As previously mentioned, both influx and efflux Aβ transporters seem to be dysregulated in AD and have shown to be key players in Aβ accumulation. In this context, these transporters have been presented as potential therapeutic targets for AD. RAGE blockage have showed to slow down Aβ pathology and lower the rate of cognitive decline in animal models (Bell et al., 2012). However, the phase III clinical trial with a RAGE inhibitor (NCT02080364) was terminated due to lack of efficacy. Statins have been shown to upregulate LRP-1 at the BBB and reduce Aβ brain levels in vitro models. Although clinical trials have failed to show their efficacy, a recent reanalysis of these studies suggested that these drugs may benefit AD patients with potentially greater therapeutic efficacy in those homozygous for the APOE ε4 allele (Storck and Pietrzik, 2017), though it is not known whether the modulation of LRP-1 plays any role. In vitro studies support the role of P-gp in removing amyloid proteins (McCormick et al., 2021) and in vivo studies showed that prevention of P-gp degradation lowers Aβ brain levels (Hartz et al., 2018). However, to our knowledge, no LRP-1 or P-gp modulators have been clinically tested in AD yet.

The pathogenesis of AD has also been associated with a decreased expression of GLUT1, the major glucose transporter in the BBB. Liraglutide, a GLP1 analog, has showed to delay memory decline in a mouse model of AD (Hansen et al., 2015). Recently, it was demonstrated that the same drug slowed down memory decline in a group of patients with obesity and type 2 diabetes (Vadini et al., 2020) and more recently, GLP1 receptor agonists were shown to prevent glucose transport decline through BBB in AD patients, although no conclusions were drawn on cognitive decline (Gejl et al., 2016). Although little is known about the specific mechanism, it is thought that they could act by restoring the levels of GLUT1 at the BBB (Gejl et al., 2017). Also, a recent study pulling out data from randomized controlled trials and a nationwide cohort concluded that dementia incidence was reduced in patients with type 2 diabetes treated with GLP1 receptor agonists (Nørgaard et al., 2022). Based on these results, a phase IIb clinical trial (NCT01843075) demonstrated liraglutide improved cognitive function and MRI volume in patients with mild to moderate AD (Edison et al., 2021). Taken together, these data have been supporting GLP1 analogues as potential disease modifying therapy for AD. A phase III randomized controlled trial evaluating semaglutide in people with early AD is ongoing (NCT04777396).

Lastly, inhibition of CypA-MMP-9 cascade, a pathway involved in BBB breakdown, in addition to repairing the BBB, also slowed down and reversed neurodegenerative changes in animal models (Bell et al., 2012). Table 1 presents BBB transporters/pathways implicated in AD and how they can be targeted. Figure 2 outlines BBB targeting strategies in AD.

Despite positive outcomes from pre-clinical studies in vitro and in vivo, the scarcity of clinical trials emphasizes the necessity for additional research to bring these discoveries from the laboratory to the patient bedside.

Although the genetic underpins of BBB dysfunction are relatively understudied, genetic data has given a number of possible links between BBB dysfunction and AD. For instance, APOE ε4, the strongest genetic risk factor for AD (Farrer et al., 1997), has also been shown to lead to brain blood dysfunction and cognitive decline, independently of AD pathology (Montagne et al., 2020). This supports a link between BBB dysfunction and AD, but also suggests that BBB dysfunction may add cognitive decline independently. More recently, it has been shown that 30 of the top 45 genes that have been linked to AD risk by genome-wide association studies (GWAS) are expressed in the human brain vasculature, suggesting that vascular and perivascular systems are highly involved in AD pathogenesis (Yang et al., 2022). Genetic variants such as PICALM, BIN1, CD2AP, and RIN3 have been linked to AD in GWAS and seem involved in BBB Aβ transcytosis pathways (Juul Rasmussen et al., 2019). A recent study has shown these variants to increase the risk of AD and vascular dementia, suggesting that there is a shared mechanism, possibly related to vascular or perivascular dysfunction (Juul Rasmussen et al., 2019).

Several monogenic mendelian diseases leading to brain disorders and cognitive impairment appear to originate in individual cell types of the neurovascular unit. Some of the genes are linked to specific roles in BBB development, function, and regulation (Zhao et al., 2015) but have also been linked with AD associated changes, such as OCLN (Romanitan et al., 2007), COL4A1 (Pang et al., 2019; Xiao et al., 2021), NOTCH3 (Sassi et al., 2018), CSF1R (Sassi et al., 2018), TREM2 (Guerreiro et al., 2013), and TYROBP (Audrain et al., 2021). Furthermore, there is some evidence of BBB dysfunction happening in a number of genetically determined degenerative disorders, such as AD, but also Parkinson’s disease, Huntington’s disease and amyotrophic lateral sclerosis (Katt et al., 2019).

Other studies such as those enumerated in Table 2 have identified other potential genetic links between BBB regulation and development and AD pathogenesis, providing additional potential drug targets. For example, GLUT1 is the most important energy carrier of the brain across the BBB, and shown to be critical in BBB development (Zheng et al., 2010). It has been reported that GLUT1 reductions exacerbate AD pathology (Winkler et al., 2015). Although the syndromes currently associated with GLUT1 pathogenic variants start at a very young age (Koch and Weber, 2019), small-effect changes with low frequency in the population can potentially be associated with the risk of developing AD. This is further supported by some genetic data in drosophila models (Shulman et al., 2011). Likewise, MFSD2A, another major transporter with a key role in modulating transcytosis to regulate the BBB (Andreone et al., 2017), expression is reduced in AD patients (Sánchez-Campillo et al., 2019).

A GWAS (Lee et al., 2022) has looked into FML2, a formin-related protein expressed in astroglial cells that regulares glia and vasculature interaction and partakes in the homeostasis and clearance of amyloid β. It has shown that FML2 is overexpressed in AD and in cerebrovascular pathology independently. Expression of FMNL2 increases in the presence of vascular risk factors among individuals with AD, and it appears that this gene plays a significant role in the progression of AD pathology when both conditions are present.

Several other transporters may also link both disorders. PRF1, for example, which has been linked with BBB dysfunction (Willenbring et al., 2016), has also been linked with amyloid processing, being shown to promote amyloid-beta internalization in neurons (Lana et al., 2017). Recently, UNC5B has been suggested to control BBB integrity (Boyé et al., 2022). This gene is of the same family of UNC5C, which has been repeatedly linked with AD. Another link between AD and BBB dysfunction comes from a model of multiple sclerosis, where FOXO1 downregulation has also been linked to BBB changes (Mora et al., 2020). Interestingly, FOXO1 upregulation has been shown to reduce Aβ production and tau phosphorylation in vitro (Zhang et al., 2020).

This data gives support to the idea that BBB dysfunction is an important step in AD pathophysiology, possibly by facilitating known mechanisms but perhaps also by adding additional damage to AD patients’ brains by other mechanisms.

Nanomaterials for targeted drug delivery across the BBB have been extensively studied in the medicine field, particularly in AD (Faiyaz et al., 2021). In order to cross effectively the BBB and target directly the AD brain, nanoformulations should have: stability in blood circulation, proper surface modification, amyloid direct targeting that does not interfere with BBB targeting (multifunctionality) and ability to release at the target site (Tosi et al., 2019).

A vast variety of nanotechnology-based approaches fulfilling these characteristics have been and are currently being adapted to surpass the BBB in AD (Ordóñez-Gutiérrez and Wandosell, 2020; Srivastava et al., 2021; Chopra et al., 2022). Some relevant examples of AD nanoformulations include lipoprotein-based, polymers or metallic nanoparticles (Karthivashan et al., 2018; Bilal et al., 2020; Faiyaz et al., 2021).

The rationale behind these nanotechnology-based approaches is to take advantage of the intrinsic transport strategies of the BBB to efficiently reach brain tissue and target AD’s pathological processes. For instance, lipophilic nanoparticles directed toward the brain endothelial cells (BECs) could allow the transport of therapeutic compounds through endocytosis or lipophilic transcellular pathways (Karthivashan et al., 2018). In this context, solid lipid nanoparticles loaded with curcumin (a potent antioxidant, anti-inflammatory compound with anti-Aβ accumulation effects) have shown to efficiently surpass the BBB due to their lipidic nature and to reduce oxidative stress in the hippocampal tissue improving spatial memory in an AD rodent model (Sadegh Malvajerd et al., 2019). Other strategies such as ionized-nanomaterials could take advantage of the negatively charged BECs membrane to surpass the BBB through adsorptive transcytosis. For example, the conjugation of deferasirox (an iron-chelating agent) to cationized human serum albumin showed correct brain uptake though adsorptive transcytosis with attenuated amyloid beta-induced learning deficits in a rat model of AD (Kamalinia et al., 2015). Another strategy comprises functionalized nanomateriales including liposomes, polymeric or metallic nanocarriers. This strategy relies on receptor-mediated transcytosis and carrier protein-mediated pathways taking advantage of the already existing receptors in the cells surface to transport the therapeutic cargo across the BBB (Karthivashan et al., 2018). A good example of this mechanism is the use of LPR-1 as delivery system across the BBB. Polymeric nanoparticles functionalized with Angiopep-2 (a specific ligand to LPR-1) and loaded with Prussian blue showed efficient BBB crossing and restoration of mitochondrial function along with reduced neurotoxic Aβ aggregation (Zhong et al., 2022). A gold standard in receptor mediated transcytosis is the transferrin receptor (TfR). The TfR is highly expressed in the BBB and has been prove to be functional in the BBB during AD (Bourassa et al., 2019), hence conforming a great candidate for brain delivery in AD. In this context, transferrin-functionalized liposomes loaded with gallic acid where proven to efficiently bypass the BBB through TfR-mediated transcytosis and efficiently decrease the number of Aβ fibrils formed by 56% (Andrade et al., 2022).

Despite all, none of these strategies aiming to a single AD process has achieved the expected results in terms of disease treatment. To this aim, and due to the complex nature of AD and the multifactorial regulation of Aβ aggregation, multitarget nanotherapeutics seem to be a promising strategy to target not only this hallmark but also complementary pathways involved in this disease (Ibrahim and Gabr, 2019). In this context, self-assembled Fmoc-Trp-Fe^{2 +} -Que NPs have shown successful suppression of Aβ plaques, reduction of reactive oxygen species (ROS) generation and suppression of the neurotoxicity induced by Aβ (Zhu et al., 2022). Furthermore, chondroitin sulfate-selenium NPs have also shown to be very promising as multitarget AD therapy showing reduction in ROS levels and cytoskeleton damage, attenuation of hyperphosphorylation of tau and Aβ aggregation inhibition (Gao et al., 2020).

All in all, nanoformulations present themselves as one of the most promising tools for AD treatment, taking advantage of the selective BBB characteristics and using it as a delivery platform to selectively target the brain. Nonetheless, albeit very promising, its use is limited due to safety concerns and in most cases failure to achieve adequate concentrations of the delivered compounds to the brain tissue (Lipsman et al., 2018). Moreover, BBB transport mechanisms can be altered during neurodegeneration which will, in turn, affect the delivery of therapeutic compounds and hamper the use of this mechanism as a delivery platform (Wasielewska and White, 2022).

In this context, controlled opening of the BBB has been brought up as an alternative potential mechanism to facilitate paracellular drug delivery into the brain using the BBB as a platform.

Controlled BBB opening should be transient and selective to avoid unwanted accumulation in the brain and any potential side effects (Han, 2021). Several pre-clinical and clinical studies have studied this controlled opening of the BBB as a way to the brain in the treatment of AD. For instance, mannitol has been used in vitro to reversibly open the BBB (Haluska and Anthony, 2004). Nonetheless, focused ultrasound in combination with microbubbles currently constitutes the only truly transient, localized, and non-invasive technique for opening the BBB (Konofagou et al., 2012). Significant progress has been made in the pre-clinical validation and development of focused ultrasound which eventually lead to the initiation of clinical trials examining its application for delivery in AD (Wasielewska and White, 2022). In fact, this technique has been shown to reversibly open BBB in AD patients. A clinical trial investigated BBB opening in the white matter of the superficial dorsolateral prefrontal cortex of AD patients demonstrating that this opening is safe in humans (Lipsman et al., 2018). Hippocampal BBB opening has been also demonstrated safely in a proof-of-concept study that also reported an Aβ decrease after opening (Rezai et al., 2020). In this line, a recent phase I/II trial in early stage AD showed that repeatedly opening the BBB with ultrasound is well-tolerated and may be associated with a reduction of amyloid burden (Epelbaum et al., 2022). Extensive opening of the BBB (above 20 cm³) has also been shown safe and potentially beneficial with a potential anti-amyloid effect by itself (Park et al., 2021).

While the potential benefits of BBB opening by itself are still not clear, the proof of safety in transiently opening the BBB of AD patients could additionally allow the delivery of larger molecules such as antibodies or growth factors (Lipsman et al., 2018) that directly target the AD pathology taking advantage of the increased permeability. Nonetheless, BBB opening raised several concerns such as potential brain toxicity due to the non-specific accumulation of neurotoxic substances from the blood that could produce neuronal damage and degenerative changes (Han, 2021).

Interestingly, BBB disruption has been reported to occur in AD even before the onset of hippocampal atrophy (Montagne et al., 2015; Sweeney et al., 2018) and hence the evidence of a naturally permeable BBB in the early phases of the disease could allow the delivery of therapeutic compounds even at the early stages of AD without invasive measures.

Currently, no disease-modifying therapies for AD are approved by the European Medicines Agency (EMA). In the USA, Aducanumab (Aduhelm™) has been given Food and Drug Agency (FDA) approval under the agency’s accelerated approval pathway for AD treatment, meaning a post-approval trial has to demonstrate clinical benefit to keep the approval. Aducanumab, a human IgG1 anti-Aβ monoclonal antibody selective for Aβ aggregates, has become the first FDA-approved drug to reduce Aβ levels, a decision that was not devoid of criticism among the scientific community (Tampi et al., 2021). More recently, another accelerated approval has been granted by FDA to Lecanemab (Leqembi™) after the results of an 18-month, phase 3 trial that involved 1795 patients with early AD. This humanized IgG1 monoclonal antibody which selectively binds to large, soluble Aβ protofibrils has shown a reduction of Aβ in early AD patients and less decline in measures of cognition and function with modest effect sizes (van Dyck et al., 2022; Gandy and Ehrlich, 2023). Other Aβ-targeted monoclonal antibody therapies have failed to show positive clinical outcomes. Several reasons have been pointed out for such failure one being poor antibody brain penetration (< 0.1%) (van Dyck, 2018; Bajracharya et al., 2021; Nehra et al., 2022). In fact, it was only in the high dose intervention arm of the Aducanumab EMERGE trial that clinical benefit was found, suggesting the need of higher accumulated doses of the antibody to compensate for the low penetrance (Schneider, 2020; Leinenga et al., 2021). As pointed above, ultrasound has the potential to open BBB (Leinenga et al., 2016) constituting an option to alternatively enhance the clearance of amyloid by facilitating para and transcellular transport across this barrier (Pandit et al., 2020) and to facilitate drug delivery to the central nervous system. That is why, just recently, Leinenga et al. (2021) have investigated separately Aducanumab and scanning ultrasound in an animal model, showing that each one of these strategies has a comparable potential to reduce plaque burden. This suggests that the latter may be a treatment option in AD and hypothesizes that both should be studied in a combination trial as an approach to increase the brain levels of Aducanumab and to treat AD. As mentioned in a study cited above, magnetic resonance-guided focused ultrasound was able to open transiently and non-invasively the BBB in 5 patients with Aβ AD (Lipsman et al., 2018). A non-eloquent brain region was targeted and this technique was not complemented with any treatment which may soon be tested.

Other strategies to increase the delivery of drugs across BBB, particularly monoclonal antibodies, that are currently under development include bi-specific antibodies (a group that included the above-mentioned LRP-1 and other receptor-mediated transporters), nanoparticles (including liposomes, metallic nanoparticles and dendrimers), exosomes, viral vectors (Bajracharya et al., 2021)and drug re-engineering as fusion proteins that interact with BBB transcytosis systems, particularly transferrin receptor monoclonal antibodies (Pardridge, 2015). This latter strategy has been coined Brain Shuttle-mAb technology which has been shown to target Aβ in a mouse model of AD 55-fold compared to the parent antibody and significantly improves Aβ reduction (Niewoehner et al., 2014).

A pitfall associated with immunotherapy against Aβ in AD has been the adverse effects, particularly Aβ-related imaging abnormalities (ARIA) such as cerebral edema (ARIA-E) and microhemorrhages (ARIA-H). These indicate a state of increased barrier permeability and leakage in these patients (Nehra et al., 2022) also observed spontaneously in patients with cerebral amyloid angiopathy. This means that BBB permeability is a double-edged sword: on the one hand, it represents a barrier needed to be crossed over for efficacious delivery of immunotherapy and on the other hand, its dysfunction may precipitate the occurrence of adverse effects associated with such immunotherapy.

Research on disease modifying therapies in Alzheimer’s has been focused on Aβ, as it is the most common target of phase 2 and phase 3 clinical trials (Scheltens et al., 2021). However, anti-tau therapies have also been gaining momentum through phase 1 and 2 trials (Cummings et al., 2019). Similar to anti-amyloid antibodies, focused ultrasound has been used as an effective strategy to enhance the delivery of anti-tau antibodies in AD models (Jordão et al., 2010; Nisbet et al., 2017; Janowicz et al., 2019; Xhima et al., 2020).

Targeting the BBB may be a therapeutic strategy in itself. A leaky BBB is part of the pathophysiological conundrum of AD. Either as a consequence of inflammation and Aβ-induced cerebral amyloid angiopathy or part of the AD pathophysiological cascade that results in neurodegeneration and dementia (Noe et al., 2020). Therefore, drugs that repair BBB represent a window of opportunities. In pre-clinical AD models, thrombin inhibitor dabigatran (Cortes-Canteli et al., 2019), antiplatelet agents such as dipyridamole, cilostazol, tadalafil (Fisher et al., 2011; García-Barroso et al., 2013; Hattori et al., 2016) and recombinant activated protein C (Lazic et al., 2019) have shown barrier-sealing effects (Nehra et al., 2022). Another class of drugs that may prevent barrier dysfunction is angiotensin-II receptor blockers, particularly olmesartan. It has been demonstrated to act on tight junction mRNA levels and limit oxidative stress in mice (Nakagawa et al., 2017a,b).

Stem cell therapy has also the potential to target BBB. Stem cells can be differentiated into mesenchymal stem cells, neural stem cells, and induced pluripotent stem cells. All of these have been studied in AD with several purposes beyond BBB repair. The most used has been mesenchymal stem cells (Kim et al., 2020). To repair BBB, both mesenchymal stem cells and induced pluripotent stem cells both have the potential to differentiate into endothelial cells and regenerate blood vessels. In an APP/PS1 mouse model, bone marrow mesenchymal stem cells—VEGF treatment improved endothelial dysfunction, neovascularization and reduced senile plaques in the hippocampus with impact on cognitive dysfunction of AD transgenic animals (Garcia et al., 2014). Another lineage of stem cells that has been studied in BBB dysfunction related to AD is endothelial progenitor cells. They derive from the bone marrow and are thought to represent the main cell lineage involved in endothelial repair mechanisms (Custodia et al., 2021). In an APP/PS1 mouse model, endothelial progenitor cells transplantation has been shown to repair BBB tight junction function, increase microvessel density and decrease Aβ senile plaque deposition (Zhang et al., 2018).