Cancer treatments can produce debilitating symptoms that have a profound impact on patients’ quality of life. The most common, severe, and consistently reported side effect of chemotherapy is fatigue (O’Regan et al., 2019), which is defined as a sense of tiredness or exhaustion that is not proportional to activity nor sleep. No approved therapies for cancer-related fatigue (CRF) have been developed, in part because the etiology explaining the development of the symptom is poorly understood.

Brain-derived neurotrophic factor (BDNF) is an important signaling molecule in the central nervous system known for its role in neuronal growth and plasticity. The Val66Met single nucleotide polymorphism (SNP) of BDNF involves the substitution of a methionine for a valine residue in the prodomain of BDNF, and it affects the trafficking and release of BDNF (Egan et al., 2003). The prevalence of the met allele varies widely across populations, with estimates ranging from 0.5% in Sub-Saharan Africa to 43.6% in Asia (Petryshen et al., 2010). The polymorphism is most widely known for its association with higher rates of depression and mood disorders (Notaras et al., 2015; Youssef et al., 2018), but recent studies have found a protective effect of the Val66Met polymorphism on cancer-related fatigue. Specifically, one study found a relationship between the Val66Met polymorphism and fatigue in prostate cancer patients after undergoing radiation therapy (Feng et al., 2020) and other studies found a relationship between the Val66Met polymorphism and chemotherapy-related cognitive impairment in patients undergoing chemotherapy treatments for breast cancer (Ng et al., 2015; Tan et al., 2019).

A transgenic mouse model expressing the human Val66Met polymorphism of BDNF was developed and published in 2006 (Chen et al., 2006). Mice with the Val66Met polymorphism show a number of differences from wild-type mice, including increased food intake and increased bodyweight, increased anxiety and depression-like behavior, and a blunted response to antidepressant drugs like fluoxetine (Chen et al., 2006; Bath et al., 2012a,b; Liu et al., 2012; Yu et al., 2012; Ieraci et al., 2020).

Though the Val66Met polymorphism appears related to fatigue symptoms based on clinical observational data, animal models allow for more controlled study of this phenomenon. In this study, we used a mouse model to test whether mice with the Val66Met polymorphism are protected from fatigue-like effects of the commonly used chemotherapy drug, 5-fluororuracil (5FU), and assess whether age and sex played a role in modulating this effect. Similar to our previously published mouse model of radiation-induced fatigue (Wolff et al., 2017), mice in this study received three daily injections of 50 mg/kg 5FU, and we measured changes in voluntary wheel running activity (VWRA) as a fatigue-like outcome measure.

Our main finding is that after treatment with 5FU, the fatigue-like phenotype is less prominent in the Val66Met mice than in the WT controls. This reproduces observations in clinical cancer patients, where the Met allele seems to protect against CRF (Feng et al., 2020). Reproducing this in an animal model is an important step, because correlational findings in a clinical population can be difficult to interpret. Importantly, the animal model advances our understanding of the role of Val66Met in CRF by allowing us to further tease out contributions of variables such as age and sex in a controlled environment, which can be difficult to achieve in clinical studies. The Val66Met polymorphism is associated with other symptoms in humans, including mood disorders like depression (Notaras et al., 2015; Youssef et al., 2018). It is possible that other symptoms may have an impact on self-reported CRF, rather than BDNF being a direct contributor. This is an especially important consideration, as fatigue is considered as part of a network of complex symptom clusters including mood disorders that can be difficult to distinguish in clinical practice (Donovan and Jacobsen, 2007). Further, the prevalence of the Val66Met polymorphism is highly variable across different populations, most notably race, but also subpopulations within racial categories (Petryshen et al., 2010); other differences between these subpopulations, genetic or otherwise, could result in an apparent relationship of the Val66Met SNP with fatigue. The significance of our study is that it provides evidence for a more direct role of the Val66Met SNP on fatigue and fatigue-like behavior. In our mouse model, we compared the behavior of genetically similar mice within an inbred colony, where genetics and other factors like diet, light exposure, exercise, etc. are more controlled. The finding that the Val66Met mice show less fatigue-like behavior suggests that BDNF itself may be mechanistically involved in CRF warrants further study and could point to new targets for intervention.

Interestingly, the 2-week drop in VWRA after 5FU injections appeared to have two phases. The first phase was less severe in mice with the Val66Met polymorphism, but otherwise was consistent between young and middle-aged mice. The second phase was unaffected by the polymorphism but less severe in young mice. One interpretation of this is that the severity of the fatigue-like behavior is affected by the Val66Met polymorphism, but the speed of recovery gets slower with increased age. However, clinical evidence tends to show the opposite, that fatigue is if anything less severe in older patients who receive chemotherapy (Winters-Stone et al., 2008; Bischel et al., 2016). This may be because fatigue is less bothersome in older patients and therefore reported as less severe, but the relationship between age and fatigue needs more study to be properly understood.

Our 5FU-induced fatigue model used three daily doses of 50 mg/kg 5FU in mice on a C57BL/6N background to induce fatigue-like behavior measured by VWRA. Importantly, there appeared to be a full recovery from this dose during the third post-injection week, suggesting this dose did not produce any long-term injury that would affect physical activity. This dose is significantly lower than that used in another study (Mahoney et al., 2013) that used five injections of 60 mg/kg in C57BL/6 mice to produce a very similar outcome. We speculate that fatigue-like outcomes in mice may have a substantial dependence on mouse substrain, much like what we saw in a previous study with a radiation-induced fatigue model, where C57BL/6J mice showed considerably less fatigue-like behavior than C57BL/6N mice (Wolff et al., 2021). Further supporting this is a study by Dougherty et al. (2019) that used the same dosing paradigm as Mahoney et al. (2013) with C57BL/6N mice, but Dougherty et al. (2019) saw a much more severe fatigue-like behavior with a nearly complete cessation of VWRA.

Our baseline VWRA and body weight results match well with what exists in the scientific literature. Earlier studies found that older mice run less than younger mice (Turner et al., 2005), that Val66Met mice are larger than their WT littermates (Chen et al., 2006), and that there are no difference in VWRA between Val66Met mice and WT (Ieraci et al., 2016). We found the same. However, a novel finding in our study is that in the older mice, the Val66Met mice lost more weight than WT controls during VWRA behavior. The reason for the increased weight loss is unclear. VWRA is commonly used experimentally to study the effects of exercise, and it is well-established that exercise induces increased expression of BDNF (Youssef et al., 2018). An earlier study found that Val66Met mice did not show exercise-related benefits in novelty-suppressed feeding and in the forced swim test, and Val66Met mice had lower levels of hippocampal BDNF expression after exercise (Ieraci et al., 2016). In addition, a recent study showed that there are differences in white adipose tissue in Val66Met mice both before and after exercise (Sandrini et al., 2019). These studies demonstrate a clear effect of the BDNF genotype on exercise-related biological changes, so it is possible exercise-induced weight loss in our study is affected in much the same way. However, it is also possible that the increased weight loss is simply due to their increased baseline bodyweight; there is simply more weight to lose. Our linear regression model did not report baseline bodyweight to have a significant effect on weight loss, but that might be due to a non-linear relationship, insensitivity of the model, and/or high variability in the data.

We did not see differences in open field behavior in mice treated with 5FU, which matches prior studies using 5FU (Dougherty et al., 2019) or doxycycline (Zombeck et al., 2013). However, prior studies comparing Val66Met mice to WT have found an anxiety-like phenotype (Chen et al., 2006; Bath et al., 2012a) that we did not see. In fact, the trend we saw was in the opposite direction, where Val66Met mice were spending more time in the center of the arena. We think this discrepancy may be due to a relative lack of stress, as Yu et al. reported effects of Val66Met on open field center time only in stressed mice (Yu et al., 2012). We did not use a stress model in our study; while mice are individually housed and received three injections, both of which are presumably stressful, there is substantial evidence that VWRA alleviates effects of stress on mice (Huang et al., 2017; Pagliusi et al., 2020).

In conclusion, we found that the fatigue-like decline in VWRA after receiving 5FU injections is smaller in mice with the Val66Met polymorphism of BDNF. This may indicate promising new directions for studying both the mechanisms behind cancer-related fatigue and possible clinical interventions.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: 10.17605/OSF.IO/GQU4N; https://osf.io/gqu4n/?view_only=f5c01c7bebfb47ba946f5fa1ab7036af.

The animal study was reviewed and approved by the National Heart Lung and Blood Institute (NHLBI) Animal Care and Use Committee.

BW: conceptualization, methodology, formal analysis, supervision, and writing—original draft. HA: investigation, methodology, and writing—reviewing and editing. LF: conceptualization, and writing—reviewing and editing. LS: resources, conceptualization, project administration, and writing—reviewing and editing. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.