When my mother was treated for inflammatory breast cancer 20 years ago, I watched as she forgot appointments, where she had put her keys, and whether or not she had taken her medications. Once she even put a chicken still in its wrapping into the oven and didn’t realize the mistake until the plastic began to smoke. As a psychologist and a pharmacologist, I suspected what many patients complain of to their oncologists: that the chemotherapy was affecting her ability to remember and to reason.

The American Cancer Society estimates that more than 1.65 million new cancer cases will be diagnosed in 2013. For many cancers, the prognosis will be very good. For example, 90 percent of breast cancer patients will survive their cancers for at least 5 years. But while cancer chemotherapy is lifesaving, it also has a number of post-treatment adverse effects. Despite the...

Little preclinical or animal research definitively linked chemotherapy to cognitive effects, and some researchers didn’t believe that chemotherapy was really responsible for the observed decline.

In part because of my own experience and in part because my colleague Robert Raffa at Temple University suggested I collaborate with him, I began to research the phenomenon of chemo fog. What I learned at the start was that little preclinical or animal research definitively linked chemotherapy to cognitive effects, and some researchers didn’t believe that chemotherapy was really responsible for the observed decline. In recent years, however, more clinical and preclinical researchers have begun to tease out the contribution of chemo, and have created models that are a useful first step in developing interventions that could mitigate the cognitive decline.

From the clinic

It is challenging to pinpoint whether the drugs used to kill cancer cells are really causing cognitive deficits, because there are so many confounding factors that cloud the observations. Patients are often administered multiple cancer chemotherapeutic agents in different phases of treatment, as well as additional drugs to mitigate the accompanying nausea, fatigue, suppressed immune function, and anemia. Aside from this battery of drugs, surgery can also affect cognition.

In addition to factors that can affect the brain directly, chemotherapy has been known to induce menopause in premenopausal women and bring on depression and anxiety, both of which can affect cognition.1,2 What makes it even more difficult to pinpoint the effects specifically caused by chemotherapy is that not all cancer patients will exhibit cognitive deficits—the prevalence ranges from 17 to 75 percent in breast cancer patients, for instance, depending on the drugs used and the cognitive tests employed by the study. And patients can experience the effects years or even a decade after treatment has ended. Thus, controversy remains over the existence, extent, and underlying mechanisms of chemotherapy-induced cognitive dysfunction.

A number of clinical trials and animal studies have tried to address some of the confounding factors and reveal the precise contribution of chemotherapy treatments. Part of the difficulty in identifying cognitive changes in humans can be attributed to how and when baseline measures of cognitive function are taken. Chemotherapy-induced cognitive deficits are best measured with a full battery of neuropsychological tests administered by a psychologist trained in performing such surveys. However, these tests can be expensive and time-consuming, so about half the clinical studies in this area have used simpler, but less-informative, screening measures. In longitudinal studies, the baseline control tests can be administered before diagnosis to rule out the emotional impact of the news on cognition, before surgery to remove tumors, and after surgery but before chemotherapy, which together provide results that aren’t easily comparable. These methodological differences add to the variability in the field, and researchers have begun to call for consensus in the timing and quality of testing.

CHEMO’S FINGERPRINT: Breast cancer patients imaged before and 3-4 months after chemotherapy treatment showed changes in several white matter bundles, including the corpus callosum, which connects the left and right cerebral hemispheres (1); the superior longitudinal fasciculus (2) and (3), which carries neuronal traffic between, for example, the frontal and parietal association cortices, which are needed for the execution of complex cognitive tasks; and the forceps major (4), which connect the occipital lobes—areas that processes vision. Brain alterations in these areas, measured by magnetic resonance diffusion tensor imaging, tracked with changes in cognitive function.J CLIN ONCOL, 30:274-81, 2012wDespite these difficulties, neuroscientists have been making progress in identifying some of the underlying brain structures that chemo may be impacting—information that will be helpful in categorizing the deficits and knowing what to test for in cognitive experiments. Imaging studies have revealed volume decreases in areas of the brain that correlated with poor attention and impaired memory. Functionally, even years after chemo, patients showed lower levels of responsiveness in both the prefrontal cortex—which is involved in executive functions such as decision-making and social behavior—and the parahippocampal gyrus, which feeds into the hippocampus and is involved in memory formation and retrieval. Using magnetic resonance diffusion tensor imaging, which produces a structural picture of the functionality of the neural tracts, researchers have also found decreased structural integrity of white matter in a number of brain regions, including frontal cortex, as well as white-matter microstructure damage in breast cancer survivors.3 (See image above.)

Researchers have been making progress in identifying some of the underlying brain structures that chemo may be impacting.

The majority of clinical studies examining chemotherapy-induced cognitive deficits have been performed in women with breast cancer because of how long they tend to survive after treatment and because the chemo regimens are somewhat more standardized than for other cancers. However, the Childhood Cancer Survivor Study has also been following patients treated as children with chemotherapy for acute lymphoblastic leukemia (ALL) and measuring the prevalence of chemotherapy-related cognitive deficits later in life, known as late effects. These effects, which include impairments in attention, working memory, IQ score, and processing speed, manifest in adolescence and young adulthood in 40–70 percent of ALL survivors. Evidence from neuropsychological and physiological testing has indicated involvement of the hippocampus, prefrontal cortex, and white matter, but few imaging studies have been done in children to confirm the areas implicated by cognitive tests.

From the lab

Whereas clinical studies are critical for demonstrating a link in humans and identifying the extent of chemotherapy-induced cognitive deficits, they cannot provide much information on molecular and cellular mechanisms, nor do they allow us to explore interventions to prevent these deficits. For this, we turn to animals.

Animal models of learning and cognition have been in existence since the time of Pavlov and his dogs. But recently, researchers have developed more complex cognitive tests that span the functional range of cognition, including learning new tasks, consolidating this information into long-term memory, and recalling and using learned skills later.

AUTOSHAPING: Mice smell the vanilla-flavored Ensure hidden behind the central opening, and can explore the chamber freely (top two panels), but are only given access to the treat after a tone is sounded. They quickly learn to associate the tone with food (bottom panel) and will return to the opening promptly. Mice treated with chemotherapy, however, are slower to learn the association and slower to remember it the following day when they are tested again. PHOTOS BY ELLEN WALKERWe adapted one such test, called the autoshaping assay, which measures the ability of a mouse to rapidly learn a novel response and then recall or remember this task the next day, to examine how different schedules of chemo administration might affect learning and memory. In this assay, healthy male and female mice were first injected with either a single dose of a chemotherapeutic agent or a combination of two. The mice then went into an experimental chamber with a recessed opening containing a cup filled with vanilla-flavored Ensure, a favorite treat for mice.  (See photos on right.) Access to the cup holding the Ensure was only available to the animals for the 6 seconds after a tone was sounded. The mice were trained by trial and error to respond within the tone period on the first day, then tested on the second day for their ability to remember that the tone preceded the Ensure. Drug-free mice easily learned these tasks within an hour on the first day and performed 2–4 times faster on the second day, indicating that they remembered the task. Mice given the chemotherapeutic agent 5-fluorouracil, however—both alone and especially in combination with methotrexate—were slower to learn on the first day, and slower to remember the lesson on the second.4,5 Many other chemotherapeutic agents appear to have similar effects.

In order to learn more about how these drugs might cause the cognitive deficits, a number of researchers have begun to supplement these behavioral studies with neuroanatomical mapping, physiological monitoring, and molecular and cell-culture research. Several preclinical studies have shown that chemotherapy can produce oxidative stress, and reduce vascularization, which decreases the amount of oxygen and nutrients carried to the brain by the blood, as well as cause inflammation of brain tissue.

Cellular studies have shown that chemotherapeutic agents can disrupt the function of CNS progenitor cells, stem cells that give rise to the main cell types of the central nervous system, such as neurons and myelin-sheath-producing oligodendrocytes in the brain’s white matter.6 Several types of chemo can also reduce or block nerve growth factors, and can suppress neurogenesis and proliferation in cultured hippocampal cells and in the brains of mice.6,7 In addition, some research has shown that chemo decreases levels of catecholamines in the brain, which could be related to a reduced capacity for attention and some types of learning. Taken together, these changes could result in decreased hippocampal function and signaling, and subsequently lead to decreased learning and memory, as well as reduced frontal cortex functioning—in humans as well as in rodents.

New therapies

NOVEL-OBJECT RECOGNITION: Mice in holding cages first explore two identical objects, and are then removed from the cage. When mice are returned to a cage that now contains a new object together with the familiar one, they recognize the familiar block and spend more time exploring the novel one. This response, however, is dampened in chemotherapy-treated mice. PHOTOS BY EMILY BISEN-HERSHThe most exciting and promising use of animal models has been to test interventions that might prevent learning and memory deficits from occurring. Researchers have already learned that treatment with the antidepressant fluoxetine can increase neurogenesis in the hippocampus in rat models and prevent deficits assessed by novel location tasks, which test the animals’ ability to respond to and remember new objects introduced to their cage. Fluoxetine also reversed the effects on hippocampal function of methotrexate or 5-fluorouracil.8,9 Exercise appears to have similar effects. Running in an exercise wheel after administration of the drugs prevented 5-fluorouracil- and oxaliplatin-caused deficits in novel-object recognition tasks and in the navigation of the Morris water maze, which examines components of spatial learning and recognition of a newly learned object or place.10 In an aversive-conditioning model where animals learn to avoid an environment in which a noxious stimulus is presented, N-acetylcysteine, an antioxidant being tested for use in humans to counteract chemo toxicity, prevented the deficits produced by the chemotherapeutic agents adriamycin and cyclophosphamide. This suggests that the antioxidant effects of this drug could ameliorate the cognitive decline associated with oxidative damage.11

Despite these promising insights, there are limitations to rodent models. Most studies are performed in healthy young adult male animals, examine a limited number of chemotherapeutic agents, and use dosing and routes of administration that are different from real treatment regimens. But improvements are being made. For example, our group has found additional deficits associated with 5-fluorouracil, methotrexate, cyclophosphamide, tamoxifen, and paclitaxel in female mice using closely matched breast cancer chemotherapy regimens, and has begun some pharmacokinetic drug-drug interaction studies to understand how some drug combinations may interact additively to worsen cognitive decline.

More informative models

Researchers are developing new models altogether, ones specifically aimed at detecting the late effects of chemotherapy. For example, for her thesis project in my laboratory one of my graduate students, Emily Bisen-Hersh, now a postdoc at Vanderbilt University, developed a new animal model to study childhood chemotherapy treatments. While the literature on the effects of cancer chemotherapeutics in adult rodents is growing, these adult models are not adequate for studying the effects of childhood treatment, and studies using young rodents remain limited. Bisen-Hersh’s model allows researchers to study chemotherapy treatment across a developmental trajectory, in order to better understand the cognitive effects that emerge in later adolescence or young adulthood in childhood cancer survivors.

She started by injecting methotrexate and cytarabine into mouse pups at 14, 15, and 16 days after birth—roughly equivalent to a human child aged 5–7 years. She then performed cognitive tests on groups of these pups on postnatal day 35 (equivalent to a human age of approximately 15–18 years). Prior to behavioral testing, there were no significant differences in body weights or health observed in pups that had been injected with either saline control or chemotherapeutic agents earlier. Using the Ensure autoshaping task described above, as well as the novel-object recognition test, we could see deficits when each agent was used alone and in certain dose combinations. However, in another task that required a mouse to spend more time learning to discriminate the difference between a 2-second and an 8-second tone/light combination by poking its nose into a hole associated with tone length, we only saw mild deficits, suggesting that additional training or practice could help reduce the cognitive decline. This result is in line with cognitive remediation programs for childhood cancer survivors that are already in use.

However, not all current therapies used to mitigate cognitive decline in young patients are appropriate, and some of our models have been extremely useful in demonstrating some shortfalls. Young ALL cancer survivors are often treated with the same psychostimulants that are used to treat attention deficit hyperactivity disorder (ADHD). Although psychostimulants such as methylphenidate (Ritalin) are popular medications for treating cognitive late effects in childhood cancer survivors, little is known about the impact these substances have on this medically vulnerable population. We therefore used our model to examine methylphenidate, an amphetamine (Adderall), and a third-tier ADHD medication, atomoxetine (Strattera). On postnatal day 35, 19 days after treatment with methotrexate or cytarabine, chemotherapy-treated mice exhibited increased sensitivity to the rewarding properties of amphetamine and, to a lesser extent, of methylphenidate. In addition, a greater percentage of chemotherapy-treated mice began to self-administer cocaine offered via an IV drip, and maintained a higher number of infusions per session, suggesting that early exposure to chemotherapeutics methotrexate and cytarabine in rodents can produce long-lasting changes that manifest later as impaired cognition and increased sensitivity to novelty, stimulatory effects, and drug reward. Taken together, these results stress the importance of studying chemotherapy treatment in developing rodent brains, since toxins or stress during early development can have more severe long-term effects than identical damage in the mature adult rodent brain.

Going forward, it will be critical to focus on bringing the clinical and animal models closer together in order to identify those patients at the greatest risk of cognitive decline. Equally important will be efforts to develop treatments to prevent post-chemotherapy cognitive deficits. The studies in cancer survivors reveal deficits in attention, memory, processing speed, and executive function; yet testing methodologies available in the clinic can preclude examining the efficacy of cognitive interventions. Although the rodent studies allow us to probe the underlying mechanisms of learning and memory deficits and to test treatments, limitations exist in the translation of preclinical findings to long-term clinical outcomes. The fact that the evidence from rodents to date indicates damage to hippocampal functioning, particularly a disruption of neurogenesis, whereas human studies emphasize cognitive deficits associated with impairments in frontal cortical function, highlights the need for more experimental consensus. Because the reason for this discrepancy may lie in the tests selected by preclinical and clinical researchers, and in the choice of biochemical endpoints to measure, it is important to reconcile these models by working together. I’m optimistic. Between the clinical and preclinical studies, the scientific community is triangulating potential mechanisms and treatments to mitigate this adverse consequence of cancer chemotherapy. 

Ellen Walker is a professor of pharmaceutical sciences at Temple University.


  1. K. Hermelink et al., “Cognitive function during neoadjuvant chemotherapy for breast cancer: results of a prospective, multicenter, longitudinal study,” Cancer, 109:1905-13, 2007.
  2.  C. Scherling et al., “Prechemotherapy differences in response inhibition in breast cancer patients compared to controls: a functional magnetic resonance imaging study,” J Clin Exp Neuropsychol, 34:543-60, 2012.
  3. J.S. Wefel, S.B. Schagen, “Chemotherapy-related cognitive dysfunction,” Curr Neurol Neurosci Rep, 12:267-75, 2012.
  4. J.J. Foley et al., “Effects of chemotherapeutic agents 5-fluorouracil and methotrexate alone and combined in a mouse model of learning and memory,” Psychopharmacology, 199:527-38, 2008.
  5. E.A. Walker et al., “Effects of repeated administration of chemotherapeutic agents tamoxifen, methotrexate, and 5-fluorouracil on the acquisition and retention of a learned response in mice,” Psychopharmacology, 217:539-48, 2011.
  6. R. Han et al., “Systemic 5-fluorouracil treatment causes a syndrome of delayed myelin destruction in the central nervous system,” J Biol, 7:12, 2008.
  7. J. Dietrich, “CNS progenitor cells and oligodendrocytes are targets of chemotherapeutic agents in vitro and in vivo,” J Biol, 5:22, 2006.
  8. L. Lyons et al., “The effects of cyclophosphamide on hippocampal cell proliferation and spatial working memory in rat,” PLOS ONE, 6:e21445, 2011.
  9. L. Lyons et al., “Fluoxetine counteracts the cognitive and cellular effects of 5-fluorouracil in the rat hippocampus by a mechanism of prevention rather than recovery,” PLOS ONE, 7:e30010, 2012.
  10. J.E. Fardell et al., “Cognitive impairments caused by oxaliplatin and 5-fluorouracil chemotherapy are ameliorated by physical activity,” Psychopharmacology, 220:183-93, 2012.
  11. G.W. Konat et al., “Cognitive dysfunction induced by chronic administration of common cancer chemotherapeutics in rats,” Metab Brain Dis, 23:325-33, 2008.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?