With an estimated 297,790 new cases in 2023, breast cancer is the most commonly diagnosed cancer worldwide and in women to date (Arnold et al., 2022). In the United States, the 5-year survival rate of breast cancer is 91%; however, in low- and middle-income countries (LMICs), as defined by the World Bank based on a country’s gross national income bracket (Lencucha and Neupane, 2022), this rate is significantly lower. Countries such as Samoa, Gambia, Fiji, and Brazil have survivorship rates ranging from as low as 25% to 60% (Rivera-Franco and Leon-Rodriguez, 2018; Lei et al., 2021). Nevertheless, the implementation of diagnostics, adjuvant chemotherapy, and further technological advancements have propagated the survival rates of breast cancer globally.

Patients with cancer frequently report cognitive problems that can affect their quality of life. Studies have shown that patients experience cognitive dysfunction during and up to 20 years after treatment (Koppelmans et al., 2012). As breast cancer survivorship rates increase, the need to investigate the short- and long-term effects of breast cancer chemotherapy has become increasingly more prevalent. Cognitive impairments are one of the most common side effects of chemotherapy (Henderson et al., 2019). Also referred to as “chemobrain,” chemotherapy-induced cognitive impairments pose significant functionality challenges for breast cancer survivors. Patients experiencing chemobrain report deficits in memory retention, learning, executive function, and processing speed (Yang et al., 2010; Koppelmans et al., 2012; Brown et al., 2021).

The first efficacious chemotherapy regimen for breast cancer was developed in the 1970s: a polytherapy combining cyclophosphamide, methotrexate, and 5-fluorouracil (CMF). Cyclophosphamide (CYP) is a cytotoxic, alkylating antineoplastic drug that is commonly used as an immunosuppressive (Yang et al., 2010). Synthesized in 1958, CYP was the eighth cytotoxic anticancer drug approved by the Food and Drug Administration (Li et al., 2022). The alkylating nature of CYP has been found to create alkyl crosslinks between DNA, inducing apoptosis (Matalon et al., 2004). CYP has been shown to permeate the blood–brain barrier (BBB), although the underlying mechanisms remain understudied [11]. Methotrexate (MTX), developed in 1949, is an antifolate that inhibits antioxidant activity by binding to dihydrofolate reductase (Skubisz and Tong, 2012; Taran et al., 2023). This consequently hinders the amelioration of oxidative stress (Hess and Khasawneh, 2015), leading to high levels of cytotoxicity in clinical applications (Sekeres et al., 2021). Only at high doses (i.e., 1–3 g/m²) has MTX been shown to cross the BBB (Angelov et al., 2009; Prodduturi and Bierman, 2012). 5-Fluorouracil (5-FU) is an anti-metabolite that interferes with thymidine synthesis and DNA replication. Discovered in 1957 (Heidelberger et al., 1957), 5-FU is part of the first class of thymidylate synthase inhibitors used in clinical application (Chu et al., 2003). Designed as an analog of RNA nucleotide base uracil, 5-FU inhibits thymidylate synthase, which consequently disrupts proper DNA formation and instigates cytotoxicity (Wigmore et al., 2010). Similar to CYP, 5-FU has also been shown to penetrate the BBB via passive diffusion (Wigmore et al., 2010). The combination of CYP, MTX, and 5-FU modeled in 1973 by Bonadonna et al (Bonadonna et al., 1995). created a regimen that was instrumental for implementing chemotherapeutic treatments and developing safer surgical procedures for patients with breast cancer (Verrill, 2009).

As the use of CMF chemotherapy to treat breast cancer became more pervasive, the need to study its long-term effects became more pressing. Wieneke and Dienst found that women treated with adjuvant chemotherapy, primarily CMF, displayed mild cognitive impairment compared to test norms (Wieneke and Dienst, 1995). Later work by Schagen et al. revealed that patients treated with CMF and tamoxifen reported significant impairments in concentration, memory, and locomotion compared to control (Schagen et al., 1999). During this time, however, the use of taxanes in combination with anthracyclines (ACs) became the standard for breast cancer polytherapy in the United States (Brown et al., 2021). The introduction of AC regimens denoted a shift away from using CMF moving into the 21st century.

In low- and middle-income countries (LMICs), access to chemotherapeutic treatments for breast cancer remains limited (Sandelin et al., 2002). Newer taxane and AC agents are not widely available in these countries; therefore, the use of CMF remains highly prevalent globally (Tfayli et al., 2010). Furthermore, most chemobrain studies involving patients are conducted in the United States, Europe, Canada, and Australia—all high-income countries (Ribi, 2012). LMICs, particularly those in Africa, such as Uganda, South Africa, and Tanzania, still use the CMF to treat breast cancer (Serra et al., 2020; Menon et al., 2021; Keetile et al., 2023). The urgency to understand the cytotoxic effects of this regimen persists. A study in Gauteng, South Africa conducted in 2021 found that 33% of Black African women (n=10) treated with CMF reported significant cognitive impairment (Keetile et al., 2021). Another South African study in 2023 revealed that 77.4% of Black African women (n=53) diagnosed with stage II or III breast cancer treated with CMF displayed significant cognitive impairment (Keetile et al., 2023). These studies demonstrate the necessity of continual investigation of how CMF dysregulates brain functionality and how to mitigate the physiological symptoms harming patients with breast cancer and survivors.

Chemotherapy is also known to affect the microbiome–gut–brain (MGB) axis. The gut microbiota–immune–brain axis is becoming increasingly important for the treatment of a myriad of neurological and cognitive disorders (O’Riordan et al., 2025). Studies have found that chemotherapy alters microbiota populations, activates neuroimmune responses, and promotes glial dysregulation (Song and Bai, 2021; Rowaiye et al., 2024). For example, it is well known that the gut microbiome can enhance or mitigate the effectiveness of therapeutic treatments of patients with cancer (Viaud et al., 2013; Alexander et al., 2017). For example, a variety of microbial taxa are important mediators of immune checkpoint inhibitors, while others promote increased toxicity of anticancer drugs, as reviewed by Huange et al (Huang et al., 2022). Furthermore, several studies have shown various levels of responses by the gut microbiota in patients with breast cancer (Wu et al., 2020; Aarnoutse et al., 2022). While the underlying connection between the MGB and chemobrain is poorly understood, some studies have suggested that gut microbiota assists in regulating hippocampal-dependent cognition (Kuijer and Steenbergen, 2023). Furthermore, it is also possible that the MGB contributes to the development of chemobrain, overall fatigue, and other chemotherapy-induced symptoms (Grant et al., 2021; Xiao et al., 2021). For example, MTX influences MGB composition and can engender both gut dysbiosis (Chen et al., 2024; Rowaiye et al., 2024) and proteomic alterations (Letertre et al., 2020).

This study aims to investigate how combination CMF chemotherapy affects hippocampal-dependent behavior and gut-microbiome diversity. This work may provide further insight into possible underlying mechanisms of long-term deficits in hippocampal-dependent functioning due to CMF and the role of the MGB in chemobrain.

Our prior work has found that 2 weeks post-chemotherapy, 6-month-old female mice treated with CMF showed significant long-term memory impairments in the Morris water maze and a decrease in hippocampal arborization (Anderson et al., 2020). We have also shown that 24 h post-chemotherapy, 12-week-old female mice treated with CMF displayed depressive-like behaviors with deficits in social novelty in the Forced Swim and Three Chamber Sociability tests, respectively (Corley et al., 2023). Both studies use the same drug doses and injection paradigm as presented herein. This study was designed to establish if cognitive impairments still occur 30 days post-CMF. We tested spatial memory using the Morris water maze task and found that CMF-treated mice exhibited an inability to retain memory in the probe trials. Our proteomics analysis identified the protein Attractin-like 1 (ATRNL-1), which was found to have an undefined connection with three proteins related to cognitive impairments, spatial learning, and memory deficits. When probing for immediate-early genes (IEGs) via Western blots, a significant decrease in c-Fos, Zif268, and Arc expression was observed in the CMF treatment group. Lastly, there were no notable changes in gut morphology between treatment groups, mirroring the lack of detectable differences in microbial diversity between the treatment groups.

The Morris water maze is a spatial learning task that is hippocampal-dependent (Morris et al., 1982). The hippocampus has been vastly studied and shown to be critical for long-term episodic memory (Bird and Burgess, 2008; Lisman et al., 2017). Further research has shown that the anatomical formation of the hippocampus supports its role in mediating spatial cognition and memory (Chiba et al., 1994; Fortin et al., 2002; Bird and Burgess, 2008; Hartley et al., 2014; Lisman et al., 2017). Both CMF- and saline-treated groups demonstrated the ability to learn during the Morris water maze. Notably, however, the saline group on day 1 swam significantly faster to the platform than the CMF group measured via mean velocity. Days 3–5 of testing revealed that CMF-treated mice were unable to distinguish between the target and non-target quadrant, conveying a lack of memory retention. This observation follows similar findings to that of several studies. A study conducted by Kinra et al. reported that mice administered CMF once a week for 3 weeks displayed significantly lower retention time and higher latency to platform in the Morris water maze (Kinra et al., 2021). Anderson et al. showed in the Morris water maze that CMF-treated mice could not distinguish between the target and non-target quadrants on days 3 and 4 (Anderson et al., 2020). As monotherapies or doublets, CYP, MTX, and 5-FU have been shown to affect spatial cognition as well. Mishra et al. found that mice treated with a single dose of CYP were highly immobile and had higher latency to platform during the Morris water maze compared to control mice (Mishra et al., 2022). Mice given MTX (250 mg/kg) showed longer latency time to the platform during probe trials compared to control mice, despite showing the ability to learn (Seigers et al., 2008). A combination of 5-FU and MTX was administered to 2-month-old female BALB/C mice and were then tested in the water maze 1 week after treatment. 5-FU+MTX-treated mice were more latent to the platform and made more errors in an attempt to find the target than control mice on days 1 and 2, showing that learning was initially impaired (Winocur et al., 2006). Interestingly, on day 5, our study shows that CMF-treated mice indicated improved memory retention by spending significantly more time in the target quadrant than in non-target quadrants than on previous days.

Using QIAGEN Ingenuity Pathway Analysis (IPA), a network of proteins was rendered to depict the relationship between dysregulated proteins found within our study. The network compared the CMF group to the saline group. ATRNL 1 is a transmembrane protein that is thought to be integral to membrane integrity (Walker et al., 2007) and was the only protein predicted to be downregulated. This protein was shown to have an ambiguous relationship with proteins c-Fos, TAFA2, and HNRNPU. c-Fos, a proto-oncogene and one of the most extensively studied IEGs (Herrera and Robertson, 1996), encodes a protein rapidly upregulated in response to membrane depolarization and voltage-gated calcium influx, enabling neural action potentials. Beyond its role as a marker of neuronal activation, c-Fos has been implicated in learning, memory consolidation, and synaptic plasticity, particularly in spatial memory circuits. For example, female rats trained in acrobatic conditioning displayed a significant increase in c-Fos-positive cells within the motor cortex compared to control and inactive animals, underscoring its role in motor learning (Kleim et al., 1996). Méndez-Couz et al. found an increase in c-Fos protein expression in the amygdala following spatial memory extinction from the Morris water maze (Mendez-Couz et al., 2014). c-Fos knockout mice have been shown to exhibit impairment in hippocampal-dependent spatial memory and long-term potentiation (LTP). Fleischmann et al. showed that c-Fos knockout mice could not discriminate between target and non-target quadrants during the probe trial of the Morris water maze. Moreover, there was a decrease in LTP in hippocampal synapses in the CA3–CA1 subregions (Fleischmann et al., 2003).

HNRNPU encodes for the heterogeneous nuclear ribonucleoprotein U and is critical for RNA splicing as well as chromatin organization (Sapir et al., 2022). HNRNPU has been found to be important for development. For example, an HNRNPU mutation within mice resulted in embryonic growth retardation and early mortality (Roshon and Ruley, 2005). Sapir et al. investigated the role of HNRNPU in brain development. Their work revealed that deletions to HNRNPU resulted in the splicing of genes involved in neuronal survival and synaptic formation. This study also supported the findings of Roshon and Ruley, discovering that the developing mouse brain expresses high levels of HNRNPU in the cortical plate and mitotic cells (Sapir et al., 2022). TAFA2 is a central nervous system (CNS)-specific cytokine that aids in CNS regulation (Liang et al., 2023). Wang et al. found that TAFA2 knockout mice showed increased anxiety-like behaviors in the open field test and elevated plus maze. The study also used the Morris water maze and novel object recognition paradigm to assess spatial, short- and long-term memory—finding significant deficits in each (Wang et al., 2018). Another study has suggested that TAFA2 is vital for neuronal cell survival via binding to the ADGRL1 protein to activate the cAMP-PKA-CREB-BCL2 signaling pathway and prevent apoptosis (Liang et al., 2023). Consequently, we thought it important to probe similar IEGs in hopes of deepening our mechanistic understanding of the behavioral results from the hippocampal-dependent, spatial learning task of the Morris water maze on a neuronal level. Arc and Zif268 were ideal candidates to do so.

Brain IEGs are a class of genes that are rapidly and transiently activated by neuronal activity (Abraham et al., 1991). IEGs can be further divided into two functional classes: regulatory or effector. Regulatory IEGs encode proteins that indirectly affect neural physiology by increasing or decreasing “downstream” gene expression. Conversely, effector IEGs encode proteins that have a more defined role at the synapse (Guzowski et al., 2001; Davis et al., 2003). Both classes of IEGs are necessary in supporting the mechanisms that underlie neuronal plasticity such as LTP, kindling, cellular regeneration, and learning. In particular, Arc is an effector IEG. Arc, activity-regulated cytoskeleton-associated protein, is thought to be essential for memory consolidation and contributes to synaptic plasticity via α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic receptor (AMPAR) regulation. AMPARs are dynamic ionotropic glutamate receptors that mediate most of the excitatory synaptic transmission in the brain (Diering and Huganir, 2018). As an effector IEG, the mRNA of Arc can be localized to post-synaptic dendrites via cytoplasm permeation after action potential prompted by behavior (Minatohara et al., 2015). It is a plasticity protein. Zif268, like c-Fos, is a regulatory IEG. Part of the early growth response (Egr) family, Zif268 is a transcription factor that controls gene expression—aiding in cell development and function (Veyrac et al., 2014). Furthermore, it is believed that the Egr genes serve to connect neurotransmitter action with altered gene expression. Zif268 is highly expressed in the hippocampus, particularly in the CA1 region. Robert et al. found that there was a significant increase in Zif268 levels in the hippocampal CA1 following the induction of LTP in rats compared to control animals (Roberts et al., 1996). In another study, rats were conditioned using the Avoidance Shuttle Box training paradigm. An increase in Zif268 mRNA was subsequently observed (Nikolaev et al., 1992). A couple of studies have presented congruent results. Briones and Woods showed that CMF had deleterious effects on hippocampal cell proliferation as well as spatial and long-term memory in the Morris water maze task (Briones and Woods, 2011). Anderson et al. observed long-term memory deficits in CMF-treated mice. Further investigation revealed that CMF reduced hippocampal dentate gyrus dendritic length and mushroom spines. Importantly, mushroom spines are considered “memory” spines (Bourne and Harris, 2007). These studies provide context to our findings. Our results are consistent with prior and current research, supporting the notion that CMF deteriorates cognition as well as learning and memory function on the neural level.

Emerging research over the last decade has identified the microbiome as a contributing factor to outcomes for neurological changes (Luczynski et al., 2016; Jordan et al., 2018; Grant et al., 2021). Gastrointestinal mucositis (GIM), an inflammation of the intestinal mucosa, is an adverse side effect that can be induced by chemotherapy. GIM has been shown to decrease quality of life, decrease survivorship, and increase the commodity rate of patients with cancer (Sougiannis et al., 2021). Women are at higher risk for developing GIM, as well as patients treated with 5-FU mono- or polytherapy (Basile et al., 2019). MTX can also induce GIM (Higuchi et al., 2020). In fact, one of the most common adverse effects of MTX is gastrointestinal toxicity (Zhou et al., 2018). To further characterize CMF chemotherapy, we evaluated the alpha- and beta-diversity of the microbiome 30 days post-treatment. Differences in alpha-diversity between the CMF and saline groups were not observed ( Figures 7 , 8 ). We analyzed beta-diversity quantitatively (Bray–Curtis and Weighted UniFrac) and qualitatively (Jaccard and Unweighted UniFrac) (Lozupone et al., 2007). Quantitative analyses revealed no distinct separation or clustering between CMF and the saline treatment group, meaning that the two treatment groups share many of the same phylogenetically related and abundant taxa. Qualitative beta-diversity analysis revealed that the two treatment groups were significantly different, indicating that the microbial communities of these two treatment groups still share many low abundant taxa after 1 month. The reduction of altered microbial communities presented here after 1 month is in stark contrast to that observed after 24 h, in which the CMF and saline-treated mice differed significantly in both alpha- and beta-diversity (Corley et al., 2023). These vastly different results may suggest that the microbiome, in the absence of CMF, displayed resilience and was able to return to a presumed “near-healthy” state after 1 month. Notably, the study at present only evaluated microbiome alpha- and beta-diversity 30 days following chemotherapeutic treatment. A limitation of this study is the absence of reagent blanks and Zymo/ATCC mock communities as positive controls, which may limit the ability to distinguish true biological signals from potential contamination or sequencing artifacts. Future studies will address this by including an aliquot of mock community DNA and either a reagent solution blank or PCR template blank as a negative control for sequencing, with all controls processed in triplicate.

Comparable studies have used shorter timelines—typically ranging from 1 to 2 weeks post-treatment. Lu et al. showed that following five daily doses of 5-FU, 5-FU GIM-induced mice had lower ACE, Chao, and Shannon indices than control (Lu et al., 2022). A similar result was presented by Li et al., such that rats injected with 5-FU for 3 days displayed a decrease in alpha- and beta-diversity as measured by Unweighted UniFrac principal coordinates analysis (PCoA) as well as Chao and Shannon indices (Li et al., 2017). Another study using MTX found that rats treated with MTX every 3 days for 7 or 14 days showed a significant decrease in alpha-diversity compared to control (Zhou et al., 2018). Chao1, ACE, Observed species, and Shannon indices were used for analysis. Recently, Chen et al. showed rats that initially received MTX via intrathecal injection and then via intraperitoneal injection once a week for 2 weeks had a significant decrease in beta gut microbiota diversity as measured by Unweighted UniFrac (Chen et al., 2024). Interesting, Chao1 and Shannon indices did not reveal any significant difference in alpha-diversity. A study conducted by Chen, Sun et al. administered 80 mg/kg of CYP to saline-treated mice after 10 days. After receiving CYP on days 10, 12, and 14, animals were sacrificed and tissues were collected; Chao1 and ACE indices showed that CYP-treated mice had lower alpha-diversity than control in the cecum (Chen et al., 2021). Additional research using multiple time points of analysis may prove useful in unraveling the complex underpinnings of the microbiome brain axis in tandem with the tight interactions of host gut morphology and microbial communities. Research on the long-term effects of chemotherapy on the MGB remains limited. Growing interest in this realm of cancer research is promising and is essential to identifying how the microbiome may facilitate long-term chemobrain and the amelioration of its side effects.

Notably, taxanes and alkylating agents, such as the adjuvants that compose CMF, are well known to induce chemotherapeutic-GIM (Akbarali et al., 2022). Various studies have shown that administration of CTX, MTX, and 5-FU poly- and monotherapies yields an increase in interleukin-1β (IL-1β) expression in response to chemotherapeutic-induced GIM (Wu et al., 2011; Xiang et al., 2011; Yucel et al., 2016). IL-1β is a pro-inflammatory cytokine that is critical to immune system regulation. Evidence has also suggested that IL-1β is a neuromodulator required for healthy brain activity (Schneider et al., 1998). Interestingly, some research proposes that sustained expression of IL-1β may have an antagonistic relationship with hippocampal neurons. Utilizing the transgenic mouse model IL-1β^XAT (Shaftel et al., 2007) for hippocampal IL-1β overexpression, work by Moore et al. has shown that IL-1β^XAT mice displayed spatial memory retention deficits in the Morris water maze task following 2 weeks of IL overexpression (Moore et al., 2009). Both male and female IL-1β^XAT mice exhibited significantly longer path lengths to the platform than control mice. Using the same time frame, Hein et al. demonstrated similar results and found that male and female IL-1β^XAT mice spent significantly more time in non-target quadrants during probe trials of the Morris water maze (Hein et al., 2010). Hein et al. also showed that in unconditioned IL-1β^XAT mice, there was a significant decrease in basal Arc mRNA expression compared to control. Surprisingly, contextual conditioning did not significantly increase basal Arc mRNA expression in IL-1β^XAT mice, regardless of gender. These studies may provide a different lens of interpreting our results. It could be that IL-1β expression was upregulated and potentially overexpressed due to the CMF chemotherapeutic regimen, which was shown to be possible by previous studies (Briones and Woods, 2014). The increase in IL-1β expression may also explain the difference in beta-diversity, such that Wu et al. have shown that IL-1β can impact commensal microbiota in mice (Wu et al., 2022). Moreover, the possibility of IL-1β overexpression may have persisted during the Morris water maze task, which may provide an ancillary reason for the observed cognitive deficits. Subsequent studies could aim to elucidate the mechanistic relationship between the MGB and pro-inflammatory cytokines, with particular emphasis on IL-1β and its antagonist IL-1Ra.

This research provides novel insights into what deficits occur after an extended time frame post-treatment with combination CYP, MTX, and 5-FU chemotherapy. The present study shows that treatment with CMF induced impairments in spatial memory and reduced c-Fos, Arc, and Zif268 expression without producing many protein changes in the hippocampus. Our study emphasizes the capability of CMF to engender persistent hippocampal-dependent cognitive dysfunction, represented by lack of memory retention in the Morris water maze probe trials 30 days post-treatment. These findings are novel and have not been demonstrated in previous CMF studies because of their shorter timelines. Additionally, our results suggest that c-Fos, Arc, and Zif268 may play a significant role in these observed cognitive deficits. Modulation in IEG expression in a female mice chemobrain model in the future could provide novel insights to further investigate the relationship between chemotherapy, IEGs, and cognition. An abundance of research has shown that our targeted IEGs contribute, directly or indirectly, to synaptic plasticity and stabilization of excitatory synapses at major information hubs in the brain. The reduction of these IEGs may explain the continual cognitive impairments seen in this study. Finally, we observed no changes in alpha-diversity and limited significant changes qualitative beta-diversity. Given the lack of differences in gut morphology between the treatment groups, it is not surprising that there were limited differences observed in overall microbial community structure between the treatment groups, indicating that the microbial community structure is following improvements in gut health and morphology. One limitation of our study is that we only examine one time point, which only provides a snapshot into the cognitive changes. Future studies should consider utilizing multiple time points of analysis. Given the varying results of similar studies, investigating the short- and long-term effects of chemotherapy in parallel is perhaps paramount in uncovering the underlying mechanistic relationships within the MGB axis. In addition, future studies incorporating immunostaining for tight junction markers or cytokine profiling could more precisely assess gut barrier integrity and its relationship to the observed behavioral and molecular outcomes. This is particularly true for women in LMICs who do not have the same access to modern breast cancer chemotherapeutic treatments as their Western counterparts. Continued research on CMF chemotherapy is imperative for improving the knowledge, survivorship, and quality of life for breast cancer survivors across the globe.