In this study, we aimed to provide a theoretical background on how administered SP reaches the brain and increases BH4 levels in the brain parenchyma. We focused on the permeability of SP across barriers that define the major fluid compartments, built a kinetic model, and visualized it in a block diagram (Fig. 2). In this model, the increase in BH4 in brain cells was expressed by a linear equation using SP dose AUCplasma as the independent variable. In the in vivo experiment, SP was administered (ip) to mice, and if the dose exceeded a certain threshold, the increase in brain BH4 was proportional to the dose, consistent with the model's prediction. We observed that peripherally administered SP disappears within the first 30 min but nevertheless reaches the brain and effectively increases BH4 levels in the brain parenchyma.
Inherited deficiency of BH4 causes various disorders, mainly through hyperphenylalaninemia and insufficient production of CNS monoamines. Systematic protocols for the diagnosis and treatment of BH4 deficiency have been standardized [40, 41]. Peripheral administration of 6RBH4 is an established therapy to alleviate hepatic BH4 deficiency. Nevertheless, the lack of CNS monoamine production remains an unresolved problem due to the limited permeation of 6RBH4 across the BBB. Previous papers have shown that peripherally administered 6RBH4 increases brain BH4, but the promotion of monoamine synthesis was limited [8–11],as reviewed [41, 42]. We are concerned that they considered the measured increase in BH4 to be in the "whole brain." The present study showed that peripherally administered 6RBH4 enters the "whole brain," but since the uptake clearance of BH4 is 1/20 of SP, it hardly enters brain cells. The message from the previous studies using 6RBH4 should be revised: exogenous BH4 does not enter cells and therefore does not promote monoamine synthesis. Incidentally, we considered that the cell membrane of brain parenchymal cells is a critical barrier (Barrier 2) for solute uptake (BH4 or SP). We then derived an equation based on our model such that the answer is the amount that crosses this barrier and enters the cell (Eq. 7). Our experimental results (Fig. 6) showed that the in vivo events are consistent with the predictions obtained from this equation. Since brain cell BH4 can be increased by SP administration rather than 6RBH4, we have regained the BH4 replenishment as a realistic strategy of promoting monoamine production for the treatment of many acquired and genetic deficiencies of the brain monoamine neurotransmitters.
SP permeability across endothelial cell monolayer Barrier 1, CLin,SP
We calculated the inward component CLin by measuring the apical (AP) to the basolateral (BL) permeation constant across the RBEC monolayer. Given that the measured flow was in the direction of solute-free BL chamber and the permeation continued linearly, we concur with the argument made by Chen et al. [28] that the AP→BL flow measured in this apparatus does not involve a pushback counter flow. Accordingly, the estimated clearance CLin,SP thus 17.0 µL/(min·g-brain), which is close to that reported for atenolol (CLin,alenolol = 19.3 µL/(min·g-brain)) [28]. They further argued that this inward clearance of atenolol (CLin,alenolol) was not stringent enough to block the entry of atenolol into the brain but that the net flow into the brain was limited by a pushback of CLout,alenolol of 425 µL/(min·g-brain). Therefore, it was necessary to clarify the pump back of SP to assess its BBB crossing in vivo.
Pump out clearance through the endothelial layer Barrier 1, CL(out,SP)
Candidates for efflux transporters involved in the BBB include various ABC transporters and SLC transporters [23, 43], which also overlap with uptake transporters [22]. Biological pumping at the BBB is carried out through transcellular transport, involving the coordinated action of multiple transporters. Probenecid-sensitive transporters and cyclosporin A-sensitive transporters act primarily in tandem fashion in the excretion of 6RBH4 in the kidney [18, 21, 30]. Among the diverse probenecid-sensitive transporters [24], OAT1 and OAT3 enable membrane permeation of SP, BH2, and BH4 [19]. Transport by these SLC transporters is determined by the electrochemical environment and its relationship with the counter-solute. We did not measure the back flow across the RBEC monolayer, as the validity of this assembly as a model for in vivo outflow events remains unclear, primarily owing to the challenges in replicating the counter-solute conditions in vitro. To calculate CLout, we used the data from the study by Kapatos et al. [34], even though their study may not have specifically aimed to quantify this efflux. They measured the conversion of SP to BH4 via BH2 after the direct SP injection into the lateral ventricle. We assumed that the injected SP filled the ISF, and that the remaining SP dose after subtracting the amount converted to BH4 would represent the amount lost through outward flow. This method of loading exogenous solute by direct injection to the ventricle was well established as an experimental procedure. However, this method is not always recommended as a method of treatment [44]. A keen study of their data showed that the increase in BH2 + BH4 peaked at approximately 1 h and then declined almost exponentially. This is likely attributed to the non-negligible amount of BH2 + BH4 excreted into the systemic circulation, although the excretion rate is not as rapidly as that of SP. Therefore, the actual amount of SP converted to BH2 + BH4 was greater than that of BH2 + BH4 measured at 1 h, and our calculations overestimated the efflux. While we are unable to perform numerical corrections for this, it is reasonable to limit our estimates to Xout < 353 µL/(min·g-brain). We also considered information on the efflux of SP injected into the ventricles from the report by Bowers et al. [33]. Since the initial velocity of the exponential process is faster than the velocity measured after 40 s, we limited our estimates to CLout > 220 µL/(min·g-brain). In summary, the actual CLout for rats was estimated to range from 220 to 353 µL/(min·g-brain), while considering the possibility of a considerable large margin of errors owing to the technical difficulties of this experiment.
Uptake clearance of brain parenchymal cells Barrier 2, CLuptake,SP
We estimated CLuptake,SP and CLuptake,6RBH4 in CTX TNA2 cells. However, we used these cells without any evidence that they represent all cells, including monoaminergic neurons. The measured uptake rates were consistently similar and not particularly unique to any cells we used, including RBL2H3, HeLa, Caco-2, and LLC-PK1, and even a primary culture of hepatocytes. The uptake is gated by equilibrative transporters ENT1 and more probably by ENT2 of which specificity for BH4, BH2, and SP is approximately 1:1.92:15.6, respectively [17]. The similarity may be attributed to the fact that ENT2 is commonly distributed in a wide range of cells requiring uptake of nucleic acids or nucleobases [20]. Therefore, estimation of CLuptake,SP and CLuptake,6RBH4 using CTX TNA2 cells is appropriate as a representative example. The mechanism by which extracellular BH4 is assimilated into intracellular BH4 has been described previously [7]: 6RBH4 is unstable under extracellular conditions due to its low redox stability. When exposed to cells, the resulting BH2 is preferentially taken up because it is more permeable than BH4 and is then assimilated back to BH4 through the salvage pathway (see Fig. 5). On this basis, the BH4 salvage pathway precursor SP or BH2 is a better source than 6RBH4 to push BH4 into cells. In this study, we measured the uptake clearance of SP, which has the highest cell membrane permeability among the above three, and compared it with 6RBH4, which has the lowest permeability.
Given that the SP conversion to BH2 + BH4 never occurs extracellularly, an increase in cellular BH2 + BH4 is stoichiometrically equivalent to the amount of SP taken up by the cells [32]. Accordingly, we estimated the amount of SP that was taken up by brain cells in vivo as Xuptake,SP = X(SP→BP), instead of physically tracking peripherally administered SP. This method would be applicable to a wide range of compounds, such as vitamins and nucleic acid precursors; once taken up by cells, these compounds are assimilated into chemical group carriers or coenzymes via dedicated enzymatic pathways. SP is a compound that typically falls into this category. This property of rapid uptake and elimination of SP makes mass transfer within ISF faster than normal diffusion. This is because SP molecules near cells move only in the forward direction, resembling a "sink," rather than a "back and forth" motion of diffusion [45]. At the same time, this limits the opportunities for SP to flow out of the ISF. SP leaves BH2 + BH4 in brain cells, whereas atenolol, for example, is not metabolized and is eventually washed away [28], despite SP and atenolol having similar CLin and CLout.
Brain-BH4 replenishment by peripheral SP administration, the composite clearance, CL(plasma→cell)
The results in Fig. 6 showed that the model expressed by Eq. 7 is applicable to SP doses above the threshold (13.3 mg/kg-body, ip for mice). At these doses, the peripherally administered SP reaches the brain, passes through BBB (Barrier 1) and parenchymal cell membrane (Barrier 2), and then gets converted to BH4. The estimated net clearance from plasma into the cell interior, CL(plasma→cells),SP = 5.42 ~ 10.2 µL/(min·g-brain) (Eq. 8), was applicable over a realistic dose range of 20 to 125 mg/kg. This experiment showed that suprathreshold doses are critical for a successful increase in brain parenchymal BH4 levels by SP administration. Thresholds vary depending on the target species (human, rat, or mouse) and route of administration (intraperitoneal, oral, or intravenous). We now understand why previous SP administration attempts (data not shown) were not always successful in enhancing monoamine synthesis despite consistently increasing whole brain BH4 content.
On the other hand, according to Eq. 7, Fig. 6 shows that the AUCplasma was zero at the SP doses below the threshold, suggesting the disappearance of SP by conversion to BH2 + BH4 before reaching the brain. BH2 + BH4 is produced at any dose of SP and delivered systemically over several hours (Fig. 3). For 6RBH4, the inflow clearance across Barrier 1 as BBB (CLin,6RBH4) was similar to SP, although the Barrier 2 clearance of cell membrane (CLuptake,6RBH4) was approximately 1/20 that of SP (Table 1). In a recent study [10], a 2.5-fold increase in brain BH4 was observed 3 h after 6RBH4 administration (50 mg/kg, rats, ip). The authors stated that peripherally administered 6RBH4 crossed the BBB with a transport coefficient of 0.083 µL/(s·g) (measured by L-[6,7-3H]-BH4 perfusion) which corresponds to a CL = 4.98 µL/(min·g-brain). However, they did not mention whether the BH4 that entered the brain was intracellular or extracellular. Instead, potentiation of dopamine release by the administered 6RBH4 was observed as a CNS event. However, the dopamine release effect is not evidence of BH4 uptake into cells, as this dopamine release is known to be caused exclusively by extracellular BH4 [46]. Their estimate of BH4 transport clearance does not consider the difficulty of BH4 uptake into brain parenchymal cells. BH4 that enters ISF will remain in that fluid as long as it is continually supplied but will be flushed out when the supply is halted. Indeed, in the case of SP administration, the supply of BH2 + BH4 lasted for several hours (Fig. 3).
Retention of elevated BH4 in the brain
The peak level of intracellular BH4 immediately after SP administration, e.g., at 30 min, should have been much higher than the amount measured after 6 h. Assuming an exponential decline in intracellular BH4 levels, the half-life (T1/2) of BH4 is approximately 150 min (as reported in primary cultures of superior cervical ganglion neurons [47]) or 120 min (as evidenced in RBL2H3 and CTX TNA2 cells in the current study, data not shown). It decreases exponentially to the initial value multiplied by (1/2)n during time t, where t = n·T1/2. For example, at the SP dose of 125 mg/kg (Fig. 6), the initial BH4 level calculated 6-h back with T1/2 = 2 h is approximately 30 times the baseline (= 3.71 × 23), which is not very realistic.
The amount of intracellular BH4, once increased by the influx of SP, decreases at a rate slower than the normal first-order reaction rate depending on the conditions described below. Given that intracellular BH2 + BH4 crosses the cell membrane by the bidirectional passive transporter ENT2; first, in passive transport, solute passage is generally slow when the concentration difference across the membrane is small. Indeed, with peripheral SP administration, SP-derived BH2 + BH4 with its long tail was delivered to the brain over several hours (Fig. 3), resulting in a high concentration of BH2 + BH4 in ISF and a small intracellular-extracellular difference. This is an environment that slows down the efflux of intracellular BH2 + BH4. Second, a low BH2/BH4 ratio slows down the decline of BH2 + BH4, as the permeability of BH4 by ENT2 is approximately half of that of BH2. BH2 and BH4 are in a dynamic equilibrium determined by the balance between intracellular BH4 oxidation and BH4 production via the salvage pathway as well as the de novo pathway. Due to this equilibrium, BH2 leakage preferentially drives the decrease of BH2 + BH4. However, on the contrary, actual brain BH2/BH4 decreased significantly (Fig. 3, d), which must have slowed down the decrease. These in vivo-specific conditions synergistically slowed down the actual leakage compared with the typical exponential decline. Note that the intracellular BH2/BH4 ratio is approximately 0.05 to 0.10 under in vitro culture conditions. Incidentally, the significant decrease in the BH2/BH4 ratio upon SP administration in vivo is consistent with previous observations [13]. Taken together, the intracellular BH4 increased once and then decreased slowly and persisted for more than 6 h. In general, intracellular coenzyme small molecules are chemically much more stable as long as localized in the cell than the holoenzyme macromolecule. We recall that BH4 is retained longer than TPH1 in RBL-2H3 cells, where the T1/2 of TPH1 is around 20 min [48, 49].