INTRODUCTION
P-glycoproteins belong to a family of plasma membrane proteins encoded by the MDR (multidrug resistance) gene(s) that are well conserved in nature. P-glycoprotein (P-gp) functions as a membrane-localized drug transport mechanism that has the ability to actively pump out all currently prescribed HIV-protease inhibitors (PIs) from the intracellular cytoplasm. This effect may result in limited oral bioavailability of PIs as well as a decreased ability of the drugs to cross blood-tissue barriers such as the blood-brain barrier (BBB), the blood-testis barrier (BTB) and the materno-fetal barrier (MFB). MDR1 encoded P-glycoprotein has also been implicated in the cytotoxicity process and with the induction of immune responses during HIV infection. It also plays a role in oxidative and inflammatory processes and it may be involved in lipid transport and metabolism.

WHAT IS P-GLYCOPROTEIN?
P-gp, a 170-kDa phosphorylated and glycosylated plasma membrane protein belonging to the ATP-binding cassette superfamily of transport proteins, was first described by Juliano and Ling in 1976. (Juliano 1976) Humans have two MDR genes, MRD1 and MDR3 (also called MDR2), and rodents have three B mdr1a (also called mdr3), mdr1b (also called mdr1), and mdr2. Only the MDR1 gene (responsible for P-gp synthesis) in humans and the mdr1b and mdr1a genes in rodents appear to be involved in drug transport and the development of drug resistance. (Fardel 1996 ­ Gen Pharmacol) The human MDR3 and murine mdr2 genes encode a P-glycoprotein which does not seem to have a role in drug transport, but has a role in the secretion of phosphatidylcholine into bile and is probably a phospholipid transport protein or flippase. (Smit 1993 and 1994).

MDR1 P-gp (referred to simply as P-gp in this article) is a transmembrane protein which is 1280 amino acids long and consists of two homologous halves of 610 amino acids joined by a flexible linker region of 60 amino acids. Each half has an N-terminal hydrophobic domain containing six transmembrane domains followed by a hydrophilic domain containing a nucleotide binding site. The nucleotide binding sites can bind ATP and its analogues and both sites are essential since inactivation of either site inhibits substrate-stimulated ATPase activity. However, the two sites are likely to be functionally independent and cleavage probably only occurs at one site at a time. (Loo 1999 ­ J Biol Chem and Hrycyna 1998) Each nucleotide binding siteconsists of three conserved regions called Walker A, B and C motifs. (Gottesman 1996)

When viewed from above the plasma membrane, P-gp is donut shaped with 6-fold symmetry, a diameter of about 10nm and a large central pore of about 5nm in diameter. It has a thickness in the plane of the plasma membrane of about 8nm. Since the depth of the plasma membrane lipid bilayer is about 4nm, about half of the molecule is within the plasma membrane. The central pore is closed at the inner (cytoplasmic) end, forming an aqueous chamber within the membrane that is open to the extracellular medium. There is an opening from this chamber to the lipid phase. Two lobes of about 3nm each are exposed at the cytoplasmic end and probably correspond to the nucleotide binding domains. (Rosenberg 1997)

MECHANISM OF ACTION, P-GP SUBSTRATES AND INHIBITORS [TABLE 1]
The majority of published data suggest that P-gp acts as a transmembrane pump which removes drugs from the cell membrane and cytoplasm. ATP hydrolysis provides the energy for active drug transport, which can occur against steep concentration gradients. It was initially hypothesized that P-gp forms a hydrophilic pathway and drugs that are transported from the cytoplasm to the extracellular medium through the central pore, thereby shielding the substrate from the hydrophobic lipid phase. It has been further proposed that P-gp acts like a hydrophobic vacuum cleaner or flippase. In this model P-gp intercepts the drug as it moves through the lipid membrane and flips the drug from the inner leaflet of the plasma membrane lipid bilayer to the outer leaflet and into the extracellular medium. (Gottesman 1996, Higgins 1992 and Johnstone 2000)

Some studies support a model of P-gp in which there is a region or multiple regions of interaction rather than one or two simple binding sites. Molecules interacting with P-gp may be classified as substrate or antagonist. Compounds in the substrate group are characterized by a >4-fold increase in cytotoxicity in MDR cells. Compounds in the antagonist group increase the intracellular accumulation of the P-gp substrates, rhodamine-123 and vinblastine, and display <4-fold reversal of cytotoxicity by the anti-cancer drug, cyclosporin A. Antagonists may bind P-gp more tightly and, failing to be transported, prevent the transport of other compounds, while substrates, in being transported, do not block the transport of other substrates. (Scala 1997)

P-gp substrates ­ chemicals transported by P-gp ­ have very diverse structures, which only share the properties of being hydrophobic amphipathic molecules (molecules having two sides with characteristically different properties) that are not negatively charged, and may be 200-1800 Da in size. They include cancer drugs, such as doxorubicin, daunorubicin, vinblastine, vincristine, actinomycin D, paclitaxel, teniposide and etoposide; immunosuppressive drugs, including cyclosporin A and FK506; steroids like aldosterone, hydrocortisone, cortisol, corticosterone and dexamethasone; HIV PIs, such as amprenavir (APV), indinavir (IDV), nelfinavir (NFV), ritonavir (RTV) and saquinavir (SQV); the antihistamine terfenadine; cardiac drugs, such as digoxin and quinidine; the lipid lowering agent lovastatin; the dopamine antagonist domperidone; the antiemetic ondansetron; the anti-diarrheal agent loperamide; the anti-gout agent colchicine; the antibiotic erythromycin; the anti-tuberculous agent rifampicin; the anti-helminthic agent ivermectin; and the fluorescent dye rhodamine-123. P-gp activity decreases the intracellular concentration of cancer drugs, enabling resistance to develop to them; the same may be true for PIs. Under certain circumstances, P-gp may be able to transport hydrophilic negatively charged compounds, such as methotrexate. (Gottesman 1996, Huisman 2000 ­ AIDS, R. Kim 1999, Mayer 1997, Schinkel 1996 and 1997, and Schuetz 1996 and Yumoto 1999)

P-gp inhibitors include the immunosuppressant cyclosporin A and its non-immunosuppressive analogue PSC833 (valspodar); the calcium channel blocker verapamil; the PIs RTV, SQV, NFV and possibly IDV; the progesterone antagonist mifepristone (RU486); the antiarrhythmic agent quinidine; the sedative midazolam; the anti-estrogen tamoxifen; the antibiotic erythromycin; the antifungal ketoconazole; the peptide chemosensitizers reversins 121 and 205; the acridonecarboxamide derivative GF120918 (or GG918); and the cyclopropyldibenzosuberane LY335979. (Drach 1996, Fardel 1996 ­ Anticancer Drugs, Huisman 2000 ­ AIDS, Mayer 1997, Sharom 1999, Twentyman 1991 and Yumoto 1999)

There have been various attempts to classify compounds based on their effect on or interaction with P-gp. A number of chemicals, including anticancer drugs, have been categorized based on their effect on ATPase activity of human P-gp. Class I compounds in low concentrations stimulate ATPase activity and in high concentrations inhibit it. Kinetic analyses show they have high affinity for the active site and low affinity for the inhibitory site. They include vinblastine, verapamil and taxol. Class II compounds stimulate ATPase activity in a dose-dependent manner without any inhibition and interact only with the active site. They include bisantrene, valinomycin and diltiazem. Class III compounds, which bind to the inhibitory site with high affinity, inhibit both basal and verapamil-stimulated ATPase activity. They include cyclosporin A, rapamycin and gramicidin D. (Gottesman 1996) [TABLE 2]

Studies by Shapiro and colleagues have suggested the existence of three drug binding sites of P-gp. Two seem to be involved in drug transport, and substrate binding to one stimulates transport by the other. The H site is selective for Hoechst 33342 and colchicine and the R site is selective for rhodamine 123 and anthracyclines. The third site is likely not to be capable of drug transport. (Shapiro 1999)

P-gp-dependent drug transport activity depends on the level of expression of the MDR1 gene as well as on the functionality of the MDR1-encoded P-gp. Shapiro and colleagues also showed that P-gp-mediated drug transport was stimulated by the antihypertensive prazosin and the hormone progesterone. (Shapiro 1999) Rifampicin has been shown to induce MDR1 expression. Induction of intestinal P-gp by rifampicin has been shown to be the major mechanism responsible for reduced digoxin levels during concomitant rifampicin therapy. In healthy male volunteers, the oral bioavailability of digoxin decreased by 30% and intestinal P-gp levels were induced 3.5 fold during rifampin therapy. (Fromm 2000 and Greiner 1999)

It has been demonstrated that one possible mechanism of action for P-gp-mediated resistance to chemotherapeutic agents has been through gene rearrangement. Mickley et al found that in several drug-resistant cancer cell lines as well as samples from two leukemic patients who had developed drug resistance, gene rearrangements had occurred resulting in initial activation or increased expression of MDR1. (Mickley 1999)

Gene amplifications, rifampicin induction and probably other factors cause MDR-1 over-expression. Polymorphism in exon 26 (C3435T) of MDR1 is significantly correlated with levels of expression and function of MDR1. Individuals homozygous for this polymorphism (T/T allele) showed significantly lower duodenal MDR1 expression and higher digoxin plasma levels than volunteers with the C/C genotype. Evaluation of maximum plasma concentrations (Cmax) during steady state conditions of digoxin administration showed a statistically significant (p=0.006) mean difference of 38% in digoxin Cmax in the homozygous T/T genotype compared with the C/C genotype. (Hoffmeyer 2000)

P-gp is unlikely to be functional immediately after synthesis. Loo and Clarke showed that immune, core-glycosylated P-gp could be prevented from reaching the cell surface during biosynthesis by processing mutations or by proteasome inhibitors such as lactacystin or MG-132. This immature P-gp exhibited no detectable drug-stimulated ATPase activity, did not exhibit ATP-induced conformational changes as found in the mature transporter, and was more sensitive to the proteolytic digestive enzyme, trypsin, than the mature form. (Loo 1999 ­ FASEBJ)

WHERE IS P-GP FOUND? [TABLE 3]
Monoclonal antibody MRK16 was used to localize P-gp in normal human tissues. Most tissues examined revealed very little P-gp. However, P-gp was found in:

• liver in hepatocytes on the biliary canalicular front and on the apical (luminal or central canal) surface of epithelial cells in small biliary ductules;
• pancreas on the apical surface of the epithelial cells of small ductules, but not larger pancreatic ducts;
• kidney on the apical surface of epithelial cells of the proximal tubules;
• colon and jejunum in high concentrations on the apical surfaces of superficial columnar epithelial cells; and in the
• adrenal gland in high levels on the surface of cells in both the medulla and the cortex.
  (Thiebaut 1987)

In the small intestines of rats and mice, regional variation in P-gp expression has beendemonstrated, with moderate P-gp expression found in the duodenum and jejunum and maximal expression in the ileum. (Yumoto 1999)

P-gp is also found in the sub-apical surface of the epithelium of the choroid plexus of the brain (which forms the blood-cerebrospinal fluid (CSF) barrier) as well as the luminal surface of the endothelium of blood capillaries of the brain (the blood-brain barrier). (Wijnholds 2000 and Rao 1999)

In mice, mdr mRNA expression levels increase dramatically during pregnancy and are expressed at extremely high levels in the gravid compared with the non-gravid uterus. This expression is specifically localized to the luminal surface of the secretory epithelial cells of the endometrium. (Arceci 1998) Murine placental P-gp is present in the fetus-derived epithelial cells that make up the exchange border between the fetal and maternal blood circulation; the fetally derived P-gp faces the maternal blood side. (Smit 1999) P-gp is also expressed in the testis and ovaries of mice and in the steroid-producing endometrial glands of the pregnant uterus. (Schinkel 1997) P-gp expression was detected in the capillary endothelial cells of the 7th and 8th peripheralnerves of the guinea pig by immunohistochemical staining and Western bolt analysis. Levels of immunoreactivity were similar to those in the brain. Somewhat lower levels of immunoreactivity were also found in the sciatic nerve. (Saito 1997)

P-gp has been found in normal bone marrow in hematopoietic stem cells and in peripheral blood mononuclear cells (PBMCs), mature macrophages, natural killer (NK) cells, antigen-presenting dendritic cells (DCs) and T- and B-lymphocytes. (Gottesman 1996 and Johnstone 2000) P-gp and MDR1 are expressed to different levels in normal leukocytes. Klimecki and colleagues demonstrated relatively high levels in CD56+ cells (NK cells), high to moderate levels in CD4+, CD8+ and CD15+ cells (T-helper cells, T-suppressor cells and granulocytes, respectively) and lower levels in CD19+ and CD14+ cells (B-lymphocytes and monocytes, respectively). (Klimecki 1994)

WHAT IS THE FUNCTION OF P-GP?
The normal physiological function of P-gp in the absence of therapeutics or toxins is unclear. Studies of mdr1 knock out (KO) mice (mice lacking the mdr1 genes) show that they have normal viability, fertility and a range of biochemical and immunological parameters. Predictably, they do have delayed kinetics and clearance of vinblastine and they accumulate high levels of certain drugs (vinblastine, ivermectin, cyclosporin A, dexamethasone, ondansetron, loperamide and digoxin) in their brains. These mice also demonstrated marked increases in the levels of these drugs in the testis, ovary and adrenal gland compared with wild type mice. (Schinkel 1996 and 1997) It has been reported that some mdr1a KO mice develop a severe, spontaneous intestinal inflammation similar to human inflammatory bowel disease. However, this has not been observed by other researchers. (Schinkel 1997 and Wijnholds 2000)

1. Protection from Drugs and Toxins
Expression of P-gp on the luminal surfaces of the epithelial cells of the small and large intestine, biliary ductules, and proximal tubules of the kidney suggest a role in decreasing the absorption from the gut and/or the excretion of endogenous and exogenous hydrophobic amphipathic toxins. P-gp plays a role in the intestinal excretion and, hence, in the reduced bioavailability after oral ingestion of several drugs, including digoxin, paclitaxel, and HIV PIs. (Schinkel 1997) The plasma concentrations after oral administration of IDV, NFV and SQV were 2-5-fold higher in mdr-1a KO mice compared with wild-type (WT) or normal mice. (R. Kim 1998)

Expression in the capillary endothelial cells of the brain, nerves, testis and placenta suggest a barrier function to keep toxins out of the nervous system, gonads and fetus. Many relatively hydrophobic drugs that were expected to diffuse easily across lipid membranes did not readily enter the brain. P-gp has been shown to prevent or decrease the entry of certain drugs into the brain and to contribute to the BBB. These drugs include ivermectin, vinblastine, digoxin, cyclosporin A, loperamide, domperidone, ondansetron, and HIV PIs. (Gottesman 1996, Fromm 2000, Saito 1997, Schinkel 1996 and 1997) Compared with WT mice, mdr KO mice show substantially higher levels of the P-gp substrates B digoxin, ondansetron, and loperamide B in the brain (27-, 4- and 14-fold, respectively), adrenal (6-fold for digoxin), ovary (8-fold for digoxin) and testis (2- and 4-fold respectively for ondansetron and loperamide, respectively). (Schinkel 1996 and 1997)

Paradoxically, apical localization of P-gp in choroid plexus epithelial cells seems to prevent trafficking of certain substrates out of the CSF and opposes the action of P-gp at the BBB. The function of P-gp at this location is yet to be determined, but one possibility is that it functions to secrete natural substances into the CSF or to enhance cholesterol uptake and esterification within the choroid plexus epithelia. To complicate the matter even further, multidrug resistance-associated protein (MRP), another ATP-binding cassette transporter localizes basolaterally in choroid plexus epithelium and opposes the action of P-gp at this site. (Rao 1999)

In very elegant studies of pregnant heterozygous mdr mice with fetuses of three genotypes (mdr WT, homozygous mdr KO and heterozygous mdr mice), Smit and colleagues demonstrated that P-gp activity reduced the placental transfer of the P-gp substrate drugs digoxin, SQV and paclitaxel. They further demonstrated that this could be abolished by orally administered P-gp blockers PSC833 or GG918. They note that P-gp can be detected in human placental trophoblasts from the first trimester of pregnancy to full term, making it very likely that placental P-gp protects the developing embryo and fetus from toxic insult in humans as well. (Smit 1999)

P-gp has been found in hematopoietic stem cells and probably contributes significantly to the removal of drugs and toxins from the bone marrow. Pleuripotent hematopoietic stem cells from P-gp KO mice demonstrated markedly decreased rates of rhodamine 123 efflux compared with WT cells and confirming the substantial contribution of P-gp to the extrusion of drugs from the bone marrow. (Schinkel 1997)

2. Steroid Metabolism
The presence of P-gp in the adrenal and in steroid-producing cells of the endometriumsuggest it may also have a role in the handling of steroids, possibly providing a protective function for the plasma membranes of steroid-producing cells. Furthermore, it has been found that P-gp expressing epithelial monolayers of cells are able to transport steroids and that some lymphoid cells expressing P-gp are resistant to the cytotoxic effects of steroids. Cultured mouse adrenal Y-1 adrenal cells in which a P-gp allele has been insertionally mutagenized show decreased secretion of steroids. Yeast MDR1 homologue PDR5 (an ABC transporter with broad substrate specificity similar to MDR1) prevents uptake of steroids by yeast. (Gottesman 1996)

3. Cholesterol Metabolism
P-gp appears to have a role in cholesterol metabolism. Cholesterol esterification is one of the mechanisms that cells use to control the amount of free cholesterol, which is toxic to cells. Under conditions of excess cholesterol, cholesterol is transported from the plasma membrane, where 90% of it resides, to the endoplasmic reticulum (ER), where it is esterified by acyl-CoA:cholesterol acyl transferase (ACAT). The rate of cholesterol esterification is limited by the availability of cholesterol substrate in the ER, rather than by ACAT. Within a given cell type, greater expression of P-gp was correlated with increased esterification of plasma membrane cholesterol. P-gp functions to increase esterification of cholesterol derived from plasma membrane by facilitating the movement of cholesterol from the plasma membrane to the ER. The exact mechanism by which P-gp does this in not known, but expression of P-gp does not seem to confer any change in cholesterol content of plasma membrane as examined with cholesterol oxidase. (Debry 1997 and Luker 1999)

The levels of cholesterol ester increase with age in arteries prone to atherosclerosis and become predominant in advanced atherosclerotic lesions. The mRNA levels of ACAT and MDR1 showed the same correlation with age and reached the highest levels in atherosclerotic specimens, suggesting that P-gp may be involved in the accumulation of intracellular cholesterol ester in atherosclerotic lesions. (Batetta 1999 ­ J Vasc Res)

Batetta and colleagues have suggested that cholesterol esters may be involved in the regulation of cell growth and division and that MDR1 gene expression may contribute to this regulation by modulating intracellular cholesterol ester levels. (Batetta 1999 ­ Cell Prolif)

4. Immune System
The role of P-gp in the hematopoietic and immune systems needs to be elucidated. Does it protect stem cells against toxic endogenous compounds and xenobiotics? Is it important for T cell function? Is hematopoietic development or function dependent on P-gp? P-gp expression in leukocytes varies depending on the lineage and is not associated with transport of the P-gp substrate, rhodamine 123, out of the cell. (Gottesman 1996 and Klimecki 1994)

Immune responses in a peripheral organ like skin are initiated when antigen-presenting cells, especially dendritic cells (DCs), capture antigens locally. The DCs then migrate via lymphatic vessels to draining lymph nodes where they select T lymphocytes that bear receptors for the presented antigen. In vitro models have demonstrated that P-gp facilitates this migration of DCs and that in the presence of P-gp antagonists, DCs are retained in the epidermis. (Randolph 1998)

There is also evidence that P-gp may be involved in the transport of some cytokines (CKs), particularly interleukin-1 (IL-1), IL-2, IL-4 and interferon-gamma (IFN-y) out of activated normal lymphocytes. However, P-gp does not seem to transport IL-6. The biological importance of P-gp to CK secretion during an immune response is still to be clarified. (Drach 1996 and Johnstone 2000)

5. Cell Death and Cell Differentiation
The high level of P-gp expression on NK and CD8+ T cells has raised the question ofthe role of P-gp in the function of these cells. Inhibition of P-gp results in a reduction in NK and CD8+ T cell cytolytic activity. (Johnstone 2000)

The key cell death proteins are the cystein-aspases (caspases), which are initially expressed as inactive zymogens that are activated by proteolytic processing to produce the active enzyme. Cell death initiated by a range of stimuli ­ including FasL, tumor necrosis factor (TNF), ultraviolet radiation, granzyme B (GzB) and chemotherapeutic drugs ­ function by activating the caspase pathway. The majority of physiological cell death pathways appear to involve caspases.

Cell death due to membrane and cytosolic perturbations by cytotoxic granules occurs in the absence of activation of the caspase pathway, whereas nuclear damage requires caspase activation. NK and cytotoxic CD8+ T cells bind their target cells and induce death either via the Fas/Fas ligand system or, usually, by release of cytotoxic granules, such as perforin and granzymes, into the target cell.

In tumor cells, P-gp confers resistance to Fas-mediated apoptosis. It has been demonstrated that tumor cells expressing P-gp are resistant to a wide range of stimuli that activate the caspase apoptotic cascade, but are not resistant to caspase-independent cell death mediated by pore-forming proteins and GzB. Inhibition of P-gp completely reverses this resistance to caspase-dependent cell death. (Johnstone 1999 and 2000, and Smyth 1998)

Murine stem cells transduced with an MDR1 vector expanded in vitro, while mock-transduced cells did not. The mechanism by which MDR1 facilitates stem cell expansion in culture is unknown. It has been postulated that this may occur through an anti-apoptotic effect of P-gp, or through prevention of stem cell differentiation. However, mice transplanted with the expanded MDR1-transduced stem cells, but not those with unexpanded MDR1-transduced cells, developed a myeloproliferative disorder with very high peripheral white blood cell counts and splenomegaly. It thus appears that the myeloproliferative syndrome is the result of ex vivo expansion, rather than the MDR1 transduction per se, but this remains to be confirmed. (Bunting 1998)

6. Chloride channel
P-gp does not seem to have intrinsic channel activity, but may regulate an endogenous chloride channel which is yet to be identified. The physiological importance of this indirect modulation of chloride channels remains controversial. (Gottesman 1996 and Johnstone 2000)

7. Cytochromes
Many of the same drugs that are transported by P-gp are metabolized by some of the cytochromes (CYPs), especially CYP450 3A which constitutes 30% and 70% of the total CYP450s in most human livers and intestines, respectively. CYP3A substrates include HIV PIs, the antibiotic erythromycin, the lipid-lowering agent lovastatin; the calcium channel blocker nifedipine; rifampicin; and the sedative midazolam. CYP3A can be induced by many agents, including glucorticoids, barbiturates and rifampicin. Human variation in CYP3A expression is believed to influence drug response for up to one-third of all drugs.

P-gp has a role in modulating expression of CYP3A. Schuetz and colleagues demonstrated that, by regulating the intracellular concentration of rifampicin, P-gp modulated the extent to which rifampicin was able to induce CYP3A. It is likely that because P-gp can influence the intracellular concentration of many CYP3A substrates, it may also affect the availability of those substrates to CYP3A and so the extent of CYP3A metabolism of those substrates. P-gp thus plays an important role in modulating expression of CYP3A and this is likely to complicate the prediction of drug interactions among drugs that are substrates for both P-gp and CYP3A systems. (Schuetz 1996)

In mice, the amount of CYP3A protein in liver is inversely correlated with the gene dose of the normal mdr1a allele. An increased expression of CYP2B and CYP3A was seen in mdr1a KO mice housed in Amsterdam. This was in contrast to the genetically identical mice housed in the United States which showed no correlation between CYP3A activity and genotype. Measurement of CYP3A catalytic activity in the Amsterdam mice revealed the following rank order of activities: mdr1a/1b KO > mdr1a KO > mdr1b KO > WT. This suggests that MDR1 may regulate the expression of liver cytochromes. (Schuetz 2000)

P-GP AND HIV
All HIV PIs currently in use (IDV, SQV, NFV, RTV and APV) are transported by P-gp, which can actively expel these PIs from cells. Transport of HIV PIs can be inhibited by P-gp inhibitors like cyclosporin A, quinidine, verapamil, and PSC833. (Clayette 2000, A. Kim 1998, R. Kim 1998, Lee 1998, Profit 1999 and van der Sandt 2000) Srinivas et al have shown by calcein flux assays that PIs interact with P-gp with affinities in the order RTV>NFV>IDV>SQV. (Srinivas 1998) RTV, SQV, NFV and possibly IDV have also been shown to inhibit transport of some of the known P-gp substrates. Except for RTV and maybe SQV, their effects were weaker than established inhibitors like verapamil or cyclosporin A. (Garraffo 2000, Gutmann 1999, Profit 1999 and Washington 1998)

Kim et al demonstrated that plasma levels of IDV, SQV and NFV were 2-5 times higher in mdr1a KO mice compared with WT mice. This strongly suggests that P-gp transport at the intestinal and/or hepatic level limits the systemic bioavailability of these drugs. Studies in Caco-2 cells, which exhibit many of the morphological and biochemical characteristics of human small intestine, suggest that P-gp transports absorbed PIs (APV, RTV, IDV, NFV and SQV) back into the intestinal lumen, thus limiting oral bioavailability. (R. Kim 1998 and van der Sandt 2000)

Jones et al demonstrated that the rank order of in vitro intracellular accumulation of PIs was SQV>RTV>IDV. P-gp, MRP (multidrug resistance protein, another drug transporter), protein binding and HIV infection all decreased the intracellular accumulation of PIs. (Jones 2000) Compared with human erythroleukemia cells that don t express P-gp, those over-expressing P-gp demonstrated a 10-fold reduction in APV and IDV intracellular concentrations, 3- and 6-fold reductions for RTV and SQV, respectively, while there was no difference in NFV intracellular concentrations. (Garraffo 2000)

P-gp may also limit the penetration of PIs into several tissue compartments in the body, thereby possibly creating sanctuary sites, such as the brain and gonads. (Huisman 2000 ­ 1st Intl Wkshp on Clin Pharm of HIV Therapy) Brainpenetration of IDV, SQV and NFV were increased 7-, 10- and 36-fold respectively in mdr1a KO mice compared with WT mice. (R. Kim 1998) Using an in vitro BBB model, van der Sandt and colleagues demonstrated that APV, RTV and IDV are actively transported by P-gp across the BBB. (van der Sandt 2000) Choo et al demonstrated that brain and testis levels of NFV, APV, IDV and SQV were significantly increased in mice when the potent P-gp inhibitor LY335797 was administered intravenously and that this increase was not due to increased PI plasma levels. (Choo 2000) Taylor and colleagues found that while IDV achieves good penetration into the semen (seminal plasma (SP):blood plasma (BP) = 0.9), RTV and SQV penetrated very poorly (SP:BP were 0.02 and 0.03, respectively). (Taylor 2000)

The effect of P-gp on limiting oral bioavailability and tissue distribution of PIs has obvious implications for the effectiveness of PI-containing regimens. Poor penetration of PIs into the brain, testis and other sanctuary sites may result in de facto compartmental mono- or dual antiretroviral therapy with ongoing HIV replication and development of resistance.

HIV PIs do not cross the placental barrier appreciably and placental P-gp may be an important factor in this low penetration. PIs are therefore generally considered unsuitable for prevention of mother-to-infant transmission. Smit, Huisman and colleagues demonstrated that after intravenous administration of SQV to pregnant mice, the ratios of SQV concentration in fetal tissue to that in maternal plasma were 5-7 fold higher in mdr 1a/1b KO mice than in WT mice. (Smit 1999) They also demonstrated that P-gp fetal and blood-brain barriers are not abolished by co-administration of high doses of RTV. (Huisman 2000 ­ 1st Intl Wkshp on ClinPharm of HIV Therapy)

The effects of P-gp on the distribution, metabolism and excretion of drugs, including PIs, in the body is great. Blockage of P-gp may prove useful in facilitating greater intestinal absorption, bioavailability and penetration of PIs into HIV sanctuary sites as well as in reducing PI excretion. It may also simplify PI containing regimens by reducing the oral doses of PIs and the frequency at which they are taken. Higher PI levels in these sites may result in greater suppression of viral replication in these sanctuary sites, but they may also result in unwanted adverse effects. The effects of P-gp inhibition may not be limited to PIs but may extend to other co-administered drugs. For example, the antidiarrheal agent loperamide is a peripherally acting opiate which and penetrates the brain poorly. However, in mdr1a KO mice, loperamide exhibits strong morphine-like effects on the central nervous system. (Schinkel 1996 and Mayer 1997) Similarly, mice pre-treated with the P-gp inhibitors cyclosporin A and PSC833 exhibited increased sensitivity to the opiate fentanyl. (Mayer 1997)

In HIV infected cells with high P-gp expression, both accumulation and antiviral efficacy of IDV, SQV and RTV are diminished. (Lee 1998 and Jones 2000) Lucia and colleagues found that 90% of all peripheral blood lymphocyte subsets (CD4+, CD8+, CD56+ and CD10+ cells) expressed surface P-gp in both HIV-infected patients and controls. However, P-gp function was significantly reduced in CD16+ NK cells and CD19+ B-cells from HIV+ patients compared with controls. This reduced function significantly correlated with decreased NK cytotoxicity observed in HIV+ patients. (Lucia 1995 ­ Immunol Lett and AIDS Res Hum Retroviruses) P-gp can also be detected on an intracellular level in different peripheral blood monocyte subpopulations ­ mainly CD8+ T cells, CD16+ NK cells and CD14+ monocytes. This intracellular expression was decreased in CD8+ T cells and CD16+ NK cells from HIV-infected patients. (Malorni 1998)

In addition, a significantly increased proportion of CD4+ T-cells from HIV-infected patients expressed P-gp compared with controls. This resulted in a significantly increased ratio of the proportions of CD4+P-gp+ to CD8+P-gp+ cells. This ratio was significantly higher in patients with CD4+ cell counts of <200/ml than in those with CD4+ cell counts >200/ml. However, both CD4+ and CD8+ T-cells from HIV-infected patients accumulated more of the P-gp substrate rhodamine 123 compared with controls. P-gp inhibitors failed to increase further this intracellular accumulation in HIV+ patients. This suggests that in HIV infection there is increased expression of a functionally defective P-gp in CD4+ and CD8+ T-cells that appears to increase with disease progression. (Andreana 1996)

P-gp expression may affect HIV-infectivity. Lee and colleagues demonstrated a reduction in virus production when P-gp was over-expressed at the surface of 12D7, a continuous CD+ human T cell leukemia cell line, infected with HIV-1 NL4-3, a T-tropic molecular clone of HIV-1. Reduction in infectivity occurred both during the fusion of viral and plasma membranes and at subsequent steps in the HIV life cycle. P-gp over-expression did not significantly alter the surface expression or distribution of either the CD4 receptor or the CXCR4 coreceptor. (Lee 2000)

PIs cause hypercholesterolemia and, as explained earlier, P-pg plays a role in cholesterol metabolism and possibly in atherogenesis. Whether P-gp plays any role in PI-mediated dyslipidemia is not known.

Further Research on P-gp and HIV [TABLE 4]
The P-gp transport system clearly has major implications for HIV infection and its treatment. There is still much left to be understood. P-gp expression and function in HIV-infection needs to be studied ­ both in different tissues as well as in various stages of disease. The possibility of using P-gp function and expression as another surrogate marker for HIV disease progression should be explored. The effects of P-gp expression ­ and alterations in P-gp expression ­ on HIV infectivity, on the immune and other systems of HIV-infected individuals and on HIV therapy should be fully evaluated. Does P-gp play a role in the failure of antiretroviral therapy and the development of resistance to PIs?

The safety and efficacy of P-gp modulation in the management of HIV disease, especially in the use of PI-containing regimens, require further study. This includes the use of recognized P-gp inhibitors like PSC833, LY335979, and PIs like RTV, as well as the possibility of using P-gp maturation inhibitors (proteasome inhibitors). The optimal therapeutic dose of RTV required to inhibit P-gp; its effects on intracellular concentrations of PIs in HIV infected cells and on tissue penetration of PIs; its effects on concomitantly administered drugs; and the clinical value of using RTV as a P-gp inhibitor in the treatment of HIV disease remain to be evaluated. Are the different P-gp modulators site specific ­ do they inhibit P-gp to different degrees depending on location?

The interactions and interdependence of the P-gp transport and the cytochrome metabolic systems need further elucidation. Both are important causes of drug-drug interactions and HIV PIs interact with both systems. It may be necessary in the future to determine the interactions of HIV drugs not only with the cytochrome system, but also with the P-gp transport system.

It is necessary to investigate properly whether the co-administration of P-gp inhibitors with PIs is safe and effective for prophylaxis of mother-to-child transmission. Administration of P-gp inhibitors may be best done in later pregnancy to minimize the adverse effects of drugs and toxins on the developing fetus.

Studies also need to be undertaken to discern the mechanism of action of PI-induced dyslipidemias and what role, if any, P-gp plays. Does HIV disease itself affect the role of P-gp in cholesterol metabolism?

The P-gp transport system is complex and poorly understood. It is even less well understood in HIV disease, in which it may play a significant role. The role of P-gp in HIV disease pathogenesis and its effect on HIV drugs are undoubtedly deserving of greater study. It may become routine in the future to determine the interactions of HIV drugs not only with the cytochrome system, but also with the P-gp transport system.

TABLES

REFERENCES
A. Andreana, S. Aggarwal, et al. AIDS Res Hum Retroviruses 1996 Oct; 12(15):1457-62. Abnormal expression of a 170-kilodalton P-glycoprotein encoded by MDR1 gene, a metabolically active efflux pump in CD4+ and CD8+ T cells from patients with human immunodeficiency virus type 1 infection.
R. Arceci, J. Croop, et al. Proc Natl Sci USA 1988 Jun; 85(12):4350-4. The gene encoding multidrug resistance is induced and expressed at high levels during pregnancy in the secretory epithelium of the uterus.
B. Batetta, S. Dessė, et al. J Vasc Res 1999; 36(4):261-71. MDR1 GeneExpression in Normal and Atherosclerotic Human Arteries.
B. Batetta, A. Pani, et al. Cell Prolif 1999 Feb; 32(1):49-61. Correlation between cholesterol esterification, MDR1 gene expression and rate of cell proliferation in CEM and MOLT4 cell lines.
K. Bunting, J. Galipeau, et al. Blood 1998 Oct; 92(7):2269-79. Transductionof Murine Bone Marrow Cells With an MDR1 Vector Enables Ex Vivo Stem Cell Expansion, but These Expanded Grafts Cause a Myeloproliferative Syndrome in Transplanted Mice.
E. Choo, B. Leake, et al. Drug Metabol Disposition 2000 Jun; 28(6):655-60. Pharmacological Inhibition of P-glycoprotein Transport Enhances the Distribution of HIV-1 Protease Inhibitors into Brain and Testes.
P. Clayette, S. Jorajuria et al. AIDS 2000 Feb; 14(3):235-6. Significance ofP-glycoprotein for the pharmacology and clinical use of HIV protease inhibitors. (Editorial Comment)
P. Debry, E. Nash, et al. J Biol Chem 1997 Jan; 272(2):1026-31. Role of Multidrug Resistance P-glycoproteins in Cholesterol Esterification.
J. Drach, A. Gsur, et al. Blood 1996; 88(5):1747-54. Involvement ofP-glycoprotein in the transmembrane transport of interleukin-2 (IL-2), IL-4, and interferon-gamma in normal human T lymphocytes.
O. Fardel, A. Courtois, et al. Anticancer Drugs 1996 Aug; 7(6):671-7. Inhibition of P-glycoprotein activity in human leukemic cells by mifepristone.
O. Fardel, V. Lecureur, et al. Gen Pharmacol 1996 Dec; 27(8):1283-91. TheP-glycoprotein multidrug transporter.
M. Fromm. Int J Clin Pharmacol Ther 2000 Feb; 38(2): 69-74. P-glycoprotein:a defense mechanism limiting oral bioavailability and CNS accumulation of drugs.
R. Garraffo, J. Durant, et al. 1st International Workshop on Clinical Pharmacologyof HIV Therapy , Noordwijk, Netherlands, 2000; Abstract 4.5. Protease inhibitor(PI) kinetics in a human cell line expressing or not MDR1 gene and role of ritonavir (RI) to increase intracellular PI accumulation.
B. Greiner, M. Eichelbaum, et al. J Clin Invest 1999 Jul; 104(2):147-53. Therole of intestinal P-glycoprotein in the interaction of digoxin and rifampin.
M. Gottesman, I. Pastan, et al. Curr Opin Genet Develop 1996; 6:610-617. P-glycoprotein and multidrug resistance.
H. Gutmann, G. Fricker, et al. Mol Pharmacol 1999 Aug; 56(2):383-9. Interactions of HIV protease inhibitors with ATP-dependent drug export proteins.
C. Higgins and M. Gottesman. Trends Biochem Sci 1992 Jan; 17(1):18-21. Isthe multidrug transporter a flippase?
S. Hoffmeyer, O. Burk, et al. Proc Natl Acad Sci USA 2000 Mar;97(7):3473-78. Functional polymorphisms of the human multidrug-resistance gene: Multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo.
C. Hrycyna, M. Ramachandra, et al. J Biol Chem 1998 Jul; 273(27):16631-34. Mechanism of Action of Human P-glycoprotein ATPase Activity.
M. Huisman, J Smit, et al. AIDS 2000 Feb; 14(3):237-42. Significance of P-glycoprotein for the pharmacology and clinical use of HIV protease inhibitors. (Editorial Review)
M. Huisman, J. Smit, et al. 1st International Workshop on Clinical Pharmacologyof HIV Therapy, Noordwijk, Netherlands, 2000; Abstract 4.6 The role ofp-glycoprotein on oral bioavailability, brain and fetal penetration of the HIV protease inhibitor saquinavir.
R. Johnstone, E. Cretney, et al. Blood 1999 Feb; 93(3):1075-85. P-glycoprotein Protects Leukemia Cells Against Caspase-Dependent, but not Caspase-Independent, Cell Death.
R. Johnstone, A. Ruefli et al. Trend Biochem Sci 2000; 25(1):1-6. Multiplephysiological functions for multidrug transporter P-glycoprotein?
K. Jones, P. Hoggard, et al. 1st International Workshop on Clinical Pharmacologyof HIV Therapy, Noordwijk, Netherlands, 2000; Abstract 4.1. Factors Influencingthe Intracellular Accumulation of Protease Inhibitors.
R. Juliano and V. Ling. Biochim Biophys Acta 1976 Nov 11; 455(1):152-62. Asurface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants.
A. Kim, J. Dintaman, et al. J Pharmacol Experim Therapeut 1998 Sep;286(3):1439-45. Saquinavir, an HIV protease inhibitor, is transported by P-glycoprotein.
R. Kim, M. Fromm, et al. J Clin Invest 1998 Jan; 101(2):289-94. The DrugTransporter P-glycoprotein Limits Oral Absorption and Brain Entry of HIV-1 Protease Inhibitors.
R. Kim, C. Wandel et al. Pharm Res 1999 Mar; 16(3):408-14. Interrelationshipbetween substrates and inhibitors of human CYP3A and P-glycoprotein.
W. Klimecki, B. Futscher, et al. Blood 1994 May; 83(9):2451-8. P-glycoprotein expression and function in circulating blood cells from normal volunteers.
C. Lee, M. Gottesman, et al. Biochemistry 1998 Mar; 37(11):3594-601. HIV-1 protease inhibitors are substrates for the MDR1 multidrug transporter.
C. Lee, M. Ramachandra, et al. FASEBJ 2000 Mar; 14(3):516-22. Effect ofABC transporters on HIV-1 infection: inhibition of virus production by the MDR1 transporter.
T. Loo and D. Clarke. J Biol Chem 1999 Aug; 274(35):24759-65. TheTransmembrane Domains of the Human Multidrug Resistance P-glycoprotein Are Sufficient to Mediate Drug Binding and Trafficking to the Cell Surface.
T. Loo and D. Clarke. FASEBJ 1999; 13(13):1724-32. The human multidrugresistance P-glycoprotein is inactive when its maturation is inhibited: potential for a role in cancer chemotherapy.
M. Lucia, R. Cauda, et al. AIDS Res Hum Retroviruses 1995 Aug;11(8):893-901. Transmembrane P-glycoprotein (P-gp/P-170) in HIV infection: analysis of lymphocyte surface expression and drug-unrelated function.
M. Lucia, R. Cauda, et al. Immunol Lett 1995 Sep; 47(3):223-6. P-170glycoprotein (P-170) is involved in the impairment of natural killer cell-mediated cytotoxicity in HIV+ patients.
G. Luker, K. Nilsson, et al. J Biol Chem 1999 Mar; 274(11):6979-91. Multidrug Resistance (MDR1) P-glycoprotein Enhances Esterification of Plasma Membrane Cholesterol.
W. Malorni, M. Lucia, et al. Haematologica 1998 Jan; 83 (1): 13-20. Intracellular expression of P-170 glycoprotein in peripheral blood mononuclear cell subsets from health donors and HIV-infected patients.
U. Mayer, E. Wagenaar, et al. J Clin Invest 1997 Nov; 100(10):2430-6. FullBlockage of Intestinal P-glycoprotein and Extensive Inhibition of Blood-brain Barrier P-glycoprotein by Oral Treatment of Mice with PSC833.
L. Mickley, B. Spengler, et al. J Clin Invest 1997 Apr; 99(8):1947-57. GeneRearrangement: A Novel Mechanism for MDR-1 Gene Activation.
L. Profit, V. Eagling, et al. AIDS 1999 Sep; 13(13):1623-7. Modulation ofP-glycoprotein function in human lymphocytes and Caco-2 cell monolayers by HIV-1 protease inhibitors.
G. Randolph, S. Beaulieu, et al. Proc Natl Acad Sci USA 1998 Jun; 95:6924-9. A physiologic function for p-glycoprotein (MDR-1) during the migration of dendritic cells from skin via afferent lymphatic vessels.
V. Rao, J Dahlheimer, et al. Proc Natl Acad Sci USA 1999 Mar; 96:3900-5. Choroid plexus epithelial expression of MDR1 P-glycoprotein and multidrug resistance-associated protein contribute to the bloodBcerebro-spinal-fluid drug permeability barrier
M. Rosenberg, R. Callaghan, et al. J Biol Chem 1997 Apr; 272(16):10685-94. Structure of the multidrug resistance P-glycoprotein to 2.5 nm resolution determined by electron microscopy and image analysis.
T. Saito, Z. Zhang, et al. Neurosci Lett 1997 Aug; 232(1):41-4. P-glycoprotein expression in capillary endothelial cells of the 7^th and8^th nerves of guinea pig in relation to blood-nerve barrier sites.
S. Scala, N. Akhmed, et al. Mol Pharmacol 1997; 51:1024-33. P-Glycoprotein Substrates and Antagonists Cluster into Two Distinct Groups.
A. Schinkel, E. Wagenaar, et al. J Clin Invest 1996 Jun; 97(11):2517-24. P-glycoprotein in the Blood-Brain Barrier of Mice Influences the Brain Penetration and Pharmacological Activity of Many Drugs.
A. Schinkel, U. Mayer. et al. Proc Natl Acad Sci USA 1997 Apr; 94(8):4028-33. Normal viability and altered pharmacokinetics in mice lacking mdr1-type (drug transporting) P-glycoproteins.
E. Schuetz, A. Schinkel, et al. Proc Natl Acad Sci USA 1996 Apr; 93:4001-5. P-glycoprotein: A major determinant of rifampicin-inducible expression of cytochrome P4503A in mice and humans.
E. Schuetz, D. Umbenhauer, et al. Mol Pharmacol 2000 Jan; 57(1):188-97. Altered expression of hepatic cytochromes P-450 in mice deficient in one or more mdr1 genes.
A. Shapiro, K. Fox, et al. Eur J Biochem 1999; 259(3):841-50. Stimulation ofP-glycoprotein-mediated drug transport by prazosin and progesterone: Evidence for a third drug-binding site.
F. Sharom, X. Yu, et al. Biochem Pharmacol 1999 Aug; 58(4):571-86. Interaction of the P-glycoprotein multidrug transporter (MDR1) with high affinity peptide chemosensitizers in isolated membranes, reconstituted systems, and intact cells.
J. Smit, A. Schinkel, et al. Cell 1993 Nov; 75(3):451-62. Homozygousdisruption of the murine mdr2 P-glycoprotein gene leads to a complete absence of phospholipid from bile and to liver disease.
J. Smit, A. Shinkel, et al. Lab Invest 1994 Nov; 71(5):638-49. Tissuedistribution of the human MDR3 P-glycoprotein.
J Smit, M Huisman, et al. J Clin Invest 1999 Nov; 104(10):1441-7. Absence orpharmacological blocking of placental P-glycoprotein profoundly increases fetal drug exposure.
M. Smyth, E. Krasovskis, et al. Proc Natl Acad Sci USA 1998 Jun; 95:7024-9. The drug efflux protein, P-glycoprotein, additionally protects drug-resistant tumor cells from multiple forms of caspase-dependent apoptosis.
R. Srinivas, D. Middlemas, et al. Antimicrob Agents Chemother 1998 Dec;42(12):3157-62. Human Immunodeficiency Virus Protease Inhibitors Serve as Substrates for Multidrug Transporter Proteins MDR1 and MRP1 but Retain Antiviral Efficacy in Cell Lines Expressing These Transporters.
S. Taylor, D. Back, et al. 1st International Workshop on Clinical Pharmacology ofHIV Therapy, Noordwijk, Netherlands, 2000; Abstract 4.3. Antiviral DrugConcentrations in Semen of HIV-1 Infected Men are Drug Specific.
F. Thiebaut, T. Tsuruo, et al. Proc Natl Acad Sci. USA 1987 Nov; 84(21):7735-8. Cellular localization of the multidrug resistance gene product P-glycoprotein in normal human tissues.
P. Twentyman and N. Bleehen. Eur J Cancer 1991; 7(12):1639-42. Resistance modification by PSC-833, a novel non-immunosuppressive cyclosporin.
I. van der Sandt, I. Vos, et al. 1st International Workshop on ClinicalPharmacology of HIV Therapy, Noordwijk, Netherlands, 2000; Abstract 4.4. Activetransport of the HIV protease inhibitors amprenavir, ritonavir, and indinavir across the in vitro blood-brain barrier.
C. Washington, G. Duran, et al. J Acquir Immune Defic Syndr Hum Retrovirol 1998 Nov; 19(3):203-9. Interaction of anti-HIV protease inhibitors with the multidrug transporter P-glycoprotein (P-gp) in human cultured cells.
J. Wijnholds, E. de Lange, et al. J Clin Invest 2000 Feb; 105(3):279-85. Multidrug resistance protein 1 protects the choroid plexus epithelium and contributes to the blood-cerebrospinal fluid barrier.
R. Yumoto, T. Murakami, et al. J Pharmacol Experim Therapeut 1999;289(1):149-55. Transport of Rhodamine 123, a P-Glycoprotein Substrate, across Rat Intestine and Caco-2 Cell Monolayers in the Presence of Cytochrome P-450 3A-Related Compounds.

TABLES