Single Cell Transcriptomic Profiling of Patient-Derived Xenografts of Mantle Cell Lymphoma Reveals a Special Tumor Cell Population: Seed Cells for Disease Dissemination and Progression

Introduction

Mantle cell lymphoma (MCL) patients often presents at later stages and progress through its disease course by frequent involvement of multiple dissemination sites including spleen, liver, bone marrow (BM), peripheral blood (PB), and gastrointestinal tract (GI). This devious behavior translates into high degree of clinicopathologic heterogeneity, which may compromise therapies and promote relapse. Therefore, dissecting the cellular and molecular profiling and trafficking is critical in understanding the role of tissue tropism and evolution patterns contributing to its biological behavior. Since it is almost unfeasible to perform spatiotemporal collection in patients, in this study we took advantage of PDX models with serial samples and single cell transcriptomic profiling to address this important biology issue for the first time on MCL.

Method

Orthotopic PDX models (n = 6) were established via intravenous (IV) inoculation of primary MCL patient samples collected from PB (n = 5) or from LN (n = 1). These mouse models displayed similar dissemination patterns as the parental tumors. Cells from the predominant site of generation 1 (G1) were used to pass onto next generations (up to G9). For heterotopic PDX models, subcutaneous (SC) models were generated in parallel from two independent lines (up to G6) and exhibited predominant tumor growth at primary injection site with tumor spread to secondary sites only at very late stage. PDX samples from IV models (spleen, liver, BM, PB) and SC models across generations (n = 36) were collected and subjected to scRNA-seq profiling together with parental patient samples (n = 6) and healthy donor PBMC samples (n = 2).

Results

All six PDX models at G1 faithfully mirrored parental samples by displaying similar cancer hallmarks. Interestingly, MYC and OXPHOS signaling were predominantly and progressively augmented with each IV passage, and to a lesser extent across SC passages, suggesting a higher degree of selection and evolution processes during orthotopic passage. With spatial collection at distinct dissemination sites (spleen, liver, BM and PB) within same generations, we revealed that heterogenous transcriptomic profiles were more evident across tissues than generations. Specifically, cancer hallmarks such as MYC (NES = 8.4, FDR < 0.01), OXPHOS (NES = 8.9, FDR < 0.01) and mTORC1 (NES = 6.6, FDR < 0.01) signaling were highly enriched in cells from PB, and to a lesser extent in spleen and liver when compared to the cells in BM. More intriguingly, 55-60% of tumor cells in PB clustered together and showed enhanced cancer hallmarks for tumor migration and invasion (NES = 7.9, FDR < 0.01), higher de-differentiation scores (cytoTRACE) and G0/G1 cell cycle stage. This suggests that these cells are quiescent, de-differentiated and disseminative. Importantly, a small fraction of cells from spleen (5-18%) and liver (12-18%), but not in BM, showed similar characteristics and clustered together with those from PB. Histopathologic analysis showed that tumor cells could be detected in blood only after cells settled and expanded in the spleen, liver or BM, whereas dissemination to LN, GI tract, lung and kidney were even later events. Therefore, it is likely that these disseminative MCL cells originate from tissues and represent the tumor seed cells for disease dissemination. More interestingly, the top differential expressed genes (DEGs) in these seed cells were also significantly upregulated in ibrutinib-resistant patients (p < 0.01), compared to that in ibrutinib-sensitive patients based on bulk RNA sequencing (n = 69). This indicates that these seed cells are more resistant to ibrutinib and may drive therapeutic relapse. Targetable molecules are under active investigation to eradicate this ibrutinib-resistant seed cells.

Conclusion

MCL tissue tropism results in distinct transcriptomic profiles. A special cell population of tumor seed cells was identified to be quiescent, de-differentiated and disseminative, and may drive tumor spread, disease progression and therapeutic resistance (Figure 1). These observations provide biological insights into MCL disease progression in multiple MCL sites.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal