Introduction

Sickle cell disease (SCD) corresponds to an autosomal recessive hemoglobinopathy in which structurally abnormal hemoglobin (HbS) leads to chronic hemolytic anemia and to a variety of severe clinical manifestations. The disorder is caused by a point mutation. A single DNA base change leads to substitution of valine for glutamic acid at the sixth position on β globin chain. Patients with homozygous hemoglobin (SS) often present with severe symptoms, while those with a heterozygous mutant allele (SA) demonstrate minimal clinical symptoms. The combination of hemoglobin S with another type of β subunit gene mutation, such as hemoglobin C or β thalassemia, forms a compound heterozygous hemoglobinopathy (SC or Sβ0). With the exception of the compound Sβ0, the heterozygous genotypes are usually less clinically severe than hemoglobin SS [1].

Since Herrick’s description of SCD in 1910 [2], a wide variety of malignancies, including hematological neoplasms, have been reported in both children and adults with SCD. However, the exact incidence of malignancy has not been accurately determined due to a lack of long-term follow-up. The first description of SCD coexisting with acute leukemia has been reported by Goldin et al. in 1953 in a 38-year-old black man with SCD and acute myeloid leukemia (AML) [3]. Since then, the occurrence of acute leukemia has been reported in several cases of patients with SCD.

We reported here a new case of SCD patient who developed AML and reviewed extensively the literature in order to better understand the relationship between the two diseases. This review leads to the hypothesis of a mechanism involving multifactorial causes through the pathophysiologic mechanisms of the clinical manifestations of SCD.

Patients and methods

Case selection

A sole case of acute leukemia in the setting of SCD was retrieved from the pathology database of the “Centre de Référence Constitutif des pathologies du globule rouge et de l’érythropoïèse” in Lyon (France), including a pool of more than 600 adults and children with SCD. The diagnosis of leukemia was confirmed according to the World Health Organization classification [4]. Informed consent for reporting this case was obtained from this patient in accordance with the declaration of Helsinki. Clinical history and laboratory data, including flow cytometry analysis, cytogenetics, and molecular biology, were collected as well as data regarding SCD history.

Literature data sources

The PubMed database was searched on October 2022 for case reports previously published involving both SCD and acute leukemia. The relevant keywords used were: “sickle cell disease” or “sickle cell anemia”, combined with “acute leukemia”, or “myelodysplastic syndrome” (MDS). Fifty-one previously published cases were identified since 1972, and the relevant data regarding acute leukemia and SCD were collected and analyzed (Tables 1 and 2). These cases did not included those only mentioned in the epidemiologic reports from California and the United Kingdom [5, 6].

Table 1 Acute leukemia characteristics and outcome of the 51 patients identified in the literature
Full size table
Table 2 SCD characteristics of the 51 patients identified in the literature
Full size table

Results

Case report

A 27-year-old woman of African origin with known SCD (Sβ0), previously complicated by recurrent severe vaso-occlusive crisis (VOC) and acute chest syndrome despite hydroxyurea (HU) therapy (1000 mg/day for 7 years) and regular exchange transfusions, presented on May 2022 a progressive bicytopenia with anemia to 50 g/dL (based-hemoglobin level under compliant treatment with HU around 75 g/dL) and thrombocytopenia to 50 × 109/L, leading to HU discontinuation. A suspicion of MDS was confirmed by a first bone marrow sample showing a hypercellular marrow with signs of dyserythropoiesis with demonstration of ring sideroblasts on a Perls’stain, dysmegacaryopoiesis, and dysgranulopoiesis, but no leukemic cells. The patient was referred to the Hematology Department and a repeat bone marrow aspirate, performed sequentially showed a progressive blast increase up to 25% leading to the diagnosis of AML-MRC (myelodysplastic related changes). The immunophenotypic profile was CD34+/− CD38+/− CD123+/−, CD13+/− CD33+/− CD117++, HLADR+/−, CD36+/− CD71++, CD7− CD19− CD56−, MPO−. Seventeen percent of myeloblasts expressed a multipotent progenitor-like leukemia stem cell (LSC) profile CD34+ CD38− CD90++/− CD45RA+/− CLL1/TIM3/CD97+/−. Cytogenetic analysis showed a complex karyotype: 44–46, XX, der(1)t(1;12)(q31;q15)add(1)(p11), − 3, der(5)t(5;7)(q13;q31), − 7, del(12)9q21), der(15)t(?3;15)(q21;p13), add(16)9p13), − 22, + 3-6mars[cp18]/46, XX [2]. Molecular study by next-generation sequencing (NGS) identified the presence of a TP53 mutation c.1024C > T with a variant allele frequency (VAF) of 0.64. The patient received induction chemotherapy with Vyxeos (daunorubicin/cytarabine) at a dose of 44 mg/m2 on days 1, 3, and 5. On day 22, peripheral blood showed 18% blasts signing remnant leukemia. Salvage chemotherapy combined mitoxantrone 6 mg/m2/day, etoposide 80 mg/m2/day, and intermediate-dose cytarabine 1 g/m2/day from day 1 to day 6. Salvage chemotherapy was complicated by infections including inguinal cellulitis requiring large spectrum antibiotics and white blood cell infusion therapy, pericarditis, and posterior reversible encephalopathy syndrome (PRES) leading to a transitory hospitalization in intensive care unit. Cytological remission was achieved, but measurable residual disease (MRD) remained positive at 0.09% based on leukemia associated immunophenotype (LAIP)/LSC. Allogeneic phenol-identical hematopoietic stem cell transplantation (HSCT) with one mismatch was performed on January 2022 based on thiotepa-busulfan-fludarabine (TBF) conditioning regimen followed by post-transplant cyclophosphamide and everolimus for graft-versus-host prophylaxis. The hospitalization was complicated by a septic shock (Klebsiella pneumonia) and by invasive pulmonary aspergillosis and severe hepato-splenic candidosis (Candida glabrata). Bone marrow evaluation at one month and two months post-transplant confirmed the cytological remission with MRD negativity assessed by multi-parameter flow cytometry and total donor chimerism.

Review of the literature

Fifty-two cases of acute leukemia in SCD patients (including our case report) were identified in the literature since 1972 (Tables 1 and 2). Among patients with available data, male/female sex ratio was 0.45. Median age was 23.5 years (range: 3 – 61 years). Thirteen patients (25%) had acute lymphoblastic leukemia (ALL), one patient an undifferentiated acute leukemia, and 38 patients (73%) a myeloid neoplasm, including 16 AML, 6 MDS and 16 MDS/AML. Among the 26 patients studied for genetic markers, two patients with ALL showed a Philadelphia chromosome (Ph +) (#9, #49), one patient with AML had a normal karyotype (#12), two had acute promyelocytic leukemia (#21, #26), and 20 patients with MDS and/or AML displayed unfavorable cytogenetics [− 5, − 7, del(17), 11q23, chromosome 3 abnormality] and/or molecular abnormalities of poor prognosis [TP53, KMT2A, RAS, RUNX1, PTPN11] (#6, #11, #22–25, #28-#32, #37, #43–46, #48, #50, #51, #52). Most of the patients (84%) displayed a SS homozygous hemoglobin, while only 9% were SC and 7% Sβ0. Data were available in 38 patients regarding the potential use of long-term SCD therapy with HU: 16 (43%) did not receive any HU, while 22 (57%) received HU prior to acute leukemia diagnosis (median duration of treatment: 6.5 years; range: 0.05 – 17 years). Ten patients underwent allogeneic HSCT, after conditioning regimen including alkylating agents and/or total body irradiation, as treatment of SCD. Four patients were allografted from a matched sibling donor (#16, #32, #36, #45) and six patients from a haploidentical donor (#28, #33–35, #44, #46) (Table 2). The median time between HSCT and acute leukemia diagnosis was 2.5 years (range: 0.26 – 7 years). Two cases of AML developed in SCD patients who had been treated by gene therapy with LentiGlobin [34, 36], which required myelo-ablation with an alkylating agent.

Overall acute leukemias in SCD patients were of dismal outcome with overall survival (OS) ranging from few days to 2.5+ years (median in patients with available data: 7 months).

Discussion

Historically, the development of malignancy in children and adults patients with SCD has been documented by several small series [7, 12, 38]. On the basis of a single institution study, the cancer incidence in SCD patients has been estimated to be 1.74 cases per 1,000 patient-years [39]. Malignancies mainly included hematological neoplasias, especially acute leukemias. In the 1970s, Jackson reported, among 58 black children treated for acute leukemia, 4 ALL and 3 AML with sickle cell trait, and one ALL with homozygous HbS [7]. In a low-income country, the association of acute leukemia with SCD was even reported in 8.6% of cases [30]. Actually, the risk of hematologic malignancies is 2 to 11 times as high as that in the general population. This was established by three recent epidemiology reports [5, 6, 19]. The first study used a standardized incidence ratio (SIR) to compare individuals with SCD to the general population. One hundred and fifteen on 6423 SCD individuals were diagnosed with cancer, with a total of 6 AML cases (SIR, 3.59; 95% confidence interval: 1.32–7.82) and 3 cases of ALL (1.83; 0.38–5.35) [5]. In the second study, 8 cases of AML on 7512 individuals with SCD were reported (11.05; 3.86–30.17). Among hematological malignancies, the risks remained elevated for all conditions studied, except for lymphoid leukemia [6]. The third study identified 52 cases of cancer in 49 patients among 16,613 individuals with SCD, 40% of cases occurring in children [19]. The most frequent malignancy was acute leukemia (8 cases).

The vast majority of SCD patients receive conservative therapy. In this setting, HU has greatly improved the survival of SCD patients in developed countries, due to its efficacy in preventing VOC via an inhibitory effect on HbS polymerization by increasing the synthesis of fetal hemoglobin, and an improvement of blood flow in the microcirculation through the expression or activity of several adhesion molecules on red blood and endothelial cells [40, 41]. Three randomized placebo controlled trials have demonstrated the efficacy of HU in SCD, with an excellent safety profile and up to a 40% reduction in mortality after 9 years of follow-up [42,43,44]. HU is an inhibitor of DNA synthesis that may theoretically lead to an accumulation of acquired DNA mutations and eventually leukemic transformation. Whether acute leukemia in SCD patients with long-term exposure to HU is a co-incidental or related to therapy has been a major issue debated in many reports. The leukemogenic risk could theoretically increase with the duration of drug exposure. The index of DNA damage in peripheral blood leukocytes from HU-treated patients with SCD was demonstrated higher than in controls and was confirmed influenced by the duration and the dose of HU treatment, and by the HbS genotype [45, 46]. The leukemic risk of HU has never been confirmed in patients with chronic myeloproliferative diseases [47, 48], and no increased risks of malignancy were reported in large series of SCD patients [49,50,51]. Among 278 SCD children receiving long-term treatment with HU, only one developed acute leukemia [15, 52, 53]. If one study in pediatric SCD patients treated with HU showed that genotoxicity increased with HU administration [54], it was demonstrated that individuals may have different susceptibilities to HU, and that this occurred in a patient population that may already have an elevated risk for malignancy evaluated at baseline by a greater Damage Index [55]. Overall the genotoxicity results clearly demonstrate that HU does not directly bind DNA and is not mutagenic [56]. In vitro, HU can result in the accumulation of somatic mutations and chromosomal damages due to interference with DNA repair, but the number of acquired mutations did not increase in patients with long-term exposure to the drug [57]. On another hand, HU therapy can alleviate the risk of chronic hemolysis by increasing the fetal hemoglobin content in the blood, and potentially reduce the accompanying marrow stress in these patients. The recent prospective observational study ESCORT-HU (NCT02516579), which evaluated the long-term safety and effectiveness of HU in SCD patients across several European centers, confirmed the benefit-to-risk ratio of HU in children and adults [58]. Only one incident hematological malignancy was reported.

In contrast to life-long supportive care measures, HSCT offers a curative option but may be followed by various severe complications. It is therefore being reserved to patients who are refractory to conventional therapy. The 5-year OS ranges from 91 to 95% in children who underwent HLA-identical HSCT after myeloablative conditioning, while disease-free survival, rate of rejection, and incidence of chronic graft versus host disease are approximately 82%, 8%, and 12%, respectively [59, 60]. Nine percent of patients died of complications related to transplantation [59]. Peripheral blood stem cells, which would come from AA or AS donors, have also been proposed as a source of stem cells for allogeneic HSCT. The results after related (5-year OS: 97%) and unrelated (2-year EFS: 90%) donor umbilical cord transplantation (UCT) have also been encouraging [61]. However recent results were more discouraging showing a high incidence of graft rejection (50% to 62%) after unrelated UCT [62], although updated data using a reduced intensity conditioning combining HU, alemtuzumab, fludarabine, thiotepa, and melphalan were more impressive [63]. Haploidentical-related donor transplantations are under study. It is becoming a viable alternative curative option for SCD, extending the availability of HSCT as a treatment option to eligible SCD patients. Overall survival was high (91%) in all studies included in a recent meta-analysis [64]. One study has suggested that HSCT for SCD does not increase the risk of developing acute leukemia, compared with patients who have not undergone SCT [60]. However, transplanted patients are generally exposed to alkylating agents and ionizing radiation as part of a conditioning regimen, and intervals, found in the literature, between the procedure and the diagnosis of leukemia are falling in the range of latency reported in other diseases. Furthermore, therapy-related MDS/AML is a well known event after autologous transplantation for lymphomas, with cumulative risks as high as 15%.

Trials in gene therapy are under way and also offer great promise. However, the largest lentiviral vector-mediated β-globin replacement gene therapy trial in SCD reported two cases of adult patients diagnosed with AML [34, 36, 65, 66]. These two cases shared similar cytogenetic and molecular abnormalities with monosomy 7 and RUNX1 and PTPN11 mutations, which were not found in patients pre-conditioning bone marrow samples. The first case was considered to be related to busulfan conditioning [34]. The second case showed vector present in leukemia blast cells, which suggests that blast cells originated from a transduced hematopoietic stem cell and not from residual host cells exposed to busulfan [36]. However, several lines of evidence showed that the development of this case of AML most likely occurred independently of insertional oncogenesis [36].

If a coincidental event between acute leukemia and SCD could be evocated in several cases from the literature that resemble de novo acute leukemia, an increased risk for acute leukemia is suggested by the significant underlying MDS features of most reported cases compared to leukemic patients from the general population of the same age. Extensive literature review demonstrates at least 18 patients with presence of complex structural rearrangements involving complete or partial loss of chromosome 5 and/or chromosome 7 and/or 17p deletions. TP53 gene mutations have also been shown frequently implicated [67]. Furthermore, several cases were classified as AML6 or AML7 and/or presented bone marrow fibrosis. Those facts are not in favor of a simple coincidence between the occurrence of acute leukemia and SCD, but are generally considered as a marker of secondary leukemia.

The exact underlying connection between acute leukemia and SCD is not clearly understood. Beside therapy for SCD, other potential cancer risk factors might exist for SCD patients and have been discussed in a recent published commentary [68]. Red blood cell transfusions can lead to increased iron levels and non-specific immunomodulation that could increase the risk of malignancy. However, heavily transfused patients with thalassemia only show a few cases of cancer [69]. Chronic inflammation implies the potential involvement of inflammasomes in SCD pathogenesis [70]. Chronic organ damage with inflammation could also cause cellular damage with subsequent malignant transformation. The pro-tumorigenic role of inflammasomes is associated with promoting cell proliferation, inhibition of apoptosis, and an immunosuppressive effect on the immune cells. Constant hematopoietic hyperplasia, stimulated by a hemolysis-induced cytokine storm, may increase the risk of somatic mutations, resulting in transformation of myeloid precursors [71]. Other factors associated with the increased risk include increased risk of infections, and increased bone marrow turnover, which form the pathophysiologic mechanisms of the clinical manifestations of SCD [5, 72]. The accumulation of multiple genetic abnormalities over years, due to a high degree of proliferative activity of bone marrow cells, may be responsible of the increased risk of cancer.

After myeloablation, the bone marrow niche undergoes extensive proliferation of hematopoietic stem cells, generating proliferative stress that may lead to mutations as part of the normal engraftment process [73]. After HSCT, myeloid malignancy was only seen within patients who did not engraft [28, 32]. In case of graft failure, the need for more replication cycles is required to repopulate the bone marrow, increasing the probability of acquiring a mutation that could lead to AML. TP53 mutations were detectable in blood before transplantation and increase until therapy-related myeloid malignancy diagnosis [32]. The progression of baseline high-risk TP53 clonal abnormalities into AML in patients with SCD has been reported after unsuccessful allogeneic HSCT. It has been previously demonstrated that the TP53 mutated clones specially expanded after chemotherapy exposure [74]. Because of erythropoietic stress and systemic inflammation, SCD patients may have been predisposed to developing clonal hematopoiesis. As these clones may be more resistant to radiation and/or chemotherapy, it has been suggested that they may preferentially expand after a failed transplant, leading to the myeloid malignancy detected after graft rejection.

TP53 is the most commonly mutated gene in therapy-related MDS/AML. Low folic acid, associated with an increased risk for leukemia, can make cells vulnerable to mutagenesis and can affect the genetic and epigenetic integrity of TP53 [75]. TP53 plays a central role in regulating cellular responses to genotoxic stress, and loss of TP53 provides a selective advantage for neoplastic growth [76]. The specific TP53 mutation has been shown to be present at low frequencies (0.003–0.7%) in blood leucocytes in some cases 3–6 years prior to the development of therapy-related MDS/AML and prior any chemotherapy [74]. TP53 mutations have also identified in small populations of peripheral blood cells of healthy chemotherapy-naïve elderly individuals. Chromosomal aberrations were demonstrated in some SCD patients with no evidence of hematological disease [27]. Furthermore, murine bone marrow chimeras containing wild type and TP53+/− hematopoietic stem/progenitor cells preferentially expanded after exposure to chemotherapy [74]. These data suggest that TP53 mutations precede the development of AML and the acquisition of other mutations, such as TET2, NUP98, or RUNX1.

Despite limitations coming from the retrospective nature of our study involving missing data and biases related to cases reported over an extended period, our review of the literature tend to suggest that chronic hemolysis, increased iron levels, and increased bone marrow turnover, which form the pathophysiologic mechanisms of the clinical manifestations of SCD are mainly responsible for a situation in which cells are undergoing constant hematopoietic hyperplasia, leading to the increased risk of acute leukemia by inducing genomic damage and somatic mutations [77]. The effects of SCD on progenitor cells have not been fully determined [78]. SCD may promote accelerated aging of hematopoietic cells and oncogenic somatic mutations [79]. Further studies are needed to identify risk factors for developing acute leukemia by pre-screening individuals with SCD. Next-generation DNA sequencing can be used to detect expanded peripheral blood progeny of a mutant clone and clonal hematopoisis of indeterminate potential (CHIP), which is a risk factor for subsequent hematologic malignancy [80]. Recent large studies have tried to address clonal hematopoiesis in SCD [81, 82]. Despite different conclusions related to the technique used, the control cohort chosen, and the value of VAF defined for considering clonal hematopoiesis, a small percentage of cases were identified as having somatic variants of TP53, DNMT3A, ASXL1, and/or TET2.

In conclusion, several cases of MDS/AML have been reported in SCD leading to the hypothesis that SCD may lead to the development of hematopoietic malignancies, even in the absence of disease-modifying treatments. The increased risk of leukemogenesis is certainly multifactorial and related to the pathophysiologic mechanisms of the clinical manifestations of SCD, which may promote accelerated aging of hematopoiesis. A prevalence of clonal hematopoiesis in SCD patients should demonstrate a higher risk than in the general population.