“Sowing dragon's teeth?” Myeloma and AML

In this issue of Blood, Mailankody et al demonstrate that patients with multiple myeloma (MM) or monoclonal gammopathy of uncertain significance (MGUS) have an increased incidence of developing acute myeloid leukemia (AML) or myelodysplastic syndromes (MDS).¹ 

Greek mythology describes the separate legends of Cadmus and Jason who, in the face of challenges, sowed dragons' teeth that grew into warriors. These initially uncontrollable and destructive warriors were eventually subjugated by Cadmus and by Jason in the respective legends. These stories have led to the phrase “sowing dragons' teeth,” which has been changed over time to mean that one's actions can result in unintended consequences.

Chemical agents have been associated with risk of cancer, often AML/MDS.²  Andersen and Videbaek³  and Kyle et al⁴  in 1970 reported the first associations between melphalan use and the development of AML in 4 MM patients described in each study. However, Kyle et al described 1 case of AML presenting with monoclonal gammopathy without clinical features of MM and in the absence of prior therapy. This case might have been MGUS. Indeed, other cases of monoclonal plasma cell disorders presenting simultaneously with AML in the absence of prior therapy have been described.⁵  Simultaneous or near simultaneous presentation of MM/MGUS and AML in the same patient were thought initially to be coincidental. In retrospect, it was reasonable to postulate that mutagenic chemotherapy such as melphalan would result in the development of second cancers such as AML/MDS. However, the Seligmann group might be considered prescient when they proposed that a common etiologic agent could be responsible for some cases of concomitant AML and MM/MGUS.⁵  This was based on clinical observation and preclinical research.^5,6  The latter, by Warner et al, demonstrated the induction of an AML cell line (the Walter and Eliza Hall Institute [WEHI] series) in mice by intraperitoneal injection of mineral oil and testosterone that normally would cause MM and plasmacytomas.⁶ 

The Mailankody et al paper is exciting because it demonstrates the value of a formal national tumor registry system for tracking all cancers and “premalignant” cases, including MM, AML, MGUS, and MDS diagnoses, and the power of an epidemiologic study that will guide future research in mechanisms of carcinogenesis. Their study confirms the strong association between MM and the development of AML/MDS. As a retrospective cohort study, it is limited by lack of information regarding patient treatment and timeline assumptions regarding the type of therapy. However, because of the introduction of therapy concomitantly over time in Sweden, timeline estimates might be made regarding the use of low-dose melphalan chemotherapy, high-dose melphalan chemotherapy with autologous hematopoietic cell transplantation (HCT), and immunomodulatory drug (IMiD) use. These registry data demonstrate that the increased risk (11.51-fold) of AML/MDS in MM patients has remained relatively constant over time, regardless of changes in MM therapy.

A novel finding by Mailankody et al is the 8.01-fold increased risk of AML/MDS in MGUS patients with IgG/IgA monoclonal proteins and that this risk was highest in patients with higher quantities of M-protein (> 1.5g/dL). Thus, without chemotherapy exposure, the hypothesis is that another underlying mechanism induces AML in MGUS patients. In preclinical mice experiments, mineral oil will induce plasmacytomas in nongerm-free mice. However, germ-free mice when injected with mineral oil do not develop plasmacytomas.⁷  Instead, the germ-free mice went on to develop more primitive lymphoreticular neoplasms several months later. Thus, environmental factors modulated hematologic tumor generation in these earlier experiments. Examining common signaling pathways in AML and MM cells may allow for better understanding of their genesis. An example of shared signaling is the mitogen-activated protein (MAP) kinase system (reviewed in Platanias⁸ ). The Raf/Mek/Erk signaling cascade mediates both leukemia and myeloma proliferation and may be important in malignant transformation. Previous work has demonstrated that there is a multistep transformation from MGUS to MM.⁹  Using microarray expression analysis, Davies et al demonstrated that there is a continuum of genetic changes that occurs from cells that are normal to MGUS to MM.⁹ 

Lastly, as noted in Mailankody et al, preliminary reports have noted an increased incidence of second cancers in MM patients receiving lenalidomide in randomized clinical trials as maintenance therapy after autologous HCT or primary induction therapy.^10-12  Three studies, 2 transplant and 1 nontransplant, have reported superior progression-free survival or time to progression for patients receiving long-term lenalidomide maintenance therapy. Concern regarding the second cancer incidence led to the closure of the lenalidomide arm of the Intergroupe Francophone du Myelome IFM 2005-02 study, whereas patients remain on lenalidomide in the US (Cancer and Leukemia Group B 100104) and Italian MY-015 cooperative group studies. The results reported by Mailankody et al, that MGUS is associated with an increased incidence of AML/MDS, suggests that endogenous factors (host and plasma cell/hematopoietic stem cell) may play a role in susceptibility to additional cancers. This susceptibility may be independent of or related to exposure to therapeutic agents or other environmental factors. When challenged with chemotherapy, the risk for development of AML/MDS may be increased as demonstrated by the increased incidence of AML/MDS in the Swedish MM registry cohort in the Mailankody study.¹¹  Thus, are chemotherapy agents used for MM therapy “dragons' teeth,” or does the beast of AML lie within the MM stem cell? This observational study from Mailankody et al provides evidence that therapeutic, environmental, and host factors/mechanisms all are involved. Further studies on the biology and pathogenesis of AML, MDS, MGUS, and MM will be the next step forward.