The development of a secondary hematologic malignancy is one of most serious risks of cytotoxic cancer chemotherapies. Therapy-related myeloid neoplasms (t-MN) including myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) are associated with a particularly poor prognosis. Broadly, there are two main mechanisms by which cytotoxic therapies are proposed to cause t-MN. First, cytotoxic therapy–induced DNA damage might directly lead to cancer-driving mutations through inappropriate DNA repair. Alternatively, chemotherapy exposure may apply selective pressure on specific subpopulations of pre-existing abnormal hematopoietic stem cells (HSCs), which then gain a survival advantage, thus promoting subsequent clonal evolution. Clonal hematopoiesis (CH) takes place when mutant stem cells occur in healthy individuals and contribute disproportionately to the production of blood; it is increasingly recognized that individuals with CH are at markedly increased risk of subsequent development of hematologic malignancies, including t-MN.
PPM1D (protein phosphatase Mn2+/Mg2+–dependent 1D) is involved in the DNA-damage response pathway and is a regulator of p53 activity. Recurrent truncating mutations of PPM1D with gain-of-function activity have been detected in CH in normal populations, but also following chemotherapy1 and in patients with solid organ cancers. This association prompted Dr. Joanne Hsu and colleagues to investigate the PPM1D gene in patients with t-MDS/AML in a study led by the Goodell Laboratory. Using targeted-capture sequencing of 295 cancer genes on bone marrow samples of 156 t-MDS/AML patients, investigators found PPM1D mutations in 20 percent of patients (31 of 156) with only TP53 mutations occurring more frequently (28.8%; 45 of 156 patients). Strikingly, mutated PPM1D occurred in only one of 228 patients in a matched de novo MDS/AML cohort. Perhaps surprisingly, for two mutations proposed to act through the same pathway, presence of TP53 and of PPM1D mutations were not mutually exclusive, suggesting a distinct mechanism of action.
A significant association was observed with prior exposure to platinum agents and etoposide and topoisomerase inhibitors but not with other cytotoxic agents such as 5-fluorouracil. To investigate the mechanism by which platinum agents might promote t-MN associated with mutant PPM1D, in an elegant series of experiments the authors created PPM1D-mutant cell lines using a CRISPR-Cas9 system and showed mutant cells competitively expanded from 20 percent to more than 80 percent after exposing a mix of wild-type (WT) and mutant cells to cisplatin. This expansion was tempered by treating the cells with a combination of a PPM1D inhibitor and cisplatin, demonstrating the pivotal role of increased PPM1D expression in mutant cell growth advantage during cytotoxic therapy. This chemoresistance was specific to cisplatin, doxorubicin, and etoposide, with no mutant cell growth advantage over WT cells after exposure to vehicle, vincristine, or 5-fluorouracil. This mutant cell fitness advantage was likely due to resistance to apoptosis, rather than enhanced proliferation. The authors next explored the effect of PPM1D mutations on hematopoiesis in vivo by establishing a heterozygous PPM1D-mutant mouse model that was used for competitive serial transplantation experiments. PPM1D mutant cells showed normal contribution to all lineages in the hematopoietic hierarchy but the contribution to blood production increased over time after cisplatin exposure. In contrast, with different HSC stressors such as serial bone marrow transplantation without cisplatin exposure, PPM1D mutant cells reconstituted less effectively than WT cells with no selective growth advantage.
In this impressive work, Dr. Hsu and colleagues have clearly delineated that hematopoietic clones carrying PPM1D mutations are selected for following exposure to certain cytotoxic therapies, clearly implicating an important role for PPM1D in therapy-related leukemogenesis. Presence of PPM1D mutations in healthy individuals would suggest that this mutation alone is insufficient to drive leukemogenesis but is nevertheless enough to drive a degree of clonal expansion. This is somewhat at odds with the findings from the mouse model in this study, raising the possibility that certain environmental exposures or other factors might influence the impact of PPM1D mutations on hematopoiesis. Furthermore, the observation that t-MN patients typically had PPM1D mutations with a low variant allele frequency implies that the mutation is only present in a subclone of t-MN cells in most patients. Consequently, the exact role of this mutation in disease pathogenesis remains uncertain. Nonetheless, this study clearly advances our understanding of mechanisms of t-MN development and highlights the importance of specific mutation-therapy interactions during the development of secondary hematologic cancers.
Dr. O'Sullivan and Dr. Mead indicated no relevant conflicts of interest.