INTRODUCTION
Colorectal cancer (CRC) is one of the most frequently diagnosed malignancies in high-income countries [1]. Beyond its physical impact, patients with CRC experience psychological burden, with a study indicating that up to 19% of patients develop depression [2]. The risk is particularly elevated in the first year after diagnosis—typically the period of active anticancer treatment—during which patients with CRC have an approximately 12-fold higher hazard of developing depression than cancer-free individuals [3]. Indeed, among patients who have undergone surgical resection, depression and anxiety are significantly higher in patients receiving 5-fluorouracil chemotherapy compared to those who do not [4].
The clinical impact of these depressive symptoms is substantial, as they are associated not only with poorer quality of life and reduced physical functioning [2], but also with increased mortality risk and accelerated CRC progression [5]. These associations may reflect lifestyle factors or treatment non-adherence, but growing evidence suggests that the underlying biological mechanism involves the dysregulation of the immune system [6,7].
Indeed, depression can dysregulate immunity and influence multiple phases of tumour progression, particularly in inflammation-associated malignancies such as CRC [6,7]. Specifically, pro-inflammatory activation is often heightened in depression [8], and dysfunction of the hypothalamic-pituitary-adrenal (HPA) axis can weaken cellular immunity by inhibiting the proliferation of T cells and natural killer (NK) cells, both of which are critical for anti-tumour immunity [6,9]. Furthermore, depression-associated behaviours (e.g., smoking, physical inactivity, sleep disturbance) may further exacerbate immune dysregulation [9].
Based on these theoretical foundations, researchers have investigated the association between depression and immune markers; however, the results have been inconsistent. For instance, several studies have found that depression can result in both immune suppression and immune activation [9]. Reduced NK cell cytotoxicity, along with elevated levels of interleukin (IL)-6 and C-reactive protein, has been reported in individuals with depression [10]. In contrast, another study found no association between depression and increased levels of immune markers in a different cohort [11]. These inconsistencies may be due to the heterogeneous nature of the populations studied, sex differences, but an increasingly recognized explanation is the moderating role of genetic factors [9].
This perspective aligns with the Gene-Environment (G×E) Interaction model, which posits that an individual’s genetic predisposition can moderate their response to environmental stressors. A classic example is the serotonin transporter (5-HTT or SERT) gene polymorphism, which has been shown to moderate the influence of stressful live events on depression [12]. This specific gene polymorphism has also been identified as a predictor of poor survival in CRC patients [13,14], and is known to have a significant impact on immune parameters such as IL-6 levels [15,16].
Therefore, based on the G×E model, this study aims to investigate whether the relationship between changes in depression and changes in lymphocyte percentages is moderated by 5-HT transporter gene-linked polymorphism (5-HTTLPR) in a clinical setting. Our primary hypothesis is that the 5-HTTLPR genotype moderate the association between changes in depression and changes in lymphocyte parameters among CRC patients undergoing chemotherapy.
METHODS
Study design and recruitment
This longitudinal cohort study with a prospective design collected data on the impact of depression and sleep disturbances in patients with CRC undergoing 5-fluorouracil, leucovorin, and oxaliplatin (FOLFOX)-based chemotherapy. Results concerning sleep disturbances and circadian genes within this cohort have already been published [17]. In the current study, we focused on data related to depression and the 5-HTTLPR gene.
Participants were followed at two time points: the first (baseline) and the final cycles of FOLFOX chemotherapy. Blood samples for lymphocyte calculation were collected one day before chemotherapy at each follow-up point to minimise the effects of chemotherapy on lymphocyte percentages.
The study recruited participants who were admitted to receive their first cycle of FOLFOX chemotherapy at Chonnam National University Hwasun Hospital, South Korea, from December 2012 to November 2017. Inclusion criteria included being aged 18 or older, diagnosed with stage II CRC at high risk for systemic recurrence or stage III CRC, and scheduled to receive adjuvant FOLFOX treatment (8-12 cycles). Participants also needed to be fluent in Korean. Exclusion criteria included plans for transfer to another hospital, underlying intellectual disabilities, dementia, delirium, or psychotic disorders. Patients diagnosed with hematologic diseases or double primary cancers were also excluded. This study was approved by the Chonnam National University Hwasun Hospital review board (2012-169), and written informed consent was obtained from all participants.
We confirm that all methods were performed in the present study are in accordance with the relevant guidelines and regulations. The study was performed in accordance with the approved guidelines and the Declaration of Helsinki.
Demographic and clinical characteristics
Demographic characteristics such as age, sex, and education level were collected. Clinical status data included cancer stage; disease site, classified as colon or rectum; time elapsed since cancer diagnosis and subsequent surgery; history of insomnia, depression, or anxiety; medication history for sleep pills and antidepressants; number of comorbid physical illnesses; Eastern Cooperative Oncology Group performance status; mean dose of FOLFOX chemotherapy; and the number of chemotherapy cycles, ranging from 9 to 12.
Measurement of depression
We determined the clinical status of depression by measuring changes in scores using the Korean version of the Hospital Anxiety and Depression Scale-depression subscale (HADS-D) [18,19]. This scale includes 7 items, each scored from 0 to 3, with total scores ranging from 0 to 21. A higher score indicates more severe depression. The HADS-D is recommended for assessing depression in medically ill patients because it excludes physical symptoms of depression, such as fatigue, appetite loss, or weight changes [20], helping to reduce the overestimation of depression symptoms due to underlying medical conditions such as cancer. In this study, we measured changes in HADS-D scores between two follow-up points. A positive “delta HADS-D score”—calculated as the follow-up score minus the baseline score—indicated a worsening of depression during the study period.
Immunological assays
We analysed the percentages of CD4+ and CD8+ T lymphocytes, CD19+ B lymphocytes, and CD16/56+ NK cells at baseline and follow-up. In detail, 5 mL of peripheral blood was drawn in the ethylenediamine tetraacetic acid tubes prior to the initiation of chemotherapy. To minimise the potential influence of diurnal lymphocyte fluctuations, samples were collected in the afternoon (14:00 to 18:00). Total white blood cell (WBC) counts and lymphocyte percentage were measured as reference markers to assess the effect of chemotherapy-induced myelosuppression, which could affect the reliability of the assays.
Immunophenotyping was conducted using a “3-colour panel” of monoclonal antibodies, including CD45-PerCP for lymphocyte gating (Becton Dickinson Biosciences) and FITC/PE-labelled IgG1/IgG1 for isotype control. Specific antibodies used included CD3/CD4, CD3/CD8, CD3/CD19, and CD3/CD16+56 (Beckman Coulter, Inc.). Five tubes were used for the immunophenotyping: IgG1/IgG1/CD45, CD3/CD4/CD45, CD3/CD8/CD45, CD3/CD19/CD45, and CD3/CD16+56/CD45. The samples were then analysed using a FACSCaliburTM flow cytometer and CellQuest software (Becton Dickinson Biosciences).
Genotyping
The 5-HTTLPR is a variant of the 5-HTT gene located in the promoter region of the serotonin transporter gene (SLC6A4). This functional polymorphism presents two types of alleles, ‘s’ and ‘l,’ differentiated by a 44-bp insertion or deletion. The s allele is known to have lower transcriptional efficiency of the promoter compared to the l allele. As a result, the s allele leads to less serotonin reuptake compared to the l allele due to reduced 5-HTT mRNA expression [21].
For genotyping, DNA was extracted from blood samples. The primers 5-HTTLPR-F 5' (GGC GTT GCC GCT CTG AAT GC) 3' and 5-HTTLPR-R 5' (GAG GGA CTG AGC TGG ACA ACC AC) 3' were used for the PCR. The PCR reaction was performed with 20 ng of genomic DNA as the template in a 30 μL reaction mixture using Dr. HOT DNA Polymerase. The procedure included activation of Taq polymerase at 95°C for 15 minutes, followed by 35 cycles at 95°C for 1 minute, 64°C for 1 minute, and 72°C for 1 minute, concluding with a 30-minute extension at 72°C. The DNA samples containing the extension products were added to Hi-Di formamide and GeneScanTM 600 LIZ® Size Standard (Applied Biosystems). The mixture was incubated at 95°C for 5 minutes, followed by cooling on ice for 5 minutes, and then analysed using a 3730XL DNA Analyzer (Applied Biosystems). The genotypes were categorised as s/s, s/l, and l/l based on allele status.
Statistical analysis
We calculated changes in lymphocyte percentages (CD4+, CD8+, CD19+, and CD16/56+) and delta HADS-D scores between two time points. Participants with more than 5% missing data across the predictors were excluded from the analysis. Among the remaining participants, missing responses on individual HADS-D items were addressed using multiple imputation, implemented in SPSS version 27.0 (IBM Corp.) with the fully conditional specification method. Imputations were performed using linear regression modeling, and five imputed datasets were generated and pooled for the final analysis. Auxiliary variables, including age, sex, and cancer stage, were incorporated into the imputation model to improve estimation precision. The dyadic associations of changes in depression and lymphocytes, as well as the analysis of 5-HTTLPR polymorphism and lymphocytes, were analysed using linear regression analysis with covariates.
Finally, to test our G×E hypothesis, a moderated regression analysis was performed to investigate the moderating effect of 5-HTTLPR on the relationship between depression and lymphocyte percentages. To reduce potential bias related to differential attrition, all candidate predictors were included as covariates in the regression models, thereby accounting for variables that significantly differed between study completers and dropouts. The sociodemographic and clinical data were entered as covariates in step 1, followed by delta HADS-D scores in step 2, 5-HTTLPR allele status in step 3, and the interaction between delta HADS-D scores and allele status in step 4. Our primary analysis employed an additive model, coding the number of ‘s’ alleles as 0, 1, and 2. A sensitivity analysis was then conducted using a dominant model, specifically to address the statistical power issue arising from the small l/l group in the Korean population. We compared ‘l’ allele carriers (l/l and s/l) against s/s homozygotes. Delta HADSD scores were mean-centred to reduce multicollinearity. To control for type I errors in multiple testing, we reported false discovery rate (FDR) adjusted p-values in the moderated regression analysis [22]. Statistical analyses were performed using SPSS software (version 27.0; IBM Corp.).
RESULTS
Recruitment
Of the 278 patients who met the inclusion criteria during the study period, 25 could not be contacted, and 91 refused to participate. Of the 162 patients (58.3%) who enrolled in the study, 105 completed the follow-up. Among those, one participant had complete missing data on the HADS-D scale and was excluded based on the predefined threshold of >5% missing data. The final analytic sample thus included 104 (62.2%) participants (Figure 1). Of these, two participants (1.9%) had one missing item each on the HADS-D scale, which were imputed using multiple imputation.
Attrition was significantly associated with the following baseline characteristics: a shorter time elapsed since cancer diagnosis (8.2 weeks vs. 9.5 weeks, p=0.022), lower doses of oxaliplatin per cycle (68.6 mg vs. 73.6 mg, p=0.017), and a lower number of chemotherapy cycles (10.0 cycles vs. 11.6 cycles, p=0.001) compared to those who completed the study.
Demographic and clinical characteristics
Table 1 presents the demographic and clinical characteristics of the patients based on their allele type. More than half of the patients were female (51%), and the mean age was 60.8 years. Most patients were diagnosed with TNM stage III CRC (86.5%) and had cancer originating in the colon (80.8%) rather than the rectum (19.2%). Only 6 patients (5.8%) reported a past history of depression. Although we surveyed the medication history for antidepressants and sleep pills, no patient reported a history of antidepressant use. Therefore, the medication history for antidepressants was excluded from this analysis. In the univariate analysis using analysis of variance, no variable was significantly associated with the allele type of 5-HTTLPR (p>0.05).
Depression
The baseline mean (SD) score of HADS-D was 4.58 (2.9), with a range of 0 to 17.0. The mean (SD) HADS-D score during the follow-up period was 4.49 (2.7), with a range of 0 to 16.0. The mean (SD) change in HADS-D scores between baseline and follow-up was -0.09 (3.0), with a range of -9.0 to 6.0. The baseline HADS-D scores and the changes in HADS-D scores during the follow-up period were not significantly different among the 5-HTTLPR alleles (p>0.05).
Genotyping
Among the 104 participants, 69 (66.3%) had the s/s allele, 29 (27.9%) had the s/l allele, and 6 (5.8%) had the l/l allele. The allele frequency of the 5-HTTLPR in this Korean sample was comparable to that reported in previous studies [23]. The genotype distributions in this sample complied with the Hardy-Weinberg equilibrium (p=0.341).
Immunological assays.
The WBC count at baseline and follow-up was 6.17 (1.37) and 5.06 (1.63) mean (SD) ×103/μL, indicating that the count remained within the normal range at both time points. Similarly, the percentage of total lymphocytes remained within the normal range: 34.5 (10.2) and 33.3 (10.7) mean (SD) % at baseline and follow-up, respectively.
For dyadic associations between lymphocytes and allele type, the results are summarised in Table 2. Changes in the percentages of CD4+ and CD19+ cells were significantly associated with the number of s alleles of the 5-HTTLPR polymorphism (p=0.025 and p=0.032, respectively). In contrast, changes in percentages of other lymphocytes (CD8+ and CD16/56+) were not significantly different by 5-HTTLPR allele type (all p>0.05). Additionally, no significant associations were observed between changes in lymphocyte percentages (CD4+, CD8+, CD19+, and CD16/56+) and changes in HADS-D scores (all p>0.1).
Moderated regression analysis
Moderated regression analyses were conducted to examine the moderating role of 5-HTTLPR polymorphism on the relationship between HADS-D scores and changes in lymphocyte percentages. Two significant results were found after FDR correction: the interactions significantly influenced the changes in CD4+ and CD16/56+ lymphocyte percentages during FOLFOX chemotherapy (B=-2.21, p=0.002 and B=1.82, p=0.016, respectively). The inclusion of each interaction in step 4 of the analysis explained 8.0% and 5.7% of the variance in changes in CD4+ and CD16/56+ lymphocyte percentages, respectively. These results are described in Table 3 and Figure 2.
In detail, for patients without the s allele of 5-HTTLPR, an increase in the HADS-D score corresponded to an increase in the CD4+ lymphocyte percentage during FOLFOX chemotherapy. However, the presence of an s allele tended to attenuate the slope of the regression lines. For patients with two s alleles of 5-HTTLPR, the slope was reversed, indicating that the s allele negatively moderated the effect of changes in the HADS-D score on the changes in CD4+ lymphocyte percentages.
In the case of CD16/56+ lymphocytes, an increase in the HADS-D score decreased the lymphocyte percentage in the group without the s allele of 5-HTTLPR during chemotherapy. The presence of an s allele negatively moderated this relationship. In patients with two s alleles, the slope of the regression line reversed, indicating that increases in the HADS-D score were associated with an increase in CD16/56+ lymphocyte percentages during chemotherapy.
Sensitivity analysis
The results of the sensitivity analysis were largely consistent with our primary analysis. The interaction between changes in the HADS-D score and the 5-HTTLPR genotype remained highly significant for the change in CD4+ lymphocyte percentage (B=-2.42, p=0.005). The interaction also remained significant for the change in CD16/56+ lymphocyte percentage (B=1.87, p=0.039), although this result did not withstand FDR correction. Similarly, no significant interaction was observed for changes in CD8+ or CD19+ lymphocyte percentages (all p>0.5), consistent with the primary analysis.
DISCUSSION
This prospective study provides preliminary evidence for a G×E interaction between the 5-HTTLPR polymorphism and depressive symptoms on immunological changes among CRC patients receiving FOLFOX chemotherapy. Our primary finding is that 5-HTTLPR consistently moderated the relationship between worsening depression and changes in the percentages of CD4+ cells. A similar, supportive interaction was also observed for CD16/56+ NK cells. These findings are noteworthy given that a variation in the 5-HTT gene has already been reported as a predictor of poor survival in CRC patients [14]. Our results suggest a potential immune-mediated pathway for this clinical observation, establishing a crucial role for 5-HTTLPR in the complex relationship between depression and the immune system in this specific clinical context.
It is interesting that the dyadic analysis for the pair of “depression- lymphocytes” shows no significant interaction. However, under the moderation of 5-HTTLPR, significant GxE interactions emerged. This was most pronounced for CD4+ T cells, where the interaction was statistically significant in both our primary additive model and a subsequent dominant model sensitivity analysis, underscoring the consistency of this core finding. The interaction for CD16/56+ cells also showed a consistent pattern, with the percentage increasing as depression symptoms worsened in CRC patients with the s allele in both models. This suggests genetic moderation for the patients under the stress of cancer and FOLFOX chemotherapy.
This phenomenon highlights the potential impact of 5-HTTLPR on the immune system among CRC patients. In a twin study, the significant role of this gene in the shared genetic liability for neural-immune dysregulation was demonstrated, where the 5-HTT gene accounted for about 10% of the correlation between depressive symptoms and IL-6 levels [16]. The 5-HTTLPR polymorphism is related to SERT activity. SERT activity plays a crucial role in the regulation of serotonin levels, which is known for its signalling role in the proliferation and activity of immune cells such as T, B, and NK lymphocytes [24]. Those lymphocytes have critical roles for anti-tumour immunity in cancer patients.
Generally, the s allele of 5-HTTLPR is considered the ‘risk allele’ for depression and psychiatric conditions due to decreased SERT activity [25]. Existing meta-analyses have shown that depressed patients had higher mean absolute CD4+ T cell and NK cell counts, indicating enhanced inflammation [26]. Therefore, it is plausible that participants with the s allele exhibited an increase in the percentage of NK cells as their depression worsened in this study. Enhanced activity in the amygdala and HPA axis observed in individuals with the 5-HTTLPR s/s genotype may be involved in this correlation [15,27].
Meanwhile, for CD4+ T cells, the percentage increased as depression worsened under the moderation effect of the l allele but decreased under the s allele. These results may indicate that 5-HTTLPR has different roles in the central and peripheral systems. For instance, the l allele has been reported as a risk factor for IBS development due to enhanced SERT activity, which decreases gut secretion and motility [28]. Additionally, in a case-control study in Brazil, the l allele was associated with increased numbers of CD4+ T lymphocytes among fibromyalgia patients [22]. These findings partially align with our results, showing that the l allele is associated with increased CD4+ T lymphocytes as depressive symptoms increase.
Despite the complexity of the findings, the results hold important implications. First, our results highlight the critical importance of applying the G×E model to clinical cancer research for the relationship between depression and immunity. To our knowledge, this is the first report on the moderating role of 5-HTTLPR in the associations between depression and immune parameters in a clinical cancer sample. Second, investigating clinical depression and immune parameters enhances clinical interest, although the clinical impact of these findings on cancer progression remains unclear. It is noteworthy that CD4+ T cells and NK cells are well known to play important roles in anti-tumour immunity [29]. Previous reports have indicated that the s/s allele of 5-HTTLPR was a predictor of poor survival in CRC patients [14]. Therefore, the immune modulation by 5-HTTLPR polymorphism under depression warrants further investigation.
This study has several methodological strengths. The longitudinal design enabled a comprehensive investigation of clinical depression and immune parameters. Additionally, we conducted statistical analyses that accounted for potential confounding factors, such as demographic and clinical information. Using the HADS-D scale to assess depression is another methodological strength, as this scale does not include physical symptoms of depression such as fatigue, appetite, and weight loss, which could be confused with symptoms of inflammation. Finally, the study was performed within a relatively homogeneous group, specifically advanced CRC patients who were receiving FOLFOX chemotherapy.
Meanwhile, the conclusions of this study should be interpreted cautiously due to several limitations. First, this study was conducted with CRC patients undergoing FOLFOX chemotherapy, so the results should not be generalised to the general population. Second, the depressive symptoms observed in our cohort were, on average, in the mild to moderate range, and the mean change in scores over study was subtle. However, the changes in scores spanned a substantial range (e.g. -9.0 to 6.0 for mean change), which provided the necessary statistical power to detect the interaction in our regression model. Nevertheless, as the cohort was not primarily composed of patients with major depressive disorder, the findings should not be generalized to patients with severe depression. Third, ethnic differences should be considered. The high frequency of the s allele of 5-HTTLPR in the Korean population could affect the statistical power of the study. Further studies in various ethnic populations are needed. Fourth, the small sample size and resulting low minor allele frequency created a potential risk of bias, although the consistency of our primary finding was supported in the sensitivity analysis. Fifth, our study did not measure functional immune markers. Future studies incorporating functional assays, such as NK cell cytotoxicity, are warranted to elucidate the clinical implications of the findings. Sixth, this study did not include an a priori sample size calculation. A post hoc power analysis indicated limited statistical power (0.54), and thus the findings should be interpreted as exploratory. Finally, the potential impact of FOLFOX chemotherapy on lymphocyte subsets should be considered. Although the mean WBC count remained within the normal range in the follow-up sample, the myelosuppressive effects of FOLFOX chemotherapy may have influenced the lymphocyte subset proportions. Previous studies have reported increases in CD8+ T cells but decreases in CD19+ B cells after FOLFOX chemotherapy [30]. Finally, although a broad range of clinical and psychological predictors was included, the possibility of unmeasured confounding or model misspecification due to omitted variables cannot be excluded. These limitations may affect the validity and generalizability of the results.
In conclusion, the changes in the percentages of CD4+ T cells and CD16/56+ NK cells under depression were moderated by 5-HTTLPR alleles among CRC patients undergoing FOLFOX chemotherapy, suggesting a G×E interaction between the brain and the body. A decreasing percentage of CD4+ T cells and an increasing percentage of NK cells were related to the s allele as depression worsened. Considering the high comorbidity of depression among cancer patients, greater attention and further research on the role of 5-HTTLPR in the immune system under depression among cancer patients are warranted.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
