In Vivo and In Vitro Models of Drug-Drug Interaction Research
Chair: Louis Cantelina, M.D., Ph.D.; Co-Chair: Jerry Collins, Ph.D.
Clinical Issues in Drug-Drug Interaction Studies
Chair: Gerald Friedland, M.D.; Co-Chair: Elinore McCance-Katz, M.D., Ph.D.
Friday, September 10, 1999
Interactions Between Drugs of Abuse and Drug Addiction Treatment Pharmacotherapeutics
Benowitz, Neal L, M.D.: Cigarette Smoking: Drug Interactions
Cigarette smoking is nearly ubiquitous among illicit drug users, and correspondingly in HIV-infected drug abusers. Cigarette smoking can affect illicit drug use and/or drug therapy by both pharmacokinetic and pharmacodynamic mechanisms.
Pharmacokinetic interactions occur primarily by the effects of cigarette smoking on the activity of drug metabolizing enzymes. Components of cigarette smoke that may affect drug metabolism include polycyclic aromatic hydrocarbons, carbon monoxide, nicotine, and heavy metals. Cigarette smoking accelerates metabolism by the liver enzymes CYP1A1, 1A2 and to a lesser extent 2E1, and accelerates glucuronidation by some UGT isoenzymes. Cigarette smoking also appears to inhibit metabolism via CYP2A6, the enzyme primarily responsible for the metabolism of nicotine. Little is known of the effects of cigarette smoking on the metabolism of antiretroviral drugs or other drugs that are used to treat patients with AIDS.
Pharmacodynamic interactions between cigarette smoking and other drugs are less well understood, but possibly more important than the pharmacokinetic interactions. With respect to drugs of abuse, most heroin and cocaine abusers indicate the cigarette smoking enhances the pleasure they get from their illicit drug use and vice-versa. Nicotine could augment neurotransmitter release triggered by various drugs of abuse. Nicotine can also modulate alterations in arousal produced by other drugs. Animal studies indicate cross-sensitization between nicotine and the stimulants cocaine and amphetamine, which could provide a pharmacological basis for the proposition that nicotine is a "gateway" drug for stimulant abuse.
Nicotine addiction must be considered in planning drug abuse therapy. Cigarette smoking may serve as a conditioned stimulus for relapse to drug abuse after drug abuse therapy. Research is needed to ascertain whether drug abuse therapy is facilitated by simultaneous smoking cessation therapy.
Cantilena, Louis R., M.D.: In vitro prediction of the in vivo drug-drug interaction between methadone and protease inhibitors
Methadone is used in maintenance detoxification programs for former intravenous heroin users, many of whom are HIV-infected. CYP3A appears to play a role in the N-demethylation of methadone in vitro. Thus, there is the potential for indinavir (IDV), a CYP3A substrate and inhibitor, to inhibit the metabolism of methadone. In vitro study of the drug-drug interaction between methadone and selected protease inhibitors has demonstrated significant inhibition of methadone biotransformation at pharmacologically relevant concentrations of both substrate and inhibitor. The in vitro data would suggest a strong possibility for a clinical interaction to occur. To determine if the in vitro effect accurately predicted the in vivo outcome, the in vivo effect of indinavir on the pharmacokinetics of methadone was determined. Measurement of methadone and methadone metabolite concentrations with and without concurrent indinavir in a two-period, double blind crossover study in a clinical pharmacology inpatient research unit was conducted. Study design: a multiple dose, 2- period, crossover study in 12 male and female HIV-negative patients on stable methadone maintenance therapy due to narcotic addiction. Subjects received their standard methadone (20 to 60 mg/day) dose administered q.a.m. for 8 doses and in crossover fashion (with washout of at least 10 days) and either: Treatment A: 800 mg IDV q8h for 8 days (24 doses) or Treatment B: IDV placebo q8h for 8 days (24 doses) IDV was administered at 8 am, 4 pm, and midnight each day in a fasted state. A plot of dose-adjusted methadone AUC 0-24 hr vs. methadone dose shows no obvious trend indicating that methadone pharmacokinetics are sufficiently linear over the dose range studied to allow pooled comparison between treatments of dose-adjusted parameters. For methadone, combination /monotherapy geometric ratios (90% CI) were 0.96 (0.86-1.06) for AUC 0-24 and 0.93 (0.84-1.03) for Cmax. For EDDP the corresponding ratios were 1.04 (0.91-1.19) for AUC 0-2 and 0.97 (0.83-1.12) for Cmax. These results indicate that co-administration with IDV has no discernible effect on the pharmacokinetics of methadone or its metabolite. Thus, the in vitro interaction data did not predict the in vivo results under the conditions tested. Due to a lack of significant effect on methadone or methadone metabolite concentrations during the 8-day interaction study, it appears that IDV can be administered with methadone without dose modification of methadone.
Collins, Jerry M, Ph.D.: An FDA Perspective on Metabolic Drug-Drug Interactions
Because high-quality metabolic data have always been generated during some drug development programs, it's possible to overlook the revolutionary nature of changes, which have occurred in the use of human tissue for metabolism. The predominance of idiosyncratic reactions has been supplanted by systematic utilization of prospective predictions and intervention. From the rapid deployment of human tissue in vitro, an overall consensus has emerged regarding the reliability for predictions of human metabolism and drug interactions. Some areas of apparent discord remain between studies in vitro and in vivo. The use of unbound drug concentration could avoid confusion. The relationship between plasma concentration of drug and concentration at the site of the enzyme within the hepatocytes requires further research. Subcellular preparations of liver continue to be useful for screening, but some important limitations need to be appreciated. The excellent work with CYP enzymes needs to be generalized to other metabolic pathways.
Flockhart, David A., M.D., Ph.D.: Clinical Pharmacogenetics in Drug Abuse
Genetic polymorphisms in metabolic enzymes and in receptors that mediate the effects of drugs can have important effects on efficacy and toxicity. This is particularly important in the area of drug abuse where vulnerability to specific patterns of abuse may be increased by specific metabolic genotypes and where rational pharmacotherapy may be reasonably improved by consideration of pharmacogenetic influences.
Friedland, Gerald H., M.D.: Summary of Other Opiate and Antiretroviral Studies
Studies of pharmacokinetic interactions between the nucleoside reverse transcriptase inhibitors (AZT, d4T, ddI) have shown variable and probably clinically important interactions. These drugs are the backbone of many antiretroviral therapy regimens, but not the most potent agents in these regimens. Currently, only anecdotal information exists on the most potent agents- the non-nucleoside reverse transcriptase inhibitors and the protease inhibitors. These drugs are even more likely to have important interactions with methadone because of shared metabolic pathways. Anecdotal information indicates that both nevirapine and efavirenz administration may result in methadone withdrawal. Seven patients who experienced methadone withdrawal soon after being placed on nevirapine have been reported. Clinical symptoms were verified by sub-therapeutic methadone levels in three patients. Escalation of methadone was required in order to continue nevirapine and four of the seven patients discontinued antiretroviral therapy. Nine patients enrolled in a clinical trial of efavirenz who were also on methadone were reviewed retrospectively. Among the seven who received both drugs, a mean increase in methadone of 23 mgs was required during the study. Finally, in a trial of once a day supervised antiretroviral therapy among methadone recipients (nevirapine, ddI and 3TC), almost 50 % of individuals withdrew from the study, many because of methadone withdrawal symptoms and 45% of subjects required increases in methadone doses. These studies suggest an important effect of these agents on methadone and formal interaction studies are greatly needed. Studies with nucleoside and non-nucleoside reverse transcriptase inhibitor therapy indicate that these agents have important interactions with methadone. These interactions may affect efficacy, toxicity and adherence to therapies. Performing formal studies in this area is difficult but essential if the benefits of antiretroviral therapy are to be extended fully to substance abusers with HIV disease. These studies must move forward quickly but multiple questions regarding their performance need to be addressed. As this effort proceeds. Among them are: Must all drugs be studied in vivo or just classes of drugs with reliance on in vitro studies and knowledge of metabolic pathways? Is it possible to study drugs in interaction studies in the combinations now being administered clinically or only as single agents? How can other drugs of abuse be included in these studies: cocaine, heroin, amphetamines, alcohol, marihuana, caffeine and nicotine? And probably most important, what is the clinical relevance of drug interaction study findings? How do these translate to effectiveness of therapies in real world clinical practice and how can this be measured?
Gerber, John G., M.D.: CYP Isozymes in Interactions Between Drugs of Abuse and Antiretrovirals
Both HIV-1 protease inhibitors and the non-nucleoside reverse transcriptase inhibitors (NNRTI), in clinical use for the treatment of HIV infection, are lipophilic molecules that require oxidative metabolism for ultimate elimination. These drugs utilize various isozymes of the cytochrome P450 (CYP) for this purpose, and some of these drugs also inhibit these isozymes, others induce them, and still others do both. The HIV protease inhibitors are large organic bases that utilize mainly CYP3A4 for metabolism but also inhibit this isozyme because of their attachment with high affinity. Ritonavir is the most potent inhibitor of CYP3A4 in this class because of the high attraction to the substrate binding site as well as the electron-rich thioazole side group that competes with oxygen at the iron site of CYP. Ritonavir and nelfinavir, besides inhibiting CYP3A4, also induce this isozyme to induce their own metabolism. Ritonavir also inhibits CYP2D6. Indinavir, amprenavir, and saquinavir are pure inhibitors of CYP3A4. Of the NNRTIs, delavirdine is a large molecule and inhibits CYP3A4 and CYP2C9. It is metabolized by multiple CYPs. Nevirapine, a small molecular weight compound, is a pure inducer of CYP3A4 and CYP2B6 and utilizes both isozymes for metabolism. Efavirenz is also an inducer of CYP3A4 and CYP2B6 but also inhibits CYP2C19 and possibly CYP3A4. In order to predict drug-drug interactions between drugs of abuse and antiretroviral drugs, one needs to understand how both classes of drugs interact with specific CYPs. For example, methadone is metabolized by CYP3A4 in vitro to an inactive metabolite, EDDP, while LAAM is metabolized by CYP3A4 in vitro to the more active metabolites, norLAAM and dinorLAAM. MDMA (Ecstasy) is metabolized by CYP2D6 to an inactive metabolite. In order to understand the pharmacokinetic interactions between drugs of abuse and antiretroviral drugs, studies have to be performed in vitro as well as in vivo so that both groups of drugs can be used safely together.
Greenblatt, David J., M.D.: Interactions of HIV Therapies and Abusable Drugs: Clinical Importance and Research Approaches
Substance abuse among patients with human immunodeficiency virus (HIV) infection is a matter of major public health concern. Some 25-30% of HIV-infected patients have acquired the disease through I.V. drug use; a principal therapeutic objective is to sustain abstinence, both for patientsŐ health and to minimize risk of disease dissemination, through maintenance programs using methadone or buprenorphine. For these and other HIV-infected patients, the emotional burden of the disease produces comorbidity with depression, anxiety, and sleep disorders. Benzodiazepine agonists are commonly prescribed for anxiety and insomnia, and there is concern regarding abuse of this class of drugs. Highly active antiretroviral therapies (HAART), including the HIV protease inhibitors (PIs) and nonnucleoside reverse transcriptase inhibitors (NNRTIs), constitute major advances in the treatment of HIV infection, but greatly complicate the link between HIV/AIDS and substance abuse. HIV PIs and NNRTIs may inhibit and/or induce the activity of human Cytochrome P4503A (CYP3A) enzymes, responsible for the metabolism of buprenorphine, methadone, many benzodiazepines, and for the formation of the hepatotoxic metabolite of cocaine. CYP3A inhibition could promote methadone toxicity, or enhance abusability of benzodiazepines; CYP3A induction could precipitate methadone withdrawal and opiate relapse, or enhance cocaine hepatotoxicity. This presentation will review the clinical and experimental data base on pharmacokinetic and pharmacodynamic interactions of HAART components with potentially abusable drugs, and with pharmacological treatments of drug abuse. In vitro-in vivo correlations are discussed, as well as methodological concepts and problems.
Jatlow, Peter, M.D.: Methodological Issues and Problems in Performing Drug-Interaction Studies with Opiates in Intravenous Drug Users
Certain problems must be recognized and addressed when planning drug interaction studies in intravenous drug users. These issues impact on study design as well as on the recruitment and retention of subjects. The obstacles to employing a within subject design for studies of the effects of opioid drugs used to treat opiate abuse on the pharmacokinetics of antiretroviral drugs may be overwhelming. The resources and logistic difficulties and recruitment disincentives when subjects must be detoxified prior to performing the antiretroviral alone phase may preclude a within subject design. The alternative between subject design, with its larger intersubject CVŐs and lesser statistical power, necessitate a larger sample size, and an opiate free control group which may not be optimally matched. On the other hand, a within subject design is feasible and desirable when evaluating the effects of antiretrovirals on opiates. Venous abscesses are a major problem, which can disqualify otherwise eligible subjects and/or interrupt the progress of a pharmacokinetics study. Other considerations include the high incidence and confounding consequences of poly drug use, including tobacco, by drug abusers, the need for closed wards and specially trained and experienced support personal, the often chaotic life style of intravenous drug users and the high incidence of undercurrent disease. Despite these several constraints, such studies can and should be done. However it is strongly recommended that the research team include one or more experienced drug abuse clinical researchers.
McCance-Katz, Elinore, M.D., Ph.D.: Enhancing the Clinical Care of Drug Abusers with HIV Disease: Drug Interactions and Implications for Treatment
Nationwide, approximately 30% of people with AIDS are injection drug abusers, and current data indicates that over half of new AIDS cases in Northeast urban areas occur in this population. Injection drug users represent the major route for heterosexual and perinatal transmission of HIV. While these infected drug abusers are eligible for antiretroviral therapy, only a small percentage actually receive treatment in part because of concern regarding the potential for drug interactions in opioid maintained injection drug abusers and out of fear of patient non-adherence to complicated clinical regimens and the generation of antiretroviral resistant HIV. There is limited information available on the extent and clinical implications of drug interactions in substance abusers with HIV disease. However, several studies have been published to date. Methadone has been shown to significantly increase zidovudine levels. Interim results from a study of the interaction of zidovudine with LAAM, buprenorphine, or naltrexone indicates that LAAM does not affect zidovudine levels, while buprenorphine may decrease zidovudine levels nonsignificantly. Preclinical studies have shown that protease inhibitors inhibit the metabolism of drugs utilizing CYP 450 3A and 2D enzymes. However, a recent case report indicates that nelfinavir may induce these enzymes. The non-nucleoside reverse transcriptase inhibitors nevirapine and efavirenz induce CYP 450 enzymes important to methadone metabolism resulting in the precipitation of withdrawal symptoms in patients treated with these drugs. The consequences of undetected drug interactions include patient non-adherence with treatment regimens, development of resistant viruses, increased illicit drug abuse, drug toxicity, and lack of efficacy. It will be important to further define the drug interactions of importance to the treatment of the substance abusing population with HIV disease. Studies are ongoing toward this end.
Morse, Gene D., Pharm. D.: Influence of HIV Infection on the design and interpretation of drug-drug interaction studies
The University at Buffalo Laboratory for Antiviral Research has been conducting clinical drug-drug interaction studies for over 10 years. In a parallel series of studies we have been interested in evaluating the influence of HIV infection on antiviral pharmacokinetics including issues such as reduced gastric pH, menstrual cycle effects, cytokine patterns and renal function. In an additional series of studies conducted with the adult AIDS Clinical Trials Unit, drug interaction studies between indinavir-efavirenz, indinavir-nelfinavir and efavirenz-nelfinavir have been completed. Data from these studies will be discussed.
Rainey, Petrie, M.D., Ph.D.: Studies with Other Nucleoside Agents and Methadone.
When didanosine or stavudine were administered to 16 subjects on methadone maintenance, the mean AUCŐs of the nucleosides were reduced by approximately 60% (p = 0.11) and 25% (p=0.02), respectively, in comparison with values in 10 controls. These changes appeared to be largely due to decreases in bioavailability. Absorption was delayed, suggesting increased spontaneous or metabolic degradation during increased residence time of the drugs in the gastrointestinal tract. Methadone was previously observed to delay absorption of zidovudine. In contrast, zidovudine, didanosine, stavudine, or lamivudine did not have a significant effect on methadone disposition.
Unadkat, Jashvant, Ph.D.: Prediction of in-vivo Drug Interactions of Anti-HIV and Anti-Opportunistic Infection Drugs from in-vitro Data.
The focus will be on P450 and P-glycoprotein interactions. Studies were conducted to determine if clinically significant drug interactions between anti-HIV and anti-opportunistic (anti-OI) infection drugs can be predicted from in vitro studies in human liver microsomes. The inhibitory capacities of anti-HIV and anti-OI drugs towards drug metabolizing cytochrome P450 (CYP) enzymes were determined in a panel of human liver microsomes using isoform selective probes. As expected, the protease inhibitors (PIs) were found to inhibit CYP3A4/5. Using the in vitro Ki of PIs towards CYP3A4/5, the extent to which the PIs will inhibit the in vivo clearance of other CYP3A4/5 substrates was predicted. The rank order of in vivo inhibitory capacities of the PIs was better predicted from in vitro data if the total rather than the unbound plasma concentrations of the PIs were used. Similarly, in another study, the inhibition of in vivo hydroxylamine formation of both dapsone and sulfamethoxazole by fluconazole could be semi-quantitatively predicted from in vitro data in human liver microsomes. In summary, clinically relevant in vivo drug interactions can be predicted from in vitro studies in human liver microsomes. Experimental factors (e.g. protein binding) critical to the ability of human liver microsomes and hepatocytes to semi-quantitatively predict other clinically relevant metabolic drug interactions should be elucidated.
Recommendations for Future Research
In terms of global issues, the panel made a strong plea that NIH/NIDA should establish a special review committee (study section) or bring the necessary expertise to an existing committee to review this type of research, and provide sufficient funds to support research in this new area.
Neal Benowitz, M.D., Ph.D.
Louis R. Cantilena, Ph.D., Director
Jerry Collins, Ph.D.
David A. Flockhart, M.D., Ph.D.
Henry L. Francis, M.D.
Gerald H. Friedland, M.D.
John G. Gerber, M.D.
David J. Greenblatt, M.D.
Peter Jatlow, M.D.
Jag H. Khalsa, Ph.D.
Deborah B. Leiderman, M.D.
Alan I. Leshner, Ph.D.
Elinore McCance-Katz, M.D., Ph.D.
Gene D. Morse, Pharm. D.
Petrie Rainey, M.D., Ph.D.
Jashvant Unadkat, M.D./Ph.D.
Frank Vocci, Ph.D.