Autoimmune Disorders and the Development of Myeloid Disease: Should We Always Blame the Therapy?

A 64-year-old woman was referred to our clinic with pancytopenia and fatigue. The patient reported a four-year history of systemic lupus erythematosus (SLE; synovitis, malar rash, oral ulcers, pleural effusions, anti-Smith antibody) that had previously been treated with pulse cyclophosphamide and concomitant oral prednisolone followed by maintenance azathioprine for severe lupus nephritis. The bone marrow (BM) aspirate at diagnosis showed 56 percent blasts consistent with acute myeloid leukemia (AML). BM cytogenetics revealed a chromosome 7 deletion. Azathioprine was promptly discontinued. Molecular testing revealed lack of NPM1, FLT3, CEBPA, or TP53 mutations and presence of pathologic mutations in DNMT3A and ASXL1. After detailed discussion with the patient about lower- and higher-intensity therapy options in the context of adverse cytogenetics with monosomy 7, she elected to proceed with intensive chemotherapy, to be followed by allogeneic hematopoietic stem cell transplantation.

This case illustrates one of the dreaded downstream events in individuals with severe autoimmune disease (AD) — the development of therapy-related myeloid neoplasm (t-MN). An elevated risk of AML/myelodysplastic syndromes (MDS) has been noted across several ADs, including SLE, rheumatoid arthritis (RA), inflammatory bowel disease, and multiple sclerosis. Large scale population-based studies have shown the risk to be between 1.3 and 2.1 higher in patients with a previous history of AD compared to population baseline.^1,2  Historically, this risk has been tied to given cytotoxic therapies, but that is likely too simple. Rather, current thought is that genetic susceptibilities, immune stimulation of the BM, and immune surveillance defects cooperate in the development of t-MN. Supporting this is the early emergence of myeloid malignancies in the natural course of AD and development of MDS/AML among patients who have had no prior treatment exposure for their AD.

Etiologies

One of the first questions that this patient and others will ask is why t-MN occurs? For years, many hematologists blamed prior treatment; however, it is increasingly understood that there is a multifactorial pathogenesis to AD-associated MN development. The risk of t-MN varies by the type of AD and the duration and type of genotoxic agent exposure. Among the commonly used drug classes in the management of AD, secondary leukemogenesis has been described extensively with the use of alkylating agents (cyclophosphamide), antimetabolites (azathioprine), and topoisomerase inhibitors (mitoxantrone).^3-5  For example, prior azathioprine exposure has been associated with as high as a sevenfold elevated risk compared to the population baseline, possibly due to direct DNA mutagenic effects. However, it is not just the treatment. Leukemogenic susceptibility to genotoxic agent exposure is strongly influenced by individual genetic predisposing factors. These factors include metabolization phenotypes, DNA damage repair pathways, and possibly the concomitant presence of clonal hematopoiesis of indeterminate potential (CHIP). Patients with preleukemic CHIP have a significantly higher risk of developing t-MNs than patients without clonal hematopoiesis, and screening for CHIP at the time of primary cancer diagnosis can help identify patients at risk for t-MN.⁶  The risk of t-MN in breast cancer and Hodgkin lymphoma survivors is correlated with the extent of radiotherapy and the duration and type of chemotherapy.^7,8  Importantly, the elevated risk of MN with conventional cytotoxic therapiess does not seem to be shared by biologic therapies such as tumor necrosis factor-alpha (TNF-α) inhibitors.⁹ 

What about treatment-naïve patients? One hypothesis is that HLA-associated susceptibility can partially explain the occurrence of AML in patients with AD who have never been treated or who avoid cytotoxic agents. For example, HLA-B27 carrier status has been associated with an increased predisposition to both AD and AML.¹⁰  Another important molecule implicated in hematologic malignancies (HMs) involving the myeloid lineage is interleukin-1. Polymorphisms within the interleukin-1 receptor antagonist gene have been associated with both AD and secondary AML.¹¹ 

The BM microenvironment is known to play an important role in AML/MDS pathogenesis and progression. A proinflammatory tumor microenvironment can contribute to overall tumor progression by promoting various aspects of cancer cell proliferation and survival. A central proinflammatory mediator, implicated in both ADs¹²  and leukemias,¹³  is nuclear factor-κ B (NF-κB). Constitutive NF-κB signaling can be achieved either through intracellular autogenic activation or extrinsically by cytokine factors produced in the tumor microenvironment milieu. Finally, leukemic myeloid cells develop a variety of escape mechanisms to evade peripheral T-cell immune surveillance and destruction. Further impairment of a disrupted T-cell immunologic surveillance by exposing the BM to immunosuppressive agents could potentially lead to their clinical emergence.

SLE and t-MN

Only a few studies have specifically investigated the risk of myeloid leukemia in SLE, which is the pre-existing condition in this case. Analyses by Dr. Lena Björnådal and colleagues¹⁴  and Dr. Arti Parikh-Patel and colleagues¹⁵  reported an elevated risk of myeloid leukemia in SLE, with a standard incidence ratio of 3.4 and 2.96, respectively. Similarly, Dr. Björn Löfström and colleagues reported an increased risk of AML in a Swedish national cohort of 6,438 patients with SLE.¹⁶  Notably, the leukemia risk in this study was confined to the patient subset with preceding prolonged cytopenias, especially leukopenia. The median latency period between SLE and leukemia diagnoses was five years. The risk of myeloid leukemia was restricted to the subgroup characterized by more men and older age at onset of SLE. Further, the study did not identify a difference in the frequency of cytotoxic exposure between the case and control cohorts, suggesting that prior cytotoxic exposure is not a major cause for AML development in SLE.¹⁶  Another study investigated the effect of latency of AML development in SLE and showed a decremental risk for myeloid leukemias with longer SLE latency.¹⁴ 

Newer Treatments for AD?

Another question that arises, especially during consultation on cases of AD, is whether or not there are increased risks when the agents dubbed “disease-modifying antirheumatic drugs” (DMARDs) are used. These medications are commonly used in patients with RA and in conditions such as ankylosing spondylitis, psoriatic arthritis, and SLE. While data on the increased risk of development of certain cancer types (e.g., melanoma, nonmelanomatous skin cancers, and lymphomas in patients with RA treated with anti–TNF-α therapy) are conflicting, there is no substantial increase in the risk of leukemias in patients treated with anti–TNF-α therapy as compared with those treated with any nonbiological DMARDs.¹⁷  In a Swedish Cancer Registry study by Dr. Johan Askling and colleagues, a significant association between RA and AML risk was observed only in the inpatient advanced and early-arthritis cohorts, but not in the TNF-α–blocker group, which argues for a DMARD approach as a critical risk factor in AML development.¹⁸  Data from the same study also suggested no effect of RA latency on the risk for AMLs. While a significant, albeit weak, association between DMARD use and HM risk has been noted with azathioprine and cyclophosphamide therapies, the occurrence of leukemia cases even among patients who had had no prior exposure to cyclophosphamide or azathioprine suggests a link to RA-related immune stimulation itself.¹⁹ 

Future Directions

The excess risk of MNs in AD varies by the particular type of AD, reflecting biologic differences between AD entities. Unfortunately, studies designed to investigate the role of AD-directed therapy in MN risk have yielded conflicting results, and data correlating patient and AD features with subsequent MN development are not forthcoming. Additionally, we still do not have a complete understanding of the molecular defects underpinning secondary leukemogenesis in AD. Important future research agendas include evaluation of molecular defects and identification of risk factors associated with MN development in AD. A more detailed characterization of biologic mechanisms through focused efforts directed toward delineating pathophysiologic pathways will not only further our understanding of the association between the two entities but also help identify patients who are at a higher risk for developing post-AD MNs and who may therefore benefit from preemptive strategies, especially now with the availability of potentially safer and more effective biologic treatment alternatives.