-derived proteins, the central hurdle is immunogenicity: will the human immune system recognize these non-mammalian scaffolds as foreign and mount neutralizing antibody responses? Ditto's stated strategy—building a tissue biobank to map pre-existing immune memory against parasite proteins—addresses this risk head-on. Whether it works is another question entirely.
A Crowded Field, Different Bets
Ditto isn't alone in exploring evolutionarily inspired immunomodulation. Holoclara, based in Pasadena, raised an oversubscribed $16 million Series A in mid-2024 and dosed its first patient in a Phase 1 trial in Australia late last year, targeting eosinophilic esophagitis with a roundworm-derived molecule. The choice of EoE—a relatively small, well-characterized allergic and autoimmune indication—reflects a pragmatic development strategy: prove safety and mechanism in a focused population before expanding to larger diseases like inflammatory bowel disease or rheumatoid arthritis.
Parallel to parasite mining, a cohort of antigen-specific tolerance platforms is advancing toward registration. COUR Pharmaceuticals' CNP-104, a tolerizing nanoparticle for primary biliary cholangitis, earned FDA Orphan Drug Designation in January 2025 following positive Phase 2a topline data last November. Anokion's KAN-101, an "inverse vaccine" for celiac disease, showed supportive Phase 1b/2 signals in May 2024. Topas Therapeutics published preclinical data in Frontiers in Immunology on March 17, 2025, demonstrating that liver sinusoidal endothelial cell-targeted nanoparticles induce antigen-specific tolerance in CD4+ T-cell disease models; the company holds partnerships with Eli Lilly and Boehringer Ingelheim. Parvus Therapeutics announced a collaboration with AbbVie in March 2024 to develop peptide-MHC-based tolerogens for inflammatory bowel disease. The 9th Antigen-Specific Immune Tolerance Summit, held in Boston this past March 3–5, convened late-stage programs from Diamyd, COUR, Aravax, and others—a sign that an ecosystem is coalescing around shared biomarkers and regulatory pathways.
Cell therapy, too, is entering the tolerance race. Sonoma Biotherapeutics received a $45 million milestone from Regeneron in September 2024 for its engineered regulatory T-cell programs; the company installed a new CEO last November. Quell Therapeutics presented preclinical data for QEL-005, a CAR-Treg candidate, at ACR Convergence in October 2025. Immuthera licensed City of Hope's CD6/CTLA-4 CAR-Treg platform this January, with an IND allowed in November 2025. These modalities—cellular, nanoparticle-based, and protein-derived—compete and cross-pollinate, each navigating distinct manufacturing, regulatory, and reimbursement challenges.
Ditto's bet is that parasite proteins offer unique advantages: multi-target activity (a single helminth molecule may modulate multiple immune checkpoints simultaneously), proven in vivo efficacy across millions of years of evolution, and potentially lower manufacturing complexity than cell therapies. Emily Weiss, who holds a PhD in microbiology and genomics, posted on LinkedIn in early March that the team had moved from "zero lab infrastructure" to active immunology assays after hiring a new immunologist. The startup lists a team size of three on its YC profile—a lean operation consistent with other pre-seed computational biology ventures—but claims to have built partnerships for wet-lab validation and is assembling patient-derived tissue samples to assess immunogenicity risk before IND-enabling studies.
Data Will Decide
The near-term outlook for tolerance and "reset" therapies hinges on clinical data, as it always does. Diamyd Medical, pursuing HLA-stratified glutamic acid decarboxylase immunotherapy in type 1 diabetes, received Fast Track designations in 2024 and is enrolling its DIAGNODE-3 trial with an eye toward accelerated U.S. approval; however, the program's history includes mixed outcomes across multiple trials, and precision biomarker selection will be critical. COUR, Anokion, and Topas must demonstrate that tolerizing platforms translate to hard clinical endpoints—remission, steroid sparing, disease modification—not just transient immunologic signals. For engineered Tregs, manufacturing scale, persistence, and safety profiles in chronic autoimmune settings (as opposed to acute transplant or graft-versus-host disease) remain open questions.
Parasite-derived therapeutics face additional hurdles. First, immunogenicity: even with computational deimmunization and humanization, regulators will demand robust anti-drug antibody assays and risk-mitigation strategies, per FDA and EMA guidance updated in recent years. Academic data on molecules like ES-62 and FhHDM-1 span roughly two decades, with some models showing protection and others showing no effect, underscoring the need for target validation and mechanistic clarity. Second, most published helminth protein data are preclinical; the translational gap between mouse colitis models and human Crohn's or ulcerative colitis is wide, as Coronado's 2013 TSO failure made painfully clear. Third, Big Pharma's recent acquisitions—Prometheus, Telavant, Dren Bio—set a high bar: acquirers will demand late preclinical or Phase 1 data, clear IP, and differentiated mechanisms before writing checks. Ditto's current stage—a million proteins screened, early binders identified, no disclosed IND timelines—places it years from that inflection point.
Yet the structural incentives favor bold bets, perhaps more than the founders expected when they started. Dupixent's $17.8 billion in sales for 2025 prove that novel immunomodulators can reach blockbuster scale. IQVIA forecasts immunology among the top three therapy areas by 2028, with the psoriasis market alone valued around $40 billion in 2024 and projected to grow roughly 5% annually through 2029. Biosimilar pressure and IRA pricing are forcing incumbents to pursue differentiation through durability and curative intent, creating tailwinds for tolerance platforms—provided they deliver. The confluence of AI-enabled protein discovery (structure prediction, generative design, developability optimization), a maturing regulatory pathway for first-in-class immunomodulators, and multi-billion-dollar pharma appetite for pipeline replenishment suggests that the tolerance/reset wave is entering its build-out phase.
For Ditto, the next 18 to 24 months will likely determine viability. The company must transition from computational hits to validated preclinical leads, demonstrate that engineered parasite proteins avoid catastrophic immunogenicity, and show disease-relevant activity in at least one animal model with translational credibility. It will need to raise a Series A—likely in the range of Holoclara's $16 million or higher, given the capital intensity of biologics development—and recruit experienced drug-development talent beyond its founding trio. If early data hold, partnerships with contract research organizations, academic medical centers, or pharma scouts could accelerate timelines. If immunogenicity or off-target effects emerge, the program could stall before reaching the clinic.
The broader question is whether mining evolution's "back catalog" of immune evasion strategies—parasites, viruses, ticks, and other organisms that have mastered immune modulation—can deliver transformative therapies faster or more reliably than purely rational design or repurposed biologics. Holoclara's first-in-human milestone suggests the concept is at least testable. Ditto's AI-first approach, targeting a million-protein search space that remains largely uncharacterized, represents a more speculative—but potentially higher-upside—variant.
The race to reset autoimmune diseases is crowded with nanoparticles, inverse vaccines, engineered cells, and next-gen antibodies. Whether parasite proteins join that list, or become a footnote in translational medicine, depends on data yet to be generated, assays yet to be run, and patients yet to be dosed. For now, Weiss and Borges are betting that worms know something we don't.
