Synlogic raised over $400 million to engineer bacteria as medicine. They called it Synthetic Biotics™. On January 21, 2026, Nasdaq delisted them. One employee remains. The company that was supposed to prove engineered living therapeutics could work instead proved how fast they can die.
Two weeks later, on February 12, Seres Therapeutics paused its SER-155 program — a live biotherapeutic with FDA Breakthrough Therapy designation that had shown a 77% relative risk reduction in bloodstream infections among bone marrow transplant patients. They cut 30% of their workforce. Cash runway: Q3 2026. The program that proved engineered microbes could prevent lethal infections couldn't find $50 million to start its Phase 2 trial.
Against this backdrop, five technologies are converging on the same radical idea: engineering biology to fight biology. Not new chemical compounds. Not slightly modified drugs. Bacteria reprogrammed to kill bacteria. Viruses loaded with CRISPR payloads. Cell-free factories that manufacture antimicrobial cocktails on demand. Gene drives that turn resistance plasmids against themselves.
The science has never been stronger. The funding environment has never been more hostile. Both of these things are true at the same time.
The Maturity Gradient
These five platforms exist on a steep gradient from patient bedside to laboratory bench. Only two have reached human clinical trials. The most innovative — and potentially most transformative — remain entirely preclinical. The gap between them is not just time; it's the distance between proof-of-concept and the grinding reality of GMP manufacturing, regulatory classification, and commercial survival.
I. The CRISPR Vanguard
The two platforms closest to patients share a common architecture: bacteriophages armed with CRISPR payloads. They represent the first generation of truly programmable antimicrobials — viruses engineered not just to kill bacteria, but to destroy specific resistance genes while doing it.
LBP-EC01: The Registration Trial
Locus Biosciences is running what may be the most important phage trial in history. ELIMINATE (NCT05488340) is a registration-enabling Phase 2/3 trial — the first statistically powered trial of a bacteriophage therapy ever conducted. LBP-EC01, a six-phage cocktail enhanced with CRISPR-Cas3, is being tested against multidrug-resistant E. coli urinary tract infections.
Part 1 data, published in The Lancet Infectious Diseases: 87.5% microbiologic cure rate. 100% symptom resolution in evaluable patients. Part 2 — the randomized, controlled, double-blind efficacy portion — is now enrolling up to 288 patients across US and EU sites.
The backing is unprecedented for phage therapy. BARDA has committed up to $93 million as part of a $152 million program supporting Phase 2, Phase 3, and FDA marketing approval activities. In January 2026, Locus received a separate $3.3 million NIH award (up to $28 million total) for LBP-PA01, an AI-designed phage therapeutic targeting antibiotic-resistant Pseudomonas aeruginosa pneumonia.
CRISPR-Cas3 is the key differentiator. Unlike Cas9, which makes precise cuts, Cas3 shreds DNA — it processively degrades the target, making resistance through simple point mutations far more difficult. The system is designed to destroy the bacterial chromosome at multiple essential sites simultaneously.
SNIPR001: The Gut Decolonizer
SNIPR Biome is solving a different problem: prophylactic elimination of dangerous bacteria from the gut before they cause bloodstream infections. Their target population — blood cancer patients undergoing hematopoietic stem cell transplant — faces devastating infection rates during neutropenia.
Final Phase 1 results, published in The Lancet Microbe on March 3, 2026: SNIPR001 — a CRISPR-Cas-armed phage cocktail delivered orally — achieved a 78% reduction in gut E. coli colonization at the highest dose. Gut-restricted (detected in stool but not plasma or urine). Microbiome preserved. No serious adverse events.
The Phase 1b trial (NCT06938867) in 24 HSCT patients across 8 US centers is now more than halfway enrolled. FDA has granted Fast Track designation for prophylaxis of bloodstream E. coli infections in hematological malignancy. CARB-X is co-funding $5.48 million.
This is the first randomized, placebo-controlled trial of a CRISPR-based antimicrobial in humans. The concept — selective pathogen elimination while preserving commensals — is the therapeutic logic that distinguishes engineered biology from conventional antibiotics. Antibiotics carpet-bomb the microbiome. CRISPR phages are guided missiles.
II. The Biosafety Breakthrough
The single biggest objection to engineered living therapeutics has always been: what if they escape? Release genetically modified bacteria into a patient's gut, and they could persist in the environment, transfer their engineered genes to wild-type organisms. Every regulatory framework collides with this question.
SimCells eliminate it by removing the chromosome entirely.
Published in PNAS in March 2026, SimCells are chromosome-free bacterial cells — stripped of all replicative DNA by controlled endonuclease activity, leaving only engineered plasmid-borne programs and intact cellular machinery. They cannot divide. They cannot replicate. They cannot transfer chromosomal genes. Plating yields zero colonies, with an escape frequency below 10⁻⁸, meeting NIH biosafety guidelines.
But they can still act. The cellular machinery — ribosomes, enzymes, metabolic pathways — remains functional. Glycolysis (10 genes) continued operating in chromosome-free SimCells, regenerating ATP and NADH to power synthetic circuits for up to 10 days.
The SimCell kill chain against MDR E. coli ST131:
- Target recognition: Surface-displayed anti-OmpA nanobodies bind the target pathogen
- Toxic injection: Heterologous Type VI secretion system (T6SS) injects lethal effectors directly into the bacterial cytoplasm
- Chemical warfare: Salicylate hydroxylase converts aspirin into catechol, generating sustained local H₂O₂ production
- Result: >97% elimination of MDR E. coli ST131 within 48 hours. 10³-fold selective reduction in mixed communities with minimal disruption to non-target bacteria
SimCells come in two sizes: standard (1–2 µm) and mini-SimCells (100–400 nm, generated through asymmetric division via minD deletion). The platform is modular — swap the nanobody to target a different pathogen, swap the effector to deliver a different payload.
The conceptual leap matters as much as the data. SimCells occupy a new category: not a living cell, not a cell-free system, but something between — a "smart bioparticle" that may sidestep the GMO classification entirely because it cannot propagate. If the regulatory void surrounding engineered living therapeutics has a solution, it may look like this: organisms that aren't organisms anymore.
III. The Trojan Horse
The most subversive technology in this landscape doesn't kill bacteria. It turns their own resistance machinery against them.
pPro-MobV, developed by Ethan Bier's group at UC San Diego, is a conjugal CRISPR gene drive published in npj Antimicrobials and Resistance in February 2026. The concept: engineer a ~65 kb plasmid that hijacks conjugation — the same horizontal gene transfer mechanism I described in my last post as the engine of resistance spread — to deliver CRISPR-Cas9 payloads that destroy antibiotic resistance genes.
The plasmid carries an arabinose-inducible operon encoding three λRed recombination proteins plus Cas9, a guide RNA targeting the bla (ampicillin resistance) gene flanked with homology arms, and the full IncP RK2 conjugative transfer machinery. It spreads from cell to cell through the same conjugation tunnel that normally transfers resistance. But instead of carrying resistance genes, it carries their executioner.
| Metric | Result |
|---|---|
| Transfer efficiency (72-hour liquid conjugation) | ~40% |
| Reduction in ampicillin-resistant CFU (RecA⁻ recipients) | ~100,000-fold (10⁵) |
| Gene disruption: precise sgRNA insertion | 25% of events |
| Gene disruption: homology-based deletion (HBD) | 75% of events |
| HBD deletion size | ~1.2 kb (entire bla coding region) |
| Biofilm activity | Confirmed — transfers through cell-cell contacts |
The dual inactivation pathway is elegant. Cas9 cuts the resistance gene. Then either the sgRNA cassette inserts at the cut site (25%), or the λRed recombination system mediates a larger homology-based deletion that removes the entire resistance gene (75%). The target plasmid stays intact — only its resistance payload is destroyed.
Two critical features: it works in biofilms (where 65–80% of clinical infections originate and antibiotics routinely fail), and the system includes a built-in safety mechanism — the same HBD process can remove the gene drive cassette itself, preventing uncontrolled environmental spread. The team also demonstrated phage delivery of pPro-MobV components, opening a dual-vector strategy.
This is early-stage work — all in vitro, all E. coli, no animal models. But if plasmids are the real enemy, this is the first weapon designed to fight them on their own terms.
IV. Factories Without Cells
The four platforms above all face a shared manufacturing nightmare: producing living or semi-living biological systems under GMP conditions. Cell-free protein synthesis (CFPS) eliminates the cell entirely.
A 2025 Communications Biology study demonstrated multiplexed bacteriocin production using cell-free gene expression: rapid synthesis of antimicrobial peptide cocktails from DNA templates. These tailored combinations eradicated MDR pathogens while preventing resistance development — the cocktail approach blocks single-mutation escape routes that doom individual antimicrobials.
Why CFPS works for bacteriocins: these antimicrobial peptides are toxic to the bacteria that would normally produce them, making cellular production self-limiting. Cell-free systems don't need to keep cells alive. The toxicity that makes bacteriocins powerful as drugs makes them impossible to manufacture in conventional bioreactors. CFPS dissolves this contradiction.
The economics are shifting fast. A 2026 Nature Communications study reported protein production costs dropping from ~$4,080/g to ~$39–60/g — a 95% reduction. At 4-mL scale with oxygen supplementation, yields reached 3.7 g/L. The platform has been demonstrated at 100-liter scale for other therapeutic proteins.
The convergence with AI-driven design accelerates everything. A Nature Communications pipeline combined deep learning with cell-free synthesis to screen 500 computationally designed antimicrobial peptide candidates, identifying 30 functional AMPs — six with broad-spectrum activity against MDR pathogens and no detectable resistance development.
Cell-free bacteriocin cocktails won't replace antibiotics for systemic infections. They're best suited for topical, wound, and device-associated applications. But as a rapid-response platform — producing custom antimicrobial cocktails within hours of pathogen identification — they offer something no other technology can: speed matched to the pace of resistance evolution.
V. The Valley of Death
Between these five platforms and the patients who need them lies a killing field that has already consumed hundreds of millions in invested capital.
The Corporate Graveyard
Synlogic wasn't an accident. The engineered therapeutics space is littered with companies that proved the science worked, then died before reaching market. PHAXIAM liquidated in 2025 despite €150 million invested. BiomX discontinued its lead program despite Phase 2 data showing 500-fold pathogen reduction. The pattern is structural: antimicrobial development generates $46 million average annual revenue against development costs exceeding $1 billion. The economics that killed conventional antibiotics are even more lethal for complex biological therapies that cost more to manufacture.
Seres is the cautionary tale in real time. SER-155 has everything a drug candidate could want: Breakthrough Therapy designation, Fast Track designation, Phase 1b data showing 77% reduction in bloodstream infections, a $3.6 million CARB-X grant. What it doesn't have is money. The company paused the program, cut 30% of staff, and pivoted to earlier-stage inflammatory disease programs. A 15-patient investigator-sponsored trial at Memorial Sloan Kettering is the only remaining clinical activity. Stock down 75% year-over-year.
The Regulatory Void
No genetically modified bacterial therapy has ever been approved by the FDA. The two approved live biotherapeutics — Rebyota and Vowst — are donor-derived microbiota products for C. difficile, not engineered organisms. The regulatory framework was designed around strain characterization, not synthetic gene circuits.
The gaps are severe. The EMA has no harmonized definition for "engineered living materials." The FDA's CBER has 19 therapeutic product guidances in its 2026 agenda but nothing specific to engineered antimicrobial organisms. International harmonization through ICH doesn't cover LBPs at all. A company developing SimCells for the US, SNIPR001 for the EU, and pPro-MobV globally would navigate three entirely different regulatory regimes, none designed for what they're building.
Manufacturing at the Edge
GMP production of living therapeutics confronts challenges small-molecule drugs never face: cell viability throughout manufacturing, batch-to-batch consistency from biological materials, short shelf lives, continuous sterility monitoring. The engineered living therapeutics market reached $177.3 million in 2024 — a rounding error against the $50+ billion antibiotics market — projected to reach $1.03 billion by 2032. But growth depends on solving manufacturing problems that remain unsolved.
Where This Converges
Look at these five platforms together and a pattern emerges. Each attacks antibiotic resistance at a different point:
They're also converging technically. SimCells use the T6SS — the same secretion system being engineered in P. putida and V. natriegens as modular delivery platforms. pPro-MobV demonstrated phage delivery alongside conjugation. CFPS bacteriocin design is accelerating through the same AI pipelines that drive phage engineering. Engineered E. coli Nissle outer membrane vesicles could serve as oral delivery vehicles for any of these payloads. The platforms aren't competing — they're assembling a modular toolkit.
The question isn't whether these technologies work. The lab data is unambiguous. The question is whether they can cross the valley of death that has already consumed Synlogic, PHAXIAM, BiomX's lead program, and now SER-155.
The answer depends on three things: government funding (BARDA's $152 million for Locus is the model — private capital alone cannot sustain this), regulatory adaptation (frameworks must evolve to accommodate synthetic gene circuits, chromosome-free chassis, and conjugal gene drives), and the willingness to accept that the most innovative solutions may not look like drugs at all.
Biology built the resistance crisis. Plasmids weaponized conjugation. IS elements rewired genomes. Environmental reservoirs seeded resistance into clinical pathogens. The five technologies in this article represent the first serious attempt to turn that same biological ingenuity around — to make evolution a weapon rather than an obstacle. They are here. They work. Whether they reach the patients who need them is not a scientific question anymore. It's an economic and political one.