Thiopurine Use During Pregnancy Has Deleterious Effects on Offspring in Nudt15 R138C Knock-In Mice.

概要

Thiopurines are key immuno- suppressants for the treatment of inflammatory bowel disease. Thio- purine use during pregnancy has not been prohibited, but its safety is still debated.1,2 Additionally, it has been revealed that several genetic poly- morphisms are associated with thio- purine toxicities3; however, the effects of parental or offspring genotype on the safety of thiopurine use during pregnancy have not been investigated. NUDT15 (nucleoside diphosphate- linked moiety X-type motif 15) is responsible for the inactivation of thiopurines by converting thio- guanosine-5’-triphosphate to thio- guanosine-5’-monophosphate, and single-nucleotide polymorphisms of NUDT15 are strongly associated with cytopenia during thiopurine use.4,5 In particular, the c.415C>T single- nucleotide polymorphism, which in- duces p.Arg139Cys (R139C) and the loss of normal enzymatic activity, is clinically important in East Asians because it penetrates more than 10% of them and frequently causes severe cytopenia.4 We recently established knock-in mice harboring a p.Arg138Cys mutation (Nudt15R138C), which corre- sponds to human NUDT15 R139C, and demonstrated that thiopurine administration causes hematopoietic stem cell (HSC) toxicity in Nudt15+/ R138C or Nudt15R138C/R138C mice (see Supplementary Methods).6 In this study using our mouse model, we investigated whether thiopurine use during preg-nancy differentially affects offspring, based on their NUDT15 genotype.

Our previous report demonstrated that the long-term (>2 months) sur- vivable dose of mercaptopurine (MP) is 1.0 mg/kg for Nudt15+/+, 0.5 mg/kg for Nudt15+/R138C, and 0.2 mg/kg for Nudt15R138C/R138C adult mice.6 Thus, we administered the same doses of MP to Nudt15+/R138C or Nudt15+/+ preg- nant mice, respectively, and then characterized the Nudt15 genotypes of the neonatal mice. Nudt15+/R138C fe- male mice that were mated with Nudt15+/R138C male mice generated neonatal mice in a Mendelian fashion under 0 mg/kg and 0.1 mg/kg MP treatment (see Supplementary Methods). However, few to zero Nudt15R138C/R138C neonatal mice were generated under 0.2 mg/kg or 0.5 mg/ kg MP treatment, respectively (Figure 1A). Similarly, Nudt15+/+ fe- male mice that were mated with Nudt15+/R138C male mice failed to generate Nudt15+/R138C neonatal mice under 1.0 mg/kg MP treatment (Figure 1B). These data indicate that the therapeutic MP dose for pregnant mice could be deleterious to offspring harboring more Nudt15R138C allele than the female parent.

Next, to investigate fetal abnormal- ities caused by MP treatment during pregnancy, we analyzed the Nudt15 ge- notypes of fetal mice from Nudt15+/R138C pregnant mice that were administered 0.2 mg/kg or 0.5 mg/kg MP after mating with Nudt15+/R138C male mice. On em- bryonic day 14.5, the number of Nudt15R138C/R138C fetal mice was reduced under 0.2 mg/kg MP treatment and eliminated under 0.5 mg/kg MP treatment, indicating that thiopurine use during pregnancy can lead to embryonic lethality in Nudt15R138C/R138C offspring (Supplementary Figure 1). Nudt15R138C/R138C fetal mice that sur- vived 0.2 mg/kg MP treatment during pregnancy tended to be paler (Figure 2A) and were significantly smaller in size than Nudt15+/+ fetal mice (Figure 2B). We previously reported that hematopoietic tissue is promptly injured by MP treatment in Nudt15R138C/R138C adult mice, and HSCs are damaged by MP in a Nudt15R138C allele number- dependent manner.6 Therefore, we determined the number of HSCs in each fetal liver, the center of hematopoiesis in fetal mice. To do this, fetal HSCs that are phenotypically enriched in the CD48–CD150highLineage–c-Kit+Sca-1+ population using multicolor staining7 were counted on embryonic day 14.5 (see Supplementary Methods and Supplementary Figure 2). The number of fetal HSCs was not altered in any of the Nudt15 genotypes without MP treatment (Figure 2C). However, it was significantly reduced by MP treatment in Nudt15+/R138C fetal mice (P < .01; mean number 1104 in dimethyl sulfoxide and
577 in MP) and in Nudt15R138C/R138C fetal mice (P < .01; mean number 927 in dimethyl sulfoxide and 402 in MP), but not in Nudt15+/+ fetal mice (P = .39; mean number 1004 in dimethyl sulf-oxide and 820 in MP). Finally, the num- ber of HSCs in Nudt15R138C/R138C fetal livers was significantly reduced to less than 50% of that in Nudt15+/+ fetal livers by MP treatment during gestation (P = .0089).

The current study clearly demon- strates that thiopurine use during pregnancy can cause serious damage to fetal mice, depending on the Nudt15 genotype of the offspring. In particular, Nudt15+/R138C offspring in Nudt15+/+ pregnant mice and Nudt15R138C/R138C offspring in Nudt15+/R138C pregnant mice are not safe when the pregnant mice are exposed to therapeutic MP dose. Because the placental perme- ability of thiopurines and metabolites is unknown in mice, our data may possibly overestimate thiopurine fetal toxicity in humans. However, it has been reported that thioguanines including thio-gua- nosine-5’-triphosphate, which is the active thiopurine metabolite for cyto- toxicity and is directly metabolized by NUDT15, can cross human placenta.8 In addition, cases of anemia in thiopurine- exposed infants have been reported, although there is no description of NUDT15.9 Our data also show that fetal HSCs can be damaged by thiopurine use during pregnancy in a Nudt15 allele number-dependent manner.

In summary, given the ethical diffi-culty of conducting further prospective clinical studies, thiopurine use during pregnancy should be considered with caution based on the NUDT15 genotype. Particularly, the paternal NUDT15 R139C alle is recommended to be examined in the decision to use thiopurine during pregnancy because it is critical to deter- mine the Nudt15 genotype of offspring.

この論文で使われている画像