CHIP and gp78-mediated ubiquitination of CYP3A4

The cytochrome P450 (CYP) isoform 3A4 (CYP3A4) mediates the biotransformation of more than half of the pharmacological armamentarium,¹ including widely prescribed anticancer agents such as tamoxifen, docetaxel and imatinib.² CYP3As account for ∼30% of liver CYP content,³ and like most proteins they have a limited life span. In order to maintain sufficient levels of properly folded and functional enzymes, cellular degradation processes are activated to proteolyze and recycle CYP proteins.⁴ It was previously demonstrated that members of the CYP3A family of enzymes (either native or inactivated) are substrates for ER-associated degradation (ERAD), which includes phosphorylation, ubiquitination, extraction from the ER into the cytosol, and finally degradation by the 26S proteasome.^5–13 Therefore, although it remains poorly explored, intracellular CYP3A content is likely to depend heavily on ERAD; this process may be behind much of the variation in the pharmacology of anticancer agents that are also CYP3A4 substrates.

The key factor in ERAD, ubiquitin, is a polypeptide that attaches to substrate proteins by covalently attaching its carboxy terminus to the ε-amino group of lysines on ER-associated proteins, thereby fating them for degradation. Before ubiquitin can be conjugated to proteins, it must be activated by E1 proteins that form a thioester complex (E1-Ub). This is followed by forming a complex with E2 (ubiquitin-conjugating proteins; Ubc) and/or E3 (ubiquitin-ligase proteins; Ub-ligase) proteins that attach activated ubiquitin to the substrate protein via transthiolyation.¹⁴ It was previously reported that CYP3A4 is targeted for ERAD in yeast by E2 (UBc7p/Cue1p)-dependent Ub-ligases, but not E3 Ub-ligases.^15–17 The process of ERAD for CYPs is conserved between yeast, mammals and most eukaryotes;¹⁸ however, minor sequence differences may have accounted for varying CYP specificity for members of the mammalian compared to yeast E3 Ub-ligases. The tumor autocrine motility factor receptor (AMFR) or glycoprotein 78 (gp78), was discovered by Fang et al. to be an E3 Ub-ligase, with subsequent domain structure elucidation by Chen et al.²⁰ Furthermore, Correia et al.¹⁰ previously demonstrated that the mammalian E3 Ub-ligases gp78 and C-terminus of Hsp70-interacting protein (CHIP) are involved in ERAD of CYP3A4 in vitro.

By infecting rat hepatocytes with vectors containing short hairpin RNA (shRNA) strands that transcriptionally inactivate gp78 and CHIP genes, Correia et al. recently confirmed that gp78 and CHIP are relevant mammalian E3 Ub-ligases that target CYP3A4 in vivo.¹⁰ Two slightly different shRNAs (shRNA-1 or shRNA-2) significantly decreased gp78 mRNA levels and gp78 protein content in primary rat hepatocytes. The cellular CYP3A4 protein content was increased in this model when dexamethasone (a CYP3A4 inducer) was added. The results were similar for four different shRNAs directed towards different exons of CHIP (i.e., CHIP-1, -2, -3 and -4) in the presence of dexamethasone. The present results suggest for the first time that hepatocellular CYP3A content is regulated by gp78 and CHIP.

Correia et al. next confirmed that gp78 and CHIP are CYP3A-ubiquitin ligases and elucidated the mechanism behind ERAD of CYP3A4 through gp78. Ubiquitin immunoblotting analyses on rat hepatocyte lysates after infection with the aforementioned vectors designed against gp78 and CHIP demonstrated that CYP3A ubiquitination was significantly decreased in all cases. Moreover, when an inducer of CYP ERAD (i.e., 3,5-dicarbethoxy-2,6-dimethyl-4-ethyl-1,4-dihydropyridine; DDEP) was added to the above model along with dexamethasone only, the baseline level of CYPs decreased due to ERAD induction. However, when shRNA-1/2 was added to dexamethasone and DDEP, CYPs were significantly increased to levels comparable to the absence of DDEP. This suggested that gp78 plays a critical role in CYP ubiquitination resulting in ERAD.

The authors next ascertained which gp78 domains were most important for CYP3A4 ERAD. HepG2 cells were infected with vectors containing CYP3A and either wild-type gp78 (gp78-WT), a RING finger mutant (gp78-RM), or just the cytosolic portion of gp78 (gp78-C) that contains the RING finger Ub-ligase and all relevant binding sites for E3 activity. Controls using no gp78 vector or an empty gp78 vector demonstrated baseline CYP3A levels. When the vector containing gp78-WT was co-infected with the vector containing CYP3A, a significant reduction in CYP3A content was observed. However this was abrogated when using vectors containing gp78-RM, suggesting that the RING finger Ub-ligase is a functionally important part of gp78. Surprisingly, when the vector containing gp78-C was co-infected with CYP3A, a very significant decrease in CYP3A content was observed that surpassed gp78-WT. This suggested that the cytosolic portion of gp78 is even more active in ERAD for CYPs than the native polytopically ER-bound gp78, and is consistent with a previous study that demonstrated ERAD activity for the C-terminus of gp78.²¹ The results were similar for vectors containing CYP3A and either no vector, an empty vector or a vector containing CHIP. As expected, the vector-containing CHIP significantly decreased levels of CYP3A, further suggesting a role for CHIP in ERAD for CYPs.

The authors then obtained in situ evidence of changes in CYP3A ubiquitination in the absence of gp78 or CHIP due to shRNA knockdowns. This was accomplished via infection with vectors containing the shRNAs into rat hepatocytes, followed by confocal immunofluorescence microscopy to detect and quantify a fluorescent antibody for CYP3A. After infection with shRNAs, cell morphology was undisturbed; however, CYP3A levels were significantly elevated, which corresponds with previous experiments performed by this group.

Next, the authors determined the intracellular location of the CYP3As over time using pulse-chase experiments. Under normal ERAD conditions, parent CYP3A was found in the ER and Ub-CYP3A in both the ER and cytosol, where it would subsequently be degraded by the 26S proteasome. When shRNAs for gp78 or CHIP were added, parent CYP3A in the ER was significantly elevated compared to normal ERAD conditions. Also, the amount of Ub-CYP3A in ER and cytosol was significantly decreased, further suggesting the vital roles gp78 and CHIP have in ERAD for CYP3As. CHIP actually demonstrated superior activity compared to gp78, although this does not necessarily mean CHIP is more relevant or active than gp78 and could have been due to experimental conditions.

A final experiment performed examined the functional activity of CYP3As in rat hepatocytes that were stabilized due to the knockdown of gp78 or CHIP. CYP3A4 activity was assessed by measuring the fluorescence of the 7-benzyloxy-4-trifluoromethylcoumarin (BFC) metabolite (HFC) after a knockdown of gp78 or CHIP. A time-dependent increase in the amount of the fluorescent metabolite HFC was measured in the culture medium, suggesting extracellular transport after biotransformation by CYP3A4. In rat hepatocytes that were infected with shRNAs for gp78 or CHIP, a 2.2 fold- and 2.8 fold-increase in HFC was observed, respectively. This suggested that after knockdown of ERAD-associated proteins, elevated levels of CYP3As were still functional.

The present results clearly show that both gp78 and CHIP play critical roles in ubiquitination and subsequent ERAD of CYP3As, suggesting that they may play important roles in regulating the hepatocellular expression of functional CYP3As in vivo. This may have significant pharmacological significance given that there is ∼20–60% unexplained inter-individual variation in hepatocellular CYP3A content or activity that significantly complicates the administration of anticancer therapies.²² Interestingly, it was reported that ∼90% of the variability in CYP3A activity is due to genetic factors.²² A recent study explored the genetic contribution to variability in CYP3A activity, focusing on variation in those factors that regulate CYP3A expression or often contribute to CYP3A-mediated substrate clearance such as: FoxA2, FoxA3, HNF4α, PXR, ABCB1 and the CYP3A4 promoter.²³ Accounting for these factors, along with gender, explained approximately 25% of the variation in hepatic CYP3A expression; however, none have yet studied potential genetic variants that may influence CYP-degradation pathways that may explain even more variability in CYP3A expression and activity.

These results may also be very important in anticancer drug pharmacology and therapy as anticancer drugs have a narrow window in which therapeutic aspects are balanced with the toxicity that these drugs induce.²⁴ Thus, inter-individual variation may make these drugs less efficacious in some and highly toxic in others. For instance, mutations that alter the RING finger moiety of gp78 (albeit artificially-induced) were shown by Correia et al. to be a vital factor in ERAD of CYP3As, resulting in the accumulation of hepatocellular CYP3As and increasing hepatic CYP3A4 metabolism. Should functional germline polymorphisms in vital gp78 domains be discovered, these genetic variables could alter the exposure of drugs that are directly inactivated by CYP3A4 (i.e., docetaxel),²⁵ or interfere with the formation of active drug metabolites where CYP3A4 participates in drug activation (i.e., endoxifen formation from tamoxifen),²⁶ or competes with drug activation (i.e., SN-38 formation from irinotecan).²⁷ The pharmacological consequences of CYP3A variation are further complicated by the high incidence of comorbidity in patients with cancer who take many drugs that potentially interact. The present results could have serious implications regarding drug-drug interactions (DDIs) because CYP3A4 contributes to metabolism of at least half of all pharmacologically-relevant xenobiotics (reviewed in ref. 1 and 28). For example, the importance of CYP3A4 levels regarding DDIs was underscored when imatinib was co-administered with a CYP3A4 inducer (phenytoin). The 24-hour exposure (AUC_{0–24 h}) of imatinib was reduced to a fifth of the exposure when imatinib was administered alone, whereas when the CYP3A4 inhibitor ketoconazole was co-administered with imatinib, exposure was significantly increased.²⁹ Future studies must explore genetic variability in the ERAD pathway and identify new factors that influence CYP3A ERAD. As our understanding of CYP3A ERAD develops, it will hopefully lead to a better understanding of CYP3A variability in drug pharmacology, especially in patients treated with anticancer agents.