Genetic determinants of leukemic transformation in MPN

Myeloproliferative neoplasms (MPN) represent a heterogeneous group of disorders with variable risk of progression to leukemic, blast phase disease. The leukemic transformation of MPN is associated with a very poor prognosis.

At the 62nd American Society of Hematology (ASH) Annual Meeting and Exposition, two studies investigated the genes and mutations influencing leukemic transformation in patients with MPN. Christian Marinaccio reported on the impact of loss of the serine threonine kinase 11 (STK11) gene on the progression of MPN in murine models,¹ and the data has subsequently been published in Cancer Discovery.² Clemence Marcault presented clinical data demonstrating a role for nuclear factor erythroid 2 (NFE2) mutations in leukemic transformation of MPN.³ Here we report the findings from these studies.

Loss of STK11 facilitates progression of MPN¹

A CRISPR-Cas9 screen was performed to identify the genetic determinants of MPN progression using mice harboring the JAK2 V617F mutation with constitutive Cas9 expression in hematopoietic cells transduced with lentiviral CRISPR libraries containing short guide RNAs against various tumor suppressor genes. The screen showed short guide RNA enrichment for the STK11 gene, also known as liver kinase B1 (LKB1), in a serial replating assay.

In vitro observations

• Loss of STK11 promoted serial replating of both JAK2 V617F and MPL W515L murine MPN cells.
• RNA sequencing analysis of MPL W515L cells with loss of STK11 expression showed dysregulation of ~1,500 genes, compared with MPL W515L cells expressing STK11. Of note, the hexokinase 2 gene was upregulated and there was enrichment for hypoxia-related genes in MPL W515L/STK11^-/- cells.
• STK11 loss resulted in stabilization of hypoxia-inducible-factor 1A (HIF1A), which enabled serial replating of MPL W515L cells.

Murine model results

• STK11 knockout enhanced the MPN phenotype in MPL W515L mice, with significantly increased lethality compared with MPL W515L/STK11^+/+ mice (p < 0.0001).
• Histologic assessment revealed osteosclerosis and immature cells in the bone marrow of MPL W515L/STK11^-/- mice.
• Loss of STK11 promoted the engraftment of human MPN (CD34⁺) peripheral blood cells in MPL W515L mice, with a significant increase in the number of human CD34⁺ cells in the peripheral blood and bone marrow of transplanted STK11-knockout mice compared with STK11-wild type mice (p ≤ 0.05).

MPN patient samples

• STK11 was downregulated, whilst HIF1A expression was increased, in patients with blast phase MPN compared with patients with chronic phase MPN.
• Chronic and blast phase MPN had distinct genetic profiles, and blast phase MPN was enriched for the HIF1A-induced genes.
• There was considerable concordance between human RNA sequencing data (for blast phase vs chronic phase MPN) and murine RNA sequencing data (for STK11-knockout vs wild type MPL W515L mice) when looking at enrichment scores for hypoxia and oxidative phosphorylation.

Summary

Taken together, these results suggest that STK11 is a tumor suppressor in MPN, and loss of this gene is associated with progression to leukemic, blast phase disease. Hypoxia-inducible proteins could be potential therapeutic targets in blast phase MPN.

NFE2 mutation increases the risk of leukemic transformation in MPN³

Although NFE2 mutations occur in only a small subset of patients with MPN, epigenetically induced overexpression of NFE2 is found in most patients, and mutation or overexpression of NFE2 in murine models is associated with an MPN phenotype.

To determine the clinical impact of NFE2 mutation, a total of 1,243 patients diagnosed with MPN at a single institution were followed between January 2011 and May 2020. Diagnostic and follow-up samples from 707 of these patients were analyzed by next-generation sequencing, targeting 36 myeloid genes.

Patient characteristics

• Median age at MPN diagnosis was 51 years (IQR, 40–63 years).
• Breakdown of MPN subtypes in the overall cohort of 1,243 patients was as follows: essential thrombocythemia (46%), polycythemia vera (33%), primary or pre-fibrotic myelofibrosis (15%), unclassified MPN (5%), and myelodysplastic syndromes (MDS)/MPN (1%).
• An NFE2 mutation was present in 64 (9.1%) of 707 patients analyzed by next-generation sequencing; 36 patients (5.1%) had variant allele frequency ≥ 5%.
• There was no significant association between the presence of an NFE2 mutation and MPN subtype or type of driver mutation; most patients (73.1%) had a JAK2 V617F driver mutation.
• The following clinical characteristics were also similar between NFE2‑mutated and NFE2‑non-mutated patients: hemoglobin levels, platelet counts, white blood cell counts, and percentage of peripheral blasts; however, hematocrit was slightly higher for NFE2‑mutated patients than for NFE2‑non-mutated patients (50% vs 42.1%; p = 0.032).

Mutational landscape of NFE2-mutated MPN

• Two-thirds of NFE2 mutations were frameshift mutations, mostly occurring between the transactivation domain and binding domain.
• The most common mutation, occurring in around one-third of cases, was a truncating E261A frameshift mutation.
• Patients tended to harbor multiple NFE2 mutations; the most frequently co-occurring mutations (excluding driver mutations) were seen in TET2, DNMT3A, and ASXL1.

Response evaluation

• Patients harboring an NFE2 mutation had lower rates of complete response, complete hematological response, and clinical improvement, compared with patients with non-mutated NFE2 (p = 0.026; Table 1).
• Incidence of thrombotic/hemorrhagic events was not impacted by an NFE2 mutation.

Table 1. Median follow-up and response rates for NFE2-mutated vs non-mutated patients³

CHR, complete hematological response; CI, clinical improvement; CR, complete response; IQR, interquartile range; PD, progressive disease; PR, partial response; SD, stable disease.
	All patients (N = 1,243)	NFE2‑mutated patients (n = 36)	NFE2‑non-mutated patients (n = 671)	p value
Median follow-up, months (IQR)	103.8 (47.2–175.6)	181.3 (78.4–251.8)	119.8 (68.2–184.1)	—
Response, %
CR/CHR	46.10	41.67	49.18	0.026
PR	11.83	11.11	8.79
CI	4.57	0	3.73
SD	12.23	25.00	14.61
PD	7.00	22.22	6.71

NFE2 mutation is associated with acute myeloid leukemia/MDS transformation and inferior overall survival (OS)

• The presence of an NFE2 mutation was significantly associated with both transformation to acute myeloid leukemia/MDS and shorter OS (Table 2).
• Median transformation-free survival was 216.1 months for patients with an NFE2 mutation vs not-reached for patients without the mutation (HR, 10.29; 95% CI, 3.58–29.61; p < 0.001).
• Median OS was 144.2 months for patients with an NFE2 mutation, however it was not reached for patients without an NFE2 mutation (HR, 8.24; 95% CI, 3.67–18.52; p < 0.001).

Table 2. Factors associated with leukemic transformation and OS in multivariate analysis³

AML, acute myeloid leukemia; CI, confidence interval; HMR, high-molecular risk; HR, hazard ratio; MDS, myelodysplastic syndrome; MPN, myeloproliferative neoplasms; NFE2, nuclear factor erythroid 2; OS, overall survival; PMF, primary myelofibrosis.
	HR	95% CI	p value
Factors associated with AML/MDS transformation
Age at MPN diagnosis	1.08	1.05–1.12	< 0.001
PMF subtype	6.44	2.62–15.81	< 0.001
NFE2 mutation	10.29	3.58–29.61	< 0.001
HMR mutations	2.51	1.16–5.43	0.020
Factors associated with OS
Age at MPN diagnosis	1.09	1.07–1.12	< 0.001
PMF subtype	3.68	1.96–6.91	< 0.001
MDS/MPN subtype	5.73	1.27–25.82	0.023
NFE2 mutation	8.24	3.67–18.52	< 0.001
HMR mutations	2.13	1.22–3.70	0.007

Clonal changes during leukemic transformation

• Several cytogenetic abnormalities were acquired at the time of leukemic transformation in NFE2-mutated patients, including trisomy 8, and chromosome 5 and 7 deletions.
• Clonal evolution was mapped in four transformed patients and was consistent with that observed in murine models of NFE2 mutation-induced myeloid disorders, suggestive of NFE2 mutation-induced genetic instability in leukemic transformation.

Summary

The presence of an NFE2 mutation is associated with a greater risk of leukemic transformation, and shorter OS in patients with MPN, presenting a rationale for routine assessment for NFE2 mutations during MPN diagnosis and follow-up.

Conclusion

These studies identified loss of STK11, a putative tumor suppressor, and mutated or overexpressed NFE2, a potential oncogene, as genetic factors that may both independently contribute to leukemic transformation in MPN. These findings support our increased knowledge of the genetic determinants of progression in MPN and may help to better predict blast phase transformation and identify patients at high risk who would benefit from early intervention.