Now that I have reached the age at which the likelihood of my having clonal hematopoiesis (CH) exceeds 10%, I have become a bit more interested in the phenomenon. Thus, I was fascinated to read the novel, unexpected, and potentially important observation made by Imus et al in this issue of Blood Advances that preallogeneic transplant presence of CH in the recipient (yes, recipient, not donor) leads to significantly higher posttransplant nonrelapse mortality (NRM) and worse overall survival (OS) without affecting relapse rates.1
Focusing on 97 patients with lymphoid malignancies over the age of 60 years who underwent allogeneic transplantation, they detected pretransplant CH in 62% of the recipients. They restricted the study to lymphoid malignancies to avoid confusion with myeloid measurable residual disease. Compared with those without CH, individuals with CH were threefold more likely to die of nonrelapse causes (35% vs 11%) and had a 3-year OS of 47% compared with 78% for those free of CH. The authors hypothesize that the negative effect of recipient CH is the result of persistent CH-derived macrophages secreting proinflammatory cytokines, resulting in an exuberant inflammatory response and thus an increase in cytokine release syndrome (CRS) and graft-versus-host disease (GVHD). These observations raise the possibility that efforts to eliminate CH before transplantation or to block the proinflammatory effects of CH-derived cells might be of benefit.
These results raise several questions. First, is the association between recipient CH and worse outcomes real, and is it generalizable? The number of patients in each cohort was small (67 with CH and 30 without CH), and the cohorts were not perfectly balanced. A multivariable analysis was conducted and found that the strongest predictor of NRM was recipient CH, thus supporting the author’s argument. However, the number of patients was limited, as were the number of factors included in the multivariable analysis. Additionally, most patients were recipients of haploidentical transplants. The association of recipient CH with worse outcomes needs to be confirmed in a larger cohort so that obvious variables such as patient age, disease status, and prior therapy and less obvious variables (including smoking and obesity, both of which increase the incidence of CH) can be more confidently accounted for, and to determine whether the association persists in other transplant settings.
A larger confirmatory study would have 2 other important benefits. First, many different types of CH exist; the Hopkins study alone included 20 different mutational varieties. It seems unlikely that the proinflammatory profile of macrophages derived from DNMT3A, TP53, and TET2 mutations would all be identical. Furthermore, variant allele fractions vary. Thus, an analysis of a larger cohort, beyond confirming the overall association of recipient CH with worse outcome, would have the benefit of identifying those forms and levels of CH associated with greater or lesser risk, thus guiding clinical decision-making and providing a glimpse into the inflammatory pathways involved.
A second benefit would be to gain more insight into the clinical consequences of recipient CH. Although NRM was greater and OS was worse, there was no easily identifiable single clinical reason for these results. There were more deaths from acute GVHD and a trend toward an increased incidence of CRS, sepsis, and multiorgan failure. No analysis of the inflammatory markers or subsequent immune recovery was provided. A study of a larger cohort would allow for a better definition of the laboratory and clinical impact of recipient CH, which would be invaluable when considering interventions.
Assuming that the correlation between recipient CH and worse outcomes is confirmed, then a second question is one of cause and effect: is recipient CH the cause of the increased NRM and poorer OS, or are both CH and worse outcomes the result of something else, that is, is CH simply a surrogate for a more fundamental abnormality?
The hypothesis that CH is a driver is attractive. Innumerable recent studies have linked the presence of CH to an increased incidence and severity of many inflammatory conditions, including cardiovascular disease, chronic obstructive lung disease, type 2 diabetes, and osteoporosis.2 Although these associations do not prove that CH is the cause, elegant mouse models have found that, for example, disabling Tet-2 specifically in the bone marrow leads to accelerated atherosclerosis in mice challenged with high-fat diets or pulmonary disease after exposure to cigarette smoke, demonstrating that in these models, the altered marrow population is helping to drive the disorder.3
An alternative explanation is that recipient CH is not the driver of increased toxicity; instead, both CH and increased toxicities are the result of a deeper problem. Although the analogy is not perfect, consider Fanconi anemia (FA). Germ line mutations in FA increase the risk of both the development of leukemia and toxicities seen after hematopoietic cell transplantation (HCT). No one argues that the malignant population in FA is causing the increased toxicities. At least 2 dozen germ line risk loci associated with the development of CH have been identified, many of which, like FA, are involved in DNA damage response or telomere maintenance.4 It seems at least plausible that CH and increased post-HCT toxicity have a shared germ line predisposition rather than one causing the other.
In addition to germ line predisposition, 2 other major factors that influence the development of CH are inflammation and aging. Both animal models and human studies have demonstrated that CH development and progression are associated with a proinflammatory state. However, whether an existing proinflammatory state, be it germ line, microbiome, environmental, or infectious in origin, augments the establishment and progression of CH or whether CH creates its own proinflammatory environment remains an open question. Animal models exist supporting either hypothesis.5 Both may be true, with inflammation promoting the outgrowth of CH, and CH further contributing to the inflammation.
The single factor by far most associated with CH is aging.6 Overall shortened life span and many diseases associated with aging, including cancer and dementia, have been linked to accelerated epigenetic aging, a phenomenon in which a person’s DNA methylation profile predicts an older phenotypic age than the calendar.7 A recent study involving over 5000 individuals found that the development of CH was strongly associated with epigenetic age acceleration. Furthermore, the increased all-cause mortality and cardiovascular disease associated with CH were restricted to the subset of individuals who had both CH and accelerated epigenetic age; CH by itself had no effect.8 These results suggest that CH, rather than being causative, may be a marker of biologic aging, which could explain the increased posttransplantation toxicities seen in recipients with CH.
The distinction of whether CH is the driver or passenger is important because, if the latter is true, then efforts to decrease post-HCT toxicities by specifically eliminating CH pretransplant would likely be futile. In contrast, efforts to block proinflammatory cytokines during the peritransplant period might be beneficial no matter whether the inflammatory state comes from CH cells, or the existence of CH is simply a surrogate for the presence of such a state. Preliminary studies using JAK inhibitors in a peritransplant setting are encouraging.9 Whether or not the soiled soil hypothesis is correct, it provides fertile ground for a large crop for future studies with the potential to improve the outcome of allogeneic HCT.
Conflict-of-interest disclosure: F.R.A. declares no competing financial interests.
