INTRODUCTION

Morphine is the analgesic of choice for the treatment of moderate to severe cancer pain.¹ However, 10–30% of patients do not respond to morphine, achieving poor analgesic response or intolerable adverse effects.² At present, we cannot predict which patients are likely to achieve good analgesia or develop adverse effects. In patients who do not tolerate morphine, it is becoming increasingly common to prescribe other 'alternative' strong opioids. Even with the use of these alternative opioids, outcomes are often variable and unpredictable.³ Moreover, it can be difficult to identify true opioid intolerance in cancer patients, where symptoms such as nausea or cognitive impairment may have multiple underlying aetiologies.

We have previously hypothesized that two-thirds of the interindividual variability in response to morphine is due to genetic variation.⁴ Response to a drug may depend on a number of factors including drug absorption, distribution, metabolism and elimination. Drug concentration at the target site, the number and morphology of target receptors, together with variation in multiple downstream events will also influence response. Thus, genetic variation in the alleles of multiple candidate genes may potentially influence response to a given opioid. The reason for variable response to codeine has already been established; 6–7% of Caucasians have polymorphisms of the cytochrome P4502D6 gene, which prevent metabolism of codeine to the active drug morphine.⁵ Morphine is metabolized by uridine diphosphate-glucuronysltransferase 2B7 (UGT2B7). Serum morphine and morphine-metabolite concentrations have not been shown to correlate with pain relief or side effects in clinical practice.^{6, 7} Genetic variation in the promoter region of the UGT2B7 gene correlated with serum morphine and morphine-6-glucuronide concentrations in one study,⁸ but this was not seen in a larger Norwegian study.⁹ Therefore, whether or not polymorphisms in UGT2B7 influence clinical response to morphine remains unclear.

The primary site of action of morphine and other commonly used opioid analgesics is the -opioid receptor.^{10, 11} The -opioid receptor is a G-protein-coupled receptor.¹² These receptors are expressed both at a spinal level in the dorsal horn and in multiple regions within the brain.¹³ They are also widely expressed in the periphery and found on various circulating immune cells.¹⁴ In -opioid receptor knockout mice, spinal and supraspinal analgesic models show loss of analgesic activity and virtually all other effects of morphine.¹⁵ Nociceptive thresholds vary in gene dose-dependent manner. Mice with no -opioid receptor gene have lower thresholds than heterozygous knockouts (with 50% of wild-type receptor densities), which have lower thresholds than wild-type mice with intact -opioid receptors.

Changes in -opioid receptor densities, potentially contributed by allelic variants, can produce changes in nociceptive responses and affect opioid response.¹⁵ Binding studies to post-mortem brain samples and in vivo positron-emission tomography radioligand analyses suggest 30–50% or even larger ranges of differences in human -opioid receptor densities.^{16, 17} Interindividual variation in human -opioid receptor densities may be due to altered -opioid receptor gene expression. In the cis-acting DNA promoter and enhancer sequences of the -opioid receptor, there are recognition sites for regulatory DNA-binding proteins. These proteins include various transcription factors. A number of transcription factor recognition sites have been postulated in the human -opioid receptor gene.^{18, 19} Functional studies have highlighted the role of stat6, which causes an increase in -opioid receptor gene expression.²⁰

Genetic variation in the gene encoding the -opioid receptor has been shown to alter the binding affinities of different opioids.^{21, 22, 23, 24} Different opioids may therefore cause differential activation and desensitisation of the -opioid receptor.

Following agonist-induced activation, a complex cascade of intracellular events results in inhibition of neuronal transmission of painful stimuli. This signal is then terminated by receptor phosphorylation, desensitisation and internalisation. Arrestin2 is an intracellular protein, involved at multiple points in regulating this sequence of events.^{12, 25} Animal studies support the physiological role of arrestin2 in -opioid receptor desensitisation. The analgesic effect of morphine is both increased and prolonged in arrestin2 knockout mice.²⁶

This prospective case/control study was part of a larger project to identify cancer patients who achieved good analgesic benefit with morphine (controls) and to compare them with those who did not tolerate morphine, but required switching to alternative opioids (cases). Furthermore, we aimed to correlate the need to switch with genetic variation in various candidate genes. Full data on the clinical phenotype of each group and its contribution to the need to switch has been described previously.²⁷ This paper presents novel data on the genetic variation in four candidate genes, the -opioid receptor, arrestin2, stat6 and UGT2B7 genes, and their contribution to variation in response to morphine in cancer patients.

RESULTS

Patient Demographics and Blood Results

In all, 186 patients were recruited to the primary study, 138 controls who had taken morphine for at least 1 month with good clinical benefit and 48 switchers who despite adequate dose escalation had not achieved analgesic control but experienced intolerable side effects. In all, 87% of the switchers achieved acceptable pain control with minimal side effects on an alternative opioid; the majority switched to oxycodone as per departmental guidelines. DNA was available for genetic analysis on 156 Caucasian patients (117 controls and 39 switchers). There were no differences in age (mean 57.612.9 years), gender (45% male) or cancer diagnoses between switchers and controls, and full details of this cohort have been described elsewhere. Morphine dose ranged from 10 to 1060 mg/24 h and there were no differences between groups in either serum morphine and metabolite levels, or basic biochemical and haematological parameters (Table 1). As expected from clinical assessment, switchers had higher pain and side-effect scores than controls. The reasons for switching are given in Table 2.

Table 1 - Morphine dose and blood results in the 186 patients recruited (138 controls who tolerated morphine and 48 switchers who required an alternative opioid).

Full table

Table 2 - Pain scores and reason for switching in the 186 patients recruited (138 controls who tolerated morphine and 48 switchers who required an alternative opioid).

Full table

Genotype and Haplotype Data

The genotype and haplotype data for each of the candidate genes arrestin2, stat6, -opioid receptor and UGT2B7 are presented in Tables 3a–d and 4a–d, respectively. Genotype/phenotype associations for each gene are summarised below. Genetic associations were independent of each other; no gene–gene interactions were identified in the data-mining analysis.

Table 3 - Genotype data for 156 Caucasian patients (117 controls who tolerated morphine and 39 switchers who required an alternative opioid): (a) -opioid receptor gene, (b) arrestin2 gene, (c) stat6 gene and (d) UGT2B7 gene.

Full table

Table 4 - Phase constructed haplotypes from genotype data on 156 Caucasian patients (117 controls who tolerated morphine and 39 switchers who required an alternative opioid): (a) -opioid receptor, (b) arrestin2, (c) stat6 gene and (d) UGT2B7 gene.

Full table

Arrestin2

We found a significant difference in genotype and allele frequencies for the T8622C SNP in the arrestin2 gene, in switchers compared with controls (²=6.22, P=0.045 and ²=4.41, P=0.036, respectively). Switchers were more likely to carry the common T allele at this position (²=6.14, P=0.013). In addition, although differences in genotype frequencies did not reach statistical significance, switchers were more likely to carry the common allele at the following positions: 1182 G/A, 5864 G/A and 11143 G/A (²=5.3, P=0.021, ²=4.1, P=0.043, ²=4.1, P=0.043 respectively).

All SNPs across the 19.3 kb region studied were in linkage disequilibrium (D'>0.5) using the anchor point -8130 in the promoter. Four haplotypes occurred with a frequency of >5% and accounted for 86% of the population. Seven haplotypes with a frequency of 1–3% and four haplotypes with a frequency <1% described the remaining 14% of the population. There were no significant differences in haplotype frequencies in switchers compared with controls.

Stat6 Genotype

For stat6, there was increased carriage of the variant T allele at -1714 C/T (²=4.95, P=0.026) and a significant difference in the allelic frequency at 9065 C/T (²=3.86, P=0.049).

The SNPs in stat6 covered a 15.5 kb region. One SNP in intron 14 (8334 A/C) was not in linkage disequilibrium. Haplotypes were constructed firstly using all eight SNPs and secondly from the remaining seven SNPs. Data are presented for the seven SNPs, five haplotypes had a frequency of >5% and accounted for 92% of the population. The resultant four haplotypes with a frequency of 1–2% and four haplotypes with a frequency <1% described the remaining 8% of the population. There were no significant differences in haplotype frequencies in switchers compared with controls.

-Opioid Receptor Genotype

No significant differences in genotype or allelic frequencies were seen for SNPs in the -opioid receptor gene.

All seven SNPs across the 80.7 kb region of the -opioid receptor gene were in linkage disequilibrium (D'>0.5) using the position 80547 in the 3'UTR as the anchor point. Six haplotypes had a frequency of >5% and accounted for 91% of the population. A further six haplotypes described the remaining 9% of the population. There were no differences in the haplotype frequencies in switchers compared to controls.

UGT2B7 Genotype

No significant differences in genotype or allelic frequencies were seen for SNPs in the UGT2B7 gene and there were no correlations between genotype and morphine or morphine metabolite levels.

Three of the four SNPs across the 16 kb region were in complete linkage and 100% of the population could be described by three haplotypes. There were no differences in haplotype frequencies between switchers and controls.

DISCUSSION

Improving pain control in cancer patients is a priority both for clinicians and patients. In this study, all patients were treated with morphine first-line to control cancer-related pain, as per WHO and departmental guidelines. For those patients who did not tolerate morphine, a significant proportion (87%) improved on an alternative opioid. This study identified significant differences in the genotype of switchers compared with controls, which may explain part of the observed variation in response to morphine.

The phenotype of switchers is characterised by higher pain and side-effect scores compared with controls. In clinical practice, morphine doses are titrated until the patient achieves adequate analgesia. Most patients who do not achieve analgesic control with morphine will reach a point where side effects prevent further increases in morphine dose. However, both the dose at which this occurs and the predominant limiting side effect varies between individuals. Therefore, while this group can be treated as a whole, we recognise that there are different phenotypes within the switcher group itself. Accurate clinical information collected in this study allows these subgroups to be more clearly identified and considered as separate factors in data-mining analysis, but larger numbers of patients will need to be recruited to enable meaningful statistical evaluation of these subgroups. One of the limitations of this study therefore is that if genetic variation contributes to a single side effect, it will have been diluted in the analysis by variation in the switcher group as a whole and may not have been identified in data mining due to small numbers.

Every effort was made to identify whether side effects were truly morphine related rather than due to the cancer itself or other concomitant medications, but this can be difficult in the clinical setting. Biochemical and haematological data were included in the analysis to help identify any confounding clinical reason for side effects. A retrial of morphine would clearly be unethical in these patients, but the fact that a high percentage of switchers' (87%) side-effect profile did indeed improve when switched to an alternative opioid supports side effects being morphine related. This compares favourably with outcomes of switching in smaller clinical studies from other groups studying a similar patient population: Ashby 1999 (n=49) 53–72% improvement in different side effects on switching; and Gagnon 1999 (n=63) 25–35% improvement in different side effects on switching.^{28, 29}

The dose of morphine required to achieve analgesia is known to vary widely between individuals and this is reflected in our population; morphine doses ranged from 10 to 1060 mg/day. It is usual practise to titrate with short-acting opioids (oramorph or sevredol) until adequate pain relief is achieved with acceptable side effects, and then to convert to long-acting equivalents (MST, MXL, Zomorph). This explains why a greater proportion of switchers, by definition not yet stable, were taking short-acting opioids.

Variation in the frequency of individual polymorphisms between different ethnic populations is well recognised and therefore different ethinic groups should be considered separately. We report genotype data on the Caucasian patients in our population.

Arrestin2 is an intracellular protein, which is important in desensitisation of the -opioid receptor and intracellular trafficking of internalised receptors. As such, it can regulate the number of functional receptors expressed on the cell surface at a given time. Studies in arrestin2 knockout mice show increased and prolonged analgesia in animals without arrestin2, despite similar concentrations of morphine in the blood and similar number and affinity of cerebral -opioid receptors in both populations.³⁰ We hypothesised that genetic variation in the arrestin2 gene might be important in affecting response to morphine in our patients.

This is the first report linking genetic variation in the arrestin2 gene with analgesic response in humans. Comparing switchers and controls, we found a significant difference in genotype (²=6.22, P=0.044) and allele frequency (²=4.41, P=0.035) for the T8622C polymorphism. Switchers were more likely to carry the common T allele at this position (²=6.14, P=0.013). This SNP is in the coding region, exon 11, but does not lead to a change in the amino-acid sequence of the resultant arrestin2 protein. It is possible, however, that this SNP may be in linkage with another functional SNP in the arrestin2 gene and this requires further evaluation. In addition, ² analysis confirmed a significant difference in allele carriage between switchers and controls for three noncoding SNPs in the arrestin2 gene (A1082G, A8864G, A11143G). The significance of these findings needs to be confirmed on larger patient numbers.

A number of addiction studies have focused on genetic variation in candidate genes that may influence tolerance to or dependence on different opioids. The most widely studied SNP in the -opioid receptor gene is the A118G nucleotide substitution. This SNP codes for the amino-acid change Asn to Asp, resulting in loss of a putative glycosylation site in the extracellular N-terminal domain of the receptor. It has been reported in association with both addiction and pain studies.

Addiction studies have published conflicting results. The mutant allele was found to be increased in both a Hispanic subgroup, protecting against drug abuse,³¹ and a Caucasian population, protecting against alcohol³² abuse. However, other studies found no association.^{33, 34, 35} In pain studies, case studies have suggested that the mutant allele may decrease the potency of morphine or morphine-6-glucuronide in cancer patients.^{33, 34, 35, 36, 37} A study in normal volunteers (n=11) showed reduced pupil constriction in response to morphine-6-glucuronide but not morphine in subjects carrying the G allele.³⁸ This may be explained by differences in the binding sites for morphine and M6G.

We found no difference in the genotype or allelic frequencies for this SNP between switchers or controls. In addition, there was no correlation between genotype and the average doses of morphine required to achieve adequate pain relief, two-third of the nine patients requiring >500 mg/day were homozygous for the A allele and none were homozygous for the rare allele. Of note, the frequency of the mutant allele (G) in our population was 0.15, comparable to reported frequencies of 0.12 in other Caucasian populations.³¹

The C17T nucleotide substitution in the -opioid receptor gene has also been weakly associated with opioid addiction.^{31, 39, 40} Genotype for this SNP was determined by direct sequencing in 84 patients, only one heterozygote was identified. This polymorphism was not considered further as part of this study, but may be important in other ethnic groups such as African Americans where it can be found in 20% of the population.⁴¹

A number of putative SNPs have been identified in the promoter region of the -opioid receptor. Of five SNPs studied in our group, only one -172 G/T was polymorphic (minimum 80 Caucasian patients studied). Promoter polymorphisms may upregulate or downregulate gene expression. Importantly, the binding site for the transcription factor stat6 has been reported to contain an SNP that alters IL-4 upregulation of the -opioid receptor.²⁰ This nucleotide was not polymorphic in our Caucasian population.

Stat6 is a transcription factor that can alter -opioid receptor gene expression. The stat6 gene is known to be highly polymorphic and assays were developed for 24 SNPs that had been validated on SNP databases. Seven of these were polymorphic in our Caucasian population. For stat6, there was increased carriage of the variant T allele at -1714 and a significant difference in the allelic frequencies at 9065 C/T in intron 16. The polymorphism at -1714 does not correspond to a known transcription factor binding site. It will be therefore important to both validate this finding in a separate group of patients and consider functional studies to determine whether this SNP, or one in linkage to it, is important in affecting response to morphine.

Patients who did not respond to morphine were switched to oxycodone as second-line opioid. Oxycodone is metabolised by a different pathway to morphine, cytochrome P450 (2D6, 3A4, 3A5), but like morphine, acts primarily at the -opioid receptor. This would suggest that pharmacokinetic rather than pharmacodynamic variation would be more likely to mediate the difference in analgesic response between switchers and controls. As such, it was surprising to find no differences in UGT2B7 genotype or morphine/metabolite ratios between switchers and controls. This confirms the findings of a large Norwegian study, which found no correlation between serum morphine/metabolite ratios and the functional nonsynonymous SNP (2099 C/T; H268Y) in exon 2 of UGT2B7.

Studies that correlate SNPs with disease outcome by this method are known to generate false-positive but rarely generate false-negative results. As such, while the positive findings of this study require further validation in other patient groups, these results suggest that SNPs in the prime candidate genes of UGT2B7 and the -opioid receptor are unlikely to have clinical relevance in predicting the response to morphine. Previous work by our group showed that 20% of variation in response could be explained by nongenetic factors⁴ and therefore other genes need to be considered in future studies to account for the as yet unexplained variation in clinical response to morphine.

In summary, our study presents novel data suggesting that variation in genes involved in -opioid receptor signalling influence clinical response to morphine. We have presented data in a Caucasian population, for the genotype and haplotype frequencies of 26 SNPs across four candidate genes, which we hypothesised might influence response to morphine. We found that genetic variation in the arrestin2 gene was associated with the need to switch from morphine to an alternative opioid; switchers were more likely to carry the T allele at position 8622, P=0.013. Polymorphisms in the -opioid receptor gene, which have shown conflicting results in previous studies, were not significantly different in switchers compared with controls. Further work is needed to validate and determine the functional significance of these findings and to explore the significance of genetic variation in switcher subgroups and different ethnic populations.

METHODS

This prospective case/control study was approved by the local science and ethics committees and all patients gave informed written consent.

Patients, Inclusion and Exclusion Criteria

Eligible patients were identified by the Specialist Palliative Care Team. Two broad phenotypes were identified. Controls were defined as patients who had already been taking morphine for at least 1 month, with good pain relief and minimal side effects. Switchers (or cases) were those who despite adequate dose escalation had inadequate analgesia or intolerable side effects on morphine and required switching to an alternative opioid. Patients either under the age of 18 years, with predominately neuropathic pain, unable to give informed consent, or with serum creatinine >1.5 times the upper limit of normal, were excluded.

Data Collection

Data were collected at a single time point. For controls, this could be at any time point as long as they had been taking morphine for at least 1 month with good clinical effect. For switchers, it was the time point at which they were reviewed by the Specialist Palliative Care team because of inadequate response to morphine and the need for an alternative opioid. Demographic data were collected including, age, ethnicity, cancer diagnosis and dates of recent chemotherapy. Patients were asked to complete a modified brief pain inventory and toxicity scores. We recorded current opioid use and opioid history, and current medications, and for switchers, the reason for switching and whether or not change in therapy was successful.

Blood samples were taken for routine haematology and biochemistry, DNA extraction, and measurement of morphine and metabolite levels. DNA was extracted using a modified salting out method⁴² and stored at -20°C for genotype analysis. Plasma was separated and stored at -20°C for measurement of serum morphine, morphine-6-glucuronide and morphine-3-glucuronide using reversed phase ion paired high-performance liquid chromatography.⁴³

Genotyping

Putative SNPs in candidate genes were identified from the literature and SNP databases (http://www.ncbi.nlm.nih.gov, http://snpper.chip.org/bio). Figure 1 shows the SNPs identified in each of the candidate genes (-opioid receptor, arrestin2, stat6 and UGT2B7) for inclusion in this study. All assays were used initially to genotype 80–100 Caucasian subjects. All study patients were then genotyped for SNPs with a variant allele frequency >1%. Genotypes were determined using sequence-specific primers in a polymerase chain reaction (SSP-PCR). SSPs with mismatches at the 3'-end were designed to identify each variant, which in combination with a consensus primer produced a product of known size (to be made available from authors or as web table, Appendix A). Exon 1 of the -opioid receptor gene has multiple putative polymorphic sites, and this region was sequenced.

Figure 1.

(a–d) Schematic diagrams of (a) arrestin2 gene, (b) stat6 gene, (c) -opioid receptor gene and (d) UGT2B7 gene. Arrows show the positions of SNPs studied (minimum 80 patients); red arrows show SNPs that are polymorphic in this Caucasian population. SNPs were identified from SNP databases (http://www.ncbi.nlm.nih.gov, http://snpper.chip.org/bio) and published studies.³⁹

Full figure and legend (176K)

SSP-PCR⁴⁴

Amplification conditions

Primer concentrations were titrated to ensure amplification only occurred with exact matching of the primer with genomic DNA. A measure of 5 l of primer mix was added to a 8 l PCR reaction mixture containing buffer (67 mM tris base, pH 8.8; 16.6 mM ammonium sulphate; 2 mM magnesium choride; 0.01% (v/v) Tween-20 (Bioline Ltd, London, UK)); 200 mM of each dATP, dTTP, dGTP and dCTP (Bioline Ltd, London, UK); 0.32 U of Taq polymerase (Biotaq TM, Bioline Ltd, London, UK) and 0.01–0.1 g DNA. This 13 l reaction was dispensed under 10 l of mineral oil and amplified in a PCR machine (PTC-200 machine, MJ Research, Waltham MA, USA). Cycling parameters were set as 1 min at 96°C, followed by: five cycles of 96°C for 25 s, 70°C for 45 s and 72°C for 45 s; 21 cycles of 96°C for 25 s, 65°C for 50 s and 72°C for 45 s; and four cycles of 96°C for 25 s, 55°C for 60 s and 72°C for 120 s. PCR products were then electrophoresed on 1.5% agarose gels (Bioline Ltd, London, UK) containing 0.14 mg/ml ethidium bromide (Sigma Ltd, Poole, UK), at 200 V/cm² in 0.5% tris borate EDTA buffer (Sigma Ltd, Poole, UK). Products were visualised with a UV illuminator and photographed with a Polaroid camera. The presence of an allele-specific band of the expected size, in conjunction with a control band, was used to identify an allele. A reaction was considered negative if the control band was visualised without a specific band of the expected size.

Sequencing

SSPs were designed to amplify a 696 bp fragment of the -opioid receptor gene, containing Exon 1. SSP-PCR was performed using 4 volumes described above. A QIAquick^® PCR purification kit (Qiagen, Crawley, UK) was used to purify the product. CEQ™ 8000 dye terminator cycle sequencing kit (Beckman Coulter, High Wycombe, UK) was used to prepare samples. A measure of 8 l of quick start, 3.2 pM of primer, 100 fM of PCR product and sterile water, to a final volume of 20 l, was amplified in a PTC-200 machine (MJ Research, Waltham MA, USA). Cycling parameters were set as: 30 cycles of 96°C for 20 s, 50°C for 20 s and 60°C for 4 min, followed by cooling to 5°C. A measure of 2 l of 100 mM EDTA (Sigma Ltd, Poole, UK), 2 l of 3 M NaOAc (Sigma Ltd, Poole, UK) and 1 l of glycogen (CEQ sequencing kit) was added.

Ethanol clean-up of samples (20 l reaction in 96-well plate) was as follows: 60 l of ice-cold 95% ethanol was added and the sample centrifuged at 3000 rpm, 4°C for 30 min. Supernatant was discarded and three further washes of 200 l of ice-cold 70% ethanol performed. After the final wash, the supernatant was discarded and the plate centrifuged upside down at 300 rpm, 4°C for 15–20 s. The remaining pellets were air-dried for 20 min to remove remaining ethanol, and then resuspended in 40 l of sample loading solution (CEQ sequencing kit), with an overlay of mineral oil.

Samples were sequenced on a Beckman Coulter CEQ8000 capillary sequencer as per the manufacturer's instructions.

Genotype/haplotype analysis

Individual SNP associations were examined by comparison of genotype and allele frequencies and allele carriage between switchers and controls. The allele frequency is the fraction or percentage of loci that the allele occupies within the population. The allele frequency is calculated as 2 number of homozygote individuals+number of heterozygote individuals/total number of alleles. The allele carriage (whether an individual carries the allele—regardless of homo/heterozygosity) is calculated as number of homozygotes+number of heterozygotes/total number of individuals. All genotype and allele frequencies were checked for Hardy–Weinberg equilibrium.

Different SNPs (within a gene or between different genes) may work in combination having additive or opposing influences on outcome. As such, interactions between SNPs were considered and haplotypes of SNPs within genes constructed as this might have functional implications for the resultant protein. Linkage disequilibrium was calculated and phase-constructed haplotypes determined using computer programs Phase and Arlequin version 2.0 (http://anthropologie.unige.ch/
arlequin/).

Identification of rare intragene haplotypes can also identify genotype errors and so samples identified as having rare intragene haplotypes were regenotyped to reduce any errors.⁴⁵ While all haplotypes were included in the analyses, only the most common are tabulated for brevity.

Statistical Analysis

The database of clinical, laboratory and genetic data were analysed using data mining software, Knowledge Studio (www.angoss.com) and standard statistical software (Stata version 8 www.stata.com). Data mining uses decision tree modelling to examine effects between different variables (including clinical, laboratory and genetic variables) using a step-wise linear regression approach. Phenotype associations were examined and potential gene–gene interactions considered and explored.

Notes

DUALITY OF INTEREST

None declared.

References

References

1.	World Health Organisation. Cancer Pain Relief 2nd edn. World Health Organisation: Geneva, Switzerland 1996;.
2.	Cherny N, Ripamonti C, Pereira J, Davis C, Fallon M & McQuay H et al.. Strategies to manage the adverse effects of oral morphine: an evidence-based report. J Clin Oncol 2001; 19: 2542−2554. | PubMed | ChemPort |
3.	Cherny NJ, Chang V, Frager G, Ingham JM, Tiseo PJ & Popp B et al.. Opioid pharmacotherapy in the management of cancer pain: a survey of strategies used by pain physicians for the selection of analgesic drugs and routes of administration. Cancer 1995; 76: 1283−1293. | PubMed | ChemPort |
4.	Riley J, Ross JR, Rutter D, Shah S, Gwilliam B & Wells AU et al.. A retrospective study of the association between haematological and biochemical parameters and morphine intolerance in patients with cancer pain. Palliat Med 2004; 18: 19−24. | Article | PubMed | ChemPort |
5.	Sindrup SH & Brosen K. The pharmacogenetics of codeine hypoalgesia. Pharmacogenetics 1995; 5: 335−346. | PubMed | ISI | ChemPort |
6.	Quigley C, Joel S, Patel N, Baksh A & Slevin M. Plasma concentrations of morphine, morphine-6-glucuronide and morphine-3-glucuronide and their relationship with analgesia and side effects in patients with cancer-related pain. Palliat Med 2003; 17: 185−190. | Article | PubMed |
7.	Klepstad P, Borchgrevink PC, Dale O, Zahlsen K, Aamo T & Fayers P et al.. Routine drug monitoring of serum concentrations of morphine, morphine-3-glucuronide and morphine-6-glucuronide do not predict clinical observations in cancer patients. Palliat Med 2003; 17: 679−687. | PubMed |
8.	Sawyer MB, Innocenti F, Das S, Cheng C, Ramirez J & Pantle-Fisher FH et al.. A pharmacogenetic study of uridine diphosphate-glucuronosyltransferase 2B7 in patients receiving morphine. Clin Pharmacol Ther 2003; 73: 566−574. | Article | PubMed | ChemPort |
9.	Holthe M, Rakvag TN, Klepstad P, Idle JR, Kaasa S & Krokan HE et al.. Sequence variations in the UDP-glucuronosyltransferase 2B7 (UGT2B7) gene: identification of 10 novel single nucleotide polymorphisms (SNPs) and analysis of their relevance to morphine glucuronidation in cancer patients. Pharmacogenom J 2003; 3: 17−26. | Article | ChemPort |
10.	McQuay H. Opioids in pain management. Lancet 1999; 353: 2229−2232. | Article | PubMed | ChemPort |
11.	Satoh M & Minami M. Molecular pharmacology of the opioid receptors. Pharmacol Therap 1995; 68: 343−364. | Article | ChemPort |
12.	Connor M & MacDonald JC. Opioid receptor signalling mechanisms. Clin Exp Pharmacol Physiol 1999; 26: 493−499. | Article | PubMed | ChemPort |
13.	Stein C. Peripheral mechanisms of opioid analgesia. Anesth Analg 1993; 76: 182−191. | PubMed | ISI | ChemPort |
14.	Bidlack JM. Detection and function of opioid receptors on cells from the immune system [Review] [67 refs]. Clin Diagn Lab Immunol 2000; 7: 719−723. | Article | PubMed | ChemPort |
15.	Clarke S & Kitchen I. Opioid analgesia: new information from gene knockout studies. Curr Opin Anaesthesiol 1999; 12: 609−614. | Article |
16.	Ravert HT, Bencherif B, Madar I & Frost JJ. PET imaging of opioid receptors in pain: progress and new directions. Curr Pharm Des 2004; 10: 759−768. | Article | PubMed | ChemPort |
17.	Zubieta JK, Dannals RF & Frost JJ. Gender and age influences on human brain mu-opioid receptor binding measured by PET. Am J Psychiatry 1999; 156: 842−848. | PubMed | ChemPort |
18.	Wendel B & Hoehe MR. The human mu opioid receptor gene: 5' regulatory and intronic sequences. J Mol Med 1998; 76: 525−532. | Article | PubMed | ChemPort |
19.	Borner C, Hollt V & Kraus J. Involvement of activator protein-1 in transcriptional regulation of the human mu-opioid receptor gene. Mol Pharmacol 2002; 61: 800−805. | Article | PubMed | ISI | ChemPort |
20.	Kraus J, Borner C, Giannini E, Hickfang K, Braun H & Mayer P et al.. Regulation of mu-opioid receptor gene transcription by interleukin-4 and influence of an allelic variation within a STAT6 transcription factor binding site. J Biol Chem 2001; 276: 43901−43908. | Article | PubMed | ISI | ChemPort |
21.	Wang JB, Imai Y, Eppler CM, Gregor P, Spivak CE & Uhl GR. Mu opiate receptor: cDNA cloning and expression. Proc Natl Acad Sci USA 1993; 90: 10230−10234. | PubMed | ChemPort |
22.	Chaturvedi K, Shahrestanifar M & Howells RD. Opioid receptor: role for the amino terminus as a determinant of ligand binding affinity. Mol Brain Res 2000; 76: 64−72. | Article | PubMed | ChemPort |
23.	Surratt CK, Johnson PS, Moriwaki A, Seidleck BK, Blaschak CJ & Wang JB et al.. Mu opiate receptor. Charged transmembrane domain amino acids are critical for agonist recognition and intrinsic activity. J Biol Chem 1994; 269: 20548−20553. | PubMed | ChemPort |
24.	Pil J & Tytgat J. The role of the hydrophilic Asn230 residue of the mu-opioid receptor in the potency of various opioid agonists. Br J Pharmacol 2001; 134: 496−506. | Article | PubMed | ChemPort |
25.	Keith DE, Anton B, Murray SR, Zaki PA, Chu PC & Lissin DV et al.. mu-Opioid receptor internalization: opiate drugs have differential effects on a conserved endocytic mechanism in vitro and in the mammalian brain. Mol Pharmacol 1998; 53: 377−384. | PubMed | ISI | ChemPort |
26.	Bohn LM, Lefkowitz RJ, Gainetdinov RR, Peppel K, Caron MG & Lin FT. Enhanced morphine analgesia in mice lacking beta-arrestin 2. Science 1999; 286: 2495−2498. | Article | PubMed | ISI | ChemPort |
27.	Riley J, Ross JR, Rutter D, Wells AU, Goller K, DuBois R & Welsh KI. No pain relief from morphine? Individual variation in sensitivity to morphine and the need to switch to an alternative opioid. Support Care Cancer 2005; June 11 2005; Epub ahead of print.
28.	Ashby MA, Martin P & Jackson KA. Opioid substitution to reduce adverse effects in cancer pain management. Med J Aust 1999; 170: 68−71. | PubMed | ChemPort |
29.	Gagnon B, Bielech M, Watanabe S, Walker P, Hanson J & Bruera E. The use of intermittent subcutaneous injections of oxycodone for opioid rotation in patients with cancer pain. Support Care Cancer 1999; 7: 265−270. | Article | PubMed | ChemPort |
30.	Bohn LM, Lefkowitz RJ & Caron MG. Differential mechanisms of morphine antinociceptive tolerance revealed in (beta)arrestin-2 knock-out mice. J Neurosci 2002; 22: 10494−10500. | PubMed | ChemPort |
31.	Bond C, LaForge KS, Tian M, Melia D, Zhang S & Borg L et al.. Single-nucleotide polymorphism in the human mu opioid receptor gene alters beta-endorphin binding and activity: possible implications for opiate addiction. Proc Natl Acad Sci USA 1998; 95: 9608−9613. | Article | PubMed | ChemPort |
32.	Town T, Abdullah L, Crawford F, Schinka J, Ordorica PI & Francis E et al.. Association of a functional mu-opioid receptor allele (+118A) with alcohol dependency. Am J Med Genet 1999; 88: 458−461. | Article | PubMed | ISI | ChemPort |
33.	Bergen AW, Kokoszka J, Peterson R, Long JC, Virkkunen M & Linnoila M et al.. Mu opioid receptor gene variants: lack of association with alcohol dependence. Mol Psychiatry 1997; 2: 490−494. | Article | PubMed | ChemPort |
34.	Sander T, Gscheidel N, Wendel B, Samochowiec J, Smolka M & Rommelspacher H et al.. Human opioid receptor variation and alcohol dependence. Alcohol Clin Exp Res 1998; 22: 2108−2110. | Article | PubMed | ISI | ChemPort |
35.	Li T, Liu X, Zhu ZH, Zhao J, Hu X & Sham PC et al.. Association analysis of polymorphisms in the opioid gene and heroin abuse in Chinese subjects. Addict Biol 2000; 5: 181−186. | Article | ChemPort |
36.	Lotsch J, Zimmermann M, Darimont J, Marx C, Dudziak R & Skarke C et al.. Does the A118G polymorphism at the mu-opioid receptor gene protect against morphine-6-glucuronide toxicity? Anesthesiology 2002; 97: 814−819. | Article | PubMed |
37.	Hirota T, Ieiri I, Takane H, Sano H, Kawamoto K & Aono H et al.. Sequence variability and candidate gene analysis in two cancer patients with complex clinical outcomes during morphine therapy. Drug Metab Dispos 2003; 31: 677−680. | Article | PubMed | ChemPort |
38.	Lotsch J, Skarke C, Grosch S, Darimont J, Schmidt H & Geisslinger G. The polymorphism A118G of the human mu-opioid receptor gene decreases the pupil constrictory effect of morphine-6-glucuronide but not that of morphine. Pharmacogenetics 2002; 12: 3−9. | Article | PubMed | ChemPort |
39.	Hoehe MR, Kopke K, Wendel B, Rohde K, Flachmeier C & Kidd KK et al.. Sequence variability and candidate gene analysis in complex disease: association of mu opioid receptor gene variation with substance dependence. Hum Mol Genet 2000; 9: 2895−2908. | Article | PubMed | ChemPort |
40.	Berrettini WH, Hoehe MR, Ferraro TN, DeMaria PA & Gottheil E. Human opioid receptor gene polymorphisms and vulnerability to substance abuse. Addict Biol 1997; 2: 303−308. | Article | ISI | ChemPort |
41.	LaForge KS, Yuferov V & Kreek MJ. Opioid receptor and peptide gene polymorphisms: potential implications for addictions [Erratum appears in Eur J Pharmacol 2001 Aug 24; 426(1−2): 145]. Eur J Pharmacol 2000; 410: 249−268. | Article | PubMed | ISI | ChemPort |
42.	Miller SA, Dykes SS & Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 1988; 16: 1215. | PubMed | ISI | ChemPort |
43.	Joel S, Osborne R & Slevin M. An improved method for the simultaneous determination of morphine and its principal glucuronide metabolites. Chromatography 1988; 430: 394−399. | ChemPort |
44.	Dowman R. Topographic analysis of painful laser and sural nerve electrical evoked potentials. Brain Topogr 2004; 16: 169−179. | Article | PubMed |
45.	Ahmad T, Neville M, Marshall SE, Armuzzi A, Mulcahy-Hawes K & Crawshaw J et al.. Haplotype-specific linkage disequilibrium patterns define the genetic topography of the human MHC. Hum Mol Genet 2003; 12: 647−656. | Article | PubMed | ChemPort |

Appendices

Appendix A