Decreased blood–brain Barrier Expression of P-glycoprotein in Alzheimer‘s Disease: Impact on Pathogenesis and Brain Access of Therapeutic Agents

Keywords::

• β-amyloid
• Alzheimer‘s disease
• blood–brain barrier
• CNS
• efflux transporter
• P-glycoprotein

The blood–brain barrier (BBB), formed by the endothelial cells lining the cerebral microvasculature, was once thought to act as a simple physical barrier protecting the brain parenchyma from blood-borne agents. However, with the revitalization of cerebral vascular research, it is becoming more recognized that the BBB is a dynamic interface that plays a critical role in communication between the sensitive CNS and the periphery. This regulatory role of the BBB is largely mediated through specialized transport proteins expressed at the luminal (blood-facing) and abluminal (brain-facing) membranes of brain endothelial cells [1,2], in addition to the inter-endothelial tight junction proteins [3]. The transporter proteins normally act to regulate the movement of endogenous molecules across the BBB (through both influx and efflux mechanisms), and drug-delivery scientists have attempted to manipulate these transporters to enhance the entry of therapeutic agents into the CNS. Of significant interest to pharmaceutical scientists is P-glycoprotein (P-gp), the 170 kDa ATP-binding cassette efflux protein expressed at the luminal surface of brain endothelial cells. Due to its ability to prevent the CNS entry of various structurally-unrelated drugs such as dixogin, loperamide, paclitaxel and cyclosporin A, it has been referred to as the ‘gatekeeper‘ at the BBB [4].

The expression and function of P-gp has been reported to be altered in various neurological and peripheral disorders, including Alzheimer‘s disease (AD), Parkinson‘s disease, epilepsy, HIV infection [5], and as described by Ronaldson and Davis in a forthcoming issue of Therapeutic Delivery, peripheral inflammatory pain [6]. Being the gatekeeper at the BBB, alteration of the functional activity of P-gp may have an impact on the CNS disposition of systemically-administered therapeutic agents. While there are reports demonstrating altered CNS disposition of drugs in some of the aforementioned diseases, there are limited studies addressing this issue in AD. The purpose of this editorial, therefore, is to provide an overview of P-gp expression and function in AD, and highlight the potential implications of this disorder on access of drugs into the CNS.

AD & P-gp: what is the link?

The cognitive decline observed in AD is associated with brain parenchymal accumulation of the neurotoxic peptide β-amyloid (Aβ) and intraneuronal lesions comprised of hyperphosphorylated tau protein [7]. The amyloid cascade hypothesis suggests that the accumulation of Aβ species of both 40 and 42 amino acid length (i.e., Aβ_1–40 and Aβ_1–42) leads to microglial activation, release of inflammatory mediators and reactive oxygen species, hyperphosphorylation of tau and subsequent neuronal death [7]. Various mechanisms have been suggested to be responsible for the accumulation of Aβ, including increased production (albeit overexpression of the amyloid precursor protein is rare in most AD cases), decreased metabolism and/or decreased clearance of Aβ across the BBB. The ‘neurovascular hypothesis of AD‘ suggests that accumulation of Aβ is a result of dysfunctional BBB efflux mechanisms [8], mainly as a result of decreased expression and function of the abluminal Aβ transporter, low density lipoprotein receptor-related protein (LRP)-1 [9]. While most research assessing BBB efflux of Aβ has focused on LRP-1, there are also suggestions that P-gp is involved in the BBB clearance of Aβ, facilitating the egress of Aβ from the brain endothelial cytoplasm into the systemic circulation. These suggestions are based on the following observations:

▪	Aβ interacts with purified hamster P-gp reconstituted into vesicles [10];
▪	Basolateral-to-apical transport of Aβ is higher in kidney epithelial cells stably transfected with human P-gp (relative to parental cells), which is reversed in the presence of the P-gp inhibitor cyclosporin A [11];
▪	BBB efflux of exogenous Aβ is significantly reduced in P-gp-deficient mice (relative to wild-type mice) [12];
▪	Administration of the P-gp inhibitor XR9576 leads to a 30% increase in interstitial fluid concentrations of endogenous Aβ in a mouse model of AD [12];
▪	Administration of a pregnane X receptor (PXR) agonist, which leads to a specific increase in P-gp expression, results in enhanced abluminal-to-luminal transport of exogenous Aβ and reduced endogenous Aβ levels in an AD mouse model [13].

In addition, others have demonstrated that P-gp restricts the luminal uptake of Aβ into human and bovine brain endothelial cells [14,15], suggesting that P-gp may limit plasma-derived Aβ from entering the brain parenchyma. Despite these findings, there are also reports suggesting a limited role of P-gp in mediating endothelial uptake and BBB efflux of Aβ in rat models [16,17]; however, this may be due to species differences in P-gp binding. Similarly, another study demonstrated no difference in the basolateral-to-apical transport of Aβ between parental Madin-Darby canine kidney cells and P-gp-transfected cells [18]. Therefore, no definitive conclusion can be made on whether P-gp mediates the BBB clearance of Aβ in human AD, albeit ample data is available to implicate P-gp in the transport of Aβ across the mouse BBB.

BBB expression of P-gp in AD

If P-gp is a critical component in the clearance of Aβ across the BBB, it might be expected that the expression and/or function of this transporter is attenuated in AD, and indeed, this is supported by literature studies. Suggestions that P-gp expression may be lower in AD patients stem from an immunohistochemical study demonstrating an inverse relationship between vascular P-gp immunoreactivity and Aβ-positive plaques in the brains of nondemented, elderly humans [19]. Further studies demonstrated that brain endothelial P-gp and vascular Aβ were never co-localized in nondemented, elderly subjects with cerebral amyloid angiopathy, which is one of the pathological features of AD [20]. These studies demonstrate a relationship between vascular P-gp and Aβ in nondemented subjects; however, an appreciation of what occurs in patients with confirmed AD diagnoses has also recently become available. In one recent study, AD subjects exhibited significantly lower hippocampal microvascular expression of P-gp compared with age-matched non-AD subjects [21], and similarly, another study identified a negative relationship between P-gp-positive capillaries and Aβ₄₀-positive staining in the superior temporal cortex of AD subjects [22]. As has been observed in humans, reduced brain microvascular expression of P-gp has also been reported in a mouse model of AD [13]. Therefore, there is ample evidence demonstrating that microvascular P-gp expression is reduced in AD patients; however, whether this occurs prior to, or as a result of Aβ accumulation, remains unknown. Furthermore, the underlying biochemical mechanisms responsible for this downregulation in P-gp expression currently remain to be identified. If P-gp is indeed involved in mediating the BBB clearance of Aβ in humans, unravelling the underlying mechanisms regulating its expression may lead to novel approaches for reducing parenchymal Aβ burden, as demonstrated in mice using the PXR agonism approach [13].

Impact on CNS drug exposure in AD

Despite the controversial role of P-gp in Aβ accumulation, the attenuated brain microvascular expression in AD is expected to be associated with reduced protective function of the BBB. This would be even more pronounced if the tight junctions sealing the paracellular route of the BBB are disturbed, albeit direct evidence demonstrating enhanced paracellular diffusion in a relevant AD model or in AD patients is still lacking. As outlined by Ronaldson and Davis in a forthcoming issue of Therapeutic Delivery, alteration in P-gp expression at the BBB leads to altered brain disposition of therapeutics with affinity to P-gp [6]. For example, associated with limbic (kainic acid induced) seizures is an upregulation of hippocampal P-gp expression, and this leads to decreased brain uptake of the weak P-gp substrate phenytoin [23]. In addition, the BBB transport of morphine, another P-gp substrate, is decreased in a rat model of peripheral inflammatory pain where brain microvascular P-gp expression is enhanced [24]. By contrast, conditions associated with decreased microvascular expression or function of P-gp (e.g., systemic inflammation, late-stage Parkinson‘s disease, and normal aging) have been associated with increased CNS exposure of the P-gp substrate, verapamil [25–27]. It is likely, therefore, that the BBB transport of similar P-gp substrates would be enhanced in AD where the expression of this efflux pump is reduced. Indeed, a very recent communication reports that the brain deposition or binding potential of ¹¹C-verapamil, as assessed by PET, is significantly higher in AD patients compared with healthy controls [28], likely as a result of decreased P-gp functionality at the BBB. This is the first study to demonstrate enhanced brain disposition of a therapeutic agent in AD, albeit it has been suggested by many in the past despite the paucity of data to support these claims, as reviewed recently in Therapeutic Delivery[29]. It is likely that the brain uptake of other P-gp substrates, such as digoxin, loperamide and quinidine, may also be enhanced in AD, potentially placing AD patients at risk of enhanced exposure and possibly neurotoxicity from therapeutic agents that normally do not gain access to the CNS. However, one needs to keep in mind that the reduced expression (and possibly function) of P-gp in AD may be offset by other disease-related BBB alterations, including increased expression of the other major efflux transporter breast cancer resistance protein, increased thickness of the microvascular basement membrane and reduction in cerebral blood flow rate [29]. Therefore, to gain a greater understanding of the impact of attenuated P-gp expression on the CNS disposition of therapeutics in AD, a systematic evaluation of the BBB transport of various molecules with varying affinity to P-gp is recommended. Such studies would first need to be undertaken in a relevant mouse model of AD, and supported by human PET studies, in order to clearly identify whether AD patients are indeed at greater risk of unnecessary CNS exposure from systemically acting drugs.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.