Modulation of glucocorticoid receptor in human epileptic endothelial cells impacts drug biotransformation in an in vitro blood-brain barrier model

Abstract

Objective: Nuclear receptors and cytochrome P450 (CYP) regulate hepatic metabolism of several drugs. Nuclear receptors are expressed at the neurovascular unit of patients with drug-resistant epilepsy. We studied whether glucocorticoid receptor (GR) silencing or inhibition in human epileptic brain endothelial cells (EPI-ECs) functionally impacts drug bioavailability across an in vitro model of the blood-brain barrier (BBB) by CYP-multidrug transporter (multidrug resistance protein 1, MDR1) mechanisms.

Methods: Surgically resected brain specimens from patients with drug-resistant epilepsy, primary EPI-ECs, and control human brain microvascular endothelial cells (HBMECs) were used. Expression of GR, pregnane X receptor, CYP3A4, and MDR1 was analyzed pre- and post-GR silencing in EPI-ECs. Endothelial cells were co-cultured with astrocytes and seeded in an in vitro flow-based BBB model (DIV-BBB). Alternatively, the GR inhibitor mifepristone was added to the EPI-EC DIV-BBB. Integrity of the BBB was monitored by measuring transendothelial electrical resistance. Cell viability was assessed by glucose-lactate levels. Permeability of [³ H]sucrose and [¹⁴ C]phenytoin was quantified. CYP function was determined by measuring resorufin formation and oxcarbazepine (OXC) metabolism.

Results: Silencing and inhibition of GR in EPI-ECs resulted in decreased pregnane X receptor, CYP3A4, and MDR1 expression. GR silencing or inhibition did not affect BBB properties in vitro, as transendothelial electrical resistance and P_sucrose were unaltered, and glucose metabolism was maintained. GR EPI-EC silencing or inhibition led to (1) increased P_phenytoin BBB permeability as compared to control; (2) decreased CYP function, indirectly evaluated by resorufin formation; (3) improved OXC bioavailability with increased abluminal (brain-side) OXC levels as compared to control.

Significance: Our results suggest that modulating GR expression in EPI-ECs at the BBB modifies drug metabolism and penetration by a mechanism encompassing P450 and efflux transporters. The latter could be exploited for future drug design and to overcome pharmacoresistance.

Keywords: P450 enzymes; bioavailability; drug-resistant epilepsy; multidrug transporter; nuclear receptors.

Figure 1

GR silencing in EPI‐EC and BBB integrity. A, ECs isolated from drug‐resistant epileptic brain resections were transfected (EPI‐EC GR siRNA; n = 6). EPI‐EC non‐siRNA and control human brain microvascular endothelial cells (HBMECs) were compared. Immunocytochemistry shows successful GR silencing by a decrease in GR levels in EPI‐ECs. Decreased PXR, CYP3A4, and MDR1 were observed in EPI‐EC GR siRNA. A1, Quantification of fluorescent signals showed a significant decrease in expression levels of GR (P ≤ 0.01), PXR (P ≤ 0.05), CYP3A4 (P ≤ 0.01), and MDR1 (P ≤ 0.01) in EPI‐EC GR siRNA (mean ± SEM). B, Transendothelial electrical resistance (TEER) measurement showed development of a barrier characterized by an increase in TEER (~600‐800 Ω/cm²) normalized with the baseline TEER reading. No statistical difference was found among TEER readings in EPI‐EC GR siRNA, EPI‐EC non‐siRNA, and HBMEC DIV‐BBB. All experiments were repeated in triplicate. C, The cellular metabolic pattern within the DIV‐BBB setup showed no significant difference in the pattern of glucose and lactate levels over time during the DIV‐BBB formation

Figure 2

GR inhibition in EPI‐ECs and BBB integrity. A, EPI‐ECs were treated with the GR inhibitor mifepristone, MF (n = 4 donors in triplicate). Untreated EPI‐ECs (n = 4 donors in triplicate) and control HBMECs (n = 4 in triplicate) were compared. Decreased levels of PXR, CYP3A4, and MDR1 were observed in EPI‐EC GR inhibitor compared with EPI‐EC no inhibitor. A1, Quantifications are provided (mean ± SEM). B, No statistically significant difference in TEER readings was found between EPI‐EC GR inhibitor, EPI‐EC no inhibitor, and HBMEC DIV‐BBB modules

Figure 3

Permeability of [³H]sucrose and [¹⁴C]phenytoin in EPI‐EC GR siRNA, EPI‐EC non‐siRNA, and HBMEC DIV‐BBB. A‐A2, The tightness of the barrier evaluated by [³H]sucrose (a paracellular marker) shows no significant difference across the endothelium(s) used. B‐B2, [¹⁴C]phenytoin showed a 2.1‐fold increase in permeability levels (1.08 × 10⁻⁷ cm/s) in EPI‐EC GR siRNA DIV‐BBB (P ≤ 0.05) compared with EPI‐EC non‐siRNA BBB (5.16 × 10⁻⁸ cm/s). The permeability pattern in EPI‐EC GR siRNA at the DIV‐BBB was comparable to that of HBMEC DIV‐BBB (1.23 × 10⁻⁷ cm/s)

Figure 4

Permeability of [³H]sucrose and [¹⁴C]phenytoin evaluated in EPI‐EC GR inhibition (MF), EPI‐EC no inhibition, and HBMEC DIV‐BBB. A‐A2, [³H]sucrose permeability was comparable across experimental conditions. B‐B2, [¹⁴C]phenytoin permeability was higher (1.20 × 10⁻⁷ cm/s) in EPI‐EC GR‐inhibited DIV‐BBB (P ≤ 0.05) compared with EPI‐EC untreated BBB (4.90 × 10⁻⁸ cm/s)

Figure 5

CYP drug biotransformation is supported by resorufin and OXC metabolism in EPI‐ECs after GR silencing. A, Formation of resorufin is the highest when using EPI‐EC non‐siRNA. GR silencing resulted in decreased resorufin conversion comparable to that of control DIV‐BBB (HBMECs). B, Similar results were obtained when analyzing the abluminal compartment. C, OXC metabolism is increased in EPI‐EC non‐siRNA in the luminal side compared with EPI‐EC GR siRNAs and HBMECs. C1, A relative increase in OXC penetration across the BBB was observed in EPI‐EC GR siRNA and HBMECs compared with EPI‐EC non‐siRNA DIV‐BBB. C2, Elevated LiCBZ levels in the abluminal side of EPI‐EC non‐siRNA suggest that OXC is metabolized by a GR‐CYP pathway. Such metabolism is lower in EPI‐EC GR siRNA or HBMECs. D, The levels of resorufin positively correlated with the amount (micrograms per milliliter) of OXC metabolized. Results are expressed as mean ± SEM (by analysis of variance). Asterisks indicate P ≤ 0.05

Figure 6

CYP drug biotransformation is supported by resorufin and OXC metabolism in EPI‐ECs after GR inhibition. A, Resorufin formation was the highest when using untreated (no MF) EPI‐ECs compared with controls. B, Similar results were obtained when analyzing the abluminal compartment of the DIV‐BBB modules. C, OXC metabolism was increased in noninhibited EPI‐EC compared with EPI‐EC GR‐inhibited and HBMECs. C1, A relative increase in OXC penetration across the BBB was observed when using EPI‐EC GR‐inhibited and HBMECs compared with EPI‐EC non‐inhibited DIV‐BBB. C2, LiCBZ formation in the abluminal side of noninhibited EPI‐ECs, indicating OXC metabolism by a GR‐CYP mechanism. OXC metabolism was lower or negligible in EPI‐ECs after GR inhibition or in HBMECs. D, Levels of resorufin positively correlated with the amount of OXC metabolized. Results are expressed as mean ± SEM. Asterisks indicate P ≤ 0.05