Galinpepimut-S
A four-peptide cancer vaccine aimed at the WT1 protein, and the first therapeutic cancer vaccine in a decade to meet a survival endpoint in a randomised phase 3 trial. In REGAL, reported on 17 November 2026, patients with acute myeloid leukaemia in second remission lived a median 24.0 months against 14.7 on best available treatment.
Galinpepimut-S is a cancer vaccine made of four short fragments of a protein called WT1, which leukaemia cells make in large amounts and healthy blood cells barely make at all. The idea is to teach the immune system to recognise that protein and kill the cells carrying it. On 17 November 2026 its maker, SELLAS, reported that the phase 3 REGAL trial hit its goal: 126 patients with acute myeloid leukaemia in a second remission who could not have a stem-cell transplant lived a median 24.0 months on the vaccine against 14.7 months on whatever their doctor would otherwise have given them, a 41% reduction in the risk of death (hazard ratio 0.59, p=0.035). Side effects were mostly sore injection sites. Read it with the right caveats: this is a company press release, not a published paper, and the release did not give a confidence interval around that hazard ratio. The trial was open-label, it is small for a phase 3, and the p-value of 0.035 sits close enough to 0.05 that the result is real but not overwhelming. The company plans to file with the FDA in the first half of 2027. Nothing is approved anywhere, and the vaccine has no use outside a trial.
What is Galinpepimut-S?
Galinpepimut-S is a therapeutic cancer vaccine built from four peptides derived from WT1, the Wilms tumour 1 protein. It is given as a subcutaneous injection alongside two immune adjuvants, and it is designed to provoke a T-cell response against cells that display WT1 fragments on their surface. It is investigational: not approved by the FDA, the EMA or any other regulator. WT1 was chosen for a specific reason. In 2009 a National Cancer Institute consensus exercise ranked candidate cancer antigens by how well they met criteria such as therapeutic function, immunogenicity, specificity and expression level, and placed WT1 first out of 75. The protein is strongly overexpressed in acute myeloid leukaemia, mesothelioma, ovarian cancer and multiple myeloma, and is close to absent in normal adult tissue outside a few sites. That combination — present on the tumour, nearly absent on everything else — is what a vaccine target needs. The difficulty with any self-antigen is tolerance: the immune system has already been trained not to attack it. Galinpepimut-S addresses that with heteroclitic peptides, meaning peptides carrying a deliberate single amino-acid substitution that makes them bind the presenting HLA molecule more tightly than the natural sequence, and so provoke a stronger T-cell response — with the intention that the T cells raised against the altered peptide still recognise the natural one on the tumour. Two of the four peptides are heteroclitic and two are the native sequence. The four components are WT1-A1, a nine-residue CD8 epitope with the sequence YMFPNAPYL covering WT1 residues 126 to 134 and carrying an arginine-to-tyrosine substitution at position 126; WT1-122A1 long, a longer heteroclitic peptide that contains the A1 sequence within it and stimulates both CD4 and CD8 responses; and WT1-331 long and WT1-427 long, two native long peptides directed at CD4 helper responses. Because the long peptides are processed and presented by a range of HLA molecules, the vaccine is described as non-HLA-restricted — patients are not screened by HLA type to receive it, which is a practical advantage over single-epitope vaccines. The vaccine originated at Memorial Sloan Kettering Cancer Center, where David A. Scheinberg is named as its inventor, and has been developed clinically by SELLAS Life Sciences Group.
What Galinpepimut-S Is Investigated For
Everything that matters about galinpepimut-S right now comes from one trial, with a long tail of small supporting studies behind it. REGAL is the trial. It enrolled adults with acute myeloid leukaemia who had reached a second or later complete remission — or complete remission with insufficient platelet recovery — after second-line or later salvage therapy, and who were ineligible for or unable to undergo allogeneic stem-cell transplantation. That is a population with no approved maintenance option and a short expected survival. 126 patients were randomised 1:1 to galinpepimut-S or investigator's choice of best available treatment, which could be observation, a hypomethylating agent, venetoclax, low-dose cytarabine, or a combination. The topline reported on 17 November 2026 gave a median overall survival of 24.0 months on the vaccine against 14.7 months on best available treatment, a hazard ratio of 0.59 and a two-sided p-value of 0.035. The analysis fired at a pre-specified 80 deaths. Treatment-related adverse events were mostly grade 1 or 2, led by local injection-site reactions, and patients who went on to transplant were few and evenly split between the arms. Four caveats belong next to that result rather than underneath it. The data are a company press release; the full dataset has not been published or peer-reviewed, and no confidence interval was given for the hazard ratio. The trial was open-label, which overall survival tolerates better than most endpoints but does not render irrelevant, since the control arm included observation and treatment decisions downstream were unblinded. 126 patients is small for a registrational trial, and the p-value of 0.035 against a 0.05 threshold describes a result that is positive but not robust to much. And a median survival difference of 9.3 months in a disease this aggressive is a large clinical claim resting on a modest statistical margin. The other indications are earlier and considerably weaker. In malignant pleural mesothelioma, a randomised phase 2 trial after surgery and a second modality enrolled 41 of a planned 78 patients before the control arm closed for futility and the vaccine arm closed because the resulting unblinding made it uninterpretable; one-year progression-free survival was 45% on vaccine against 33% on control, with median overall survival 22.8 against 18.3 months, in a trial explicitly not powered to compare the arms. In ovarian cancer, an 11-patient phase 1 combining the vaccine with nivolumab produced T-cell responses in 10 of 11 patients and a one-year progression-free survival of 70% among evaluable patients who received more than two treatments. A matching 10-patient mesothelioma study with nivolumab was the weakest of the set: only 3 of 10 patients mounted a vaccine-specific T-cell response, there were no partial responses, and median overall survival was 7.4 months. A phase 1/2 study in multiple myeloma after autologous transplant and a 26-patient combination study with pembrolizumab round out the programme. There is no circumstance in which this is a compound an individual obtains or uses outside a clinical trial. It is an injected biologic given with two adjuvants on a fixed schedule under oncology supervision, and SELLAS operates an expanded-access programme for AML and MDS patients who qualify.
History & Discovery
The vaccine's lineage runs from a basic-immunology problem to a phase 3 result over roughly twenty years, and the turning points are all about tolerance. WT1 had been recognised as a leukaemia-associated antigen since the 1990s, but the obstacle was that it is a self-protein: immunising against it produced weak responses because the strongly reactive T cells had already been deleted. The solution came from Memorial Sloan Kettering Cancer Center, where work published in Leukemia in 2006 by Javier Pinilla-Ibarz, Randall May, Tatyana Korontsvit and colleagues in David Scheinberg's group showed that WT1 nonamers carrying single substitutions at HLA anchor positions stabilised MHC class I better than native sequences, elicited stronger T-cell responses, and — the critical point — produced T cells that still recognised the native peptide and killed HLA-matched leukaemia cells. In 2009 a National Cancer Institute consensus exercise ranked WT1 first among 75 candidate cancer antigens, which did a great deal to attract development interest to the target. The formulation that became galinpepimut-S combined that heteroclitic CD8 epitope with long peptides aimed at CD4 help, deliberately avoiding the HLA restriction that limits single-epitope vaccines. Scheinberg is named as its inventor. Memorial Sloan Kettering ran the early trials — a pilot study in AML, then the 22-patient phase 2 in first remission published in Blood Advances in 2018, and the randomised mesothelioma phase 2 published in Clinical Cancer Research in 2017. SELLAS Life Sciences took the programme forward commercially, arriving on the public markets through a 2017 reverse merger with Galena Biopharma. The FDA granted Fast Track designation in AML and, in September 2016, in malignant pleural mesothelioma; orphan drug designations followed from both the FDA and the EMA in AML, mesothelioma and multiple myeloma. REGAL opened in February 2021 and read out on 17 November 2026 — a positive overall-survival result in a field where therapeutic cancer vaccines have failed repeatedly and publicly for three decades. Whether it becomes an approved product depends on data that have not yet been shown to anyone outside the company and the regulators.
How It Works
Leukaemia cells make large amounts of a protein called WT1 and chop pieces of it onto their surface, where the immune system can in principle see them. Normal cells make almost none. Galinpepimut-S injects four of those pieces, slightly modified to be more noticeable, mixed into an oily emulsion that holds them at the injection site and a growth factor that draws immune cells in. The aim is to raise an army of T cells trained on WT1, which then circulate and kill any cell displaying it — in practice, the leukaemia cells left behind after chemotherapy has put the disease into remission.
The vaccine is a mixture of four synthetic WT1-derived peptides delivered with two adjuvants, and its design reflects three separate problems in cancer vaccinology. The first is tolerance. WT1 is a self-antigen, so the high-avidity T cells that would respond to it have largely been purged. The heteroclitic strategy, established for WT1 in work published in Leukemia in 2006, introduces single amino-acid substitutions at HLA-anchor positions. The modified peptide binds the presenting MHC molecule more stably, pushing the response above the threshold that tolerance would otherwise suppress, and the T cells raised are selected to cross-react with the native sequence displayed by the tumour. WT1-A1, the vaccine's CD8 epitope, is YMFPNAPYL — WT1 residues 126 to 134 with arginine replaced by tyrosine at position 126. The second is HLA restriction. A nine-residue CD8 epitope is presented by one specific HLA allele, so a vaccine built from one would only be usable in patients carrying it. Galinpepimut-S includes two native long peptides, WT1-331 and WT1-427, and one long heteroclitic peptide, WT1-122A1, which contains the A1 sequence embedded in it. Long peptides are taken up by antigen-presenting cells and processed internally, so they can be presented on a range of class I and class II molecules. This is why the vaccine is described as non-HLA-restricted and why patients are not HLA-screened before receiving it — and it is the reason the formulation targets CD4 helper responses alongside CD8 killing, since a CD8 response without CD4 help tends to be short-lived. The third is getting an immune response to a peptide at all. Short synthetic peptides injected alone are cleared rapidly and are poorly immunogenic. The peptides are emulsified one-to-one with Montanide ISA 51 VG, a water-in-oil adjuvant that forms a depot releasing antigen over days, and sargramostim (recombinant GM-CSF) is injected at the same site two days before and on the day of each vaccination to recruit and mature dendritic cells there. A structural caveat on the heteroclitic design emerged in 2026. The R1Y substitution raised the melting temperature of the peptide-HLA complex from 48.9 to 54.0 °C — a real stabilisation — but weakened recognition by the naturally selected C4 T-cell receptor from a dissociation constant of 13.8 micromolar to worse than 150 micromolar, because the substituted tyrosine reorients on engagement and destroys salt bridges at the interface. Greater HLA stability and better T-cell recognition are separable properties, and in this case they moved in opposite directions for one well-characterised receptor. Whatever clinical activity the vaccine has is therefore likely to run through T-cell clones other than that one.
Evidence Snapshot
Human Clinical Evidence
One positive randomised phase 3 trial, reported in topline form only. REGAL (NCT04229979) randomised 126 patients with AML in second or later complete remission, ineligible for transplant, to galinpepimut-S or investigator's choice of best available treatment, and reported median overall survival of 24.0 versus 14.7 months, hazard ratio 0.59, p=0.035, at a pre-specified 80 deaths. Behind it sit small studies: a 22-patient phase 2 in AML first remission with median disease-free survival of 16.9 months and an immune response in 9 of 14 tested patients; a 41-patient randomised phase 2 in mesothelioma that closed early and was not powered for comparison; an 11-patient ovarian phase 1 and a 10-patient mesothelioma phase 1 with nivolumab; and a 20-patient study in multiple myeloma after transplant. Total clinical experience is a few hundred patients.
Animal / Preclinical
The preclinical foundation is immunological rather than animal-efficacy based. The key experiments were human in vitro work: analog WT1 peptides stabilised HLA-A*02:01 better than native sequences and elicited T cells that killed HLA-matched leukaemia cell lines, and later work extended the epitope set to peptides active in the context of both HLA-A*02:01 and HLA-A*24:02, broadening potential population coverage. Structural and biophysical characterisation of the peptide-HLA-TCR interface has continued through 2026 and has complicated as much as it has confirmed the heteroclitic rationale.
Mechanistic Rationale
Strong on target choice, reasonable on mechanism. WT1 was ranked the highest-priority cancer antigen in a formal National Cancer Institute consensus exercise of 75 candidates, and its expression pattern — high in AML blasts, near-absent in normal haematopoietic cells — is close to ideal for an antigen-directed therapy. The multivalent, non-HLA-restricted, CD4-plus-CD8 design addresses the standard failure modes of peptide vaccines in a considered way. What remains unresolved is whether the measured immune responses are what produce the survival effect; no trial has yet linked immunological response to outcome.
Research Gaps & Open Questions
What the current literature has not yet settled about Galinpepimut-S:
- 01The REGAL dataset is unpublished. Until the full analysis appears in a peer-reviewed journal or at a major meeting, no one outside SELLAS and the regulators can evaluate the confidence interval around the hazard ratio, the survival curves' shape, the balance of post-progression therapy between arms, or the complete safety table.
- 02No trial has linked immune response to clinical benefit. The phase 2 AML study found a response in 9 of 14 tested patients, but no study has been powered to ask whether the responders are the ones who live longer — which is the question that would establish the mechanism rather than assume it.
- 03The 2026 structural work showing that the heteroclitic R1Y substitution weakens recognition by a clinically relevant T-cell receptor raises an unanswered question about which T-cell clonotypes actually mediate any benefit, and whether a different epitope design would do better.
- 04Durability is unknown. The median survival difference was measured at a single analysis triggered by 80 deaths; whether the curves continue to separate, converge, or cross has not been reported.
- 05The optimal duration of maintenance has never been studied. Trial schedules continued until relapse, but whether a fixed course would work as well, or whether immune responses wane with repeated dosing, is untested.
- 06Nothing is known about the vaccine in AML patients with specific molecular subtypes — FLT3, NPM1, TP53 and the rest — even though those drive prognosis and increasingly drive therapy choice.
- 07The checkpoint-inhibitor combination produced a 10-of-11 T-cell response rate in ovarian cancer and a 3-of-10 rate in mesothelioma. Why the difference, and whether it reflects the tumour, the prior therapy or chance in two ten-patient studies, is unresolved.
Forms & Administration
Galinpepimut-S is given as a subcutaneous injection, with sites rotated between extremities. Immediately before administration the four peptides are mixed one-to-one with Montanide ISA 51 VG to form a water-in-oil emulsion. Sargramostim (GM-CSF) 70 micrograms is injected subcutaneously at the planned vaccination sites on day minus two and day zero of each dose, to prime the site with dendritic cells before the antigen arrives. The schedule used across the trials is a loading phase followed by maintenance. In the phase 2 AML study, patients received six vaccinations every two weeks over ten weeks, with up to six further monthly doses if they remained in remission. REGAL used the same shape: six injections every two weeks through week ten, a four-week rest, then continued dosing until relapse. There is no oral, intranasal or topical form, and the preparation requirements — a multi-peptide emulsion made up at the point of care, with a separate recombinant growth factor given at the same sites on a staggered schedule — place it firmly outside anything that could be self-administered.
Common Questions
Did galinpepimut-S actually work in its phase 3 trial?
On the trial's own terms, yes. REGAL met its primary endpoint of overall survival: a median of 24.0 months on galinpepimut-S against 14.7 months on best available treatment, hazard ratio 0.59, two-sided p=0.035, in 126 patients with AML in second or later remission who could not have a transplant. That is a 41% reduction in the risk of death over the analysis period. What should temper enthusiasm is the form the result arrived in rather than its direction. It is a press release issued on 17 November 2026, not a paper; the company did not publish a confidence interval around the hazard ratio, which for a trial this size would be wide; the study was open-label; and p=0.035 is a pass, not a comfortable one. SELLAS says it will meet the FDA to discuss an accelerated filing and is preparing a biologics licence application for the first half of 2027. Until the full dataset is presented and reviewed, the honest description is a positive topline from a small registrational trial.
What is WT1 and why target it?
WT1 — Wilms tumour 1 — is a transcription factor that is strongly overexpressed in acute myeloid leukaemia, mesothelioma, ovarian cancer and several other tumours, while being almost absent from normal adult tissue. A 2009 National Cancer Institute consensus exercise that ranked 75 candidate cancer antigens against criteria including therapeutic function, immunogenicity and specificity placed WT1 at the top of the list. The practical consequence is that a T-cell response directed at WT1 should hit tumour cells and largely spare healthy ones. Several different modalities target it — peptide vaccines like this one, engineered T-cell receptors, and TCR-mimic antibodies — and galinpepimut-S is the furthest advanced.
What does “heteroclitic” mean here?
It means the peptide has been deliberately mutated to be a better immune stimulus than the natural sequence. WT1 is a self-protein, so the T cells that would respond strongly to it have mostly been deleted or suppressed during development. Substituting a single amino acid at a position that anchors the peptide into the HLA groove makes the peptide-HLA complex more stable and the resulting T-cell response stronger — and the hope is that those T cells then cross-react with the unmodified peptide displayed by the tumour. In galinpepimut-S, two of the four peptides carry an arginine-to-tyrosine substitution at WT1 position 126; the other two are native sequences aimed at CD4 helper responses. The strategy dates to work published in Leukemia in 2006 showing that analog WT1 peptides stimulated T cells more effectively than native ones and that those T cells still killed HLA-matched leukaemia cells.
Is the heteroclitic approach free of downsides?
No, and a 2026 structural study in the Journal of Biological Chemistry made the trade-off explicit. Crystal structures of the clinically studied C4 T-cell receptor bound to HLA-A*02:01 presenting either the native WT1 peptide RMFPNAPYL or its heteroclitic R1Y variant YMFPNAPYL showed that the substitution does what it was designed to do — the variant complex is markedly more thermally stable, with a melting temperature of 54.0 °C against 48.9 °C — but at a cost. The tyrosine reorients when the T-cell receptor engages, eliminating two salt bridges and weakening the interface, so that C4 bound the native complex with a dissociation constant of about 13.8 micromolar and the heteroclitic complex at worse than 150 micromolar. Better HLA binding is not automatically better T-cell recognition. The authors are careful to note this does not preclude the variant eliciting other cross-reactive T-cell clones, which is in effect what the clinical results suggest is happening, but it argues that the two properties need to be evaluated separately rather than assumed to move together.
Can I get galinpepimut-S?
Not as a prescription anywhere in the world, and not through any other route. It is unapproved by the FDA and every other regulator, and it is not the kind of product that exists outside a controlled supply chain: four separate synthetic peptides, emulsified one-to-one with Montanide ISA 51 VG immediately before injection, with sargramostim injected at the same sites two days before and on the day of each dose. The access route that exists is clinical: the REGAL trial, or SELLAS's expanded-access programme for patients with AML or myelodysplastic syndrome who meet its criteria. Any vendor offering a “WT1 peptide vaccine” outside those channels is selling something else.
How does it compare to checkpoint inhibitors?
They do different jobs, and the trials that combined them are instructive. A checkpoint inhibitor removes a brake on T cells that already exist; a vaccine tries to create the T cells in the first place. The combination ought to be complementary, and galinpepimut-S has been studied with both nivolumab and pembrolizumab on that logic. Results were mixed: in ovarian cancer, 10 of 11 patients mounted T-cell responses and one-year progression-free survival among evaluable patients reached 70%, while in mesothelioma only 3 of 10 patients responded immunologically, no partial responses occurred and median overall survival was 7.4 months. Both were single-arm studies of about ten patients each, which is too small to conclude much in either direction. The phase 3 result that exists is for the vaccine alone, as maintenance monotherapy.
Who Galinpepimut-S Is NOT For
- •Active, measurable leukaemia — every trial studied the vaccine as maintenance in patients already in remission, and nothing supports its use against detectable disease
- •Patients eligible for allogeneic stem-cell transplantation, which remains the only potentially curative option in relapsed AML and which the REGAL population was selected for being unable to receive
- •Known hypersensitivity to any vaccine component, including the Montanide ISA 51 VG emulsion or sargramostim
- •Significant active autoimmune disease, on the general principle that applies to all active immunotherapies — the vaccine's trials excluded such patients and its effect on them is unstudied
- •Patients on systemic immunosuppression, including meaningful corticosteroid doses, where a vaccine intended to generate a T-cell response has no rational basis
- •Pregnancy and breastfeeding — no data exist in either setting
Drug & Supplement Interactions
No formal drug-interaction studies have been published. The interactions that matter are pharmacodynamic rather than pharmacokinetic, because the vaccine acts by generating an immune response rather than by circulating as an active drug. Systemic corticosteroids and other immunosuppressants are the clearest concern. Anything that suppresses T-cell priming works directly against the vaccine's mechanism, and the trials were designed around that constraint. Checkpoint inhibitors are the deliberate combination. Nivolumab and pembrolizumab have both been studied with galinpepimut-S on the rationale that a vaccine generates T cells while a checkpoint inhibitor keeps them from being switched off. Results were inconsistent across tumour types, and the combination's toxicity is not the monotherapy's: grade 3 or higher events occurred in 2 of 10 mesothelioma patients and a dose-limiting toxicity in 1 of 11 ovarian cancer patients. In REGAL the comparator arm included hypomethylating agents, venetoclax and low-dose cytarabine, but patients received either the vaccine or those agents rather than both, so the trial generated no data on combining them. Whether vaccination can be layered onto a hypomethylating agent — a plausible pairing, since hypomethylating agents can upregulate tumour antigen expression — is untested. Sargramostim is part of the regimen rather than an interaction, but it is an active biologic in its own right and carries its own profile.
Safety Profile
Common Side Effects
Cautions
- • Investigational everywhere — not approved by the FDA, the EMA or any other regulator, with a biologics licence application planned for the first half of 2027 at the earliest
- • Administered only under oncology supervision as part of a defined regimen with two adjuvants; there is no self-administered form
- • The vaccine is a maintenance therapy studied in patients already in remission, not a treatment for active disease — none of the trials tested it against measurable leukaemia
- • Safety experience outside AML and mesothelioma is limited to studies of ten to twenty-six patients
- • The combination with checkpoint inhibitors produced grade 3 or higher events in 2 of 10 mesothelioma patients and a dose-limiting toxicity in 1 of 11 ovarian cancer patients, so the combination's safety profile is not the monotherapy's
What We Don't Know
The REGAL safety dataset has not been published, so the full adverse-event table by type and grade is not yet available to anyone outside the company and regulators. Longer-term questions are entirely open: whether the survival separation holds beyond the analysis point, whether repeated vaccination produces autoimmune effects against the small number of normal tissues that express WT1 (notably renal podocytes and some haematopoietic progenitors), and whether immune responses measured in the laboratory predict which patients benefit. The phase 2 AML study found an immune response in 9 of 14 tested patients, but no trial has yet been large enough to link that response to survival.
Legal Status
United States
Investigational. Galinpepimut-S is not approved by the FDA for any indication and does not appear in the FDA's approved-drug databases. It holds Fast Track designation in acute myeloid leukaemia and malignant pleural mesothelioma, and orphan drug designation in AML, mesothelioma and multiple myeloma. Following the November 2026 REGAL topline, SELLAS has said it will meet the FDA to discuss an accelerated filing strategy and is preparing a biologics licence application for the first half of 2027. Lawful access in the United States is through a clinical trial or the company's expanded-access programme for AML and myelodysplastic syndrome; there is no other legal route.
International
Not approved in any jurisdiction. The European Medicines Agency has granted orphan designation in AML, mesothelioma and multiple myeloma, which is a development incentive rather than any form of authorisation. REGAL was run as a global trial, so sites existed outside the United States, but availability outside a trial does not follow from that.
Sports & Competition
Not a sports-doping concern in any practical sense. Galinpepimut-S is an antigen-specific cancer vaccine with no anabolic, stimulant or oxygen-carrying properties, no performance rationale, and no plausible route to an athlete outside oncology care. Athletes receiving it would be doing so as cancer patients, in which case the relevant process is a therapeutic use exemption handled with the treating oncologist.
Regulatory status changes over time. Verify current local rules with a qualified professional.
Myths & Misconceptions
Myth
Galinpepimut-S is an approved cancer treatment.
Reality
It is not approved anywhere in the world. A positive phase 3 topline was reported on 17 November 2026 and SELLAS says it plans a biologics licence application in the first half of 2027. Between a topline press release and an approval sit the full dataset, FDA review, and a decision — none of which has happened.
Myth
A hazard ratio of 0.59 means the vaccine cuts your risk of death by 41%.
Reality
It means that over the trial's follow-up, in this population, the rate of death in the vaccine arm was about 59% of the rate in the control arm. That is a real and meaningful result, but it is a group-level rate ratio measured at one analysis point, not an individual's risk reduction, and the company did not release the confidence interval. In a 126-patient trial analysed at 80 deaths, that interval would be wide — wide enough that the plausible range of true effects runs from substantial to slight.
Myth
Cancer vaccines do not work, so this result is probably noise.
Reality
The scepticism is earned — therapeutic cancer vaccines have a long record of failure, and plenty of positive phase 2 results have evaporated in phase 3. But the right response to a prior is to examine the new evidence, not to dismiss it. REGAL was randomised, had overall survival as its primary endpoint, used a pre-specified analysis trigger, and met its endpoint. The appropriate caution concerns the size of the trial, the open-label design and the absence of published data, not the premise that this class can never work.
Myth
It works by stimulating the immune system generally, like an immune booster.
Reality
It does the opposite of general stimulation. The whole design is about specificity: four peptides from one protein, selected so that the T cells raised recognise cells displaying WT1 and ignore everything else. The adjuvants — a Montanide emulsion and GM-CSF at the injection site — are local, intended to make the antigen visible to dendritic cells at that spot rather than to raise systemic immune activity.
Myth
You could get the same effect from a WT1 peptide bought from a supplier.
Reality
No. The vaccine's activity depends on four specific peptides, two of them carrying deliberate substitutions, emulsified one-to-one with a specific water-in-oil adjuvant immediately before injection, with GM-CSF given at the same sites on a staggered schedule. A raw peptide injected without that machinery is cleared within minutes and is close to non-immunogenic — which is exactly the problem the formulation was built to solve.
Published Research
15 studiesEnhanced HLA stability of a heteroclitic WT1 vaccine exacts a kinetic penalty in recognition by a natural TCR.
Yang, Wei and Wu, Journal of Biological Chemistry, September 2026 — the most important recent paper on this vaccine, and a genuine complication to its design logic. Crystal structures of the clinically studied C4 T-cell receptor bound to HLA-A*02:01 presenting either native RMFPNAPYL or the heteroclitic R1Y variant YMFPNAPYL show that the substitution stabilises the peptide-HLA complex substantially, raising its melting temperature from 48.9 to 54.0 °C, while weakening C4 binding from a dissociation constant of 13.8 micromolar to worse than 150 micromolar, because the substituted tyrosine reorients on engagement and destroys salt bridges at the interface. HLA stabilisation and T-cell recognition are separable and here they moved in opposite directions. The authors are explicit that this does not preclude the variant eliciting other cross-reactive clonotypes, but it undercuts the assumption that a tighter-binding peptide is a better vaccine peptide.
Combining a WT1 Vaccine (Galinpepimut-S) With Checkpoint Inhibition (Nivolumab) in Patients With WT1-Expressing Diffuse Pleural Mesothelioma: A Phase 1 Study.
Agrawal, Offin, Lai and colleagues, JTO Clinical and Research Reports, January 2025 — the negative study in the set, and the reason the checkpoint-combination story should not be told only through the ovarian data. Ten patients with previously treated diffuse pleural mesothelioma received two doses of the vaccine followed by six doses with nivolumab. Only 3 of 10 mounted a vaccine-specific T-cell response, there were no partial responses, three patients had prolonged stable disease with at most a 17% reduction in tumour volume, and median progression-free and overall survival were 3.9 and 7.4 months. Two patients had grade 3 or higher adverse events. The authors attribute the limited benefit partly to sample size, which is fair, but the immunological response rate is itself much lower than in the other trials.
REGAL: galinpepimut-S vs. best available therapy as maintenance therapy for acute myeloid leukemia in second remission.
Phase I Study of a Multivalent WT1 Peptide Vaccine (Galinpepimut-S) in Combination with Nivolumab in Patients with WT1-Expressing Ovarian Cancer in Second or Third Remission.
Galinpepimut-S (GPS): an investigational agent for the treatment of acute myeloid leukemia.
Phase 2 trial of a multivalent WT1 peptide vaccine (galinpepimut-S) in acute myeloid leukemia.
Maslak, Dao, Bernal and colleagues, Blood Advances, February 2018 — the study that justified taking the vaccine into phase 3, and the source of the dosing schedule and adjuvant regimen still used. 22 adults with AML in first complete remission, median age 64, received six vaccinations over ten weeks with up to six monthly doses after. 14 completed six or more doses and 9 received all twelve. Median disease-free survival from first remission was 16.9 months; median overall survival from diagnosis had not been reached and was estimated at 67.6 months or more. Nine of 14 tested patients showed an immune response by CD4 proliferation, CD8 interferon-gamma ELISPOT or tetramer assay. Toxicity was grade 1/2 injection-site reactions (46%), fatigue (32%) and skin induration (32%). The paper discloses that D.A. Scheinberg is the vaccine's inventor and a consultant to SELLAS, and that two authors are SELLAS employees.
A Randomized Phase II Trial of Adjuvant Galinpepimut-S, WT-1 Analogue Peptide Vaccine, After Multimodality Therapy for Patients with Malignant Pleural Mesothelioma.
Zauderer, Tsao, Dao and colleagues, Clinical Cancer Research, December 2017 — the mesothelioma trial, and a useful lesson in reading a positive-looking result correctly. 41 of a planned 78 patients were randomised after surgery and a second treatment modality to vaccine plus GM-CSF and Montanide or to the adjuvants alone. One-year progression-free survival was 45% on vaccine against 33% on control, median progression-free survival 10.1 against 7.4 months and median overall survival 22.8 against 18.3 months. Every one of those comparisons favours the vaccine and none of them means what it appears to: the control arm closed early on a pre-specified futility rule, the vaccine arm then closed because the unblinding made it uninterpretable, and the authors state plainly that the study was neither designed nor powered for between-arm comparison.
An immunogenic WT1-derived peptide that induces T cell response in the context of HLA-A*02:01 and HLA-A*24:02 molecules.
Improved human T-cell responses against synthetic HLA-0201 analog peptides derived from the WT1 oncoprotein.
Pinilla-Ibarz, May, Korontsvit and colleagues, Leukemia, November 2006 — the founding paper for the heteroclitic strategy this vaccine is built on. Single amino-acid substitutions at HLA-A*02:01 anchor positions in WT1 nonamers produced peptides that stabilised MHC class I better than the native sequences and elicited stronger T-cell responses. The result that mattered for drug development was the last one: T cells stimulated with the analogs cross-reacted with the native WT1 sequence and killed HLA-matched chronic myeloid leukaemia cell lines. Without that cross-reactivity the whole approach would raise T cells against a peptide no tumour displays.
SELLAS Life Sciences Announces Positive Pivotal Phase 3 REGAL Trial Data Showing Galinpepimut-S (GPS) Demonstrated a Statistically Significant and Clinically Meaningful Improvement in Overall Survival in Patients with Acute Myeloid Leukemia
SELLAS Life Sciences, 17 November 2026 — the source for every REGAL figure on this page, and the only source that exists for them. 126 patients randomised 1:1 to galinpepimut-S or investigator's choice of best available treatment (observation, hypomethylating agent, venetoclax, low-dose cytarabine, or a combination), in AML in second or later complete remission after second-line or later salvage, ineligible for allogeneic transplant. Median overall survival 24.0 versus 14.7 months, hazard ratio 0.59, two-sided p=0.035, at a pre-specified 80 deaths. Treatment-related adverse events predominantly grade 1/2, led by injection-site reactions; transplant recipients few and evenly split. No confidence interval was disclosed, and secondary endpoints are described as still being analysed. The ClinicalTrials.gov record lists 127 patients as actual enrolment against the company's 126 randomised, the usual gap between enrolled and randomised. SELLAS announced a proposed $150 million public offering the same day, which is context a reader should have when weighing the framing rather than the numbers.
Galinpepimut-S Versus Investigator's Choice of Best Available Therapy for Maintenance in AML CR2/CRp2 (REGAL)
Expanded Access Program (EAP) for Galinpepimut-S (GPS) in Patients Diagnosed With AML or MDS
Galinpepimut-S in Combination With Pembrolizumab in Patients With Selected Advanced Cancers
WT1 Vaccine Treatment of Patients With Multiple Myeloma After Autologous Stem Cell Transplantation
SELLAS Life Sciences Receives FDA Fast Track Designation of Galinpepimut-S for the Treatment of Malignant Pleural Mesothelioma
Quick Facts
- Class
- Multivalent Heteroclitic Peptide Cancer Vaccine
- Tier
- B
- Evidence
- Moderate
- Safety
- Moderate Data
- Updated
- Nov 2026
- Citations
- 15PubMed
Also known as
Tags
Evidence Score
Clinical Trials
View Clinical TrialsLinks to ClinicalTrials.gov for reference. Listing does not imply endorsement.