AUTHORS
Alexander J Silver, MD, PhD, Sarah Olmstead, BS2, Yara Almadani, MD3, Maria P Arrate, BS4, Caitlyn Vlasschaert, MD5, Yash Pershad, BS6, Taylor South7, Yixiang Gao8, Brian Sharber, MS9, Haley E Ramsey, PhD10, Yaomin Xu, PhD11, Alexander Bick, MD, PhD12, Michael R. Savona, MD13
AFFILIATIONS
1 Vanderbilt University Medical Center, Nashville, Tennessee, United States
2 Vanderbilt University Medical Center, Nashville, Tennessee, United States
3 Vanderbilt University Medical Center, Nashville, Tennessee, United States
4 Vanderbilt University Medical Center, Nashville, Tennessee, United States
5 Queen's University, Kingston, Ontario, Canada
6 Vanderbilt University Medical Center, Nashville, Tennessee, United States
7 Vanderbilt University Medical Center, Nashville, Tennessee, United States
8 Vanderbilt University Medical Center, Nashville, Tennessee, United States
9 Vanderbilt University Medical Center, Nashville, Tennessee, United States
10 Vanderbilt University Medical Center, Nashville, Tennessee, United States
11 University of Texas Southwestern Medical School, Dallas, Texas, United States
12 Vanderbilt University Medical Center, Nashville, Tennessee, United States
13 Vanderbilt University Medical Center, Nashville, Tennessee, United States
ABSTRACT
Background: Clonal hematopoiesis (CH) is an age-associated phenomenon that increases risk for malignancy, cardiovascular disease, and death. Exogenous exposures such as tobacco use, diet, sleep, exercise, chemotherapy, and radiation are known to influence the prevalence of CH. However, the extent to which cannabinoids, a commonly used recreational drug which can signal through cannabinoid receptors on hematopoietic precursors, affect the prevalence of CH is unknown. Therefore, we investigated whether cannabinoid use is associated with the prevalence of CH, and if cannabinoid exposure in the presence of CH could lead to competitive advantage.
Methods: We first examined the association of CH with cannabis use in human cohort studies. We identified CHIP mutations in the All of Us (AoU; v8) and UK Biobank (UKB) cohorts and used EHR and survey data to identify individuals with any prior cannabis use. We defined frequent cannabis use as at least weekly in the prior 3 months (AoU) or 100+ times over the lifetime (UKB). Excluding individuals with incomplete survey or EHR data or those with hematologic malignancy resulted in an analytic set of 85,146 (AoU) and 142,126 (UKB) individuals.
We asked how exogenous administration of cannabinoid affects stemness and proliferative capacity of mutant HSPCs, specifically examining Dnmt3a as the prototypical and most commonly mutated CHIP gene. Finally, we investigated whether cannabinoid exposure would provide a selective advantage to Dnmt3a-knockout HSPCs in an in vivo competitive transplant model. Competitive transplantations with 10% CD45.1 CRISPR-treated cells (Dnmt3a vs Rosa26) and 90% wild-type CD45.2 marrow were performed. Mice received either cannabinoid compound CP-55,940 (CP) or DMSO vehicle every other day from weeks 3-24, then were sacrificed. To determine whether transient or continuous exposure was required for selective advantage, secondary and tertiary transplants (repeated with 1:10 CD45.1:CD45.2 ratio for each) were performed in which half of the mice (N = 5 per group) received the same treatment as the primary transplant group, while half were switched to the other condition (e.g., CP in primary and DMSO in secondary and tertiary transplants).
Results: In both AoU and UKB, historical cannabis use was inversely related to age. In a meta-analysis of multivariable logistic regression models (adjusted for age, age-squared, sex, tobacco use, and genetic ancestry), frequent cannabis use was associated with increased risk for CHIP overall (meta-analysis OR = 1.25; 95% CI 1.09-1.43], p=0.0013).
In CRISPR-Cas9 knockout (KO) in vitro experiments, mouse HSPCs were treated with Dnmt3a- vs Rosa26-targeted sgRNAs followed by treatment with CP (100 nM) vs DMSO vehicle. Dnmt3a-KO cells exposed to CP cells exhibited dramatically increased in vitro stemness and proliferative capacity compared to controls as assessed by a serial replating colony forming unit assay (mean of 100+ colonies per 500 cells through 6 replatings).
In transplantation experiments, there was trend toward greater expansion of Dnmt3a-KO cells in all groups ever treated with CP vs not. After the second transplant, endpoint CD45.1 chimerism was highest in the Dnmt3a group treated continuously with CP and lowest in the Rosa26 group never treated with CP (13.0% vs 0.7%; p = 0.041). There was statistically significant increase in chimerism in the Dnmt3a group with late transient CP treatment in comparison to the matched DMSO treated group (6.2% vs 0.5%; p = 0.021). At the tertiary takedown, CD45.1 chimerism was higher in mice transplanted with Dnmt3a-KO cells treated continuously with CP than mutant transplant controls never treated with CP (13.5% vs 2.0%; p = 0.016), and early transient CP treatment of Dnmt3a-KO cells compared with the CP treated control group (29.5% vs 1.0%; p = 0.011).
Conclusion: Cannabis is a new potential risk factor for CHIP. In human cohorts, frequent cannabis use is associated with increased prevalence of CHIP. In experimental models, smoke-free administration of exogenous cannabinoid led to increased long-term competitive advantage of Dnmt3a-deficient hematopoietic cells. These findings raise the possibility that recent increases in the rates of cannabinoid use may translate to a higher prevalence of CH, and that continued use of cannabis in setting of CH may lead to increased selective advantage of mutant cells.
DISCLOSURES
Alexander J Silver, MD, PhD: No relevant conflicts of interest to declare.