Cytokine Genes and Pain Severity in Lung Cancer: Exploring the Influence of TNF-α-308 G/A IL6-174G/C and IL8-251T/A

Cancer pain affects 17 million people worldwide. Lung cancer is one of the most common cancers to cause pain. Up to 50% of patients with small cell lung cancer (SCLC) report pain, and estimates increase to 65% for patients with non-SCLC (NSCLC; ref. 1). Cancer pain results from several mechanisms, including the primary activation of visceral or somatic nociceptors by a primary or metastatic tumor (nociceptive pain), impingement of the tumor on adjacent tissues, obstruction of blood vessels, or as a result of actual damage to the peripheral or central nervous system (neuropathic pain). More recently, studies suggest that inflammation caused by tumor-induced mediators, such as cytokines (2), may also be a potential mechanism for cancer-related pain.

Cytokines are soluble proteins or glycoproteins that act as mediators of cell-to-cell communications and are integral to the function of immune cells. Cytokines are pleiotropic, and their multiple overlapping functions depend on their local concentration, the type and the maturational stage of the responding cell, and the presence of other cytokines and their mediators.

The role of cytokines in the biology of various neoplastic disorders has been extensively studied. Cytokines are aberrantly produced by cancer cells, macrophages, and other phagocytic cells associated with tumor-producing cytokines (3, 4). These tumor-induced mediators are a potential mechanism for cancer-related cachexia, loss of appetite, or enhanced energy expenditure (5). More recently, research in the neurosciences suggests an important role of cytokines in cancer-related pain as well (6-9).

Variants in genes encoding for cytokines have been suggested as candidates for risk of a variety of cancers (3, 10-12), but their role has not been assessed in the epidemiology of cancer-related symptoms. We therefore explored the role of such variants in explaining the variability in pain in lung cancer patients.

Tumor necrosis factor-α (TNF-α) is a proinflammatory cytokine involved in the regulation of a wide spectrum of biological processes, including cell proliferation, differentiation, apoptosis, lipid metabolism, and coagulation. A neuroprotective function of this cytokine (13-15) and an important role in cancer-related symptoms including sickness-induced pain facilitation and enhancement have been suggested (8, 9). Genetic polymorphisms in the promoter region of the TNF-α gene could modulate protein expression. Specifically, the G to A transition in the promoter region at position -308 has been shown to result in higher expression levels of TNF-α in vitro and in vivo (10, 16).

Interleukin-6 (IL-6) is a proinflammatory cytokine released in response to infection, trauma, and neoplasia. It has a key role in immune and acute-phase response and hematopoesis. IL-6 exhibits a significant amount of pleiotropism and can modulate the effects of other cytokines and interact with glucocorticoids. The-174 G/C polymorphism is a single nucleotide polymorphism (SNP) that affects the transcription of the IL-6 gene, thus altering the serum levels of IL-6. Homozygotes for the G allele have been shown to have higher plasma IL-6 levels, higher IL-6 gene transcriptional activity, and higher inducible IL-6 responses than subjects homozygous for the C allele. The C allele has also been shown to be associated with adverse clinical outcomes such as less favorable prognosis after abdominal aortic aneurysms and poorer survival after sepsis (17-19).

IL-8 is a chemokine and one of the major mediators of the inflammatory response. IL-8 is secreted by several cell types, functions as a chemoattractant, and is also a potent angiogenic factor. IL-8 has also been shown to induce the production of other cytokines, including IL-6 and TNF-α (20). The IL-8 gene is believed to play a role in the pathogenesis of cancer. Animal and human studies have begun to suggest its role in pain facilitation and enhancement. However, its association with symptom distress and other cancer-related symptoms is still speculative (21). The A allele of an IL-8 SNP in the promoter region (-251T/A) has been associated with increased IL-8 production (22).

In this preliminary analysis, we explored the extent to which functional polymorphisms in TNF-α -308 G/A, IL-6 -174G/C, and IL-8 -251T/A are associated with the severity of pain using a sample of 446 Whites, 125 African-Americans, and 35 Hispanics with newly diagnosed, previously untreated patients with histologically confirmed lung cancer enrolled in an epidemiologic study of lung cancer. Lung cancer is a lethal disease marked by debilitating symptoms; understanding genetic factors associated with the severity of pain in lung cancer is crucial for early symptom intervention and may help inform treatment options.

The study sample was drawn from an ongoing previously described case-control study of lung cancer (23). Case patients with histologically confirmed lung cancer who were newly diagnosed were recruited at the time of initial registration at the cancer center and before initiation of radiotherapy or chemotherapy. There were no restrictions with regard to age, sex, ethnicity, or disease stage. All cases were residents of the United States recruited from 1999 to 2005 and for whom pain and genetic data were available. (We selected this start date because from 1999, all patients are asked to respond to self-administered questionnaires on pain and quality of life during initial registration at the M.D. Anderson Cancer Center). The overall response rate for the study was 80%. This study was approved by the Institutional Review Board at M.D. Anderson Cancer Center, and all participants provided written informed consent.

Trained M.D. Anderson Cancer Center staff interviewers collected data on demographics, smoking history, and history of cancer. Patients rated their pain on an 11-point numerical scale, (0 = “no pain” and 10 = “pain as bad as you can imagine”). This item was taken from the Brief Pain Inventory, a recommended standard for pain assessment in clinical studies of pain (24). Because studies show a high correlation between depression, fatigue, and pain, we also assessed depressed mood and fatigue using the following items “during the past 4 weeks, have you felt downhearted and blue?” and “during the past 4 weeks, did you have a lot of energy?” These items were taken from the SF-12. The SF-12 is a validated measure of quality of life and is extensively used in studies of cancer patients (25-28). Patients rated these symptoms upon presentation at the M.D. Anderson Cancer Center before any cancer treatment. Clinical data, including stage of disease and history of comorbid conditions (heart disease, stroke, diabetes, etc.), were abstracted from patients' charts.

After the interview was completed, a 40-mL blood sample was drawn into coded heparinized tubes. Genomic DNA was extracted from peripheral blood lymphocytes by proteinase K digestion, followed by isopropanol extraction and ethanol precipitation. DNA samples were stored at −80°C. We did genotyping using a 5′-nuclease assay with dual fluorescent reporter probes VIC and FAM. The primer and probe sequences for each SNP can be found in the report by Landi et al. (29). Reactions were completed and read in a 7900 HT TaqMan sequence detector system (Applied Biosystems). Amplification mixes (5 μL) contained sample DNA (5 ng), 1× TaqMan buffer A, deoxynucleotide triphosphates (200 μmol/L), MgCl₂ (5 mmol/L), AmpliTaq Gold (0.65 units), each primer (900 nmol/L), and 200 nmol/L of each probe. The thermal cycling conditions consisted of one cycle for 10 min at 95°C, 40 cycles for 15 s at 95°C, and 40 cycles for 1 min at 60°C. Sequence Detector System (SDS) version 2.1 software (Applied Biosystems) was used to analyze end point fluorescence according to the allelic discrimination technique. Water control, ample internal controls, and previously genotyped samples were included in each plate to ensure the accuracy of genotyping.

Descriptive statistics were used to summarize patient characteristics. The Kolmogorov-Smirnov Z test was used to assess normality distribution for pain severity. Because normality was not met, we used the National Comprehensive Cancer Network (NCCN) cutoff score for severe pain (30). A score ≥7/10 is considered as a pain emergency, and treatment is initiated with short-acting opioids.

Comparisons across specific genotypes and levels of pain were done using Pearson's χ² analyses. Multivariable logistic regression analyses were done to assess the extent of association between the genotypes and severe pain. Variables found in the univariate analysis to be significantly associated with pain at a level of P < 0.20 (31) were included in the multivariable model. A P value of 0.20 was used as the cutoff because using a more traditional level (P < 0.05) might fail to identify variables known to be important (31). Further variable selection in the model was conducted by using a backward elimination approach. With the goal of having the most parsimonious model, only variables with P < 0.05 were included in the final multivariable model.

A significance level of 5% (two-sided) was used for the final analyses. Because of the multiplicity of tests conducted, this significance level can result in false-positive results; however, this level was considered acceptable for this exploratory study (32).

Of the 695 patients with previously untreated and histologically confirmed lung cancer, 78% were White Caucasians and 21.7% were Hispanic, African-American, Asian, American-Indian, and of other ethnicities. A majority (95%) of the sample had histologically confirmed NSCLC, and only 5% (n = 30) presented with SCLC. Table 1A shows pain distribution for Caucasian whites, African-Americans, and Hispanics with NSCLC. The mean (4.33) and median (5.0) pain was highest for African-Americans, followed by Hispanics (mean = 3.2; median = 2.0) and lowest for Whites (mean = 2.9; median = 2.0). The Kolmogorov-Smirnov Z test showed that pain severity was not normally distributed for White (Z = 4.400; P = 0.0001), African-American (Z = 1.854; P = 0.0001), or Hispanic patients (Z = 1.660; P = 0.0001). Using NCCN guidelines for categorizations of pain severity, we observed that African-Americans (31.5%) had the highest proportion of patients reporting severe pain, followed by Hispanics (20%) and Whites (17%; P < 0.001).

Table 1B shows that a majority of African-Americans were carriers of the TT genotype in IL-8, whereas the majority of Whites and Hispanics were heterozygotes in IL-8. Most patients were GG carriers in TNF-α, irrespective of ethnicity. The majority of African-Americans (83.6%) and Hispanics (70.5%) were IL-6 GG carriers, but there were more GC carriers (45.5%) among Whites.

Table 2 shows the distribution of severe pain by genotype groups among Caucasian Whites. Assuming a dominant model, we observed that although a higher proportion of patients with variant alleles in TNF-α, IL-6, and IL-8 reported severe pain, only IL-8 (TT = 13%; TA + AA = 87%; P = 0.04) was significantly associated with severe pain.

We also present genotype and pain data for African-Americans and Hispanics in Table 3. We did not observe any significant association between genotypes in TNF-α, IL-6, and IL-8 and severe pain for either subgroups.

Given the small number of patients from other ethnicities and to avoid issues related to population stratification, we restricted our subsequent analyses to non-Hispanic whites with NSCLC and with complete information on the variables of interest (n = 446). The majority of these White patients presented with adenocarcinoma (62%), followed by squamous cell carcinoma (20%). Fifty-two percent presented with advanced-stage disease (stages IIIB and IV). There were more males than females (53% versus 47%, respectively), and a majority were aged 50 or more (78%). Hypertension was the most prevalent (36%) comorbid condition.

Table 4 shows that the prevalence of severe pain significantly varied by stage of disease (early stage, 11%; advanced stage, 22%; P = 0.001); age (<50 years, 26%; ≥50 years, 14%; P = 0.01); sex (male, 14%; female, 20%; P = 0.05); depressed mood (none to mild, 15%; moderate to severe, 41%; P = 0.001); and fatigue (none to mild, 8%; moderate to severe, 28%; P = 0.001). None of the comorbid conditions were associated with severe pain.

Table 5 presents the results of the multivariable logistic regression analyses. We found that variant alleles in IL-8 [odds ratio (OR), 2.35; 95% confidence interval (95% CI), 1.10-5.03;P = 0.027] persisted as a significant factor for severe pain. Other factors included moderate to severe levels of depressed mood (OR, 3.67; 95% CI, 1.57-8.54;P = 0.003) and fatigue (OR, 3.40; 95% CI, 1.91-6.05; P = 0.0001), advanced stage of disease (OR, 2.25; 95% CI, 1.3-4.0; P = 0.006), and age <50 years (OR, 1.87; 95% CI, 1.03-3.40; P = 0.039).

We also evaluated whether the influence of polymorphisms in IL-8 on pain severity varied by the level of the other factors (age, depressed mood, and fatigue) and found that there were no significant interactions between these factors and IL-8.

Studies of cancer-related pain have traditionally focused on the influence of disease-related variables (stage of disease), clinical health status (performance status, comorbid conditions), and sociodemographic characteristics (age, gender, race, and marital status). This study assessed the contribution of genetic variation in select cytokine genes to the severity of pain in a large sample of patients newly diagnosed with NSCLC. We found that both genetic and nongenetic factors were relevant in the severity of pain and specifically found an association between polymorphisms in IL-8 and severe pain in whites. To our knowledge, this is among the first studies to provide preliminary evidence that functional polymorphisms in cytokine genes associated with increased gene expression are associated with higher levels of pain in a large sample of White patients newly diagnosed with lung cancer.

Although the exact molecular mechanism by which cytokines influence pain has not been fully elucidated, studies suggest that cytokines released during inflammation or tissue damage (as in the cancer process) modify the activity of nociceptors contributing to pain hypersensitivity. Clinical studies show elevated IL-8 levels in patients with chronic pain conditions such back pain (33), post-herpetic neuralgia (34), and unstable angina (35). In animal studies, Cunha et al. (36) measured the hyperalgesic effect of IL8 in a rat paw pressure test and found that IL-8 evoked dose-dependent hyperalgesia. The hyperalgesia was blocked by specific anti–IL-8 serum, suggesting that IL-8 causes hyperalgesia. Ribeiro et al. (37) also observed that i.p. administration of a specific antiserum against IL-8 partly blocked the nociceptive response in mice. Other studies suggest that chemokines including IL-8 may produce enhanced sensitivity to pain via direct actions on chemokine receptors expressed by nociceptive neurons (38).

The observed allele frequencies in our population were similar to other studies (39, 40). Consistent with our findings, the presence of the -251A allele in IL-8 has been associated with adverse clinical outcomes, including colorectal adenoma (40) and gastric cancer (39). Studies have also suggested that other cytokines, such as proinflammatory IL-6 and TNF-α, cause hyperexcitability in pain transmission neurons, and the exaggerated release of substance P and excitatory amino acids from presynpatic terminals produces an exaggerated pain response (7, 41). However, we did not observe a relationship between TNF, IL-6, and pain severity in this study. Of note is that IL-8 has been shown to induce the production of other cytokines such as IL-6 and TNF-α (20).

We found depressed mood and fatigue as significant correlates of pain in the univariate analyses, and this pattern persisted even when we took into account the impact of coexisting medical conditions and sociodemographic factors. Pain, depression, and fatigue are typically experienced as the most prevalent symptoms among patients with cancer, and many have argued that because they occur as a cluster of symptoms or as a pain syndrome, there may be a shared biological mechanism for these symptoms. Because inflammatory cytokines have been associated not only with pain but also with depression and fatigue, it might be possible that in a polygenic model for pain, other cytokines involved in the modulation of nociceptive input are also related to fatigue and depression. Future studies should measure the frequency and severity of depressed mood/depression and fatigue in patients with cancer pain.

It is also important to note that other studies have addressed the relationship of depression and pain and have found depression as having either a causal or a mediating effect on pain (42). However, we were unable to assess the directionality of the relationship between depressed mood and severe pain due to the cross-sectional nature of our data set. Prospective studies are needed to assess if these symptoms covary over the course of the disease process. Importantly, our analyses showed that genetic polymorphism in IL-8 remained as a significant factor for pain severity, independent of the effect of depressed mood and fatigue.

Consistent with other studies, we found that ethnic minorities, those with advanced stage of disease, and younger age were more likely to report severe pain relative to their counterparts. In a population-based study of cancer survivors, non-Hispanic African-Americans (OR, 1.78; 99% CI, 1.33-2.37) and Hispanics (OR, 1.80; 99% CI, 1.26-2.56) were found to have higher risk for severe pain compared with non-Hispanic whites (43). Studies also show that up to 40% of patients receiving active cancer treatment report pain, and these increase to 80% for patients with advanced cancer (44, 45). It is also documented that whereas older patients are more likely to experience pain, they are less likely to complain of pain (46).

There are limitations to our study. We lack data on analgesic intake, and thus, we were unable to adjust for its potential confounding effect in the analyses. Another limitation is that using a single item to measure pain, depressed mood, and fatigue may be an inadequate assessment of these symptoms. It should be noted, however, that other studies have found that single-item questions correlate well with assessment tools for fatigue and depression (47, 48).

Many have also argued for the need for a polygenic model for pain and pain-related phenotypes. Ideally, many genes with functional significance should be assessed, especially because it is known that there is considerable interaction between different cytokines, forming networks that initiate gene activation and suppression (49). It is also important to note that inclusion of multiple SNPs per gene would have allowed examination of haplotypes, which may have provided different outcomes. In addition, the tested SNP may not necessarily be the functional SNP, but may be in linkage disequilibrium with the true functional SNP. We also have a very small sample of African-Americans and Hispanics in our study. Given that ethnic minorities are at higher risk for severe pain, additional studies are needed to further explore the association between severe pain and cytokine genes in these populations.

In conclusion, modest improvements in the survival rates of patients with lung cancer have been observed in recent years. However, the majority of patients still suffer from both disease- and treatment-related symptoms. Previous studies have focused on clinical health status and sociodemographic characteristics in understanding cancer-related pain. Because genetic polymorphisms are stable markers, our observation that genetic differences in cytokine genes is a significant correlate of pain may prove useful in developing risk models for pain and, importantly, in developing personalized therapies. Furthermore, with experimental studies on cytokine receptor antagonists or other cytokine inhibitors currently under way, the use of genetic polymorphisms, arguably stable markers for cytokine function, could help in understanding patients who might benefit most from symptom intervention using cytokine products. Our findings need to be validated in large, prospectively accrued populations and incorporating additional genetic markers in the cytokine pathway.