The Discovery And Significance Of A Tissue-specific Proteasome Inhibitor-gambogic Acid

Intracellular protein degradation is mainly carried out by two pathways: one is thelysosome pathway, which degrades membrane-associated proteins and long-lived proteins; theother is the ubiquitin-proteasome system (UPS), this system mediates the degradation of morethan80%intracellular protein, including abnormal proteins and short-lived proteins. UPSregulates cellular processes, such as cell cycle, apoptosis, gene expression, inflammation,immune regulation, protein quality control. Since many diseases, especially some malignanttumors, are found to have abnormal proteasome activity, and therefore, proteasome inhibitionhas emerged as a novel approach to anticancer therapy. Velcade (Vel) as the first proteasomeinhibitor anticancer drug has been approved by US FDA for the treatment of multiplemyeloma, but its effectiveness and administration have been limited by toxic side effects,including peripheral neuropathy, low platelet, gastrointestinal events, fatigue, and suppressionof the hematopoietic and immune system. As the proteasome exists in all the cells, specificproteasome inhibitors would definitely inhibit the proteasome function to some extent.Therefore, one important approach is to discover new proteasome inhibitors with efficientanti-tumor effects and cell-specific proteasome inhibition to decrease the toxic side effects.Gambogic acid (GA) is a natural compound derived from Chinese herbs. Recent studies have demonstrated that GA has anticancer effects and inhibits the growth of multiple types ofhuman cancer cells in vitro and in vivo, with little toxicity on immune and hemopoieticsystems and it has been approved by the Chinese FDA for the treatment of different cancers inclinical trials. Some potential molecular targets of GA have been reported that may contributeto its cytotoxicity and anticancer activity, including binding to the transferrin receptor andsuppressing nuclear factor-κB (NF-κB) signaling pathway, and inhibiting VEGFR2, but itsmolecular targets have not been thoroughly studied. Thus, identification of the specificmolecular targets responsible for GA-mediated anti-cancer effect should have great clinicalsignificance. Here we studied the specific mechanism of GA and carry out the followingtrials:1. GA indirectly inhibits proteasome activitiesMost recently we have reported that the combination of GA with the classic proteasomeinhibitor MG132or MG262synergistically inhibited tumor cell growth and induced cell death.To elucidate the involved molecular mechanism, we measured the levels of the proteasomeinhibition, we found that GA not only increased the levels of ubiquitinated proteins inducedby MG132or MG262, but also GA alone caused accumulation of both ubiquitinated proteinsand GFP dose-dependently, confirming that GA is able to inhibit the cellular proteasomeactivity. But it was found that GA at up to5μΜ failed to inhibit the chymotrypsin (CT)-likeactivity of the purified20S proteasome. Only at10-50μΜ doses, GA exhibited a partialinhibition (IC50~25μΜ). These results demonstrated that GA itself is not a potent proteasomeinhibitor.2. Computational modeling studies predict MT1has the potential to interact with theproteasomal β5chymotryptic subunitThe fact that the proteasome inhibition activity of GA was high in the cell compared to inpurified proteasome system suggests that one or more cellular metabolites of GA are potential proteasome inhibitors and that GA serves as a pro-drug of a proteasome inhibitor. Consistently,the computational modeling studies predict that the GA metabolite MT1but not GA and MT2has the potential to interact with and inhibit the proteasomal β5chymotryptic subunit.3. MT1and CYP2E1are responsible for GA-induced proteasome inhibition in tumor cellsTo provide experimental evidence for MT1as a direct proteasome inhibitor, MT1waschemically synthesized and purified, and it inhibited the purified20S proteasomal CT-likeactivity with IC50value of~0.5μΜ, while in a sharp contrast with GA (IC50:~25μΜ). Wethen determined which isotype of CYP enzymes is responsible for metabolizing GA to MT1and whether inhibition of this CYP enzyme could mitigate GA-induced proteasome inhibitionby using a cell-based peptidase assay. We found that only diethyldithiocarbamate (DDC), aCYP2E1inhibitor, dramatically rescued GA-induced proteasome inhibition, suggesting thatCYP2E1may be responsible for metabolizing GA into MT1. Indeed, DDC rescuedGA-induced proteasome inhibition in K562, Jurkat T and P388cells. We further found thatCYP2E1siRNA was able to partially decrease the CYP2E1protein in human HepG2cells,associated with decreased levels of CT-like activity inhibition by GA. In addition, it wasfound the level of CYP2E1was different between cancer cell lines with normal cell (hMSC).4. GA selectively inhibits cellular proteasomal CT-like activityFurthermore, DDC was also able to suppress GA-induced proteasome inhibition in HepG2cells. We have also noticed that GA only slightly inhibits the proteasomal caspase-like activityand dose not have any effect on the proteasomal trypsin-like activity, indicating that GA (orMT1) selectively inhibits cellular proteasomal CT-like activity.5. GA has a similar gene expression profile with VelIt has been reported that proteasome inhibition could induce typical gene expression profilein many cancer cell lines. We then compared the gene expression profiles between GA andVel treatment. We found that GA and Vel yielded not only a similar gene expression profile but also the similar proteasome inhibition-specific genes.6. Proteasome inhibition induced by the metabolite produced by CYP2E1is required for GA’scytotoxicityWe next determined whether proteasome inhibition contributes to GA-induced cytotoxicity.We found that inhibition of CYP2E1by DDC not only partially rescued GA-inducedproteasome inhibition but also inhibited GA-induced cell death. Our further study found thatCYP1A2inhibitior-α-naphthoflavone (ANF) treatment caused a significant increase the levelsof ubiquitinated proteins induced by GA, thereby significantly enhanced GA-induced celldeath.7. Silencing CYP2E1partially rescued GA-induced cell death whereas silencing CYP1A2enhanced GA-induced cell deathTo further confirm that the cell death induction by GA is due to CYP2E1, CYP2E1-andCYP1A2-siRNA were used to silence CYP2E1or CYP1A2, respectively. We found that,similar to proteasome inhibition, silencing CYP2E1partially inhibited GA-induced cell deathwhereas silencing CYP1A2enhanced GA-induced cell death. These results clearly showedthat GA-induced cytotoxicity relies on its proteasome-inhibitory activity, which is mediatedmainly by CYP2E1and its metabolite MT1.8. GA induces proteasome inhibition and cytotoxicity which is reversed by DDC treatment inNCI-H929and U266myeloma cancer cellsSince Vel as the first proteasome inhibitor anticancer drug has been approved by US FDAfor the treatment of multiple myeloma. Next we study the effect of GA in NCI-H929andU266myeloma cancer cells. We found that GA-induced proteasome inhibition andcytotoxicity could be partially reversed by DDC-mediated CYP2E1inhibition in myelomacancer cells. 9. ER stress responses associate with cytotoxicity induced by GAWe also found that similar to Vel, GA was able to induced ER stress, as measured byincreased levels of ER stress-related proteins, which demonstrated that similar to Vel, GAinduced the ER stress responses which are associated with proteasome inhibition-mediatedcytotoxicity.10. The anti-tumor effect of GA treatment increases the survival of P388-bearing mice andinhibits H22tumor growth and proteasome function in vivoWe next determined the anti-cancer effect of GA in vivo. We found that GA treatmentsignificantly improve the survival of mice bearing P388tumors, associated with proteasomeinhibition at early hours and inhibits H22solid tumor growth. Next we found that, GAselectively inhibited proteasome function of H22solid tumors, but not muscle tissue. GA andVel, both could inhibit proteasome function in tumor and liver tissue, and selectively inducePARP cleavage in tumor tissue, but had no effect on muscle tissue. To further test whether GAis a tissue-specific proteasome inhibitor in vivo, we detected GA distribution in some of therelevant normal tissues after GA injection in mice. It was found that GA could be detected inliver, muscle and spleen tissues, similar to Vel. We found that Vel, but not GA, coulddramatically accumulate ubiquitinated proteins. These results have further demonstrated thatGA induces tissue-specific, such as tumor tissue-specific proteasome inhibition, compared toVel.11. The reduced form of GA fails to induce proteasome inhibition and cytotoxicityTo investigate the requirement of the C9-C10double bond of GA for proteasome inhibition,C9-C10-disrupted GA (GA~), a reduced form of GA, was chemically synthesized. It was foundGA~lost its ability to inhibit proteasome activity and induce cytotoxicity. These resultsclearly demonstrated that GA-induced cytotoxicity and cell death definitely depend on theexistence of C9-C10double bond, which is required for its mediated proteasome inhibition. 12. GA does not decrease lymphocyte survival in CYP2E1-deficient peripheral blood cells invitro and in vivoTo further study the importance of CYP2E1in mediating GA-induced proteasomeinhibition and cytotoxicity, we detected the effect of peripheral blood cells induced by GA andVel. And found that only Vel but not GA at the tested doses inhibits the proteasomal CT-likeactivity in the peripheral blood cells by in vitro peptidase assay, and found that GA did notinhibit CT-like activity in whole blood cell culture either. Also, GA did not show any effectson the survival of blood cells during7days of culture while Vel inhibited white blood cell andlymphocyte survival. To further confirm whether therapeutic dose of GA could affect whiteblood cell number, the CYP2E1and CYP1A2protein distribution in mouse and human bonemarrow cells were compared with cancer cell lines. It was found that both normal mouse andhuman bone marrow cells weakly express CYP2E1protein, indicating an inability for thebone marrow cells to metabolize GA.13. GA induces more cytotoxicity and proteasome inhibition in cancer cells from leukemicpatients than in human peripherial mononuclear cellsNext we further compared the effects of GA on cytotoxicity and proteasome inhibition incancer cells obtained from10leukemia patients and in peripheral mononuclear cells from6normal volunteers. It was found that Vel markedly induced cytotoxicity or accumulation ofubiquitinated proteins and PARP cleavage both in leukemic cells and normal mononuclearcells. While GA at all the doses dramatically induced cytotoxicity or accumulation ofubiquitinated proteins and PARP cleavage in leukemic cells, but only slightly induced thesechanges in normal cells. These results demonstrated that GA, compared to Vel, selectivelyinduced proteasome inhibition and cytotoxicity in leukemic cancer cells. Conclusions: First, GA acts as a prodrug of a proteasome inhibitor and only gainsproteasome-inhibitory function after being metabolized by intracellular CYP2E1; Secondly,GA-induced proteasome inhibition is a prerequisite for its cytotoxicity and anti-cancer effectwithout off-targets; Finally, because expression of CYP2E1gene is very high in tumor tissuesbut low in some of the normal tissues, GA could therefore produce tissue-specific proteasomeinhibition and tumor-specific toxicity, providing great clinical significance for designingnovel strategies for cancer treatment.