Novartis has halted development of VHB937 in amyotrophic lateral sclerosis after the Phase 2 ASTRALS trial failed to meet its primary or secondary endpoints, the company confirmed this week. The trial enrolled 251 people with early-stage ALS, and the failure is the latest in a long line for a disease that has spent three decades confounding every mechanism the field has tried.
The loss is also a specific blow to one of the field's most carefully constructed hypotheses. VHB937 targets TREM2, a receptor on the brain's immune cells, the microglia, that is thought to control how those cells clear debris and regulate inflammation. The drug is designed to push microglia toward a protective state, calming the immune-driven damage that accompanies motor neuron death. In a Phase 1 study of 64 healthy volunteers, the drug did what it was designed to do: it raised the TREM2 marker in cerebrospinal fluid and lowered pro-inflammatory signals. The biology moved. The patients did not.
The trial that ended the program
ASTRALS was a randomized, double-blind, placebo-controlled study of 251 people whose ALS symptoms began within 24 months, randomized two to one to VHB937 or placebo. The primary endpoint combined survival without permanent assisted ventilation with change on the ALSFRS-R, the standard functional scale for ALS, analyzed as a joint measure of function and survival. Secondary endpoints included breathing capacity, a blood marker of nerve damage called neurofilament light, and survival.
Neither the primary nor the secondary endpoints favored the drug. The details will be presented at the International Symposium on ALS and Motor Neuron Disease in Amsterdam in December, and the company said it will support enrolled patients in completing the trial. The program's end for ALS does not end the molecule: Novartis is continuing a separate Phase 2 trial of VHB937 in early Alzheimer's disease, where the company says TREM2 target validation is stronger. The antibody is not dead. Its ALS hypothesis is.
The biomarker trap, again
The trial's failure pattern is one the field has seen before, and it deserves to be named: biomarkers moved in the direction the hypothesis predicted, and clinical endpoints did not follow. The CSF TREM2 signal rose, the inflammatory markers fell, and the disease kept progressing. The gap between the biology and the outcome is the hard lesson of neuroinflammatory drug development, and TREM2 has now taught it to more than one company.
Alector and AbbVie's TREM2 activator failed in Alzheimer's disease in 2024. Vigil Neuroscience's TREM2 antibody also failed. The class has now produced a consistent pattern: the microglial hypothesis is biologically real, the drugs engage the target, and the clinical benefit has not materialized in either of the two diseases where it was most ardently pursued. Novartis's decision to keep the Alzheimer's trial running while abandoning ALS suggests the company reads the pattern the same way everyone else does: the hypothesis is not disproven everywhere, but it has not proven itself anywhere.
For ALS specifically, the failure joins a graveyard that is unusually full even by neurology's standards. The disease has no curative therapy, and the approved treatments slow progression modestly at best. Every mechanistic bet of the past decade, antisense approaches, cell therapies, anti-inflammatories, has failed or is still unproven, while the disease itself remains relentless. Each failure is worse than the last not because the science was bad, but because the patients who enrolled in the failed trial are the same patients who have nowhere else to go.
A fourth stumble in a month
The ALS announcement is Novartis's fourth clinical setback in under a month, a streak that has changed the tone around the company's pipeline. The company paused eight of ten trials of its rap-cel cell therapy after three patient deaths. Its Lp(a)-lowering drug pelacarsen missed in a cardiovascular outcomes trial. Its muscle-wasting therapy del-desiran missed in the HARBOR trial, a failure this site covered earlier this month. VHB937 makes four.
The streak matters beyond any single program because it concentrates in the areas Novartis has told investors are its future. The company's growth story rests on a pipeline of novel mechanisms, and novel mechanisms fail more often than follow-ons, which is the risk the story always carried. Four failures in a month is the risk arriving on schedule. The company's broader portfolio is large enough to absorb the losses, but the pattern is now visible enough that the next readout will be read against it.
The honest counterpoint belongs in the same paragraph. A company that runs novel-mechanism trials in hard diseases will fail often, and the failure rate is the price of the successes the same pipeline occasionally produces. Novartis's bet on radioligand therapies paid off over years of doubt. The pipeline setbacks of the past month could look different in a year, and the ALS program's end, while real, is one decision in a company that makes dozens. The streak is context, not verdict.
What this means for the TREM2 idea and for ALS
The TREM2 hypothesis is not dead, but it is now twice failed in two diseases, and the burden of proof has shifted. The remaining argument for the mechanism is that the Alzheimer's trial has better target validation, and the company is letting that argument run. If the Alzheimer's trial also fails, the microglial era of drug development will close the way the amyloid era almost did: real biology, engaged targets, and clinical results that never quite arrived.
For ALS, the arithmetic is harsher. The trial's 251 participants contributed two years of their lives to a hypothesis that did not survive contact with the disease, and they will get the standard disclosure: the drug did not work. The field will present the data in Amsterdam, mine it for subgroups, and move on to the next mechanism. That is how ALS research works, and it is also why the disease has defeated every mechanism so far. The patients who enrolled in ASTRALS did not get a therapy, but they did move the field's understanding, and the December presentation will show how much. For a disease with nothing to lose and everything to learn, that is not nothing. It is just not enough.
Primary sources
- Fierce Biotech for the trial failure, the program halt, and the pipeline context.
- BioSpace for the trial design, the TREM2 mechanism, and the sequence of setbacks.