The Protein That Wouldn't Be Drugged

Canonical Wnt/β-catenin signaling represents an evolutionarily conserved morphogenetic pathway regulating cell fate determination, proliferation, and tissue polarity across metazoans. In early embryonic development, graded Wnt signaling orchestrates anterior-posterior and dorso-ventral axis specification. In adult mammalian homeostasis, the pathway maintains tissue-resident stem cell compartments—most notably supporting the continuous self-renewal of intestinal crypt base columnar cells, hematopoietic stem cell niches, and osteoblast-mediated bone mineralization.

The central intracellular effector of this cascade is β-catenin, a dual-function armadillo-repeat protein functioning both at the plasma membrane in adherens junctions with E-cadherin and in the nucleus as a transcriptional coactivator. In roughly 80% of sporadic colorectal adenocarcinomas and in familial adenomatous polyposis (FAP), loss-of-function mutations in the adenomatous polyposis coli (APC) tumor suppressor disrupt the cytoplasmic destruction complex. Consequently, β-catenin escapes proteasomal clearance, translocating into the nucleus to constitutively drive proto-oncogene expression (such as MYC and CCND1).

The therapeutic challenge lies in achieving selective inhibition of oncogenic transcription while sparing essential adult stem cell niches, a hurdle that stymied first-generation Wnt-directed therapies.

The First Assault: Blocking the Cell Surface

The earliest attempts aimed at the cell surface. Antibodies like vantictumab and ipafricept were built to block the Frizzled receptors, stopping Wnt ligands before they could dock. The logic was sound — but the pathway they were silencing also builds bone.

Patients on these drugs suffered fractures. The trials closed not because the drugs didn't work, but because they worked too well on tissue no one meant to touch.

Small molecules had actually gotten there first. Back in 2009, XAV939 showed that inhibiting tankyrase could stabilize Axin and restore the destruction complex, degrading β-catenin from the inside. It worked — but it also hit the intestinal crypts hard, and the gut toxicity that showed up in early testing kept it from ever becoming a viable drug.

Researchers also moved upstream, targeting the porcupine enzyme that Wnt ligands need to leave the cell in the first place. The drug WNT974 (LGK974) ran into two separate problems: tumors that had already lost APC no longer needed to receive a Wnt signal at all, since the pathway was broken open internally — and even where the drug worked, it brought back the same bone and gut side effects, because a pathway that builds the embryo doesn't take a holiday in the adult.

Other efforts — blocking β-catenin's partnership with the coactivator CBP, or repurposing niclosamide, an anthelmintic with pleiotropic anti-cancer effects reaching beyond Wnt into mTOR and STAT3 signaling — ran into a shared obstacle: β-catenin doesn't have an obvious pocket to plug. It works through broad, flat protein-protein interfaces — handshakes rather than locks — which conventional small molecules aren't built to grip.

"Undruggable" turned out to be the wrong word. It wasn't that the target was invincible; it was that the tools were too blunt for a molecule that does its work through partnerships rather than active sites.

The Architecture of the Problem

To understand why β-catenin was so hard to drug, you have to understand what it actually does in a cell.

In its normal state, β-catenin is kept in check by the destruction complex — a group of proteins (APC, Axin, GSK3β, CK1α) that phosphorylate β-catenin and tag it for cellular disposal. When a Wnt signal arrives, it disassembles this complex, letting β-catenin escape into the nucleus. There, it partners with transcription factors like TCF and BCL9 to switch on growth-promoting genes.

In colorectal cancer, the APC gene is mutated in ~80% of cases. Without APC, the destruction complex can't form. β-catenin accumulates unchecked, flooding the nucleus and keeping growth genes permanently active. The cell doesn't need a Wnt signal anymore — the pathway is broken open from the inside.

This is why upstream drugs failed. If you block Wnt ligands at the cell surface with a porcupine inhibitor, but the tumor has already lost APC, the signal is irrelevant. The pathway is constitutively active regardless of what happens upstream. It's like trying to stop a flood by turning off a faucet when the dam has already burst.

The New Generation: Wedging Into the Handshake

The newer generation of drugs is built around exactly that insight.

FOG-001, developed by Parabilis Medicines, is a stabilized helical peptide — a short chain of amino acids locked into a rigid shape — that wedges directly into the interface between β-catenin and its partner TCF. In early trials, it has shown tumor shrinkage across multiple cancer types without reproducing the bone toxicity of the earlier drugs.

Why it's been better tolerated is still being worked out. Canonical Wnt signaling in bone also runs through β-catenin/TCF, so there's no clean branch point separating the two. The likelier explanation is some selectivity for oncogenic β-catenin/TCF complexes over normal ones — differences in cofactor availability, post-translational modification, or expression level between tumor and healthy tissue — rather than the drug simply missing bone biology altogether.

In data presented at ESMO 2025 and AACR 2025, FOG-001 showed striking results in desmoid tumors — a Wnt-driven fibromatosis where β-catenin mutations are nearly universal:

The FDA granted FOG-001 Fast Track designation in November 2025 for desmoid tumors. The drug is also being tested in microsatellite-stable colorectal cancer, hepatocellular carcinoma, and prostate cancer.

ST316, from Sapience Therapeutics, takes a related approach but targets a different partner: BCL9, another β-catenin coactivator essential for oncogenic transcription but less critical for normal tissue maintenance. In April 2026, at AACR, Sapience reported the first Phase 2 data — and the numbers turned heads:

Responses were seen across RAS-mutated and RAS-wild-type patients, in patients with liver metastases, and in those previously treated with bevacizumab — populations that typically do poorly on standard therapy. The pharmacodynamic data showed something equally important: Wnt target genes were knocked down in tumor cells but spared in adjacent normal tissue, suggesting genuine selectivity.

Beyond the Peptide: The Degrader Toolkit

Behind FOG-001 and ST316 sits a broader toolkit of technologies that didn't exist when the first Wnt drugs failed.

PROTACs (proteolysis-targeting chimeras) are bifunctional molecules: one end grabs β-catenin, the other recruits a cellular disposal system (E3 ubiquitin ligase), and the cell's own machinery destroys the target. The first β-catenin PROTAC, xStAx-VHLL, linked a β-catenin-binding peptide to a VHL ligand and achieved sustained degradation in APC-mutant colorectal cancer organoids. In 2024, nanoengineered peptide PROTACs were developed to simultaneously degrade β-catenin and STAT3, leveraging synergy between these oncogenic pathways.

Molecular glues are smaller, simpler molecules that induce proximity between a target protein and an E3 ligase without the bulky two-armed structure of a PROTAC. NRX-252114, developed in 2019, binds the E3 adapter β-TrCP and enables it to recognize mutant β-catenin — a particularly elegant solution for cancers where the mutation removes the normal degradation signal.

Biomolecular condensates represent perhaps the most novel approach. Dewpoint Therapeutics' DPTX3186 doesn't destroy β-catenin at all — it sequesters it into inactive nuclear droplets, essentially putting it in cellular time-out. The drug is currently in Phase 1 trials for Wnt-driven solid tumors.

And antisense oligonucleotides (ASOs) — short DNA strands that bind to β-catenin's mRNA and trigger its destruction before the protein is ever made — remain preclinical but offer the ultimate upstream intervention.

The Graveyard Is a Map

The graveyard of failed Wnt drugs isn't really a monument to defeat. It's a record of exactly where the pathway's dependencies live — knowledge that the newer, more selective drugs are now built on.

The shift is clear: from blocking the signal (upstream, blunt, toxic) to disrupting the partnership (downstream, precise, tolerable). The first generation treated Wnt like a faucet to be turned off. The second generation treats it like a conversation to be selectively silenced.

What Comes Next

The Wnt/β-catenin story is still unfolding, but several patterns are already clear.

First, the "undruggable" label was always about chemistry, not biology. β-catenin was never invincible — it was just incompatible with the small-molecule toolbox of the 1990s and 2000s. Peptides, PROTACs, and molecular glues are opening targets that traditional chemistry couldn't reach.

Second, selectivity matters more than potency. The first Wnt drugs were plenty potent — they just hit the wrong cells. FOG-001 and ST316's early success stems from distinguishing oncogenic β-catenin partnerships from normal ones, not from being stronger inhibitors.

Third, the clinical pipeline exhibits diverse mechanistic strategies rather than convergent chemical scaffolds. While no Wnt-pathway targeted agent has received regulatory approval to date, distinct clinical modalities are establishing proof-of-concept: FOG-001 holds FDA Fast Track designation supported by disease control in desmoid fibromatosis; ST316 has demonstrated clinical activity in second-line metastatic colorectal cancer; and biomolecular condensate modifiers represent novel non-degradative biophysical interventions.

Fourth, colorectal carcinoma remains the primary clinical indication of interest. Given that over 80% of colorectal malignancies harbor truncating APC mutations or activating CTNNB1 alterations, selective pharmacological inhibition of downstream β-catenin transcriptional complexes represents a validated, long-sought therapeutic objective capable of overcoming the toxicities that arrested earlier upstream programs.