or cryo‐electron microscopy structures of human PMAT, but structures exist for human equilibrative nucleoside transporter 1 (hENT1).
2.6.2 Tissue Distribution in Human
Though highly expressed in various tissues such as adipose and breast tissue, THTR‐2 is the major absorptive transporter for thiamine in the intestine [3]. It is located on the apical membrane of intestinal epithelia and renal proximal tubule cells, where it plays a role in renal reabsorption of thiamine. In contrast, while exhibiting low ubiquitous expression across a variety of tissues, THTR‐1 is perhaps most physiologically relevant in the bone marrow, ear, and pancreas/liver, as suggested by the manifestation of megaloblastic anemia, diabetes, and sensorineural deafness in patients with loss‐of‐function THTR‐1. Both THTR‐1 and THTR‐2 are expressed at the blood–brain barrier. In accordance with its role as a neurotransmitter transporter, PMAT is most highly expressed in the brain, and notably in the hypothalamus, cortex, substantia nigra, cerebellum, amygdala, and hippocampus regions, as well as in the heart, kidney, skeletal muscle, and liver [4] (Table 2.1).
2.6.3 Substrate and Inhibitor Selectivity
Thiamine, which consists of an aminopyrimidine and a thiazolium ring linked by a methylene bridge, is positively charged at pH 7.4. THTR‐1 transports thiamine with a K m value of approximately 7 μM, whereas THTR‐2 has a K m of approximately 4 μM. Both transporters are considered high‐affinity, low‐capacity thiamine transporters, whereas other human transporters that take up thiamine (i.e., OCT1, OCT2, MATE1) are of low affinity, high capacity [18]. Thiamine transport by both THTR‐1 and THTR‐2 is pH sensitive, optimal at neutral pH, and decreases as extracellular pH becomes acidic [55]. In November 2013, a clinical trial of fedratinib, a JAK2 kinase inhibitor, was halted due to the observation of Wernicke’s encephalopathy, a severe neurological disease induced by thiamine deficiency, in several subjects [56]. Subsequent studies revealed fedratinib as a potent inhibitor of THTR‐2, with the hypothesis that the aminopyrimidine structure present in fedratinib is responsible for the THTR‐2 interaction [57]. Fedratinib is also a substrate of THTR‐2 with a K m ~ 0.44 μM, but is not a substrate of THTR‐1. Another JAK2 kinase inhibitor, AZD1480, inhibits both THTR‐1 and THTR‐2, with IC50 values of 22 μM and 15 μM, respectively. A pharmacophore model created to identify features associated with potent inhibition of THTR‐1 and THTR‐2 validated the importance of the 2,4 diaminopyrimidine structure for substrate and inhibitor [56].
Additional prescription drugs have been identified as substrates of THTR‐1 and THTR‐2 using in vitro assays and computational models. For example, metformin has been identified as a substrate and inhibitor of THTR‐2 [55]. Additionally, MPP+, a neurotoxin associated with Parkinson’s disease, and famotidine, an over‐the‐counter histamine receptor antagonist used to treat acid reflux disease, have been identified as substrates of THTR‐2. Verapamil and chloroquine are inhibitors of THTR‐2‐mediated uptake of metformin at 1 mM. Substrate and inhibitor specificity differences exist between species, as exemplified by metformin, which is a substrate for human, but not mouse THTR‐2, and fedratinib, which demonstrates weaker inhibition of mTHTR‐2 compared with hTHTR‐2. Trimethoprim, an antibiotic used for a wide variety of infections, is not only an inhibitor but also a substrate of both THTR‐1 and THTR‐2, with K m values of 22 μM and 1 μM, respectively.
A high‐throughput THTR‐2 drug inhibitor screen evaluating over 1,300 compounds, including many prescription and over‐the‐counter medications, identified several prescription drugs as novel inhibitors of THTR‐2 at low‐ or mid‐micromolar concentrations including hydroxychloroquine sulfate (IC50 ~ 17 μM), sertraline hydrochloride (IC50 ~ 1 μM), amitriptyline hydrochloride (IC50 ~ 11 μM), amoxapine (IC50 ~ 46 μM), quinapril (IC50 ~ 34 μM). Based on the current FDA guidelines, all of these drugs are predicted to reach the threshold required for a clinical drug–drug interaction study [58].
The increasing number of approved drugs identified as potent THTR‐1 and/or THTR‐2 inhibitors initiates concerns regarding the potential for these common medications to cause clinical adverse events related to thiamine deficiency. Many of these drugs are predicted to reach concentrations in the intestine sufficiently high to significantly inhibit absorption of thiamine by the primary intestinal thiamine transporter, THTR‐2 [58]. The overwhelming majority of these drugs are taken chronically, and thus prolonged inhibition of thiamine absorption is possible. Individuals already at risk for thiamine deficiency (i.e., individuals with alcoholism, malnourished individuals, and the elderly) who take such THTR‐2 inhibitor drugs may be at risk for thiamine deficiency. This risk may be increased for individuals on multiple medications, several of which may inhibit the transporter.
PMAT substrates have been identified using an in vitro screen of a variety of radiolabeled compounds. As noted, despite its sequence similarity to the ENT transporters, PMAT does not transport most nucleosides or nucleoside analogs; instead, PMAT transports neurotransmitters such as serotonin, dopamine, epinephrine, and norepinephrine, as well as the prototypical OCT substrate MPP+. The K m and V max values for these substrates suggest that PMAT is a low‐affinity, high‐capacity neurotransmitter transporter. PMAT substrate uptake is pH‐dependent (i.e., decreased pH, increased transport). Metformin and atenolol, a prescription medication used for hypertension, are substrates of PMAT, with greatest uptake at pH ~ 6 [59, 60]). There is significant overlap in exogenous drugs that are substrates and inhibitors of PMAT and OCT transporters (e.g., HIV protease inhibitors). Lopinavir is a PMAT‐selective inhibitor that does not inhibit OCT transporters and can be used to differentiate PMAT‐mediated monoamine and organic cation transport from OCT1‐3 mediated transport [61].
2.6.4 Animal Models
Slc19a2 and Slc19a3 knockout mouse models have been generated to evaluate intestinal absorption of thiamine in vitro and in vivo [62]. Slc19a3 −/− mice are viable and fertile, but a noticeable change in gross phenotype occurs around 1 year of age. In particular, the mice exhibit progressive wasting 2–3 months prior to a premature death at 1 year of age. Chronic inflammation in hepatic parenchyma and renal cortex is observed in the mice. Notably, thiamine plasma levels in Slc19a3 −/− mice are significantly lower than sex‐matched wild‐type mice, reflecting reduced thiamine absorption. In contrast, similar experiments in Slc19a2 −/− mice revealed no difference in plasma thiamine levels compared to wild‐type littermates [63].
Slc19a2 knockout mouse models serve as experimental models of TRMA. Homozygous Slc19a2 −/− mice are physiologically normal when fed standard chow, but develop sensorineural deafness, megaloblastosis, and diabetes mellitus with reduced insulin secretion and an enhanced response to insulin when fed a thiamine‐deficient diet. Erythrocytes from Slc19a2 −/− mice lack the high‐affinity component of thiamine transport. Furthermore, bone marrow samples from thiamine‐deficient Slc19a2 −/− mice are abnormal, with megaloblastosis affecting the erythroid, myeloid, and megakaryocyte lines. All aspects of the TRMA phenotype in Slc19a2 −/− mice are reversible with doses of thiamine [63].
Pmat −/− mice are viable and fertile with no physiologic abnormalities. Further, no significant differences in standard blood chemical measurements are observed between WT and KO mice. Whole brain expression of functionally related transporters (Sert, Dat, Net, Oct3) is not changed by knockout of Pmat, and choroid plexus size and morphology is similar between Pmat +/+ and Pmat −/− mice [64]. Some studies have observed subtle, sex‐specific phenotypic changes in behavior in Pmat
