Sofi-cel

Sofi-cel (WU-CART-007) is Wugen's off-the-shelf anti-CD7 CAR-T cell therapy for relapsed/refractory T-cell acute lymphoblastic leukemia (T-ALL) and lymphoblastic lymphoma (T-LBL). The product uses gene-edited healthy donor T-cells, sidestepping the central problem that anti-CD7 CAR-T cells normally kill each other (fratricide) because they themselves carry CD7. Per Wugen's ASH 2024 presentation of the Phase 1/2 trial, 13 patients were dosed at the recommended Phase 2 dose, 11 were efficacy-evaluable, and 8 (73%) achieved a composite complete remission (6 CR + 2 CRi), with overall response in 10 of 11 (91%) and MRD-negative complete remission in 5 of 6 evaluable CRc patients (83%) [1][2]. An earlier publication cut reported 9 of 11 (81.8%) composite CRs at the same dose [3]. Median duration of response was 6.2 months, with the 95% confidence interval starting at 1.8 months and an upper bound not yet reached [1]. The FDA granted breakthrough therapy designation in January 2026 on the basis of this Phase 1/2 dataset [4]. If the response rate holds in the larger key cohort, this becomes the first off-the-shelf CAR-T for any T-cell cancer and one of the few credible options for a patient population whose median overall survival after standard chemo failure is measured in single-digit months. Two things matter for assessing the asset. The biology works (autologous CD7 CAR-T programs out of China have produced similar response rates), and the allogeneic manufacturing model is the commercial unlock if durability holds. The catch sits on the durability side: allogeneic CAR-T cells get rejected by the recovering host immune system, so the question is whether responses last long enough to bridge patients to curative stem cell transplant. A 6.2-month median DOR is encouraging but not yet definitive for that question.

Sofi-cel is a novel cell therapy in Phase 2, not approved anywhere. The program received FDA breakthrough therapy designation in January 2026 based on Phase 1/2 efficacy in heavily pretreated r/r T-ALL/T-LBL, a setting where standard treatment options are exhausted and historical outcomes are uniformly poor (composite CR rates in the single digits with conventional salvage chemo) [4]. The Phase 1 portion of the original NCT04984356 study established the recommended Phase 2 dose at 900 million cells following enhanced lymphodepletion; the Phase 2 expansion at that dose generated the response data above [3]. The key registration-supporting trial (NCT06514794, T-RRex) began dosing in March 2025 and is the trial intended to support a potential BLA filing in the 2027 timeframe, contingent on confirmation of durability [5][6]. The product is allogeneic, manufactured in batches from healthy donor cells using CRISPR/Cas9 to knock out both CD7 (to prevent fratricide) and the T-cell receptor alpha constant (TRAC) gene (to prevent graft-versus-host disease). This is the commercial unlock if it works: no per-patient apheresis, no 3-week manufacturing wait, and dosing on demand from a cryopreserved inventory. Wugen is a private St. Louis-based company spun out of Washington University; in August 2025 it announced a $115M financing to advance the key study, with cumulative reported funding above $170M led historically by RiverVest Venture Partners and including Tybourne Capital [7]. Outstanding regulatory items: a Phase 3 confirmatory design will need to be agreed with FDA, and the company will need to decide whether to file via accelerated approval (which BTD often enables) or wait for full data. No earnings calls to track since Wugen is private; data flow will come through ASH, ASCO, and EHA presentations.

CD7 is a protein sitting on the surface of essentially all T-cell cancers and most normal T-cells and NK cells (natural killer cells, a class of innate immune lymphocyte that helps kill virally infected and malignant cells) [8]. CD7's normal job is to help T-cells respond to activating signals during immune responses. From a drug-design standpoint, it is an attractive target for T-ALL precisely because it is so uniformly expressed on the tumor. The problem: a CAR-T cell engineered to attack CD7 is itself a T-cell carrying CD7, so the CAR-T cells kill each other before they can find the tumor. This is fratricide. Wugen's fix is to CRISPR out CD7 from the donor T-cells before adding the CAR construct, so the engineered cells have nothing to attack on each other. They also delete TRAC so the donor cells do not recognize the patient's tissues as foreign (which would cause graft-versus-host disease). Mechanism validation is strong by drug-development standards. Autologous anti-CD7 CAR-T programs from Chinese academic centers have produced complete remission rates in the 70-90% range across small studies [9]. The target is genetically and pharmacologically validated; what is unproven is whether allogeneic cells can persist long enough to do durable work before the patient's immune system clears them. That is the open biological question driving this asset's value.

NCT06514794 (T-RRex) is a Phase 2, single-arm, open-label, multi-cohort key study evaluating sofi-cel in adult and pediatric patients with r/r T-ALL or T-LBL [5]. The primary efficacy endpoint is the rate of complete remission (CR plus CR with incomplete hematologic recovery, CRi) within a defined post-infusion assessment window. Key secondary endpoints include minimal residual disease (MRD) negativity by flow cytometry, duration of response, event-free survival, and the proportion of responders successfully bridged to allogeneic stem cell transplant. MRD negativity means ultra-sensitive testing (typically flow cytometry at a 10^-4 sensitivity threshold) cannot detect any residual leukemic cells; in acute leukemia it predicts long-term cure and relapse-free survival far more reliably than a standard morphologic CR, which is why it is the metric most directly tied to the durability question this asset turns on. Patients receive lymphodepleting chemotherapy (fludarabine plus cyclophosphamide, enhanced regimen) before sofi-cel infusion. The single-arm design is appropriate. There is no standard of care that produces meaningful remissions in this setting, and historical outcomes are an unambiguous comparator. Enrollment target for the key has not been publicly disclosed, though registration-enabling CAR-T trials in heme malignancies typically run 50 to 100 patients. The design has no obvious flaws. The real question is whether FDA will accept CR rate as a surrogate for accelerated approval or demand a randomized confirmatory trial, which is operationally hard in this small, urgent population. Wugen will likely seek accelerated approval with a post-marketing commitment, the same regulatory path used by tisagenlecleucel in B-ALL.

Our model puts the chance of eventual approval at 9%. Phase 2 drugs in this area are approved about 21% of the time historically, and the model adjusts that starting point using ten facts about the trial and sponsor. The estimate is pulled down mainly by the sponsor's thin or weak approval record, the trial's few secondary endpoints, and weak or limited earlier-phase results, while the non-randomized design gives it a small boost. The remaining facts are close to average for this stage, so they don't move the number much either way.

Three failure modes dominate. First, durability. A CR at day 28 is not the same as a CR at month 12. The current median DOR of 6.2 months is encouraging but the confidence interval starts at 1.8 months. Allogeneic CAR-T cells get rejected by the patient's recovering immune system, and persistence windows for off-the-shelf cell products have historically been weeks rather than months. If patients relapse before they can be bridged to allogeneic stem cell transplant (the only known curative consolidation), the clinical proposition collapses to a bridge therapy with a niche role. Second, on-target toxicity and infection. The Phase 1/2 safety dataset showed cytokine release syndrome in 88.5% of patients (19.2% grade 3-4) and grade 1 ICANS in a single patient, with treatment-related grade ≥3 adverse events in 44% (8 of 18) [3]. CD7 is expressed on normal T-cells and NK cells, so anti-CD7 CAR-T causes deep, prolonged depletion of those compartments. Patients face elevated risk of opportunistic infections (CMV reactivation, EBV-driven lymphoproliferative disease, invasive fungal infections) on top of the lymphodepleting chemo already required, and infectious deaths have complicated efficacy stories in other CAR-T programs. Third, manufacturing and supply. Allogeneic CAR-T economics depend on consistent batch potency, scalable expansion, and reliable cold-chain logistics. A single contamination event or potency drift in a registration-supporting batch can stall a program for 12 to 18 months. Competitive risk is moderate: Beam Therapeutics' BEAM-201 (base-edited anti-CD7 CAR-T) is the most credible Western competitor, and several autologous Chinese programs from academic centers and Bioheng Therapeutics are advancing in parallel [11]. Commercial risk is bounded by the small addressable population (approximately 5,000-6,000 incident US T-ALL/T-LBL cases per year across pediatric and adult populations, with the addressable r/r subset materially smaller) but offset by likely CAR-T pricing of $500K to $800K per infusion if durability holds. Currently approved autologous CAR-T list prices include Kymriah at roughly $633K, Yescarta at roughly $538K, and Carvykti at roughly $465K [12]; a first-in-class allogeneic product for an indication with no approved alternatives could plausibly command the upper end of that range. At $600K-$700K per infusion and even a 1,000-patient annual addressable population, peak US revenue could reach $600M-$700M, with upside from line shifting and platform extension.

Worth watching closely. The specific signal to monitor: 6-month and 12-month duration of response data from the expanding key cohort, MRD-negativity rates by flow cytometry, and the proportion of responders successfully bridged to allogeneic stem cell transplant. If roughly half of complete responders are still in remission at one year (whether through persistence or successful transplant bridging), sofi-cel becomes a real product. If most relapse at three months, it becomes a bridge therapy with a defensible but smaller commercial niche. The January 2026 breakthrough designation is encouraging but not predictive on its own. Wugen is private, so progress visibility runs through ASH (December), ASCO (June), and EHA (June) presentations rather than earnings calls. Next obvious checkpoint is ASH 2026 for updated follow-up data on the existing cohort and any early read from the key. Broader implication: if the fratricide-resistant editing plus TRAC knockout architecture works for CD7, the same platform applies to other T-cell antigens (CD5 is the most obvious candidate), opening a series of T-cell malignancy programs. The asset is also a plausible acquisition target. Gilead (Kite, owner of Yescarta and Tecartus), Bristol-Myers Squibb (Juno, owner of liso-cel / Breyanzi), Novartis (Kymriah platform), and J&J / Legend Biotech (cilta-cel / Carvykti) all have CAR-T franchises that could absorb a Wugen-style asset and the underlying allogeneic platform. Worth tracking quarterly through next ASH at minimum.