Investigating Racially Induced Metabolic and Genomic Signatures Mediating Inferior Survival in African Americans with Acute Myelogenous Leukemia

Background: African American (AA) patients experience worse survival outcomes compared to other patients with acute myelogenous leukemia (AML), independent of socioeconomic and molecular factors.¹ Indeed, our group previously demonstrated inferior survival in AA associated with the proxy for (1) genomic instability and (2) suggestion for inability to achieve protein damaging threshold among concurrent mutations.² In study expansion, we have confirmed worse overall survival (OS) for AA patients, and detected rapid attrition of AA patients, soon after diagnosis and therapy initiation. Given high incidence of cardiometabolic complications in AA patients, this prompted us to investigate the link between deranged metabolism and clinical outcomes. Additionally we explored nucleotide modifications that influence leukemogenesis in AA patients. In previous data, obesity correlated with poor outcomes for patients with acute promyelocytic leukemia (APL), but not in other AML, including in the CALGB cohort.^3–5 Interestingly, obesity is linked to 1) FTO/IRX3 pathways that block hematopoietic stem cell differentiation, and 2) oxidative stress, which may induce DNA damage favoring C>A and G>T nucleotide changes^6,7 and aberrant bone marrow microenvironment⁸.

Methods: We conducted retrospective analysis of 272 AML patients from Baylor TX, Georgetown DC, and Washington Hospital Center DC, including demographic and clinical variables (age, sex, race, BMI, comorbidities, smoking, treatments, and outcomes). APL patients were excluded. For patients with NGS data, mutation frequencies were annotated by ethnicity and prioritized by ELN 2022 classification. Chi-square and t-test evaluated differences in categorical and continuous variables respectively. Logistic binary regression was used to detect predictors for racial induced survival. SAS software was used.

Results: Our expanded cohort has 272 AML patients, with 43 AA and 229 non-AA. AA patients showed worse OS than non-AA patients (Fig 1; 149 days vs 605 days, p=0.01). There were not any AA patients with favorable risk ELN 2022 disease status. Neither ELN 2022 risk groups nor presence of MDS-like mutations differentiated worse outcomes for AA patients, so we opted to examine the effect of obesity among AA and non-AA patients with intermediate and high risk AML (given no favorable risk AA patients in our cohort). Mean BMI for our AA patients was 28, with AA patients with BMI>28 showing significantly worse OS (Fig 1; 67 days vs 198 days, p=0.001), which was not present for non-AA patients. Metabolic derangement with high total cholesterol (p=0.02) and LDL (p=0.04) also segregated with worse OS for AAs (n=8 with cholesterol data), but not for other races. Nucleotide changes C>A and G>T characteristic of oxidative stress were overrepresented in AA patients with high BMI suggestive of clinical trend (33% vs 11%, p=0.06).

Conclusion: Metabolic derangement in BMI and lipids correlated with inferior outcomes for AA patients only, supporting a unique pathobiology of highly aggressive AML that afflicts AA patients. Only AA patients demonstrated preference for C>A and G>T nucleotide changes, a hallmark of oxidative stress, also linked to obesity. These negative impacts support further mechanistic studies, and may present therapeutic strategies to improve metabolism, reduce oxidative stress, and suppress FTO/IRX3 pathways to induce myeloid differentiation.

Figure 1: Inferior Survival For African American Patients Differentiated by BMI>28

References:

1. Bhatnagar B, et al. Cancer Discov. 2021
2. Williams AK, et al. Blood. 2022;140(Supp_
3. Kashanian SM, et al. EJHaem. 2021
4. Castillo JJ, et al. Am J Hematol. 2016
5. Castillo JJ, et al. Leuk Res. 2012
6. Usman M, et al. Int J Obes (Lond). 2021
7. Poetsch AR. Comp and Struct Biotech J. 2020
8. Cheng F, et al. Trends in Cancer. 2023