Reduced Enterohepatic Recirculation of Mycophenolate and Lower Blood Concentrations Are Associated with the Stool Bacterial Microbiome after Hematopoietic Cell Transplantation

Mycophenolate mofetil (MMF) is an important immunosuppressant used after allogeneic hematopoietic cell transplantation (HCT). MMF has a narrow therapeutic index, and blood concentrations of mycophenolic acid (MPA), the active component of MMF, are highly variable. Low MPA concentrations are associated with the risk of graft-versus-host disease (GVHD), whereas high concentrations are associated with toxicity. Reasons for variability are not well known and may include the presence of β-glucuronidase-producing bacteria in the gastrointestinal tract, which enhance MPA enterohepatic recirculation (EHR) by transforming MPA metabolites formed in the liver back to MPA. This study was conducted to determine whether individuals with high MPA EHR have a greater abundance of β-glucuronidase-producing bacteria in their stool and higher MPA concentrations compared with those with low EHR. We conducted a pharmacomicrobiomics study in 20 adult HCT recipients receiving a myeloablative or reduced-intensity preparative regimen. Participants received MMF 1 g i.v. every 8 hours with tacrolimus. Intensive pharmacokinetic sampling of MMF was conducted before hospital discharge; total MPA, MPA glucuronide (MPAG), and acyl-glucuronide metabolite (acylMPAG) were measured. EHR was defined as the ratio of MPA area under the concentration-versus-time curve (AUC)_4-8 to MPA AUC_0-8. Differences in stool microbiome diversity and composition, determined by shotgun metagenomic sequencing, were compared above and below the median EHR (22%; range, 5% to 44%). The median EHR was 12% in the low EHR group and 29% in the high EHR group. MPA troughs, MPA AUC_4-8, and acyl-glucuronide metabolite (acylMPAG) AUC_4-8/AUC_0-8 ratio were greater in the high EHR group compared with the low EHR group (1.53 μg/mL versus .28 μg/mL [P = .0001], 7.33 hour·μg/mL versus 1.79 hour·μg/mL [P = .0003], and .33 hour·μg/mL versus .24 hour·μg/mL [P = .0007], respectively). MPA AUC_0-8 was greater in the high EHR group than in the low EHR group, and the difference trended toward significance (22.8 hour·μg/mL versus 15.3 hour·μg/mL; P = .06). Bacteroides vulgatus, Bacteroides stercoris, and Bacteroides thetaiotaomicron were 1.2- to 2.4-fold more abundant (P = .039, .024, and .046, respectively) in the high EHR group. MPA EHR was positively correlated with B. vulgatus (⍴ = .58; P ≤ .01) and B. thetaiotaomicron (⍴ = .46; P < .05) and negatively correlated with Blautia hydrogenotrophica (⍴ = -.53; P < .05). Therapeutic MPA troughs were achieved in 80% of patients in the high EHR group but in no patients in the low EHR group. There was a trend toward differences in MPA AUC_0-8 and MPA concentration at steady-state (μg/mL) between the high EHR group versus the low EHR group (P = .06). MPA EHR was variable. Patients with high MPA EHR had greater abundance of Bacteroides species in stool and higher MPA exposure compared with patients with low MPA EHR. Therefore, Bacteroides may be protective against poor outcomes, such as graft-versus-host disease, in some patients but may increase the risk of MPA adverse effects in others. These data need to be confirmed and studied after oral MMF therapy.