Cholecystokinin (CCK): a neuromodulator with therapeutic potential in Alzheimer ’ s and Parkinson ’ s disease

Cholecystokinin (CCK) is a neuropeptide modulating digestion, glucose levels, neurotransmitters and memory. Recent studies suggest that CCK exhibits neuroprotective effects in Alzheimer ’ s disease (AD) and Parkinson ’ s disease (PD). Thus, we review the physiological function and therapeutic potential of CCK. The neuropeptide facilitates hippocampal glutamate release and gates GABAergic basket cell activity, which improves declarative memory acquisition, but inhibits consolidation. Cortical CCK alters recognition memory and enhances audio-visual processing. By stimulating CCK-1 receptors (CCK-1Rs), sulphated CCK-8 elicits dopamine release in the substantia nigra and striatum. In the mesolimbic pathway, CCK release is triggered by dopamine and terminates reward responses via CCK-2Rs. Importantly, activation of hippocampal and nigral CCK-2Rs is neuroprotective by evoking AMPK activation, expression of mitochondrial fusion modulators and autophagy. Other benefits include vagus nerve/CCK-1R-mediated expression of brain-derived neurotrophic factor, intestinal protection and suppression of inflammation. We also discuss caveats and the therapeutic combination of CCK with other peptide hormones.


Introduction
Cholecystokinin (CCK) is a gut-brain peptide with many roles.It modulates metabolic processes, acts as a neuromodulator and as a growth factor in the brain (Rehfeld, 2017).CCK is widely synthesised across the body, such as by intestinal endocrine l-cells mainly located in the duodenum, jejunum and ileum, as well as by peripheral nerves and immune cells.Furthermore, greater than in the periphery, CCK is expressed at high levels in almost every brain area, e.g. the hypothalamus, hippocampus, cerebral cortex, striatum and spinal cord (summarised in (Rehfeld, 2017)).While region-specific, the level of cerebral CCK expression is much greater than that of other neuropeptides (Crawley, 1985;Rehfeld, 1978), highlighting the importance of CCK.An exception is the cerebellum, where only pro-CCK is expressed especially in fetal tissue, but conversion to bioactive CCK (with α-amidated COOH terminus 'Trp-Met-Asp-Phe') is virtually absent (Rehfeld, 1978;Rehfeld et al., 1992).
On a physiological level, CCK acts as a hormone that promotes digestive processes in the periphery.This includes the stimulation of bile release from the liver and gallbladder, production and secretion of pancreatic enzymes, improvement of gut motility and blood flow as well as satiety-signalling across CCK-1Rs located on vagal afferent nerves (Rehfeld, 2017).CCK also inhibits gastric acid secretion through CCK-1R-mediated somatostatin release (Chen et al., 2004).
Importantly, CCK plays key roles in mood, plasticity and memory formation, as evident by co-localisation with cell bodies of glutamatergic pyramidal (Morino et al., 1994), GABAergic (Somogyi et al., 1984), dopaminergic (Hokfelt et al., 1980) and serotonergic neurons (van der Kooy et al., 1981).Our group has recently demonstrated that CCK has protective effects in animal models of neurodegenerative diseases (Zhang et al., 2022;Zhang et al., 2023).As such, this review aims to illustrate the implications of CCK in the brain.This includes the neuropeptide's physiological impact on hippocampal plasticity and memory formation and regulation of the nigrostriatal and mesolimbic dopamine release.We further describe the neuroprotective effects, with a focus on AD and PD.Finally, we cover related key aspects of CCK, including colonic protection, brain-derived neurotrophic factor (BDNF)/ nerve growth factor (NGF) expression and anti-inflammatory properties, as well as the therapeutic opportunity for growth factor combinations.

Cholecystokinin and receptor expression in the hippocampus
CCK, particularly the CCK-8S form, is one of the most expressed genes in the hippocampus (cornu ammonis (CA)1), indicating a key role in memory formation (Meyer, 2014).The cell bodies of CCKimmunoreactive neurons are located in the entorhinal cortex, subiculum, stratum pyramidale, amongst dentate granule cells and within the inner molecular layer of the dentate gyrus (Greenwood et al., 1981).CCK-producing neurons are exclusively GABAergic and regulated by incoming CCK-positive synaptic boutons, while CCK neurons form synapses with local pyramidal and non-pyramidal neurons (Nunzi et al., 1985;Somogyi et al., 1984).CCK binding sites and expression of both CCK-1Rs and CCK-2Rs has been reported for all major subregions of the hippocampus, including the (pre/para-)subiculum, dentate gyrus, CA1, CA2 and CA3 (Honda et al., 1993;Kritzer et al., 1988).CCK-2Rs are present on excitatory pyramidal neurons and GABAergic interneurons (Lee et al., 2011) and expressed at greater levels compared to CCK1Rs (Honda et al., 1993).Hippocampal CCK-2Rs principally contribute to memory formation, whereas peripheral CCK-1Rs expressed on vagal afferent fibres seem to be indirectly involved by enhancing hippocampal BDNF expression (section 5.1) (Lemaire et al., 1994a;Peters et al., 2006;Tirassa & Costa, 2007;Tirassa et al., 1998).

Transient facilitation of glutamatergic neurotransmission
CCK modifies both hippocampal presynaptic glutamate release and postsynaptic excitation.CCK-8, CCK-8S or a CCK-2R agonists were shown to elicited the endogenous release of glutamate, but not GABA, in hippocampal slices or into synaptosomes in a CCK-2R and extracellular Ca 2+ -dependent manner (Breukel et al., 1997;Migaud et al., 1994).In vivo, however, peripheral CCK-8S administration resulted in reduced hippocampal glutamate levels 30 min post injection (Acosta, 1998).This suggests that CCK-mediated glutamate release is time-dependent and initially heightened, but then decreased.A study using cultured rat brain slices has confirmed that the addition of CCK, or endogenously released CCK, acutely elevates synaptic vesicle numbers and glutamate release probability in the hippocampus (Deng et al., 2010).These effects were conveyed by CCK-2Rs, involving phospholipase C beta (PLCβ), Ca 2+ and protein kinase C (PKC)-signalling.CCK-indued PKC activation in the hippocampus likely prevents neuropeptide Y receptor 1/2/5-mediated inhibition of voltage-gated Ca 2+ channels and presynaptic glutamate release (Fig. 1) (Silva et al., 2003;Silva et al., 2007).
There are time-associated effects of CCK on hippocampal plasticity in vivo, however.CCK-8S was shown to transiently enhance tonic excitatory postsynaptic potential and population spikes 5 -10 min post i.p. injection, followed by their return to baseline from 15 min (Dolatabadi & Reisi, 2014).Moreover, relative to controls, CCK-8S-treated animals showed enhanced paired-pulse responses in the dentate gyrus at 40 ms, but deteriorated ones between 70 -300 ms.Because CCK was shown to promote presynaptic glutamate release, but not vesicle recovery (Deng et al., 2010), this suggests that the neuropeptide acutely triggers a burst release of glutamate that leads to depletion of this neurotransmitter in response to prolonged stimulation.
In both the CA1 (ex vivo) and dentate gyrus (in vivo), CCK-2R activation on interneurons was demonstrated to induce K + channel inhibition, spontaneous GABA release and increased inhibitory post-synaptic current (iPSC) frequencies 2 -4 min post administration, followed by a persistent iPSC reduction from 10 min (Deng & Lei, 2006;Miller et al., 1997).Interestingly, via CCK-2Rs, either CCK, conotoxin (N-type Ca 2+ channel inhibitor) or endocannabinoids were shown to depress iPSCs induced by carbachol in CA1-located interneurons, whereas iPSCs mediated by CCK were inhibited by agatoxin (a P/Q-type Ca 2+ channel inhibitor) (Karson et al., 2008).This indicates that CCK has a bidirectional effect on hippocampal GABA transmission.Other studies specified that CCK-2R activation on parvalbumin + basket cells enhanced GABA release, whereas GABA exocytosis by their CCK + counterparts was inhibited by CCK through endocannabinoid-signalling (Foldy et al., 2007).In this context, the cannabinoid receptor 1 (CB1R) is exclusively present on axonal terminals of CCK + , but not parvalbumin + , basket cells, while pyramidal neurons also express dendritic CCK-2Rs.Postsynaptic receptor activation on pyramidal neurons was shown to result in Gα q/11 /PLCβ-mediated endocannabinoid synthesis (Fig. 1), retrograde activation of CB1Rs on presynaptic boutons of CCK + basket cells and, thus, inhibition of GABA release (Lee & Soltesz, 2011;Neu et al., 2007).As such, the activity of CCK + basket cells is inversely correlated with that of both parvalbumin + basket and pyramidal cells in the hippocampus (Dudok et al., 2021).Furthermore, the CCK-mediated gating towards parvalbumin + basket cell activity potentially enhances performance in tasks needing precise timing and integration of information in short time windows (Armstrong & Soltesz, 2012).
CCK-2R agonist was injected 30 min prior to the retrieval trial and 6 h after the initial training session.Interestingly, endogenous CCK immunoreactivity was also greater during the acquisition and retrieval phase, while the hippocampal administration of a CCK-2R antagonist at these time points impaired object memory.In aged Fischer 344 rats, daily systemic injections of CCK-8S prior to behavioural tests did not acutely influence cognitive or motor performance, but enhanced spatial memory retention in the probe trial of the Morris Water Maze (MWM) (Voits et al., 2001).Another study tested memory improvements in i.p. injected rats for 14 days with CCK-8S, with MWM training and the probe trial on days 10 -14 (Reisi et al., 2015).On MWM days, CCK-8S was administered 3 -4 h prior to the trials.Using this administration schedule, CCK-8S had no effect on spatial learning, suggesting that the neuropeptide was given too far away from the trial.Interestingly, combining CCK-1/ 2R dual agonist or, more strongly, CCK-4 injections with scopolamine, a sedative drug, before training sessions impaired food pellet (spatial working) memory in the radial maze (Harro & Oreland, 1993).There is evidence that hippocampal CCK-1R activation promotes, but that of CCK-2R suppresses, olfactory recognition in rats, however (Lemaire et al., 1994a;Lemaire et al., 1994b).This suggests that CCK-4 (CCK-2R agonist) given prior to trials did not impede the learning process in the study of Harro and Oreland (1993), but reduced radial maze performance by inhibiting sense of smell.On the other hand, administration of either a CCK-1/2R agonist or antagonist immediately after every training session mostly had no impact on spatial working memory in the radial maze over a 2 week period, suggesting that memory consolidation was unaffected (Harro & Oreland, 1993).In fact, prolonged peripheral, daily CCK-8S treatment after passive avoidance learning sessions lowered memory on the test day, suggesting that the gut-brain peptide could inhibit the consolidation step in the hippocampus (Sadeghi et al., 2015).
The chemogenetic activation of CCK-GABA neurons in mice resulted in improved hippocampus-associated memory and cognitive function, including greater social and object recognition, contextual fear conditioning, contextual discrimination and problem-solving, but with a mild increase in anxiety (Whissell et al., 2019).As such, enhancing the activity of GABAergic CCK + interneurons facilitates memory formation.On the other hand, loss of CCK-2R-expressing parvalbumin + and CCK + basket cells in hippocampal regions has been associated with learning deficits in a traumatic brain injury rat model (Zhang et al., 2011).Rodent studies have further confirmed that genetic knockout of either CCK (Lo et al., 2008), CCK-1R (Otsuka Long-Evans Tokushima fatty/OLETF rats; also a T2DM model (Kawano et al., 1994)) (Li et al., 2002;Matsushita et al., 2003;Nomoto et al., 1999) or CCK-2R (Sebret et al., 1999) impairs hippocampus-associated memory formation in passive avoidance, object recognition and spatial MWM paradigms.
Collectively, this suggests that the presence of endogenous or administered CCK in the hippocampus, or CCK-2R stimulation close to a training session, enhances declarative learning.This is in agreement with the fact that systemic CCK-8S administration elevates spontaneous glutamate release via CCK-2Rs (Breukel et al., 1997;Migaud et al., 1994) and increases baseline postsynaptic plasticity at 5 -10 min post injection (at least in the dentate gyrus) (Dolatabadi & Reisi, 2014).Additionally, CCK diminishes activity of GABAergic CCK + , but enhances that of parvalbumin + , basket cells.Since plasticity in hippocampal early-and late-born parvalbumin + basket cells mediates knowledge exploitation during learning and memory acquisition, respectively (Donato et al., 2015), basket cell gating by CCK further facilitates learning.Endogenous CCK release appears to enhance hippocampal memory retrieval, but CCK does not affect, or potentially impairs, memory consolidation.More studies are necessary to investigate these aspects.
The neocortical levels of CCK are 10-fold higher than that in the duodenum (Sanders et al., 1982), suggesting that the neuropeptide also plays a memory-regulating role in the neocortex.CCK was shown to be endogenously released in response to N-methyl-D-aspartate receptor (NMDAR) activity in the neocortex (Chen et al., 2019).CCK immunoreactive neurons were identified in the perirhinal and entorhinal cortices, with entorhinal CCK + neurons projecting to the auditory cortex in order to potentiate auditory, and visual cortex-derived, responses (Li et al., 2014;Zhang et al., 2020b).By activating CCK-2Rs, CCK infusion into the neocortex or auditory cortex enhanced longterm potentiation and visuo-auditory associative memory formation, which was inversely impaired after intracortical administration of a CCK-2R antagonist or CCK knockout (Chen et al., 2019;Li et al., 2014).Thus, the CCK system plays a key role in cortical visuo-auditory memory.While the amygdaloid impact of CCK on anxiety and fear memory is not in scope of this review (see e.g.Ballaz & Bourin (2021)), entorhinal CCK+ neurons also project CCK to the lateral amygdala, contributing to auditory (trace) fear memory formation (Feng et al., 2021).Interestingly, CCK-expressing basket cells in the cortex equally coordinate pyramidal neuron activity in the hippocampus (Del Pino et al., 2017).Genetic disruption of cortical CCK + basket cells was demonstrated to negatively affect specific aspects of spatial memory, including novel object recognition, spatial learning and short-term memory organisation.
There are also human studies investigating the cognitive impact of CCK.Peripheral or oral administration of the CCK analogue ceruletide or a CCK-2R antagonist was well tolerated and had no effect on attention, short-term and remote memory (Grasing et al., 1996;Hommer et al., 1985b), but evoked fatigue (Hommer et al., 1985b).Intravenous ceruletide infusions reduced food reward formation memory, but elevated neutral recognition memory (as expected for an anorexigenic peptide) (Pietrowsky et al., 1994), while CCK-4 impaired recall and recognition of words during the administration process (Shlik et al., 1998).Schneider et al demonstrated that intranasal CCK-8S administration promoted automatic (familiarity-based) recognition memory shortly after administration (Schneider et al., 2009), but impaired controlled (recollection-based) recognition memory when administered during the consolidation phase (Schneider et al., 2005).These results are in agreement with animal studies, which show that entorhinal CCK facilitates relay of visuo-auditory responses (i.e.recognition) in the cortex (Chen et al., 2019;Li et al., 2014;Zhang et al., 2020b), whereas the presence of CCK might block hippocampal memory consolidation (Sadeghi et al., 2015).Reflective of improvements in cortical processing, intranasal CCK-8S was reported to transiently increase auditory eventrelated brain potentials (P3 complex) for about 120 min in humans (Denecke et al., 2002;Denecke et al., 2004;Pietrowsky et al., 2001;Pietrowsky et al., 1996).

Pathological impairments in cholecystokinin secretion and hyperphagia
Bile acids are known to inhibit the intestinal release of CCK triggered by ingestion of triglycerides, amino acids or full liquid meals (Gomez et al., 1986;Gomez et al., 1988;Koop et al., 1989).Interestingly, plasma levels of certain bile acid metabolites have been shown to be elevated in MCI and AD patients (Greenberg et al., 2009;Marksteiner et al., 2018;Olazaran et al., 2015).Serum levels of some bile acid metabolites have been associated with amyloid beta (Aβ) 1-42 , total Tau, phospho-Tau (Thr 181 ) in the CSF, glucose metabolism and neuronal degeneration in MCI and AD patients (Nho et al., 2019).Moreover, a higher ratio of circulatory deoxycholic acid, a bacterial product, versus cholic acid, as reduced in AD, was associated with cognitive decline (Mahmou-dianDehkordi et al., 2019;Pan et al., 2017).Bile acids seem to cross the BBB and their deregulation has additional adverse effects in AD, PD and other neurodegenerative diseases (reviewed in Xing et al. (2023)).The latter authors suggested that the bile acid-induced impairment of CCK release, possibly due to intestinal dysbiosis or poor nutrition, might negatively affect CCK levels and disease progression.
N. Reich and C. Hölscher Regarding CCK secretion, dietary changes affect the gut microbiota composition in as little as 24 h (Ley et al., 2006).In addition to inhibitory bile acids, there is in vivo evidence that nutritional changes alter the intestinal composition of bacteria and their secretion of metabolites, e.g.propionate, which increases the enteroendocrine expression of CCK (Zhang et al., 2019).Interestingly, ketones seem to promote intestinal CCK release (Paoli et al., 2015).This indicates that some of the cognitive, and possibly motor, benefits of ketogenic diets in AD and PD (Choi et al., 2021;Lilamand et al., 2021) might be due to peripheral production of CCK and associated benefits.
Of the circulatory CCKs, animal studies suggest that CCK-8, especially the sulphated form, might pass the BBB (Acosta, 1998(Acosta, , 2001;;Dolatabadi & Reisi, 2014;Reisi et al., 2015).Once in the brain, CCK exerts direct protective effects, but also indirect ones, such as through vagus nerve-mediated upregulation of BDNF and NGF expression in the brain (section 5.1).However, CCK-8 levels in the blood stream are low (~10 % of all CCKs (Rehfeld et al., 2001)), suggesting that they only mildly contribute to cerebral CCK levels.On the other hand, due to its protective effects in the gut, CCK indirectly benefits cognitive health by reducing intestinal inflammation and enhancing growth factor secretion.For example, a CCK analogue improved BDNF release in the gut of a PD animal model (Su et al., 2022), while BDNF can exert its neuroprotective effects by crossing the BBB (Bathina & Das, 2015;Pan et al., 1998).It is also well established that intestinal, or microbiota-gut-brain axis, dysfunction is linked to the development of AD and PD (Wang et al., 2021).
Interestingly, there is a pathologic connection between loss of CCK, hyperphagia, obesity and AD.The persistent ingestion of energy-dense foods can affect endocannabinoid-signalling, the gut microbiome, release of their metabolites (e.g.bile acids or propionate) and promote chronic inflammation, thus resulting in downregulation of vagal CCK-1Rs, desensitisation to circulatory CCK, reduced satiety and obesity (reviewed in Cawthon & de La Serre ( 2021)).In rats, supplementation with CCK was shown to reduce meal size and induce weight loss, but these satiating effects only persisted while peptide was administered (Gibbs et al., 1973;West et al., 1984).Inversely, CCK-1R knockout rats display more aggressive food intake and become obese (Bi et al., 2007;Moran et al., 1998), although lack of this receptor can be compensated for by other satiety mechanisms (Kopin et al., 1999).Notably, genetic deletion of CCK-2Rs also results in obesity due to loss of feeding inhibition in the hypothalamus (Clerc et al., 2007;Weiland et al., 2004).Likewise, a high-fat diet was demonstrated to lower cerebral CCK expression in hypothalamic regions in mice, indicating decreased satiety-signalling (Morris et al., 2008).In humans, metabolic syndrome evokes reductions in postprandial CCK secretion (Zwirska-Korczala et al., 2007), while plasma levels of bile acids were found to correlate with a poorer ability to abstain from food and greater body mass indexes (Prinz et al., 2015).Possibly linked to CCK, some AD animal models also show hyperphagia (Knight et al., 2012;Pugh et al., 2007;Vloeberghs et al., 2008), and so do ~18.6 % of AD patients (mostly male subjects) (Shea et al., 2018).Moreover, 3xTgAD mouse model were resistant to the anorexigenic effects of CCK injections and exhibited a lower c-Fos response in the paraventricular hypothalamic nucleus (PVN), which was inversely correlated with meal intake (Adebakin et al., 2012).C-Fos activities in other feeding-regulating regions, including the supraoptic nucleus of the hypothalamus, NTS and area postrema, were not affected, however.In summary, this suggests that poor dietary choices may promote obesity by impairing intestinal CCK secretion, thus reducing satiety-signalling across CCK-1Rs on vagal afferent nerves, and by decreasing hypothalamic CCK expression.Obesity is a clear risk factor for AD (Zhuang et al., 2021;Zuin et al., 2021), but independent of body weight, circulatory and/or cerebral CCK reductions in AD might promote hyperphagia.

Evidence for pathological alterations of cerebral cholecystokinin and its receptors
Recently, it has recently been suggested that CCK levels could serve as a biomarker for AD (Plagman et al., 2019).This correlation with disease progression suggests that CCK-signalling might be part of the mechanisms that drive the disease process.Furthermore, evidence for reductions in the cerebral CCK levels was found in both animal and human AD studies.Aging has been shown to diminish cerebral CCK expression and number of binding sites in the rat hippocampus (Greenstein et al., 1991;Harro & Oreland, 1992).Additionally, the hippocampal CCK mRNA levels in APP/PS1 mice were reported to be halved relative to age-matched wild-type littermates, suggesting that lack of CCK might predispose to neurodegeneration and negatively affects cognition in AD (Liu et al., 2021b).Indeed, dysfunction of hippocampal CCK + interneurons, which are the local source of CCK, was associated with more advanced stages of AD (Reid et al., 2021).Similar to AD animals, post-mortem studies in AD patients have revealed that CCK expression in the cerebral cortex is lower compared to healthy controls (Perry et al., 1981), with a 24 -38 % downregulation of CCK immunoreactivity in some, but not all, cortical regions (Mazurek & Beal, 1991).However, unaltered levels of CCK derivates or CCKR binding in the cerebral cortex of AD patients have also been reported (Lofberg et al., 1996).Importantly, higher CSF levels of CCK were positively correlated with a reduced risk of MCI and AD, better cognitive function and more gray matter volume in several brain areas, such as the posterior cingulate cortex, parahippocampal gyrus and medial prefrontal cortex (Plagman et al., 2019).CSF CCK pools further correlated with Tau and phospho-Tau (Thr 181 ) levels, while greater Tau loads attenuated the memory-improving association of higher CCK levels.In agreement with a beneficial role of CCK, genetic screens in MCI or AD patients have identified a downregulated expression of cerebral CCK-1Rs and hippocampal CCK-2Rs relative to healthy controls (Hokama et al., 2014;Lin et al., 2014), and a link between hippocampal CCK expression and cognition (Liu et al., 2021b).

Neuroprotective effects of cholecystokinin in Alzheimer's disease
Classically, the neuropathology in AD has been summarised as the triad of extracellular Aβ plaques, intracellular neurofibrillary Tau tangles and synapse loss (Breijyeh et al., 2020).This is only a subset of the AD pathology promoting neurodegeneration and cognitive decline, however.Other pathologic events include, for example, oxidative stress, mitochondrial damage, dysfunctional autophagy, loss of hippocampal neurogenesis and BBB and microvascular injury (Reich & Holscher, 2020).Importantly, chronic inflammation likely plays a key role in AD, resulting in cerebral insulin resistance, growth factor desensitisation and glucose hypometabolism (see also Fig. 2) (Holscher, 2019;Neth & Craft, 2017).Such insulin receptor (IRS-1/phosphoinositide 3-kinase (PI3K)/ Akt) pathway defects have been confirmed in the brains of AD patients, which led to the designation of AD as 'type 3 diabetes', even though the mechanism driving insulin desensitisation is different from that in true diabetes (Moloney et al., 2010;Steen et al., 2005;Talbot et al., 2012).The consequences of cerebral insulin resistance in AD patients include decreased brain blood flow and glucose hypometabolism (Drzezga et al., 2003;Hoyer et al., 1988;Lying-Tunell et al., 1981;Mosconi et al., 2008;Ogawa et al., 1994), impaired expression of glucose-metabolising enzymes and enhanced formation of oxidation products (Iwangoff et al., 1980;Zhao et al., 2015).T2DM is also a well-established risk factor for AD (Athanasaki et al., 2022).Therefore, it is intuitive to test drugs that N. Reich and C. Hölscher re-sensitise insulin-signalling in peripheral tissue and the brain.This has been done with success, using the incretin hormones GLP-1 and glucosedependent Insulinotropic polypeptide (GIP), or both, in AD animal models and clinical trials (Reich & Holscher, 2022b).CCK and gastrin also exert a limited incretin, i.e. insulin-releasing, effect (see Rehfeld (2011) and Rehfeld (2019)).Nevertheless, studies employing peptide co-treatment or dual agonists have shown that gastrin potentiates the beneficial effects of other incretins (GLP-1) in diabetes models (Dalboge et al., 2014;Fosgerau et al., 2013;Suarez-Pinzon et al., 2008).As such, CCK could have therapeutic potential in AD.
In vitro, in both hippocampal CCK neurons derived from wild-type and APP/PS1 mice, CCK-8S exposure for multiple weeks promoted the density of dendritic filopodia and spines in a CCK-2R-dependent manner (Zhang et al., 2013).Treatment with CCK-8S also changed membrane parameters, increased firing frequency, facilitated excitatory synaptic transmission and heightened postsynaptic density protein-95 (PSD-95) expression, while decreasing synaptic inhibition.This suggests that CCK can potentially improve the synaptic pathology in AD.There is also evidence that cortical CCK-2R activation protects from glutamate-, kainate or NMDA-induced excitotoxicity (Akaike et al., 1991;Tamura et al., 1992), which is another pathologic event in the cortex (and other brain regions) in AD (Chen et al., 2021;Ong et al., 2013).
In the hippocampus, Gα q/11 -recruiting CCK-2Rs are expressed on excitatory pyramidal neurons and both parvalbumin + and CCK + basket cells (Lee et al., 2011).For the first time, our group has demonstrated that a CCK-2R-binding, unsulphated CCK-8 analogue has neuroprotective effects in an Aβ-based AD animal model.Initially, we demonstrated that a proteolytically resistant, carboxyfluoresceinlabelled CCK analogue crossed the BBB after injection, diffusing into the hippocampus and cortex (Zhang et al., 2023).Similar to in vitro studies (Zhang et al., 2013), hippocampal dendritic spine density, synapse numbers, morphology, various synaptic proteins (microtubuleassociated protein 2, synaptophysin and postsynaptic density protein 95 (PSD-95)) and long-term potentiation were recovered in CCK analoguetreated APP/PS1 mice relative to untreated littermates (Zhang et al., 2023).Synaptic protection resulted in improved spatial learning and memory (Morris Water Maze), working memory (Y Maze) and exploratory behaviour (Hao et al., 2023;Zhang et al., 2023).These procognitive effects of CCK involved a reduction in Aβ 1-42 production and deposition (Hao et al., 2023;Zhang et al., 2023).Moreover, in contrast to untreated APP/PS1 mice, CCK administration normalised the downregulated phosphorylation (activation) of the PI3K/Akt and PKA/cAMP response element-binding protein (CREB) pathways and expression of both BDNF and tyrosine kinase B (TrkB) in the hippocampus (Zhang et al., 2023).

Physiological modulation of the nigrostriatal and mesolimbic dopamine pathways
CCK is involved in modulating dopamine transmission across the movement-inducing nigrostriatal (SN pars compacta (SNpc) to caudateputamen) and reward-regulating mesolimbic (ventral tegmental area (VTA) to nucleus accumbens) pathways (Luo & Huang, 2016).CCK-1Rdependent activation by CCK-8S, but not by non-sulphated CCK-4, has been shown to increase dopamine, aspartate and dynorphin B levels in the SN in vivo (from 10 µM) (You et al., 1996).At higher CCK-8S concentrations (100 µM), dopamine, but also glutamate and GABA, pools were heightened in the neostriatum (caudate, putamen and nucleus accumbens).Vice versa, the nigrostriatal release of CCK was dependent on dopamine (Sierralta & Gysling, 1990), likely mediated by dopamine D2 receptors (D 2 Rs) located on axons and nerve terminals in the caudate-putamen (Ding & Mocchetti, 1992;Meyer & Krauss, 1983).Curiously, CCK-1R knockout mice displayed higher basal dopamine levels in the caudate-putamen, but not nucleus accumbens (Feifel et al., 2003).This could indicate that CCK-1Rs stimulate dopamine production and transmission in the SNpc, but suppress dopamine release in the caudate-putamen.Alternatively, baseline dopamine release might be increased as a compensatory measure due to the lack of CCK-1Rs.
Regarding the mesolimbic pathway, when dopamine-releasing drugs (amphetamine and methamphetamine) were administered, CCK was shown to be expressed in the VTA and projected to the medial nucleus accumbens, with extracellular CCK levels paralleling those of dopamine (Hurd et al., 1992;Martin et al., 2012).As such, likely similar to the nigrostriatal pathway, endogenous CCK production and accumbal neuropeptide release are caused by a rise in dopamine levels (Hurd et al., 1992;Sierralta & Gysling, 1990).Animal studies have further revealed the principal mesolimbic function of CCK-1Rs and CCK-2Rs.CCK-1Rs present in the posterior, but not anterior, part of the nucleus accumbens enhanced stimulus-evoked dopamine release (Marshall et al., 1991).Furthermore, injection of CCK into the nucleus accumbens promoted locomotion and stereotypy in a CCKR-dependent manner, when a dopamine agonist was co-administered (Crawley et al., 1985).Because this effect was not observed when CCK was desulphated or a nonsulphated CCK analogue was employed (Crawley et al., 1985), this suggests that accumbal CCK-1Rs selective for sulphated versions of CCK facilitate reward responses.Congruently, studies employing CCK-1R antagonists demonstrated that systemic or nucleus accumbens-specific inhibition of this receptor counteracts locomotion induced by (chronically administered) amphetamine (DeSousa et al., 1999;Phillips et al., 1993;Wunderlich et al., 2004).On the other hand, amphetamineevoked dopamine release in the neostriatum of mice was blocked following central or peripheral CCK-8S injections via CCK-2Rs, but not CCK-1Rs (Altar & Boyar, 1989).Furthermore, CCK-2R agonism reinforced amphetamine ingestion, indicating that the effect of this dopamine-increasing drug was impaired (Bush et al., 1999).In line with inhibitory properties, CCK-2R antagonism prevented the CCK-induced reduction in dopamine release in the anterior nucleus accumbens following stimulation (Marshall et al., 1991).Pharmacological blockade of CCK-2Rs also potentiated locomotor behaviour triggered by amphetamine (Higgins et al., 1994).Therefore, CCK-1Rs located in the posterior nucleus accumbens augment dopamine-induced reward responses, whereas CCK-2Rs in the anterior nucleus accumbens inhibit them.The ultimate effect of CCK was suggested to be dependent on the magnitude of its release or the presynaptic vs. postsynaptic and regionspecific presence of stimulatory CCK-1Rs vs. inhibitory CCK-2Rs (Ma & Giardino, 2022).Additionally, given that the neuropeptide's sulfation state was e.g.shown to influence gastric acid secretion (Maeda et al., 2000), it matters whether a sulphated (CCK-1&2R-binding) or nonsulphated (only CCK-2R-interacting) form of CCK is released.
Since the mesolimbic CCK release reflects that of, and is seemingly triggered by, dopamine (Hurd et al., 1992;Martin et al., 2012;Sierralta & Gysling, 1990), the physiological purpose of CCK appears to be the termination of dopamine-induced reward behaviour via CCK-2Rs (Ma & Giardino, 2022).This is supported by the fact that CCK-2Rs suppress neostriatal dopamine release, especially when dopamine pools are elevated (Altar & Boyar, 1989).Moreover, VTA neurons were shown to somatodendritically release CCK in response to optogenetic stimulation (Martinez Damonte et al., 2023).This physiological release, or VTA infusion of CCK, encouraged long-term potentiation in GABA synapses in a CCK-2R-dependent manner, lowered local dopamine (Ca 2+ )-signalling and reduced the appetite of mice.
Collectively, particularly at higher neuropeptide concentrations, CCK-1R activation by sulphated CCKs can augment extracellular dopamine levels across the movement-coordinating nigrostriatal pathway.By contrast, baseline dopamine levels are not affected by CCK in the mesolimbic pathway.However, CCK-2Rs in the VTA and nucleus accumbens terminate dopamine release and reward behaviour following a stimulus.Selective expression of CCK-1Rs in the posterior nucleus accumbens may promote reward-associated behaviours, however.This indicates that CCKR-modulating drugs could be promising agents for the treatment of PD, but also other disorders such as schizophrenia or addiction, as reviewed elsewhere (Ballaz, 2017;Ma & Giardino, 2022).

The disease-associated impact of cholecystokinin
Compared to AD, the pathological role of CCK is less prominent in PD.In the 6-hydroxydopamine (6-OHDA) animal model of PD, increased levels of two CCK derivatives were found in the lesioned hemisphere (Nilsson et al., 2009).Similarly, in 6-OHDA-injected rats, CCK expression in the SN was enhanced and further potentiated via l-3,4-dihydroxyphenylalanine (L-DOPA), while L-DOPA administrations elevated striatal CCK levels ipsilateral to the lesion (Taylor et al., 1992).This lesion-induced increase in CCK could be a consequence of feedback loss from dopaminergic neurons, or a protective compensatory response, as suggested for Tau pathology in AD (Plagman et al., 2019).Notably, L-DOPA treatment also heightened CCK levels in the SN of sham-operated animals (Taylor et al., 1992).Because nigrostriatal CCK expression is stimulated by the activation of neuronal dopamine D 2 Rs (X.Z. Ding & Mocchetti, 1992; D. K. Meyer & Krauss, 1983), dopamine replacement therapy could upregulate CCK levels in relevant brain regions in PD patients (at least transiently, before desensitisation to L-DOPA etc. occurs (Beckers et al., 2022)).On the other hand, lack of dopamine impairs striatal CCK production (Sierralta & Gysling, 1990), which could negatively affect disease progression.
Interestingly, it was shown that DJ-1 interacts with Ras-responsive element-binding protein 1 to induce the transcription of CCK from rev response elements and Sp1-binding sites (Yamane et al., 2013).DJ-1 upregulates antioxidant expression, thus detoxifying reactive oxygen species (ROS), induces intracellular dopamine degradation and regulates the immune system, while a loss-of-function mutation in DJ-1 (L166P) leads to early-onset PD and other complications (Liu et al., 2023;Moore et al., 2003;Zhang et al., 2020a).Considering the antiinflammatory actions of CCK (section 5.2), this DJ-1-associated reduction in CCK expression might participate in deregulation of the immune system (Zhang et al., 2020a).
Genetically, in humans, CCK, CCK-1R or CCK-2R polymorphisms do not increase PD risk.However, a genetic variant of CCK (-45C>T), even more so when combined with the CCK-1R 779T>C polymorphism, has been linked to a higher occurrence of visual hallucinations amongst L-DOPA-treated PD patients (Fujii et al., 1999;Goldman et al., 2004;J. Wang et al., 2003).

Protection of dopaminergic neurons by cholecystokinin in Parkinson's disease
PD is characterised by the selective loss of SNpc-located dopaminergic neurons that project to the putamen to control movements (nigrostriatal pathway towards dorsal striatum) (Dickson, 2012).Up to ~30 % of dopaminergic SNpc neurons, ~60 % of their striatal processes and ~70 % of dopamine release in the dorsal striatum are lost when PD motor symptoms emerge (Cheng et al., 2010).The adjacent VTA is also negatively affected, but regresses slower than the SNpc (Alberico et al., 2015).There are numerous non-motor symptoms in PD (see Schapira et al. (2017)), such as apathy, which is related to dysregulation of mesolimbic (and mesocortical) reward pathways emanating from the VTA.A side product of neuronal degeneration may be the intracellular formation of Lewy bodies or Lewy neurites consisting of alpha-synuclein (α-syn) (Dickson, 2012).While all neurons are subjected to the aging process and associated complications, dopaminergic SNpc neurons are specifically vulnerable due to high energy demands, oxidative phosphorylation and concomitant mitochondrial ROS production, catecholamine synthesis (thus increasing danger of reactive dopamine quinone formation), modest ROS and Ca 2+ -buffering capabilities and poorly myelinated axons (Bose & Beal, 2016;Giguère et al., 2019;Pacelli et al., 2015;Sulzer & Surmeier, 2013).Likewise, toxins (i.e.pesticides) and genetic mutations in various genes, including DJ1, leucine-rich-repeat kinase 2, PTEN-induced kinase 1, Parkin, glucocerebrosidase or α-syn, N. Reich and C. Hölscher cause early-or late-onset PD by exacerbating oxidative stress, mitochondrial dysfunction, autophagy defects, protein aggregation and inflammation (Simon et al., 2020).
A root cause of sporadic PD, however, might be metabolic deficiency (Muddapu et al., 2020).Responsible for such an impairment is the development of cerebral insulin resistance, as evident in the SNpc, basal ganglia and other brain regions in PD patients, with a loss of insulin receptor expression and downstream pathway inactivation by proinflammatory cytokines (Fig. 2) (Holscher, 2019;Moroo et al., 1994;Morris et al., 2014;Takahashi et al., 1996;Tong et al., 2009).Additionally, expression of glucose-metabolising enzymes, i.e. those of the pentose phosphate pathway, are downregulated at an early stage during PD, which aggravates depletion of antioxidants and promotes glucose hypometabolism in the dorsal striatum (Dunn et al., 2014;Xu et al., 2015).Similar to AD, T2DM is also a risk factor for PD (Hassan et al., 2020).Insulin re-sensitising GLP-1 agonists have shown great promise in treating PD in animal models and human clinical trials (reviewed in Reich & Holscher (2022b)), suggesting that CCK, like GLP-1, could also be used therapeutically.
We have shown that CCK has therapeutic potential in PD animal models.Biodistribution studies confirmed that an unsulphated CCK-8 analogue, which only binds CCK-2Rs, translocated into the SN and striatum following i.p. injection, suggesting that key brain regions were reached (Zhang et al., 2022).In MPTP-administered mice, 14 days of CCK analogue treatment restored the levels of the mitochondrial biogenesis marker PGC-1α and its translation product, the fusionmodulator Mfn2, while preventing deformation and loss of mitochondria in the SNpc (Fig. 1) (Zhang et al., 2022).Use of the CCK analogue also prevented endoplasmic reticulum (ER) stress (downregulation of inositol-requiring transmembrane kinase/endoribonuclease 1α/IRE1α expression), the MPTP-evoked abnormal increase in the macroautophagy markers autophagy-related 7 and Beclin 1 and the accumulation of damaged, autophagosome-engulfed mitochondria in the SNpc, further supporting that mitochondria were protected.Moreover, CCK mimetic administration blunted the MPTP-induced reduction in CREB phosphorylation and prevented the pro-apoptotic increase in the Bcl-2associated X protein (Bax)/B-cell lymphoma 2 (Bcl-2) ratio.Microglial ionized calcium-binding adaptor molecule 1 (IBA-1) and astrocyte glial fibrillary acidic protein (GFAP) immunoreactivity in the SNpc were also reduced by the CCK analogue in MPTP-treated mice.Jointly, these antiinflammatory, anti-apoptotic and mitoprotective effects of CCK resulted in reduced synaptic degeneration and death of tyrosine hydroxylasepositive dopamine neurons in the SNpc, leading to improved motor function and exploratory behaviour.The underlying neuroprotective signalling pathways of CCK in the SNpc were likely mediated by local CCK-2Rs and Gα q/11 -signalling in our study (see Fig. 1), since the CCK-8 analogue was unsulphated and, thus, had low affinity for CCK-1Rs (Zhang et al., 2022).As such, the protective effects of the CCK-8 analogue in the MPTP mouse model parallel those seen in the hippocampus of APP/PS1 mice (Fig. 1) (Hao et al., 2023;Zhang et al., 2023).
However, CCK-1Rs could play an indirect role in neuroprotection by modulating dopamine release.The study of Hill et al (1990) suggests that CCK-1Rs are expressed at high densities in the SNpc (and VTA), with higher densities in primates compared to rodents, whereas low or no expression of CCK-2Rs was detected in the SNpc (more pronounced in the striatum) (Hill et al., 1987;Hill et al., 1990;Honda et al., 1993;Ito et al., 1993;Moran et al., 1986;Van Dijk et al., 1984).CCK-1R, but not CCK-2R, activation of nigral neurons leads to their depolarisation and local dopamine release (Wu & Wang, 1994b;You et al., 1996).Treatment with higher CCK-8S concentrations also stimulated dopamine release across the neostriatum in vivo (You et al., 1996).Therefore, low CCK concentrations induce CCK-1R-mediated somatodendritic dopamine release by SNpc neurons, whereas higher doses of CCK trigger axonal dopamine release in striatal regions.Mechanistically, similar to the CCK-2R/Gα q/11 -mediated increase in excitability in some hippocampal subregions (Fig. 1) (Gronier & Debonnel, 1995;Wilke et al., 2014), CCK-1R activation appears to induce Gα q/11 -signalling, cationic currents (TASK-1/3 opening) and enhanced excitability in SN dopaminergic neurons (at least in vitro) (Wu & Wang, 1994a).It has been suggested that dopamine release by SNpc neurons is impeded by inhibition of vesicular monoamine transporter 2 (VMAT2; responsible for dopamine-packing into vesicles), while N-type and P/Q-type Ca 2+ channel inhibition blocks the axonal liberation of dopamine (Rice & Patel, 2015;Sulzer et al., 2016).Because VMAT2 activity and N-type Ca 2+ channel opening seem to be decreased by Gα q/11 -signaling (Holtje et al., 2003;Zamponi & Currie, 2013), this might explain why lower CCK concentrations induce local somatodendritic dopamine release in the SNpc, but not axonal release in the striatum.Depending on the conditions, CCK-1Rs can recruit a range of G proteins, including Gα q/11 , Gα s , Gα i , and Gα 13 (Ding et al., 2022;Liu et al., 2021a;Zhang et al., 2021b).Because high extracellular CCK levels elicit axonal dopamine release in the striatum in vivo (You et al., 1996), there could be a CCK concentration-dependent, or temporally regulated, switch in G proteins downstream of the CCK-1R in SNpc dopaminergic neurons.
Importantly, CCK-evoked somatodendritic dopamine release by SNpc neurons will stimulate local dopamine DRs and, thus, influence downstream signalling and neuroprotection post CCKR activation.In the striatum, dopamine D 1 Rs and D 2 Rs are almost exclusively present on separate cells, with striatonigral neurons predominantly expressing dopamine D 1 Rs (Gerfen, 2022).In vivo, 6-OHDA injections were shown to reduce dopamine D 1 R levels in the SNpc, which was reversible with repeated administration of a dopamine D 1 R agonist (Gerfen et al., 1990), suggesting that CCK could do the same by stimulating endogenous, somatodendritic dopamine release (You et al., 1996).By contrast, in PD patients, the expression of dopamine D 1 Rs was shown to be upregulated amongst other cell survival-associated proteins in SNpc neurons (Simunovic et al., 2009).Dopamine D 1 Rs are Gα olf -coupled and result in induction of PKA and CREB (Jones-Tabah et al., 2021).Activation of the latter transcription factor is neuroprotective and e.g.counteracts 6-OHDA-induced oxidative stress in vitro and in the SNpc in vivo (Kim et al., 2020).While mostly present in striatopallidal neurons (Gerfen, 2022), dopamine D 2 R mRNA has also been detected in the SN (Le Moine et al., 1990).Dopamine D 2 Rs are usually G i -coupled and inhibit cAMP generation (Gerfen, 2022).Interestingly, in SN neurons, dopamine D 2 Rs were shown to form heterodimers with the receptor for acylated ghrelin (growth hormone secretagogue receptor 1 alpha; GHS-R1α), changing downstream signalling (Tang et al., 2023).Activation of these receptor heterodimers by quinpirole (dopamine D2/3R agonist) protected from MPTP in vitro and in vivo, and enhanced TH and VMAT2 levels as well as CREB phosphorylation above baseline even in the SNpc of MPTPuntreated rats.Therefore, CCK-8S-induced dopamine release could be neuroprotective by activating dopamine D 1 Rs and D 2 Rs in the SNpc.Striatal dopamine release triggered by high CCK-8S doses could also acutely enhance motor function (You et al., 1996).Further studies are necessary to clarify the signalling mechanisms of nigral CCK-1Rs, receptor-driven dopamine release and the putative neuroprotective effects of dopamine D 1/2 R activation, however.
Related to PD, methamphetamine use greatly increases the risk for developing Parkinsonism, resulting in oxidative and ER stress, neuroinflammation, excitotoxicity and degeneration of dopaminergic and other neurons (see Jayanthi et al. (2021)).In a respective mouse model, pre-treatment with CCK-8 decreased the methamphetamine-induced loss of tyrosine hydroxylase expression in the striatum and SN as well as that of dopamine transporters (striatum only) (Gou et al., 2015).This provides further evidence that CCK protects dopaminergic neurons from various toxic events.
A small trial tested the immediate impact of a single intranasal CCK-8 administration on motor function in PD patients, which had no effect (Smolnik et al., 2002).Curiously, acute intranasal CCK improved auditory brain potential (P3 complex; a measure of attention and information processing) in healthy controls, but impaired it in PD patients (Smolnik et al., 2002).Other studies support that intranasal CCK-8S N. Reich and C. Hölscher transiently enhances the P3 complex in normal adults (Denecke et al., 2002;Denecke et al., 2004;Pietrowsky et al., 2001;Pietrowsky et al., 1996).These attentional deficits specifically in CCK-8S-treated PD patients might be the consequence of CCK-induced GABA release in the striatum (Acosta, 1998;You et al., 1996), but without stimulation of the already lost dopaminergic SN/VTA neurons in PD.

Improvement of the gut pathology
PD patients show severe intestinal damage, including permeabilisation, changes in the gut microbiota, degeneration of enteric neurons and Lewy body formation in their nerve terminals (Scheperjans et al., 2018).These events are causative for the typical gastrointestinal symptoms in PD, usually preceding motor symptoms (Warnecke et al., 2022).The Intestinal pathology also contributes to that in the brain, for example by triggering systemic inflammation or via propagation of alpha synuclein (α-syn) seeds from the enteric nervous system and across the vagus nerve to the CNS (Klann et al., 2021;Rani & Mondal, 2021).
Besides the brain, CCK also ameliorates the gastrointestinal pathology in MPTP or human A53T α-syn transgenic PD mouse models.The 5 week-long administration of either a non-sulphated CCK-8 analogue or liraglutide was shown to prevent loss of the tight junction proteins occludin and zonula occludens-1 (ZO-1) and reduced the levels of the stress enzyme inducible nitric oxide synthase and pro-inflammatory cytokine TNF-α (Su et al., 2022).Moreover, administration of either of the peptides lowered the density of mono-and oligomeric α-syn species, while elevating TH levels, as indicative of decreased degeneration of enteric nervous system dopaminergic neurons, in the myenteric plexus (distal colon).The colonic reduction in BDNF levels in the PD mouse models was also partially prevented by neuropeptide injections.This restoration of gastrointestinal BDNF expression, which is impaired by inflammation, likely accounted for most of the protective effects of CCK and liraglutide, including the observed improvements in intestinal permeability, bowel and motor function and survival of dopaminergic neurons (Ahn et al., 2021;Li et al., 2018).An anti-inflammatory effect is supported in the lipopolysaccharide-induced endotoxemia rat model, where injection of CCK-8S preserved gut integrity, prevented proinflammatory cytokine production and decreased bacterial invasion (Saia et al., 2020).CCK likely suppresses intestinal inflammation by acting on CCK-2R-expressing peripheral immune cells (Iwata et al., 1996;Schmitz et al., 2001b), while sulphated peptide variants may additionally stimulate the vagus nerve (details in section 5.2) (Bozkurt et al., 2003;Luyer et al., 2005).Notably, in the mouse colon, BDNF transcription has been identified in epithelial cells and myenteric (GABAergic or dopaminergic) neurons (Lommatzsch et al., 1999;Lucini et al., 2002;Rao & Gershon, 2018).Our study in PD animal models employed a non-sulphated CCK-8 mimetic, suggesting that intestinal BDNF expression was restored by downregulating inflammation (Su et al., 2022).However, myenteric neurons express CCK-1Rs, indicating that sulphated CCKs might show additional benefits (Sternini et al., 1999).

Enhancement of nerve growth factor and brain-derived neurotrophic factor expression
BDNF is expressed in many tissues, including the brain and intestines, and acts as a protective growth factor, modulator of neurotransmission, plasticity and memory formation, neurogenesis-enhancer and more (see Bathina & Das (2015)).The neuroprotective effects of BDNF are exerted by binding to TrkB receptors, activating IRS 1/2 /PI3K/ Akt, src homology collagen peptide (Shc)/growth factor receptorbinding protein 2 (Grb2)/rat sarcoma (Ras)/rapidly accelerated fibrosarcoma (Raf)/mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase 1/2 (ERK 1/2 ) and PLCβ-mediated IP3 and PKC signalling cascades.Stimulation of TrkB further leads to plasticity, stress resistance and survival gene expression through activation of CREB.BDNF expression is impaired in the hippocampus in AD, contributing to plasticity defects and memory decline (Banerjee & Shenoy, 2023), and the nigrostriatal pathway in PD, promoting neuronal atrophy (Palasz et al., 2020).Enhancers of BDNF expression have been recognised as a potential therapeutic option for AD and PD (Banerjee & Shenoy, 2023;Palasz et al., 2020).
As BDNF, pro-NGF and its more mature form NGF are present in booth peripheral and cerebral tissues, both mediating regeneration and survival-signalling across TrkA receptors via the same neuroprotective downstream pathways as TrkB (Marlin & Li, 2015;Mitra et al., 2019).However, defects in cortical and hippocampal pro-NGF release as well as extracellular processing to NGF by plasmin in the forebrain negatively affect survival of local cholinergic neurons and cognition in AD (reviewed in Mitra et al. (2019)).This involves loss of principal TrkA receptor expression, pathologic binding of these receptors by Aβ (which also induces inflammation), lowered (apoptosis-provoking) ratios of TrkA vs. p75 neurotrophin receptors (p75NTRs) and poor retrograde NGF transport from axons to the nucleus in cholinergic neurons.Interestingly, peripheral NGF production was shown to be elevated in MCI and AD patients (Shen et al., 2019), whereas pro-NGF generation was impaired, but NGF elimination enhanced, in cortical regions (Pentz et al., 2021).NGF seems to be less involved in the cerebral pathology of PD compared to AD, although some studies have shown reduced pro-NGF and NGF levels in the serum of PD patients (Lorigados Pedre et al., 2002;Xu et al., 2018).NGF delivery or upregulation strategies in the brain and NGF mimetics have been tested as treatment options for AD (see Mitra et al. (2019)).
CCK also affects cerebral NGF expression.Peripheral CCK-8S injections transiently elevated NGF transcription and concentrations in the hypothalamus and pituitary (3-fold increased), hippocampus (+ 50 -60 %) and septum, but not cortex, 15 -60 min post administration or following prolonged treatment from 3 days (Tirassa et al., 1999;Tirassa & Costa, 2007;Tirassa et al., 1998).Interestingly, CCK-1Rs were responsible for elevating NGF expression in the hypothalamus and pituitary and CCK-2Rs in the hippocampus, with vagus nerve inhibition by atropine reducing NGF expression in all brain regions (Tirassa et al., 1998).Because NGF expression in the CNS (cortex) was found to be dependent on promotor binding by both CREB and CCAAT/enhancerbinding protein delta (C/EBPdelta) (Fig. 1) (McCauslin et al., 2006), this suggests that CCK-evoked CREB expression elicits that of NGF in the same brain region.Interestingly, CCK-8S administrations were shown to reduce expression of TrkA and p75NTR, but elevate that of TrkB, in the hippocampus and septum (Tirassa & Costa, 2007).Thus, relevant in the context of AD, CCK modifies hippocampal neurotrophin receptors and weakens NGF-and apoptosis-, but potentiates BDNF-, signalling.
CCK exerts direct immunosuppressive effects in the body.Peripheral immune cells, including human myeloid cells (e.g.THP-1 monocytes), T and B lymphocytes and mononuclear blood cells isolated from patients, were shown to express CCK-2Rs (Iwata et al., 1996;Schmitz et al., 2001b).Lymphocyte exposure to CCK-8S inhibited mitogen-induced proliferation and mobility, while CCK-8S decreased the expression of pro-inflammatory cytokines (TNF-α) via CCKR activation on THP-1 cells in vitro (De la Fuente et al., 1998;Miyamoto et al., 2012).Administration of the neuropeptide also prevented systemic and peripheral tissue (kidney, lung, spleen, gut) as well as immune cell inflammation in a T1DM mouse model or following lipopolysaccharide injection in vivo (Meng et al., 2002;Miyamoto et al., 2012) (Saia et al., 2020).A fundamental exception is the pancreas, however.In this organ, CCK evokes inflammation by overstimulating local acinar cells and stellate cells, which can induce an inflammatory response (see Smith & Solomon (2014) and section 6.2).As such, CCK has been used to induce a model of pancreatitis (Gorelick & Thrower, 2009;Kim, 2008).
CCK also stimulates anti-inflammatory signalling across the vagus nerve.A recent drug screen identified both CCK and dopamine agonists as intestinal anti-inflammatory agents in a zebrafish enterocolitis model.
Reversely, use of CCK receptor antagonists aggravated colitis (Oehlers et al., 2017).Interestingly, CCK-1Rs are present on vagal afferent neurons projecting from the gut, in particular the duodenum, to the brain (Peters et al., 2006).CCK-8S treatment was demonstrated to quench colonic inflammation in a CCK-1R-dependent manner, but only when vagal afferent nerves were intact (Bozkurt et al., 2003).Additionally, electrical, anti-inflammatory vagus nerve activation, which involved participation of the hypothalamic-pituitary-adrenal (HPA) axis, was inhibited by cerebral administration of a CCK-2R antagonist (Gulpinar et al., 2004).There are three anti-inflammatory vagus nerve pathways (reviewed in Bonaz et al. (2016)), including the intestine-associated HPA axis-cortisol, cholinergic and splenic sympathetic pathways.In this context, food (fat) intake stimulates CCK release (Liddle, 1997), and it was shown that the enteral infusion of dietary fat prior to haemorrhagic shock prevented systemic pro-inflammatory cytokine production and subsequent intestinal permeabilization (Luyer et al., 2005).This anti-inflammatory effect was abolished when CCK-1R, CCK-2R or nicotinic receptor antagonists were co-administered, however (Luyer et al., 2005).As such, this suggests that intestinal CCK-1R-signalling across vagal afferent fibres results in reciprocal, vagal efferent nerve-mediated activation of enteric neurons, acetylcholine release and antiinflammatory α-7-nicotinic acetylcholine receptor activation on macrophages (Bonaz et al., 2016).On the other hand, plasma corticosterone (rodent variant of cortisol) levels were heightened by haemorrhagic shock, but unaltered by co-injection of CCKR antagonists (Luyer et al., 2005).This indicates that CCKR-signalling does not induce the antiinflammatory HPA axis-cortisol pathway, but CCK-2Rs may be involved in its execution (Gulpinar et al., 2004;Luyer et al., 2005).
Other studies also shed light on the anti-inflammatory mechanisms of CCK, which may combat the development of insulin resistance in the brain (Fig. 2).The expression of CCK-2R was confirmed in N9 murine microglial cells (Gou et al., 2020).Moreover, CCK-8S treatment prevented methamphetamine-induced inflammation, such as reduced activation of the pro-inflammatory transcription factor NF-κB, both in vitro and in vivo.Similar to microglia, the presence of functional CCK-2Rs has been confirmed in primary mouse and rat astrocytes (W.Muller et al., 1997).Furthermore, astrocyte CCK-signalling was shown to play a bidirectional role in hippocampal neurogenesis in vivo (Asrican et al., 2020).The latter group showed that CCK released by local interneurons stimulates glutamatergic input from astrocytes to neural stem cells in the dentate gyrus, leading to progenitor neuron proliferation.Inversely, the depletion of CCK induced astrogliosis (increased GFAP immunoreactivity) and activated pro-inflammatory genes, with inflammation known to inhibit neural stem cell proliferation (Asrican et al., 2020;Ekdahl et al., 2003).A direct anti-inflammatory effect is supported, since microglial IBA-1 and astrocyte GFAP immunoreactivity were reduced in CCK-8 analogue-treated MPTP mice (Zhang et al., 2022).CCK might also quench inflammation indirectly, by reducing neuronal apoptosis and the associated release of DAMPs, such as MMP3 derived from dopaminergic neurons (Choi et al., 2008).
In summary, CCK directly quenches systemic inflammation by activating CCK-2Rs on peripheral immune cells, microglia and astrocytes, while stimulating the cholinergic anti-inflammatory vagus nerve pathway via CCK-1Rs.By contrast, in the pancreas, CCK evokes inflammation.

Side effects
There are some caveats regarding the therapeutic use of CCK.Representative of the difficulties CCK mimetics face, pharmaceutical companies have previously developed CCK-1R agonists for the treatment of obesity (see L. J. Miller & Desai (2016) and Miller et al. (2021)).Such CCK-1R analogues were tested in clinical trials, but have not been approved due to insufficient effects and, crucially, a range of side effects.
The latter include acceleration of cancer growth (see next section), nausea, diarrhea, stomach cramping/peptic ulcers and gallbladder complications (cholelithiasis) (Rehfeld, 2019).Furthermore, in most studies, CCK analogues were administered through parenteral or intracerebral routes.While orally available CCKR agonists have been reported, e.g. by Elliott et al. (2010), injection will likely be necessary to achieve significant therapeutic effects.

Pancreatitis and cancer
Another caveat is the known association between CCK, pancreatitis and, consequentially, a greater risk of pancreatic cancer (Smith & Solomon, 2014).In fact, supraphysiological CCK concentrations (at least 10-fold greater than following meal intake) and the CCK-1R/CCK-2R agonist cerulein have been used to generate pancreatitis models in vitro and in vivo (Gorelick & Thrower, 2009;Kim, 2008).In the pancreas, the collective evidence suggests that CCK-1Rs are dominant in humans, whereas rodents mainly express CCK-2Rs (reviewed in Smith & Solomon (2014)).However, both receptors are involved in exerting different pathologic responses.Activation of CCK-1Rs on pancreatic acinar cells is presumably linked to digestive enzyme secretion and, thus, driving inflammation (pancreatitis), while stimulation of CCK-2Rs drives hyperplasia.Moreover, stimulation of stellate cells by CCK results in collagen production, fibrosis and an acetylcholine-mediated augmentation of enzyme secretion by acinar cells.However, it has been suggested that high dose CCK administrations are not sufficient to induce (pancreatic) cancer, and that an additional pathologic factor must be present, for example co-exposure to carcinogens, a high fat diet, obesity, other inflammatory diseases or genetic predispositions.
From a therapeutic standpoint, CCK-2R agonists are preferable for long-term treatment, given that peripheral CCK-1Rs are linked to a range of other adverse effects (see section 6.1).Use of a CCK-2R agonist will likely diminish the side effect profile relative to a CCK-1R/CCK-2Rco-stimulating CCK mimetic in the pancreas, but the higher expression of CCK-2R in humans might lead to unforeseen complications (Smith & Solomon, 2014).Besides the pancreas, adenocarcinomas in the stomach, liver, rectum, colon, oesophagus, lung and medullary thyroid were shown to overexpress CCK-1Rs and/or CCK-2Rs, with CCK and gastrin aggravating pathogenesis (compiled in Zeng et al. (2020)).Whether long-term administration of CCK-2R mimetics is safe in elderly AD and PD patients remains to be determined, and concomitant cancer screens might be necessary.

Panic and anxiety
By activating central CCK-2Rs (Bradwejn et al., 1994), human studies have shown that exogenous CCK-4 has panicogenic effects in humans (Bradwejn, 1993;Bradwejn et al., 1990;de Montigny, 1989).In fact, intravenous CCK-4 injections have been validated as a human panic model (see also Rehfeld (2021)) (Eser et al., 2007;Goettel et al., 2023).The intensity of such panic attacks is dose-dependent.Moreover, as Rehfeld describes, panic lasts about 20 -28 min following CCK-4 administration and includes 'intense anxiety, with a fear of dying, and a strange sense of the world sliding away, accompanied by palpitations, sweating, and faintness' (Rehfeld, 2021).Some studies suggest that patients with various disorders (social phobia, major depression, anxiety, panic and obsessive compulsive disorder) are more susceptible to the panicogenic effects of CCK-4 and its analogue pentagastrin compared to healthy individuals (Brawman-Mintzer et al., 1997;deLeeuw et al., 1996;Koszycki et al., 2004;McCann et al., 1997;van Vliet et al., 1997).The cause of this enhanced sensitivity is not fully understood, and could be due to greater CCK-2R sensitivity and/or a synergistic effect with pre-existing anxiety in these patients.In the context of panic disorder, genetic studies have further shown that CCK polymorphisms may facilitate panic (Garvey et al., 1998;Hansen et al., 2000;Hosing et al., 2004;Wang et al., 1998), but some can also be protective (Koefoed et al., 2010).Furthermore, an aggravating (Hosing et al., 2004) impact of genetic CCK-2R variants on panic disorder has been reported in one study, but not in another (Hattori et al., 2001).
The CCK system has not only been associated with panic, but also with anxiety.Unlike CCK-4, there does not seem to be any evidence for panicogenic effects of CCK-8.However, CCK-8S, or its mimetic caerulein, may trigger anxiety in vivo (Koks et al., 2000;Pérez de la Mora et al., 2007;Rex et al., 1994).An in vivo study administering cerulein showed that anxiety only materialised when rats were pre-stressed, suggesting that these anxiogenic effects of the CCK mimetic synergise with other sources of anxiety (Koks et al., 2000).Overexpression and knockout studies suggest that, at least in the forebrain, the magnitude of CCK-2R activation correlates with the expression of anxiety in rodents (Chen et al., 2006;Raud et al., 2005).
Interestingly, when directly compared in vivo, CCK-8S is less anxiogenic than CCK-4 (Rex et al., 1994).It was hypothesised that CCK-4 is more potent due to exposition of its N-terminal tryptophan residue, possibly allowing binding of additional receptors, such as GPR142 on pancreatic islets (Rehfeld, 2021).Additionally, sulphated CCK-8 mimetics might be less anxiogenic than unsulphated CCK-8 variants due to CCK-1R activation.The available evidence supports that CCK-1Rs are anxiolytic (compiled in Zwanzger et al. (2012)), while a CCK-1R haplotype has been associated with reduced risk of panic disorder in women (Koefoed et al., 2010).Activation of CCK-1Rs has further been linked to enhanced reward and lessened anxiety-like behaviour following drug withdrawal (see Ma & Giardino (2022)).
Regarding anxiety-associated brain regions, CCK is strongly expressed in various subregions of the amygdala (a key player in fear expression and memory, and modulated by gut-brain peptides (Reich & Holscher, 2022a)), the hippocampus and cerebral cortex.CCK is also present at notable levels in the hypothalamus, thalamus and olfactory bulb.Similarly, CCK-2Rs are expressed in these brain areas, and multiple CCK projections between regions are involved (mapped in Zwanzger et al., (2012)).
Collectively, the evidence suggests that, unlike CCK-4, CCK-8 analogues (especially if sulphated) are less likely to trigger panic, and rather elicit anxiety.However, these dose-dependent anxiogenic effects may still interfere with the clinical utility of (unsulphated) CCK-2R analogues, given that an estimated 40 % of AD and 20 -50+ % of PD patients already exhibit symptoms of anxiety (J.J. Chen & Marsh, 2014;Mendez, 2021).
N. Reich and C. Hölscher

Sex differences
Some studies have shown gender-specific differences in the CCK system.Comparison of both sexes has shown that CCK (intestinal) and CCK-1R (pancreatic) transcription shows an age-dependent decline in male, but not in female rats (Miyasaka et al., 1995).In ovariectomised animals, estradiol injections were shown to increase the number of CCK-1R binding sites in the pancreas (Geary et al., 1996), suggesting that sex hormones affect the CCK system.As such, considering that CCK-1R activation on acinar cells has been linked to pancreatitis (Smith & Solomon, 2014), susceptibility to this adverse effect could differ between male and female subjects of varying age.
It has also been reported that only male lean and obese Zucker rats, but not their respective female littermates, showed increased food intake in response to administration of a CCK-1R antagonist (Strohmayer & Greenberg, 1996).This sex difference is likely also explained by the higher circulatory levels of estrogens in females, which enable and potentiate CCK-mediated satiation in the estrus, but not in other phases of the ovarian cycle (reviewed in Geary (2001)).Estrogens also modulate female reproductive behaviour via the CCK and endogenous opioid systems, involving estrogen-driven CCK expression in the limbichypothalamic lordosis circuit (see Micevych & Sinchak (2001)) (Holland et al., 1998).
It is not known whether such gender-associated differences in e.g.estradiol levels affect the cognitive or neuroprotective effects of CCK.It must be noted that our studies testing a non-sulphated CCK-8 analogue in AD and PD models employed only male mice (Hao et al., 2023;Z. Zhang et al., 2022;Z. Zhang et al., 2023).As such, it cannot be ruled out that female rodents, possibly in an ovarian cycle-associated manner, show a different response to exogenous CCK in the brain.

Combination of growth factors as a therapeutic opportunity
Interestingly, the protective effects of the CCK analogue in the brain and colon of AD or PD animal models were on par with those of liraglutide (Hao et al., 2023;Su et al., 2022;Zhang et al., 2022).The latter is a GLP-1 receptor agonist activating the neuroprotective PI3K/Akt, ERK 1/2 and cAMP/PKA pathways, with robust neuroprotective effects in AD/PD in vivo models and in clinical trials (reviewed in Reich & Holscher (2022b)).In T2DM animal models, co-administration of CCK and the GLP-1 analogue exendin-4 was more effective in ameliorating weight loss and T2DM-associated plasma glucose deregulation (Irwin et al., 2013;Trevaskis et al., 2015).Moreover, a GLP-1/gastrin (CCK-2Rbinding) dual agonist has shown promising results in db/db (leptin receptor-deficient) mice (Dalboge et al., 2014;Fosgerau et al., 2013).CCK also synergises with leptin, whose receptor is co-expressed on vagal afferent nerves (at least in the absence of food (Burdyga et al., 2002;Buyse et al., 2001)), in reducing gastric emptying, food intake and inflammation (Bozkurt et al., 2003;Peters et al., 2006).Importantly, GLP-1/GIP receptor dual agonists have demonstrated superior neuroprotective effects in AD and PD animal models compared to single receptor agonists such as liraglutide (P.Feng et al., 2018;Lv et al., 2021;Maskery et al., 2020;Salles et al., 2020;Yuan et al., 2017;Zhang et al., 2021a).This suggests that a combination of incretins and related growth factors released following food intake or during fasting, i.e. in form of multi-receptor agonists, are a promising therapeutic strategy for AD and PD (Reich & Holscher, 2022b).Such combinations can be beneficial in various ways.First, neuroprotective pathways can be synergistically activated to achieve a greater effect.For example, co-activation of hippocampal or nigral CCK-2Rs and GHS-1Rα (ghrelin receptor) could theoretically potentiate neuroprotective Gα q/11 -signalling, AMPK/PGC-1α activation and autophagy (Lee et al., 2011;Reich & Holscher, 2020).This requires that there are no counterregulatory effects, however.Our own observations support that the activation of opposing pathways in the same neurons, e.g.mTORC1-stimulating Akt vs. mTORC1-inhibiting AMPK activation (Saxton & Sabatini, 2017), should be avoided.Second, dependent on receptor distributions in the brain, a wider range of neurons and brain regions may be covered, if multiple GPCR agonists are utilised.Third, side effects could potentially be reduced by using lower doses (due to synergistic effects), or possibly by employing growth factors that reciprocally cancel adverse effects.An example could be the possible neutralisation of the anorexigenic and glucose-lowering effects of CCK (Pathak et al., 2018) by the appetite-and hyperglycemiastimulating properties of co-administered ghrelin (Reich & Holscher, 2020).

Conclusion
CCK is a multi-functional peptide hormone.Impaired release and/or expression of CCK, CCK-1Rs and CCK-2Rs could contribute to neurodegenerative processes in AD, while deregulation of the CCK system seems to be involved in visual hallucinations in PD.Brain-penetrant forms, in particular long-lasting synthetic CCK8 analogues, bind to neuronal CCK-2Rs in the hippocampus (AD) and SNpc (PD), resulting in AMPK activation and neuroprotection.This includes synaptic protection, restoration of the mitochondrial fusion/fission dynamics, autophagy and a reduction in systemic and cerebral inflammation.Similar to peptide hormones such as GLP-1, the anti-inflammatory effects of glial CCK-2R activation could prevent insulin resistance in neurons, a key pathologic feature in AD and PD.Sulphated CCK-8 likely offers additional therapeutic benefits by inducing CCK-1Rs, but at the cost of long-term side effects such as gallstones.Improved weight loss, glucose regulation and intestinal protection indirectly benefit AD and PD.Combinatorial use of CCK-8 and other growth factors, such as acylated ghrelin, could yield synergistic therapeutic benefits.Although CCK may stimulate pancreatitis, tumour growth, anxiety and panic attacks, studies testing CCK-8 in patients with neurodegenerative disorders are warranted.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.