A “gut feeling” about precursor B-ALL

In this issue of Blood, Vicente-Dueñas et al¹  elegantly demonstrate the interplay of genetic predisposition, altered gut microbiome, and delayed infection in development of precursor B-cell acute lymphoblastic leukemia (pB-ALL) in Pax5^+/− mice. Many childhood leukemia-initiating events, such as chromosomal translocations, originate in utero (reviewed in Greaves² ), and studies so far suggest that these “first hits” occur at a far higher frequency than the prevalence of childhood leukemia itself.^3,4  Why only a fraction of children born with such preleukemic “hits” develop full-blown leukemia is a puzzle. Although the causes are likely to be multifactorial,⁵  1 theory put forward has been that of a disordered immune response either upon delayed exposure to infections in childhood or to altered natural microbiota (reviewed in Greaves⁵ ). Vicente-Dueñaset al show that the gut microbiome is distinct in genetically predisposed mice (Pax5 heterozygosity or ETV6-RUNX1 fusion), and that alteration of this microbiome can trigger pB-ALL in Pax5^+/− mice even in the absence of infection (see figure).

Previous reports by the authors⁶  suggested that exposure to infections is a causal factor in the development of pB-ALL in a mouse model of Pax5-inherited susceptibility. In this new study, the authors have investigated the role of the gut microbiome in the development of acute lymphoblastic leukemia (ALL) in genetically predisposed mice in specific pathogen-free (SPF) and conventional animal housing facilities, with or without antibiotic treatment. In addition to confirming their previous results that Pax5^+/− mice moved from SPF to CF conditions develop pB-ALL, the authors demonstrate that the gut microbiome of both wild-type (WT) and Pax5^+/− mice changed dramatically over time after transfer from SPF to CF facilities. However, the gut microbiome of Pax5^+/− mice remained different from WT controls in both conditions. The authors were able to use machine learning to predict the genotype of WT and Pax5^+/− mice from the microbiome data. These differences in the gut microbiome were also seen in mice with ETV6-RUNX1 fusion, although the small number of mice studied meant that no conclusions could be drawn about pB-ALL development in CF. Hematopoietic cells transplanted from Pax5^+/− mice to WT mice altered the gut microbiome of recipient mice, suggesting that preleukemic cells could play a specific role in shaping the gut microbiome. In addition, Pax5^+/− mice had a genotype-specific metabolomics profile that correlated with the altered gut microbiome, when compared with WT controls.

Depleting the gut microbiome by antibiotic treatment did not prevent infection-driven pB-ALL development in Pax5^+/− mice in CF, but remarkably, this intervention promoted pB-ALL development in an SPF environment, even in the absence of infectious stimuli. This reinforces the fact that an intact microbiome is crucial in preventing leukemic transformation in predisposed mice, with or without exposure to infections.

These results bring us 1 step closer to understanding the potential role of the microbiome and infections in the transformation of preleukemic clones in early childhood. The authors of this study⁷  and others (Valeria Cazzaniga, Anthony Ford, Mel Greaves, written personal communication, August 2020) have been using ETV6−RUNX1⁺ mouse models to understand the implications of delayed infections in childhood, and how an abnormal immune response to these might provide the “second hit” for leukemic transformation. However, this current study has demonstrated an increased risk of leukemia even in the absence of infections, by merely depleting the microbiome in genetically predisposed mice. This has implications for antibiotic usage in early life⁸  especially in those infants with a genetic predisposition to pB-ALL (these, including germline PAX5 mutations,⁹  are only a very rare subgroup in childhood ALL), or more importantly those born with a preleukemic “first hit.”

Although we must bear in mind that the defect in B-cell development in a Pax5-deficient mouse model may in itself drive the changes evident in the microbiome rather than Pax5^+/− being a preleukemic condition, similar changes seen in ETV6−RUNX1⁺ mice and in WT mice transplanted with hematopoietic stem cells from Pax5^+/− mice suggests that an altered microbiome may be present in several preleukemic states. This warrants further studies to investigate the presence of an altered microbiome, and the development of leukemia in other prelukemic mouse models.

The results of this study have exciting epidemiological and public health implications. Is the gut microbiome altered in all children with a preleukemic clone or only in those that will transform to leukemia? If changes in the gut microbiome can be detected before the development of overt leukemia, could it potentially be used as a biomarker to identify those children with preleukemic “first hits” at high risk of developing leukemia? The gut microbiome in babies and children is likely to be quite different from that of mice housed in controlled biomedical facilities.¹⁰  The relevance of this study therefore needs to be confirmed in a systematic manner in a human setting first and could involve, for example, prospective screening of newborns to first identify those at risk of leukemia, and then studying their gut microbiome (although such population-based screening comes with its own ethical implications). Perhaps in the future, the human microbiome could even be used to identify children predisposed to ALL and/or become a modifiable therapeutic target for preventing childhood leukemia.