Improving our understanding of the complex relationship between cancer-related cognitive decline and Alzheimer’s disease
Year: 2024
Type: editorial
Abstract: As the population continues to age, incidence and prevalence of many aging-related conditions, including both cancer and Alzheimer’s disease (AD), will continue to increase (1-3). Moreover, cancer survivors also continue to age, with an anticipated 74% of survivors being greater than or equal to 65 years old by 2040 (4). Thus, there is an urgent need to investigate and understand health outcomes in cancer survivorship, with particular attention to long-term and late effects of cancer and its treatments. Of particular concern to older adults with cancer is the potential for cancer-related cognitive decline (CRCD), a common long-term effect of cancer and its treatments characterized by deficits in attention, memory, and/or executive function (5-8). However, causes of CRCD remain elusive. Given similar cognitive symptomatology, a potential overlap between CRCD and AD seems plausible (3). Many studies have observed a bidirectional inverse relationship between cancer and AD and have indicated that observed cognitive deficits in CRCD are generally milder than those observed in AD (9-12). However, understanding the biology underlying a potential relationship between CRCD and AD is needed. In this issue of the Journal, Mandelblatt and colleagues explore the association between blood-based AD-related biomarkers and cognition both at baseline and up to 60 months of follow-up using data from a case-control study of older women with nonmetastatic breast cancer (13). Their analysis includes 236 breast cancer survivors and 154 friend or frequency-matched community-based controls. They collect baseline blood-based biomarkers of AD, including neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), beta-amyloid 1-42 (Aβ42), beta-amyloid 1-40 (Aβ40), and phosphorylated tau at threonine 181 (p-tau18). They also collect subjective cognition and objective neurocognitive testing at baseline and annually thereafter, including assessments of attention, processing speed, and executive function (via a battery of 6 tests) and learning and memory (via a battery of 5 tests). Overall, results indicate that breast cancer survivors have lower attention, processing speed, and executive function scores at baseline and over time compared to controls, but AD-related biomarkers are not associated with these scores over time. Learning and memory scores at baseline and over time are similar between survivors and controls with still no association between AD-related biomarkers and cognitive scores. These results are consistent with prior studies indicating an inverse association between AD and cancer (9-12). Studies consistently report that the risk of AD is lower after a cancer diagnosis, and, conversely, the risk of cancer is lower after an AD diagnosis (9,10). Moreover, studies also consistently report that observed cancer-related cognitive deficits are milder than those observed in AD (14). Additional studies reported that, among those with cancer who did develop subsequent AD, cognitive decline was slower among those with history of cancer compared to those without (12,15). These results are contrary to what may be expected given similar cognitive symptoms and the role of aging in driving risk for both cancer and AD. Thus, one biological hypothesis suggests that because cancer and AD operate in opposing biological pathways, with cancer being proliferative and AD being degenerative, it requires substantial bodily resources to overcome one and enter the other pathway (16). However, AD involves not only neurodegeneration but also includes accumulation and/or proliferation of amyloid plaques and tau tangles in the brain (3). In the current paper, Mandelblatt and colleagues seek to elucidate the potential role of these and other AD-related biomarkers in explaining CRCD (13). This is particularly important given that evidence suggests amyloid accumulation begins decades before AD symptom onset (17). Thus, perhaps adults with cancer and pre-existing evidence of amyloid accumulation are at greater risk for CRCD. Results from Mandelblatt and colleagues’ study suggest otherwise (13). These results are also consistent with prior studies suggesting that the use of donepezil, a commonly prescribed medication for mild-to-moderate AD, is not associated with any improvement in CRCD (18,19). Although donepezil does not act directly on amyloid or tau, it does slow cognitive decline in AD (19). Moreover, newer monoclonal antibodies approved for AD such as lecanemab do operate directly on amyloid plaque accumulation, but given no evidence of an association between amyloid and CRCD in this study, it seems futile to investigate the role of lecanemab on CRCD (20). Despite no evidence of an association between AD-related biomarkers and CRCD in this study, there remain plausible biological connections between the two conditions. For example, proinflammatory cytokines and/or reactive oxygen species associated with both cancer and AD may lead to reduced cognitive function (eg, tumor necrosis factor-α, interleukin-6) (21-24). This study points research in a direction away from AD-specific biomarkers and toward more general inflammatory markers of cognitive decline. Additionally, other plausible mechanisms for development of CRCD include inhibited generation of hippocampal neurons, damaged white matter, reduced activity in the hypothalamic-pituitary-adrenal axis, and impaired blood flow in the brain (23). Most of these hypotheses are also present in AD, and yet there remains a consistent inverse relationship between the two conditions (25-29). Developing an understanding of the biological link (or lack thereof) between CRCD and AD may assist in drug or intervention development for either condition. For example, if there seems to be a protective factor for AD due to cancer, basic scientists may be able to leverage it for drug development to prevent AD (provided the protective factor is not a risk factor for cancer). Given the lack of success of many AD drug trials over the past several years, this may provide a solid new direction in an increasingly prevalent disease with no known cure. Similarly, if AD seems to be a protective factor for cancer, understanding and leveraging this mechanism for drug development to prevent cancer may also be plausible. However, this would not address CRCD directly. Evidence suggests that CRCD is a function of cancer and/or its treatments with perhaps some additional psychosocial components (30). However, it is not clear why some, but not all, cancer survivors develop CRCD across the cancer continuum and/or why some survivors have persistent, long-term CRCD. Given current results from Mandelblatt and colleagues, it seems unlikely that pre-existing AD-related biomarkers are potential predictors or causes of CRCD (13). However, the majority of cases in this sample are those with early-stage disease, and only 19% received chemotherapy. Thus, additional studies are needed to determine the consistency of these results across cancer types and treatment characteristics (eg, type and/or dose of chemotherapy, radiation, targeted therapies). Perhaps baseline AD-related biomarkers have a differing role across cancer types and/or more advanced disease requiring greater systemic treatment burden. Moreover, although there is an observed difference in cognitive performance at baseline between cases and controls (not due to AD-related biomarkers), there does not appear to be a significant difference in longitudinal decline, which may be due to the sample composition or due to practice effects over time. Therefore, studies should be conducted to understand potential practice effects among both cases and controls and how these effects may contribute to longitudinal cognitive performance. Given the lack of an association between AD-related biomarkers and baseline cognitive status, it seems unlikely that there would be a longitudinal association in this sample even accounting for practice effects, but it may be very important when investigating other potential biological mechanisms. Finally, the results reported by Mandelblatt and colleagues support that other risk factors such as education and/or cognitive reserve, psychosocial dysfunction, and/or genetic markers may be prevailing risk and protective factors for cognitive differences between those with and without cancer (30). Again, no inferences specific to longitudinal decline can be made, but these additional factors may contribute to the baseline difference observed between cases and controls. For example, higher education and pre-morbid cognitive reserve have demonstrated protective effects for CRCD, whereas pre-morbid depression or anxiety may contribute to worse cognitive deficits during or after cancer (8,30-32). Additionally, some genetic markers (some of which are also associated with AD risk like APOE ε4) are associated with greater burden of cognitive impairment among adult cancer survivors (30,33). However, to our knowledge, a direct comparison of longitudinal cognitive performance between those with CRCD, those with AD, and healthy controls has not been done accounting for such factors as education or other indicators of cognitive reserve, psychosocial dysfunction, and/or genetics. Understanding these risk factors as potential contributors to CRCD will allow identification of those most at risk for CRCD and/or inform intervention targets. In conclusion, as the older adult population continues to expand and cancer, CRCD, and AD become increasingly prevalent, it is essential to investigate the overlap between common age-associated conditions that are key drivers of quality of life and disability. Thus, we believe that continued investigation into the complex biological mechanisms between CRCD and AD as done by Mandelblatt and colleagues will result in additional targets for translation from the bench to the bedside and population. Moreover, given the limited availability of data sources with both robust cancer and robust AD data, leveraging existing studies either through addition of further measures as done in Mandelblatt and others’ study or through secondary data analyses provides an excellent opportunity to conduct these analyses more quickly and efficiently. No new data were generated or analyzed for this editorial. Mackenzie E. Fowler, PhD (Conceptualization; Writing—original draft; Writing—review & editing), Michael Crowe, PhD (Conceptualization; Writing—review & editing). No funding was used for this editorial. M.E.F. has no disclosures. M.C. has no disclosures.
DOI: 10.1093/jnci/djae146
Authors: Mackenzie E Fowler, Michael Crowe
Language: English
Subfield: Oncology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4400409454