|Home | About | Journals | Submit | Contact Us | Français|
A major dose-limiting side effect of human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) chemotherapies, such as the nucleoside reverse transcriptase inhibitors (NRTIs), is a small-fiber painful peripheral neuropathy, mediated by its mitochondrial toxicity. Co-morbid conditions may also contribute to this dose-limiting effect of HIV/AIDS treatment. Alcohol abuse, which alone also produces painful neuropathy, is one of the most important co-morbid risk factors for peripheral neuropathy in patients with HIV/AIDS. Despite the prevalence of this problem and its serious impact on the quality of life and continued therapy in HIV/AIDS patients, the mechanisms by which alcohol abuse exacerbates highly active antiretroviral therapy (HAART)-induced neuropathic pain has not been demonstrated. In this study, performed in rats, we investigated the cellular mechanism by which consumed alcohol impacts antiretroviral-induced neuropathic pain. NRTI 2',3'-dideoxycytidine (ddC) (50 mg/kg) neuropathy was mitochondrial dependent and PKCε independent, and alcohol-induced painful neuropathy, PKCε dependent and mitochondrial independent. At low doses, ddC (5 mg/kg) and alcohol (6.5% ethanol diet for one week), which alone do not affect nociception, together produce profound mechanical hyperalgesia. This hyperalgesia is mitochondrial dependent but PKCε independent. These experiments, which provide the first model for studying the impact of co-morbidity in painful neuropathy, support the clinical impression that alcohol consumption enhances HIV/AIDS therapy neuropathy, and provide evidence for a role of mitochondrial mechanisms underlying this interaction.
Alcohol abuse is one of the most important co-morbid risk factors for peripheral neuropathy in patients being treated for human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) (Moyle & Sadler, 1998; Nath et al., 2002; Lopez et al., 2004; Nicholas et al., 2007). Despite the prevalence of this problem and its serious impact on quality of life and ability to continue treatment, the mechanisms by which alcohol abuse exacerbates highly active antiretroviral therapy (HAART)-induced neuropathic pain has not been investigated. To create a foundation for the development of rational therapeutic strategies to treat alcohol-exacerbated neuropathic pain in HIV/AIDS patients, we investigated the cellular mechanisms by which consumed alcohol aggravates antiretroviral-induced neuropathic pain. We employed well-established, clinically relevant, rodent models of HIV/AIDS therapy-induced painful peripheral neuropathy (Joseph et al., 2004; Joseph & Levine, 2004; 2006), and neuropathic effects of alcohol abuse and withdrawal (Dina et al., 2000; Dina et al., 2006) to create a model for their co-morbidity, and to evaluate the underlying mechanism.
The experiments were performed on adult male Sprague–Dawley rats (200–220 g, Charles River, Hollister, CA, USA). Animals were housed in the Laboratory Animal Resource Center of the University of California, San Francisco, under a 12-h light/dark cycle. All experimental protocols were approved by the UCSF Institutional Animal Care and Use Committee (IACUC), and conformed to NIH guidelines for the care and use of experimental animals. Effort was made to limit the numbers of animals used and their discomfort.
The chemicals used in this study were: the broad spectrum caspase inhibitor Z-Val-Ala-Asp(OMe)-fluoro methyl ester (Z-VAD-FMK, R&D Systems, Minneapolis, MN); the antioxidant α-lipoic acid, the mitochondrial respiratory complex (mETC) selective inhibitors rotenone (complex I) and oligomycin (complex V), the nucleotide antagonist of ATP-dependent mechanisms P1,P4-di(adenosine-5') tetraphosphate (Ap4A) (Sigma, St. Louis, MO), and PKCεV1-2 a PKCε specific translocation inhibitor peptide (PKCε-I, Calbiochem, La Jolla, CA) (Johnson et al., 1996; Khasar et al., 1999). Stock solution (1 μg/μl) of PKCε-I (in 0.9% saline) was stored at −20°C and the injections [1 μg/2.5 μl, using a 10 μl microsyringe (Hamilton, Reno, NV)] were preceded by injection of distilled water (2.5 μl) in the same syringe, separated by a small air bubble, to produce hypo-osmotic shock, thereby enhancing cell membrane permeability to these cell agents (Tsapis & Kepes, 1977; West & Huang, 1980; Taiwo & Levine, 1989; Khasar et al., 1995; Widdicombe et al., 1996). Drug dose selection was based either on the results of previous studies (Dina et al., 2000; Joseph et al., 2004; Joseph & Levine, 2004; 2006) or on preliminary experiments carried out for this study. All inhibitors were diluted with distilled water before intradermal injection into a hind paw. The mETC inhibitors, Z-VAD-FMK, α-lipoic acid and Ap4A (each 5 μg), were administered intradermally (i.d.) on the dorsum of the hind paw, in a volume of 5 μl, via a 30-gauge hypodermic needle. Rotenone, oligomycin and Ap4A were dissolved in 10% DMSO. All the other drugs were dissolved in saline. Paw withdrawal threshold was determined before and 30 minutes after inhibitor administration. The effect of each chemical was determined on different groups of rats.
Mechanical nociceptive threshold was quantified using the Randall–Selitto paw pressure test (Randall & Selitto, 1957), in which a force that increases linearly over time is applied to the dorsum of the rat's hind paw (Taiwo et al., 1989; Taiwo & Levine, 1989), using an Ugo Basile Algesymeter® (Stoelting, Chicago, IL, USA). Rats were placed in cylindrical acrylic restrainers designed to provide adequate comfort and ventilation, to allow extension of the hind leg from the cylinder, and to minimize restraint stress. All rats were acclimatized to the testing procedure, and testing was performed in parallel across groups. Rats were placed in individual restrainers for 1 h prior to starting each study and for 30 min prior to experimental manipulations. Nociceptive threshold was defined as the force at which the rat withdrew its paw. The baseline paw-withdrawal threshold was defined as the mean of three readings. Each paw was treated as an independent measure and each experiment performed on a separate group of rats. All behavioral testing was done between 10:00 and 17:00 h.
The protocol used to study the effects of the interaction between chronic ethanol consumption and nucleoside reverse transcriptase inhibitor (NRTI) was based on two models of neuropathic pain described previously (Dina et al., 2000; Joseph et al., 2004) and used as controls in the current experiments:
Previous studies from our laboratory have shown that a single intravenous (i.v.) injection (50 mg/kg) of the NRTI 2',3'-dideoxycytidine (ddC, Sigma, St Louis, MO) produces an ~25% reduction in paw withdrawal threshold (Joseph et al., 2004) 1 day after administration, with maximum intensity (~35%) on the fifth day. The ddC was dissolved in normal saline, and the volume adjusted to 1 ml/kg for i.v. administration. Before removal of the injection needle, administration of this drug was followed by a bolus injection of an equal volume of saline.
Previous studies from our laboratory have established a model of alcoholic painful peripheral neuropathy in the rat (Dina et al., 2000). Male Sprague Dawley rats (200–220 g), individually caged and maintained under a 12 hr light/dark cycle, were fed Lieber–DeCarli liquid diet (Dyets Inc., Bethlehem, PA) (Lieber & DeCarli, 1982; 1989; Lieber et al., 1989) with ethanol (6.5%) for 3 weeks, in a regimen of 4 days of diet with ethanol/3 days normal diet. After the second week the mechanical nociceptive threshold was significantly lower in the rats on the ethanol diet (ED) than in the control group. After the third week of ED, the animals showed persistent hyperalgesia that lasted for at least five weeks.
The protocol used to study the effects of the interaction between ethanol consumption and NRTI therapy consisted in the administration of a low dose of ddC (5 mg/kg), which does not induce changes in mechanical threshold (Joseph et al., 2004) in rats submitted to ED (6.5%) for 4 days, which also does not produce a change in nociceptive threshold. The ddC was intravenously injected on the 4th day of ED.
The effect of pharmacological inhibitors on the hyperalgesia induced by the neuropathic pain models was determined in three different groups of rats, i.e., in ddC-treated rats, in rats on ED for 3 weeks, and in the co-morbidity neuropathy model (4 days ED + low-dose ddC). Mechanical paw withdrawal threshold was measured immediately before the administration of the pharmacological inhibitors and again 30 min afterwards. For the groups that received intravenous ddC, the inhibitors were tested five days post-ddC injection. The tests with the inhibitors in the ethanol-fed groups were performed on the fourth week after the ED has started (one week after finishing ED). The inhibitors were tested, in the animals submitted to the combination protocol (ED/ddC), one day after the ddC injection.
Oligodeoxynucleotide (ODN) antisense (AS) and mismatch (MM) to PKCε mRNA were prepared as described previously (Parada et al., 2003a). The AS ODN, 5'-GCC AGC TCG ATC TTG CGC CC-3', was directed against a unique sequence of rat PKCε mRNA. The corresponding GeneBank (National Institute of Health, Bethesda, MD) accession number and oligodeoxynucleotide position within the cDNA sequence are XM345631 and 226–245, respectively. We have previously shown that spinal intrathecal administration of AS ODN with this sequence decreases PKCε protein in dorsal root ganglia (Parada et al., 2003b; Parada et al., 2003a). The sequence of the MM ODN, 5'-GCC AGC GCG ATC TTT CGC CC-3', corresponds to the PKCε AS sequence with 2 bases mismatched (in bold typeface).
Prior to use, lyophilized ODN was reconstituted in nuclease-free 0.9% NaCl to a concentration of 10 μg/μl and stored at −20°C until use. A dose of 40 μg of AA or MM oligodeoxynucleotide was administered intrathecally once daily in a volume of 20 μl. For this study, the animals were treated for 3 consecutive days before the ED was started, and daily until the 4th day, when the ddC was administered. Prior to each injection, rats were anesthetized with 2.5% isoflurane in oxygen. ODN was injected using a 30-gauge hypodermic needle inserted between the fifth and sixth lumbar vertebrate, at the level of the cauda equina; intrathecal location of the injection needle was confirmed by a flicking of the rat's tail (Papir-Kricheli et al., 1987).
In all experiments, the dependent variable was paw withdrawal threshold expressed as percent change from baseline. One-way ANOVA of the pre-intervention (baseline) paw withdrawal threshold values of all groups (N=198) showed no significant difference (F32,165=1.058; p=0.395). Average baseline paw withdrawal threshold was 103.4 ± 0.59 g (standard error of the mean - SEM). For the data presented in figure 1, a three-way repeated measures ANOVA with two between-subjects factors (diet with two levels and drug with two levels) and one within subjects factor (time with five levels) was performed. Because there was a significant three-way interaction, separate two-way repeated measures ANOVAs were performed for each of the between subjects factors, diet and drug, in order to determine the basis of the three-way interaction. For the data presented in figure 2, one-way ANOVAs with one between-subjects factor (drug with seven levels) were performed, followed by Scheffé post-hoc analyses to identify the significant differences. Because there was a significant interaction, separate one-way ANOVAs were performed for each of the drug groups to determine the basis of the difference. For data presented in figure 3, a two-way ANOVA with two between-subjects factors (drug group with two levels) and ODN treatment group (two levels) was performed. For the data presented in figure 4, two-way repeated measures ANOVAs with one between-subjects factor (drug with two levels) and one within-subjects factor (time with 10 levels) were performed. For all repeated measures ANOVAs, the Mauchly criterion was tested to determine if the assumption of sphericity for the within-subjects effects was met; if the Mauchly criterion was not satisfied, Greenhouse-Geisser adjusted p-values are presented. Data are presented in figures as mean ± SEM.
We developed an experimental model to test the changes in mechanical threshold induced by ethanol consumption and NRTI therapy in the same animals, using doses (ddC) or duration of administration (ethanol) that alone do not cause sensory changes. Rats submitted to ED (6.5% of ethanol) for four days did not show changes in pain threshold. However, when a low dose of ddC was administrated (5 mg/kg, i.v.) on day 4, the mechanical threshold decreased precipitously by ~30% (Figure 1), thus demonstrating an interaction between ethanol consumption and the NRTI in the induction of a painful peripheral neuropathy. To evaluate mechanisms mediating this hyperalgesia, we used this model to test the effect of drugs that affect each type of neuropathic model separately and when administrated to the animals submitted to the combination.
We first confirmed that inhibitors of the mitochondrial electron transport chain, rotenone (complex I) and oligomycin (complex V) and the antioxidant α-lipoic acid, as well as the ATP-dependent mechanism antagonist P1,P4-di(adenosine-5') tetraphosphate (Ap4A), inhibited the hyperalgesia induced by ddC (50 mg/kg, i.v.) (rotenone 76% inhibition, oligomycin 72%, α-lipoic acid 76%, and Ap4A 79%) (Figure 2A). In addition, the non-specific caspase inhibitor Z-VAD-FMK also inhibited ddC hyperalgesia (94%). However, the PKCε translocation inhibitor (PKCε-I) had no effect in this model.
In the ethanol-induced neuropathy model drugs that inhibit mitochondrial processes (rotenone, oligomycin, α-lipoic acid and Ap4A, and non-selective caspase inhibitor Z-VAD-FMK) did not affect the ED (6.5% ethanol for 3 weeks, in a regimen of 4 days ED/3 days normal diet)-induced decrease in mechanical nociceptive threshold, while PKCε-I decreased hyperalgesia (70%, Figure 2B).
In the painful peripheral neuropathy model induced by low doses of ddC plus short duration ED, when we administered the same pathway inhibitors, we observed a profile more similar to that observed in ddC- than ethanol-induced neuropathy, as hyperalgesia was decreased by rotenone (98% of inhibition), oligomycin (74%), α-lipoic acid (63%) and Ap4A (76%), and PKCε-I had no effect (Figure 2C). However, Z-VAD-FMK, effective in the ddC painful peripheral neuropathy model, had no effect in the co-morbidity model.
We also confirmed the lack of a role of PKCε in the co-morbidity model by spinal administration of oligodeoxynucleotides antisense or mismatch to PKCε. In the ED model of painful peripheral neuropathy, as previously reported (Dina et al., 2006), ODN AS but not MM to PKCε markedly inhibited hyperalgesia (Figure 3, ED for 2 weeks, two right bars). However, in the co-morbidity model PKCε AS did not significantly affect hyperalgesia (Figure 3, two left bars).
Finally, we examined the impact of repeated exposure to ED in the co-morbidity model. We found that repeated cycles of ethanol exposure further enhanced hyperalgesia in the co-morbidity model (Figure 4), as ethanol diet in ddC-treated animals that received ED for 2 weeks still showed significant decrease in mechanical nociceptive threshold, at least 2 weeks after the interruption of the ED, when compared to rats treated only with ED for 2 weeks (Figure 4B). Animals submitted to ED plus ddC that received ED for only one week, still showed decreased mechanical threshold (~15%) 20 days after interruption of ED, when compared to control animals fed ED for 1 week (Figure 4A).
The most effective treatment for HIV/AIDS is “highly active anti-retroviral therapy” (HAART), which consists of combinations of nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors and protease inhibitors. While it is a highly effective therapy for the treatment of HIV/AIDS, HAART can induce a painful peripheral neuropathy, a distal symmetric small fiber dying back axonal neuropathy (Dubinsky et al., 1989; Simpson & Tagliati, 1995; Dalakas, 2001; Pardo et al., 2001; Reliquet et al., 2001), also known as antiretroviral toxic neuropathy, a cause of significant morbidity in HIV/AIDS patients (Berger et al., 1993; Simpson & Tagliati, 1995; Dalakas, 2001; Cohen et al., 2002; Quasthoff & Hartung, 2002). This therapy-induced peripheral neuropathy compromises adherence to treatment and may alter a clinically effective regimen or even necessitate its discontinuation (Dieterich, 2003). Among the drugs that comprise HAART, the NRTIs (e.g., zidovudine (AZT), zalcitabine (ddC), didanosine (ddI), and stavudine (d4T)) clearly play a role in HAART-induced painful peripheral neuropathy (Brinkman et al., 1998; Moyle & Sadler, 1998; Dalakas, 2001; Dalakas et al., 2001; Simpson, 2002; Gerschenson & Brinkman, 2004; Hulgan et al., 2005), being associated with a three-fold increase in the incidence of peripheral neuropathy in AIDS patients (Moore et al., 2000). This type of neuropathy occurs in up to two-thirds of patients taking NRTIs (Simpson & Tagliati, 1995; Oh et al., 2001) and limits the amount of time HAART can be administered (Sharma et al., 2004). It produces clinically significant morbidity in 10–35% of HIV-positive individuals (Hall et al., 1991; Kieburtz et al., 1998; Sacktor, 2002).
The toxic effects of chronic ethanol consumption on the peripheral nervous system are also well-documented (Juntunen et al., 1978; Bosch et al., 1979; Scott & Edwards, 1980; Oakes & Pozos, 1982; Juntunen et al., 1983; Massarotti, 1983; Riopelle et al., 1984; Scott et al., 1986; McLane, 1987; 1990; Diamond & Messing, 1994; Hundle et al., 1997; Wu & Kendig, 1998; Dina et al., 2000). Neuropathic pain syndromes occur as a result of ethanol-induced peripheral neuropathy (Foster et al., 1999). Importantly, abuse of ethanol is one of the most important co-morbid risk factors for peripheral neuropathy in patients with HIV/AIDS, and the painful peripheral neuropathy induced by HAART (Moyle & Sadler, 1998; Nath et al., 2002; Lopez et al., 2004; Nicholas et al., 2007). Of note, alcohol abuse is especially prevalent in the population of HIV patients; for example, most HIV patients (53–63%) regularly drink ethanol (Galvan et al., 2002; Miguez et al., 2003), and it has been reported that 4–41% of HIV patients in various cohorts are alcoholic (Atkinson et al., 1988; Brown et al., 1992; Rosenberger et al., 1993; Zenilman et al., 1994; Lefevre et al., 1995; Dew et al., 1997; Cook et al., 2001; Galvan et al., 2002; Samet et al., 2004). This condition led to a number of studies involving the effects of ethanol consumption on HIV-positive patients, including how it affects the progression of the infection (Cook et al., 1997; Cook et al., 2001; Liu et al., 2003; Brailoiu et al., 2006) and the negative impact upon the patient's response to NRTI therapies (Giancola et al., 2006). However, few studies have focused on the effects of the interaction of ethanol consumption and NRTI therapy in sensory systems.
We have previously established a model of ethanol-induced painful peripheral neuropathy in rats by feeding them a Lieber-DeCarli diet, which simulates human chronic alcohol consumption while assuring normal micronutrient intake (Dina et al., 2000; Dina et al., 2006). In protocols in which rats underwent intermittent withdrawal, painful peripheral neuropathy developed much more rapidly (Dina et al., 2006); the C-fiber mechanical threshold was lowered and the number of action potentials elicited during sustained mechanical stimulation increased in ethanol fed rats. In our current study, ethanol-containing diet with normal micronutrient levels was used to study the interaction of ethanol consumption and NRTIs on the function of the peripheral nervous system. We first tested the hypothesis that exposure to ethanol enhances the neuropathic impact of HAART by characterizing the effect of NRTI in the setting of ethanol consumption. We then determined the second messengers mediating the hyperalgesia induced by ethanol- and nucleoside-induced hyperalgesia, and if ethanol exacerbation of nucleoside-induced hyperalgesia involves second messenger pathways implicated in ddC- and/or ethanol-induced painful peripheral neuropathy.
To address the question, what is the mechanism underlying the co-morbid effects of NRTI-induced painful peripheral neuropathy and alcohol consumption, we developed a model of co-morbid painful peripheral neuropathy. After establishing a model system for studying co-morbidity in painful peripheral neuropathy, we focused our attention on the mechanisms in sensory neurons that underlie the interaction between the hyperalgesia induced by ethanol and antiretroviral therapy. Given that the mechanical hyperalgesia induced by giving both low dose ddC and short duration ethanol consumption is mitochondria but not PKCε dependent, we suggest that ethanol consumption enhances ddC effects, rather than vice versa. Since the painful peripheral neuropathy produced by longer term consumption of ethanol is mitochondria independent, we assume ethanol acts by an indirect mechanism to enhance neuropathic effects of ddC. Of note in this regard, we have previously shown that physiological activation of neuroendocrine stress pathways play a crucial role in the neurological effect of ethanol (Dina et al., 2008). In fact, it is well established that ethanol abuse activates the neuroendocrine stress axes and its withdrawal further exacerbates neuroendocrine stress axis activation (Linnoila et al., 1987; Koob, 1999; Sofuoglu et al., 2001; Errico et al., 2002; Bruijnzeel et al., 2004; Devaud et al., 2006; Koob, 2006; Rasmussen et al., 2006). Since alcoholics consume ethanol intermittently, they may enter early withdrawal before they are able to re-administer ethanol, and recovering alcoholics report increased stress (Lamon & Alonzo, 1997; Koob, 2003; Poage et al., 2004). Importantly, recent evidence has shown that stress hormones modulate mitochondrial function (Du et al., 2009a; Du et al., 2009b; Fujita et al., 2009). Additional studies will be needed to assess the indirect mechanism by which ethanol affects the painful peripheral neuropathy induced by ddC, by impacting mitochondria-dependent mechanisms.
Finally, while mitochondrial mechanisms appear to underlie the painful peripheral neuropathy produced by sub neuropathic doses of ddC and ethanol, there was one difference between the mechanism of ddC painful peripheral neuropathy and that induced by low dose ddC and ethanol. Thus, in ddC peripheral neuropathy inhibitors of three mitochondrial functions – the mitochondrial electron transport chain, oxidative stress and caspase signaling – attenuate mechanical hyperalgesia while in the co-morbidity model the caspase inhibitor was without effect. What underlies this difference in the role of this one mitochondrial mechanism in the two peripheral neuropathies is currently unknown.
In summary, the impact of co-morbid risk factors in peripheral neuropathies is poorly understood, in large part due to lack of model systems in which to evaluate this clinically important problem. We have developed an animal model of co-morbid neuropathic insults for two common neurotoxic exposures in patients with HIV/AIDS, a nucleoside reverse transcriptase inhibitor used to treat HIV/AIDS and ethanol consumption, a common co-morbid factor in patients with HIV/AIDS. This is, to our knowledge, the first model system for studying the mechanism by which co-morbid risk factors induce painful peripheral neuropathy. Based on our studies, we suggest that ethanol consumption enhances, by an indirect mechanism, the mitochondrial-dependence underlying ddC-induced painful peripheral neuropathy. Our ultimate goal is to use this type of study to improve the medical management of painful peripheral neuropathy in patients with HIV/AIDS.
This study was funded by the National Institutes of Health (NIH). We thank Dr. Robert Gear for assistance with statistical analysis.