The NSD2 p.E1099K Mutation in Relapse Pediatric Acute Lymphoblastic Leukemia Is Linked to Mercaptopurine Resistance

Introduction: While the outcome for children with acute lymphoblastic leukemia (ALL) has improved dramatically, the prognosis for those who relapse remains poor. One of the most common alterations found at relapse is the p.E1099K missense change within the SET domain of NSD2, a histone methyltransferase that di-methylates histone 3 lysine 36 (H3K36). NSD2 has 3 isoforms, two of which, Type II (canonical) and REIIBP (C-terminal), contain the SET domain, and another, Type I (N-terminal), that does not. The p.E1099K mutation leads to increased enzymatic activity, but pathways leading to a clonal advantage are unknown in ALL.

Methods: We used short hairpin RNAs (shRNAs) to target knockdown of two combinations of NSD2 isoforms: shI/II targets Types I and II, shII/RE targets Type II and REIIBP. Three different B-cell lines (Reh, 697, and KOPN-8) with 2 wildtype (WT) copies of NSD2 were stably transduced with shII/RE. Two B-Cell lines, RS4;11 and RCH-ACV, heterozygous for the NSD2 p.E1099K mutation, were transduced with shI/II and shII/RE. As a control, each B-cell line was stably transduced with a scrambled non-targeting (NT) shRNA. NSD2 knockdown was confirmed by Western Blots.

Cell lines were treated for 5 days with chemotherapy agents commonly used in pediatric ALL treatments (mercaptopurine (MP), cytarabine, methotrexate, prednisone, and doxorubicin). Cytotoxicity was assessed by CellTiter- Glo® and significance between IC50s was determined by ANOVA and post hoc Tukey test. Cell proliferation was measured by cell counting with trypan blue. Cell cycle progression in RS4;11 lines was monitored with Edu staining and flow cytometry with and without exposure to MP.

Results: Similar to previously reported results, knockdown of NSD2 in the 3 WT B-cell lines had no effect on cell proliferation. However, shI/II reduced growth by 40% in RS4;11 and 20% in RCH-ACV, while shII/RE decreased proliferation by 45% in RS4;11 and 55% in RCH-ACV when compared to their NT control. In RS4;11, both shI/II and shII/RE led to a similar 10% decrease in cells progressing through S phase compared to NT, which could be due to either a slower progression through cell cycle or less cells entering the cell cycle.

Knockdown of NSD2 resulted in sensitivity to 6MP compared to NT in both RS4;11 and RCH-ACV lines. RS4;11 shII/RE had an IC50 3.2-fold more sensitive ( p<.01) and the RS4;11 shI/II IC50 was 1.25-fold more sensitive (NS) versus the NT control. Similarly, RCH-ACV shII/RE had an IC50 3.4-fold more sensitive (p<.01) and the RCH-ACV shI/II IC50 was 2.6-fold more sensitive (p<.01) compared to the NT control. No significant changes in drug sensitivity were noted for the 3 WT NSD2 knockdown B-cell lines compared to their NT controls.

During a 120 hour exposure to MP, 34% more RS4;11 shII/RE cells were arrested in the G phase than NT controls, while 26% more RS4;11 shI/II cells were arrested in G phase relative to NT controls. This result indicates MP exposure leads to a reduced percentage of knockdown cells able to progress through the cell cycle. Overall, simultaneously reduced expression of Type II and REIIBP had a greater effect of on cell proliferation and MP response compared to the co-reduction of Types I and II NSD2 in the p.E1099K heterozygous cell lines.

Conclusion: The p.E1099K mutation confers a growth advantage and resistance to MP, a cornerstone of ALL therapy. Concurrent reduction of Type II and REIIBP expression by shII/RE resulted in the largest impact on proliferation and MP sensitivity. Both of these isoforms include the SET domain containing the p.E1099K mutation, which indicates one or both isoforms could be responsible for changes in the chromatin state and other possible alterations that lead to a clonal advantage. Based on our findings, determining the mechanism of resistance to MP imparted by NSD2 p.E1099K is now a top priority.