Do Cancer Drugs Counteract Neurodegeneration? Repurposing for Alzheimer’s Disease

INTRODUCTION

Accumulating epidemiological evidence indicates an inverse association between Alzheimer’s disease (AD) and overall risk for cancer [1–3]; a reduced relative risk of AD was shown in patients with breast cancer, lung, bladder, colorectal, and prostate malignancies along with non melanoma skin cancer [4, 5]. Additionally, a significant reduction of AD diagnosis from 9 to 45 months before the diagnosis of cancer [6, 7] was shown, indicating that cancer history may be considered a measurable delaying factor for AD onset.

A growing body of evidence also addresses a common pathophysiological background between cancer and brain neurodegeneration. A series of recently developed biological theories support the mutual link between oncogenesis and neurodegeneration; in keeping with that, a meta synchronously action is expected to drive cancer or neurodegeneration, respectively [8].

Herein, the present review provides a summary of the key relevant biological theories, as illustrated in Table 1.

The “metabolic deregulation hypothesis” theorizes a genetic overlap between cancer and AD with the sharing of a mutual independent molecular pathway [8]; the theory indicates aging, mitochondrial dysfunction, oxidative stress, DNA damage, abnormal mitotic signaling, and aberrant cell cycle mitotic signaling as the main determinants of the differential molecular regulation [8, 9].

In particular, the inverse tuning of clustered genes expressions (p53; pin1; Wnt; UPS) [10–12] is considered to drive either the proliferative or the pro apoptotic pathway. According to this theory, on one hand, the uncontrolled cell proliferation is expected to promote cancer dissemination, exerting at the same time brain neuroprotection; on the other hand, the anti proliferative effect, is supposed to increase apoptosis, arresting neuronal mitosis and exerting neuroprotection.

The “cell-cycle theory” hypothesizes that oxidative stress and abnormal mitotic signaling are essential triggers which promote neuronal cell cycle re-entry and cancer cell uncontrolled proliferation [13–16], in order to cope with the cellular metabolic energetic crisis [17, 18]. In line with this theory, cancer cells are expected to effectively stabilize microtubule and mitosis, and neurons to stick in a deregulated mitotic state, increasing extracellular amyloid burden and tau hyperphosphorylation.

More recently, a sophisticated biological model [19] has focused on the aberrant brain cellular proliferative/regenerative response; the biosynthetic brain cycle is considered to trigger maladaptive responses (cell mitotic re-entry, glial proliferation, and aberrant neuroplasticity), with increased neurotoxicity and neuronal and synaptic loss. This same biosynthetic cycle is considered to be maintained at the expense of other biosynthetic resources, such as the cellular systems associated to tumorigenesis.

On the other hand, the increased cell anabolic activity, driving malignancies and self-maintaining aberrant cell proliferation, is expected to counteract or at least delay the onset of AD hallmarks [17, 20].

The last prominent biological theory in the field lies on the deregulated unfolded protein system (UPR) [21], which could be considered the main biological target for both cancer and AD. The UPR system mediates both survival signals (by activating anti apoptotic pathways) and cell-associated death programs. The integrity of the UPR system is essential for maintaining brain homeostasis, through the proteolytic removal of damaged proteins and new proteins synthesis, particularly in the case of brain increased oxidative stress. According to the Mallucci hypothesis [21], the deregulated UPR system is responsible for a catastrophic brain translational protein failure, with neuronal death, mediated by increased amyloid-β (Aβ) oligomerization and neurotoxicity. Besides this, other biological mechanisms, such as the cell-cycle theory are supposed to act synchronously, boosting either the brain pro amyloidogenic burden or the uncontrolled cancer cell proliferation.

So far, the pathophysiology provides some evidence to the inverse correlation between cancer and AD and lends credit to the recent discovery of a repurposing role for cancer drugs in AD. It represents a potential new effective intervention for dementia [22] and refers to the development of new uses for existing pharmaceuticals having the same role as the original mechanism or to the discovery of new drug action with disease modifying effects (MOA) [23, 24].

A cornerstone longitudinal analysis of older chemotherapy treated breast cancer survivors, has shown, as a by treatment effect, a lower risk of AD compared to controls [25]. The study has firstly introduced the concept that anticancer drugs could inhibit the uncontrolled cell proliferation, relieving, at the same time, brain amyloid degenerative burden.

So far, accumulating evidence addresses a role for repurposing cancer drugs in AD, even if with high heterogeneous results, as summarized in Table 2.

A recent review [26] has examined several anticancer drugs, to investigate their repurposing therapeutic effects for AD, based on drugs’ physicochemical characteristics and blood-brain barrier (BBB) permeability scores. The initial screening has prioritized anti cancer drugs for brain neoplasms, using in vivo models of xenografted mice with human glioblastoma cells [27]. In particular, Vinca-drugs [27] have shown poor BBB permeability while etoposide, methotrexate, and carboplatin have shown negligible BBB permeability, insufficient to exert any brain effect [28]. It must be noted that methotrexate, in an in vivo mouse model, has shown to induce hippocampal cellular dysfunction [29].

In turn, trimetazidine (TMZ), carmustine, and lomustine have shown good BBB permeability [30]. Namely, TMZ has shown that oxidative damage in an AD rat model is prevented, with a significant decrease of malondialdehyde (MDA) levels along with an increase of hippocampal superoxide dismutase (SOD) and catalase activity. TMZ administration also increased the expression of DHCR24 (Seladin-1) gene in the hippocampus, indicating a potential neuroprotective effect of TMZ in AD [31]. In addition, a Phase I trial is on going, testing TMZ plus memantine and mefloquine to assess potential therapeutic application for AD [32, 33].

Interestingly, BCNU (1,3 bis (2-chloroethyl)-1-nitrosourea or carmustine), an alkylating agent used to treat malignant gliomas was found, in cultured cells overexpressing amyloid-β protein precursor (AβPP), to strongly reduce Aβ production [34] by increasing the secretion of AβPPα (alternative α-cleavage of AβPP within the Aβ region). BCNU was able to alter the trafficking and processing of AβPP without interfering with the secretase activities. In addition, BCNU was found to increase transforming growth factor (TGF)-β1 levels, improving glial neuroinflammation [34]. Following these cell-based experiments, the same authors have performed in vivo experiments; the results have confirmed the role of carmustine in decreasing levels of Aβ and AβPPα C-terminal fragments, increasing, the levels of secreted AβPP in mouse brain as well. Additionally, carmustine suppressed microglial activation in the mouse brain. [23]. Thus, carmustine may reduce Aβ production via additive effects on AβPP trafficking and on TGF-β1 pathway. However, the anti-amyloidogenic effects of BCNU may derive from a metabolite that remains to be identified. Further, the higher drug-associated toxicity implies that a safer compound has to be developed before a clinical use.

Furthermore, JLK 1486, an interesting new chemical compound, has shown brain anti-tumor and neuroprotection on mice xenografted with human glioblastoma [35, 36]; it was found to activate the (C/EBP)-homologous protein (CHOP) signaling, overturning the deficits of UPR system, responsible for neural deficits [37].

Concerning anticancer drugs for different oncological malignancies, there is growing evidence for Epothilone D (Epo D) and bexarotene, as they both have shown the best BBB permeability scores.

Epo D was found to improve microtubule density, axonal integrity, and cognition, in a transgenic mouse model of tauopathy [38, 39], providing preliminary in vivo evidence that a brain penetrant microtubule stabilizing (MT) agent may be a therapeutic strategy for AD and frontotemporal dementia as well. Furthermore, the comparison between several MT stabilizing compounds, from the taxane and epothilone product families [40] has concluded that epothilone displayed greater brain penetration, increased brain MT stabilization, with a goodpharmacodynamics profile in mice. Further evidence has confirmed that epothilone treatment, in tau transgenic mice, improved axonal dystrophy and MT integrity along with increased hippocampal integrity and overall cognitive and behavioralperformance [41].

In addition, epothilone, at sub nanomolar concentrations, has shown to reverse Aβ-mediated dendritic spine loss, resulting in decreased microtubule destabilization and enhanced spine density as well [42].

However, epothilone exhibits potential deficiency as a candidate for MOA treatment in AD, including the intravenous route of administration and the inhibition of P-glycoprotein transporter. Recent evidence has documented the superiority of orally bioavailable and brain penetrant MT stabilizers, other than EPO, to be the ideal therapeutic choice for AD andtauopathies [43].

Interestingly, mounting evidence addresses bexarotene as a good candidate for further drug development in AD treatment, given its high BBB permeability and current safety profile [38].

Bexarotene is a compound used to treat cutaneous T-cell lymphomas [38, 44], that has shown potential MOA effects in AD mouse models. Its putative MOA mechanism relies on the retinoid X receptor agonism that upregulates the high-density lipoprotein pathway, such as apolipoprotein E (APOE) expression, affecting liver X receptor-retinoid X receptor complexes [45].

In vitro evidence has shown that bexarotene induced peroxisome proliferator-activated receptor-γ and APOE-dependent proteolytic degradation of Aβ [46]. Furthermore, bexarotene shares a similar structure with cholesterol, competing for the same binding site at the C terminal region of Aβ peptide 1-42. Bexarotene has shown to efficiently prevent cholesterol dependent increase of calcium fluxes on amyloid channel formation, hindering the oligomerization of amyloid fibrils [47]. In addition, in vitro models of human BBB [48] have shown that bexarotene not only promoted the cholesterol exchange between the brain and the blood but also decreased the apical-to-basolateral influx of Aβ peptides across BBB. The results of a recent study [49] have confirmed that bexarotene is able to prevent the impaired bidirectional exchange of Aβ, impeding the excessive Aβ brain accumulation [50–52]. Furthermore, bexarotene was shown to reverse Aβ impaired excitability in hippocampal neurons associated to K channels insulin signaling pathway dysfunction, [53] as well as to reduce network excitability in mouse models of AD disease and epilepsy [54]. This last evidence suggests a role for bexarotene in Aβ-independent increase of cortical network excitability associated with AD.

Interestingly, recent evidence, addressing a chemical kinetic approach, has shown that bexarotene was able to suppress the primary nucleation of toxic Aβ₄₂ aggregates; in keeping with that bexarotene associated MOA mechanism is expected to selectively target the primary nucleation step of Aβ [55]. Additionally, another evidence has shown that early treatment with age-dependent critical concentrations of bexarotene, in a AD mouse model, significantly improved the efficacy of bexarotene-mediated Aβ clearance [56].

Thus, a growing bulk of evidence has addressed a multi-target drug effectiveness for bexarotene in AD [47]. Regarding that, a proof of concept trial of bexarotene X, in moderate AD patients, has been recently conducted to provide effective clinical translation of potential innovative therapies for AD [57]. The primary outcome of the trial was negative as bexarotene did not reduce brain amyloid (amyloid imaging with florbetapir scans) when all the patients were included in the analysis. Elevated triglycerides represented a major side effect along with a higher cardiovascular risk; thus, bexarotene should only be administered in a research setting, warranting in depth investigations in retinoid X receptor activities in AD.

In line with that, another in vivo evidence [58] investigating a multifactorial pathological phenotype of AD mice (TASTPM) has demonstrated the failure of bexarotene in AD. In particular, bexarotene did not induce any significant memory improvement, as well as no Aβ plaque reduction or decreased microglial activation. Additionally, the brain analysis with magnetic resonance did not detect any bexarotene-mediated change in the volume reduction of the mice affected brain areas. Another original investigation has tested the potential therapeutic effects of bexarotene in two different animal species and concluded that bexarotene failed any efficacy even if administered in the preclinical period, prior to the onset of cognitive deficits reported for human AD [59].

Tamibarotene (Am80) [60] is another synthetic retinoid (retinoic acid receptor agonist) approved in Japan for the treatment of acute promyelocytic leukemia. In a senescence-accelerated mouse model (SAMP) [61], Am80 has ameliorated the decrease of cortical acetylcholine, with an immunomodulatory effect, by reducing proinflammatory cytokines and chemokines secretion from astrocytes and microglia. Further, Am80 has also shown to improve mice behavioral tests.

Am80 was also found to promote differentiation of neural stem cells. According to this evidence, a clinical study was initially conducted to evaluate the efficacy and safety of Am80 for the treatment of AD [61]. Tamibarotene has also shown to improve ADAM10 expression in SAMP8 mice models, by activating the hippocampal ADAM10-Notch-Hes 5 proliferative pathway, ultimately improving working memory.

To a greater extent, the repurposing of anticancer drugs for AD might also encompass the role of peroxisome activated receptor gamma and liver x receptors (LXR) in coordination with retinoid X receptors in AD, as an additive neuroprotective effect. A recent in vitro investigation, on the combining effects of LXR in coordination with retinoid X receptors, have not shown any significant change in brain and cerebrospinal fluid Aβ₄₀ levels of naïve rats [62].

To further entangle the clinical issue, another study has demonstrated the improvement of memory deficits and Aβ clearance in aged AβAPP 23 mice treated with a combination of anti-Aβ antibody and LXR agonist [63].

Among the remaining chemotherapy for different types of cancer, imatinib, a tyrosine kinase inhibitor, currently approved for the treatment of chronic myelogenous leukemia, targets the Bcr-Abl complex and the ATP binding site of different tyrosine kinases. Mounting in vitro and in vivo evidence has suggested that imatinib may exert therapeutic effects in AD by reducing amyloid burden and promoting neuroprotection. Imatinib was initially found to inhibit Aβ formation without affecting the Notch system [64]; further in vivo evidence has indicated an imatinib-mediated Aβ lowering effect, by preventing γ-secretase activating protein (GSAP) interaction with γ-secretase substrate AβPP carboxy-terminal fragment [65, 66].

Imatinib was shown, in vivo, to selectively block GSAP without affecting other pathways (steady state of AβPP, BACE-1, ADAM-10, or the four components of the γ-secretase complex); the results support imatinib as a viable target for an anti-Aβ therapeutic approach [67].

Interestingly, imatinib was found to upregulate protein transthyretin (TTR) and neprilysin (NEP) by the intracellular domain of AβPP (AICD), leading to enhanced Aβ clearance. These results have addressed a potential role for imatinib in the pharmacological regulation of both NEP and TTR as a viable therapeutic target in AD [68–71].

To overcome the poor BBB permeability of imatinib, an in vivo study has demonstrated that the co-administration of LPS, by increasing the inflammatory-mediated BBB leakage and permeability, led to a significant reduction of hippocampal Aβ burden, along with an overall cognitive improvement in mice [72].

The in vivo and in vitro treatment with imatinib has also shown to participate in neuronal gene expression [73]. Recently, an in vivo model has shown that blocking the production of Aβ that follows inflammation resulted in a significant reduction of tau phosphorylation. This study originally gave credibility to a model in which Aβ initiated tau phosphorylation [74].

Conversely, a recent study has questioned the role of imatinib in inhibiting amyloidogenic AβPP processing: an in vitro model of cell overexpressing AβPP with Swedish mutation (HEK293), treated with imatinib for twelve months, has not achieved any lowering of brain Aβ burden [75]. The same authors have collected plasmatic samples, in a group of treated patients with chronic myeloid leukemia, and no significant decrease of plasmatic Aβ was found.

From a clinical perspective, in spite of low cardiomiotoxicity, no major adverse event in humans has been reported with imatinib. Unfortunately, the current use in AD is hampered by its low BBB penetration and by the high rate of brain removal by P-glycoprotein system [64, 77].

Another compound, thalidomide, a chemotherapy currently approved for multiple myeloma relapse, has been proven to be effective in reducing cerebral microvasculature derangement, endothelial dysfunction, blocking astrogliosis and hippocampal neuronal apoptosis [78].

In an AD mouse model, the chronic administration of thalidomide has shown to ameliorate amyloid-like pathology through the inhibition of β-secretase, along with a decreased activation of astrocytes and microglia [79]. In addition, thalidomide has shown to block astrogliosis, and to decrease brain neuroinflammation, by inhibiting TNFα [80]. In the same mouse model, thalidomide was found to reduce multiple hallmark features of AD, including phosphorylated tau protein, AβPP, Aβ peptide, and Aβ-plaque number along with deficits in memory function present in younger adult cognitively unimpaired AD mice[81, 82].

Restraining TNFα, thalidomide was found to prevent the nitration of proteins in the hippocampus and the memory impairment of recognition memory in mice. The results have underlined the feasibility of targeting TNFα as a preventive strategy against Aβ accumulation [83].

Paclitaxel (PTx/taxol), an anti-cancer taxane drug, used in solid tumors, has demonstrated its potential application in AD and tauopathies, due to the reduction of tau phosphorylation, and Aβ neuronal death. In particular, PTx was found to prevent Aβ-mediated cell death by inhibiting Aβ-induced activation of cytosolic cdk5-p25 complex and calpain, respectively; this mechanism resulted in an overall improvement of tau function and decreased Aβ toxicity in cortical neurons [84–86].

Paclitaxel has also shown that pre treatment of neurons with a MT stabilizing drug prevented the induction of UPR with prevention of endoplasmic reticulum dysfunction and subsequent cell death by Aβ in neurons [87].

Growing evidence addresses a role for paclitaxel in clearing tau aggregation. In vitro evidence, on cultured Aplysia neurons, has shown that paclitaxel prevented MT human-tau-induced swelling of axonal segments along with the translocation of tau and microtubules. Paclitaxel has demonstrated to potentially serve as slowing down drug of tauopathies unfolding; unexpectedly, higher concentration doses even facilitated the protein unfolding [87].

Moreover, paclitaxel was found to reverse tau-induced synaptic dysfunction at the axoplasmic transport and the releasable presynaptic vesicle stores levels, rescuing tau induced synaptic transmission pathology [88, 89]. Furthermore, paclitaxel was shown to partially reverse the blocked mitochondrial transport associated to microtubules derangement, exerting a neuroprotective effect by microtubule stabilization and axonal transport maintenance [90]. Additionally, PTx has demonstrated a possible role in tau protein mediated dynamic rearrangement of cytoskeletal fibers and BDNF-induced synaptic plasticity [91].

Moreover, paclitaxel was found to phosphorylate Aβ-related death inducing protein, providing a mechanistic link between microtubule disruption, mitotic cell abnormalities (G2/M arrest), and neuronal dysfunction [92]. Paclitaxel was also shown to reduce GTP stimulated tubulin polymerization, affecting tubulin function in AD models as well as in frontotemporal dementia with Parkinsonism associated to chromosome-17 [93, 94]. Unfortunately, this drug has poor BBB penetration; feasible small chemical modifications of taxol have been generated, to reduce affinity to BBB P-glycoprotein and reach target therapeutic in central nervous system [84, 95], with inconclusive findings.

Recently, NVP-BEZ235 (BEZ; dactolisib), a dual Phosphoinositide 3-Kinase (P13K)/mTOR inhibitor under Phase I/II clinical trial, has shown promising anti-amyloid effects in both in vitro and in vivo models [96]. Substantial evidence indicates that P13K/mTOR axis is involved in brain neurodegeneration; in cell hippocampal cultures, BEZ reduced Aβ-mediated neuronal death and microgliosis. BEZ did not affect BDNF and neuronal growth factor and restored inflammatory cytokines levels (IL-10 and TNFα).

However, BEZ did not change the phosphorylation of AKT and p70s6K, indicating a controversial molecular mechanism. In addition, in animal mice models, BEZ has shown to effectively prevent Aβ-induced memory deficit in the object recognition task. To date, even if the drug is promising, further studies are needed to investigate the protective effect promoted by BEZ in transgenic AD models [96].

The potential mechanisms that link cancer and AD have not yet been fully clarified. AD is associated with the degeneration of cholinergic neurons and nicotinic acetylcholine receptors signaling. However, it was found that both could promote the synthesis of growth factors and neurotrophines boosting cancer proliferation [4, 97]. Additionally, Aβ could directly activate tumor suppressor gene p53, inducing apoptosis; the mutation of presenilins was also found to trigger p53-dependent cell death [4, 98]. Accordingly, growing evidence addresses an upregulation of p53 in AD brain [4, 99]. Interestingly, Pin1 a peptidyl-prolyl cis/trans isomerase has shown a promising dual role in cancer and neurodegeneration. PIN1 regulates several oncogenic signaling pathways involving cyclin D1 and p53, and was found to be overexpressed in human cancer. Moreover, an in vivo model has shown that Pin1 deletion in mice is associated to neurodegeneration that resembles human AD [4, 99].

So far, several molecular pathways have shown connections between cancer and neurodegeneration. Such a new set of signaling pathways, including increased oxidative stress, mitochondrial dysfunction, formation of misfolded proteins, and impaired cell metabolism can be defined as anti-aging pathways. A better understanding of the intertwined mechanism between the aging process and age related diseases may drive future research in the field of aging, cancer, and the multi level modulation of different pathological events of AD.

Currently, basic science is unable to disentangle the whole issue; the gap of knowledge relies on the poor investigation of different anticancer drugs and the lack of understanding between living animal models and human pathophysiology [100]. Future investigations are expected to understand the clinical benefit to harm ratio of such oncological compounds in the central nervous system along with the alteration of blood-brain barrier and the response of adjacent or other tissues. Additionally, the tailoring of site-specific cancers, of gender effects, and of the inter variability of the biology of aging may provide further insight to potential MOA treatment for AD.

In parallel, AD research is expected to overcome the current limited areas of therapeutic target and to encompass the disease clinical complexity; the real challenge is to effectively develop a multilevel clinical phenotype approach to AD.

The crosstalk among aging, neurodegeneration, and cancer is still far to be addressed; a deeper insight on the cellular metabolic/energetic crisis and related molecular mechanisms may foster stronger bench to bedside evidence, with the development of innovative therapeutic alternatives.

So far, the potential disease modifying effect of repurposing cancer drugs in dementia is promising; however, further investigations are needed to unravel the clinical relevance of such attractive candidate compounds for AD.

DISCLOSURE STATEMENT

Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/16-0840r2).