Introduction

Drug resistant cells represent a major threat for tumor patients as they might induce relapse and severely decrease disease outcome. Relapse represents a major drawback in patients with acute lymphoblastic leukemia (ALL), the single most frequent malignancy in children. Novel treatment options are intensively desired to remove drug resistant cells, which often additionally display dormancy.

Aim

We aimed at unraveling basic mechanisms determining drug resistance and dormancy, as basis for developing novel treatment strategies to prevent relapse.

Methods

Using cutting edge in vivo technology, we performed genetic engineering in the individualized xenograft mouse model of ALL. Primary patients' ALL cells were amplified in mice to generate patient-derived xenograft (PDX) cells. ALL PDX cells were lentivirally transduced to express transgenes. Recombinant luciferase allowed highly sensitive and reliable follow-up of leukemia growth and treatment. Recombinant surface markers enabled an unbiased approach to reliably and effectively enriching minute numbers of PDX cells from mouse bone marrow. Two independent, complementary innovative preclinical in vivo mouse models were established.

  1. In the first model, proliferation sensitive dyes allowed identifying and enriching in vivo long-term dormant PDX ALL cells.

  2. In the second model, the clinically highly relevant and challenging situation of MRD was mimicked in mice. PDX ALL cells were grown to advanced leukemia stages of above 30 % human blasts in bone marrow, when systemic chemotherapy with conventional cytotoxic drugs was initiated for prolonged periods of time, similar as applied in ALL patients. Chemotherapy reduced advanced leukemia down to 0,1 % or 10-3 leukemia cells in bone marrow, resembling not only complete morphologic remission, but even molecular remission.

This novel preclinical model allows for the first time to characterize patients' dormant and MRD cells in detail including functional in vivo assays.

Results

Using our innovative preclinical model of dormancy, we identified a novel, distinct, rare subpopulation of PDX ALL cells that displayed long term dormancy in vivo. Long-term dormant cells showed significant resistance against drug treatment in vivo, as therapy nearly exclusively targeted proliferating cells. Dormant cells showed stem cell behavior as they initiated leukemia upon re-transplantation into further recipient mice. Long-term dormant cells thus combined the three challenging characteristics of relapse-inducing cells dormancy, drug resistance and stemness with re-growth upon withdrawal of treatment pressure.

Using our second novel preclinical model, we isolated a pure, vivid fraction of rare MRD cells. These cells showed drug resistance in vivo and stemness features.

We used single cell RNA sequencing to compare the transcriptomes of dormant and MRD populations and found that they were highly similar. Both populations had further similarities with primary high-risk ALL cells and dormant sub-fractions in patients' leukemia samples.

Of high relevance for future treatment strategies, both, dormancy and drug resistance revealed transient characteristics in PDX ALL cells. When PDX long-term dormant ALL cells were distracted from their in vivo environment, they started proliferating similarly as their previously highly proliferative counterparts. When in vivo drug resistant PDX ALL cells were retrieved from murine bone marrow, they showed similar drug sensitivity in vitro as their sensitive counterparts.

Summary/Conclusion

Thus, both in vivo dormancy and drug resistance represent reversible characteristics in ALL cells which might result from the localization of ALL cells in the bone marrow niche. Dissolving ALL cells from their in vivo environment might sensitize them towards treatment. Addressing and inhibiting the interaction between ALL cells and their bone marrow niche might represent an attractive future therapeutic strategy to prevent ALL relapse.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution