The research, led by Sotaro Chikamatsu, Masaki Shiota, and colleagues at Kyushu University in collaboration with teams across Japan and the University of Tokyo, took a two-pronged approach. First, the investigators analyzed longitudinal clinical data from a prospective cohort of 199 patients with advanced prostate cancer, tracking how serum cholesterol changed as the disease evolved from a hormone-sensitive state to castration resistance. Second, they performed paired, single-cell-resolution spatial multi-omics analyses on hormone-sensitive prostate cancer specimens and matched CRPC samples taken from the very same patients. This paired design is a technical tour de force: rather than comparing tumors from different individuals, where genetic background and prior treatment histories confound the results, the researchers could observe how the metabolic landscape of a single patient’s tumor transformed as it adapted to androgen deprivation.

The clinical findings alone are striking. The team found that it was not the baseline level of serum total cholesterol that predicted risk, but rather the magnitude of its increase over time. Patients whose total cholesterol rose more steeply during treatment showed a significantly higher risk of progressing to CRPC. This distinction matters. Cholesterol is a routine clinical measurement, but it is usually interpreted as a static snapshot, a cardiovascular risk factor. Here, the data suggest that the trajectory, the dynamic change in circulating cholesterol, carries prognostic information about the tumor’s metabolic behavior. It hints at a biological dialogue in which the evolving tumor and the body’s lipid economy influence one another, with rising cholesterol levels marking, and possibly supporting, the tumor’s escape from hormone dependence.

To understand the tissue-level mechanisms behind this clinical signal, the researchers turned to spatial multi-omics, a family of technologies that maps gene expression and protein markers onto intact tissue sections at single-cell resolution. Using a targeted spatial transcriptomics platform profiling hundreds of curated genes, combined with multiplex immunofluorescence imaging, the team dissected both the malignant epithelial compartment and the surrounding tumor microenvironment, including macrophages, fibroblasts, and other stromal populations. Formalin-fixed, paraffin-embedded archival tissue, the standard material stored in hospital pathology labs, could thus be interrogated with unprecedented molecular depth while preserving the crucial dimension of location: knowing not just which cells express which genes, but which cells sit next to which.

What emerged was a picture of remarkable heterogeneity. Within the same tumor, the researchers identified distinct cholesterol acquisition strategies among cancer cells. Some tumor cells adopted an import-dominant program, ramping up expression of the scavenger receptor class B type 1, known as SCARB1, a cell-surface receptor that takes up cholesterol from circulating lipoproteins. Others favored de novo synthesis, building cholesterol from scratch through the mevalonate pathway. These metabolic subtypes were not randomly distributed. Instead, the spatial analyses revealed that import-dominant tumor cells showed a striking tendency to localize in close proximity to stromal cells with active cholesterol efflux programs, a spatial association that held across the paired HSPC-to-CRPC comparison and became more pronounced as the disease progressed.

The efflux-active stromal populations were themselves a discovery. The team identified expanded populations of macrophages characterized by high expression of ABCA1, the ATP-binding cassette transporter that pumps cholesterol out of cells, and fibroblasts marked by high expression of ABCA8, a related transporter associated with lipid handling. Macrophages are well known elsewhere in the body as cholesterol shuttlers; in atherosclerotic plaques, foam-cell macrophages loaded with cholesterol use ABCA1 to offload their lipid cargo. The new study suggests that an analogous lipid-handling program operates within the prostate tumor microenvironment, where macrophages and fibroblasts enriched for efflux machinery accumulate near tumor cells that are simultaneously upregulating cholesterol import. The spatial logic is compelling: stromal cells that export cholesterol could create a local lipid-rich niche, and tumor cells positioned nearby, equipped with SCARB1 and other import machinery, would be ideally placed to harvest it.

This model reframes the tumor microenvironment as a metabolic marketplace. Rather than viewing stromal cells as passive bystanders or merely as sources of growth factors, the study positions them as active participants in the tumor’s lipid supply chain. The coordinated geography, efflux-active macrophages and fibroblasts clustered around import-hungry cancer cells, is exactly the pattern one would expect if cholesterol were being transferred across the tumor-stroma interface. The researchers support this interpretation with enrichment analyses showing that the genes upregulated in ABCA1-high macrophages and ABCA8-high fibroblasts are significantly associated with cholesterol efflux-related biological processes, and with correlations between transporter expression and lipid-associated markers within the tissue. While such spatial associations demonstrate correlation rather than direct transport, the consistency of the pattern across patients and disease states makes a functional relationship highly plausible.

The implications for therapy are considerable. Current strategies targeting cholesterol in prostate cancer remain investigational, and one reason they have struggled is that tumors can compensate: block uptake, and cells may simply synthesize more; inhibit synthesis, and they may import instead. The spatial framework developed here suggests a more nuanced target. If import-dominant tumor cells depend on their efflux-active stromal partners, then disrupting the tumor-stroma metabolic coupling, rather than tumor cholesterol handling in isolation, might undermine the adaptation more effectively. It also raises the possibility that the stromal efflux programs themselves could serve as biomarkers, identifying patients whose tumors are primed for lipid scavenging and who might benefit most from metabolic interventions. The authors are appropriately cautious, framing these as potential therapeutic opportunities for future investigation rather than immediate clinical recommendations.

The study also carries a broader methodological message for cancer research. Bulk omics, which average molecular signals across millions of cells, would have averaged away the very phenomenon at the heart of this discovery. Only by preserving spatial context, and by pairing samples from the same patient before and after castration resistance emerged, could the team detect the coordinated choreography of cholesterol programs. As spatial transcriptomics and multiplex imaging become more accessible, similar metabolic geographies are likely to be uncovered in other cancers, where nutrient sharing between malignant and stromal cells is increasingly recognized as a hallmark of tumor ecology. The Japanese consortium’s work, conducted under the KYUCOG-1401 prospective trial framework with appropriate ethical oversight and patient consent, demonstrates what is possible when rigorous clinical cohort design meets cutting-edge spatial technology.

For patients with advanced prostate cancer, the road from these findings to the clinic will take time, and the authors emphasize that their results define a framework rather than a treatment. But the conceptual shift is immediate and profound. Castration-resistant prostate cancer is not simply a tumor that has learned to grow without androgens; it is a restructured ecosystem in which metabolic roles are divided among cell types and organized in space. Rising serum cholesterol during therapy, once dismissed as an incidental laboratory value, now appears as a potential echo of that restructuring, a systemic signature of a local metabolic bargain being struck between cancer cells and their neighbors. Decoding that bargain, and learning how to break it, may open one of the next frontiers in the treatment of this stubborn disease.

Subject of Research: Spatially organized cholesterol metabolism between tumor and stromal cells in castration-resistant prostate cancer

Article Title: Spatial multi-omics reveal spatially coordinated cholesterol metabolic programs in castration-resistant prostate cancer

Article References: Chikamatsu, S., Shiota, M., Fukuchi, G., Tanegashima, T., Seki, M., Kanai, A., Suzuki, Y., Matsuyama, H., Kamoto, T., Enokida, H., Fujimoto, N., Sakai, H., Igawa, T., Kamba, T., Yokomizo, A., Naito, S., & Eto, M. (2026). Spatial multi-omics reveal spatially coordinated cholesterol metabolic programs in castration-resistant prostate cancer. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-08816-5

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08816-5

Keywords: prostate cancer, castration-resistant prostate cancer, spatial multi-omics, cholesterol metabolism, tumor microenvironment, metabolic reprogramming, SCARB1, ABCA1, ABCA8, macrophages, cancer-associated fibroblasts, spatial transcriptomics

The researchers, led by Qingyu Zhang and Xiaolin Lu of the Capital Institute of Pediatrics together with colleagues at Capital Medical University, the Chinese People’s Liberation Army General Hospital Eastern Medical Branch, and Beijing University of Technology, combined metabolomics with transcriptome-wide sequencing to interrogate both sides of the puzzle at once. On the metabolic side, they used mass spectrometry to measure concentrations of amino acids and related metabolites in the serum of women with anembryonic pregnancy and compared those profiles against healthy pregnant controls. On the genomic side, they sequenced RNA from chorionic villi, the placental precursor tissue obtained from the affected pregnancies, and used a computational tool called rMATS to systematically catalog differences in alternative splicing between patients and controls.

Alternative splicing is the cell’s strategy for squeezing remarkable diversity out of a finite genome. After a gene is transcribed into precursor messenger RNA, segments called exons can be stitched together in different combinations while others are skipped, producing protein variants with distinct functions. The process is a fundamental driver of development and tissue identity, and errors in splicing have been implicated in conditions ranging from neurodegenerative disease to cancer. Previous work had already hinted that splicing matters in reproduction: animal studies have shown that splicing regulators shape placental growth, and disruptions of splicing factors have been linked to impaired placental development and spontaneous abortion in experimental models. What remained unknown was whether splicing defects contribute to the metabolic disturbances characteristic of anembryonic pregnancy specifically.

The scale of the splicing disruption the team documented was striking. Compared with healthy controls, patients with anembryonic pregnancy carried 903 genes showing differential alternative splicing and 558 differentially expressed genes. Among the splicing changes, exon skipping emerged as the predominant event type, meaning that in the affected tissue, specific exon segments were disproportionately excluded from the mature messenger RNA. Mass spectrometry independently confirmed that amino acid metabolism was indeed dysregulated in the patients’ serum, consistent with earlier metabolomic studies of early pregnancy loss and reinforcing the idea that the metabolic environment of the pregnancy goes awry when embryonic development fails.

Three genes stood out from the splicing analysis with the most significant alterations: PRKCD, which encodes a protein kinase involved in cell signaling and apoptosis; SCARB1, the scavenger receptor class B member 1 that serves as the major cell-surface receptor for high-density lipoprotein cholesterol; and METTL2B, a member of the methyltransferase family. In all three, the affected samples showed preferential exon inclusion rather than skipping, a shift that would change the protein products these genes yield in the chorionic villi. The team validated these findings using splicing-specific polymerase chain reaction, a technique that amplifies and distinguishes the exon-included and exon-excluded transcript forms, confirming that the sequencing results reflected genuine biological differences rather than computational artifacts.

The choice of SCARB1 as a headline finding carries considerable biological weight. The receptor it encodes, often called SR-BI, was identified in landmark work in the 1990s as the high-density lipoprotein receptor, and it plays a well-established role in cholesterol homeostasis. Intriguingly, prior research in mice had already shown that the splicing regulator RBFOX2 in the liver controls cholesterol levels by modulating Scarb1 alternative splicing, and independent studies have tied SR-BI to metabolic phenotypes in other contexts. The new study extends that logic to human early pregnancy, suggesting that when SCARB1 splicing goes wrong in chorionic villi, the consequences ripple outward into both lipid and amino acid metabolism.

To understand what might be driving the splicing changes, the investigators turned to motif analysis, a computational approach that searches DNA sequences near splicing events for binding signatures of transcription factors. The analysis predicted that SCARB1 is a potential target gene of TEAD4, a transcription factor with an established and critical role in early embryology. TEAD4 is a key effector of the Hippo signaling pathway, and in mouse studies it is required for the specification of the trophectoderm, the cell lineage that gives rise to the placenta. Work in human preimplantation embryos has similarly shown that TEAD4 regulates trophectoderm differentiation upstream of the gene CDX2, and separate research has demonstrated that loss of TEAD4 function impairs postimplantation development and has been implicated in early human pregnancy loss.

Functional experiments in the new study revealed that both SCARB1 and TEAD4 participate in the regulation of metabolism, including amino acid and lipid metabolism, linking the transcriptional regulator, its predicted splicing-sensitive target, and the metabolic derangements measured in patient serum into a coherent mechanism. The authors further examined correlations between the degree of SCARB1 exon exclusion and the serum concentrations of nineteen amino acids and metabolites, integrating transcription factor prediction with publicly available datasets to build a picture of how splicing-mediated changes could feed into systemic metabolic disruption. Protein-protein interaction analysis of pregnancy-related TEAD4 targets undergoing exon-skipping events added another layer, highlighting placenta-expressed genes within this regulatory network.

The study’s implications cut in two directions. Scientifically, it elevates alternative splicing-mediated metabolic reprogramming from a suspected mechanism to a documented feature of anembryonic pregnancy, connecting a transcription factor central to trophectoderm specification, a cholesterol receptor with newly appreciated metabolic reach, and the amino acid disturbances measurable in maternal blood. Clinically, the findings suggest potential avenues that were previously invisible: if the splicing pattern of SCARB1 or the serum amino acid profile proves robust in larger cohorts, such molecular signatures could eventually inform earlier diagnosis of anembryonic pregnancy or point toward metabolic interventions, although the authors’ work remains at the mechanistic stage and no diagnostic or therapeutic application follows directly from these results.

Important caveats temper the enthusiasm. The validation experiments compared eight patients with eight controls, a sample size that underscores the need for replication in larger and more diverse populations before the SCARB1-TEAD4 axis can be considered a confirmed driver of anembryonic pregnancy rather than a strongly supported hypothesis. The study also documents correlation and predicted regulation rather than demonstrating, through direct perturbation experiments, that correcting SCARB1 splicing rescues metabolic function. The RNA sequencing datasets have been deposited in the National Genomics Data Center under accession PRJCA041061, allowing other researchers to interrogate the data independently, and the work received support from the Beijing Research Ward Excellence Program, the National Natural Science Fund of Beijing, and the National Natural Science Foundation of China. Even with these limitations, the study marks a meaningful step toward explaining why some pregnancies stall before the embryo ever develops, and it reframes early pregnancy loss as a disorder that can begin with a misstep in how RNA is cut and pasted.

Subject of Research: The role of SCARB1 alternative splicing in amino acid metabolism dysregulation in anembryonic pregnancy

Article Title: Dysregulation of SCARB1 Alternative Splicing Disturbs Amino Acid Metabolism in Anembryonic Pregnancy

Article References: Zhang, Q., Wu, X., Liu, X., Wei, J., Wang, L., Cao, J., & Lu, X. (2026). Dysregulation of SCARB1 Alternative Splicing Disturbs Amino Acid Metabolism in Anembryonic Pregnancy. Reproductive Sciences. https://doi.org/10.1007/s43032-026-02179-4

Image Credits: AI Generated

DOI: 10.1007/s43032-026-02179-4

Keywords: anembryonic pregnancy, alternative splicing, SCARB1, TEAD4, amino acid metabolism, chorionic villi, RNA sequencing, miscarriage, metabolomics, transcription factors, early pregnancy loss, reproductive biology