Association between prior lenalidomide exposure and TP53-mutated therapy-related MDS

Therapy-related myeloid neoplasms (t-MN) are secondary malignancies, such as acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS).¹ t-MN arise as a result of prior exposure to chemotherapy and/or radiation therapies (CRTs) for a primary condition. Generally, t-MN develop 3–7 years after exposure to CRTs and confer an adverse prognosis. Recent data suggest that t-MN may arise due to CRTs inducing clonal selection of preexisting mutant hematopoietic stem and progenitor cells (HSPCs).¹

The thalidomide analogs lenalidomide and pomalidomide facilitate the degradation of Ikaros transcription factors IKZF1 and IKZF3.¹ Lenalidomide has also been shown to promote the degradation of the protein kinase CK1α.¹ Sperling et al.¹ recently published an article in Blood suggesting that prior exposure to lenalidomide promotes the development of TP53-mutated t-MN, and we are pleased to summarize the key findings below.

Study design

Patients and mutational profiling

A retrospective analysis was performed of data from 416 patients diagnosed with t-MN (based on the 2016 World Health Organization classification) at MD Anderson Cancer Center, US, between 2008 and 2019. A comparison was made with data from 1,021 patients with de novo MN who were diagnosed in the same time period.

Next-generation sequencing was used to detect somatic mutations in bone marrow (BM) and peripheral blood samples using a 300-gene panel (n = 156).

Mouse model experiments

Trp53 mutations and resistance to thalidomide analogs

Immortalized mouse HSPC cell lines with an estrogen-inducible Hoxb8 transgene were generated using Crbn^I391V knock-in mice.

Hoxb8 Crbn^I391V;Rosa26-Cas9 cells were engineered to carry recurrent clonal hematopoiesis mutations using the CRISPR (clustered regularly interspaced short palindromic repeats)-Cas9 system and treated with lenalidomide and pomalidomide.

Selective advantage of HSPCs in the presence of thalidomide analogs

Live, lineage^lo, Sca1⁺, c-Kit⁺ (LSK) cells from Crbn^I391V;Rosa26-Cas9 were virally transduced with concentrated lentiviruses encoding single-guide RNAs targeting Dnmt3a, Tet2, Asxl1, Trp53, Ppm1d, and Ezh2 genes, a tag-red fluorescent protein, and two control single-guide RNAs in parallel. Mutant cells were transplanted into lethally irradiated wild-type (WT) recipient mice and validated through next-generation sequencing and flow cytometry (Figure 1).

c-Kit⁺ cells were transplanted with donor marrow consisting of 20% CD45.2;Trp53^−/−;Crbn^I391V cells mixed with 80% CD45.1;Crbn^I391V WT cells. Mice were then treated with dimethyl sulfoxide, lenalidomide 50 mg/kg, or pomalidomide   20 mg/kg. 

BID, twice a day; BM, bone marrow; HSPC, hematopoietic stem and progenitor cell; LSK, Live, lineage^lo, Sca1⁺, c-Kit⁺; PB, peripheral blood; PO, taken orally.
*Adapted from Sperling, et al.¹ Created with BioRender.com.

Differential toxicity of thalidomide analogs in HSPCs

Competitive BM transplant experiments were carried out with Csnk1a1 heterozygous knock-out mouse cells and WT cells in the Crbn1^391V background. Transplanted mice were treated with dimethyl sulfoxide , lenalidomide 50 mg/kg, pomalidomide 20 mg/kg, or iberdomide 20 mg/kg.

Results

Patient characteristics

Of the 416 patients diagnosed with t-MN, a total of:

• 40% were diagnosed with therapy-related AML and 60% with therapy-related MDS
• 63% had a primary diagnosis of solid tumors and 37% had non-myeloid hematologic cancers
• 45% received prior treatment with chemotherapy alone, 17% with radiotherapy alone, and 39% received both
• 17% underwent hematopoietic stem cell transplantation

Statistically relevant patient characteristics are listed in Table 1.

Table 1. Characteristics of patients diagnosed with t-MN*

PLT, platelet; t-AML, therapy-related acute myeloid leukemia; t-MDS, therapy-related myelodysplastic syndromes; WBC, white blood cell. *Adapted from Sperling, et al.1 †Median time from initial chemotherapy and/or radiation therapies to myeloid neoplasms diagnosis. ‡Cytogenetic abnormalities were recorded as a single abnormality or with multiple other abnormalities. §19% of patients had concurrent chromosome 5 and 7 abnormalities.
Characteristic, % (unless otherwise specified)	All patients (N = 416)	t-AML (n = 167)	t-MDS (n = 249)	p value
Median age (range), years	68 (17–91)	65 (17–89)	69 (22–91)	0.002
Median latency† (range), years	6.0 (0.1–40)	5.0 (0.1–40)	6.4 (0.3–45)	0.028
Median WBC (range), K/μL	3.2 (0.1–267)	3.8 (0.1–267)	3 (0.2–85.7)	<0.001
Median PLT (range), K/μL	58 (3–895)	41 (3–389)	65 (6–895)	0.002
Active primary malignancy	15	7	21	<0.001
Cytogenetics‡
Del 5q/-5§	32	25	37	0.013
Del 7q/-7	34	20	43	<0.001
Inv 16/t (16;16)	1	4	0	0.003
11q23	7	16	1	<0.001
t (15;17)	1	2	0	0.036
t (8;21)	1	2	0	0.036

Mutational profiling

At least one gene mutation was detected in 85% of patients with t-MN. The most common gene mutations detected were TP53 (37%), PPM1D (19%), TET2 (16%), DNMT3A (15%), RUNX1 (13%), ASXL1 (13%), and SRSF2 (10%). Comparison with the de novo AML and MDS cohort (N = 1,021) found that:

• TP53 and PPM1D mutations were significantly associated with t-MN
• STAG2 and ASXL1 mutations were more likely to occur in de novo AML/MDS
• NPM1, IDH1/2, FLT3, CEBPA, and NRAS mutations occurred more frequently in de novo AML
• TET2, PHF6, SRSF2, SF3B1, and U2AF1 mutations were more common in de novo MDS

Analysis based on treatment

Complex karyotype was associated with prior exposure to platinum agents (odds ratio [OR], 1.88; 95% confidence interval [CI], 1.23–2.89; false discovery rate [FDR] = 0.052).

Abnormalities in chromosome 7 were associated with prior exposure to alkylating agents (OR, 1.64; 95% CI, 1.08–2.49; FDR = 0.057) and platinum agents (OR, 1.65; 95% CI, 1.06–2.57; FDR = 0.057).

Patients who received radiation therapy were more likely to harbor NPM1 mutations, splicing gene mutations, and normal karyotype.

Mutations in TP53 were associated with treatment with proteasome inhibitors (OR, 3.06; 95% CI, 1.52–6.15; FDR = 0.025), and thalidomide analogs (OR, 2.62; 95% CI, 1.36–5.05; FDR = 0.035). Multivariate logistic regression analysis also found TP53 mutations were more common in patients treated with thalidomide analogs (OR, 3.14; 95% CI, 1.60–6.18; p = 0.009) or vinca alkaloids (OR, 1.76; 95% CI, 1.05–2.93; p = 0.031), and less common in patients treated with topoisomerase inhibitors (OR, 1.76; 95% CI, 1.05–2.93; p = 0.031).

In patients treated with lenalidomide, TP53 mutations were associated with a longer duration of exposure.

Mouse model experiments

In vivo and in vitro mouse models demonstrated that lenalidomide, but not pomalidomide, conferred a selective advantage to Trp53-mutant HSPCs.

• Multiplexed CRISPR-Cas9 experiments showed that Trp53 mutations conferred a selective advantage to HSPCs in the presence of lenalidomide, whereas mutations in Dnmt3a, Tet2, Asxl1, Trp53, Ppm1d, and Ezh2 genes did not.
• In long-term in vitro competition assays, mutated Trp53 cells demonstrated a strong resistance to lenalidomide compared with control cells, whereas pomalidomide only had a mild proliferation advantage even at high doses.
• In competitive BM transplant experiments with Csnk1a1^−/+ cells, differential HSPC toxicity occurred mainly due to lenalidomide-specific CK1α degradation.

Key findings

• Mutational profiling revealed that genes enriched in patients with t-MN differed from patients with de novo AML/MDS, with TP53 and PPM1D mutations occurring more frequently in patients with t-MN, suggesting selective pressure from prior therapies may drive t-MN.
• A positive association was found between exposure to thalidomide analogs and the presence of TP53 mutations.
• Mouse model experiments demonstrated that lenalidomide, but not pomalidomide, exerted selective pressure on Tp53-mutant leukemic cells and HSPCs, but was not associated with the proliferation of cells with other common clonal hematopoietic mutations. Moreover, lenalidomide was linked to Tp53 mutant selection, likely due to the degradation of CK1α.
• Further experiments with pomalidomide are necessary, as pomalidomide exposure normally occurs after prior exposure to lenalidomide, making it challenging to extract the independent effect of pomalidomide alone.
• This study also highlights the importance of genetic screening prior to initiation of treatment, to identify potential mutations that may exert selective pressure and to improve risk stratification.