Activation of PKC Isoform β I at the Blood–Brain Barrier Rapidly Decreases P-Glycoprotein Activity and Enhances Drug Delivery to the Brain

Abstract

P-glycoprotein is an ATP (adenosine triphosphate)-driven drug efflux transporter that is highly expressed at the blood–brain barrier (BBB) and is a major obstacle to the pharmacotherapy of central nervous system diseases, including brain tumors, neuro-AIDS, and epilepsy. Previous studies have shown that P-glycoprotein transport activity in rat brain capillaries is rapidly reduced by the proinflammatory cytokine, tumor necrosis factor-α (TNF-α) acting through protein kinase C (PKC)-dependent signaling. In this study, we used isolated rat brain capillaries to show that the TNF-α-induced reduction of P-glycoprotein activity was prevented by a PKCβ_I/II inhibitor, LY333531, and mimicked by a PKCβ_I/II activator, 12-deoxyphorbol-13-phenylacetate-20-acetate (dPPA). Western blotting of brain capillary extracts with phospho-specific antibodies showed that dPPA activated PKCβ_I, but not PKCβ_II. Moreover, in intact rats, intracarotid infusion of dPPA potently increased brain accumulation of the P-glycoprotein substrate, [³H]-verapamil without compromising tight junction integrity. Thus, PKCβ_I activation selectively reduced P-glycoprotein activity both in vitro and in vivo. Targeting PKCβ_I at the BBB may prove to be an effective strategy for enhancing the delivery of small molecule therapeutics to the brain.

Introduction

P-glycoprotein at the blood–brain barrier (BBB) is a major obstacle to central nervous system (CNS) pharmacotherapy (Miller et al, 2008). This ATP (adenosine triphosphate)-driven efflux transporter limits the ability of a large number of therapeutic drugs to cross the BBB and thus hampers effective treatment of several CNS diseases, including bacterial/viral infection, human immunodeficiency virus encephalitis, epilepsy, multiple sclerosis, neurodegenerative diseases, and brain tumors (Miller et al, 2008). Specific inhibition of P-glycoprotein transport activity (e.g., with cyclosporin A (CSA) or PSC833 (valspodar)) can increase the brain accumulation of clinically important therapeutic compounds in both animal models and humans (Fellner et al, 2002; Hubensack et al, 2008). However, these drugs have consistently failed as chemosensitizers in clinical trials because of significant side effects and organ toxicity (Friedenberg et al, 2006; Hubensack et al, 2008; Pein et al, 2007).

We previously suggested an alternative strategy for improving delivery of P-glycoprotein substrates to the brain: targeting signaling pathways that maintain transporter activity (Hartz et al, 2004, 2006). Studies from this laboratory have shown that exposing isolated rat brain capillaries to tumor necrosis factor (TNF)-α rapidly and reversibly reduces P-glycoprotein transport activity (Hartz et al, 2006). Tumor necrosis factor-α-induced loss of P-glycoprotein activity is signaled through TNF-α receptor 1, endothelin-1 (ET-1), ET_B receptor, nitric oxide synthase, and protein kinase C (PKC). The effect of TNF-α or ET-1 is mimicked by treatment with 10 nmol/L PMA (phorbol-12-myristate-13-acetate), and is prevented by treatment with 10 nmol/L BIM-1 (bis-indolylmaleimide), indicating involvement of a classic PKC isoform (α/β_I/II/γ) (Hartz et al, 2004, 2006).

Using both in vitro and in vivo techniques, in this study, we identify the specific PKC isoform that connects TNF-α signaling to rapid loss of P-glycoprotein transport activity. Using isolated rat brain capillaries, we show that TNF-α signals through PKCβ_I to rapidly reduce P-glycoprotein transport activity. Using in situ brain perfusion in rats, we further show that intracarotid administration of 12-deoxyphorbol-13-phenylacetate-20-acetate (dPPA), a selective activator of PKCβ_I, substantially increases delivery of the P-glycoprotein substrate and Ca²⁺ channel blocker, [³H]-verapamil, to the brain. Therefore, selective activation of PKCβ_I provides a novel and effective strategy for suppressing P-glycoprotein activity at the BBB and improving CNS drug delivery.

Materials and methods

Reagents

[N-ε(4-nitrobenzofurazan-7-yl)-D-Lys(8)]-CSA (NBD-CSA) was custom synthesized by R Wenger (Basel, Switzerland) (Schramm et al, 1995). PSC833 (valspodar) was a kind gift from Novartis (Basel, Switzerland). [¹⁴C]-sucrose (specific activity 300 mCi/mmol) and [³H]-verapamil (specific activity 80 Ci/mmol) were obtained from American Radiolabeled Chemicals (St Louis, MO, USA). Protein kinase C isoform-specific antibodies were anti-phospho-PKCβ_I (targets threonine 642; GeneTex, Irvine, CA, USA), anti-phospho-PKCβ_II (targets serine 660; Epitomics, Burlingame, CA, USA), anti-phospho-PKCα (targets serine 657), and anti-PKCβ_I, β_II or α (Santa Cruz Biotechnology, Santa Cruz, CA, USA). dPPA was obtained from Biomol (Enzo Life Sciences, Plymouth Meeting, PA, USA). Recombinant human TNF-α and all other chemicals were obtained from Sigma (St Louis, MO, USA).

Brain Capillary Isolation

All experiments were carried out in compliance with NIH (National Institutes of Health) animal care and use guidelines and approved by the Animal Care and Use Committee of National Institute of Environmental Health Sciences. Brain capillaries were isolated from adult male retired breeder Sprague–Dawley rats as described previously (Hartz et al, 2004). Briefly, rats were killed by CO₂ inhalation, decapitated, and their brains were removed immediately to 4°C phosphate-buffered saline (PBS) buffer (in mmol/L: KCl 2.7, KH₂PO₄ 1.5, NaCl 136.9, Na₂HPO₄ 4.3, CaCl₂ 0.7, MgCl₂ 0.5, D-glucose 5, and pyruvate 1, pH 7.4). Cortical gray matter was dissected, homogenized, and centrifuged in 15% Ficoll PM 400 for 20 mins (5,800 × g at 4°C). Capillary-enriched pellets were resuspended in PBS with 1% bovine serum albumin (BSA), filtered using a 300-μm nylon mesh and passed over a glass bead column supported by a 30-μm nylon mesh. Capillaries were collected from the filter support and glass beads by mechanical agitation and aspiration. They were washed three times in BSA-free PBS and immediately used for transport studies, membrane isolation, or immunostaining.

Transport Assays

Procedures are described in detail in the studies by Bauer et al (2007) and Hartz et al (2006). Briefly, freshly isolated brain capillaries in PBS buffer were transferred into Teflon microscope chambers with glass coverslip bottoms. In the chambers, they were exposed to activators and inhibitors of signaling and to fluorescent substrates (see figure legends for details of individual experiments). In a typical experiment, capillaries were pretreated for 30mins with inhibitors of signaling, followed by exposure to the inhibitors plus TNF-α, ET-1, or dPPA for 1 or 6 h (indicated in the description of each experiment). During the last hour, fluorescent substrates (2 μmol/L NBD-CSA for P-glycoprotein or 2 μmol/L Texas red (sulforhodamine 101) for multidrug resistance associated protein 2 (MRP2) were present in the chamber. We previously showed that 1-h incubation of capillaries with these fluorescent substrates results in steady-state luminal accumulation. Confocal images of capillaries were acquired using a Zeiss 510 meta-laser scanning confocal microscope (Zeiss, Jena, Germany) fitted with a × 40 water-immersion objective (numerical aperture 1.2). Luminal fluorescence intensity was measured as described previously (Hartz et al, 2004; Miller et al, 2000) using NIH Image J software (WS Rasband, Image J, US National Institutes of Health, Bethesda, MD, USA, http://rsb.info.nih.gov/ij/, 1997 to 2008).

Tissue Fractionation

Isolated capillaries obtained from 10 rat brains were divided into 2 to 4 samples. After 30 mins of exposure to dPPA or TNF-α, capillaries were flash frozen in liquid nitrogen and resuspended in 100 μL solution A: PBS (at 4°C) with protease and phosphatase inhibitors (Mini Complete Protease Inhibitor tablet (Roche, Indianapolis, IN, USA) and 2 mmol/L EGTA, 5 mmol/L EDTA, 30 mmol/L NaF, 20 mmol/L pyrophosphate, 10 mmol/L orthovanadate, and 8.6 mg/mL β-glycerophosphate). The suspension was centrifuged at 15,000 × g, and the supernatant was used for western blot analysis of aqueous proteins. The pellet was washed once with solution A, then resuspended in CellLytic MT mammalian tissue lysis reagent with 10% solution A, followed by sonication (30 secs), rest on ice (10 mins), and centrifugation (15,000 g, 10 mins, 4°C). The resulting membrane extract was used for western blot analysis.

Western Blotting

Western blots were performed using the NuPage Bis-Tris electrophoresis system (Invitrogen, Carlsbad, CA, USA). Polyvinylidene fluoride transfer membranes were incubated in Odyssey (Li-Cor, Lincoln, NE, USA) blocking solution for 1 h (room temperature) and incubated overnight at 4°C with primary antibodies. Membranes were washed in PBS with 0.1% Tween and incubated in 50 ng/mL secondary antibody (Odyssey Goat anti-mouse IR Dye 800CW or Goat anti-rabbit IR dye 680CW) for 30 to 60mins. Alternatively, blocking and antibody incubation steps were performed using the Millipore Snap i.d. blotting system according to the manufacturer's specifications (Millipore, Billerica, MA, USA). Western blot infrared fluorescence was detected using the Li-Cor Odyssey infrared imaging system. Integrated band intensities were quantified using the NIH Image J software. For each sample, PKC isoform band intensities were normalized to β-actin and results are expressed as the normalized treatment to control ratio.

Immunohistochemistry

Isolated capillaries adhering to glass coverslips were fixed in 3% paraformaldehyde/0.2% glutaraldehyde for 10 mins, as described previously (Bauer et al, 2007; Hartz et al, 2006). Briefly, fixed capillaries were washed in PBS, permeabilized in 1% Triton X-100 in PBS, and blocked for 1 h with 1% BSA in PBS. Capillaries were incubated in 0.5 μg/mL monoclonal C219 anti-P-glycoprotein antibody (Signet, Dedham, MA, USA) for 1 h at 37°C and washed in 1% BSA, followed by exposure to 2 μg/mL Alexa Fluor 488-conjugated goat anti-mouse IgG (Invitrogen) for 1 h at 37°C. For negative controls, the primary antibody was omitted in the first incubation step. Immunostained capillaries were visualized by confocal microscopy (×60 oil objective; N.A. 1.4), and membrane fluorescence intensity was measured as before (Bauer et al, 2007), using the NIH Image J software.

In Situ Brain Perfusion

In situ brain perfusion was carried out as described previously (Hawkins et al, 2007). Briefly, male Sprague–Dawley rats were anesthetized by intraperitoneal injection with 1 mL/kg ketamine cocktail (79 mg/mL ketamine, 3 mg/mL zylazine, 0.6 mg/mL acepromazine) and administered heparin (10 kU/kg). The common carotid arteries were exposed by midline incision at the neck and perfused with oxygenated Ringer's solution (37°C) (in mmol/L: NaCl 117, KCl 4.7, MgSO₄ 0.8, NaHCO₃ 24.8, KH₂PO₄ 1.2, CaCl₂ 2.5, D-glucose 10, plus 39 g/L 170 kDa dextran, 1 g/L BSA, and 0.055 g/L Evans Blue) at 3 mL/min. [¹⁴C]-sucrose (0.5 μCi/mL) or [³H]-verapamil (0.1 μCi/mL) were infused into the circuit using a syringe pump at 0.5 mL/min for 20 mins. Samples of perfusate were collected at the end of each experiment. Brains were removed and stripped of meninges, midbrain, and choroid plexuses, and minced by hand. Tissue and 100 μL perfusate samples were incubated for 2 days with tissue solubilization solution (Hyamine Hydroxide, MP Biomedicals, Santa Ana, CA, USA). Solubilized samples were prepared for scintillation counting by adding 100 μL of 30% acetic acid and 4mL liquid scintillation cocktail (CytoScint ES, MP Biomedicals), and incubating overnight in the dark. Results of scintillation counting were expressed as the ratio of radioactivity in the brain to that of perfusate (Rbr μL/g):

$\begin{array}{c}(R_{\mathrm{b}\mathrm{r}}=C_{\mathrm{b}\mathrm{r}\mathrm{a}\mathrm{i}\mathrm{n}}/C_{\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{f}\mathrm{u}\mathrm{s}\mathrm{a}\mathrm{t}\mathrm{e}})\end{array}$

(1)

C_brain represents radioactivity in the brain tissue (d.p.m./g) and C_perfusate represents radioactivity in the perfusate (d.p.m./μL).

Statistical Analyses

Data are expressed as mean ± s.e. Differences between mean values were deemed statistically significant when P < 0.05 using one-way ANOVA (analysis of variance) with a Tukey–Kramer post hoc test (for multiple comparisons) or Student's unpaired t-test (GraphPad Prism version 4.0c for Macintosh; GraphPad Software, San Diego, CA, USA; http://www.graphpad.com).

Results

P-Glycoprotein-Mediated Transport in Isolated Rat Brain Capillaries

To determine P-glycoprotein transport activity, we used confocal microscopy and quantitative image analysis to measure steady-state accumulation of NBD-CSA, a fluorescent P-glycoprotein substrate, in the capillary luminal space (Bauer et al, 2007; Hartz et al, 2004, 2006; Miller et al, 2000). As before, luminal accumulation of NBD-CSA was highly concentrative (lumen-to-bath fluorescence ratios exceed 10) and was reduced by NaCN (inhibits energy metabolism) and by the specific P-glycoprotein inhibitor, PSC833 (Figures 1A and 1B). As in previous reports, in this study, we define specific P-glycoprotein transport activity as NaCN- or PSC833-sensitive luminal NBD-CSA accumulation; residual NBD-CSA accumulation represents diffusive entry of NBD-CSA into the luminal space plus nonspecific binding to the capillary tissue (Hartz et al, 2004, 2006; Miller et al, 2000).

Figure 1

LY333531 prevents rapid loss of P-glycoprotein transport activity (steady-state luminal NBD-CSA fluorescence) induced by tumor necrosis factor (TNF)-α and endothelin-1 (ET-1) in rat brain capillaries. (A) Representative confocal images of capillaries showing reduced luminal accumulation with exposure to 5 μmol/L PSC833 or 1 mmol/L NaCN. (B) Quantitative image analysis of luminal NBD-CSA accumulation in capillary lumens showing significant decreases with exposure to 5 μmol/L PSC833 or 1 mmol/L NaCN. (C) Representative confocal images of brain capillaries exposed for 1 h to 100 ng/mL TNF-α or 100 ng/mL TNF-α plus 10 nmol/L LY333531. (D) A volume of 10 nmol/L LY333531 blocks the effects of 100 ng/mL TNF-α or 100 nmol/L ET-1 on P-glycoprotein activity. (E) LY333531 blocks the effects of PMA on P-glycoprotein activity. Pooled data from 3 to 5 experiments, each involving measurements of 10 to 15 capillaries within each treatment are shown. Statistical comparisons: *Significantly different from control, P < 0.05; Significantly different from control, P < 0.001.

Rapid Loss of P-Glycoprotein Activity By Tumor Necrosis Factor-α is Signaled Through Protein Kinase C Isoform β

Previous studies have shown reduced P-glycoprotein transport activity in rat brain capillaries exposed to TNF-α or ET-1. (Hartz et al, 2004, 2006). Figures 1C and 1D confirm those findings. For both effectors, the reduction in transport activity was abolished when capillaries were pretreated with LY333531, a specific PKCβ_I/II inhibitor. LY333531 also abolished the reduction in transport activity caused by exposure to the PKC activator, PMA (Figure 1E). Consistent with PKCβ_I/II signaling reducing P-glycoprotein transport activity, dPPA, an activator of these PKC isoforms, mimicked the effects of TNF-α, ET-1, and PMA (Figure 2A). dPPA stimulation of transport was antagonized by the isoform-specific inhibitor, LY333531 (Figure 2B). It is noteworthy that the estimated EC₅₀ for LY333531 reversal of dPPA stimulation of transport was 6nmol/L, which agrees well with the half-maximal inhibitory concentration (IC₅₀ = 4 to 6nmol/L) of LY333531 for PKCβ_I/II (Jirousek et al, 1996). It is also noteworthy that this IC₅₀ is 1 to 2 orders of magnitude lower than that of LY333531 for any other PKC isoform (Jirousek et al, 1996).

Figure 2

12-Deoxyphorbol-13-phenylacetate-20-acetate (dPPA) reduces P-glycoprotein transport activity. (A) Dose response. (B) Reversal of dPPA effect by LY333531. The half-maximal effective concentration (EC₅₀) for LY333531 is 6 nmol/L. (C) Lack of effect of dPPA on Mrp2-mediated Texas red accumulation in capillary lumens. Texas red accumulation was reduced when metabolism was inhibited by NaCN and when tight junctions were disrupted by hypertonic mannitol. Pooled data from 3 to 5 experiments, each involving measurements of 10 to 15 capillaries within each treatment are shown. Statistical comparisons: Significantly different from control, P < 0.001.

Although these results are consistent with the loss of specific transport activity, they could also be explained by PKCβ_I/II activation increasing tight junctional permeability and thus causing leak of pumped NBD-CSA from the luminal space to the bath. We previously showed that tight junction opening did not underlie the reduction in luminal accumulation of NBD-CSA caused by TNF-α or ET-1 (Hartz et al, 2004, 2006). In those experiments, we reasoned that increased tight junctional permeability would affect luminal accumulation of any solute pumped into that space. We showed that neither TNF-α nor ET-1 affected luminal accumulation of the fluorescent organic anion sulforhodamine 101 (Texas red; Hartz et al (2004, 2006)), a substrate for, Mrp2, another ATP-driven efflux pump in rodent brain capillaries, (Bauer et al, 2008a). Figure 2C shows that luminal accumulation of Texas red was not affected by dPPA, but was reduced by the metabolic inhibitor, NaCN and by the addition of 100 mmol/L mannitol to the medium. Exposure of rat brain capillaries to hypertonic mannitol solution increases tight junctional permeability, inducing loss of Texas red and other transported substrates from the luminal space (Hartz et al, 2004, 2006). Thus, dPPA, similar to TNF-α and ET-1, reduced luminal accumulation of NBD-CSA by impairing specific transport rather than by increasing junctional permeability. Taken together, these experiments with isoform-selective PKC effectors identify PKCβ_I/II as a negative modulator of P-glycoprotein transport activity in rat brain capillaries.

Tumor Necrosis Factor-α Exposure Increases the Amount of Active Protein Kinase Cβ_I/II at the Membrane

Although activation of PKCβ_I/II was required for the rapid loss of P-glycoprotein activity induced by TNF-α, ET-1, or PMA, the present transport experiments by themselves did not identify which of the two PKCβ isoforms was responsible for the loss of transport, nor the mechanism by which these agents triggered activation of PKC. In general, PKC activation requires translocation of the affected isoform to the plasma membrane, and PDK1-dependent threonine (T500) phosphorylation, followed by autophosphorylation of the isoform. For PKCβ_I/II, autophosphorylation occurs at serine 661/660 (β_I and β_II, respectively) and threonine 642/641 (β_I and β_II, respectively) (Parekh et al, 2000; Steinberg, 2008). Thus, changes in PKC isoform localization and specific residue autophosphorylation could disclose the critical isoform.

We fractionated brain capillaries to determine the aqueous versus membrane distribution, as well as the autophosphorylation status of PKCβ_I/II. For these experiments, capillaries were treated with or without (control) 10 ng/mL TNF-α for 30 mins before fractionation. Western blots of aqueous and membrane fractions were probed with antibodies for total PKCα, β_I, or β_II, as well as for phosphorylated PKCα (S657), β_I (T642), or β_II (S660). Tumor necrosis factor-α treatment did not change the amounts of total PKCα, β_I, or β_II in cytosolic or membrane fractions of brain capillaries (for PKCβ_I or β_II, see Figures 3A and 3B; PKCα data not shown). Therefore, TNF-α did not increase protein expression or membrane translocation of PKCα, β_I, or β_II in brain capillaries. In contrast, TNF-α exposure did increase (by 1.3-fold) the phosphorylation of PKCβ_I and PKCβ_II in membrane fractions (Figures 3C and 3D). Both PKCβ_I and PKCβ_II phosphorylation levels did not change in aqueous fractions (Figures 3C and 3D), and TNF-α did not increase the phosphorylation of PKCα in aqueous or membrane fractions (not shown). Therefore, rapid inhibition of P-glycoprotein transport activity in response to TNF-α coincided with activation (autophosphorylation) of membrane-localized PKCβ_I and PKCβ_II.

Figure 3

Phosphorylation and membrane/aqueous distribution of protein kinase C (PKC)β_I (A, C) and PKCβ_II (B, D) in fractions derived from control and tumor necrosis factor (TNF)-α-treated (10 ng/mL for 30 mins) brain capillaries. Panels A and B show western blots for total PKC isoform levels. Panels C and D show blots for phosphorylated isoforms. Representative blots, and quantified band intensities normalized to actin, involving measurements from 5 to 10 separate experiments are shown. Statistical comparisons: Significantly different from control, P < 0.001.

dPPA Selectively Activates Protein Kinase C Isoform β_I in Brain Capillaries

To confirm that dPPA treatment also selectively activated PKCβ_I/II in brain capillaries, aqueous and membrane fractions of capillary lysates were analyzed by western blotting for total or phosphorylated PKCα, β_I, or β_II. After 30 mins of exposure to 10nmol/L dPPA, the amount of membrane-associated PKCβ_I significantly increased 1.5-fold (Figures 4A), indicating that PKCβ_I translocated to the membrane in response to dPPA treatment. dPPA did not increase the amount of PKCβ_II (Figure 4B) or PKCα (not shown) at the membrane. dPPA did increase (1.2-fold) the amount of phosphorylated PKCβ_I at the membrane (Figure 4C). However, phosphorylated PKCβ_II (Figure 4D) and phosphorylated PKCα (not shown) at the membrane did not increase. Therefore, dPPA treatment specifically activates PKCβ_I in brain capillaries by translocating only PKCβ_I (both phosphorylated and unphosphorylated forms) to the membrane.

Figure 4

Phosphorylation and membrane/aqueous distribution of protein kinase C (PKC)β_I (A, C) and PKCβ_II (B, D) in fractions derived from control and 12-deoxyphorbol-13-phenylacetate-20-acetate (dPPA)-treated (10 nmol/L for 30 mins) brain capillaries. Panels A and B show western blots for total PKC isoform levels. Panels C and D show blots for phosphorylated isoforms. Representative blots, and quantified band intensities normalized to actin, involving measurements from 5 to 10 separate experiments are shown. Statistical comparisons: Significantly different from control, P < 0.001.

dPPA Reduces P-Glycoprotein Activity for at Least 6 h

Tumor necrosis factor-α effects on rat brain capillary P-glycoprotein are time dependent. Tumor necrosis factor-α exposure rapidly reduces P-glycoprotein transport activity (Bauer et al, 2007; Hartz et al, 2006). However, with 3 to 5 h exposure, P-glycoprotein activity recovers to control levels, and with 6 h exposure, P-glycoprotein activity and protein expression substantially exceed control levels (Bauer et al, 2007). As both rapid and long-term effects of TNF-α exposure on P-glycoprotein regulation are inhibited by BIM-1 (Bauer et al, 2007; Hartz et al, 2004, 2006), these multiphasic effects could be mediated entirely by the activation of PKCβ_I. We exposed capillaries for 6 h to TNF-α or dPPA and then measured P-glycoprotein transport activity and expression. As shown previously (Bauer et al, 2007), 6 h of TNF-α exposure increased P-glycoprotein activity (Figure 5A) and protein expression (Figure 5B). In contrast, 6 h of dPPA exposure caused the same reduction in P-glycoprotein activity observed after 1 h (Figure 5A); immunostaining showed no change in the expression of P-glycoprotein protein after 6 h of dPPA exposure (Figure 5B). Thus, selective activation of PKCβ_I rapidly reduced P-glycoprotein activity but did not increase transporter expression over the longer term.

Figure 5

P-glycoprotein transport activity (A) and protein expression (B) in brain capillaries after 6 h of exposure to 10 ng/mL tumor necrosis factor (TNF)-α or 10 nmol/L 12-deoxyphorbol-13-phenylacetate-20-acetate (dPPA). (A) Pooled data from 5 to 7 experiments, each involving measurements of 10 to 15 capillaries within each treatment are shown. (B) Representative confocal images of capillaries immunostained for P-glycoprotein (green), counterstained with propidium iodide (red). Numbers under the images show the corresponding relative fluorescence of P-glycoprotein from 10 to 15 capillaries in each treatment, representing 3 independent experiments. Statistical comparisons: **Significantly different from control, P < 0.01; Significantly different from control, P < 0.001.

Activation of Protein Kinase Cβ_I Enhances In Vivo Drug Delivery to the Brain

To test the effects of dPPA exposure on P-glycoprotein transport function in vivo, we used in situ brain perfusion, a well-established procedure for monitoring drug transport into the brain (Hawkins and Egleton, 2008). In initial experiments, we measured the brain distribution of [¹⁴C]-sucrose, a marker for paracellular vascular permeability, to verify that the effects of dPPA treatment on brain drug distribution are not due to increased tight junction permeability. Neither the P-glycoprotein inhibitor CSA at 8 μmol/L nor dPPA at 1 μmol/L had any effect on brain distribution of [¹⁴C]-sucrose (Figure 6A). To show the effect of disrupting tight junctions on brain distribution of [¹⁴C]-sucrose, we administered an intracarotid bolus injection of 1.3 mol/L mannitol immediately before [¹⁴C]-sucrose perfusion. Perfusing the brain with hyperosmotic mannitol increased the uptake of [¹⁴C]-sucrose in the brain tissue 3.6-fold (Figure 6A); this is in good agreement with the 3.8-fold increase observed previously with 1.6 mol/L mannitol (Brown et al, 2004).

Figure 6

In situ brain perfusion with [³H]-verapamil or [¹⁴C]-sucrose. (A) Brain distribution of [¹⁴C]-sucrose was not increased in response to 8 μmol/L cyclosporine A (CSA) or 1 μmol/L 12-deoxyphorbol-13-phenylacetate-20-acetate (dPPA). A bolus of 1.3 mol/L mannitol increased [¹⁴C]-sucrose accumulation. (B) Perfusion with 8 μmol/L CSA increased [³H]-verapamil accumulation. (C) Perfusion with dPPA increased [³H]-verapamil accumulation. Statistical comparisons: *Significantly different from control, P < 0.05; Significantly different from control, P < 0.001 (data shown are n > 5 rats in each treatment group).

We next measured brain accumulation of [³H]-verapamil, a P-glycoprotein substrate. Similar to previous experiments, 8 μmol/L CSA significantly increased the brain accumulation of [³H]-verapamil 6.2-fold (Figure 6B) (Hawkins et al, 2010). When 1 to 1,000 nmol/L dPPA was added to the perfusate, we observed a concentration-dependent increase in the accumulation of [³H]-verapamil in the brain (Figure 6C). Remarkably, [³H]-verapamil uptake in the brain increased 3.4-fold during the 20 mins perfusion with 1,000 nmol/L dPPA. To confirm that increased brain distribution of [³H]-verapamil was not due to increased accumulation in the endothelium, capillary depletion was performed as described previously (Triguero et al, 1990) on brain samples obtained from a subset of animals in the control and 1,000 nmol/L dPPA groups. We found no significant difference between the radioactivity in brain homogenate and capillary-depleted supernatant in either group, nor was there any significant difference in radioactivity in the capillary pellets between groups. Thus, dPPA substantially enhanced delivery of the drug, [³H]-verapamil, to the brain in vivo, presumably through PKCβ_I activation and subsequent loss of P-glycoprotein activity.

Discussion

Limited therapeutic drug delivery to the CNS is a serious problem in the clinic. The ATP-driven drug efflux transporters, e.g., P-glycoprotein, expressed at the BBB substantially underlie our inability to deliver small molecule therapeutics. We previously speculated that TNF-α/ET-1 signaling to BBB P-glycoprotein could be used to rapidly and transiently reduce transporter activity and thus improve brain delivery of drugs that are P-glycoprotein substrates (Hartz et al, 2004, 2006). Such signaling involves TNF-R1 and ET_B receptors, nitric oxide synthase, and at least one classic PKC isoform. In this study, we used isolated rat brain capillaries to identify PKCβ_I as the specific PKC isoform that signals reduced transporter activity and showed that activating that isoform in vivo increased brain distribution of a drug that is a P-glycoprotein substrate.

We found that TNF-α- or ET-1-induced loss of P-glycoprotein transport activity in rat brain capillaries was prevented by a PKCβ_I/II inhibitor, LY333531, and mimicked by a PKCβ_I/II selective activator, dPPA. Additional analyses of PKCβ_I/II translocation and phosphorylation patterns in rat brain capillaries identified PKCβ_I as the critical isoform activated by both TNF-α and dPPA. Tumor necrosis factor-α increased the phosphorylation of both PKCβ_I (T642) and PKCβ_II (S660) in capillary membrane fractions, but had no apparent effect on the aqueous versus membrane distribution of either isoform. dPPA only increased PKCβ_I membrane translocation and phosphorylation (S657). Neither TNF-α nor dPPA affected phosphorylation or aqueous/membrane distribution of another classic isoform, PKCα. Thus, the rapid loss of P-glycoprotein transport activity induced by TNF-α or dPPA was associated with specific activation of PKCβ_I, not PKCβ_II. These findings for PKCβ_I signaling are novel in two respects: TNF-α has not been shown previously to activate PKCβ;, and PKCβ_I is not known to regulate P-glycoprotein activity in any tissue. The only reported function of PKCβ_I signaling in brain capillary endothelium is to increase Na–K–2Cl cotransport activity during hypoxia/aglycemia (Yang et al, 2006), which may contribute to ionic imbalances observed in the brain during ischemia.

It is noteworthy that we previously showed that long-term (6 h) exposure of rat brain capillaries to TNF-α increased both P-glycoprotein transport activity and transporter protein expression (Bauer et al, 2007). The present 6-h experiments confirm the previous finding for TNF-α; but they also show no increases in transport activity or transporter protein expression with 6-h exposure to dPPA. Rather, capillaries exposed to dPPA for 6 h exhibited reduced P-glycoprotein transport activity and control levels of transporter expression, just like capillaries exposed to TNF-α or dPPA for 1 h. As TNF-α signaled through both PKCβ_I and PKCβ_II and dPPA only signaled through PKCβ_I, it appears that the TNF-α-induced increase in P-glycoprotein expression observed after 6 h exposure required activation of PKCβ_II alone or in concert with PKCβ_I. At present, we cannot differentiate between these two possibilities. However, it is apparent that the rapid loss of P-glycoprotein transport activity and the delayed increase in transporter expression are signaled through divergent pathways well downstream of TNF-α.

By what sequence of events does PKCβ_I decrease P-glycoprotein activity in the brain capillary endothelium? There are two aspects to this larger question: how is the signal transmitted to P-glycoprotein, and how is the transporter inactivated? With regard to PKCβ_I signaling to P-glycoprotein, exhaustive muta-genesis of all consensus PKC phosphorylation sites on P-glycoprotein has shown that transporter phosphorylation by PKC is not required for regulation (Goodfellow et al, 1996). In preliminary experiments, we immunoprecipitated P-glycoprotein from lysates of brain capillaries treated with dPPA and used phospho-serine-, phospho-threonine-, or phosphotyrosine-specific antibodies to detect changes in transporter phosphorylation. No such changes could be detected (RR Rigor, unpublished data). Therefore, rapid loss of P-glycoprotein activity by PKCβ_I signaling likely involves phosphorylation of an unidentified accessory protein. One such protein is RBCK1, a specific adapter protein and binding partner for PKCβ_I that confers specificity by linking PKCβ_I and other regulatory proteins to their targets (Ron and Mochly-Rosen, 1995; Vallentin and Mochly-Rosen, 2007). In preliminary western blot experiments with rat brain capillary lysates, we detected strong expression of RBCK1 (RR Rigor, unpublished data). Identification of RBCK1 binding partners in brain capillaries could help elucidate additional cellular elements responsible for PKCβ_I-induced loss of P-glycoprotein activity. With regard to transporter inactivation, it is still not clear whether the transporter physically moves to an inactive compartment, e.g., trafficking into an endosome, or whether it is inactivated through protein–protein interactions within the plasma membrane.

The most exciting finding of this study is the observation that pharmacological activation of PKCβ_I increased by 3.4-fold in vivo brain uptake of verapamil, a P-glycoprotein substrate. This increase was not accompanied by an increase in sucrose uptake, indicating that the physical BBB at the level of the tight junctions was preserved. The magnitude of dPPA-induced increase in verapamil uptake was roughly 50% of what we measured when 8 μmol/L CSA was added to the perfusate, which may have come close to inhibiting the transporter fully. In agreement with this supposition, the effect of dPPA is ∼50% that of the increase in brain accumulation of verapamil observed in P-glycoprotein knockout in mice (Dagenais et al, 2001). By itself, this 3.4-fold increase in brain uptake is impressive. However, for other more lipophillic drugs, e.g., some chemotherapeutics, knocking out P-glycoprotein in mice can increase brain accumulation 20- to 30-fold (Mizuno et al, 2003). Assuming proportional increases, one can expect the same degree of PKCβ_I activation as observed here to increase brain accumulation of these drugs by at least an order of magnitude.

It is our contention that a fundamental understanding of the mechanisms that regulate expression and transport activity of P-glycoprotein at the BBB can form the basis for strategies designed to control that transporter in the context of drug delivery. We consider two different ways to target signaling to P-glycoprotein. First, disrupting the signaling pathway through which epileptic seizures increase P-glycoprotein expression has suggested ways to limit that upregulation and thus improve CNS delivery of antiepileptic drugs that are P-glycoprotein substrates, e.g., inhibit cyclooxygenase-2, block EP-1 receptors (Bauer et al, 2008b; Pekcec et al, 2009; Zibell et al, 2009). As this strategy targets signals exclusive to seizure-induced upregulation of transporter expression, its use is limited to that pathway and presumably pathways that intersect upstream of cyclooxygenase-2. Second, activating PKCβ_I, a signal responsible for establishing the set point for basal transporter activity (in this study) should have wider applicability, especially if the set point can be rapidly and reversibly modulated. It is noteworthy that multiple signaling pathways and transcription factors converge to upregulate P-glycoprotein expression in brain capillaries (Miller et al, 2008). We speculate that targeting transporter activity through PKCβ_I has the potential to enhance delivery of therapeutic drugs to the brain, even in situations in which transporter expression is upregulated through genomic mechanisms. This exciting possibility remains to be explored.

This in vitro/in vivo study provides proof of principle for the targeting of BBB P-glycoprotein activity through PKCβ_I activation. It suggests a way to do this in a rapid and transient manner, thus limiting the potential side effects of P-glycoprotein down-regulation. However, it does not address practical uncertainties associated with shifting this strategy to the clinic. First, dPPA is a phorbol ester derivative and there would certainly be problems using this class of drug in the clinic. However, when one becomes available, a cell-permeable, engineered peptide or small molecule PKCβ_I activator could be used. Second, we do not know the central or systemic consequences of PKCβ_I activation. PKCβ isoforms have been implicated in a number of diseases and PKCβ inhibitors, including LY333531, are currently in clinical trials for treatment of microvascular complications of diabetes and of certain cancers (McGill et al, 2006; Sledge and Gokmen-Polar, 2006; Tuttle, 2008). Conversely, the effects of specific activation of PKCβ_I in vivo are unknown. In vitro dPPA has been shown to, e.g., increase glucose transport in isolated muscle fibers (Wright et al, 2004) and inhibit cyclic nucleotide-induced ion transport in colonocytes (Broughman et al, 2006). Finally, P-glycoprotein is expressed in peripheral tissues, such as the liver, kidney, and gut, where it certainly influences drug xenobiotic pharmacokinetics. It is not clear to what extent (if any) transporter activity in those tissues is modulated through PKCβ_I. All of these issues will have to be addressed in preclinical studies.