Small cell lung cancer (SCLC) is one of the deadliest cancers with a five-year survival rate below 6 percent. It initially responds well to standard chemotherapy with cisplatin and etoposide, but nearly every patient relapses with drug-resistant disease within months. The genetic drivers of that resistance have remained largely unknown.
A team led by David MacPherson at the Fred Hutchinson Cancer Center has now identified one of them , and a potential way to circumvent it. Using in vivo CRISPR screens in patient-derived xenograft (PDX) mouse models, the researchers discovered that loss of the USP22 deubiquitylase enzyme is a powerful driver of acquired chemoresistance in SCLC.
The findings, published in Nature Communications on July 8, 2026 (DOI: 10.1038/s41467-026-75117-2), also reveal that USP22-null tumors develop a metabolic dependency that can be exploited therapeutically.
How the screen worked
Traditional drug-resistance studies grow cancer cells on plastic dishes, missing the tumor microenvironment and drug penetration dynamics that matter in a real body. MacPherson’s team used a more clinically relevant approach: they implanted human SCLC tumors into immunocompromised mice, treated them with cisplatin-etoposide or saline, and used a CRISPR library targeting approximately 400 genes to identify which gene knockouts were enriched in the chemotherapy-resistant tumors that regrew.
The results were striking. Five of the six most enriched guides targeted genes encoding components of the SAGA (Spt-Ada-Gcn5 Acetyltransferase) transcriptional co-activator complex , with USP22 at the top of the list.
USP22 is a deubiquitylase that removes ubiquitin from histones H2B and H2A, playing a critical role in transcriptional activation. It is well-known as part of a “death-from-cancer” gene signature associated with poor prognosis in many tumor types. This study is the first to identify USP22 loss , not gain , as a driver of chemoresistance.
Epigenetic reprogramming
The mechanism unfolds as a cascade. USP22 loss leads to increased H2AK119 monoubiquitylation , a repressive histone mark , at the gene bodies of regulators of neuronal differentiation. This epigenetic change suppresses the expression of neural and neuroendocrine genes, including targets of the transcription factor ASCL1, a master regulator of neuroendocrine identity in SCLC.
The result is an attenuated DNA damage response: cells show reduced γH2AX signaling , a marker of DNA double-strand breaks , and reduced apoptosis in response to cisplatin-etoposide treatment. In effect, the tumor cells stop responding to chemotherapy that would normally kill them.
At the same time, USP22-null tumors upregulate glycolysis and hypoxia-related genes, switching to a glycolytic (Warburg-like) metabolism.
A targetable vulnerability
That metabolic shift creates a new dependency. USP22-null tumors become reliant on GLUT1 , the primary glucose transporter in most cancer cells , to sustain their elevated glycolysis.
The MacPherson team tested BAY-876, a highly selective small-molecule GLUT1 inhibitor developed by Bayer (IC₅₀ ≈ 0.002 µM for GLUT1, with 250- to 500-fold selectivity over other GLUT isoforms). In PDX models, the combination of BAY-876 with cisplatin-etoposide re-sensitized USP22-null tumors to chemotherapy.
Crucially, the team also performed a genetic rescue experiment. A PDX model carrying a natural homozygous truncating USP22 mutation was re-sensitized to chemotherapy when USP22 expression was restored , confirming that USP22 loss is directly causal, not merely correlated.
Clinical implications
SCLC is not routinely profiled for USP22 or SAGA complex status in the clinic. These findings suggest that patients whose tumors harbor USP22 loss, or alterations in other SAGA complex components, could be identified through genomic profiling and treated with a combination of chemotherapy plus a GLUT1 inhibitor.
Several caveats apply. The findings are preclinical , all in PDX models , and human clinical validation is needed. BAY-876’s safety and efficacy in combination with cisplatin-etoposide has not been tested in people, and systemic GLUT1 inhibition carries risks: GLUT1 is expressed on red blood cells and at the blood-brain barrier. USP22 loss may explain resistance in only a subset of SCLC patients; other mechanisms, including KEAP1 loss and MYCN amplification, also drive resistance.
Still, the study demonstrates that the PDX-based in vivo CRISPR screening approach can uncover clinically relevant resistance mechanisms that traditional cell-culture screens miss. For a disease with a five-year survival rate below 6 percent, every actionable target is worth pursuing.
Sources
1. Best, S. et al., “Loss of the USP22 deubiquitylase confers resistance to chemotherapy in small cell lung cancer”, Nature Communications (2026). DOI: 10.1038/s41467-026-75117-2
2. Fred Hutchinson Cancer Center, “Researchers identify driver of chemotherapy resistance in small cell lung cancer” (8 July 2026).