Clinical Trial Protocol, IRB amendments and DSMB recommendations

Abstract Background: Liver metastasis is a major cause of mortality in small cell lung cancer (SCLC), but the organ-specific cues that enable metastatic colonization are poorly understood. We hypothesized that the hepatic microenvironment promotes a stem-like, high-plasticity state required for metastatic outgrowth. Methods: We analyzed patient-derived SCLC liver metastases using RNA-seq, ATAC-seq, and metabolomics, combined with spatial transcriptomics and orthotopic liver colonization models. Regional hypoxia, HIF1α signaling, and metabolic flux were assessed by protein quantification and 13C-glucose tracing. Functional studies included pharmacological ACLY inhibition (SB204990) and CRISPR-mediated ACLY knockout in vivo. Results: Across patient specimens and in vivo models, metastatic SCLC cells in the liver adopted a distinct stem-like transcriptional state, marked by activation of regenerative programs and chromatin accessibility at stemness loci (HNF1A, HNF4A, SOX9, ATF3). Spatial profiling revealed that tumor cells adjacent to hepatocytes experienced localized hypoxia, resulting in HIF1α stabilization and the transcriptional induction of ACLY, a key enzyme that generates nuclear acetyl-CoA. ATP-citrate lyase (ACLY)-dependent generation of nuclear acetyl-CoA, driving histone hyperacetylation and chromatin remodeling at liver-lineage transcription factors HNF1A, HNF4A, and SOX9. Liver-metastatic cells undergo glycolytic reprogramming, a metabolic shift that parallels hepatocyte regeneration during liver injury, providing acetyl-CoA and biosynthetic precursors to support epigenetic remodeling and lineage adaptation. Similar progenitor-like hepatic plasticity is observed across other epithelial cancers with liver tropism, including breast, colon, and non-small cell lung cancers, indicating a conserved mode of metastatic adaptation. ALY activity mediated histone hyperacetylation and epigenetic remodeling required for the stem-like state. ACLY was functionally essential. Pharmacologic ACLY inhibition blocked the induction of stem-like programs, reduced histone acetylation, and growth rate of liver metastatic cells. CRISPR-ACLY knockout prevented the acquisition of the reprogrammed phenotype. Similar hepatic niche-induced stemness signatures were observed in liver metastases from breast, colon, and NSCLC, suggesting a conserved mechanism across epithelial cancers. Conclusions: The liver microenvironment actively promotes metastatic competency by inducing a HIF1α-ACLY-acetyl-CoA-driven stem-like state in disseminated tumor cells. ACLY inhibition disrupts this metabolic-epigenetic reprogramming and markedly suppresses liver metastatic outgrowth, highlighting ACLY as a tractable therapeutic target for liver metastasis. Citation Format: Ajit Kumar SHARMA, Nobuyuki Takahashi, Sophie Zhuang, Amira Kazi, Michael Nirula, Abhinav Joshi, Yingying Cao, Rajesh Kumar, Kanak Parmar, Christopher Schultz, Parth Anil Desai, Samantha Nichols, Linda Sciuto, Yue Huang, Chiori Tabe, Yang Zhang, Sanghvi Neel, Nishanth Ulhas Nair, Christopher A Febres Aldana, Nir Friedman, Simone Difilippantonio, Thorkell Andresson, Eytan Ruppin, Stephen M. Hewitt, Anish Thomas. Hepatic niche driven metabolic-epigenetic reprogramming mediates metastatic colonization through liver progenitor like plasticity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6127.

Trial Schema; Gating strategy; KEY RESOURCES table

ecDNA heterogeneity in SCLC, related to Figure 4

ecDNA in SCLC, related to Figure 1

Despite mechanistic rationale for combining PARP inhibitors with topoisomerase I inhibitors, clinical use has been hindered by dose-limiting toxicities. We hypothesized that integrating tumor-targeted topoisomerase I inhibitor delivery with optimized PARP inhibitor scheduling could enable effective combination therapy while reducing toxicity. In this trial (NCT02769962), we combined CRLX101, a nanoparticle topoisomerase I inhibitor, with olaparib using a gapped dosing schedule. The primary objective was to determine the maximum tolerated dose. Secondary objectives were to evaluate pharmacokinetics, pharmacodynamics, overall and progression-free survival. Twenty-four patients with advanced solid tumors were enrolled. The maximum tolerated dose for CRLX101 was 12 mg/m² every two weeks and olaparib 250 mg twice daily on days 3-13 and 17-26. Pharmacokinetics were consistent with monotherapy of each agent, and γH2AX kinetics revealed elevated DNA damage with the combination treatment compared to CRLX101 alone, supporting mechanistic efficacy. Among 19 evaluable patients, 2 patients had partial responses, and 6 had stable disease. Median overall survival was 6.06 months, progression-free survival 2.34 months, and duration of response 7.95 months. The combination showed acceptable safety across dose levels. Targeted delivery of a topoisomerase I inhibitor and gapped scheduling allowed higher olaparib dosing, showing promising activity and supporting the strategy's potential to widen the therapeutic window of DNA-damage response inhibitors while reducing toxicity.

Amplified genes identified from cell-free ChIP-seq of patient plasma samples.

Fusion landscape and ecDNA RLF-MYCL fusions, related to Figure 3.

Expression and copy-number status of recurrently highly amplified genes in patient and cell line samples. The tables summarize copy-number signal and AmpliconArchitect curation, expression of key genes and results of the multivariate analysis.

Enhancer-enhancer interaction analysis derived from H3K27ac HiChIP in the NCI-H889 and NCI-H524 cell lines. Interaction strength of proximal enhancers was calculated from the “.hic” files generated by the JuiceBox suite.

Fusion analysis results calculated with STARfusion and summary of RLF exon1 and exon2 expression for the CCLE cohort. Table summarizes the RLF-MYCL expression qPCR results.

Validation of ecDNA in SCLC, related to Figure 2

Enhancer-enhancer interaction results of MYC and MYCL ecDNAs in the DMS-273 cell line.

Patient sample AmpliconArchitect assemblies, related to Figure 1

Cell line AmpliconArchitect assemblies, related to figure 1 and Figure 2

Experimental validation of copy-number of the MYC locus in NCI-H889, DMS-114, NCI-H446 and RA022 patient derived cell lines and number of ecDNA and HSR positive cells in the RA022 patient derived cell lines

Amplicon design used for the targeted single-cell copy-number analysis. Normalized depth of MYC and MYCL amplicons in single-cells for the DMS-273 cell line are shown, as well as comparison of MYC amplicon depths in single cells between the patient derived Adrenal gland (ecDNA+) and Cerv.LN (HSR+) cell lines.

Heterogeneity in patient derived samples, related to Figure 5