Introduction

The solute carrier (SLC)15 family of proton-coupled oligopeptide transporters (POTs) has four members, PepT1 (SLC15A1), PepT2 (SLC15A2), PhT1 (SLC15A4) and PhT2 (SLC15A3) that transport di- and tripeptides and small peptidomimetics [1]. A fifth member, SLC15A5 has been identified but its function is uncertain [2, 3]. PepT1 was originally cloned from intestine [4] where it is involved in the uptake of oligopeptides produced by digestion. PepT2 was cloned from kidney [5] where it is involved in reabsorption of peptides. Compared to PepT1, PepT2 is a high-affinity, low-capacity transporter [1]. In contrast to PepT1 and PepT2, PhT1 and PhT2 can transport L-histidine as well as oligopeptides and hence were named peptide/histidine transporters (PhTs). While all endogenous di- and tripeptides and some endogenous peptidomimetics (e.g., 5-aminolevulinic acid; 5-ALA) are POT substrates, so are some therapeutics including some beta-lactam antibiotics (e.g., cefadroxil) and antiviral prodrugs (e.g., valacyclovir) [1].

PepT2 and PhT1 are expressed in brain parenchyma [6, 7] whereas there is little or no expression of PhT2 and PepT1 [8, 9]. PepT2 is highly expressed at the choroid plexus and the arachnoid membrane, tissues forming the blood-cerebrospinal fluid (CSF) barrier [10,11,12], but no identified POT is present in the brain endothelial cells that form the blood-brain barrier [13]. Given the importance of neuropeptides in brain function and the need for developing CNS-acting therapeutics, David Smith and his collaborators began investigating the role brain of POTs around 2000. At the time of David Smith’s retirement, the purpose of this review is to highlight some of that pioneering work and then to discuss some recent studies advancing the field and opening new areas for research.

Contribution of David E. Smith to our Understanding of Oligopeptide Transport in the Brain

Pept2-Mediated Transport

David Smith has a longstanding interest in kidney transport [14,15,16] and it is interesting that his first work in the brain was on the choroid plexus (CP), a tissue that has been called the “kidney” for the brain [17]. The CPs are in the lateral, third and fourth ventricles of the brain. They are a primary site of CSF secretion and form the blood-CSF barrier. Thus, CP epithelial cells are linked by apical tight junctions and they possess an array of transporters that produce CSF and regulate its composition [18]. Using freshly isolated CP [19, 20], CP epithelial primary cultures [21] and immunohistochemistry [10], David Smith and colleagues showed that the rat and mouse CP expresses PepT2 on the apical membrane, clearing oligopeptides (GlySar, carnosine, glycyl-L-glutamine), endogenous peptidomimetics (5-ALA), and peptide mimetic drugs (cefadroxil) from the CSF into the CP epithelium [19,20,21,22,23,24,25,26,27,28] (Fig. 1). Such uptake was inhibited by other dipeptides [20,21,22, 24] and by genetic deletion of PepT2 [23,24,25,26], as well as being pH dependent [21, 24, 25].

Fig. 1
figure 1

Oligopeptide and peptidomimetic transport at the choroid plexus epithelium. PepT2 (SLC15A2) on the apical membrane of the choroid plexus epithelial cell transports a wide range of di- and tripeptides, peptidomimetics and drugs from the CSF into the cell. That transport is pH dependent. The fate of these compounds after entering the cell is uncertain. They may be metabolized or cleared across the basolateral membrane that faces stroma and fenestrated capillaries. The paracellular route between choroid plexus epithelial cells is hindered by the presence of apical tight junctions (TJ)

The development of the PepT2 knockout (KO) mouse [26] was a major development in understanding the functions of PepT2. They facilitated examination of the role of PepT2 in in vitro CP preparations [23, 24, 26, 28] but particularly allowed the role of PepT2 to be examined in vivo. Thus, deletion of PepT2 in mice caused very marked increases in the CSF to blood concentration ratio for GlySar [29], cefadroxil [30] and carnosine [31]. Chen et al. [32] further extended this using brain microdialysis and found that deletion of PepT2 increased brain interstitial fluid as well as CSF cefadroxil concentrations relative to blood. Smith et al. [33] also used quantitative autoradiography to examine the biodistribution of [14C]GlySar after intracerebroventricular administration (i.e., direct into CSF) and showed markedly lower CP uptake in the PepT2 KO mouse but also altered penetration into brain parenchyma. The potential importance of PepT2 was further underlined by experiments demonstrating that systemic administration of 5-ALA not only caused higher CSF 5-ALA concentrations but it also caused more neurotoxicity in PepT2 KO mice [34] (Fig. 2). Similarly, the analgesic effect of the neuropeptide kyotorphin given into the cerebral ventricle was greater in PepT2 KO mice than wildtype by reducing clearance from the CSF [35].

Fig. 2
figure 2

5-Aminolevulinic acid (5-ALA) and PepT2 in relation to brain health and disease. PepT2 at the choroid plexus epithelium clears 5-ALA from CSF and that protects against 5-ALA induced neurotoxicity [34]. The ependyma lining the cerebral ventricles is leaky allowing movement of 5-ALA between CSF and brain interstitial fluid (ISF). 5-ALA is the first compound in the porphyrin/heme synthesis pathway and parenchymal cell uptake of 5-ALA via PepT2 may contribute to heme synthesis in the normal brain although 5-ALA is also produced from succinyl-CoA and glycine. In gliomas (and different systemic cancers), 5-ALA administration leads to a build-up in protoporphyrin IX (PPIX) which can be used for fluorescence guided surgery. The role of PepT2 in that selective fluorescence is being investigated (e.g., [70]) as is the role of reduced ferrochelatase (which converts PPIX to heme) activity. Porphyrias are caused by mutations in the enzymes involved in heme synthesis. There is a feedback loop where heme inhibits the conversion of succinyl-CoA and glycine to 5-ALA by ALA synthase and mutations in the enzymes involved in heme synthesis cause a buildup in 5-ALA and other (depending on the mutation) early intermediates in the heme pathway. Whether PepT2 impacts porphyrias is being examined in brain [74] and systemically [73]

To investigate which cells express PepT2 at different stages of brain development, Shen et al. [10] used immunohistochemistry and immunoblot. In cerebral cortex, they found the highest expression in the fetus with an 86% reduction in adult. In neonatal brain, PepT2 colocalized with both an astrocyte marker, glial fibrillary acidic protein (GFAP), and the neuronal marker, NeuN. However, in the adult it just colocalized with NeuN. The role of PepT2 in neonatal astrocytes oligopeptide transport was confirmed by in vitro experiments, where neonatal astrocytes from PepT2 KO mice had a 94% reduction in [14C]GlySar uptake compared to those isolated from wildtype mice [36] as well as marked decreases in the uptake of 5-ALA, carnosine and L-kyotorphin [37, 38]. A role of neuronal PepT2 in oligopeptide transport has been confirmed by other groups using adult rat synaptosomes [39] and fetal neuronal cultures [40]. It should be noted that there is controversy in regard to adult astrocytes where high PepT2 mRNA expression has been reported [41]. There is a need for PepT2 transport studies in adult astrocytes.

PhT1-Mediated Transport

The peptide/histidine transporters PhT1 (SLC15A4) and PhT2 (SLC15A3) were first cloned by Yamashita et al. [7] and Sakata et al. [9]. The former was cloned from brain where it is highly expressed [7]. In contrast, Sakata et al. found PhT2 was only faintly detected in brain [9].

In initial experiments by David Smith and colleagues, L-histidine had no effect on oligopeptide uptake into rodent CP indicating that that PhT1 or PhT2 were not involved in transport at the blood-CSF barrier [20, 27]. However, they did find that PhT1 is highly expressed in the adult mouse and rat brain parenchyma at the both the mRNA and protein levels and that L-histidine markedly reduced GlySar uptake into brain (cortex, cerebellum or hippocampal) slices whereas PepT2 deletion had little effect [8]. This was reversed in neonatal brain slices where PepT2 deletion caused a marked reduction in GlySar uptake and L-histidine had little effect. This reflected opposite changes in the expression of PhT1 and PepT2 during brain development with PhT1 markedly increasing with age [8] while PepT2 decreases [10] (Fig. 3).

Fig. 3
figure 3

Results from Shen et al. [10] and Hu et al. [8] indicate opposite effects of development on PepT2 (SLC15A2) and PhT1 (SLC15A4) protein expression in the rat brain. While PepT2 is highly expressed in the fetus, it declines in the postnatal period. In contrast, PhT1 expression was very low in the neonate but progressively increase from day 14, day 21, day 28 to adult. The change in PepT2 with age reflects a reduction in astrocyte expression in the adult rat with neuronal expression being found both in neonates and adults

To examine the function of PhT1, David Smith and colleagues examined PhT1 null mice [42]. Those mice were healthy, fertile with no difference in body weight or serum chemistry from wildtype mice. This might in part be due to a compensatory increase in PepT2 expression in the brain. Loss of PhT1 resulted in lower L-histidine transport into brain slices in vitro and lower L-[14C]histidine concentrations in brain areas after intravenous administration even though plasma pharmacokinetics were not altered.

Implications

This work has been fundamental for our understanding of both the actions of endogenous neuropeptides and CNS acting drugs. Thus, PepT2 at the CP and on brain parenchymal cells and PhT1 in the brain parenchyma play a critical role in determining the extracellular concentrations of small neuropeptides and peptidomimetics which in turn effects their activity; i.e., they may be an important brain regulatory mechanism. In addition, PepT2 and PhT1 may play a role in the uptake cleavage products of larger neuropeptides either to limit their neuroactivity or for use in the synthesis of new peptides. Being present at the blood-CSF barrier, PepT2 may also serve to protect the brain from the neurotoxic effects of circulating peptides/peptidomimetics (as shown with 5-ALA). While cefadroxil is a PepT2 substrate, cephalosporins that do not contain an a-amino group are not and those are the drugs currently being used to treat meningitis [43].

Recent Advances and Potential Directions

Brain Cellular Distribution and Potential Subcellular Compartments

The blood-CSF barrier is located at the CPs in the lateral, third and fourth ventricle, and the arachnoid membrane that surrounds the brain. Historically, more focus has been on the highly vascularized CP which is a site of CSF secretion. As outlined above, David Smith and colleagues showed the importance of PepT2 in the clearance of oligopeptides and peptidomimetics from CSF to the CP epithelium. Recently, there has been growing interest in the functions of the arachnoid membrane. That has included using quantitative proteomics to examine the absolute expression of transporters in the arachnoid membrane in dog, pig and human [11, 12, 44]. Those studies have shown that PepT2 is highly expressed in the arachnoid of all three species. It is present on the CSF facing membrane of the arachnoid [12] suggesting that it has an important role in controlling peptide concentrations in CSF within the subarachnoid space. This may prevent neuropeptide release into CSF from affecting other brain areas.

While brain POT studies have focused on CP, astrocytes, neurons and now the arachnoid membrane, it is possible that they may have a role in other cell types or subsets of cells. In the periphery, Sasawatari et al. [45] have found PhT1 in dendritic cells, activated macrophages and B cells. David Smith and colleagues have shown that macrophages express PepT2, PhT1 and PhT2 where they play a role in inflammatory responses to bacteria [46,47,48,49]. In brain, the primary resident macrophages are the microglia, but there are also border-associated macrophages at the dura mater, the subdural meninges, the CP and in the perivascular space [50]. Whether these cells normally express POTs or whether they may be induced to express POTs during neuroinflammation requires investigation. Current single-cell RNA-sequencing data sets from brain may be an invaluable resource.

In addition to resident macrophages, neuroinflammatory conditions (e.g., multiple sclerosis, stroke and meningitis) result in an influx of leukocytes into brain including monocyte-derived macrophages. Based on the work in the periphery, it is likely that these cells will express PepT2, PhT1 and PhT2. The role of POTs in different neuroinflammatory responses is still unknown.

In relation to neurons, it will be important to know which cell types express POTs. Is it related to specific neurotransmitters or is there a more general expression? For astrocytes, while there is agreement as to the role of PepT2 in neonates [36, 51, 52] there are apparently contradictory results as to the role of PepT2 in adults [10, 41] that need to be investigated. In addition, adult astrocytes are not a single population. They differ by location and they display a range of phenotypes after injury (astrogliosis) which can be beneficial or detrimental [53]. Whether PepT2 expression and function differs in these diverse populations should be examined.

Most work on brain POTs has focused on plasma membrane transport. However, studies of David Smith and colleagues on peripheral macrophages has indicated that PhT2 is lysosomal, transporting substrate from the lysosome to the cytosol [48] and that while PepT2 is expressed on the plasma membrane, PhT1 is present on the endosomes [49]. A detailed analysis of the subcellular distribution of different POTs in the brain is warranted.

Regulatory Pathways

Understanding how POT activity is regulated may allow manipulation of systemic and brain drug delivery. A wide range of mechanisms that regulate POT expression and/or activity have been described in different tissues. For example, Wang et al. [54] have recently reviewed the regulation of PepT2. That includes effects of di- and tripeptides, protein restriction, hormones and corticosteroids, as well as regulation by protein kinase C, phosphoinositide 3-kinase, mitogen-activated protein kinase and nuclear factor kappa B (NFkB) pathways. POT transport is pH-gradient dependent and PDZK1 (also known as Na/H exchange regulatory cofactor NHE-RF3) enhances PepT2 activity in kidney cells [55]. These effects are likely tissue specific and there is a need studies directed at regulatory pathways at the CP and in brain parenchymal cells.

POTs may also be important in regulating other cell responses. As mentioned above, PepT2, PhT1 and PhT2 play a role in inflammatory responses of systemic macrophages to bacteria by transporting bacteria-derived peptides [46,47,48,49]. The interplay between pro-inflammatory pathways and POTs is complex. Song et al. [46] found that ligands to the toll-like receptors (TLRs), TLR2, TLR4, TLR7 and TLR9 increased expression of PhT2 in macrophages. This may be related to TLRs activating NFkB, as there are putative binding sites for NFkB on the promoter region of PhT2. However, Hu et al. [49] showed that KO of PepT2 and PhT1 reduced the cytokine release engendered by bacterial peptides suggesting that POTs can be upregulated by pro-inflammatory pathways and enhance the inflammatory response. The role of POTs in neuroinflammation is still unknown, but it is interesting that the CP is becoming increasingly recognized as an important site regulating neuroinflammation [18].

Another example of POT activity regulating other pathways has been described by Zhang et al. [56] in Dictyostelium cells. They found deletion of SLC15A reduces micropinocytosis and cell growth by limiting the availability of key amino acids. Whether reducing POTs has a similar effect on endocytotic pathways in mammalian cells (for example at the CP epithelium) merits exploration.

Transporter Structure

POTs show great promiscuity in terms of substrates (all di- and tripeptides as well as some peptidomimetics). Understanding the basis of that promiscuity may aid in developing drugs that are substrates (e.g., for greater oral bioavailability or renal reabsorption) or are not substrates (e.g., for greater CSF penetration). Most work on POT structure has been performed on procaryotes. Thus, Minhas and Newstead [57] recently examined the crystal structure of PepTsh from Staphylococcus hominis, which is a homolog of the mammalian PepT1. By examining the binding of both valacyclovir and 5-ALA to PepTsh they were able to construct a pharmacophore model to aid drug design. Similarly, Ural-Blimke et al. [58] have crystalized the peptide transporter DtpA from Escherichia coli and examined its interactions with valganciclovir. Although the structures of valacyclovir and valganciclovir are very similar, the peptide transporter binding was different, with the valganciclovir orientation in the ligand recognition site being flipped almost 180° [59]. Available information on procaryote transporters has been reviewed in [59]. However, there are differences between procaryote and mammalian peptide transporters (for example, in substrate preferences and the large extracellular domain of mammalian PepT1 and PepT2) and, thus, recent work of Parker et al. [60] using cryo-EM to elucidate the structure of the mammalian (rat) PepT2 is very important for developing a framework for mammalian transport.

Species Differences

Even between mammalian species there are differences in POT activity. For example, David Smith and colleagues expressed cDNA of human, mouse and rat PepT2 in yeast Pichia pastoris cells [61]. They found that the Km for both GlySar and cefadroxil transport was much higher for human PepT2 than for mouse or rat PepT2. One approach to examine the importance of such species differences is to produce humanized mice expressing the human but not the mouse transporter. This has been done for PepT1 [62] but not for PepT2.

There can also be differences in tissue POT expression between species. Quantitative proteomics has been used to compare PepT2 expression at the CP and the arachnoid membrane across species and have found differences in absolute transporter levels that may impact the levels of PepT2 substrates in CSF [44]. Such studies have also indicated that while the absolute level (fmol/mg protein) of PepT2 is greater in CP than the arachnoid membrane in rats, the opposite is found in human [44, 63, 64].

Disease States

5-Aminolevulinic acid (5-ALA) is a precursor for porphyrin synthesis and 5-ALA administration results in cellular accumulation of protoporphyrin IX (PPIX) in a variety of cancers making them sensitive to photodynamic therapy [65]. 5-ALA-induced PPIX accumulation is also used for fluorescence-guided surgery, including brain gliomas [66]. The mechanism(s) underlying why 5-ALA preferentially causes PPIX accumulation and fluorescence in cancer cells is still not entirely certain [67, 68]. While it may involve reduced levels of ferrochelatase, an enzyme that metabolizes PPIX, 5-ALA is a peptidomimetic and POT substrate [19, 69] raising the question of the importance of POTs in the photosensitization and location of tumors using 5-ALA (Fig. 2). Hou et al. [70] have found that PepT2 mRNA and protein levels are higher in grade II and III gliomas that fluoresce during 5-ALA fluorescence-guided surgery than those that do not. They also found that a PepT2 small interference (si)RNA reduced PPIX fluorescence in a grade III glioma cell line (SW-1783) treated with 5-ALA in vitro.

Porphyrias are caused by mutations in the genes involved in heme biosynthesis. They cause both peripheral and nervous system symptoms and porphyric attacks can cause acute encephalopathy. ALA synthase is the rate limiting enzyme in heme biosynthesis and the role of PepT2 in 5-ALA transport and 5-ALA toxicity has suggested that it might have a role in porphyrias [71] (Fig. 2). Genetic polymorphisms impact POT function. Thus, two haplotypes of PepT2 have been described in human, hPepT2*1 and hPepT2*2, with the former having a 3-fold lower Km for GlySar and different pH sensitivity [72]. Tchernitchko et al. [73] found that two copies of the high affinity haplotype (hPepT2*1/1) predicted worse porphyria-associated kidney disease. However, Pischik et al. [74] recently failed to find a significant relationship between PepT2 haplotype and encephalopathy in patients with acute hepatic porphyria although they note that this might be due to the low number of patients.

Lead binds and inactivates the enzyme ALA dehydratase, which is part of the heme biosynthesis pathway. This results in increased brain 5-ALA. In a study examining children with low-level lead exposure, Sobin et al. [75] found that children with the hPepT2*2 haplotype had worse motor dexterity and working memory that was independent of blood lead levels. It was speculated that this might relate to differences in brain 5-ALA concentrations.

Acute or chronic liver disease results in increased blood ammonia concentrations which may lead to hepatic encephalopathy and death. In brain, ammonia is detoxified by conversion of L-glutamate and ammonia to L-glutamine resulting in marked increases in brain glutamine in hepatic encephalopathy. Thus, Tossman et al. [76] reported a ~ 6-fold increase in brain extracellular L-glutamine concentration and Zanchin et al. [77] found a 2.5 to 3-fold increase in total brain tissue L-glutamine in rat models of hepatic encephalopathy, while Weiser et al. [78] reported a 4-fold increase in brain L-glutamine in patients who died after hepatic coma. These changes in L-glutamine are associated with changes in other brain amino acids, including significant increases in L-histidine [77, 78]. Whether a derangement in brain L-histidine will impact PhT1-mediated uptake of endogenous di- and tripeptides and peptidomimetics, potentially contributing to brain dysfunction in liver disease, merits investigation.

Multiple studies have shown an association between PhT1 (SLC15A4) polymorphisms and an autoimmune disease, systemic lupus erythematosus [79,80,81,82]. Rimann et al. [83] have dissected the role PhT1 and found that it is required to traffic nucleic acid-sensing TLRs and their ligands to the endolysosomes in TLR-activated plasmacytoid dendritic cells, trafficking which regulates type I interferon production. Sasawatari et al. [45] have also that SLC15A4 KO mice develop a less severe form of colitis. Such results are another reason for further examination of the role of POTs in neuroinflammation, including in meningitis and encephalitis.

Conclusions

This mini-review highlights the enormous contributions that David Smith has made to our understanding of the role of protein-coupled oligopeptide transport (SLC15) in the brain (as well as elsewhere in the body). Neuropeptides and peptidomimetics have crucial roles in brain function and the POT family has important regulatory roles. Similarly, POTs have important actions in relation to the distribution of therapeutics. His work has formed a foundation for future studies in the brain, such as examining the role of POTs in neuroinflammation and the importance of POT expression in different subcellular compartments.