The alarming increase of the aging population worldwide brings a significant increase in numerous chronic and degenerative diseases, causing substantial socio-economic burden (Ferrucci et al., 2008). In fact, Europe is expected to have the oldest populations in the world in the twenty-first century, with almost one in four Europeans projected to be aged 65 years or older by 2030 (Ferrucci et al., 2008). Because of the gradual increase in aging, there is a rise in age-linked diseases that were less prevalent in the past. Some of the most prevalent central nervous system (CNS) diseases related to aging consistent of progressive neurodegenerative cognitive impairment ultimately leading to Alzheimer’s disease (AD), with incidence of 10% in humans older than 65 years (Alzheimer's Association Report, 2023).

Even normal, healthy aging is associated with a decline of attention and cognitive performance (Glisky, 2007) leading to increasing deficits in learning and memory. However, there is a large individual variability in the onset and speed of cognitive loss: While many senior citizens still enjoy their cognitive abilities at an advanced age, others may show signs of cognitive impairment and increased risk of neurodegenerative diseases like Alzheimer’s or Parkinsons’s disease early in their life. The onset and progression of age-related decline in brain functions also differs strongly between the cognitive domains in humans (Glisky, 2007) and also in rodents (Bilkei-Gorzo et al., 2012). The reason for this large variance is partly the differences in the ability to recruit additional brain areas for task solving, which can partially counterbalance the effect of functional decline (Eyler et al., 2011). Another factor, which may contribute to the differences in the effect of aging on cognitive functions, is the large variance in the sensitivity to age-related changes between brain areas andneuronal types (Double et al., 2010; Small et al., 2012).

There are a wide range of molecular and histological changes in the aging brain, including increased levels of toxic metabolic end products, misfolded macromolecules and non-functional organelles, increased pro-inflammatory activity of glial cells and depletion of catecholaminergic neurons. Although these changes are known to underlie age-related functional deficits, it is not known whether and at what level these individual factors influence cognitive aging. Identification of the critical factors would be key for targeted therapy of cognitive deficits in aging. One of the most sensitive neuronal populations to aging is the noradrenergic neurons in the locus coeruleus., therefore in the present work we focused on this brain area. Particularly, on the potential consequences of the loss of noradrenergic neurons in the locus coeruleus (LC) on learning and attention, which are known to be impaired in aging and are influenced by the activity of the noradrenergic system.

The LC represents the major source of noradrenaline (NE) in the central nervous system, comprising just 45–50 thousand neurons in healthy young adult humans (Sharma et al., 2010) and roughly 1,500 in adult mice (Sturrock and Rao, 1985). Noradrenergic terminals richly innervate the cortex, hippocampus, amygdala, hypothalamus and brainstem (Westlund and Dan Coulter, 1980; Gaspar et al., 1989; Sadikot and Parent, 1990; Ginsberg et al., 1993; Radley et al., 2008). LC terminals have synaptic (Papadopoulos et al., 1987) and also non-synaptic (Beaudet and Descarries, 1978) release sites, where they bind and activate noradrenergic receptors. There are two subfamilies of noradrenergic receptors: those in the α subfamily are Gq (α1, often associated with excitation) or Gi (α2, often associated with inhibition) coupled, whereas members of the β subfamily are Gs-coupled metabotropic receptors (Szabadi, 2013). Activation of neuronal noradrenergic receptors leads to long-term alterations in synaptic strength and gene transcription, particularly within the hippocampus and prefrontal cortex, influencing memory formation (Sara, 2009) and consolidation (Gibbs and Summers, 2002). Most probably noradrenergic signaling influences memory not only directly but also indirectly, through the regulation of arousal. Lower or higher arousal than optimal leads to impaired learning ability. Reduced arousal is associated with drowsiness and sleepiness, whereas high arousal is associated with distractibility and anxiety. Noradrenergic neurons in the LC are responsible for setting of the optimal arousal level for different life situations and tasks, thus for the optimization of performance (Aston-Jones and Cohen, 2005). It is hypothesized that noradrenergic neurons act as a temporal attention filter enabling focused attention in the presence of stimuli associated with high reward or necessary for avoidance and for the optimal balance between goal-oriented and exploratory behavior (Aston-Jones and Cohen, 2005).

Evidence from several studies suggests that noradrenergic dysfunction is involved in the development of age-related cognitive deficits. Experimental disruption of LC-noradrenergic transmission in animal models generally led to an impaired performance in tasks designed to measure attention-related processes (Sara, 1985) supporting the notion that lesions to the dorsal noradrenergic bundle interfere with the animal’s ability to select stimuli and to maintain attention once selected (Marien et al., 2004). Indeed, noradrenaline-depleted rats performed as well as non-lesion control animals in the visual discrimination task, but when irrelevant stimuli were introduced into the discrimination array, the depleted group was severely impaired (Oke and Adams, 1978). Noradrenaline-depleted animals showed significant impairment in performing the task also when the trials were presented unpredictably and more rapidly, or when a white noise arousal stimulus was presented (Carli et al., 1983) underlying again the importance of noradrenergic signaling in selective attention. A number of animal studies reported that depletion of NA differently affects cognitive domains: it has no or minimal effect in test of spatial memory including Morris water-maze (Hagan, 1983) or radial arm maze tests (Chrobak et al., 1985) whereas in emotionally arousing models like in the inhibitory avoidance test it impaired long term retention (Prado de Carvalho and Zornetzer, 1981) and performance in the fear conditioning test (Giustino and Maren, 2018). Moreover, NA depletion enhanced the effect of the cholinergic antagonist scopolamine (Decker and Gallagher, 1987) whereas an increase of NA signaling improved performance of aged rodents in the inhibitory avoidance (Zornetzer, 1985) test.

Noradrenergic neurons in the LC can protect against brain aging and the progression of neurodegenerative diseases by influencing glial activity. The pro-inflammatory activity of glia cells increases during aging, which contributes to the aging process and has a pathogenic role in neurodegenerative processes. Noradrenalin exerts potent anti-inflammatory effects (Madrigal et al., 2005) through activating microglial and astrocytic adrenoceptors and on microglia (Szabó et al., 1997) and astrocytes (Frohman et al., 1988). In the brain, the anti-inflammatory effect of NA is mediated via β2 adrenergic receptors (Loop et al., 2004) and involves a reduced activity of NFkB (Gavrilyuk et al., 2002; Heneka et al., 2003). The age-related decline in the number of LC noradrenergic neurons can contribute to the enhanced glial activity and to the increasing pro-inflammatory milieu in the aging brain. Indeed, in our previous work we showed that depletion of NA neurons with DSP-4 treatment in APP23 mice enhanced pro-inflammatory activity of glia cells, augmented amyloid deposition and neurodegeneration, increased memory deficits (Heneka et al., 2010). Whether decreased noradrenergic cell number contributes to the neuroinflammatory changes associated with normal aging is not known.

It seems well-known that the disruption of the noradrenergic system is linked to the onset and progression of neurodegenerative diseases. It has been demonstrated that the LC undergoes an approximately 60% neuronal loss in patients suffering from Parkinson’s disease (PD) and Alzheimer’s disease (AD) even from the early stage of the disease (Braak et al., 2003, 2011; Ordureau et al., 2015; Hou et al., 2018; Reddy and Oliver, 2019). However, whether or not LC neurons are also lost during physiological aging is still debated (Flood and Coleman, 1988; Mouton et al., 1994; Manaye et al., 1995). A recent in vivo human study showed a correlation between LC signal intensity values and age, revealing an age-related decline in LC signal intensity values for the first time (Liu et al., 2019). For mice the results were controversial (Leslie et al., 1985; Sturrock and Rao, 1985; Tatton et al., 1991; Szot et al., 2009) but in our recent study, using unbiased stereological counting, we found an age-related approximately 20% decline in the number of noradrenergic neurons in the LC of C57BL/6 J mice (Gargano et al., 2021).

In the present study, we wanted to understand whether this limited decrease in the number of LC noradrenergic neurons per se affects cognitive abilities. To do this, we used the noradrenergic neuron-specific toxin DSP-4 to reduce the number of noradrenergic neurons in young animals to levels typical of old mice. We first determined the dose of DSP-4 required to induce a loss of LC neurons in young mice similar to that seen in old mice. We then investigated whether this approximately 20% reduction in the number of LC noradrenergic neurons in DSP-4-treated young animals leads to similar deficits in fear memory using contextual fear conditioning (CFC) and in attention using the 5-choice serial reaction time test (5-CSRTT) as observed in aged mice. We focused on these cognitive domains because they are impaired in old individuals and are regulated by the noradrenergic system.

Although detailed stereological analyses of the aging brain revealed that the histological structure of the brain is largely preserved, there is a significant neuronal loss in several brain areas such as the (LC) (Flood and Coleman, 1988). In patients with neurodegenerative diseases such as Alzheimer’s disease (AD) (German et al., 1992) or with mild cognitive impairment (Grudzien et al., 2007) a substantial loss of LC neurons has been clearly demonstrated. There is a significant loss of LC neurons even in the prodromal symptom-free phase of AD (Arendt et al., 2015), with further reductions in the successive Braak stages (Theofilas et al., 2017). Reduced noradrenergic signaling is known to contribute to the pathogenesis of Alzheimer’s disease. The reduction in the number of LC noradrenergic neurons is more moderate in normal aging. It was not known whether this approximately 20% loss of LC neurons had any effect on cognitive function. In our present work, we now show that a reduction in LC noradrenergic neurons in young animals, i.e., in the absence of all other age-related histological changes such as neuroinflammation, impairs cognitive functions.

Aging is known to be associated with a progressive decline in episodic memory (Levine et al., 2002), working memory (Gazzaley et al., 2005), speed of information processing and task switching. An age-related decline in attention could significantly contribute to these deficits (Robertson, 2013). Some studies have also proven how aging selectively impairs the process of memory extinction and the capacity to use context to modulate learned responses to threat (Battaglia et al., 2018). These cognitive domains that seem to be deeply influenced by aging have also been shown to be heavily regulated by the noradrenergic system (Arnsten et al., 1997; Seo et al., 2021). For these reasons we selected a battery of models (CFC and 5-CSRTT) sensitive to both age-related decline and noradrenergic signaling. In this way, we could selectively investigate memory formation and extinction, working memory and attention in our animals.

Our behavioral data pointed out that old C57BL/6 J mice experience difficulty extinguishing fearful memory compared to young controls in the contextual fear conditioning test (Figure 3A; Supplementary Figure S1). Different studies confirmed our data and provided us with additional insights to explain what indeed happens during aging (Shoji et al., 2016; Battaglia et al., 2018; Yanai and Endo, 2021). For example, Shoji et al. (2016) conducted a large-scale analysis of age effects on behavior in C57BL/6 J wild-type mice. In particular, they examined the mice in the startle response/prepulse inhibition test and in the fear conditioning test, both known to be associated with the activity of the noradrenergic system (Kokkinidis and Anisman, 1978; Saitoh et al., 1986; Fendt et al., 1994; Giustino and Maren, 2018), finding an age-related increased prepulse inhibition and especially an increase in freezing behavior. Therefore, they suggested that this could be explained by an age-related increase in generalized fear/anxiety or age-related deficits in the ability to discriminate between contexts. Moreover, a study from Yanai and Endo (2021) also showed results consistent with our own; he also hypotheses that, based on their data, the intense freezing observed in older aged mice may reflect elevated anxiety states and/or reduced locomotor activity.

Importantly, in our study we could also demonstrate that old C57BL/6 J animals show a significant impairment in working memory in the 5-CSRTT compared to young controls, which was associated with lower choice accuracy (Figures 3B,C). It is well known how noradrenaline modulates the balance of cortical excitation and inhibition and how these neuromodulatory adjustments facilitate the selective processing of prioritized information. Thus, by disrupting the coordination in attention networks, age-related LC degeneration can impair the neural processing (Dahl et al., 2022). Contextually, a study in humans from Lee et al. (2018) showed how arousal, which in young adults increased processing of salient stimuli and decreased processing of non-salient stimuli, in older adults increased processing of both low and high-salience stimuli. Older adults also showed a drop in LC functional connectivity. Thus, the lower choice accuracy that we associated with aging could be due to an impaired arousal and an age-related altered processing of non-salient stimuli (Lee et al., 2018). Additionally, a recent human study indicated that cognitive components contributing to distractibility follow different trajectories with aging and that increased distractibility in older adults seems to result from a dominance of involuntary attention processes (Hoyer et al., 2022).

The cause of these age-related pathologies is not fully understood, although this is the key to successful therapy or prevention of dementia and neurodegenerative disorders.

To clarify whether and to what extent a specific element of brain aging—in our study a reduction in noradrenergic signaling due to the loss of LC noradrenergic neurons—is responsible for these cognitive deficits, we depleted LC neuronal numbers by treating young animals with 14 mg/kg DSP-4 toxin and tested them 60 days after the treatment. This model allowed us to selectively test the effect of a similar loss of neurons as found in the aged brain, but in the absence of other age-related pathologies such as neuroinflammation.

Consistently with the literature (Evans et al., 2021), old mice showed an increased pro-inflammatory profile in the LC characterized by increased microglial volume and enhanced Iba1 and TNFα levels. On the other hand, DSP-4 treated mice showed only a moderate increase in Iba1 levels, while in other parameters they were similar as vehicle controls (Figure 7). We could therefore exclude that an extensive inflammatory change due to the neuronal death was responsible for the observed memory and attention deficits. The fact that we did not find enhanced pro-inflammatory glial activity is probably due to the relatively low dose of DSP-4. Previous studies, typically using 50 mg/kg of DSP-4, reported enhanced neuroinflammation and oxidative stress in the LC and in its main target areas (Iannitelli et al., 2023).

It is important to note, that although the extent of neuronal loss in our model and during aging is similar, the underlying mechanisms are substantially different. DSP-4 is able to cross the blood–brain barrier and block noradrenaline release at the nerve terminals by causing an initial inhibition of the noradrenaline transporter (NET) (Dudley et al., 1990; Ross and Stenfors, 2015). Administration of 50 mg/kg DSP4 leads to a rapid depletion of noradrenaline levels, followed by a degeneration of noradrenergic axon terminals within 24 h (Grzanna et al., 1989). The number of noradrenergic cell bodies remains stable for at least 18 days after the treatment, although these cells show expression changes indicative of neurodegeneration and reduced cell identity (Iannitelli et al., 2023). A reduction also in the number of LC noradrenergic neurons is present first 23 days after the DSP-4 treatment (Li et al., 2018) Although the reduction in neuronal numbers persists for at least 1-year, almost complete regeneration of the axons was observed 6 months after the treatment (Fritschy et al., 1992). On the other hand, there are several factors, which contribute to the increased vulnerability of noradrenergic neurons to neurodegenerative changes in aging: dopamine toxicity due to increased oxidative stress during the noradrenaline synthesis, high neuronal iron content, autonomous pacemaker activity (Sanchez-Padilla et al., 2014), very high axonal arborization size (Morrison et al., 1979), which together lead to energetic failure (Giguère et al., 2018), are all thought to contribute to the slow decline of LC neurons. Nevertheless, both aging and treatment of young mice with 14 mg/kg DSP-4 resulted in a similar decline in noradrenergic signaling.

As expected, our model revealed a significant decrease in the density of noradrenergic axon terminals in the PFC (Figure 2). We selected this brain region because the PFC is crucial in mediating the working memory and attentional functions, for which the key importance of noradrenaline has been extensively demonstrated (Aston-Jones et al., 1991, 1994; Arnsten et al., 1997; Sara, 2009). PFC receives noradrenergic projections selectively from the LC. The noradrenergic system in the brain is composed of two primary ascending projections that originate from the brainstem: the dorsal noradrenergic bundle (DNB) and the ventral noradrenergic bundle (VNB). The DNB originates from A6 LC, and it functions as the principal site of norepinephrine production in the central nervous system. It sends projections to the cerebral cortex, hippocampus, and cerebellum exclusively, while the VNB, originating from nuclei in the pons and medulla, sends projections to innervate mostly amygdala, hypothalamus, and areas of the midbrain and medulla (Weinshenker and Schroeder, 2007; Montoya et al., 2016; Rho et al., 2018). Thus, the reason for the lower density of noradrenergic terminals in the PFC of DSP-4 treated mice must be the reduced number of LC noradrenergic neurons. Indeed, in our previous study, we demonstrated that the neuronal loss in the LC during aging similarly affects the density of axon terminals in all the efferent regions analyzed [cortex, hippocampus, amygdala, and hypothalamus (Gargano et al., 2021)]. We also excluded the involvement of the peripheral sympathetic system in the observed cognitive deficits, since NE (and in general catecholamines) are unable to cross the blood-brain-barrier. In fact, the blood-brain-barrier is both a physical and an enzymatic barrier, containing the catecholamine metabolizing enzymes monoamine oxidase and catechol-O-methyltransferase (Bertler et al., 1963; Hamberger and Masuoka, 1965; Bertler et al., 1966; Bartholini et al., 1971). Although the blood-brain-barrier is fairly permeable to NE at birth (Glowinski et al., 1964), its permeability begins to decrease as early as 2 weeks after birth (Loizou, 1970), and becomes fully competent in adulthood (Axelrod et al., 1959; MacKenzie et al., 1976; Kostrzewa, 2007).

Our data did not show any significant difference between DSP-4 treated and vehicle control animals in either the process of fear learning or extinction, indicating that the age-related neurodegeneration of LC neurons alone is not liable for the deficit in fear extinction characteristic of old age (Figure 4A). This could be explained by the major role played by other brain regions in contextual conditioning, for instance the hippocampal dentate gyrus. Indeed, despite the β-adrenergic-mediated modulation of hilar interneurons is necessary to promote aversive generalization (Seo et al., 2021), a reduction as small as that associated with aging is probably not sufficient to disrupt the whole process. In contrast to our mice treated with 14 mg/kg DSP-4, mice repeatedly treated with 50 mg/kg DSP-4 showed deficits in contextual fear memory (Chalermpalanupap et al., 2018). Not surprisingly, the reduction in the NET-positive noradrenergic axon terminals was much higher using higher amount of DSP-4: it was 50% in Chalermpalanupap’s study and 20% in our study. We hypothesize that a moderate reduction in noradrenergic signaling, which occurs during normal, healthy aging, does not significantly affect memory abilities. A more severe reduction, such as that observed in Alzheimer’s and Parkinson’s disease, may impair learning and memory per se.

Yet, when comparing DSP-4 treated and vehicle animals in the 5-CSRTT, we found that DSP-4 treated animals needed significantly more time to learn the task and to reach the last phase of the test compared to vehicle injected controls, indicating an impaired working memory (Figure 4B). Interestingly, our data also show that, when a distractor was introduced in the test, the DSP-4 treated mice exhibited a poorer choice accuracy compared to controls which did not seem to be disturbed by it (Figures 4C, 5). Our data also showed that, in both vehicle and DSP-4 treated groups, the presence of the distractor influenced the number of correct/incorrect answers but not the number of omissions. This could be explained by the fact that the distractor does not impact the drive of the animal per se (which is mostly due to the food restriction) but influences the choice accuracy during test. It is also important to remember that DSP-4 treated animals are considered young and that the only brain region altered the way aging would is the LC. Therefore, these data seem to be particularly interesting as, while DSP-4 treated animals did not show a major impairment as old animals did, their performance was compromised compared to vehicle controls. Thus, these data demonstrated that even a small reduction in noradrenergic signaling is able to impact working memory and attention as these cognitive domains are majorly controlled by the noradrenergic system.

In conclusion, we have found that a reduction in LC neurons may directly contribute to impaired working memory and greater distractibility in the performance of attentional tasks. Thus, even in the absence of other age-related changes in the brain, a reduced number of LC noradrenergic neurons per se impairs cognitive functions. Taken together, our findings help to elucidate the role of reduced noradrenergic signaling in age-related cognitive deficits and provide a druggable target for alleviating negative symptoms of brain aging.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was approved by Landesamt fuer Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen. The study was conducted in accordance with the local legislation and institutional requirements.

AG: Formal Analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. BO: Methodology, Writing – original draft. MP: Investigation, Methodology, Writing – original draft. AZ: Funding acquisition, Writing – review & editing. AB-G: Conceptualization, Funding acquisition, Resources, Supervision, Writing – original draft, Writing – review & editing.

The author(s) declare financial support was received for the research, authorship, and/or publication of this article.

The work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—project numbers 426320013 and 461228630 to AB-G and to AZ under project number 324087152 as well as under Germany’s Excellence Strategy—EXC2151—390873048.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.