Abstract
Background:
Fatigue is a common side effect of cancer and its treatment and is thought to be driven in part by activation of the proinflammatory cytokine network. However, the cellular and molecular underpinnings of cancer-related fatigue (CRF) have not been determined, nor have immune pathways beyond inflammation been carefully investigated. The goal of this study was to examine the association between CRF and activation of canonical proinflammatory gene regulation pathways and Type I interferon (IFN) signaling pathways in breast cancer patients during and after treatment.
Methods:
Women diagnosed with early-stage breast cancer (n = 181) completed assessments before and after treatment with radiation and/or chemotherapy and at 6, 12, and 18-month post-treatment follow-ups. Assessments included self-reported fatigue (Multidimensional Fatigue Symptom Inventory – Short Form) and expression of pre-specified sets of Type I IFN and pro-inflammatory immune response genes determined from mRNA sequencing of PBMCs. Mixed effect linear models examined changes in fatigue and immune gene expression over time and tested the hypothesis that fatigue would be associated with increased expression of Type I IFN and inflammatory response genes.
Results:
There were significant changes in fatigue and immune gene expression across the assessment period; all measures increased from pre- to post-treatment but showed diverging patterns over the follow-up, with declines in fatigue and persistent elevations in Type I IFN and proinflammatory gene expression. In mixed effect linear models, expression of Type I IFN response genes was elevated in association with fatigue across the assessment period, from pre-treatment to 18-month follow-up. In contrast, pro-inflammatory gene expression was associated with fatigue only at 6, 12, and 18-month follow-ups. Analyses controlling for changes in leukocyte subsets continued to show a significant association between fatigue and Type I IFN gene expression but reduced the time-dependent association with pro-inflammatory gene expression to non-significant.
Conclusions:
Results revealed unexpected complexity in the immune underpinnings of CRF and identify a novel role for IFN signaling as a robust contributor to this symptom before, during, and after treatment. Pro-inflammatory gene expression emerged as a predictor of fatigue later in the cancer trajectory, and that effect was primarily accounted for by a concurrent increase in monocyte prevalence.
INTRODUCTION
Fatigue is a common side effect of cancer diagnosis and treatment that interferes with all aspects of quality of life.1 Fatigue can be elevated before treatment onset, typically increases after receipt of surgery, radiation, and/or chemotherapy, and may persist for months after treatment completion in a substantial minority of patients.2 There is considerable interest in identifying the biological mechanisms underlying cancer-related fatigue (CRF), with a focus on the immune system.3 Based on animal and human experimental studies documenting effects of peripheral inflammation on fatigue and other “sickness behavior”, our group and others have hypothesized that CRF may be driven in part by activation of the proinflammatory cytokine network.4–6 Indeed, studies have shown that fatigue is associated with elevations in markers of inflammation among cancer patients and survivors.2,7
The majority of studies examining links between inflammation and CRF have been cross-sectional in design and have used blood protein measures of canonical pro-inflammatory cytokines such as IL-1β, TNF-α, and IL-6 (or reporter proteins for these cytokines such as IL-1ra, sTNF-RII, and CRP) to index inflammatory activity.2 A smaller number of longitudinal studies have also been conducted, typically while patients are undergoing treatment and again focusing on blood protein cytokine markers.8–11 Findings from these studies generally support a link between inflammation and cancer-related fatigue, but provide little insight into the cellular or molecular mechanisms involved or how these processes might change over the course of treatment and into survivorship. Initial evidence suggests activation of the pro-inflammatory transcription factor NF-κB in survivors with persistent fatigue,12,13 but only two longitudinal studies have used an RNA-based approach to identify molecular underpinnings of CRF, both conducted with head and neck cancer patients.14,15
To date, there has been less focus on immune pathways beyond inflammation that might potentially contribute to CRF. However, there is good reason to suspect that other pathways may be involved given findings from early clinical studies implicating Type I interferons (IFNs; IFN-α, −β, and −ω) in fatigue during cancer treatment.16 Type I IFNs mediate a complex array of pro- and anti-inflammatory effects on the immune system17 and act via multiple signaling pathways beyond NF-κB, including the Interferon Response Factor (IRF) and Signal Transducers and Activators of Transcription (STAT) family of transcription factors. Type I IFNs are physiologically activated by infection and are well-known for triggering innate antiviral responses and promoting cellular immunity. In the context of cancer, Type I IFNs can also be activated by the genotoxic effects of common cancer therapies, including radiation18 and chemotherapy19, and have been shown to promote antitumor immunity.20 Of note, fatigue and depressive symptoms are often treatment-limiting side effects of Type I IFN therapy.21
The neural and behavioral effects of Type I IFNs have been elegantly documented in clinical studies of individuals undergoing treatment with IFN-α for melanoma or hepatitis C (HCV).22 Pharmacologic IFN-α leads to symptoms of fatigue that emerge early in the treatment course (along with other neurovegetative symptoms)23,24 and can persist well after treatment in a subset of patients,25 similar to CRF.26 However, it is unclear whether IFN-α has direct effects on the brain that result in fatigue, or whether these effects are driven by secondary mediators, such as NF-κB activation of canonical proinflammatory cytokines. In one report, administration of IFN-α in patients with HCV led to acute (within 4 hours) increases in plasma concentrations of IFN-α that co-occurred with increases in fatigue.24 In contrast, changes in plasma concentrations of proinflammatory cytokines were more modest, with increases seen for IL-6 but not TNF-α or IL-1β, leading the authors to conclude that IFN-α may have a direct effect on the brain (vs mediation by peripheral inflammatory processes). Further, chronic administration of IFN-α is associated with persistent elevations in plasma (and CSF) concentrations of IFN-α, but not canonical proinflammatory cytokines.27,28
The goal of this study was to examine the association between CRF and activation of 1) canonical proinflammatory gene regulation pathways (e.g., NF-κB), and 2) Type I IFN signaling pathways (IRF, STAT) in whole genome leukocyte transcriptome profiles from a prospective, longitudinal study of breast cancer patients recruited after diagnosis and tracked over 18 months post-treatment (the RISE study29,30). We have previously identified clinical and psychological correlates of fatigue at baseline29 and over the follow-up period30 in the RISE sample, as well as treatment-related changes in plasma concentrations of canonical pro-inflammatory cytokine proteins.31 In this study, we sought to understand the gene regulatory origin of these dynamics over the course of early breast cancer survivorship. Specifically, we tested the hypothesis that fatigue would be associated with increased expression of pro-inflammatory and Type I IFN genes in women with breast cancer assessed up to 5 times from diagnosis through early survivorship.
METHOD
Patients and Procedures
Patients were recruited from oncology practices between 1/2013 and 7/2015 to participate in a prospective, longitudinal study of cancer-related fatigue (RISE study).29,30 Inclusion criteria were: 1) diagnosed with Stage 0-IIIA breast cancer; 2) had not yet started adjuvant or neoadjuvant therapy with radiation, chemotherapy, or endocrine therapy; and 3) English proficient. Primary recruitment sites were UCLA and Cedars Sinai Medical Center (CSMC) in Los Angeles.
Participants completed up to 5 study assessments from diagnosis to 18 months post-treatment designed to capture acute and more chronic effects of cancer treatments. Assessments were conducted prior to receipt of (neo)adjuvant therapy with radiation (RT), chemotherapy (CT), and/or endocrine therapy (ET) (pre-treatment assessment); after completion of RT and/or CT for women who received those treatments (post-treatment assessment); and at 6, 12, and 18-month post-treatment follow-ups. At each assessment, participants completed online questionnaires and provided blood samples for immune assessment. The Institutional Review Boards at UCLA and CSMC approved the study, and all participants provided written informed consent.
The RISE study enrolled 270 patients.29 The current analyses focused on 181 women who agreed to provide blood samples, had at least one blood sample available for gene expression analyses, did not have comorbid autoimmune disease, had questionnaire data available for assessment of fatigue, and had complete data for covariates. The number of samples analyzed at each assessment was as follows: pre-treatment = 169, post-treatment = 123, 6-month follow-up = 148, 12-month follow-up = 137, and 18-month follow-up = 139. Note that the post-treatment assessment point had the fewest samples/participants because this assessment was only conducted for women who received RT and/or CT.
Measures
Demographic characteristics
Demographic characteristics were obtained from self-report at baseline and included age, race/ethnicity, marital status, income, education, and employment status.
Disease and treatment-related information
Disease and treatment-related information was obtained from medical record abstraction and included cancer stage, type of primary surgery (lumpectomy, unilateral mastectomy, bilateral mastectomy), and type of adjuvant therapy received (RT, CT, and/or ET).
Fatigue
Fatigue was assessed with the Multidimensional Fatigue Symptom Inventory-Short Form, a 30-item scale designed to assess the principal manifestations of fatigue in cancer patients and survivors.32,33 We focused on the General Fatigue subscale of the MFSI-SF, a 6-item scale that captures the core components of cancer-related fatigue (i.e., tired, fatigued, worn out, sluggish, run down, “pooped”). Supplementary analyses were conducted using the MFSI-SF total score, which includes subscales for general fatigue, physical fatigue, mental fatigue, emotional fatigue, and vigor.
PBMC gene expression
PBMC gene expression was assessed using RNA extracted (Qiagen RNeasy) from peripheral blood mononuclear cells isolated from 10 ml venipuncture samples collected into sodium heparin Vacutainers, tested for suitable input RNA mass (> 10 ng total RNA, by PicoGreen RNA), and subject to genome-wide transcriptional profiling using a high efficiency mRNA-sequencing assay (Lexogen QuantSeq 3’ FWD) in the UCLA Neuroscience Genomics Core following the manufacturer’s standard protocol. Assays were performed in a single batch and targeted 10 million sequencing reads per sample (achieved median = 12.8 million), each of which was mapped to the GRCh38 reference human transcriptome using the STAR aligner to quantify transcript abundance (achieved median mapping rate = 92.6%). Transcript abundance data were normalized counts per million mapped reads and log2 transformed for statistical analyses as described below. Samples yielding < 5 million total sequencing reads were excluded from further analysis, as were 3 of our a priori-targeted gene transcripts that yielded minimal expression levels and/or variability (average < .01 or SD < .2 log2 normalized transcript counts per million).
Statistical Analyses
Analyses were first conducted to examine changes in fatigue and in expression of Type I IFN and inflammatory response genes across the assessment period using mixed effect linear models. Mixed effect linear models were also used to test the primary hypothesis that fatigue would be associated with expression of Type I IFN and inflammatory response genes. For IFN, analyses focused on average (log2) mRNA abundance for a pre-specified set of 32 Type I IFN response genes used in previous research34 (GBP1, IFI16, IFI27, IFI27L1, IFI27L2, IFI30, IFI35, IFI44, IFI44L, IFI6, IFIH1, IFIT1, IFIT1B, IFIT2, IFIT3, IFIT5, IFITM1, IFITM2, IFITM3, IFITM4P, IFITM5, IGLL1, IRF2, IRF7, IRF8, JCHAIN, MX1, MX2, OAS1, OAS2, OAS3, OASL). Parallel analyses were conducted using a pre-specified set of 19 canonical proinflammatory response genes used in previous research (CXCL8, FOS, FOSB, FOSL1, FOSL2, IL1A, IL1B, IL6, JUN, JUNB, JUND, NFKB1, NFKB2, PTGS1, PTGS2, REL, RELA, RELB, TNF).34 Analyses controlled for correlation among indicator genes through a random participant intercept, and specified fixed effects of longitudinal time point (repeated measure: pre-treatment baseline, post-treatment follow-up, 6-month follow-up, 12-month follow-up, 18-month follow-up). All analyses included covariates that may be associated with cancer-related fatigue and/or immune parameters.2 These included participant age, race (white, non-white), ethnicity (Hispanic, non-Hispanic), body mass index (BMI), smoking history (current, former, never), cancer stage (0/1 vs. 2/3), surgery type (lumpectomy, unilateral mastectomy, bilateral mastectomy), type of adjuvant treatment received (RT only, CT only, RT + CT, and no RT or CT), and endocrine therapy. Surgery and endocrine therapy were coded as time varying. To account for potential differences in fatigue and immune activity over time related to adjuvant therapy, models were also conducted including treatment group by time interactions as indicated. To determine whether alterations in peripheral blood mononuclear cell mRNA profiles might stem from alterations in the relative abundance of major leukocyte subsets, ancillary analyses controlled for mRNA indicators of major leukocyte subsets (CD14, CD3D, CD4, CD8A, CD19, CD56/NCAM1, CD16/FCGR3A). Analyses were implemented using SAS PROC MIXED with maximum likelihood estimation.
RESULTS
Sample characteristics
Demographic, disease, and treatment-related characteristics of the 181 women included in this sample are shown in Table 1. Women were on average 55 years old, non-Hispanic White, employed, and partnered, with few medical co-morbidities, comparable to the full RISE sample.30 The majority had been diagnosed with Stage 0 or I breast cancer, had a lumpectomy, and were treated with RT either alone or with CT. Over 60% received endocrine therapy.
Table 1.
Sample characteristics
| Total sample (N = 181) | |
|---|---|
| Demographic and general health | |
| Age, mean (SD) | 55.3 (11.0) |
| BMI, mean (SD) | 25.3 (5.7) |
| Race | |
| Asian | 21 (11.6%) |
| Black | 6 (3.3%) |
| White | 136 (75.1%) |
| Other | 18 (9.9%) |
| Hispanic | 18 (9.9%) |
| Income (n = 3 missing), n (%) | |
| Under $60K | 44 (25.0%) |
| $60–100K | 36 (20.0%) |
| $100K or more | 98 (55.0%) |
| Education, n (%) | |
| College or less | 121 (66.9%) |
| Postgraduate | 60 (33.1%) |
| Employed, n (%) | 118 (65.2%) |
| Partnered, n (%) | 120 (66.3%) |
| Charlson Co-morbidity Scale, n (%) | |
| 0 | 141 (77.9%) |
| 1 | 28 (15.5%) |
| 2 or 3 | 12 (6.6%) |
| Smoking Status, n (%) | |
| Never | 121 (66.9%) |
| Former | 55 (30.4% |
| Current | 5 (2.8%) |
| Disease and treatment-related | |
| Stage, n (%) | |
| 0 or I | 109 (60.2%) |
| II or III | 72 (39.8%) |
| Receipt of adjuvant therapy, n (%) | |
| No RT or CT | 37 (20.4%) |
| RT only | 73 (40.3%) |
| CT only | 16 (8.9%) |
| CT + RT | 55 (30.4%) |
| Neoadjuvant chemotherapy, n (%) | 19 (10.5%) |
| Surgery type, n (%) | |
| Lumpectomy | 113 (62.4%) |
| Unilateral mastectomy | 17 (9.4%) |
| Bilateral mastectomy | 56 (30.9%) |
| Receipt of endocrine therapy | 112 (61.2%) |
Note: Women could have undergone more than type of surgery, so the total number of surgeries exceeds the total number of participants. BMI: body mass index; CT: chemotherapy; RT: radiation therapy
Changes in fatigue over time
Mixed effect linear models showed a significant effect of time for MFSI-SF general fatigue [F (4,512) = 2.87, p = .02], after control for participant age, race, ethnicity, BMI, smoking history, cancer stage, surgery type, type of adjuvant treatment received, and endocrine therapy. Examination of adjusted mean scores on the MFSI general fatigue subscale demonstrated an increase in fatigue from pre- to post-adjuvant treatment followed by a decrease over the post-treatment follow-up, consistent with previous research35 (see Figure 1, Panel A). Of note, in the validation study for the MFSI-SF, the mean score on the general fatigue subscale for non-cancer controls was 5.06,33 indicating that fatigue was elevated in our sample across the assessment period. Analyses including the treatment group × time interaction showed that the magnitude of change in fatigue symptoms did not differ significantly across the four adjuvant treatment groups (RT only, CT only, RT + CT, and no RT or CT; Group × Time interaction, p = .32).
Figure 1. Mean levels of fatigue and gene expression over the study period.
There were significant changes in mean levels of fatigue, Type I IFN gene expression, and pro-inflammatory gene expression across the assessment period. Adjusted mean levels of MFSI-general fatigue scores (Panel A) increased from pre-treatment to post-treatment with radiation and/or chemotherapy then declined over the 18-month post-treatment follow-up. Adjusted marginal means for log(2) mRNA abundance of Type I IFN genes (Panel B) also increased from pre- to post-treatment but remained elevated across the follow-up period. Similarly, adjusted marginal means for log(2) mRNA abundance of canonical pro-inflammatory genes (Panel C) increased from pre- to post-treatment and continued to rise over the follow-up period.
Changes in Type I IFN and inflammatory gene expression over time
In RNA sequencing data from 714 PBMC samples obtained from 181 study participants at up to 5 timepoints each, average expression of a pre-specified set of 32 Type I IFN response genes showed a significant effect of time after control for participant age, race, ethnicity, BMI, smoking history, cancer stage, surgery type, type of adjuvant treatment received, and endocrine therapy [F(4, 22E3) = 34.46, p <.0001]. Examination of adjusted marginal means showed an increase in Type I IFN gene expression from pre- to post-treatment that remained elevated over the follow-up (Figure 1, Panel B). Parallel analyses of a pre-specified set of 19 pro-inflammatory genes also showed a significant effect of time after control for covariates [F(4, 13E3) = 28.06, p <.0001), with increased expression from pre-treatment to all follow-up time points (Figure 1, Panel C). Analyses including treatment group × time interactions yielded similar effects of time and also indicated differences in gene expression over time in the four adjuvant therapy groups. Thus, the treatment group × time interaction was included in analyses examining associations between immune gene expression and fatigue.
Association between Type I IFN and inflammatory gene expression and fatigue
The primary goal of this study was to evaluate the association between expression of Type I IFN and inflammatory genes and fatigue over the study period. In mixed effect linear models, expression of a prespecified set of 32 Type I IFN response genes was elevated in association with fatigue (F(1, 22E3) = 23.71, p < .0001), controlling for participant age, race, ethnicity, BMI, smoking history, cancer stage, surgery type, type of adjuvant treatment received (and adjuvant therapy × time interactions), and endocrine therapy. There was no evidence that the magnitude of this association varied by timepoint (p > .10). As seen in Figure 2, there was a positive association between IFN gene expression and fatigue at pre-treatment (.013 ± .004, p = .004) and subsequent time points (post-treatment: .017 ± .005, p < .001; 6-month follow-up: .018 ± .005, p < .0001; 18-month follow-up: .016 ± .005, p < .001), though this was less pronounced (and non-significant) at the 12-month follow-up (.004 ± .005, p = .345).
Figure 2. Association between fatigue and leukocyte gene expression over time.
The association between fatigue and expression of Type I IFN and pro-inflammatory genes from mixed effect linear models is shown at each assessment point. Type I IFN gene expression was significantly positively associated with fatigue at pre-treatment, post-treatment, 6-month post-treatment follow-up, and 18-month post-treatment follow-up, such that women with higher levels of Type I IFN gene expression reported higher levels of fatigue at these assessment points. In contrast, pro-inflammatory gene expression was significantly associated with fatigue later in the cancer trajectory, at 6-month post-treatment follow-up (positive association), 12-month post-treatment follow-up (negative association), and 18-month follow-up (positive association).
Parallel analyses of pro-inflammatory gene expression found no significant association with average fatigue (F(1, 13E3) = 0.99, p = .321). However, this lack of average association across all timepoints masked significant differences in the nature of pro-inflammatory associations across different study timepoints (interaction: F(4,13E3) = 6.29, p < .001). As shown in Figure 2, pro-inflammatory gene expression was not significantly associated with fatigue at pre-treatment baseline (.0007 ± .006, p = .906) or post-treatment follow-up (.004 ± 0.007, p = .601), but did show significant positive associations at 6- and 18-month follow-ups (6-month: .017 ± .007, p = .006; 18-month: .015 ± .007, p = .020). Unexpectedly, pro-inflammatory gene expression showed a marked negative association with fatigue at the 12-month follow-up (with absolute magnitude similar to that seen at the 6- and 18-month follow-ups: −.018 ± .007, p = .007). We also conducted analyses using the MFSI total score in place of the MFSI general fatigue subscale, which yielded comparable results (see Supplementary Analyses).
Ancillary analyses were conducted controlling for variations in mRNA indicators of major leukocyte subsets (CD14, CD3D, CD4, CD8A, CD19, CD56/NCAM1, CD16/FCGR3A) to determine the extent to which these variations could account for the association between fatigue and immune response gene expression. Results continued to show a significant association between Type I IFN gene expression and fatigue (F (1, 22E3) = 14.23, p < 0.001) with no differences across the timepoints. For pro-inflammatory gene expression, controlling for leukocyte composition reduced the time dependent association with fatigue to marginal significance (F (4, 13E3) = 2.11, p = 0.077), suggesting that leukocyte subset abundance contributed to the general association between inflammatory gene expression during the later stages of treatment and survivorship. Analyses examining changes in leukocyte subsets indicated a significant elevation in the relative abundance of monocytes (CD14) beginning at the post-treatment follow-up and persisting thorough the 12-month follow-up. Full results for analyses examining change in leukocyte subsets across the assessment period are provided in Supplementary Analyses.
DISCUSSION
This longitudinal transcriptomic analysis of women with breast cancer was designed to confirm and quantify the role of two key immune signaling pathways that have been nominated in previous exploratory/discovery analyses as molecular mechanisms of cancer-associated fatigue: Type I interferon response genes and canonical pro-inflammatory gene transcripts. Results of this study showed that Type I IFN activity represented the dominant gene regulatory correlate of fatigue before, during, and up to 18 months after adjuvant treatment, with strong associations between fatigue and IFN gene expression across the assessment period. In contrast, canonical pro-inflammatory gene expression was correlated with fatigue only after treatment completion. These findings are generated using a robustly powered longitudinal study that was specifically designed to test previously derived hypotheses regarding the immunoregulatory correlates of fatigue in breast cancer patients, and to control for potential confounding factors (e.g., demographics, body mass index). It is notable that the effects were not confined to patients treated with genotoxic agents (i.e., chemotherapy and/or radiation), and that the association between immune regulatory dynamics and fatigue persisted for at least 18 months after initial treatment. Overall, results revealed unexpected complexity in the immune underpinnings of fatigue and confirm a significant role for IFN signaling as a robust contributor to this symptom during and after treatment.
Previous studies of immune correlates of cancer-related fatigue have primarily focused on inflammation, based on evidence that activation of the proinflammatory cytokine network signals the brain to induce fatigue in experimental and observational studies conducted in non-cancer patients.36 In contrast, few studies have examined the role of Type I interferon signaling in cancer-related fatigue, despite evidence that Type I interferons are associated with fatigue outside of the cancer context24 and their important role in tumor biology.20 Results from the current study suggest that Type I interferon activity may be the primary genomic correlate of fatigue in the early stages of diagnosis and treatment, with proinflammatory gene expression coming into play later in survivorship. Inflammatory gene expression does increase immediately following chemotherapy or radiation treatment alone, but those post-treatment increases are not directly associated with fatigue. Instead, pro-inflammatory gene expression emerges as a predictor of fatigue at 6 months following treatment, and that effect is accounted for in large part by the concurrent increase in monocyte prevalence.
Several mechanisms might account for the time-dependent change in the immunoregulatory correlates of fatigue. Our results suggest a progressive etiological process in which, 1) cancer and its treatments initially trigger Type I IFN activity (e.g., via treatment-induced genotoxicity, reactivation of latent viruses, etc.);37,38 2) those immunological reactions subsequently alter the regulatory processes controlling leukocyte subset abundance (e.g., myelopoiesis and monocyte differentiation, consistent with the observed CD14 upregulation); and 3) shifting leukocyte subset distributions underlie the association between pro-inflammatory gene expression and fatigue observed in post-treatment survivorship. The present data do not link cancer-related fatigue to any inverse association between inflammatory and Type I IFN gene expression that would be characteristic of the Conserved Transcriptional Response to Adversity gene expression profile as mediated by sympathetic nervous system signaling, nor does fatigue track any parallel reduction in inflammatory and Type I IFN activity that would be characteristic of hypothalamus-pituitary-adrenal axis signaling (which represses both of these gene sets in parallel).39 As such, the immunoregulatory correlates of fatigue documented here are more likely initiated by other non-neuroendocrine mechanisms (e.g., genotoxic effects of treatment, chronic viral reactivation, microbial alterations, etc.).
One complexity in these findings emerged at the 12-month post-treatment assessment, when we observed marked deviations from the overall pattern of results. These included an attenuated association between IFN gene expression and fatigue and an inverse association between canonical pro-inflammatory gene expression and fatigue (i.e., greater fatigue among those with lower pro-inflammatory gene expression). It is possible that these changes were due to the unanticipated elevation in overall pro-inflammatory gene expression observed at this time point. The time-dependent nature of these results highlights the importance of longitudinal studies that follow patients from diagnosis through treatment and into survivorship to capture the dynamic changes in immunity that occur across this period and may differentially contribute to fatigue.
Limitations of the current study include the relatively homogeneous population, which was comprised primarily of non-Hispanic White women with early-stage breast cancer. Although our analyses did control for age, race, ethnicity, body mass index, and key health behaviors (i.e., smoking), these factors are known to influence innate immune activity34 and may structure the association between immune cell gene expression and fatigue in cancer patients and survivors. In addition, given the nature of the study design, any causal effect of immune activity on fatigue cannot be determined here. Pharmacologic interventions to alter Type I interferon and/or canonical proinflammatory signaling pathways represent a logical next step in this area of research, as do studies more intensively mapping the changing immune cell distributions (particularly classical and non-classical monocyte dynamics). This analysis was designed specifically to test previously derived hypotheses regarding the role of canonical inflammatory signaling and Type I interferon signaling in fatigue dynamics over the course of therapy and post-treatment survivorship. Other biological pathways may well contribute to cancer-related fatigue, and those additional pathways remain to be identified in future exploratory/discovery analyses.
The identification of Type I interferon activity as the predominant correlate of fatigue early in the cancer trajectory not only clarifies the immunopathogenesis of cancer-related fatigue, but also suggests new etiological hypotheses for future research. For example, reactivation of latent viruses may be relevant for cancer-related fatigue.40 Treatment-induced genotoxicity may also represent a key trigger for interferon-mediated signaling to the brain to generate fatigue symptomatology.41 Of note, gene expression profiling may be required to interrogate activity in this system because Type I interferons are difficult to detect in circulation due to their generally low concentrations and high potency. Understanding the biology of fatigue is essential for targeted therapies, and ultimately for enhancing quality of life in the growing number of cancer survivors that are afflicted by this symptom.
Supplementary Material
Financial support:
This work was supported by National Institutes of Health/National Cancer Institute R01 CA160427 and by the Breast Cancer Research Foundation. Dr. Crespi is funded by the National Institutes of Health/National Cancer Institute P30 CA016042.
REFERENCES
- 1.Curt GA, Breitbart W, Cella D, et al. : Impact of cancer-related fatigue on the lives of patients: new findings from the Fatigue Coalition. Oncologist. 5:353–360, 2000 [DOI] [PubMed] [Google Scholar]
- 2.Bower JE: The role of neuro-immune interactions in cancer-related fatigue: Biobehavioral risk factors and mechanisms. Cancer 125:353–364, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Dantzer R, Heijnen CJ, Kavelaars A, et al. : The neuroimmune basis of fatigue. Trends Neurosci 37:39–46, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cleeland CS, Bennett GJ, Dantzer R, et al. : Are the symptoms of cancer and cancer treatment due to a shared biologic mechanism? A cytokine-immunologic model of cancer symptoms. Cancer 97:2919–2925, 2003 [DOI] [PubMed] [Google Scholar]
- 5.Bower JE: Cancer-related fatigue: links with inflammation in cancer patients and survivors. Brain Behav.Immun 21:863–871, 2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Miller AH, Ancoli-Israel S, Bower JE, et al. : Neuroendocrine-immune mechanisms of behavioral comorbidities in patients with cancer. J Clin Oncol 26:971–982, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Saligan LN, Kim HS: A systematic review of the association between immunogenomic markers and cancer-related fatigue. Brain Behav Immun 26:830–48, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang XS, Shi Q, Williams LA, et al. : Inflammatory cytokines are associated with the development of symptom burden in patients with NSCLC undergoing concurrent chemoradiation therapy. Brain Behav.Immun 24:968–974, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang XS, Williams LA, Krishnan S, et al. : Serum sTNF-R1, IL-6, and the development of fatigue in patients with gastrointestinal cancer undergoing chemoradiation therapy. Brain Behav.Immun, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Xiao C, Beitler JJ, Higgins KA, et al. : Fatigue is associated with inflammation in patients with head and neck cancer before and after intensity-modulated radiation therapy. Brain Behav Immun 52:145–152, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Shi W, Misra S, Li M, et al. : Inflammatory Biomarkers, Hematopoietic Stem Cells, and Symptoms in Breast Cancer Patients Undergoing Adjuvant Radiation Therapy. JNCI cancer spectrum 4:pkaa037–pkaa037, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bower JE, Ganz PA, Irwin MR, et al. : Fatigue and gene expression in human leukocytes: increased NF-kappaB and decreased glucocorticoid signaling in breast cancer survivors with persistent fatigue. Brain Behav.Immun 25:147–150, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Black DS, Cole SW, Christodoulou G, et al. : Genomic mechanisms of fatigue in survivors of colorectal cancer. Cancer, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Xiao C, Beitler JJ, Higgins KA, et al. : Associations among human papillomavirus, inflammation, and fatigue in patients with head and neck cancer. Cancer 124:3163–3170, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Xiao C, Beitler JJ, Higgins KA, et al. : Differential regulation of NF-kB and IRF target genes as they relate to fatigue in patients with head and neck cancer. Brain Behav Immun 74:291–295, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Capuron L, Gumnick JF, Musselman DL, et al. : Neurobehavioral effects of interferon-α in cancer patients: Phenomenology and paroxetine responsiveness of symptom dimensions. Neuropsychopharmacology 26:643–652, 2002 [DOI] [PubMed] [Google Scholar]
- 17.Ivashkiv LB, Donlin LT: Regulation of type I interferon responses. Nat Rev Immunol 14:36–49, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Burnette BC, Liang H, Lee Y, et al. : The efficacy of radiotherapy relies upon induction of type i interferon-dependent innate and adaptive immunity. Cancer Res 71:2488–96, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Moschella F, Torelli GF, Valentini M, et al. : Cyclophosphamide induces a type I interferon-associated sterile inflammatory response signature in cancer patients’ blood cells: implications for cancer chemoimmunotherapy. Clin Cancer Res 19:4249–61, 2013 [DOI] [PubMed] [Google Scholar]
- 20.Zitvogel L, Galluzzi L, Kepp O, et al. : Type I interferons in anticancer immunity. Nat Rev Immunol 15:405–14, 2015 [DOI] [PubMed] [Google Scholar]
- 21.Raison CL, Demetrashvili M, Capuron L, et al. : Neuropsychiatric adverse effects of interferon-alpha: recognition and management. CNS Drugs 19:105–23, 2005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Capuron L, Miller AH: Cytokines and psychopathology: lessons from interferon-alpha. Biol.Psychiatry 56:819–824, 2004 [DOI] [PubMed] [Google Scholar]
- 23.Capuron L, Gumnick JF, Musselman DL, et al. : Neurobehavioral effects of interferon-alpha in cancer patients: phenomenology and paroxetine responsiveness of symptom dimensions. Neuropsychopharmacology 26:643–652, 2002 [DOI] [PubMed] [Google Scholar]
- 24.Dowell NG, Cooper EA, Tibble J, et al. : Acute Changes in Striatal Microstructure Predict the Development of Interferon-Alpha Induced Fatigue. Biol Psychiatry 79:320–8, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Russell A, Hepgul N, Nikkheslat N, et al. : Persistent fatigue induced by interferon-alpha: a novel, inflammation-based, proxy model of chronic fatigue syndrome. Psychoneuroendocrinology 100:276–285, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bower JE, Ganz PA, Desmond K, et al. : Fatigue in long-term breast cancer survivors: A longitudinal investigation. Cancer, 2005 [DOI] [PubMed] [Google Scholar]
- 27.Raison CL, Borisov AS, Majer M, et al. : Activation of Central Nervous System Inflammatory Pathways by Interferon-Alpha: Relationship to Monoamines and Depression. Biological Psychiatry 65:296–303, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Capuron L, Raison CL, Musselman DL, et al. : Association of exaggerated HPA axis response to the initial injection of interferon-alpha with development of depression during interferon-alpha therapy. Am.J.Psychiatry 160:1342–1345, 2003 [DOI] [PubMed] [Google Scholar]
- 29.Bower JE, Asher A, Garet D, et al. : Testing a biobehavioral model of fatigue before adjuvant therapy in women with breast cancer. Cancer 125:633–641, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Bower JE, Ganz PA, Irwin MR, et al. : Do all patients with cancer experience fatigue? A longitudinal study of fatigue trajectories in women with breast cancer. Cancer 127:1334–1344, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bower JE, Ganz PA, Irwin MR, et al. : Acute and Chronic Effects of Adjuvant Therapy on Inflammatory Markers in Breast Cancer Patients. JNCI Cancer Spectr 6, 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Stein KD, Jacobsen PB, Blanchard CM, et al. : Further validation of the multidimensional fatigue symptom inventory-short form. J Pain Symptom Manage 27:14–23, 2004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Stein KD, Martin SC, Hann DM, et al. : A multidimensional measure of fatigue for use with cancer patients. Cancer Pract. 6:143–152, 1998 [DOI] [PubMed] [Google Scholar]
- 34.Cole SW, Shanahan MJ, Gaydosh L, et al. : Population-based RNA profiling in Add Health finds social disparities in inflammatory and antiviral gene regulation to emerge by young adulthood. Proc Natl Acad Sci U S A 117:4601–4608, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Jacobsen PB, Hann DM, Azzarello LM, et al. : Fatigue in women receiving adjuvant chemotherapy for breast cancer: characteristics, course, and correlates. J.Pain Symptom.Manage 18:233–242, 1999 [DOI] [PubMed] [Google Scholar]
- 36.Bower JE, Kuhlman KR: Psychoneuroimmunology: An Introduction to Immune-to-Brain Communication and Its Implications for Clinical Psychology. Annu Rev Clin Psychol 19:331–359, 2023 [DOI] [PubMed] [Google Scholar]
- 37.Hsiao CP, Araneta M, Wang XM, et al. : The association of IFI27 expression and fatigue intensification during localized radiation therapy: implication of a para-inflammatory bystander response. Int J Mol Sci 14:16943–57, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Felger JC, Cole SW, Pace TW, et al. : Molecular signatures of peripheral blood mononuclear cells during chronic interferon-α treatment: relationship with depression and fatigue. Psychol Med 42:1591–603, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Irwin MR, Cole SW: Reciprocal regulation of the neural and innate immune systems. Nat Rev.Immunol 11:625–632, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Fagundes CP, Glaser R, Alfano CM, et al. : Fatigue and herpesvirus latency in women newly diagnosed with breast cancer. Brain Behav.Immun 26:394–400, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Saligan LN, Hsiao CP, Wang D, et al. : Upregulation of alpha-synuclein during localized radiation therapy signals the association of cancer-related fatigue with the activation of inflammatory and neuroprotective pathways. Brain Behav.Immun 27:63–70, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.