- Email: [email protected]
Depression, inflammation, and epidermal growth factor receptor (EGFR) status in metastatic non-small cell lung cancer: A pilot study
Journal of Psychosomatic Research 99 (2017) 28–33
Contents lists available at ScienceDirect
Journal of Psychosomatic Research journal homepage: www.elsevier.com/locate/jpsychores
Depression, inflammation, and epidermal growth factor receptor (EGFR) status in metastatic non-small cell lung cancer: A pilot study☆
MARK
Jamie M. Jacobsa,⁎, Lara Traegera, Justin Eusebioa, Naomi M. Simonb, Lecia V. Sequistc, Joseph A. Greera, Jennifer S. Temelc, William F. Pirla,1 a Center for Psychiatric Oncology and Behavioral Sciences, Department of Psychiatry, Massachusetts General Hospital Cancer Center/Harvard Medical School, 55 Fruit St., Yawkey Center for Outpatient Care, Suite 10B, Boston, MA 02114, United States b Center for Anxiety and Traumatic Stress Disorders, Massachusetts General Hospital/Harvard Medical School, One Bowdoin Square, 6th floor, Boston, MA 02114, United States c Massachusetts General Hospital Cancer Center/Harvard Medical School, 55 Fruit St., Yawkey Center for Outpatient Care, Suite 7B, Boston, MA 02114, United States
A R T I C L E I N F O
A B S T R A C T
Keywords: Non-small cell lung cancer Genotyping Epidermal growth factor receptor Depression Inflammatory cytokines Oncology
Objective: Patients with stage IV non-small cell lung cancer (NSCLC) have high risk for depressive symptoms and major depressive disorder (MDD); however, those with epidermal growth factor receptor (EGFR) mutations may have decreased risk. The biological underpinning of this relationship is unknown. We examined differences in depression severity and MDD in patients with newly diagnosed stage IV NSCLC based on EGFR mutation status, and examined proinflammatory cytokines and growth factors known to play a role in cancer progression and depression. Methods: Fifty-five patients with newly diagnosed stage IV NSCLC completed self-report and clinician-administered depression assessments prior to receiving results of tumor genotyping. We measured serum levels of circulating biological markers of inflammation: IL-1β, IL-6, TGF-α, and TNF-α. We examined differences in depression severity, MDD, and inflammatory biomarkers in patients with and without EGFR mutations. Results: Patients with EGFR mutations (n = 10) had lower depression severity (t[43] = 2.38, p = 0.03) than those without EGFR mutations (n = 38) and fewer patients with EGFR mutations had concurrent MDD (2.08%) relative to those without mutations (27.08%). Patients with MDD had higher levels of TNF-α than those without MDD (t[40] = 2.95, p = 0.005). Those with EGFR mutations exhibited higher levels of TNF-α relative to those without EGFR mutations (t[35] = 2.17, p = 0.04). Conclusions: Patients with stage IV NSCLC harboring an EGFR mutation exhibited elevated proinflammatory marker TNF-α, yet had lower depression severity than patients without EGFR mutations. More work is warranted to examine the interaction between tumor genotyping and inflammatory cytokines in the context of depression.
1. Introduction Significant advances in our understanding of cancer biology and tumor gene mutations over the last decade has led to certain cancers, such as non-small cell lung cancer (NSCLC), to be classified by their genotype signature rather than mere histology. Particular tumor gene mutations implicated in cancer growth in subsets of patients confer sensitivity to specific targeted therapies. In the molecular landscape of cancers, patterns of cancer-related symptoms might also emerge for tumors with certain genotypes. Few studies have looked at the
molecular basis of depressive symptoms in patients with cancer. Although depression is common in patients with cancer, affecting an estimated 10–25%, it is unclear why it occurs in some patients and not others [1,2]. Patients with medical co-morbidities may be less likely to respond to pharmacological depression treatments, suggesting that there may be differences in depression between those with and without specific co-morbidities [3,4]. In addition, depression is associated with elevated inflammation [5], and risk for depression may be higher in patients with medical illness and heightened inflammatory processes. Tumor gene mutations may represent a novel opportunity to examine
☆ Conflict of interest: Dr. Simon discloses spousal equity in G1 Therapeutics and Gatekeeper. Dr. Temel discloses research funding from Pfizer. All other authors declare that they have no conflicts of interest or disclosures. Acknowledgement: This work was supported by PC-1116401, American Cancer Society (Pirl). ⁎ Corresponding author. E-mail address: [email protected] (J.M. Jacobs). 1 Present address: Sylvester Comprehensive Cancer Center, Courtelis Center, University of Miami Miller School of Medicine, 1120 NW 14th Street, Miami, FL 33156.
http://dx.doi.org/10.1016/j.jpsychores.2017.05.009 Received 13 January 2017; Received in revised form 6 May 2017; Accepted 11 May 2017 0022-3999/ © 2017 Elsevier Inc. All rights reserved.
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
under the direct supervision of a psychiatrist. Patients who completed the interview, self-report instruments, and blood draw were compensated $50.00 for their time.
risk of developing depression after a cancer diagnosis. One of the first targetable mutations to be identified in NSCLC was the family of epidermal growth factor receptor (EGFR) mutations [6]. Patients with these mutations respond well to the EGFR tyrosine kinase inhibitors (TKIs; e.g., gefitinib, erlotinib and afatinib) and thus have a better prognosis than patients without an EGFR mutation [7–9]. We previously observed that patients with stage IV NSCLC with EGFR mutations report less depression than those without EGFR mutations, even when depression was measured before the results of genotyping were known [1]. Yet, without a biologically plausible link between NSCLC EGFR mutation status and depression, the meaning of any observed association is limited. One hypothesis is that NSCLC cells with different genotypes may produce different magnitudes of inflammatory cytokines and growth factors, which are associated with and may contribute to depression [10]. The goal of this work was to conduct a new study as a first step towards understanding potential links between EGFR mutation status and depression in patients with newly diagnosed stage IV NSCLC. The objective of this pilot study was to gather initial data on possible relationships between depression, inflammation, and EGFR mutation status. Specifically, we hypothesized that 1) Patients harboring an EGFR mutation would have lower depression severity and be less likely to be diagnosed with major depressive disorder (MDD) than patients without an EGFR mutation, and 2) EGFR mutation status and MDD status would be associated with serum levels of inflammatory cytokines and growth factors. Patients harboring an EGFR mutation and those without major depression would have lower levels of inflammation than those without an EGFR mutation. If a signal for relationships exists, this would justify larger studies adequately powered to adjust for potential confounders.
2.3. Demographics and clinical characteristics Study staff collected patient data from the electronic health record, including demographic information, cancer staging and histology, pain level, smoking history (i.e., current smoking status and pack years), prescribed medications (i.e., antidepressants), and Eastern Cooperative Oncology Group (ECOG) performance status [12]. 2.4. Depression severity To evaluate the severity of depression, we used the PHQ-9, [11] a self-report measure designed to screen for MDD according to the criteria of the Diagnostic and Statistical Manual-IV (DSM-IV [13]). The PHQ-9 has been validated for use with patients with cancer and has been used in clinical trials [14,15]. Patients are asked how often they have been bothered by symptoms of depression over the past 2 weeks, with response options ranging from 0 (Not at all) to 3 (Nearly every day). Scores are summed to obtain a total severity score ranging from 0 to 27. The PHQ-9 demonstrated good reliability in the current sample (α = 0.79). 2.5. Major depressive disorder Presence of current MDD was assessed with the Mini-International Neuropsychiatric Interview (MINI [16,17]); a brief, structured, clinician-administered diagnostic interview that screens for major Axis I psychiatric disorders according to the DSM-IV [13] and ICD-10 criteria [18]. The widely used MINI has high reliability and validity comparable to the Structured Clinical Interview for the DSM-IV Axis I Disorders [16]. Interview questions assess each dimension of the disorder, including time frame, frequency, and severity. Items require “yes” or “no” answers. Patients are given a diagnosis of MDD if symptom criteria for MDD were met, and criteria for other disorders with depressive features are not met.
2. Materials and methods 2.1. Participants Eligible patients were those who had a new diagnosis of stage IV NSCLC; had genotyping ordered on a biopsy specimen; had no prior knowledge of their NSCLC genotypes; possessed the ability to understand and sign informed consent; were not at their appointment for a second opinion and/or planning to receive care elsewhere; had not previously received chemotherapy or radiation for current stage IV NSCLC; had no dementia or cognitive impairment; and had no diagnosis of a psychotic or bipolar disorder.
2.6. Genotyping Tissue from patients' biopsies was genotyped in a CLIA-certified clinical laboratory at MGH as part of standard care with SNaPshot, multiplex PCR-based allele-specific analysis of known cancer hotspots across 13 genes [19]. Genotype results were available 2–4 weeks later and were posted in the electronic health record. For the present analysis, patients who harbored an EGFR mutation were compared to patients without an EGFR mutation.
2.2. Procedures This study took place at Massachusetts General Hospital (MGH) Cancer Center and was approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board. All procedures contributing to this work comply with ethical standards of the relevant national and institutional committees on human subjects' research and with the Helsinki Declaration of 1975, as revised in 2008. Consecutive new patients presenting at the MGH Multidisciplinary Thoracic Oncology Clinic were recruited to participate in this study at their initial appointment between July 26, 2011 and August 8, 2013. Research assistants reviewed preliminary information on patients collected by oncology nurses, and those who had suspected advanced NSCLC were approached to participate. Patients interested in participating were screened for eligibility criteria. Eligible and interested patients signed informed consent after review with study staff. Prior to receiving results of their genotyping and before initiating any cancer treatments, patients completed a self-report survey, a clinician-administered diagnostic interview, and blood collection. Within the self-report survey, the Patient Health Questionnaire-9 [11] was always administered in person by paper or by telephone. The diagnostic interview was administered by a trained Clinical Research Coordinator
2.7. Serum cytokines and growth factors Blood samples were taken by a staff phlebotomist or in the MGH Thoracic Oncology Clinic prior to the initiation of chemotherapy. Samples were collected by venipuncture in 10 mL tubes containing EDTA and placed on ice. Serum levels of circulating Tumor Necrosis Factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6) were quantified measured using a ultra-sensitive multiplex immunoassay system “Human Proinflammatory II 4-Plex Assay” with electrochemiluminescent detection from Meso Scale Discovery® Multi-array Technology (Meso Scale Discovery, Gaithersburg, MD). Biomarkers were measured by commercially available reagents and were performed according to the manufacturer's instructions. No dilutions were required or used. The mean lower limits of detection for TNF-α, IL-6, IL1β were 0.37 pg/mL, 0.26 pg/mL, and 0.18 pg/mL, respectively. Serum levels of Transforming Growth Factor-α (TGF-α) was measured with the Quantikine® Human TGF- α Immunoassay, a 4.5 hour solid-phase 29
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
ELISA (R & D Systems, Inc., Minneapolis, MN). Samples were stored at ≤− 20 °C and assays were performed according to the manufacturer's instructions. Samples were not diluted per protocol instructions. The mean detectable limit for TGF-α was 21.9 pg/mL.
Table 1 Sociodemographic and clinical characteristics in the full sample (n = 55) and by EGFR mutation status. Characteristic
n (%) or M (SD)
EGFR mutation (n = 10)
EGFR wild-type (n = 38)
Age (n = 54; range = 42–83) Female gender Race White Asian Black or African American Other Missing Ethnicity Non-Hispanic Unknown Marital status Married or living as if married Single Divorced or separated Loss of long-term partner/ widowed Unknown Education < 12th grade High school graduate or GED Some college/vocational/ technical College graduate or postgraduate Unknown ECOG performance status score 0 1 2 3 Missing Brain Metastases Smoking Status Current or former smoker Never smoker Missing Not on antidepressants EGFR mutation (vs. wildtype) Deletion exon 19 L858R G719A Exon 20 insertion
62.09 (11.73) 22 (40.0%)
62.60 (12.45)
62.32 (10.12)
8 (80%)
11 (28.9%)
51 (92.7%) 1 (1.8%) 1 (1.8%) 1 (1.8%) 1 (1.8%)
8 1 0 1
(80%) (10.0%) (0.0%) (10.0%)
37 (97.4%) 0 (0.0%) 1 (2.6%) 0 (0.0%)
53 (96.4%) 2 (3.6%)
10 (100.0%) 0 (0.0%)
37 (97.4%) 1 (2.6%)
35 (63.6%)
6 (60.0%)
26 (68.4%)
4 (7.3%) 8 (14.5%) 6 (10.9%)
1 (10.0%) 1 (10.0%) 1 (10.0%)
2 (5.3%) 6 (15.8%) 4 (10.5%)
2 (3.6%) 3 (5.5%) 18 (32.7%) 9 (16.4%)
1 (10.0%)
0 (0.0%)
0 (0.0%) 0 (0.0%)
3 (7.9%) 15 (39.5%)
21 (38.2%)
2 (20.0%)
6 (15.8%)
7 (70.0%)
12 (31.65)
4 (7.3%)
1 (10.0%)
2 (5.3%)
18 (32.7%) 30 (54.5%) 2 (3.6%) 3 (5.5%) 2 (3.6%) 18 (37.5%)
4 3 1 1 1 2
(40.0%) (30.0%) (10.0%) (10.0%) (10.0%) (20.0%)
13 (34.2%) 22 (57.9%) 1 (2.6%) 2 (5.3%) 0 (0.0%) 16 (42.1%)
44 (80.0%) 10 (18.2%) 1 (1.8%) 45 (81.8%) 10 (18.2%)
2 (20.0%) 8 (80.0%)
37 (97.4%) 1 (2.6%)
9 (90.0%)
29 (76.3%)
2.8. Statistical analyses Data were analyzed with the Statistical Package for the Social Sciences (SPSS, Version 21.0). All biological data were screened for outliers and cleaned by converting values outside three standard deviations from the mean to equal the next highest value [20]. To meet assumptions of normality, biological data were log-transformed, and a value of one was added to maintain positive numbers. Due to the small sample size, two-sided Fisher's exact test was used to test for differences in categorical dependent variables. Independent samples t-tests were conducted to assess differences in continuous dependent variables, and estimates were interpreted at a two-tailed significance level (p < 0.05) and 95% confidence intervals. Levene's test was used to evaluate the assumption of equality of variance, and a correction was used if this assumption was not met. Corresponding effect sizes (0.20 = small; 0.50 = medium; 0.80 = large) were evaluated to assess the magnitude of the observed significant differences [21]. First, in an effort to replicate prior findings [1], independent samples t-tests were conducted to test for group differences in depression severity on the PHQ-9 based on EGFR mutation status. Power analyses based on this work suggested that, with 50 subjects and a 5% Type I error rate, there is 80% power to detect a minimum mean group difference of 3.3 points in depression severity on the PHQ-9 in patients with EGFR mutations versus those with wild-type. Given the clinical meaningfulness and usefulness of the MINI to identify presence of MDD, the remainder of analyses examined differences based on whether patients met criteria for current MDD. Fisher's exact test was used to test whether presence of MDD differed by EGFR mutation status. Next, independent samples t-tests were used to test for differences in inflammatory cytokines and growth factors based on EGFR mutation status. In addition, independent samples t-tests were used to test for differences in inflammatory cytokines and growth factors in patients who carried a diagnosis of MDD vs. those who did not. 3. Results Fifty-eight patients enrolled in the study (Table 1). Three patients were subsequently excluded after further oncology work-up determined them to be at an earlier stage of disease. The remaining 55 patients were an average of 62.09 years (standard deviation [SD] = 11.73 years), and more than half were male (58.2%). Most of the sample (96.4%) self-identified as non-Hispanic white. Overall, patients had good performance status, with ECOG performance status scores of 0 or 1 (90.6%), and 80% were current or former smokers. Approximately one quarter of the sample (n = 14) had current MDD on the MINI, and the mean PHQ-9 depression severity score for those with current MDD was 12.17 (SD = 4.74). The mean PHQ-9 depression severity score for the total sample was 7.28 (SD = 5.42); see Table 2. Of the 55 patients, 38 (69.1%) had cancers that were wild-type for EGFR, ten (18.2%) harbored EGFR mutations, and seven (12.7%) lacked tissue suitable for genotyping. Patients who were wild-type for EGFR were more likely to be male, while those with EGFR mutations were more likely to be female (Fisher's Exact Test, p = 0.008). Those who were wild-type for EGFR mutations were more likely to be current or former smokers than those harboring EGFR mutations (Fisher's Exact Test, p < 0.001). No other differences were found in any other sociodemographic or clinical characteristics (Table 1). A total of 50 patients had serum available for biological assays, with 45 having both biological and genotype data available for analysis. One value was adjusted for each inflammatory marker due to being greater than three standard deviations above the mean. This sample size is consistent with
5 3 1 1
(50%) (30%) (10%) (10%)
Abbreviations: SD, standard deviation; EGFR, epidermal growth factor receptor; ECOG, Eastern Cooperative Oncology Group. Table 2 Psychosocial and biological characteristics (n = 55). Characteristic
N (%) or mean (SD)
Range
Current major depressive disorder (MINI) Depression severity (PHQ-9; n = 50) TNF-α in pg/mL (n = 50) TNF-α (log transformed) IL-1β in pg/mL (n = 50) IL-1β (log transformed) IL-6 in pg/mL (n = 50) IL-6 (log transformed) TGF-α in pg/mL (n = 50) TGF-α (log transformed)
14 (25.5%) 7.28 (5.42) 6.49 (5.81) 1.66 (0.35) 9.53 (26.96) 0.85 (0.60) 6.64 (8.51) 1.51 (0.53) 12.45 (15.27) 1.72 (0.48)
0–22 1.0–24.4 1.0–2.3 1.0–130.0 0.4–2.4 0.0–44.0 0.2–2.4 2.0–88.0 1.1–2.7
Abbreviations: SD, standard deviation; MINI, Mini-International Neuropsychiatric Interview; PHQ-9, nine-item Patient Health Questionnaire; TNF-α, tumor necrosis factorα; IL-1β, interleukin-1β; IL-6, interleukin-6; TGF-α, transforming growth factor-α.
30
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
Table 3 Differences in major depression status by EGFR mutation status.
Table 5 Differences in inflammatory cytokines and growth factors by major depression status in patients not taking antidepressants (N = 42).
EGFR mutation status
Depression status
Major depressive disorder (n = 14) No major depressive disorder (n = 34)
EGFR mutation (n = 10)
EGFR wild-type (n = 38)
1 (2.08%)
13 (27.08%)
9 (18.75%)
25 (52.08%)
prior cancer studies on the associations between biological markers and depression [22,23]. Given mixed findings about the extent to which antidepressants modulate inflammatory processes, there is evidence that antidepressant use should be considered to be an exclusion criterion, especially in cases when examining the relationship between psychiatric morbidity and inflammation [24]. Therefore, ten patients on antidepressants were not included in analyses involving inflammatory cytokines and growth factors.
3.2.2. Serum factors and depression In the subsample of patients not taking antidepressants, patients with MDD had significantly higher TNF-α levels than patients without Table 4 Differences in inflammatory cytokines and growth factors by EGFR mutation status in patients not taking antidepressants (N = 37).
EGFR wildtype (n = 29)
t (df)
95% CI
TNF-α
1.94 (0.27)
1.64 (0.36)
[0.02, 0.58]
TGF-α
1.73 (0.60)
1.70 (0.47)
IL-1β
1.13 (0.73)
0.88 (0.57)
IL-6
1.68 (0.44)
1.48 (0.53)
2.17 (35)⁎ 0.05 (35) 1.03 (35) 0.94 (35)
⁎
No major depressive disorder (n = 35)
t (df)
95% CI
TNF-α
2.01 (0.22)
1.61 (0.34)
[0.12, 0.66]
TGF-α
1.94 (0.43)
1.74 (0.50)
IL-1β
1.18 (0.41)
0.84 (0.61)
IL-6
1.74 (0.49)
1.45 (0.52)
2.95 (40)⁎ 0.95 (40) 1.38 (40) 1.34 (40)
[− 0.22, 0.61] [− 0.15, 0.82] [− 0.14, 0.71]
p < 0.05.
This is the first study to our knowledge examining relationships among depression, EGFR mutation status, and proinflammatory cytokines and growth factors in patients with NSCLC. Additionally, this is the third patient cohort in whom we have found an association between depression and EGFR mutation status [1,25]. In the current study, we identified further that patients with EGFR mutations had lower depressive symptom severity, as we hypothesized. However, contrary to our expectations, patients with EGFR mutations and lower depressive symptom severity had significantly higher levels of TNF-α than patients without EGFR mutations. Patients with EGFR mutations were also less likely to meet criteria for MDD on the diagnostic interview, but this difference did not reach statistical significance. In general, patients meeting criteria for MDD had higher levels of TNF-α compared to patients without MDD. No meaningful differences were observed with regard to other inflammatory biomarkers (i.e., TGF-α, IL-1β, or IL-6) in this small sample. Although biological pathways of inflammation and depression in patients with cancer have been postulated, EGFR has not previously been considered in these models. Constitutive activation of EGFR via oncogenic mutations leads to upregulation of gene expression that may heighten the inflammatory response in certain situations [26,27]. For example, activated EGFR results in COX-2 expression; a pathway that has implications for the development of lung cancer via chronic airway inflammation [28,29]. EGFR signaling also leads to downstream production of IL-6 in human lung cancer cells [30]. A substantial portion of the tumor micro-environment is inflammatory [31] and pro-inflammatory cytokines and growth factors exacerbate cellular processes such as tumor proliferation and inhibit immune responses [32]. In turn, chronic inflammation contributes to the development of malignancies and metastases [33,34]. Integral in these processes, EGFR acts as a key regulator and a cellular hub for inflammatory cytokine signaling, thereby promoting tumor cell proliferation, invasion, migration, metastases, and survival [7–9]. While this may help explain our observed relationship between EGFR mutations and elevated TNF-α in patients with stage IV NSCLC, it makes the relationship between EGFR mutations and lower levels of depression more puzzling. Depression is associated with activation of the immune system, including increased expression of proinflammatory cytokines such as TNF-α and IL-6 [5]. In addition, patients with cancer exhibit higher levels of inflammation [31]. Importantly, evidence suggests that
3.2.1. Serum factors and EGFR mutation status In the subsample excluding patients taking antidepressants, markers of inflammation differed significantly based on EGFR mutation status (Table 4). Patients with EGFR mutations had significantly higher TNF-α levels compared to patients without EGFR mutations, t(35) = 2.17, p = 0.04, 95% CI [0.02–0.58], d = 0.93 (a large effect size). EGFR mutation status was not significantly associated with differences in levels of TGF-α (t[35] = 0.05, p = 0.96), IL-1β (t[35] = 1.03, p = 0.31), or IL-6 (t[35] = 0.94, p = 0.35).
EGFR mutation (n = 8)
Major depressive disorder (n = 7)
4. Conclusions
3.2. Serum inflammatory cytokines and growth factors, EGFR status, and depression
Log-transformed biomarker
Log-transformed biomarker
MDD, t(40) = 2.95, p = 0.005, 95% CI [0.12 = 0.66], d = 1.36 (a large effect size; see Table 5). The presence of MDD was not significantly associated with differences in levels of TGF-α (t[40] = 0.95, p = 0.35), IL-1β (t[40] = 1.38, p = 0.17), or IL-6 (t[40] = 1.34, p = 0.19).
Depression severity scores on the PHQ-9 significantly differed based on mutation status, such that patients with EGFR mutations were less depressed (M = 4.60; SD = 3.34) than patients without EGFR mutations (M = 8.03; SD = 8.03), t(43) = 2.38, p = 0.03, 95% CI [0.46–0.39], d = 0.72 (consistent with a large effect size). Although fewer of the patients with EGFR mutations (n = 10) had concurrent MDD on the MINI (2.08%) compared to those without mutations (n = 38; 27.08%), this difference did not reach statistical significance (Fisher's exact test, p = 0.24; Table 3).
Comparison
Comparison
⁎
3.1. EGFR status and depression
EGFR mutation status (M [SD])
Current depression status
[− 0.41, 0.40] [− 0.24, 0.74] [− 0.22, 0.61]
p < 0.05.
31
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
such as depression in patients with stage IV NSCLC and may ultimately lead to the use of tumor genotyping results for selection of depression therapies. However, longitudinal studies are needed to follow these patients over time while re-assessing depressive symptoms and depression status. These findings must also be replicated with a larger sample size to allow for the inclusion of other relevant contributing factors. In the long-term, with enhanced understanding of the molecular mechanisms underlying the relationships among EGFR and depression, there may be an opportunity for novel prevention or intervention strategies to improve quality of life in targeted subsets of patients with stage IV NSCLC.
patients with cancer and co-morbid depression have elevated proinflammatory markers relative to patients with either depression or cancer alone [22,35]. Yet in this study of individuals with stage IV NSCLC, patients with tumors that harbored EGFR mutations, and who had higher levels of TNF-α in their serum, had less depressive symptoms. It is possible that an unmeasured factor associated with EGFR mutations, such as EGF, may decrease the risk of depression in the setting of elevated TNF-α. For instance, a study comparing EGF levels in patients with MDD found significantly lower levels of EGF in those with MDD compared with those without MDD [36]. It is interesting to note that most work examining cytokines and depression finds consistent relationships between IL-6 and depression [37], but are more ambiguous with regard to the relationship between TNF- α and depression [38], which is the opposite of what we found. In addition, evidence suggests that there are different patterns of cytokine activation in subtypes of depression [38]. More work is needed to replicate these findings and examine potential mechanisms explaining how EGFR mutations might be protective for patients with stage IV NSCLC who would usually be at risk for depression given elevated circulating inflammatory markers. While a pilot, the current study contributes novel information regarding the potential biological pathways underlying relationships among depression, tumor genetics, and survival in patients with stage IV NSCLC that have been observed in previous work. Future research should continue in this relatively unstudied area [25]. The results also replicate and expand upon previous findings from our group showing a relationship between tumor gene mutations and depression, in which patients with EGFR mutations were less depressed than those with wildtype tumors [1,25]. Furthermore, in the current study, those that harbored EGFR mutations had similar levels of pain and performance status, ruling out the possibility that lower depression in these patients could be explained by better performance status or lower pain severity. Our findings support that crosstalk between EGFR and inflammationrelated pathways likely exists in patients with stage IV NSCLC, as has been previously observed in cellular and molecular models. A notable limitation of the current pilot study is the small sample size, precluding the use of relevant control variables (e.g., age, gender, body mass index, and smoking history) and the conduct of complex tests of mediation to expand our understanding of the relationships among EGFR, depression, and TNF-α. This is relevant given that patients with wild-type EGFR were more likely to have a smoking history, and smoking is related to depression; although, the directionality of this relationship is not clear [39]. In addition, we did not collect information on certain comorbid diseases and other clinical characteristics that may explain the differences in depression. Results were in the expected direction for the association between depression severity and EGFR status; however, we are limited in our ability to make strong inferences given that not all differences reached statistical significance. Post hoc power calculations indicated 50% power to detect the difference in depression severity between 10 patients with EGFR mutations and 38 patients without at α = 0.05 (effect size = 0.72). Nonetheless, these data can be used to adequately power future studies. Finally, the crosssectional study design prohibits conclusions regarding the direction of the findings and therefore causality. A strength of the current study is the use of an administered diagnostic interview (i.e., the MINI) to assess presence of MDD in addition to depression severity on the PHQ-9. It is noteworthy that patients completed self-report measures and were administered the MINI prior to patients' receiving the results of the genotyping. Therefore, associations between depression and EGFR mutations status were observed prior to patients' awareness of their prognosis or potential responsiveness to treatment. As a result of this timing, we can rule out the possibility that patients' perceived prognosis drove depression severity. Patients living with stage IV NSCLC experience high physical and psychological symptom burden. Despite limitations, our findings have the potential to further our understanding of the etiology of symptoms
Acknowledgements This work was supported by the American Cancer Society (PC1116401). We would also like to acknowledge the patients who dedicated their time to participating in this research study. References [1] W.F. Pirl, L. Traeger, J.A. Greer, et al., Tumor epidermal growth factor receptor genotype and depression in stage IV non-small cell lung cancer, Oncologist 16 (2011) 1299–1306, http://dx.doi.org/10.1634/theoncologist.2011-0116. [2] A.J. Mitchell, M. Chan, H. Bhatti, et al., Prevalence of depression, anxiety, and adjustment disorder in oncological, haematological, and palliative-care settings: a meta-analysis of 94 interview-based studies, Lancet Oncol. 12 (2011) 160–174, http://dx.doi.org/10.1016/S1470-2045(11)70002-X. [3] D.V. Iosifescu, B. Bankier, M. Fava, Impact of medical comorbid disease on antidepressant treatment of major depressive disorder, Curr. Psychiatry Rep. 6 (2004) 193–201 (doi). [4] G.I. Papakostas, T.J. Petersen, D.V. Iosifescu, et al., Somatic symptoms as predictors of time to onset of response to fluoxetine in major depressive disorder, J Clin Psychiatry 65 (2004) 543–546 (doi). [5] Y. Dowlati, N. Herrmann, W. Swardfager, et al., A meta-analysis of cytokines in major depression, Biol. Psychiatry 67 (2010) 446–457, http://dx.doi.org/10.1016/ j.biopsych.2009.09.033. [6] C.L. Arteaga, The epidermal growth factor receptor: from mutant oncogene in nonhuman cancers to therapeutic target in human neoplasia, J. Clin. Oncol. 19 (2001) 32S–40S (doi). [7] M. Scaltriti, J. Baselga, The epidermal growth factor receptor pathway: a model for targeted therapy, Clin. Cancer Res. 12 (2006) 5268–5272, http://dx.doi.org/10. 1158/1078-0432.CCR-05-1554. [8] H.M. Shepard, C.M. Brdlik, H. Schreiber, Signal integration: a framework for understanding the efficacy of therapeutics targeting the human EGFR family, J. Clin. Invest. 118 (2008) 3574–3581, http://dx.doi.org/10.1172/JCI36049. [9] C. Berasain, M.J. Perugorria, M.U. Latasa, et al., The epidermal growth factor receptor: a link between inflammation and liver cancer, Exp. Biol. Med. (Maywood) 234 (2009) 713–725, http://dx.doi.org/10.3181/0901-MR-12. [10] T. Fukuyama, Y. Ichiki, S. Yamada, et al., Cytokine production of lung cancer cell lines: correlation between their production and the inflammatory/immunological responses both in vivo and in vitro, Cancer Sci. 98 (2007) 1048–1054, http://dx. doi.org/10.1111/j.1349-7006.2007.00507.x. [11] K. Kroenke, R.L. Spitzer, J.B. Williams, The PHQ-9: validity of a brief depression severity measure, J. Gen. Intern. Med. 16 (2001) 606–613 (doi, 1495268). [12] M.M. Oken, R.H. Creech, D.C. Tormey, et al., Toxicity and response criteria of the Eastern Cooperative Oncology Group, Am. J. Clin. Oncol. 5 (1982) 649–655. [13] Association AP, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, DSM-IV-TR, American Psychiatric Association, Arlington, VA, 2000. [14] J.R. Fann, D.L. Berry, S. Wolpin, et al., Depression screening using the Patient Health Questionnaire-9 administered on a touch screen computer, Psychooncology 18 (2009) 14–22, http://dx.doi.org/10.1002/pon.1368. [15] K. Ell, B. Xie, B. Quon, et al., Randomized controlled trial of collaborative care management of depression among low-income patients with cancer, J. Clin. Oncol. 26 (2008) 4488–4496, http://dx.doi.org/10.1200/JCO.2008.16.6371. [16] D. Sheehan, Y. Lecrubier, K. Sheehan, et al., The validity of the Mini International Neuropsychiatric Interview (MINI) according to the SCID-P and its reliability, Eur. Psychiatry 12 (1997) 232–241 (doi). [17] D.V. Sheehan, Y. Lecrubier, K.H. Sheehan, et al., The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10, J Clin Psychiatry 59 (Suppl. 20) (1998) 22–33 (quiz 34-57. doi). [18] Organization WH, The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines, World Health Organization, Geneva, 1992. [19] D. Dias-Santagata, S. Akhavanfard, S.S. David, et al., Rapid targeted mutational analysis of human tumours: a clinical platform to guide personalized cancer medicine, EMBO Mol. Med. 2 (2010) 146–158, http://dx.doi.org/10.1002/emmm. 201000070. [20] R.R. Wilcox, Comparing dependent robust correlations, Br. J. Math. Stat. Psychol. (2016), http://dx.doi.org/10.1111/bmsp.12069.
32
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al. [21] J. Cohen, Statistical Power Analysis for the Behavioral Sciences, 2nd. ed., Academic Press, New York, NY, 1998. [22] D.L. Musselman, A.H. Miller, M.R. Porter, et al., Higher than normal plasma interleukin-6 concentrations in cancer patients with depression: preliminary findings, Am. J. Psychiatry 158 (2001) 1252–1257, http://dx.doi.org/10.1176/appi.ajp.158. 8.1252. [23] S.K. Lutgendorf, D.M. Lamkin, K. DeGeest, et al., Depressed and anxious mood and T-cell cytokine expressing populations in ovarian cancer patients, Brain Behav. Immun. 22 (2008) 890–900, http://dx.doi.org/10.1016/j.bbi.2007.12.012. [24] M.F. O'Connor, J.E. Bower, H.J. Cho, et al., To assess, to control, to exclude: effects of biobehavioral factors on circulating inflammatory markers, Brain Behav. Immun. 23 (2009) 887–897, http://dx.doi.org/10.1016/j.bbi.2009.04.005. [25] W.F. Pirl, L. Traeger, J.A. Greer, et al., Depression, survival, and epidermal growth factor receptor genotypes in patients with metastatic non-small cell lung cancer, Palliat Support Care 11 (2013) 223–229, http://dx.doi.org/10.1017/ S1478951512001071. [26] K. Liu, R.C. Gualano, M.L. Hibbs, G.P. Anderson, S. Bozinovski, Epidermal growth factor receptor signaling to Erk1/2 and STATs control the intensity of the epithelial inflammatory responses to rhinovirus infection, J. Biol. Chem. 283 (2008) 9977–9985, http://dx.doi.org/10.1074/jbc.M710257200. [27] K. Liu, G.P. Anderson, S. Bozinovski, DNA vector augments inflammation in epithelial cells via EGFR-dependent regulation of TLR4 and TLR2, Am. J. Respir. Cell Mol. Biol. 39 (2008) 305–311, http://dx.doi.org/10.1165/rcmb.2007-0458OC. [28] M. Liu, S.C. Yang, S. Sharma, et al., EGFR signaling is required for TGF-beta 1 mediated COX-2 induction in human bronchial epithelial cells, Am. J. Respir. Cell Mol. Biol. 37 (2007) 578–588, http://dx.doi.org/10.1165/rcmb.2007-0100OC. [29] J. Yoo, C.E. Rodriguez Perez, W. Nie, et al., TNF-α induces upregulation of EGFR expression and signaling in human colonic myofibroblasts, Am. J. Physiol. Gastrointest. Liver Physiol. 302 (2012) G805–G814, http://dx.doi.org/10.1152/ ajpgi.00522.2011. [30] S.P. Gao, K.G. Mark, K. Leslie, et al., Mutations in the EGFR kinase domain mediate
[31] [32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
33
STAT3 activation via IL-6 production in human lung adenocarcinomas, J. Clin. Invest. 117 (2007) 3846–3856, http://dx.doi.org/10.1172/JCI31871. F. Balkwill, L.M. Coussens, Cancer: An inflammatory link, Nature 431 (2004) 405–406, http://dx.doi.org/10.1038/431405a. M. Orozco-Morales, G. Soca-Chafre, P. Barrios-Bernal, N. Hernández-Pedro, O. Arrieta, Interplay between cellular and molecular inflammatory mediators in lung cancer, Mediat. Inflamm. 2016 (2016) 3494608, http://dx.doi.org/10.1155/ 2016/3494608. M. Dougan, D. Li, D. Neuberg, et al., A dual role for the immune response in a mouse model of inflammation-associated lung cancer, J. Clin. Invest. 121 (2011) 2436–2446, http://dx.doi.org/10.1172/JCI44796. L. Barrera, E. Montes-Servín, A. Barrera, et al., Cytokine profile determined by datamining analysis set into clusters of non-small-cell lung cancer patients according to prognosis, Ann. Oncol. 26 (2015) 428–435, http://dx.doi.org/10.1093/annonc/ mdu549. C.F. Jehn, D. Kuehnhardt, A. Bartholomae, et al., Biomarkers of depression in cancer patients, Cancer 107 (2006) 2723–2729, http://dx.doi.org/10.1002/cncr. 22294. W. Tian, J. Zhang, K. Zhang, et al., A study of the functional significance of epidermal growth factor in major depressive disorder, Psychiatr. Genet. 22 (2012) 161–167, http://dx.doi.org/10.1097/YPG.0b013e3283539550. E.P. Zorrilla, L. Luborsky, J.R. McKay, et al., The relationship of depression and stressors to immunological assays: a meta-analytic review, Brain Behav. Immun. 15 (2001) 199–226, http://dx.doi.org/10.1006/brbi.2000.0597. S.M. Gold, M.R. Irwin, Depression and immunity: inflammation and depressive symptoms in multiple sclerosis, Immunol. Allergy Clin. N. Am. 29 (2009) 309–320, http://dx.doi.org/10.1016/j.iac.2009.02.008. M. Fluharty, A.E. Taylor, M. Grabski, M.R. Munafò, The association of cigarette smoking with depression and anxiety: a systematic review, Nicotine Tob. Res. 19 (2017) 3–13, http://dx.doi.org/10.1093/ntr/ntw140.
Contents lists available at ScienceDirect
Journal of Psychosomatic Research journal homepage: www.elsevier.com/locate/jpsychores
Depression, inflammation, and epidermal growth factor receptor (EGFR) status in metastatic non-small cell lung cancer: A pilot study☆
MARK
Jamie M. Jacobsa,⁎, Lara Traegera, Justin Eusebioa, Naomi M. Simonb, Lecia V. Sequistc, Joseph A. Greera, Jennifer S. Temelc, William F. Pirla,1 a Center for Psychiatric Oncology and Behavioral Sciences, Department of Psychiatry, Massachusetts General Hospital Cancer Center/Harvard Medical School, 55 Fruit St., Yawkey Center for Outpatient Care, Suite 10B, Boston, MA 02114, United States b Center for Anxiety and Traumatic Stress Disorders, Massachusetts General Hospital/Harvard Medical School, One Bowdoin Square, 6th floor, Boston, MA 02114, United States c Massachusetts General Hospital Cancer Center/Harvard Medical School, 55 Fruit St., Yawkey Center for Outpatient Care, Suite 7B, Boston, MA 02114, United States
A R T I C L E I N F O
A B S T R A C T
Keywords: Non-small cell lung cancer Genotyping Epidermal growth factor receptor Depression Inflammatory cytokines Oncology
Objective: Patients with stage IV non-small cell lung cancer (NSCLC) have high risk for depressive symptoms and major depressive disorder (MDD); however, those with epidermal growth factor receptor (EGFR) mutations may have decreased risk. The biological underpinning of this relationship is unknown. We examined differences in depression severity and MDD in patients with newly diagnosed stage IV NSCLC based on EGFR mutation status, and examined proinflammatory cytokines and growth factors known to play a role in cancer progression and depression. Methods: Fifty-five patients with newly diagnosed stage IV NSCLC completed self-report and clinician-administered depression assessments prior to receiving results of tumor genotyping. We measured serum levels of circulating biological markers of inflammation: IL-1β, IL-6, TGF-α, and TNF-α. We examined differences in depression severity, MDD, and inflammatory biomarkers in patients with and without EGFR mutations. Results: Patients with EGFR mutations (n = 10) had lower depression severity (t[43] = 2.38, p = 0.03) than those without EGFR mutations (n = 38) and fewer patients with EGFR mutations had concurrent MDD (2.08%) relative to those without mutations (27.08%). Patients with MDD had higher levels of TNF-α than those without MDD (t[40] = 2.95, p = 0.005). Those with EGFR mutations exhibited higher levels of TNF-α relative to those without EGFR mutations (t[35] = 2.17, p = 0.04). Conclusions: Patients with stage IV NSCLC harboring an EGFR mutation exhibited elevated proinflammatory marker TNF-α, yet had lower depression severity than patients without EGFR mutations. More work is warranted to examine the interaction between tumor genotyping and inflammatory cytokines in the context of depression.
1. Introduction Significant advances in our understanding of cancer biology and tumor gene mutations over the last decade has led to certain cancers, such as non-small cell lung cancer (NSCLC), to be classified by their genotype signature rather than mere histology. Particular tumor gene mutations implicated in cancer growth in subsets of patients confer sensitivity to specific targeted therapies. In the molecular landscape of cancers, patterns of cancer-related symptoms might also emerge for tumors with certain genotypes. Few studies have looked at the
molecular basis of depressive symptoms in patients with cancer. Although depression is common in patients with cancer, affecting an estimated 10–25%, it is unclear why it occurs in some patients and not others [1,2]. Patients with medical co-morbidities may be less likely to respond to pharmacological depression treatments, suggesting that there may be differences in depression between those with and without specific co-morbidities [3,4]. In addition, depression is associated with elevated inflammation [5], and risk for depression may be higher in patients with medical illness and heightened inflammatory processes. Tumor gene mutations may represent a novel opportunity to examine
☆ Conflict of interest: Dr. Simon discloses spousal equity in G1 Therapeutics and Gatekeeper. Dr. Temel discloses research funding from Pfizer. All other authors declare that they have no conflicts of interest or disclosures. Acknowledgement: This work was supported by PC-1116401, American Cancer Society (Pirl). ⁎ Corresponding author. E-mail address: [email protected] (J.M. Jacobs). 1 Present address: Sylvester Comprehensive Cancer Center, Courtelis Center, University of Miami Miller School of Medicine, 1120 NW 14th Street, Miami, FL 33156.
http://dx.doi.org/10.1016/j.jpsychores.2017.05.009 Received 13 January 2017; Received in revised form 6 May 2017; Accepted 11 May 2017 0022-3999/ © 2017 Elsevier Inc. All rights reserved.
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
under the direct supervision of a psychiatrist. Patients who completed the interview, self-report instruments, and blood draw were compensated $50.00 for their time.
risk of developing depression after a cancer diagnosis. One of the first targetable mutations to be identified in NSCLC was the family of epidermal growth factor receptor (EGFR) mutations [6]. Patients with these mutations respond well to the EGFR tyrosine kinase inhibitors (TKIs; e.g., gefitinib, erlotinib and afatinib) and thus have a better prognosis than patients without an EGFR mutation [7–9]. We previously observed that patients with stage IV NSCLC with EGFR mutations report less depression than those without EGFR mutations, even when depression was measured before the results of genotyping were known [1]. Yet, without a biologically plausible link between NSCLC EGFR mutation status and depression, the meaning of any observed association is limited. One hypothesis is that NSCLC cells with different genotypes may produce different magnitudes of inflammatory cytokines and growth factors, which are associated with and may contribute to depression [10]. The goal of this work was to conduct a new study as a first step towards understanding potential links between EGFR mutation status and depression in patients with newly diagnosed stage IV NSCLC. The objective of this pilot study was to gather initial data on possible relationships between depression, inflammation, and EGFR mutation status. Specifically, we hypothesized that 1) Patients harboring an EGFR mutation would have lower depression severity and be less likely to be diagnosed with major depressive disorder (MDD) than patients without an EGFR mutation, and 2) EGFR mutation status and MDD status would be associated with serum levels of inflammatory cytokines and growth factors. Patients harboring an EGFR mutation and those without major depression would have lower levels of inflammation than those without an EGFR mutation. If a signal for relationships exists, this would justify larger studies adequately powered to adjust for potential confounders.
2.3. Demographics and clinical characteristics Study staff collected patient data from the electronic health record, including demographic information, cancer staging and histology, pain level, smoking history (i.e., current smoking status and pack years), prescribed medications (i.e., antidepressants), and Eastern Cooperative Oncology Group (ECOG) performance status [12]. 2.4. Depression severity To evaluate the severity of depression, we used the PHQ-9, [11] a self-report measure designed to screen for MDD according to the criteria of the Diagnostic and Statistical Manual-IV (DSM-IV [13]). The PHQ-9 has been validated for use with patients with cancer and has been used in clinical trials [14,15]. Patients are asked how often they have been bothered by symptoms of depression over the past 2 weeks, with response options ranging from 0 (Not at all) to 3 (Nearly every day). Scores are summed to obtain a total severity score ranging from 0 to 27. The PHQ-9 demonstrated good reliability in the current sample (α = 0.79). 2.5. Major depressive disorder Presence of current MDD was assessed with the Mini-International Neuropsychiatric Interview (MINI [16,17]); a brief, structured, clinician-administered diagnostic interview that screens for major Axis I psychiatric disorders according to the DSM-IV [13] and ICD-10 criteria [18]. The widely used MINI has high reliability and validity comparable to the Structured Clinical Interview for the DSM-IV Axis I Disorders [16]. Interview questions assess each dimension of the disorder, including time frame, frequency, and severity. Items require “yes” or “no” answers. Patients are given a diagnosis of MDD if symptom criteria for MDD were met, and criteria for other disorders with depressive features are not met.
2. Materials and methods 2.1. Participants Eligible patients were those who had a new diagnosis of stage IV NSCLC; had genotyping ordered on a biopsy specimen; had no prior knowledge of their NSCLC genotypes; possessed the ability to understand and sign informed consent; were not at their appointment for a second opinion and/or planning to receive care elsewhere; had not previously received chemotherapy or radiation for current stage IV NSCLC; had no dementia or cognitive impairment; and had no diagnosis of a psychotic or bipolar disorder.
2.6. Genotyping Tissue from patients' biopsies was genotyped in a CLIA-certified clinical laboratory at MGH as part of standard care with SNaPshot, multiplex PCR-based allele-specific analysis of known cancer hotspots across 13 genes [19]. Genotype results were available 2–4 weeks later and were posted in the electronic health record. For the present analysis, patients who harbored an EGFR mutation were compared to patients without an EGFR mutation.
2.2. Procedures This study took place at Massachusetts General Hospital (MGH) Cancer Center and was approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board. All procedures contributing to this work comply with ethical standards of the relevant national and institutional committees on human subjects' research and with the Helsinki Declaration of 1975, as revised in 2008. Consecutive new patients presenting at the MGH Multidisciplinary Thoracic Oncology Clinic were recruited to participate in this study at their initial appointment between July 26, 2011 and August 8, 2013. Research assistants reviewed preliminary information on patients collected by oncology nurses, and those who had suspected advanced NSCLC were approached to participate. Patients interested in participating were screened for eligibility criteria. Eligible and interested patients signed informed consent after review with study staff. Prior to receiving results of their genotyping and before initiating any cancer treatments, patients completed a self-report survey, a clinician-administered diagnostic interview, and blood collection. Within the self-report survey, the Patient Health Questionnaire-9 [11] was always administered in person by paper or by telephone. The diagnostic interview was administered by a trained Clinical Research Coordinator
2.7. Serum cytokines and growth factors Blood samples were taken by a staff phlebotomist or in the MGH Thoracic Oncology Clinic prior to the initiation of chemotherapy. Samples were collected by venipuncture in 10 mL tubes containing EDTA and placed on ice. Serum levels of circulating Tumor Necrosis Factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6) were quantified measured using a ultra-sensitive multiplex immunoassay system “Human Proinflammatory II 4-Plex Assay” with electrochemiluminescent detection from Meso Scale Discovery® Multi-array Technology (Meso Scale Discovery, Gaithersburg, MD). Biomarkers were measured by commercially available reagents and were performed according to the manufacturer's instructions. No dilutions were required or used. The mean lower limits of detection for TNF-α, IL-6, IL1β were 0.37 pg/mL, 0.26 pg/mL, and 0.18 pg/mL, respectively. Serum levels of Transforming Growth Factor-α (TGF-α) was measured with the Quantikine® Human TGF- α Immunoassay, a 4.5 hour solid-phase 29
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
ELISA (R & D Systems, Inc., Minneapolis, MN). Samples were stored at ≤− 20 °C and assays were performed according to the manufacturer's instructions. Samples were not diluted per protocol instructions. The mean detectable limit for TGF-α was 21.9 pg/mL.
Table 1 Sociodemographic and clinical characteristics in the full sample (n = 55) and by EGFR mutation status. Characteristic
n (%) or M (SD)
EGFR mutation (n = 10)
EGFR wild-type (n = 38)
Age (n = 54; range = 42–83) Female gender Race White Asian Black or African American Other Missing Ethnicity Non-Hispanic Unknown Marital status Married or living as if married Single Divorced or separated Loss of long-term partner/ widowed Unknown Education < 12th grade High school graduate or GED Some college/vocational/ technical College graduate or postgraduate Unknown ECOG performance status score 0 1 2 3 Missing Brain Metastases Smoking Status Current or former smoker Never smoker Missing Not on antidepressants EGFR mutation (vs. wildtype) Deletion exon 19 L858R G719A Exon 20 insertion
62.09 (11.73) 22 (40.0%)
62.60 (12.45)
62.32 (10.12)
8 (80%)
11 (28.9%)
51 (92.7%) 1 (1.8%) 1 (1.8%) 1 (1.8%) 1 (1.8%)
8 1 0 1
(80%) (10.0%) (0.0%) (10.0%)
37 (97.4%) 0 (0.0%) 1 (2.6%) 0 (0.0%)
53 (96.4%) 2 (3.6%)
10 (100.0%) 0 (0.0%)
37 (97.4%) 1 (2.6%)
35 (63.6%)
6 (60.0%)
26 (68.4%)
4 (7.3%) 8 (14.5%) 6 (10.9%)
1 (10.0%) 1 (10.0%) 1 (10.0%)
2 (5.3%) 6 (15.8%) 4 (10.5%)
2 (3.6%) 3 (5.5%) 18 (32.7%) 9 (16.4%)
1 (10.0%)
0 (0.0%)
0 (0.0%) 0 (0.0%)
3 (7.9%) 15 (39.5%)
21 (38.2%)
2 (20.0%)
6 (15.8%)
7 (70.0%)
12 (31.65)
4 (7.3%)
1 (10.0%)
2 (5.3%)
18 (32.7%) 30 (54.5%) 2 (3.6%) 3 (5.5%) 2 (3.6%) 18 (37.5%)
4 3 1 1 1 2
(40.0%) (30.0%) (10.0%) (10.0%) (10.0%) (20.0%)
13 (34.2%) 22 (57.9%) 1 (2.6%) 2 (5.3%) 0 (0.0%) 16 (42.1%)
44 (80.0%) 10 (18.2%) 1 (1.8%) 45 (81.8%) 10 (18.2%)
2 (20.0%) 8 (80.0%)
37 (97.4%) 1 (2.6%)
9 (90.0%)
29 (76.3%)
2.8. Statistical analyses Data were analyzed with the Statistical Package for the Social Sciences (SPSS, Version 21.0). All biological data were screened for outliers and cleaned by converting values outside three standard deviations from the mean to equal the next highest value [20]. To meet assumptions of normality, biological data were log-transformed, and a value of one was added to maintain positive numbers. Due to the small sample size, two-sided Fisher's exact test was used to test for differences in categorical dependent variables. Independent samples t-tests were conducted to assess differences in continuous dependent variables, and estimates were interpreted at a two-tailed significance level (p < 0.05) and 95% confidence intervals. Levene's test was used to evaluate the assumption of equality of variance, and a correction was used if this assumption was not met. Corresponding effect sizes (0.20 = small; 0.50 = medium; 0.80 = large) were evaluated to assess the magnitude of the observed significant differences [21]. First, in an effort to replicate prior findings [1], independent samples t-tests were conducted to test for group differences in depression severity on the PHQ-9 based on EGFR mutation status. Power analyses based on this work suggested that, with 50 subjects and a 5% Type I error rate, there is 80% power to detect a minimum mean group difference of 3.3 points in depression severity on the PHQ-9 in patients with EGFR mutations versus those with wild-type. Given the clinical meaningfulness and usefulness of the MINI to identify presence of MDD, the remainder of analyses examined differences based on whether patients met criteria for current MDD. Fisher's exact test was used to test whether presence of MDD differed by EGFR mutation status. Next, independent samples t-tests were used to test for differences in inflammatory cytokines and growth factors based on EGFR mutation status. In addition, independent samples t-tests were used to test for differences in inflammatory cytokines and growth factors in patients who carried a diagnosis of MDD vs. those who did not. 3. Results Fifty-eight patients enrolled in the study (Table 1). Three patients were subsequently excluded after further oncology work-up determined them to be at an earlier stage of disease. The remaining 55 patients were an average of 62.09 years (standard deviation [SD] = 11.73 years), and more than half were male (58.2%). Most of the sample (96.4%) self-identified as non-Hispanic white. Overall, patients had good performance status, with ECOG performance status scores of 0 or 1 (90.6%), and 80% were current or former smokers. Approximately one quarter of the sample (n = 14) had current MDD on the MINI, and the mean PHQ-9 depression severity score for those with current MDD was 12.17 (SD = 4.74). The mean PHQ-9 depression severity score for the total sample was 7.28 (SD = 5.42); see Table 2. Of the 55 patients, 38 (69.1%) had cancers that were wild-type for EGFR, ten (18.2%) harbored EGFR mutations, and seven (12.7%) lacked tissue suitable for genotyping. Patients who were wild-type for EGFR were more likely to be male, while those with EGFR mutations were more likely to be female (Fisher's Exact Test, p = 0.008). Those who were wild-type for EGFR mutations were more likely to be current or former smokers than those harboring EGFR mutations (Fisher's Exact Test, p < 0.001). No other differences were found in any other sociodemographic or clinical characteristics (Table 1). A total of 50 patients had serum available for biological assays, with 45 having both biological and genotype data available for analysis. One value was adjusted for each inflammatory marker due to being greater than three standard deviations above the mean. This sample size is consistent with
5 3 1 1
(50%) (30%) (10%) (10%)
Abbreviations: SD, standard deviation; EGFR, epidermal growth factor receptor; ECOG, Eastern Cooperative Oncology Group. Table 2 Psychosocial and biological characteristics (n = 55). Characteristic
N (%) or mean (SD)
Range
Current major depressive disorder (MINI) Depression severity (PHQ-9; n = 50) TNF-α in pg/mL (n = 50) TNF-α (log transformed) IL-1β in pg/mL (n = 50) IL-1β (log transformed) IL-6 in pg/mL (n = 50) IL-6 (log transformed) TGF-α in pg/mL (n = 50) TGF-α (log transformed)
14 (25.5%) 7.28 (5.42) 6.49 (5.81) 1.66 (0.35) 9.53 (26.96) 0.85 (0.60) 6.64 (8.51) 1.51 (0.53) 12.45 (15.27) 1.72 (0.48)
0–22 1.0–24.4 1.0–2.3 1.0–130.0 0.4–2.4 0.0–44.0 0.2–2.4 2.0–88.0 1.1–2.7
Abbreviations: SD, standard deviation; MINI, Mini-International Neuropsychiatric Interview; PHQ-9, nine-item Patient Health Questionnaire; TNF-α, tumor necrosis factorα; IL-1β, interleukin-1β; IL-6, interleukin-6; TGF-α, transforming growth factor-α.
30
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
Table 3 Differences in major depression status by EGFR mutation status.
Table 5 Differences in inflammatory cytokines and growth factors by major depression status in patients not taking antidepressants (N = 42).
EGFR mutation status
Depression status
Major depressive disorder (n = 14) No major depressive disorder (n = 34)
EGFR mutation (n = 10)
EGFR wild-type (n = 38)
1 (2.08%)
13 (27.08%)
9 (18.75%)
25 (52.08%)
prior cancer studies on the associations between biological markers and depression [22,23]. Given mixed findings about the extent to which antidepressants modulate inflammatory processes, there is evidence that antidepressant use should be considered to be an exclusion criterion, especially in cases when examining the relationship between psychiatric morbidity and inflammation [24]. Therefore, ten patients on antidepressants were not included in analyses involving inflammatory cytokines and growth factors.
3.2.2. Serum factors and depression In the subsample of patients not taking antidepressants, patients with MDD had significantly higher TNF-α levels than patients without Table 4 Differences in inflammatory cytokines and growth factors by EGFR mutation status in patients not taking antidepressants (N = 37).
EGFR wildtype (n = 29)
t (df)
95% CI
TNF-α
1.94 (0.27)
1.64 (0.36)
[0.02, 0.58]
TGF-α
1.73 (0.60)
1.70 (0.47)
IL-1β
1.13 (0.73)
0.88 (0.57)
IL-6
1.68 (0.44)
1.48 (0.53)
2.17 (35)⁎ 0.05 (35) 1.03 (35) 0.94 (35)
⁎
No major depressive disorder (n = 35)
t (df)
95% CI
TNF-α
2.01 (0.22)
1.61 (0.34)
[0.12, 0.66]
TGF-α
1.94 (0.43)
1.74 (0.50)
IL-1β
1.18 (0.41)
0.84 (0.61)
IL-6
1.74 (0.49)
1.45 (0.52)
2.95 (40)⁎ 0.95 (40) 1.38 (40) 1.34 (40)
[− 0.22, 0.61] [− 0.15, 0.82] [− 0.14, 0.71]
p < 0.05.
This is the first study to our knowledge examining relationships among depression, EGFR mutation status, and proinflammatory cytokines and growth factors in patients with NSCLC. Additionally, this is the third patient cohort in whom we have found an association between depression and EGFR mutation status [1,25]. In the current study, we identified further that patients with EGFR mutations had lower depressive symptom severity, as we hypothesized. However, contrary to our expectations, patients with EGFR mutations and lower depressive symptom severity had significantly higher levels of TNF-α than patients without EGFR mutations. Patients with EGFR mutations were also less likely to meet criteria for MDD on the diagnostic interview, but this difference did not reach statistical significance. In general, patients meeting criteria for MDD had higher levels of TNF-α compared to patients without MDD. No meaningful differences were observed with regard to other inflammatory biomarkers (i.e., TGF-α, IL-1β, or IL-6) in this small sample. Although biological pathways of inflammation and depression in patients with cancer have been postulated, EGFR has not previously been considered in these models. Constitutive activation of EGFR via oncogenic mutations leads to upregulation of gene expression that may heighten the inflammatory response in certain situations [26,27]. For example, activated EGFR results in COX-2 expression; a pathway that has implications for the development of lung cancer via chronic airway inflammation [28,29]. EGFR signaling also leads to downstream production of IL-6 in human lung cancer cells [30]. A substantial portion of the tumor micro-environment is inflammatory [31] and pro-inflammatory cytokines and growth factors exacerbate cellular processes such as tumor proliferation and inhibit immune responses [32]. In turn, chronic inflammation contributes to the development of malignancies and metastases [33,34]. Integral in these processes, EGFR acts as a key regulator and a cellular hub for inflammatory cytokine signaling, thereby promoting tumor cell proliferation, invasion, migration, metastases, and survival [7–9]. While this may help explain our observed relationship between EGFR mutations and elevated TNF-α in patients with stage IV NSCLC, it makes the relationship between EGFR mutations and lower levels of depression more puzzling. Depression is associated with activation of the immune system, including increased expression of proinflammatory cytokines such as TNF-α and IL-6 [5]. In addition, patients with cancer exhibit higher levels of inflammation [31]. Importantly, evidence suggests that
3.2.1. Serum factors and EGFR mutation status In the subsample excluding patients taking antidepressants, markers of inflammation differed significantly based on EGFR mutation status (Table 4). Patients with EGFR mutations had significantly higher TNF-α levels compared to patients without EGFR mutations, t(35) = 2.17, p = 0.04, 95% CI [0.02–0.58], d = 0.93 (a large effect size). EGFR mutation status was not significantly associated with differences in levels of TGF-α (t[35] = 0.05, p = 0.96), IL-1β (t[35] = 1.03, p = 0.31), or IL-6 (t[35] = 0.94, p = 0.35).
EGFR mutation (n = 8)
Major depressive disorder (n = 7)
4. Conclusions
3.2. Serum inflammatory cytokines and growth factors, EGFR status, and depression
Log-transformed biomarker
Log-transformed biomarker
MDD, t(40) = 2.95, p = 0.005, 95% CI [0.12 = 0.66], d = 1.36 (a large effect size; see Table 5). The presence of MDD was not significantly associated with differences in levels of TGF-α (t[40] = 0.95, p = 0.35), IL-1β (t[40] = 1.38, p = 0.17), or IL-6 (t[40] = 1.34, p = 0.19).
Depression severity scores on the PHQ-9 significantly differed based on mutation status, such that patients with EGFR mutations were less depressed (M = 4.60; SD = 3.34) than patients without EGFR mutations (M = 8.03; SD = 8.03), t(43) = 2.38, p = 0.03, 95% CI [0.46–0.39], d = 0.72 (consistent with a large effect size). Although fewer of the patients with EGFR mutations (n = 10) had concurrent MDD on the MINI (2.08%) compared to those without mutations (n = 38; 27.08%), this difference did not reach statistical significance (Fisher's exact test, p = 0.24; Table 3).
Comparison
Comparison
⁎
3.1. EGFR status and depression
EGFR mutation status (M [SD])
Current depression status
[− 0.41, 0.40] [− 0.24, 0.74] [− 0.22, 0.61]
p < 0.05.
31
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al.
such as depression in patients with stage IV NSCLC and may ultimately lead to the use of tumor genotyping results for selection of depression therapies. However, longitudinal studies are needed to follow these patients over time while re-assessing depressive symptoms and depression status. These findings must also be replicated with a larger sample size to allow for the inclusion of other relevant contributing factors. In the long-term, with enhanced understanding of the molecular mechanisms underlying the relationships among EGFR and depression, there may be an opportunity for novel prevention or intervention strategies to improve quality of life in targeted subsets of patients with stage IV NSCLC.
patients with cancer and co-morbid depression have elevated proinflammatory markers relative to patients with either depression or cancer alone [22,35]. Yet in this study of individuals with stage IV NSCLC, patients with tumors that harbored EGFR mutations, and who had higher levels of TNF-α in their serum, had less depressive symptoms. It is possible that an unmeasured factor associated with EGFR mutations, such as EGF, may decrease the risk of depression in the setting of elevated TNF-α. For instance, a study comparing EGF levels in patients with MDD found significantly lower levels of EGF in those with MDD compared with those without MDD [36]. It is interesting to note that most work examining cytokines and depression finds consistent relationships between IL-6 and depression [37], but are more ambiguous with regard to the relationship between TNF- α and depression [38], which is the opposite of what we found. In addition, evidence suggests that there are different patterns of cytokine activation in subtypes of depression [38]. More work is needed to replicate these findings and examine potential mechanisms explaining how EGFR mutations might be protective for patients with stage IV NSCLC who would usually be at risk for depression given elevated circulating inflammatory markers. While a pilot, the current study contributes novel information regarding the potential biological pathways underlying relationships among depression, tumor genetics, and survival in patients with stage IV NSCLC that have been observed in previous work. Future research should continue in this relatively unstudied area [25]. The results also replicate and expand upon previous findings from our group showing a relationship between tumor gene mutations and depression, in which patients with EGFR mutations were less depressed than those with wildtype tumors [1,25]. Furthermore, in the current study, those that harbored EGFR mutations had similar levels of pain and performance status, ruling out the possibility that lower depression in these patients could be explained by better performance status or lower pain severity. Our findings support that crosstalk between EGFR and inflammationrelated pathways likely exists in patients with stage IV NSCLC, as has been previously observed in cellular and molecular models. A notable limitation of the current pilot study is the small sample size, precluding the use of relevant control variables (e.g., age, gender, body mass index, and smoking history) and the conduct of complex tests of mediation to expand our understanding of the relationships among EGFR, depression, and TNF-α. This is relevant given that patients with wild-type EGFR were more likely to have a smoking history, and smoking is related to depression; although, the directionality of this relationship is not clear [39]. In addition, we did not collect information on certain comorbid diseases and other clinical characteristics that may explain the differences in depression. Results were in the expected direction for the association between depression severity and EGFR status; however, we are limited in our ability to make strong inferences given that not all differences reached statistical significance. Post hoc power calculations indicated 50% power to detect the difference in depression severity between 10 patients with EGFR mutations and 38 patients without at α = 0.05 (effect size = 0.72). Nonetheless, these data can be used to adequately power future studies. Finally, the crosssectional study design prohibits conclusions regarding the direction of the findings and therefore causality. A strength of the current study is the use of an administered diagnostic interview (i.e., the MINI) to assess presence of MDD in addition to depression severity on the PHQ-9. It is noteworthy that patients completed self-report measures and were administered the MINI prior to patients' receiving the results of the genotyping. Therefore, associations between depression and EGFR mutations status were observed prior to patients' awareness of their prognosis or potential responsiveness to treatment. As a result of this timing, we can rule out the possibility that patients' perceived prognosis drove depression severity. Patients living with stage IV NSCLC experience high physical and psychological symptom burden. Despite limitations, our findings have the potential to further our understanding of the etiology of symptoms
Acknowledgements This work was supported by the American Cancer Society (PC1116401). We would also like to acknowledge the patients who dedicated their time to participating in this research study. References [1] W.F. Pirl, L. Traeger, J.A. Greer, et al., Tumor epidermal growth factor receptor genotype and depression in stage IV non-small cell lung cancer, Oncologist 16 (2011) 1299–1306, http://dx.doi.org/10.1634/theoncologist.2011-0116. [2] A.J. Mitchell, M. Chan, H. Bhatti, et al., Prevalence of depression, anxiety, and adjustment disorder in oncological, haematological, and palliative-care settings: a meta-analysis of 94 interview-based studies, Lancet Oncol. 12 (2011) 160–174, http://dx.doi.org/10.1016/S1470-2045(11)70002-X. [3] D.V. Iosifescu, B. Bankier, M. Fava, Impact of medical comorbid disease on antidepressant treatment of major depressive disorder, Curr. Psychiatry Rep. 6 (2004) 193–201 (doi). [4] G.I. Papakostas, T.J. Petersen, D.V. Iosifescu, et al., Somatic symptoms as predictors of time to onset of response to fluoxetine in major depressive disorder, J Clin Psychiatry 65 (2004) 543–546 (doi). [5] Y. Dowlati, N. Herrmann, W. Swardfager, et al., A meta-analysis of cytokines in major depression, Biol. Psychiatry 67 (2010) 446–457, http://dx.doi.org/10.1016/ j.biopsych.2009.09.033. [6] C.L. Arteaga, The epidermal growth factor receptor: from mutant oncogene in nonhuman cancers to therapeutic target in human neoplasia, J. Clin. Oncol. 19 (2001) 32S–40S (doi). [7] M. Scaltriti, J. Baselga, The epidermal growth factor receptor pathway: a model for targeted therapy, Clin. Cancer Res. 12 (2006) 5268–5272, http://dx.doi.org/10. 1158/1078-0432.CCR-05-1554. [8] H.M. Shepard, C.M. Brdlik, H. Schreiber, Signal integration: a framework for understanding the efficacy of therapeutics targeting the human EGFR family, J. Clin. Invest. 118 (2008) 3574–3581, http://dx.doi.org/10.1172/JCI36049. [9] C. Berasain, M.J. Perugorria, M.U. Latasa, et al., The epidermal growth factor receptor: a link between inflammation and liver cancer, Exp. Biol. Med. (Maywood) 234 (2009) 713–725, http://dx.doi.org/10.3181/0901-MR-12. [10] T. Fukuyama, Y. Ichiki, S. Yamada, et al., Cytokine production of lung cancer cell lines: correlation between their production and the inflammatory/immunological responses both in vivo and in vitro, Cancer Sci. 98 (2007) 1048–1054, http://dx. doi.org/10.1111/j.1349-7006.2007.00507.x. [11] K. Kroenke, R.L. Spitzer, J.B. Williams, The PHQ-9: validity of a brief depression severity measure, J. Gen. Intern. Med. 16 (2001) 606–613 (doi, 1495268). [12] M.M. Oken, R.H. Creech, D.C. Tormey, et al., Toxicity and response criteria of the Eastern Cooperative Oncology Group, Am. J. Clin. Oncol. 5 (1982) 649–655. [13] Association AP, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, DSM-IV-TR, American Psychiatric Association, Arlington, VA, 2000. [14] J.R. Fann, D.L. Berry, S. Wolpin, et al., Depression screening using the Patient Health Questionnaire-9 administered on a touch screen computer, Psychooncology 18 (2009) 14–22, http://dx.doi.org/10.1002/pon.1368. [15] K. Ell, B. Xie, B. Quon, et al., Randomized controlled trial of collaborative care management of depression among low-income patients with cancer, J. Clin. Oncol. 26 (2008) 4488–4496, http://dx.doi.org/10.1200/JCO.2008.16.6371. [16] D. Sheehan, Y. Lecrubier, K. Sheehan, et al., The validity of the Mini International Neuropsychiatric Interview (MINI) according to the SCID-P and its reliability, Eur. Psychiatry 12 (1997) 232–241 (doi). [17] D.V. Sheehan, Y. Lecrubier, K.H. Sheehan, et al., The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10, J Clin Psychiatry 59 (Suppl. 20) (1998) 22–33 (quiz 34-57. doi). [18] Organization WH, The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines, World Health Organization, Geneva, 1992. [19] D. Dias-Santagata, S. Akhavanfard, S.S. David, et al., Rapid targeted mutational analysis of human tumours: a clinical platform to guide personalized cancer medicine, EMBO Mol. Med. 2 (2010) 146–158, http://dx.doi.org/10.1002/emmm. 201000070. [20] R.R. Wilcox, Comparing dependent robust correlations, Br. J. Math. Stat. Psychol. (2016), http://dx.doi.org/10.1111/bmsp.12069.
32
Journal of Psychosomatic Research 99 (2017) 28–33
J.M. Jacobs et al. [21] J. Cohen, Statistical Power Analysis for the Behavioral Sciences, 2nd. ed., Academic Press, New York, NY, 1998. [22] D.L. Musselman, A.H. Miller, M.R. Porter, et al., Higher than normal plasma interleukin-6 concentrations in cancer patients with depression: preliminary findings, Am. J. Psychiatry 158 (2001) 1252–1257, http://dx.doi.org/10.1176/appi.ajp.158. 8.1252. [23] S.K. Lutgendorf, D.M. Lamkin, K. DeGeest, et al., Depressed and anxious mood and T-cell cytokine expressing populations in ovarian cancer patients, Brain Behav. Immun. 22 (2008) 890–900, http://dx.doi.org/10.1016/j.bbi.2007.12.012. [24] M.F. O'Connor, J.E. Bower, H.J. Cho, et al., To assess, to control, to exclude: effects of biobehavioral factors on circulating inflammatory markers, Brain Behav. Immun. 23 (2009) 887–897, http://dx.doi.org/10.1016/j.bbi.2009.04.005. [25] W.F. Pirl, L. Traeger, J.A. Greer, et al., Depression, survival, and epidermal growth factor receptor genotypes in patients with metastatic non-small cell lung cancer, Palliat Support Care 11 (2013) 223–229, http://dx.doi.org/10.1017/ S1478951512001071. [26] K. Liu, R.C. Gualano, M.L. Hibbs, G.P. Anderson, S. Bozinovski, Epidermal growth factor receptor signaling to Erk1/2 and STATs control the intensity of the epithelial inflammatory responses to rhinovirus infection, J. Biol. Chem. 283 (2008) 9977–9985, http://dx.doi.org/10.1074/jbc.M710257200. [27] K. Liu, G.P. Anderson, S. Bozinovski, DNA vector augments inflammation in epithelial cells via EGFR-dependent regulation of TLR4 and TLR2, Am. J. Respir. Cell Mol. Biol. 39 (2008) 305–311, http://dx.doi.org/10.1165/rcmb.2007-0458OC. [28] M. Liu, S.C. Yang, S. Sharma, et al., EGFR signaling is required for TGF-beta 1 mediated COX-2 induction in human bronchial epithelial cells, Am. J. Respir. Cell Mol. Biol. 37 (2007) 578–588, http://dx.doi.org/10.1165/rcmb.2007-0100OC. [29] J. Yoo, C.E. Rodriguez Perez, W. Nie, et al., TNF-α induces upregulation of EGFR expression and signaling in human colonic myofibroblasts, Am. J. Physiol. Gastrointest. Liver Physiol. 302 (2012) G805–G814, http://dx.doi.org/10.1152/ ajpgi.00522.2011. [30] S.P. Gao, K.G. Mark, K. Leslie, et al., Mutations in the EGFR kinase domain mediate
[31] [32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
33
STAT3 activation via IL-6 production in human lung adenocarcinomas, J. Clin. Invest. 117 (2007) 3846–3856, http://dx.doi.org/10.1172/JCI31871. F. Balkwill, L.M. Coussens, Cancer: An inflammatory link, Nature 431 (2004) 405–406, http://dx.doi.org/10.1038/431405a. M. Orozco-Morales, G. Soca-Chafre, P. Barrios-Bernal, N. Hernández-Pedro, O. Arrieta, Interplay between cellular and molecular inflammatory mediators in lung cancer, Mediat. Inflamm. 2016 (2016) 3494608, http://dx.doi.org/10.1155/ 2016/3494608. M. Dougan, D. Li, D. Neuberg, et al., A dual role for the immune response in a mouse model of inflammation-associated lung cancer, J. Clin. Invest. 121 (2011) 2436–2446, http://dx.doi.org/10.1172/JCI44796. L. Barrera, E. Montes-Servín, A. Barrera, et al., Cytokine profile determined by datamining analysis set into clusters of non-small-cell lung cancer patients according to prognosis, Ann. Oncol. 26 (2015) 428–435, http://dx.doi.org/10.1093/annonc/ mdu549. C.F. Jehn, D. Kuehnhardt, A. Bartholomae, et al., Biomarkers of depression in cancer patients, Cancer 107 (2006) 2723–2729, http://dx.doi.org/10.1002/cncr. 22294. W. Tian, J. Zhang, K. Zhang, et al., A study of the functional significance of epidermal growth factor in major depressive disorder, Psychiatr. Genet. 22 (2012) 161–167, http://dx.doi.org/10.1097/YPG.0b013e3283539550. E.P. Zorrilla, L. Luborsky, J.R. McKay, et al., The relationship of depression and stressors to immunological assays: a meta-analytic review, Brain Behav. Immun. 15 (2001) 199–226, http://dx.doi.org/10.1006/brbi.2000.0597. S.M. Gold, M.R. Irwin, Depression and immunity: inflammation and depressive symptoms in multiple sclerosis, Immunol. Allergy Clin. N. Am. 29 (2009) 309–320, http://dx.doi.org/10.1016/j.iac.2009.02.008. M. Fluharty, A.E. Taylor, M. Grabski, M.R. Munafò, The association of cigarette smoking with depression and anxiety: a systematic review, Nicotine Tob. Res. 19 (2017) 3–13, http://dx.doi.org/10.1093/ntr/ntw140.