How lipid coating soothes the gut in AML therapy

In this issue of Blood, Renga et al,¹ in elegant experimental mouse models, compared the impact on gut microbiota, fungal colonization, and intestinal mucosa integrity of 2 intensive induction acute myeloid leukemia (AML) therapies. A surrogate for the “7+3” regimen of cytarabine infused continuously over 7 days with an anthracycline (eg, daunorubicin) injected intravenously 3 days apart was compared with repeated administration of the recently approved CPX-351, a liposomal construction encapsulating these 2 pivotal anti-leukemia drugs at a fixed molar ratio.¹ 

These experimental results are highly relevant clinically. As shown independently by several teams, intestinal microbiota of patients with AML is markedly disrupted during a classic 7+3 induction course.^2,3 This digestive microbial dysbiosis appeared to correlate with a higher risk of bloodstream infections and other microbiologically documented infections.² These changes seemed to be protracted and impact later phases of therapy.³ They partly persisted when patients with AML are referred for allogeneic hematopoietic stem cell transplantation after achievement of a complete remission. Strikingly, intestinal dysbiosis, which usually but not consistently worsened during the periengraftment period, appeared to be a strong independent predictor of nonrelapse mortality and overall survival when measured at engraftment in patients receiving a T-cell replete graft in a large multicentric cohort of patients who underwent transplantation.⁴ 

CPX-351 was recently approved for first-line treatment of AML with myelodysplastic-related changes or therapy-related AML. Experimental data produced by Renga et al would explain why CPX-351 might induce less gastrointestinal toxicity and a lower early mortality than 7+3, as suggested (but bot definitely demonstrated) by the results of the pivotal trial and a real-life retrospective study.^5,6 As already observed by Hueso et al, mice receiving several repeated injections of cytarabine and daunorubicin developed intestinal dysbiosis with the expansion of bacterial phyla associated with intestinal inflammation and the loss of other commensals.⁷ These mice also experienced a severe intestinal pathology with ileocolic infiltrates, increased intestinal permeability to endotoxins, and bacterial translocation into mesenteric lymph nodes or liver.^1,7 The latter 2 might be readout of clinical relevance correlating with the infectious risk. The mice became less resistant to colonization by Candida or Aspergillus. Furthermore, as also observed by Renga et al in mesenteric lymph nodes, expression of inflammatory cytokines was upregulated, and both T helper cell 1 (Th1) and T helper cell 2 (Th2) lymphocytes were activated. In sharp contrast, mice treated by CPX-351 did no develop intestinal dysbiosis and mucosa histologic damage or hyperpermeability. Th1 or Th2 adapted pathways were not upregulated, whereas mesenteric Forkhead box P3⁺ regulatory T cells were activated.¹ 

CPX-351 was also effective in prevention of dextran sulfate sodium–induced ileocolitis, a classic model of intestinal inflammation. This protective effect could be at least partly explained by the preservation of intestinal microbiota. Fecal microbiota transplantation (FMT) from CPX-351 treated mice prevented macroscopic and histologic damages observed in this colitis model, whereas feces from mice treated with a free 7+3-like combined chemotherapy worsened them.

Further ex vivo experiments confirmed that anthracyclines (even administered alone) have not only an antibiotic effect but are highly toxic for intestinal epithelial cells. In contrast, empty liposomes mimicking the lipidic envelope of CPX-351 recapitulated alone the protective effects to lipopolysaccharide damage. Ex vivo and in vivo experiments also helped decipher the immunologic mechanisms involved, including both innate and adoptive immune cells and explain, at least in part, the opposite impact on intestinal microbiota and mucosal barrier of 7+3 vs CPX-351. These therapies modulated in opposite directions to the interleukin (IL)-22/aryl hydrocarbon receptor (AhR) and IL-10 pathways, which are known to protect both intestinal microbiota and epithelial stem cells supporting mucosal regeneration.^8,9 Interestingly, recombinant human IL-22–IgG2 Fc chimeric protein is currently being tested as treatment of acute lower tract gastrointestinal graft-versus-host disease (NCT02406651). Upregulation of IL-10, IL-22, and AhR production induced by liposomes could be a plausible mechanism of action of CPX-351 (or empty liposomes), and explains its ability to mitigate digestive toxicity of the encapsulated daunorubicin and break the vicious circle between digestive barrier damage and dysbiosis.

Finally, the study by Renga et al explored the crosstalk between intestinal mucosa and microbiota. IL-22/AhR and IL-10 produced by immune cells in CPX-351–treated mice protect commensal bacteria. Reciprocally, microbial metabolites (including indole-3 aldehyde, which is an AhR-activating ligand) measured at higher levels in serum and/or feces of CPX-351–treated mice might protect the intestinal barrier and provide resistance to fungal colonization.¹ 

Further studies should explore how important the IL-22/AhR and IL-10 pathways are in the intestinal protection provided by CPX-351. The molecular link explaining how liposomal lipids can impact both intestinal microbiota and immunity still deserves further clarification. One puzzling aspect of this study is the correlation shown between an increase in microbiota (α) diversity induced by 7+3, intestinal dysbiosis, and damage. Loss of α diversity is usually a reasonable surrogate for intestinal dysbiosis and pathology. Moreover, FMT experiments did not document the dynamics of engraftment of donor bacterial strains, which might be different in humans than in mice.

Despite their limitations, these experimental data should encourage prospective studies comparing fecal microbiota in patients with AML treated by 7+3 or CPX-351. Such studies would document more accurately digestive or infectious complications and their impact on patient outcome, including after allogeneic transplantation. They should include other confounding factors likely to impact gut microbiota, such as antibiotic treatment. Future clinical studies may also explore how previous chemotherapy and its damage modulate the efficacy of emerging immunotherapeutic agents (eg, the impact of intestinal dysbiosis on the outcome of chimeric antigen receptor T-cell therapy).¹⁰ 

Conflict-of-interest disclosure: R.B. is an employee of the Acute Leukemia French Association collaborative group and a former advisor and employee of Da Volterra.