Crossing the Blood-Brain Barrier: Integrating Scientific Innovation With Socio-Ethical Reflection in Predictive Medicine

September 1, 2007

In this article, we discuss recent advances in drug transporters and nutrient-transporter interactions that can impact drug bioavailability in the systemic circulation and the brain. We also present emerging research strategies that may facilitate the discovery and clinical development of predictive diagnostic tests to identify patients at risk for treatment resistance.

For psychotropic drugs to produce therapeutic effects, they need to reach molecular targets in the brain at an adequate concentration. However, drug levels in the plasma and the brain differ markedly from each other. This renders the outcome of psychiatric pharmacotherapy uncertain and highly variable from patient to patient on any given dose. A deeper understanding of the mechanisms underlying this disconnect is crucial for personalization of drug therapy and development of new treatment modalities for patients whose disorders are refractory to standard doses of drugs or drug combinations.

In this article, we discuss recent advances in drug transporters and nutrient-transporter interactions that can impact drug bioavailability in the systemic circulation and the brain.1-6 We also present emerging research strategies that may facilitate the discovery and clinical development of predictive diagnostic tests to identify patients at risk for treatment resistance.

As the focus of biomedicine and public health shifts toward the prediction of future disease susceptibilities and treatment outcomes, a new field of predictive (or preemptive) medicine is rapidly emerging with the availability of molecular diagnostic tests in the clinic.7-11 This also blurs the meanings and boundaries between health and disease as well as response and resistance to drug therapy.12 Hence, we suggest ways in which bioscience and bioethics research can be synchronized in predictive medicine.

P-Glycoprotein and drug efflux: a neglected mechanism for treatment resistance

The movement of intravascular compounds from the blood to the brain parenchyma is impeded by the blood-brain barrier (BBB), which is made up of endothelial cells, pericytes, the end-feet of astrocytes, and interendothelial tight junctions. In addition, those molecules that actually penetrate the BBB are subject to extraction from the brain back into the systemic circulation by drug transporters. From an evolutionary standpoint, these biological processes make sense: they ensure that the brain and other tissues essential for life and survival are clear of foreign chemicals and environmental toxins. But they al-so present a formidable challenge to achieving adequate and sustained drug exposure in the brain.

P-glycoprotein (P-gp) is a 170-kDa transmembrane protein and a member of the adenosine triphosphate-binding cassette superfamily of cell membrane transporters. Localized in specialized cells that serve as a barrier between tissue compartments (eg, in intestinal tissue, hepatocytes in the liver), P-gp serves as a functional complement to the physical BBB.2,3 P-gp is expressed as an efflux pump in a polarized manner on endothelial cells lining the cerebral microvasculature and actively removes drugs from the brain, thereby leading to a net reduction in drug availability in the brain tissue (Figure 1).2

A very active P-gp transporter could potentially result in diminished drug delivery to the brain and, by extension, can lead to treatment resistance even though peripheral drug concentrations may appear to be within a therapeutic range.2,3 The relevance of P-gp for psychotropic drug response is not merely theoretical but is supported by recent empirical research.13-15 For example, olanzapine, risperidone, and particularly 9-OH risperidone belong to the growing list of substrates for P-gp.14,15 After intraperitoneal injection of risperidone to abcb1ab-/- P-gp knockout mice, the brain:plasma concentration ratio for risperidone and 9-OH-risperidone was markedly elevated (compared with wild-type mice) in vivo, 12- and 29-fold, respectively.15 These preclinical observations suggest that P-gp may limit the availability of certain second generation antipsychotics in the brain with important ramifications for patients who do not respond within the conventional dosage range despite pharmacotherapy adherence.14,15

At the intestinal level, P-gp mediates the active transport of drugs back into the intestinal lumen, resulting in decreased oral bioavailability of drugs.4,16 In the liver, P-gp contributes to efflux of drugs into the bile, thereby eliminating drugs from the body. A broad range of drugs-HIV protease inhibitors, cardiac glycosides, immunosuppressant agents, antifungals, statins, dietary supplements used in treatment of mental illness (eg, St John's wort), and nutrients (eg, grapefruit juice)-can interact with P-gp by virtue of being a substrate, inhibitor, and/or inducer of this clinically important transporter.1,4-5,16-17 Thus, interindividual and intraindividual variations in P-gp activity, either constitutively or through interactions with drugs and bioactive food constituents, may affect the extent of drug absorption from the intestine and the penetration of psychotropic drugs into the brain.

Interestingly, the ultrarapid activity of other pharmacokinetic pathways, such as the drug metabolizing enzyme cytochrome P-450 2D6 (CYP2D6), is known to increase the risk of treatment resistance.18,19 By contrast, interindividual or population-to-population variability in drug transporter function has received relatively little attention to date in debates on refractoriness to psychiatric pharmacotherapy. The initial research interest in P-gp was due to its expression in cancer cells, removing antitumor drugs from the intracellular to the extracellular space, and thereby conferring a multi-drug resistance phenotype to cancerous tissue. Subsequently, it became clear that P-gp occurs physiologically in healthy tissues, which broadened P-gp's relevance for clinical pharmacology and therapeutics.

Along with academics, the pharmaceutical industry is investing in developing compounds that modulate P-gp function to target and optimize drug delivery to cancers or physiologically/ anatomically isolated tissue compartments. An increase in permeability of the BBB by inhibition of P-gp also offers the promise of administering lower oral doses that may help reduce peripheral psychotropic drug exposure and toxicity. The idea of using P-gp as a site of therapeutic intervention is not new, however, and has been proposed in the literature.13-15

OATP1A2 and grapefruit juice interactions

Although the recognition of P-gp mediated drug transport provides an insight into the disconnect between plasma and brain drug concentrations, there are several conceptual and practical challenges before tangible therapeutic applications in treatment refractory patients can be realized. These caveats include the following:

  • There are often multiple drug transporters present across biological barriers such as the BBB.
  • Variability in direction of transport (eg, drug influx or efflux) can result in functional antagonism or synergy when 2 or more transporters with similar substrate selectivity are expressed in the same tissue.
  • Broad selectivity in substrates and inhibitors may result in overlaps between transporters and with drug metabolism pathways (eg, CYP3A and P-gp markedly overlap in their substrates and inhibitors).
  • For many drug transporters, functional expression at the BBB and in different regions of the brain and spatial organization in endothelial cell membranes require further characterization.
  • Lack of selective substrates and inhibitors is a major barrier to our ability to characterize population level variability in transporter function in vivo and to design interventions that result in selective modulation of transporter activity without impact on overlapping pathways (eg, consider the previously noted example of CYP3A and P-gp).

These complexities represent only a limited portion of the future challenges in drug transporter research.1-4,16,17 Recent findings concerning the organic anion transporting polypeptide superfamily of transporters (OATPs) further exemplify some of these challenges.4 The OATPs transport a broad spectrum of substrates, including endogenous compounds (eg, thyroid hormones, bile acids), and various drugs (eg, digoxin, pravastatin, methotrexate, and fexofenadine). It has recently been found that several drug influx and efflux transporters are expressed in the duodenum.5

OATP1A2, formerly thought to be expressed mostly in the brain, is expressed appreciably in the intestine.5 OATP1A2 and P-gp proteins colocalize to the apical membrane of the intestinal villi acting in an opposite fashion, increasing and decreasing drug absorption, respectively (Figure 2) [Figure restricted. Please see print edition for content].5 In the liver, both of these transporters act in a synergistic manner in hepatocytes to facilitate drug elimination from the body (Figure 2) [Figure restricted. Please see print edition for content].4 Recent data indicate that OATP1A2 and P-gp contribute to drug absorption, distribution, and delivery through coordinated effects at multiple tissues. Thus, the clinical consequences of inhibition of P-gp activity at the BBB (eg, to augment drug availability in the brain) can be mitigated or accentuated depending on interindividual variability in OATP1A2 function and respective orientations (ie, as efflux or influx transporters) of P-gp, OATP1A2, and other yet unknown drug transporters in various regions of the brain.

A second set of recent findings concern the interaction of OATP1A2 and grapefruit juice.5 It has been observed that coadministration of grapefruit juice and fexofenadine (a substrate for both OATP1A2 and P-gp) results in a 52% decrease in fexofenadine plasma exposure (area under the curve [AUC]); this appears to be consistent with inhibition of OATP1A2-mediated drug uptake in the intestine.4 Grapefruit juice consumed 2 hours before administration of fexofenadine produced a lesser effect (38% decrease in AUC) whereas grapefruit juice given 4 hours before fexofenadine did not significantly impact fexofenadine exposure.

The list of psychotropic drugs transported by OATP1A2 is not yet firmly established, but as this information accumulates, it would be essential for clinicians to keep in mind OATP1A2- dependent drug uptake as another mechanism for grapefruit juice-drug interactions. It is not known if long-term ingestion of grapefruit juice and OATP inhibitors impacts drug transport at the BBB in a clinically significant manner. However, limiting grapefruit juice consumption at least 4 hours before the administration of OATP1A2 substrates with a narrow therapeutic window could serve as a clinical guideline to reduce unpredictability in intestinal drug absorption.

While therapeutic optimization of P-gp activity is an exciting area of investigation to improve drug delivery to the brain, we suggest that a singular focus on P-gp may result in unexpected clinical outcomes and even drug toxicity if contributions made by multiple drug transport pathways are neglect- ed. Another challenge is ascertaining whether sustainable or long-term P-gp inhibition can be achieved across the BBB. When P-gp downregulation is achieved successfully, increased functional permeability of the BBB can be problematic in the event of drug overdose and accidental exposure to environmental toxins.

Dissecting the overlap in drug transport and metabolism: pituitary as a reference point

Clinical availability of highly selective and potent P-gp inhibitors is an essential first step to test the therapeutic use and safety of pharmacological interventions that alter BBB permeability.20-22 Such progress demands substantial time and effort. In the meantime, there is a need to improve the use of existing nonselective P-gp inhibitors for potential clinical applications. A number of clinically available drugs, such as verapamil and quinidine, are P-gp inhibitors. However, these compounds have low potency and also concomitantly impact drug metabolism pathways such as CYP3A or CYP2D6.20-22 This poses a significant challenge from experimental design and data analysis standpoints. After administration of such nonselective P-gp inhibitors, how do we separate and explain the changes (if there are any) in drug response due to P-gp inhibition at the BBB and inhibition of drug metabolism and attendant increase in peripheral concentrations of P-gp substrate drugs?

Advances in neuroimaging techniques for measurement of receptor occupancy in brain regions outside the BBB (eg, pituitary) may provide a forward lead in this context. For example, the dopamine D2 receptor is expressed in both the basal ganglia and the pituitary. An increase in the ratio of basal ganglia/pituitary D2 receptor occupancy would lend evidence for an increase in the permeability of the BBB after administration of a P-gp inhibitor. By contrast, this ratio would be anticipated to remain stable or unchanged when a compound inhibits drug metabolism (increasing plasma drug concentrations and thereby brain receptor occupancy) without an appreciable impact on P-gp function at the BBB. The denominator of the basal ganglia/pituitary D2 receptor occupancy ratio thus could serve to control for changes in plasma drug concentrations after nonselective inhibition of both P-gp and CYP3A.

Although similar ratios are already widely used in preclinical research concerning drug delivery to the brain, measurement of drug occupancy in the pituitary and other regions outside the BBB has not been possible until recently. This technical limitation is being overcome because an increasing number of studies are now able to characterize pituitary receptor occupancy by positron emission tomography.23,24 The use of this composite metric/ratio as an experimental end point-instead of the absolute measures of drug exposure or receptor occupancy in the brain-may help dissect the overlapping effects of drug transporter inhibitors on P-gp and drug metabolism (eg, CYP3A).

Predictive medicine, translational research, and ethical considerations

As is evident with the case of drug transporters, an emerging picture in diagnostic and predictive medicine is that an ever-growing list of therapeutically significant molecular pathways is being uncovered. It is anticipated that these advances in basic research will lead to better patient care. However, there are important gaps in the translation of basic science discoveries to clinical practice.25,26 These gaps in translational research are collectively important enough that in the view of the FDA,25 "The applied sciences needed for medical product development have not kept pace with the tremendous advances in the basic sciences. The new science is not being used to guide the technology development process in the same way that it is accelerating the technology discovery process."

Translational research has the potential to deliver many practical benefits for patients and justify the extensive investments placed by the private and public sector in biomedical research. However, translational research encompasses a complexity of scientific, financial, ethical, regulatory, legislative, and practical hurdles that need to be addressed at several levels to make the process efficient.27 Some have resisted the idea of supporting translational research because of its high costs, particularly in the bench-to-bedside direction of drug development.

As Littman and colleagues27 have shown, the complexity of translational research in human subjects is overwhelming: production and validation of products of consistent safety, potency, and quality; the necessity to keep elaborate documentation of treatments to safeguard patient safety while protecting the right to privacy; the ancillary needs associated with the care of patients with severe conditions; and the cost of validating translatable biomarkers renders this discipline uniquely expensive. Therefore, advancing translational research requires new sources of funding and education. This could be achieved through public and congressional education by a joint coalition of patients' advocacy groups, academia, drug regulatory agencies, and industry.

Discoveries in biological psychiatry, and biosciences in general, occur in response to their social, economic, political, and cultural contexts.28-33 Recognition of technical and biological complexities is crucial but not sufficient for efficient translation of scientific advances to clinically relevant and equitable treatment guidelines.

An anticipated difference of the future diagnostic tests from routine clinical chemistry (eg, complete blood counts) is that they will have to characterize several biological mechanisms and pathways in order to achieve a clinically acceptable level of predictive accuracy. These predictive tests in pharmacotherapy will comprise a battery of diagnostics (eg, looking at drug metabolism, transporters, receptors, and so forth) and will create several plausible drug response and toxicity trajectories for each patient. In this regard, a particular ethical concern is the anticipated proliferation in diagnostic patents and its impact on equitable access to new technologies and diagnostic tests. If each segment of this expanding sphere of patentable biological pathways is held by different individuals, researchers, or commercial firms, scientific advances in therapeutics can be impeded while creating hypercompetition and excessive fragmentation of the knowledge space.

Furthermore, most physicians currently receive little training in genetics, genetic counseling, or predictive medicine. The advent of predictive tests in pharmacotherapy will require physicians to familiarize themselves with genetic or proteomic influences on drug response, understand the use of genetic tests, and incorporate the results of predictive tests into their clinical decisions. Physicians will also need to obtain informed consent for such testing from patients and provide necessary counseling about their results.

Another ethical concern relates to the disclosure of undesirable or adverse effects of the drug to patients. It has been recommended that disclosing adverse events to patients be required when the adverse event has a perceptible effect on the patient but was not discussed in advance as a known risk; when it necessitates a change in the patient's care; potentially poses an important risk to the patient's future health, even if that risk is extremely small; or involves providing a treatment or procedure without the patient's consent.34

From an ethical perspective, disclosure is required and should not be limited to cases in which the injury is obvious or severe. Disclosure of near misses is also discretionary but is advisable at times. In general, disclosure by a clinician involved in the patient's care is appropriate. Organizations should develop clear policies supporting disclosure and should create supportive environments that enable clinicians to meet their ethical obligations to disclose anticipated adverse drug-related events to patients and families. In fact, patients who do not respond well or suffer adverse reactions to drugs may not be adherent or their treatment outcome may be poor. This concern may be particularly important in relation to psychiatric disorders; patients may be confronted not only with the burden of stigma but also by being told that they cannot be effectively treated for it.

A major ethical issue is the extent to which individuals with severe psychiatric disorders are capable of providing informed consent, specifically whether they are able to understand the context and implications for new predictive tests or personalized treatments, what is exactly required of them, and why. A generally accepted model of practice is one that recognizes, caters to, and protects the individual's special needs and minimizes or eliminates any potential harm associated with the test or the treatment. Moreover, an ethical requirement is that the proposed test or treatment benefits the individual. In the interest of protecting patients, it has been suggested that a patient's guardian or substitute decision maker should not be approached out of convenience if the patient is likely to be competent to give consent once symptoms of their condition have been treated.35 In the event that substituted consent is necessary, the patient's preference should be adhered to in selecting a proxy decision maker.

Patient privacy is another issue raised by the application of predictive testing for treatment response to psychotropic drugs. If appropriate safeguards are not in place to ensure patient privacy and confidentiality, patients may be subjected to stigma and discrimination, and may not accept the use of predictive testing in routine clinical practice. In some cases, a genetic variant that determines an individual's likelihood to respond to a drug may also indicate susceptibility to a disease that the drug is used to treat.36 Under such circumstances, predictive testing for drug efficacy and safety should be treated with the same level of confidentiality and privacy as disease susceptibility testing or carrier testing results.

Finally, despite the projected clinical benefits of using genetic or other predictive tests for analyzing the response to psychotropic drugs, the results of such tests could determine eligibility criteria for insurance policies. Some patients may be viewed as large liabilities by insurance companies because their genotype indicates they will require the most expensive treatments for certain conditions, rather than because they have a high likelihood of falling ill. These concerns may not be as important if the tests indicate only the risk of an individual suffering an adverse reaction or of experiencing no therapeutic benefit. Hence, we suggest that awareness of biological complexities concerning the drug transporters as well as recognition of the socio-ethical context in which science, society, and clinical practice interact are essential to develop equitable diagnostic predictive tests for people who do not respond to conventional psychiatric therapy.




Eichelbaum M, Fromm MF, Schwab M. Clinical aspects of the MDR1 (ABCB1) gene polymorphism.

Ther Drug Monit.



Doran A, Obach RS, Smith BJ, et al. The impact of P-glycoprotein on the disposition of drugs targeted for indications of the central nervous system: evaluation using the MDR1A/1B knockout mouse model.

Drug Metab Dispos.



Lee G, Bendayan R. Functional expression and localization of P-glycoprotein in the central nervous system: relevance to the pathogenesis and treatment of neurological disorders.

Pharm Res.



Kirby BJ, Unadkat JD. Grapefruit juice, a glass full of drug interactions?

Clin Pharmacol Ther.



Glaeser H, Bailey DG, Dresser GK, et al. Intestinal drug transporter expression and the impact of grapefruit juice in humans.

Clin Pharmacol Ther.



Preskorn SH. Pharmacogenomics, informatics, and individual drug therapy in psychiatry: past, present and future.

J Psychopharmacol.



Ozdemir V, Godard B. Evidence based management of nutrigenomics expectations and ELSIs.


In press.


Godard B, Marshall J, Laberge C, Knoppers BM. Strategies for consulting with the community: the cases of four large-scale genetic databases.

Sci Eng Ethics.

2004;10: 457-477.


Rose N.

The Politics of Life Itself: Biomedicine, Power, and Subjectivity in the Twenty-First Century.

Princeton, NJ: Princeton University Press; 2006.


Williams-Jones B, Ozdemir V. Pharmacogenomic promises: reflections on semantics, genohype and global justice. In: Einsiedel E, ed.

Emerging Technologies: From Hindsight to Foresight.

Calgary, AB, Canada: University of Calgary Press. In press.


Rose N. Genomics. In: Kuper A, Kuper J, eds.

The Social Science Encyclopedia

. 3rd ed. London, UK: Routledge; 2004.


Graham J, Ritchie K. Mild cognitive impairment: ethical considerations for nosological flexibility in human kinds.

Philos Psychol Psychiatry.



Sadeque AJ, Wandel C, He H, et al. Increased drug delivery to the brain by P-glycoprotein inhibition.

Clin Pharmacol Ther.



Boulton DW, DeVane CL, Liston HL, Markowitz JS. In vitro P-glycoprotein affinity for atypical and conventional antipsychotics.

Life Sci.



Wang JS, Ruan Y, Taylor RM, et al. The brain entry of risperidone and 9-hydroxyrisperidone is greatly limited by P-glycoprotein.

Int J Neuropsychopharmacol.

2004; 7:415-419.


Urquhart BL, Tirona RG, Kim RB. Nuclear receptors and the regulation of drug-metabolizing enzymes and drug transporters: implications for interindividual variability in response to drugs.

J Clin Pharmacol.

2007; 47:566-578.


Pal D, Mitra AK. MDR- and CYP3A4-mediated drug-herbal interactions.

Life Sci.



Bertilsson L, Aberg-Wistedt A, Gustafsson LL, Nordin C.Extremely rapid hydroxylation ofdebrisoquine: a case report with implication for treatment with nortriptyline and other tricyclicantidepressants.

Ther Drug Monit.



Kirchheiner J, Fuhr U, Brockmoller J. Pharmacogenetics-based therapeutic recommendations--ready for clinical practice?

Nat Rev Drug Discov.



Achira M, Suzuki H, Ito K, Sugiyama Y. Comparative studies to determine the selective inhibitors for P-glycoprotein and cytochrome P4503A4.

AAPS PharmSci

. 1999;1:E18.


Kim RB, Wandel C, Leake B, et al. Interrelationship between substrates and inhibitors of human CYP3A and P-glycoprotein.

Pharm Res.



Wandel C, Kim RB, Kajiji S, et al. P-glycoprotein and cytochrome P-450 3A inhibition: dissociation of inhibitory potencies.

Cancer Res.



Muhr C. Positron emission tomography in acromegaly and other pituitary adenoma patients.




Moresco RM, Cavallaro R, Messa C, et al. Cerebral D2 and 5-HT2 receptor occupancy in schizophrenic patients treated with olanzapine or clozapine.

J Psychopharmacol

. 2004;18:355-365.


US Department of Health and Human Services. Food and Drug Administration.

Innovation or Stagnation. Challenge and Opportunity on the Critical Path to New Medical Products.

Rockville, MD: US Food and Drug Administration; March 2004.


Ozdemir V, Williams-Jones B, Cooper DM, et al. Mapping translational research in personalized therapeutics: from molecular markers to health policy.




Littman BH, Di Mario L, Plebani M, Marincola FM. What's next in translational medicine?

Clin Sci (Lond).



Hopkins MM, Ibarreta D, Gaisser S, et al. Putting pharmacogenetics into practice.

Nat Biotechnol.

2006; 24:403-410.


Atkinson-Grosjean J.

Public Science, Private Interests: Culture and Commerce in Canada's Networks of Centres of Excellence

. Toronto, ON, Canada: University of Toronto Press; 2006.


Williams-Jones B. Knowledge commons or economic engine--what's a university for?

J Med Ethics.

2005; 31:249-250.


Graham J. Smart regulation: will the government's strategy work?




Hedgecoe A.

The Politics of Personalised Medicine: Pharmacogenetics in the Clinic.

Cambridge, UK: Cambridge University Press; 2004.


Etzkowitz H, Webster A, Gebhardt C, Cantisano-Terra BR. The future of the university and the university of the future: evolution of ivory tower to entrepreneurial paradigm.

Res Policy.



Cantor MD, Barach P, Derse A, et al. Disclosing adverse events to patients.

Jt Comm J Qual Patient Saf.



Parker LS. Ethical issues in bipolar disorders pedigree research: privacy concerns, informed consent, and grounds for waiver.

Bipolar Disord.



Hedgecoe A. Education, ethics and knowledge deficits in clinical pharmacogenetics.