CRACD loss in gastric cancer weakens immune attack

CRACD loss in gastric cancer may alter two aspects of tumor biology at once: it can push cells toward a more mucinous state while also helping them escape immune attack. The new study traces this effect from disruption of the cell’s internal actin framework to increased expression of PD-L1, a protein that can suppress antitumor T-cell activity.

CRACD is involved in regulating the actin cytoskeleton. Actin is often described as a cell’s internal scaffold, but it is actually a highly dynamic network whose filaments are constantly assembled and dismantled. These changes influence cell shape, movement, adhesion and intracellular signaling. The researchers found that losing CRACD can set off consequences that ultimately reach the immune system.

The work focused largely on mucinous gastric adenocarcinoma, a relatively uncommon histological subtype in which extracellular mucin accounts for more than half of the tumor volume. It represents roughly 2–9% of gastric cancers and is associated with features such as deeper invasion and a greater tendency to spread.

From the actin cytoskeleton to PD-L1

To investigate CRACD, the researchers engineered mouse gastric organoids carrying mutant Kras and loss of Trp53, then additionally deleted Cracd. Without Cracd, the cells became more plastic, accumulated mucin and developed features resembling mucinous gastric cancer.

The connection to immune escape was particularly striking.

Loss of CRACD disrupted the normal organization of actin filaments. This activated NF-κB/COX-2 signaling, increased production of the lipid mediator PGE2 and stimulated NADPH oxidase-dependent production of reactive oxygen species. These reactive molecules then stabilized HIF1α.

HIF1α is best known as a transcription factor that helps cells adapt to low oxygen. Here, the investigators found that activated HIF1α bound regulatory regions of Cd274, the gene encoding PD-L1, and increased its transcription.

PD-L1 is one of the molecular brakes that tumors can use to suppress T-cell responses. The pathway therefore connects an apparently structural defect to immune behavior: disrupting a protein that controls the actin cytoskeleton ultimately increases an immune-checkpoint molecule on tumor cells.

Single-cell analysis of tumors in mice supported this mechanism. Cracd-deficient tumors contained fewer T cells, while their CD8⁺ T cells showed stronger signs of exhaustion. Human gastric cancer data from TCGA provided another layer of support: across 408 tumors, HIF1A and CD274 expression were positively correlated, with a correlation coefficient of 0.44.

Evidence from human gastric tumors

The authors next asked whether the relationship was visible in human material. Patient-derived gastric cancer organoids showed an inverse pattern between CRACD and PD-L1: lower CRACD was associated with higher PD-L1.

They then examined a gastric cancer tissue microarray containing 182 patient samples. In the quantitative analysis shown for PD-L1 according to CRACD expression, 146 tumors were represented: 79 with no detectable CRACD, 48 with low expression and 19 with high expression. Tumors lacking CRACD had significantly higher PD-L1 combined positive scores than tumors retaining CRACD expression.

This human evidence is important because it links the mechanism seen in engineered organoids and mouse tumors to actual gastric cancers. It is still an association, however. The patient samples do not demonstrate that CRACD testing can already predict who will benefit from immunotherapy.

A possible therapeutic vulnerability

The researchers also tested whether interrupting the pathway could restore immune control. In mouse tumors they either inhibited HIF1α with the experimental compound PX-478 or blocked PD-L1 with an antibody.

Both approaches suppressed the growth of CRACD-deficient tumors and increased signs of antitumor T-cell activity. In one subcutaneous experiment, PX-478 was tested in 22 mice and compared with 17 vehicle-treated controls, producing a substantial reduction in tumor burden. Similar effects were observed in orthotopic models in which tumor cells were implanted into the stomach wall.

These results should not be read as evidence that PX-478 is an established treatment for people with CRACD-deficient gastric cancer. The therapeutic experiments were preclinical. Moreover, the investigators observed tumor recurrence after treatment was withdrawn in both the PX-478 and anti-PD-L1 groups, and late liver metastasis appeared after prolonged withdrawal of PX-478.

The immediate significance of the study is therefore mechanistic rather than clinical. CRACD may sit at a point connecting tumor-cell plasticity, cytoskeletal organization and immune suppression. The authors suggest that low CRACD could eventually become an additional biomarker, but explicitly do not propose replacing established markers such as PD-L1 CPS, microsatellite instability or Epstein–Barr virus status.

The next step is to test the finding in larger independent patient cohorts, particularly patients actually receiving PD-1 or PD-L1 checkpoint blockade. For now, the study illustrates an unexpected principle: changing the internal architecture of a cancer cell can alter not only how that cell looks and behaves, but also how effectively the immune system can recognize and control the tumor.