Genetics of CPSP

Abstract

Various polymorphisms in several genes appear to be involved in the development of chronic post-surgical pain (CPSP). These genes are involved in the transduction, transmission and modulation of a nociceptive impulse. Understanding the influence of such polymorphisms would lead to a better awareness of the underlying processing in CPSP, with the possibility of stratifying the risk of CPSP for individual patients. It may also identify new treatment options by targeting specific points in this pathway. We look into six genes—SCN9A, KCNS1, GCH1, COMT, OPRM1, OPRK1—that are involved in nociception, and look at current literature to support their involvement in the development of CPSP. We also describe the potential use of such information in clinical practice.

1. Introduction

Nociception involves various receptors encoded by different DNA sequences. Hence, changes in these genes could play a significant role in nociception by altering the function of receptors and other proteins involved in nociception [1].

Mutations in a gene may involve three main different mechanisms: base substitution, insertion or deletion [2]. It is more frequent to have single nucleotide changes, or polymorphisms (SNPs) than changes that involve a series of bases.

We shall be focusing on three main pathways that could be affected by different genotypes:

• Ionic channels involved in the initialization and transmission of nociceptive impulse

• Modulation of pain pathway involving catecholamines

• Pharmacogenetic response to analgesics.

2. Genetic variations in ionic channels

We shall concentrate on two ionic channels: the sodium voltage-gated channel and the potassium voltage-gated channel.

2.1 SCN9A

Voltage-gated sodium channels (VGSC) are important in the generation and transmission of an action potential. The nine different VGSC alpha subunits are encoded for by nine genes spread over four chromosomes [3]. In particular, one type of VGSC alpha subunit, Nav 1.7 is implicated in channelopathy-associated insensitivity to pain and is encoded by SCN9A. Nav1.7 is involved in the initiation of an action potential and hence it is important in setting the sensitivity for nociceptive signals to be transmitted [4]. In fact, a number of Nav1.7 inhibitors have been looked into as possible analgesics [5].

The SCN9A gene is found on chromosome 2 (2q24.3), and is 113.5-kbases long. There are 29 exons in the gene, as characterised by Raymond et al. This work also showed how SCN9A, like other genes responsible for voltage-gated sodium channels, exhibit alternative splicing of some of these exons. This mechanism allows for even more variability in the resulting protein structure. Indeed, exon 5A of SCN9A is preferentially expressed in the peripheral nerves and central nervous system, whereas exon 5A was transcripted only in dorsal root ganglion neurones [6].

SCN9A polymorphism is responsible for structural differences in Nav1.7, which may lead to differences in channel activity. Reimann et al. [7] investigated the functional effects of rs6746030, which is a mutation in exon 18 involving a substitution of an amino acid at position 1150. Although peak currents and time of activation or fast inactivation were not different, slow inactivation was shorter in subjects with the minor allele A of rs6746030. Slow inactivation regulates the firing frequency of neurons, so this could explain how this mutation predisposes to a greater sensitivity to pain.

Polymorphisms in this gene are implicated in erythromelalgia and similar neuropathic pain syndromes [8], congenital insensitivity to pain [9] and possibly epilepsy [10, 11, 12], schizophrenia [13]. SCN9A is also associated with Paroxysmal Extreme Pain Disorder, which is characterised by skin flushing and episodes of severe pain [14]. Zhong et al. [15] also related propofol sensitivity to rs6746030, with carriers of the minor allele requiring lower propofol plasma concentrations for the same effect.

Estacion et al. [16] demonstrated that the single nucleotide change from the G allele to the A allele at rs6746030 results in a structurally different Nav1.7 that is more excitable. Indeed, rs6746030 has been implicated in higher pain scores in patients with lumbar disc herniation [17]. In a study of 27 different SNP’s of the SCN9A gene, rs6746030 was the most influential in over 1200 patients investigated, including in postoperative pain [7]. Specifically in a postoperative setting, Duan et al. investigated the role of rs6746030 in the prediction of post-operative pain following gynaecological laparoscopic surgery. The presence of the minor allele of the SNP resulted in a higher Numerical Rating Score [18].

Other SNP’s investigated have been less researched. rs11898284 has been shown to be associated with increased heat pain sensitivity [19]. Patients who carry the minor allele of rs11898284 appear to have worse outcomes after total knee arthroplasty [20].

One issue with research in SCN9A is the low frequency of some of the mutations investigated. This would mean that a large number of patients would need to be enrolled in a study to see any difference, especially in homozygous carriers of these mutations.

2.2 KCNS1

Potassium voltage-gated channels do not participate directly in signal transduction but are important in modulating the resting membrane potential. In this way, these channels either facilitate or inhibit an action potential from being generated [21].

Kcns1 is a Kv9.1 channel subunit, which is electrically silent on its own, but modulates channel properties when combined with other potassium channels [22, 23]. This is coded for by the KCNS1 gene, a small gene with around 11,000 base pairs found on chromosome 20 (20q13.12).

Experimental data shows that mice that lack KCNS1 suffer from a slight increase in acute pain under normal circumstances but show an exaggerated response after nerve injury [24]. Costigan et al. [22] also explored neighbouring genes and found that nearly 80% of these were involved in membrane signalling, with nearly half of these associated with nociception. They conclude that KCNS1 is central to many pathways that are integral to pain perception.

The most common polymorphism in KCNS1 studied so far is rs734784, which is found in exon 5. This missense SNP is common in the general population (around 40–45%) and leads to one isoleucine amino acid being changed to a valine residue. rs734784 has been associated with increased pain in volunteers and in patients with sciatica [22].

Costigan et al. [22] looked into the pain of 151 patients a year after lumbar discectomy and found an association of greater pain with rs734784. The mutation accounted for around 5% of the variance in pain scores in these patients. The same authors also demonstrated that rs734784 was more frequent in patients who had suffered from chronic phantom pain after an amputation.

In a study of 345 women who underwent an elective hysterectomy, Hoofwijk et al. [25] found no correlation between polymorphisms of KCNS1, including rs734784, and CPSP at 3 and at 12 months. Similarly, in 300 patients post-mastectomy, Langford et al. [26] did not find a difference in patients with or without this SNP. Costigan et al. [22] also did not find an association between pain at 12 months following surgery and rs734783.

On the other hand, Sciberras et al. found that patients homozygous for the C allele of rs734784 had significantly less WOMAC® scores throughout the study period [20]. Clinically, this translated to a WOMAC® score of nearly 4 points less, with a similar trend in pain scores.

Such contradictory findings are common in genetic studies. Differences in methodology, such as the use of a recessive or additive model may make a difference—Sciberras et al. used a recessive model, whereas Costigan et al. employed an additive model only.

3. Modulation of pain pathways involving catecholamines

Catecholamines are integral to the modulation of nociception. Levels of noradrenaline, adrenaline and dopamine modulate the transmission of nociceptive impulses through the spinal cord [27], and affect the perception of pain in the brain [28]. For instance, in normal healthy tissue, norepinephrine has little effect. However, after injury, levels of norepinephrine may correlate with either hyperalgesia or analgesia, depending on an interplay of different receptors and neuronal pathways. Furthermore, noradrenergic neurotransmitters such as dopamine also affect the brain itself. For instance, dopamine D-1 receptors are pronociceptive, whereas stimulation of D-2 receptors appears to be effective against tonic pain [29].

3.1 GCH1

Synthesis of catecholamines starts by uptake of tyrosine [30]. This is converted to dopamine by tyrosine hydroxylase, a process that requires tetrahydrobiopterin (BH4). This cofactor is produced by GTP cyclohydrolase 1, encoded by the GCH1 gene which is found on chromosome 14 and measuring around 60,800 base pairs.

In rats, BH4 levels have been associated with pain, specifically neuropathic pain. Tegeder et al. [31] demonstrated how axonal injury increased the upregulation of GCH1 and consequently levels of BH4 in primary sensory neurons. Inhibiting the increase in BH4 levels alleviated pain, whereas administering BH4 intrathecally exacerbated the pain.

In human volunteers, subjects who carried polymorphisms of GCH1 had less pain when a topical high concentration of capsaicin was applied to their skin [32]. In this small study, GCH1 was shown to be responsible for 35% of the inter-individual response to pain.

Tegeder et al. [31] were the first to describe a pain-protective haplotype made up of 15 polymorphisms in the GCH1 gene. In a study of 523 patients attending a tertiary care outpatient pain centre, homozygous carriers of this haplotype spent less time on specialised pain therapy [33], although the effect was small. This might be due to the small number of patients who had this haplotype of 15 specific SNPs: only around 14% of patients carried this haplotype, with only 10 subjects being homozygous carriers. Lötsch et al. [34] later reduced this haplotype to three main polymorphisms, including rs3783641. Their work showed that two SNPs predicted the pain-protective haplotype with nearly 100% sensitivity. These SNPs were rs8007267 and rs3783641. We also note that the presence of rs3783641 without rs8007267 occurs infrequently (1.4%), as shown in Table 1.

SNP	Change	Haplotypes
rs8007267*	G > A	G	G	A	G	G
rs2878172	T > C	T	T	C	C	C
rs2183080	G > C	G	G	G	C	G
rs3783641*	A > T	A	A	T	A	A
rs7147286	C > T	C	C	T	T	C
rs998259	G > A	G	A	G	G	G
rs8004445	C > A	C	C	C	A	C
rs12147422	A > G	A	A	A	G	A
rs7492600	C > A	C	C	C	A	C
rs9671371	G > A	G	G	A	G	A
rs8007201	T > C	T	T	C	T	C
rs4411417	A > G	A	A	G	A	A
rs752688	G > A	G	G	A	G	G
rs7142517	G > T	G	T	G	T	G
rs10483639*	C > G	C	C	G	C	C
		31.5%	19.8%	14.6%	9.7%	7.6%

Table 1.

Pain-protective haplotype of GCH1, as per Tegeder et al. [31].

^*

SNPs investigated by Lötsch et al.

Dark grey shading: pain-protective haplotype.

Tegeder et al. [31] also showed an effect of a pain-protective haplotype on pain scores 12 months after a lumbar discectomy. 162 patients were enrolled, with successful follow-up in 147 subjects. An additive effect of the haplotype was found: patients with no copy of the haplotype fared worse, patients homozygous for the haplotype were all better, and the heterozygous patients had an intermediate response. The authors themselves note that rs3783641 and rs8007267 would have contributed most to this effect.

Kim et al. [35] also showed a protective effect of rs998259 and the above-mentioned haplotype in 69 patients after surgical treatment of lumbar disc degeneration. These patients were followed up for 12 months. Functional scores improved more in patients with the minor allele of rs998259.

Contrary to these finding, the presence of rs3783641 actually increased the odds of CPSP at 3 and at 12 months, although this was not statistically significant, in patients after elective hysterectomy [25] and in patients after a total knee arthroplasty [20].

Multiple studies were either inconclusive or showed no effect of GCH1 on CPSP [36, 37, 38]. A meta-analysis of studies involving rs3783641 concludes that any associations demonstrated so far are probably spurious [39].

3.2 COMT

The COMT gene on chromosome 22 codes for the enzyme Catechol-O-MethylTransferase (COMT). This enzyme metabolises catecholamine neurotransmitters (dopamine, epinephrine and norepinephrine), by adding a methyl group [40]. COMT itself has been extensively studied as a possible therapeutic target, most notably in Parkinsonism.

The human COMT gene was first described by Tenhunen et al. [41]. It contains six exons, spanning over around 27,000 base pairs. Two promoters control the transcription of the gene into two different mRNA: MB-COMT and S-COMT. The former is found predominantly in brain neurones, whereas the latter is found more in other tissues such as the liver, kidney and blood.

Over 8000 single point mutations in the COMT gene are currently known. The fours most commonly studied in CPSP are rs4680, rs4633, rs4818 and rs6929.

The rs4680 mutation, also known as the Val 158 Met polymorphism has been extensively studied. rs4680 causes a structural change in the COMT enzyme, which lowers enzymatic activity. Hence, patients with the A variant will be able to metabolise catecholamines at a slower rate. The two variants are co-dominant, so heterozygous individuals will have an intermediate activity level [42]. It has been implicated in more severe low back pain [43], in patients with multiple sclerosis [44], and also in predicting the opioid consumption after surgery [45]. In the case of total knee replacements, Thomazeau et al. [46] found that the rs4680 mutation was more frequent (83%) in patients reporting chronic postsurgical pain, compared with 64% in the other patients. This conferred an odds risk ratio of 3.42 upon multivariate analysis.

Similar to rs4680, rs4633 affects COMT enzyme activity, although polymorphism at this site is not associated with structural changes of the enzyme itself. The T allele is associated with lower COMT activity, and the C allele with the higher COMT activity.

rs4818 is not associated with any structural changes, but polymorphism at this allele is associated with even more variation of the COMT enzyme when compared to rs4680. Patients who are homozygous for the G variant will have increased enzymatic activity. Heterozygous individuals will have intermediate activity, and homozygous individuals with the C variant will have the least enzymatic activity [47].

With regards to CPSP, the evidence for COMT is still somewhat inconclusive. Wang et al. [48] did not find a relationship between CPSP and the genotype of women who had undergone a caesarean section, but the number of patients with CPSP was admittedly small. On the other hand, in patients after TKA, Thomazeau et al. [46] found a borderline significance between the rs4680 A allele and chronic pain, with an odds ratio of 3.2, but the authors comment that the study was most likely underpowered to find significant differences. Rut et al. [49] demonstrated a protective association of the minor allele of rs4633 (T) in patients one year after a lumbar discectomy. However, the same study showed that the G allele of rs4680 was associated with a better outcome, not the minor A allele as in other studies. It is could be that COMT variations may have a different effect on different types of surgeries.

COMT polymorphisms are increasingly being researched as a haplotype, using rs6269, rs4633, rs4818 and rs4680 respectively as a haploblock: a region on a gene that has tends to be inherited as a whole. Diatchenko et al. [50] were the first to observe that these four polymorphisms produced seven haplotypes that had a frequency of more than 0.5%, as shown in Table 2. The most common three haplotypes account for over 95% of all haplotypes: these are the GCGG, ATCA and ACCG haplotypes. Patients with the GCGG haplotype possess the rs4818 mutation only, and these patients would have the highest COMT activity. Hence GCGG is classically defined as the Low Pain Sensitivity (LPS) haplotype. Conversely, ACCG is associated with the lowest COMT activity and is defined as the High Pain Sensitivity (HPS) haplotype. Finally, the ATCA haplotype confers intermediate COMT activity and is defined as the Average Pain Sensitivity (SPS) haplotype [52].

rs6269	rs4633	rs4818	rs4680	COMT activity	Pain	Frequency (%)
G	C	G	G	High	Least pain	36.8
A	T	C	A	Intermediate	Intermediate	54.6
G	C	C	G	Low	Most pain	7.0
A	C	C	A	Unknown	Unknown	1.7

Table 2.

Various haplotypes of the COMT gene, with relative COMT activity.

Adapted from [51].

For instance, Zhang et al. showed that patients with the haplotype ACCG had a higher fentanyl consumption than in patients with the haplotypes GCGG or ATCA [52]. This effect was not seen when individual SNP’s were analysed.

Contrary to the observations by Diatchenko [50], Sciberras et al. found that the TCA haplotype was linked to lower pain scores [20]. This was a different cohort of patients, and indeed in a similar group of patients, Rut et al. [49] found that rs4633 showed a protective effect. Another study of 69 patients after lumbar spinal surgery, this time by Dai et al. [53], also found that patients with the T allele for rs4633 had better functional outcomes after twelve months. Furthermore, the ATCA haplotype was associated with better outcomes. On the other hand, Machoy-Mokryńska et al. [54] observed higher levels of pain with the TCA haplotype.

One limitation of most studies is the lack of correlation between genetic polymorphism and enzymatic activity. This has been done by Dharaniprasad et al. [55], in 216 patients after cardiac surgery. rs4680 was associated with a 14-fold lower activity in COMT activity. Indeed, patients with this polymorphism all developed CPSP.

4. Pharmacogenetic response to analgesics

Genetics also play a role in the individual response to analgesics, through changes in receptors involved in nociception, or through changes in enzymes involved in the metabolism of these analgesics.

4.1 OPRM1

The MOP receptor, previously known as the μ-opioid receptor, is a G-coupled protein receptor that binds to endomorphins and endorphins [56]. Activation of the receptor leads to reduced cAMP intracellularly which causes a hyperpolarisation of the cell membrane [57]. The MOP receptor is mainly present in the central nervous system, especially in the periaqueductal grey zone. This is involved in descending inhibitory pathways that act on second-order neurons in the spinal cord to reduce nociception and hence induce analgesia.

The OPRM1 gene resides on the long arm of chromosome 6, and it is about 230,000 base pairs long over 18 exons [58]. Given the large size of the gene, it is not surprising that there are 3324 documented polymorphisms of the OPRM1 gene. Only 1395 of these variants have a minor allele frequency greater than 1% [59, 60].

The most commonly investigated variant is rs1799971, a mutation in exon 1 of OPRM1. The change of residue 40 from asparagine to aspartic acid creates a novel CpG-methylation site that prevents the upregulation of OPRM1 [56]. This change results in a three-fold increase in the binding of β-endorphin compared to the wild-type receptor [61]. One would expect that this would mean that subjects with rs1799971 would have an augmented response to opioids, but in fact, the opposite seems to be true. Lötsch et al. [62] demonstrated that the pupils constricted less in patients with the G allele and that this response was related to the number of G alleles.

rs1799971, also known as the A118G mutation, is frequently found in Asian populations (40–60%), less so in European populations (around 15%) and very infrequently in populations of African American descent (4%) [63]. It has been linked to a poor response to opiates in several studies, both in cancer pain and postoperatively. It has also been linked to alcoholism.

Other polymorphisms also show a strong association with pain sensitivity, although more work needs to be done to confirm such findings. Shabalina et al. [58] investigated 30 candidate SNPs over OPRM1, focussing on polymorphisms in exons and promoter genes. With nearly 200 Caucasian subjects, the authors showed that rs563649 and the rs2075572- rs533586 haplotype were associated with pain sensitivity. Furthermore, they showed that morphine produced less analgesia in subjects with at least one copy of rs563649, although statistical significance was not reached.

4.2 OPRK1 gene

The KOP receptor mediates analgesia without causing respiratory depression [64]. Indeed, although all opioids act on MOP receptors, some opioids such as morphine and oxycodone exhibit some activity also on KOP receptors.

The primary ligand to KOP is dynorphin, which induces analgesia. The KOP receptor is widely distributed in the central nervous system, including in the spinal cord and brainstem [65]. Dynorphin is emerging as an important factor in the development of chronic pain [66]. The pain appears to induce an increase in dynorphin levels in the spinal cord, as shown by Wagner et al. [67] in a neuropathic pain model in rats. This increase in dynorphin occurred 21 days after injury and was observed bilaterally in the spinal cord. It is not clear if such a consequence further augments chronic pain, or if this is protective [68]. Dynorphin injected intrathecally induces analgesia, but it has only been tested in animal models—unfortunately, it is associated with paralysis of the hind limbs when used in this manner. Caudle et al. postulate that dynorphin may act to reduce pain in the initial phases of injury: this effect has also been seen in knockout mice who had the KOP receptors deleted [69]. Such mice exhibited increased hyperalgesia after injury.

The gene that encodes for the KOP receptor is the OPRK1, which is present on 8q11.23. The human gene has been characterised only in 2004, and it is the gene responsible for the KOP opioid receptor [70]. It is 26,000 base pairs long on chromosome 8, spread over 4 exons.

Literature on OPRK1 polymorphisms and pain development is still scarce. One possible candidate polymorphism would be rs6985606, but most of such literature reflects research on opioid dependence [71] and on the analgesic response to opioids. rs6985606 has been shown to be a risk factor for pre-operative pain in a study of women with breast cancer who underwent breast surgery [72].

For instance, Kringel et al. [73] explored the use of a number of biomarkers that could be used to identify patients requiring high doses of opioids. Nine potential SNP’s in the OPRK1 gene were flagged for future research. However, Sciberras et al. [20] could not find any association between rs6985606 and CPSP in a cohort of orthopaedic patients.

5. Potential use in clinical practice

So far, there is little evidence of the use of genomic testing for CPSP in clinical practice. However, this has been done for other conditions, including pharmacogenetic-guided treatment of pain post-operatively (PGx). Senagore et al. reviewed the use of PGx in a series of patients, and found better pain scores and lower use of opioids in patients who had received pharmacogenetic testing prior to surgery [74].

Given that conclusive polymorphisms that predict CPSP with confidence are still to be determined, it might be difficult to recommend a specific panel of assays for pre-operative evaluation. However, the techniques in genotyping are continuously being refined, and automated batch-testing is possible. Furthermore, the costs of such testing is becoming more commercially-viable, so it may be possible in the near future to test individual patients for a number of polymorphisms and calculate a predicted risk of CPSP.

The clinical impact of such information is still debatable. CPSP is difficult to treat, and may resolve spontaneously with time. However, if a patient is identified as having a high risk of developing CPSP, one may refer such cases to a dedicated chronic pain clinic for follow-up and treatment.

6. Conclusion

Genetic factors appear to be important in predicting the individual progression from acute to chronic post-surgical pain. However, the exact impact and the interplay between different combinations of polymorphisms are still to be determined.