Single Cell RNA Sequencing Reveals Increased Adhesion Signals in Treatment-Resistant Tumor Stem Cells in a Preclinical Mouse Model of Genetically Engineered Patient-Derived Acute Lymphoblastic Leukemia

Treatment-resistant cells determine prognosis and outcome of cancer patients as they induce relapse with poor outcome. Novel therapeutic options are urgently needed to eradicate chemo-resistant tumor cells. Towards this aim, a deep understanding is required on mechanisms determining treatment-resistance in vivo and the ability to induce relapse.

Here, we aimed to identify and characterize the challenging rare subpopulation of drug resistant cells which survive in vivo chemo-therapy and are able to induce relapse.

As technical approach, we used the individualized mouse model of acute lymphoblastic leukemia (ALL) and amplified primary tumor cells in mice to generate patient-derived xenograft (PDX) cells. Upon genetic engineering by lentiviral transduction, PDX ALL cells expressed the three transgenes NGFR, a red fluorochrome and luciferase. While recombinant luciferase was used for in vivo imaging, a combined MACS/FACS procedure based on the expression of the transgenes enabled enriching PDX ALL cells from murine bone marrow by a factor above 10,000. Staining of PDX cells with CFSE was used to discriminate between highly proliferative and dormant tumor cells in vivo.

We treated mice harboring triple-transgenic, CFSE labeled PDX ALL cells with conventional chemotherapy; while in vivo treatment decreased the number of highly proliferative cells by more than 1 order of magnitude, the amount of dormant cells remained completely unchanged. Isolated, drug resistant cells revealed leukemia propagating potential and induced leukemia upon transplantation into next generation mice. Thus and using dormancy as an anchor, we could identify, isolate and enrich a subpopulation of treatment-resistant PDX ALL cells which might mimic relapse-inducing cells at minimal residual disease in patients.

We next aimed at characterizing the expression profile of these cells and were able to isolate single cells out of the low number of dormant cells to perform single cell RNA sequencing. Dormant, drug-resistant cells showed increased expression of several adhesion molecules suggesting an increased dependence on the bone marrow environment. Upon using gene set enrichment analyses, drug-resistant cells showed a highly similar expression profile to primary high risk leukemia subpopulations such as primary high risk ALL cells, the subpopulation of CD34 positive CML cells and leukemic or benign hematopoietic stem cells.

Taken together, treatment-resistant PDX ALL stem cells isolated and enriched from mice showed increased expression of adhesion molecules and resemble primary tumor cells of high risk subpopulations. These cells represent valuable tools to increase our understanding of mechanisms in minimal residual disease and relapse in patients. Our model will help to develop novel therapies which eliminate treatment resistant cells, prevent disease relapse and increase the prognosis of patients with ALL.