Every year in biopharma brings its share of grueling defeats, and 2025 was no different, especially for companies targeting neurological diseases. Some failures split up partners, and one particularly egregious case even led to the demise of an entire company.
Failure is a fact of life for the biopharma industry: By some estimates, around 90% of all investigational medicines are doomed. This is especially true for assets targeting neurological indications, with around 85% of these assets going down in Phase II/III development.
To survive in the drug development game, companies must be comfortable with the idea of failure. They might pour years of intensive research and clinical work—not to mention millions, if not billions, of dollars—into a candidate, only for it to flounder in clinical development, sending them back to square one.
Still, some trip-ups end up hurting more than others. A particularly bad failure, for instance, can lead a biotech to abandon a drug entirely, or even lay off scores of its employees. Another could break up a long-standing partnership or even push an entire company to closure.
In this piece, BioSpace looks at some of the most costly clinical stumbles of 2025—and takes stock of the fallout.
TIGIT Troubles Push iTeos to Closure
In May, iTeos Therapeutics suffered a clinical defeat that would ultimately prove fatal for its business.
Under a 2021 contract, the former Massachusetts biotech was working with pharma giant GSK on belrestotug. Formerly in development for lung and head and neck cancers, belrestotug is a monoclonal antibody that binds to the TIGIT protein, which has a reputation as a fraught cancer target.
Several companies have tried and failed to leverage the TIGIT pathway to treat cancer, including Roche, which last year famously suffered back-to-back late-stage losses with tiragolumab. Other industry giants—Merck and Bristol Myers Squibb among them—have likewise been felled by TIGIT.
Still, iTeos and GSK stuck with belrestotug—and for good reason. At last year’s meeting of the European Society for Medical Oncology, the partners re-energized the field with mid-stage data from the GALAXIES Lung-201 trial, touting an approximately 60% confirmed objective response rate in patients with non-small cell lung cancer (NSCLC). Belrestotug also suppressed circulating tumor DNA levels by as much as 97% at seven weeks.
This optimism wouldn’t last, however. Follow-on data from the same study, released in May, showed that belrestotug had no clinically meaningful improvements in terms of progression-free survival, a key secondary outcome. Data from a parallel mid-stage study in head and neck cancer were also underwhelming, with response rates failing to show meaningful benefit.
These subpar results led GSK to bail on the partnership, prompting iTeos to enter survival mode and look for strategic alternatives for its business.
Just two weeks later, the beleaguered biotech decided to wind down operations—that is, until cleanup company Concentra Biosciences came knocking in July. At $10.047 per share, Concentra’s offer was a slight discount to iTeos valuation at the time, but given its precarious financial and pipeline position, the company’s board unanimously agreed to the buyout.
GSK, Alector Can Dementia Drug After Late-Stage Fail
In many ways, Alector is in a similar situation to iTeos: The companies once shared a powerhouse partner in GSK and earlier this year, California-based Alector also suffered a critical trial failure. The consequences for Alector, however, haven’t been as ruinous.
Alector and GSK also linked up in July 2021, with the big pharma fronting $700 million and promising up to $1.5 billion in milestones to collaborate on two neuro assets. One of these was AL001, a monoclonal antibody that would later be named latozinemab. The drug works by regulating immune activity in the brain by modulating the concentration of a protein called progranulin.
The companies sought to leverage this mechanism of action against frontotemporal dementia (FTD) and together ran the Phase III INFRONT-3 study in approximately 110 patients with or who were at risk of FTD linked to progranulin mutations. Results reported in October, however, were not what they expected.
While latozinemab indeed had a “significant effect” on progranulin levels, GSK and Alector reported at the time, it failed to significantly slow the progression of FTD versus placebo. Latozinemab also had no apparent effect on MRI scans and other fluid biomarkers, key secondary endpoints of INFRONT-3.
This disappointing mid-stage showing pushed the partners to abandon latozinemab entirely—and Alector let go of nearly half its workforce in a bid to conserve cash and “ensure continued progress across its portfolio,” according to an Oct. 22 press release.
This follows the 17% of employees cut by Alector following the high-profile failure of its AbbVie-partnered antibody AL002 in Alzheimer’s in November 2024.
But unlike iTeos, Alector still maintains its partnership with GSK. The second asset under their 2021 agreement, dubbed nivisnebart, is in Phase II development for early Alzheimer’s disease. The drug works similarly to latozinemab and acts on progranulin, increasing its concentrations to address cognitive decline.
Enrollment for nivisnebart’s mid-stage trial was completed in April and an independent interim analysis is expected in the first half of 2026.
Novo’s GLP-1 Flops in Alzheimer’s, Adding Insult to Injurious 2025
Also failing to make significant headway in the neuro space was Novo Nordisk. Last month, the obesity titan sustained two late-stage failures for its blockbuster GLP-1 drug semaglutide in Alzheimer’s disease.
In the Phase III EVOKE and EVOKE+ studies, semaglutide treatment led to improvements in disease biomarkers but did not induce clinically meaningful effects against Alzheimer’s. Novo treated more than 3,800 patients across the two trials, comparing oral semaglutide against placebo on top of standard care.
Expectations for the EVOKE program were low—but hopeful. Reacting to the readout in a Nov. 24 note, analysts at BMO Capital Markets called the studies a “higher risk, high reward opportunity” for Novo, especially given the “limited understanding of Alzheimer’s disease etiology and how GLP-1s could play a role in slowing progression.”
Still, the failure bodes poorly for the Danish pharma, which has had a difficult year marked by steep layoffs, shrinking Wegovy market share, a CEO change, a board takeover and, most recently, heavy investor skepticism over its contentious buyout offer for obesity up-and-comer Metsera. Year-to-date, Novo’s shares have lost more than 50% of their value.
Beyond its impacts on Novo’s business, the EVOKE failures last month also hint at the limits of the GLP-1 drug class, which has shown therapeutic promise in other diseases, such as liver and kidney diseases, cancer, addiction and neurodegeneration.
Indeed, evidence had been building in recent months pointing to the potential of GLP-1s for various neurodegenerative disesease. A retrospective cohort analysis in July this year, for instance, showed that patients on GLP-1 treatments were 37% less likely to develop dementia, as compared with other antidiabetic drugs. In December 2024, back-to-back papers found signals of neuroprotection with semaglutide as well as liraglutide, Novo’s older-generation GLP-1.
It is unclear how the EVOKE failures will change the industry’s approach and future appetite for testing GLP-1s against neurodegenerative diseases.
Schizophrenia Stumble Casts Clouds over BMS’s Cobenfy
Neuro proved to be a tough customer for biopharma this year. In April, Bristol Myers Squibb’s recently approved schizophrenia drug Cobenfy failed the Phase III ARISE trial, unable to significantly lower symptom severity when combined with atypical antipsychotics.
Reacting to the late-stage stumble, analysts at Leerink Partners cast some doubt on Cobenfy’s overall clinical profile. The drug, analysts told investors in an Apr. 22 note, could have “far less potential than we originally anticipated.” The underwhelming ARISE readout pushed Leerink to lower its long-term earnings forecast for Cobenfy to $2.6 billion by 2030, a 55% cut from its previous $5.8 billion estimate.
Taken orally, Cobenfy combines the muscarinic agonist xanomeline, which targets the M1 and M4 muscarinic receptors to address schizophrenia symptoms, and the muscarinic antagonist trospium, which helps manage side effects. The drug’s approval in September last year was hailed as practice-changing, marking the first new treatment class for schizophrenia in more than 30 years.
With ARISE, BMS was looking to open up the adjunctive use of Cobenfy alongside atypical antipsychotics in patients who had shown inadequate response to their current antipsychotic regimen. As of the end of July, however, Cobenfy’s prospects in this indication remain in limbo: “No conversations have occurred from the health authority perspective,” Chief Medical Officer Samit Hirawat told investors during the pharma’s Q2 earnings call.
Meanwhile, investors and analysts are eagerly anticipating the upcoming readout from the Phase III ADEPT-2 study, which is testing Cobenfy for the treatment of psychosis in patients with Alzheimer’s disease. According to a federal clinical trials database, the study ended on July 31, 2025, and analysts had been expecting a readout imminently.
However, on Dec. 3, BMS announced that “irregularities” had been detected linked to the trial’s execution “at a small number of study sites,” necessitating the recruitment of additional patients into the trial.
Data from ADEPT-2, as well as the late-stage ADEPT-1 and ADEPT-4 studies, also being run in Alzheimer’s disease psychosis, are now expected by the end of 2026.
Pfizer’s Sickle Cell Therapy Flunks Phase III, Another GBT Disappointment
Pfizer just hasn’t had any fun in clinical development this year. One major slip came in August, when the pharma’s investigational antibody inclacumab failed a late-stage study in sickle cell disease.
Pfizer was testing inclacumab in the Phase III THRIVE-131 study, which enrolled more than 240 patents who had suffered from two to 10 vaso-occlusive crises (VOC) in the previous year. THRIVE-131 compared inclacumab, a P-selectin blocker, against placebo, where the candidate failed to show significant VOC benefits over the 48-week trial.
The pharma did not provide specific data for THRIVE-131 at the time, only noting that inclacumab “did not meet our expectations.” Pfizer’s plans for inclacumab moving forward remain unclear, though the company announced a $260 million impairment charge in connection with the trial’s failure in its Q3 earnings report.
Perhaps more importantly for Pfizer, however, is that this late-stage blunder casts more doubt on its business development acumen. Inclacumab came to Pfizer through the acquisition of Global Blood Therapeutics (GBT) for $5.4 billion in August 2022.
Over the years, however, this deal hasn’t played out well for the New York pharma. In March 2024, Pfizer pulled the plug on two late-stage inclacumab studies, citing “poor accrual and associated recruitment challenges.” A few months later, in September that year, Pfizer pulled the sickle cell drug Oxbryta—the centerpiece of the GBT deal—from the market due to new safety data showing an excess of VOC episodes and risk of death.
The GBT acquisition proved to be so ill-conceived for Pfizer that activist investor Starboard Value called it a “failed” deal in an October 2024 presentation, claiming that the takeover “shocked the industry and raised serious questions about [Pfizer’s] BD capabilities.” Starboard carved out a $1 billion stake in the pharma that month with the goal of turning the business around. The firm sold its Pfizer stake last month.
