Beyond the brain: early autonomic dysfunction in Alzheimer’s disease

Abstract

Alzheimer’s disease (AD) is classically defined by central hallmarks such as amyloid-beta plaques, tau hyperphosphorylation, and synaptic failure. However, mounting evidence suggests that dysfunction outside the brain, particularly in the peripheral nervous system, may also play a significant role in disease progression. The adrenal medulla—a key regulator of systemic neurotransmission and stress response—has received little attention in this context. In this study, we investigated whether chromaffin cells (CCs) from the triple transgenic AD mouse model (3xTg) exhibit functional alterations that could contribute to peripheral neurochemical imbalance. Using electrophysiology, high-resolution amperometry, and neurotransmitter quantification, we identified early and progressive defects in CC function. Remarkably, even at two months of age—prior to cognitive decline—3xTg CCs showed impaired exocytosis, reduced vesicle release, and slower fusion pore kinetics. These changes were accompanied by diminished sodium (I_Na), calcium (I_Ca), and nicotinic (I_ACh) currents, compromising CC excitability. With age, a shift toward increased potassium (I_K) currents and enhanced catecholamine secretion may reflect compensatory adaptations aimed at preserving output. These functional deficits were paralleled by structural remodeling of the actin cytoskeleton and systemic neurotransmitter disturbances. Noradrenaline levels increased in both plasma and brain, while dopamine decreased peripherally but paradoxically increased in the prefrontal cortex and hippocampus. Serotonin levels consistently declined across compartments. These imbalances correlated with altered behavior: 3xTg mice displayed increased exploration of exposed areas and heightened behavioral despair, pointing to anxiety- and depression-like phenotypes. Together, our findings identify the adrenal medulla as a previously underrecognized site of early catecholaminergic dysregulation in AD. The observed associations between peripheral CC dysfunction, systemic neurotransmitter imbalance, and behavioral changes point to a potential link between peripheral neuroendocrine alterations and central disease features. These results broaden the current understanding of AD pathophysiology and support the adrenal medulla as a promising candidate for further investigation as a therapeutic target and source of peripheral biomarkers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40478-025-02042-8.

Introduction

The most common neurodegenerative disorder is Alzheimer’s disease (AD). Its main pathogenic features are brain deposition of amyloid-beta (Aβ) plaques and neurofibrillary tangles associated to tau hyperphosphorylation [57]. Altered neurotransmission and synaptic loss are a likely consequence of those pathogenic hallmarks; such synaptic failure seems to correlate well with AD progression and severity associated with memory loss and atrophy of the cerebral cortex [100].

The processing of amyloid precursor protein (APP) is linked to the multimolecular complex γ-secretase; its catalytic subunit presenilin 1 (PS1) regulates both Ca²⁺ circulation at the endoplasmic reticulum [83] and neurotransmitter release [117]. Also, APP is involved in metal chelation and normal synaptic function [61]. This could explain that in the familial form of AD (FAD), that accounts for only 1–2% of patients [19, 56], mutations of these proteins may combine to alter neuronal Ca²⁺ handling and synaptic transmission as shown in several mouse models of AD expressing one or more APP, PS1, or tau mutations [37, 98].

As with all neurodegenerative disorders, the pathogenesis of AD relies on brain nuclei. In AD, the hippocampus and the prefrontal cortex are central stages in its pathogenesis [11]. However, some functional features linked to such brain pathology have also been described peripherally. This is the case of fibroblasts that exhibit an altered Ca²⁺ signaling [65], decreased mitochondrial density in endothelial cells of the blood–brain barrier [105], and alterations of mitochondrial enzymes in fibroblasts and platelets [7].

Accumulating evidence indicates that the peripheral nervous system could also be affected by various neurodegenerative diseases. This has been explored at the sympatho-adrenal axis, which controls the fight-or-flight response during stressful conflicts and behaves as a gate-keeper of the body´s homeostasis [16, 17, 25, 28]. In the amplifying arm of the sympathetic nervous system, the adrenal medullary chromaffin cells (CCs) release explosive amounts of catecholamines into the bloodstream through exocytosis. [29]. This secretory process is exquisitely dependent on Ca²⁺ entry [35] and the sudden increase in cytosolic Ca²⁺ concentration ([Ca²⁺]_c), which activates the SNARE proteins of the exocytotic machinery in a way similar to the release of neurotransmitters in the central and peripheral nervous systems [44]. Thus, as neurons, CCs act as excitable cells and generate action potentials in response to acetylcholine (ACh) stimulation, relying on a surprisingly large variety of ion channels [18, 70]. Of interest is that CCs from mammalian species are readily available, they can be easily cultured and patch-clamped, and the fine kinetics of the exocytotic fusion pore of single vesicles can be characterized using amperometry [112]. Therefore, CCs have been amply used to study physiological aspects of ion channel currents, cell excitability, [Ca²⁺]_c signaling, and the role of different SNARE or cytoskeletal proteins on exocytosis [26, 82]. CCs have also been used to determine the role of some proteins linked to neurodegenerative diseases on the regulation of cell excitability, ion channel currents and exocytosis [67, 69, 72, 73, 117]. This has also been the case for some mouse models of neurodegenerative diseases [14, 27, 60, 76] and autism [13].

Several studies have explored how AD–associated factors may alter CC function. For example, a mutation of the regulator of calcineurin 1 (RCAN1), a gene that is overexpressed in Down’s syndrome (DS) and AD, altered CC exocytosis and vesicle fusion kinetics [63]. In another study, the quantal secretion of catecholamines and the kinetics of the fusion pore were studied in a mouse model expressing two mutations of FAD i.e., the APP Swedish mutation and the PS1 mutation A246E [9]; these mice present accelerated LTP decay [51], decreased synaptic plasticity and impaired LTP [45] and depressed excitatory synaptic transmission in hippocampal neurons [94]. In CCs from APP/PS1 mice, neurosecretion is similarly altered with amperometric single-vesicle spikes having smaller amplitude and half-width (t_1/2), as well as decreased quantal vesicle content of catecholamines; these changes were observed in the absence of Aβ deposition in CCs [27].

In this study, we investigated whether CCs from the 3xTg mouse model of AD exhibit functional alterations in excitability, ion channel activity, and exocytosis, and whether these changes are associated with systemic neurotransmitter imbalances and behavioral modifications. We also examined whether these alterations emerge in an age-dependent manner as the disease progresses. Given that the 3xTg model harbors three familial AD-linked mutations (APP_Swe, PS1_M146V, and tau_P301L) [87], we hypothesized that it would better recapitulate key aspects of AD pathology compared to other models. Our findings reveal significant dysregulation of exocytosis, ion currents, and neurotransmitter release, with early dysfunction at presymptomatic stages that progresses with aging. These cellular and biochemical changes coincide with memory deficits, reduced anxiety-like behaviors, and increased depressive-like behaviors, suggesting that autonomic and peripheral neurotransmitter imbalances may be linked to AD-related neuropsychiatric symptoms.

Materials and methods

Animals

Experiments were conducted according to the recommendation of the Ethics Committee from Universidad Autónoma de Madrid on the use of animals for laboratory experimentation, in accordance with the code of ethics and guidelines established by the European Community Directive (2010/63/EU) and Spanish legislation (RD53/2013). All efforts were made to avoid animal suffering and to use the minimum number of animals allowed by the experimental protocol and the statistical power of group data. Mice were housed under controlled temperature, a 12:12 h light cycle, and food and water were provided ad libitum.

Male 129/C57BL6 mice harboring three mutant transgenes linked to human FAD (PS1_M146V, tau_P301L, APP_Swe), hereafter referred to as 3xTg, were kindly provided by Dr. Javier García Sancho (Universidad de Valladolid, Spain). These mice contain the PS1_M146V (substitution of methionine by valine at codon 146), overexpress the Swedish mutation of the amyloid precursor protein (APP_Swe) and AD tau mutation P301L (substitution of proline by leucine at codon 301), under the transcriptional control of the Thy1.2 expression cassette [87].

Mice were used on postnatal days P60 (2 m), P180 (6 m), and over P360 (12 m). At 2 m age animals are cognitively unimpaired, whereas at 6 m and 12 m age animals show cognitive deficits (Supplementary Fig. 1) [2, 6, 48, 53, 86, 87, 92]. Male B6129SF2 J mice were kindly provided by Dr. Miguel Ángel del Pozo (CNIC, Madrid) and used as controls, at the same ages i.e. 2 m, 6 m and 12 m.

Primary culture of mouse chromaffin cells

Mice were euthanized by cervical dislocation. Both adrenal glands from a mouse were rapidly collected and placed into ice-cold Locke’s solution of the following composition (in mM): 154 NaCl, 5.6 KCl, 3.6 NaHCO₃, 5 HEPES and 5.6 D-glucose (pH 7.4, NaOH). Glands were fat trimmed, and medullae were isolated by carefully dissecting the cortex. Then the medullae were placed in a tube containing 200 µL Locke’s solution with papain (25 U/mL), for tissue digestion during 20–30 min at 37 °C. This solution was exchanged by 1 mL of Dulbecco's Modified Eagle Medium (DMEM), repeating the exchange 3 times and leaving finally 120 µL of DMEM. Then, medullae were minced first with a 1 mL micropipette tip and then with a 200 µL micropipette tip. Finally, the residual medulla fragments were discarded, and 10–20 µL drops of cell-containing solution from the minced extracts were plated on poly-D-lysine-coated coverslips on 24-well plates. After 30–45 min in an incubator (37 °C, water-saturated and 5% CO₂ atmosphere) to allow cell attachment at the center of coverslips, 500 µL DMEM supplemented with 5% fetal bovine serum, 50 IU/mL penicillin, and 50 µg/mL streptomycin were added to each well and remained in the incubator until the following day, during which experiments were performed.

Amperometric recordings of single exocytotic events

Quantal release of catecholamine was measured with amperometry [21, 112]. Electrodes were built as previously described [62] by introducing a 10 µm diameter graphite fiber (Amoco, now part of BP-Group, London, UK) into glass capillary tubes (Kimble-Kontes, Vineland, NJ, USA). These tubes were then pulled (Narishige PC-10 pipette puller, Narishige, Tokyo, Japan) by the application of 2 successive heats, and the carbon fiber was then cut with a pair of small scissors obtaining thus two pipettes with a carbon fiber piece sticking out of each tip. The tip was sealed by a two-component epoxy (EPIKOTE 828-Miller-Stephenson, Danbury, CT, USA; and m-phenylendiamine, 14%, Aldrich, Steimheim, Germany). The electrodes were left overnight to dry; thereafter, they were introduced into an oven at 100 °C for 2 h, and subsequently kept for another 2 h at 150 °C.

For amperometric recordings, electrodes were mounted onto the headstage of an EPC-9 patch-clamp amplifier (HEKA Elektronic, Lambrecht, Germany). Data was acquired at a sampling frequency of 20 kHz and analyzed using Pulse v8.74 software (HEKA Elektronik). A potential of 700 mV was applied to the electrode relative to an AgCl ground electrode. The electrodes were calibrated by exposing them to 50 µM adrenaline dissolved in standard Krebs-HEPES solution composed of (in mM): 145 NaCl, 5.6 KCl, 1.2 MgCl₂, 2 CaCl₂, 11 glucose and 10 HEPES (pH 7.4, NaOH) and measuring the current elicited; only electrodes that yielded a current between 200 and 400 pA were used [71]. The coverslips were mounted in a chamber on a Nikon Diaphot inverted microscope used to localize the target cell, which was continuously perfused with a Krebs-HEPES using a five-way system with a common outlet driven by electrically controlled valves. Cells were stimulated with 100 µM ACh. At the time of experiments, proper amounts of drug stock solutions were freshly dissolved into Krebs-HEPES solution to the desired concentration.

Amperometric data were analyzed on a personal computer using Excel software (Microsoft, Redmond, WA, USA) and IgorPro software (Wavemetrics, Lake Oswego, OR). Amperometric quantal size (Q) was calculated by integrating the amperometric current over time during the stimulus duration with a homemade macro written in IgorPro software. The number of spikes greater than 5 pA was manually counted on an extended graph displayed on the computer screen. Kinetic analysis of single amperometric events (spikes) was performed as previously described [39] using a macro written in IgorPro software [81]. Mean values for each parameter of each cell's spikes were obtained, and then an average of all cells belonging to the same group, namely, 2 m, 6 m, and 12 m, were pooled together for statistical comparison. This method helps overcome the large variability in spike number and kinetics by giving each cell the same weight independently of the number of spikes produced [23, 24]. All experiments were performed at room temperature (24 ± 2 °C).

Electrophysiological recording of ion currents and cell excitability

Coverslips with cultured cells were placed in an experimental chamber mounted on the stage of a Nikon Diaphot inverted microscope. Using the patch-clamp technique, inward currents through nicotinic receptors (I_ACh) and voltage-gated calcium (I_Ca) and sodium (I_Na) channels, as well as outward currents through voltage-gated (I_K(V)) and Ca²⁺-dependent K⁺ channels (I_K(Ca)) were recorded. Additionally, the membrane potential (V_m) and evoked action potentials (eAPs) were also measured [55]. Whole-cell recordings were made with fire-polished borosilicate pipettes (resistance 4–8 MΩ) mounted on the headstage of an EPC-9 patch-clamp amplifier (HEKA Electronik, Lambrecht, Germany). Data were acquired with a sample frequency of 20 kHz using the PULSE 8.74 software (HEKA Elektronik). Linear leak and capacitive components were subtracted using a P/4 protocol and series resistance was compensated by 80%. The data analysis was performed with Igor Pro (Wavemetrics, Lake Oswego, OR, USA) and PULSE softwares (HEKA Elektronik).

Cells were continuously perfused at 24 ± 2 °C with a Krebs-HEPES solution. Once the patch membrane was ruptured and the whole cell configuration of the patch-clamp technique had been established, the cell was locally, rapidly and constantly perfused with an external solution of similar composition to the chamber solution, but containing nominally 0 mM Ca²⁺ to measure I_Na and I_K(V) or 2 mM Ca²⁺ to monitor I_ACh, I_Ca, I_K(Ca), V_m, and evoked APs. To measure I_ACh, I_Na and I_Ca, cells were internally dialyzed with an internal solution containing (in mM): Cs-glutamate 100, EGTA 14, TEA.Cl 20, NaCl 10, Mg.ATP 5, Na.GTP 0.3, and HEPES 20/CsOH (pH 7.3). To record I_K currents, V_m and APs, cells were internally dialyzed with an internal solution at pH 7.4 containing (in mM): KCl 135, NaCl 10, HEPES 10, MgCl₂ 1, and EGTA 5. We used different protocols for data acquisition: for I_ACh cells were voltage-clamped at − 80 mV and a 250 ms pulse of ACh 100 µM was applied; for I_Ca, cells were voltage-clamped at − 80 mV and step-depolarizations with 50-ms depolarizing pulses were applied at 10-s intervals to minimize current rundown [38]. For I_Na, cells clamped at − 80 mV and step-depolarizations with 10-ms depolarizing pulses were applied at 15-s intervals. For I_K, 400 ms depolarizing pulses were applied at 20-s intervals to 150 mV to mouse CCs voltage-clamped at − 80 mV. Recordings of membrane potential (V_m) and APs were made under the current-clamp mode in the whole-cell configuration of the patch-clamp technique [55], which allows the observation of spontaneous variations in the V_m.

Actin cytoskeleton staining

Cultured chromaffin cells were processed as follows. DMEM was first removed from the wells, which were then washed with a pre-warmed sterile, and filtered PBS solution (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂PO₄; pH 7.4). Subsequently, the PBS was aspirated and 500 μL of a 4% paraformaldehyde (PFA) solution was added to fix the cells for 20 min at room temperature. After fixation, the PFA was discarded and the cells were washed three times with PBS. Cell membranes were permeabilized by incubating the cultures with 0.1% Triton X-100 in PBS for 10 min. The cells were then incubated with Phalloidin 546 (Sigma-Aldrich), diluted 5:200 in 0.1% Triton-PBS, for 1 h at room temperature. To label the nuclei, the cells were subsequently incubated with DAPI (1:500 in PBS) for 10 min at room temperature in the dark, followed by three PBS washes. Samples were mounted on microscope slides with 5 μl of ProLong Gold Antifade mounting medium (Thermo Fisher Scientific) and allowed to dry overnight in the dark. The slides were then stored at 4 °C in darkness until imaging. Images were analyzed using ImageJ (Fiji) [99]. A uniform threshold was applied to all images to enable comparison, and the integrated density (IntDen) was measured to obtain the mean signal relative to the labeled area.

Neurotransmitter levels measurement by high-performance liquid chromatography with mass spectrometry (HPLC/MS)

Mice were euthanized by cervical dislocation and the blood was extracted by cardiac puncture and collected in 3 mL K₂EDTA tubes to prevent blood coagulation. 25 mg/mL of ascorbic acid and 2 mg/mL of citric acid were added to the tubes for neurotransmitter preservation. Tubes were centrifugated for 10 min within the next hour at 4 °C and 1300 g to separate cells from plasma, and then plasma was stored at − 80 °C. Brains were rapidly removed over an ice bath, and prefrontal cortex and hippocampus were collected in Eppendorf tubes containing 20 µL or 50 µL of 0.5 N formic acid, respectively. Then, they were sonicated and stored at − 80 °C until HPLC–MS analysis. Concentrations of adrenaline (ADR), noradrenaline (NA), dopamine (DA) and serotonin (SER) in mouse brain tissue (hippocampus and prefrontal cortex) and plasma were measured by a high-performance liquid chromatography-tandem mass spectrometry (HPLC–MS/MS) method [113]. Isoprenaline was used as the internal standard for each analyte. For analyte extraction, the protein precipitation (PPT) method was used. Briefly, the brain samples were homogenized in ice-cold 1.89% formic acid in water at a concentration of 10 mL/g tissue and were immediately centrifuged. Acetonitrile with 1% formic acid was added to the supernatant along with the internal standard and then the proteins were precipitated in a 4:1 proportion (v/v). Subsequently, the samples were evaporated and reconstituted in mobile phase [114]. The mobile phase consisted of a combination of 0.2% formic acid in MilliQ water (solution A) and acetonitrile (solution B). Plasma samples followed the same procedure from the addition of acetonitrile with 1% formic acid. The chromatogram was run under gradient conditions at a 0.6 mL/min flow rate at 25 °C through an ACE C18-PFP reversed-phase column (SYMTA, Madrid, Spain). The total chromatographic run time was 14 min, constituting of a 10-min run time and a 4-min re-equilibration time. A volume of 5 μL eluent was injected into the HPLC–MS/MS system. Linearity was confirmed for concentration ranges of 0.25–185, 0.5–200, 0.25–200, and 10–3000 ng/mL for AD, NA, DA, and SER, respectively. After optimization, the method was validated for accuracy and precision, matrix effect, extraction recovery, stability, and carry-over based on the requirements of regulatory agencies (U.S. Food and Drug Administration and European Medicines Agency).

Behavioral tests

Novel object recognition (NOR)

The NOR test was used to assess working memory in rodents following Leger et al. [68]. Mice underwent a three-day protocol in a 40 × 40 cm arena: on Day 1, they were habituated for 10 min; on Day 2, two identical objects were placed symmetrically for a 10-min exploration; and on Day 3, one object was replaced with a novel object differing in color, texture, and shape. The arena and objects were cleaned with 70% ethanol between sessions. Exploration was defined as the mouse’s head orienting toward, contacting, or sniffing the object, and was manually timed. The discrimination index (DI) was calculated as:

DI = (Exploration Time for Novel Object − Exploration Time for Familiar Object)/(Exploration Time for Novel Object + Exploration Time for Familiar Object).

Open field test

The open field test [54] was employed to evaluate rodent behavior due to its non-aversive nature, lack of training requirements, and rapid assessment capabilities. Mice were placed in a 40 × 40 cm white arena and allowed to explore freely for 10 min. Behavior was recorded and analyzed using the MouBeat plugin in ImageJ software [4, 99], quantifying the distance traveled in the inner zone of the arena, distance traveled in the peripheral zone, time spent in the inner zone, and the number of entries into the inner zone of the arena. The arena was cleaned with 70% ethanol between trials.

Elevated plus maze test

The Elevated Plus Maze [96] is commonly used to assess anxiety-related behavior by exploiting the conflict between the drive to explore and the aversion to open, elevated spaces. The maze consists of two open arms (30 × 4.5 cm) and two enclosed arms with 15-cm-high walls, elevated 40 cm above the floor. Mice were placed at the center and allowed to explore for 10 min. The maze was cleaned with 70% ethanol between trials. Data analysis was performed using the MouBeat plugin in ImageJ [4, 99], focusing on the time spent in the open arms, the time spent in the closed arms and the exploration time in the border of the open arms.

Tail suspension test (TST)

The TST [104] was used to evaluate depressive-like behavior. A custom apparatus was constructed in our laboratory using a foam core board box (20 × 22 × 32 cm) with an aluminum bar positioned at 30 cm. Mice were suspended by attaching an 8 cm-long piece of adhesive tape to the distal end of their tail, leaving approximately 3–5 mm of the tail free. The other end of the tape was fixed to the metal bar, and the mice remained suspended for 5 min, which was the total duration of the test. Their ventral surface faced a high-resolution video camera to record their behavior. Following the protocol described by Can et al. [15], total movement time was recorded and used to calculate the percentage of immobility throughout the test.

Protein identification in tissue sections

To obtain the studied tissues, mice were anesthetized with Dolethal (50 mg/kg, intraperitoneally) and perfused via the ascending aorta with a saline solution (0.9% (w/v) NaCl), followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB, pH = 7.4). Both adrenal glands and the brain were extracted and post-fixed in 4% PFA for 24 h. The samples were then cryoprotected in 30% sucrose in 0.1 M PB. The tissues were embedded in agarose and sectioned into 40 μm serial slices using a Leica SM2400 microtome (Leica Biosystems, Nussloch, Germany). For immunostaining with specific antibodies, tissue sections were washed multiple times with PB, followed by a 1-h pre-incubation in a blocking solution (PBS containing goat serum and 2% Triton X-100). The samples were then incubated overnight at 4 °C under agitation with the following primary antibodies: anti-β-amyloid (1:200, rabbit; Invitrogen, Waltham, MA) and anti-htau (1:500, mouse; Invitrogen). The next day, samples were washed and incubated for 2 h in the dark with secondary antibodies (anti-rabbit and anti-mouse, 1:500; ThermoFisher Scientific). Finally, nuclei were stained for 5 min with DAPI (1:500). Slides were mounted using 10 μL of ProLong Gold Antifade mounting medium (ThermoFisher Scientific) and left to dry overnight in the dark. Samples were stored at 4 °C in darkness until microscopic observation. Images were captured using a Leica TCS SP5 spectral confocal microscope at the Confocal Microscopy Laboratory of SidI (UAM). Images were acquired from the adrenal medulla or hippocampal regions (CA1 and CA3), using four sequential captures (one per channel) to prevent fluorescence crosstalk. ImageJ software was used for image analysis. The same threshold value was applied to all images to ensure comparability, and IntDen (Integrated Density) values were measured to obtain the average signal relative to the labeled area.

Statistical analysis

Data is presented as the mean ± standard error of the mean (SEM). Graphs were generated using GraphPad Prism v8.0.1 (GraphPad Software, San Diego, CA, USA), and statistical analyses were performed with SigmaPlot v13.0 (Systat Software, Inc.). Data normality was assessed using the Shapiro–Wilk test. For normally distributed data, comparisons among groups were conducted using one- or two-way ANOVA followed by Tukey’s post hoc test for multiple comparisons, or by a two-tailed t-test for two-group comparisons. Non-normally distributed data was analyzed using the Kruskal–Wallis test with Dunn’s post hoc test, or the Mann–Whitney U test for pairwise comparisons. A significance level of 95% was established, with *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001 considered statistically significant.

Results

Alterations in vesicle release and exocytosis kinetics in 3xTg chromaffin cells

The selected single CC (1 DIV) was perfused in each experiment with a standard Krebs-HEPES (KH) solution containing 2 mM Ca²⁺. The tip of the carbon fiber microelectrode was smoothly touching the cell surface, and a 1-min resting period was initially allowed to reach a stable baseline amperometric recording. Cells discharging spontaneous amperometric spikes were discarded. Secretory responses were evoked by 1-min of cell perfusion with KH containing 100 µM ACh, only once (Fig. 1a). This long exposure to stimuli was necessary for two reasons: 1, to obtain a high amount of individual secretory events to increase the statistical power of the kinetic parameters monitored at the single-spike level; and 2, to explore whether CCs exhibited longer-lasting deregulated exocytotic events, upon prolonged “stress-like” cell stimulation.

Fig. 1.

Vesicle release and exocytosis kinetics are altered in 3xTg chromaffin cells. a Schematic representation of amperometric recordings in cultured chromaffin cells (CC). A carbon fiber electrode is positioned on the surface of a CC (blue) and a voltage of + 700 mV is applied (left). Catecholamines released by the application of an ACh pulse are oxidized and the resulting currents are measured as amperometric spikes (middle). A zoom-in of a representative amperometric spike, highlighting key kinetic parameters, is shown on the right (scale bar: 20 pA; 5 ms). b Representative amperometric recordings in WT (top; scale bar: 50 pA; 10 s) and 3xTg (bottom; scale bar: 100 pA; 10 s) chromaffin cells at 2, 6 and 12 months of age. Cells were stimulated with 100 µM ACh for 1 min (bottom horizontal lines). c Averaged total secretion as the number of amperometric spikes per 1-min recording in CCs of WT and 3xTg mice from different ages. d Averaged total secretion measured as the cumulative charge transfer (Q, pC) from all amperometric spikes per 1-min recording in CCs of WT and 3xTg mice from different ages. e Quantification of rise rate (pA/ms), defined as the slope of the rising phase of amperometric spikes in CCs of WT and 3xTg mice from different ages. f Quantification of Imax (pA), the peak amplitude of amperometric spikes in CCs of WT and 3xTg mice from different ages. g Quantification of the decay time (ms), the time it takes for the current signal to return to baseline after reaching its peak, of the amperometric spikes in CCs of WT and 3xTg mice from different ages. h Quantification of the spike width (t_1/2, ms), the time an event remains above 50% of its peak amplitude, of the amperometric spikes in CCs of WT and 3xTg mice from different ages. Data are presented as mean ± standard error of the mean (SEM). The number of spikes, cells, and mice per group were: 2 m WT (1925, 41, 4); 6 m WT (1152, 33, 4); 12 m WT (1230, 40, 4); 2 m 3xTg (913, 35, 7); 6 m 3xTg (1046, 34, 7); 12 m 3xTg (1234, 30, 5). Statistical significance *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001 vs. WT; #p ≤ 0.05; ##p ≤ 0.01; ###p ≤ 0.001 vs. 2 m of the same genotype

Representative secretory amperometric traces obtained from cells of WT and 3xTg mice are displayed in panel b of Fig. 1. In young mice, responses start with an initial burst of spikes, followed by lower frequency spikes. Notably, there is a significant decrease in secretory rates in 2 m 3xTg compared to WT. During the initial 15-s recording period, the spike number (SN) is similar; subsequently, while secretion ceases in 3xTg cells, the response continues growing in WT cells (Fig. 1b, c). Thus, at the end of the 60-s recording period, SN was 46.90 ± 5.56 in WT cells (n = 41 cells; 1925 spikes) and 25.77 ± 1.44 in 3xTg cells (n = 35 cells; 913 spikes) (p < 0.001) (Fig. 1c). Total secretion expressed as charge (Q, in pC) followed the same pattern as SN (Fig. 1d), although the difference did not reach the threshold for statistical significance. As WT mice age, vesicle release from CCs decreases (see the grey columns in Fig. 1c; p = 0.021). In contrast, AD 3xTg mice exhibit an increase in vesicle release (see the teal columns in Fig. 1c; p = 0.006). Thus, at 12 months, the secretory responses monitored as SN (Fig. 1c), or total charge (Fig. 1d) were surprisingly different compared to those of 2-month cells. Specifically, secretion was significantly higher in 3xTg cells than in WT cells. During the first 10-s recording period, cells fired spikes at similar rates, around 10–15 SN/5 s; however, by the end of the 60-s recording period, SN totaled 41.13 ± 4.41 in 3xTg and 30.75 ± 2.18 in WT cells (Fig. 1c). Similarly, the secretion as total charge reached 22.14 ± 3.60 pC in 3xTg cells (n = 30 cells; 1234 spikes), while it amounted to 12.46 ± 1.09 pC in WT cells (n = 40 cells, 1230 spikes; p = 0.037) (Fig. 1d).

In summary, aging in WT mice is associated with a decrease in the total number of vesicles released upon ACh stimulation, leading to a reduction in catecholamine secretion. In contrast, 3xTg mice at a pre-disease stage exhibit a lower vesicle release compared to age-matched WT mice. Interestingly, in 3xTg mice, catecholamine secretion increases with age, and at an advanced stage—when pathology is already established—both vesicle release and catecholamine secretion are significantly higher than in WT mice. This dysregulated release pattern may indicate an impaired adaptation to stressful situations in 3xTg mice.

Next, we studied the kinetics of full-spike exocytotic events (See scheme of Fig. 1a, right). The rate at which the fusion pore expands after stabilization is measured by the spike rise rate. This rate was substantially slower in 3xTg cells than WT cells across all ages (Fig. 1e). Interestingly, this difference was already evident in the 2-month-old mice (93.1 ± 6.1 pA/ms in WT versus 49.1 ± 7.7 pA/ms in 3xTg cells; p < 0.001). The values for spike amplitude (I_max) were also greater in WT compared to 3xTg cells, regardless of the mouse age (Fig. 1f). It is important to highlight that these differences existed as early as 2 months of age (WT versus 3xTg cells: 106.4 ± 6.2 pA versus 77.8 ± 14.2 pA; p < 0.001). The rate of fusion pore closure is indicated by the decay time, which was higher in 3xTg cells than in WT cells across all ages (Fig. 1g). In early 2 m mice, the decay time was about 50% higher in 3xTg compared with WT cells. Specifically, 10.9 ± 1.7 ms versus 5.7 ± 0.3 ms; p < 0.001. Consistent with this was the spike width (t_1/2) that was higher in 3xTg compared with WT cells (Fig. 1h). Again, this difference was already visible at 2 m early age (3xTg versus WT): 9.8 ± 1.0 ms versus 4.9 ± 0.2 ms; p < 0.001. Table 1 presents a detailed summary of the spike kinetic parameters analyzed.

Table 1.

Kinetic data of full-spike single exocytotic events amperometrically recorded from WT and 3xTg CCs at different ages stimulated with ACh (100 µM) for 1 min

Age	Mouse type	Event, cells, cultures	Rise rate (pA/ms)	Imax (pA)	Decay time (ms)	t1/2 (ms)	Quantal size (Q, pC)
2 m	WT	1925, 41, 4	93.06 ± 6.07	106.4 ± 6.25	5.67 ± 0.30	4.91 ± 0.24	0.60 ± 0.04
	3xTg	913, 35, 7	49.1 ± 7.7***	77.80 ± 14.17***	10.93 ± 1.68***	9.82 ± 1.05***	0.62 ± 0.04
6 m	WT	1152, 33, 4	90.26 ± 8.11	101.7 ± 9.30	4.20 ± 0.30###	3.97 ± 0.20#	0.50 ± 0.05
	3xTg	1046,34, 7	64.27 ± 8.79**	79.52 ± 9.03*	7.29 ± 0.79***	6.61 ± 0.63***	0.55 ± 0.05
12 m	WT	1230, 40, 4	81.25 ± 4.95	85.61 ± 4.48#	3.62 ± 0.21###	3.75 ± 0.19###	0.39 ± 0.02###
	3xTg	1234, 30, 5	51.52 ± 5.77***	64.46 ± 5.82***	6.21 ± 0.52***; #	5.63 ± 0.44***; #	0.42 ± 0.04##

Data are presented as means ± SEM of the number of spikes, cells and cultures indicated

^*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001 compared to WT. Differences between means were calculated with the nonparametric Mann Whitney test for non-normally distributed data and using the parametric t-Student analysis for normally distributed data

^#p ≤ 0.05; ^##p ≤ 0.01; ^###p ≤ 0.001 compared to 2 m age of the same animal type and parameter. Non-normally distributed data was compared using the nonparametric Kruskal–Wallis test, with Dunns post-test and normally distributed data using the parametric one-way ANOVA analysis of variance, with Tukey's multiple comparison test. Only cells with a minimum of 15 spikes were considered for the statistical analysis

To summarize, the release of vesicles from CCs accelerates as mice age (shorter decay times and half-width of the spike) in both WT and 3xTg mice. However, significant differences distinguish WT from 3xTg cells. In AD mice, the spikes are smaller in amplitude (spike height) and slower (longer decay time and half-width). In conclusion, the exocytotic machinery appears to be altered in 3xTg mice, resulting in a slower release of catecholamines compared to normal conditions.

When amperometrically recorded, a simple spike with a pre-spike foot may appear (see scheme of Fig. 1a, right). The foot indicates the opening of the exocytotic fusion pore that, if stabilizes, progresses to a full pore expansion; then, the whole vesicle content (or much of it) is released [21]. The foot may stabilize longer without giving rise to a spike; this mode of exocytosis is known as “flickering” [115, 118]. If this fusion pore closes and its complete expansion is not produced, the phenomenon known as “kiss-and-run” takes place, where the release of small amounts of catecholamines occurs. The frequency of flickering events in 3xTg cells was 3 to 5 times higher than in WT cells at every age studied (Table 2). The percentage of spikes presenting foot was very similar between WT and 3xTg cells, ranging from 30 to 40% (Table 2). However, as observed in the full spike, we found smaller foot amplitudes (I_foot) and longer foot duration (T_foot) in 3xTg cells compared to WT cells (Table 2). These results highlight significant alterations in fusion pore dynamics in 3xTg cells. While the initial formation of the fusion pore occurs at a similar rate in both cell types, as evidenced by the comparable percentage of spikes with a pre-spike foot, the subsequent progression of exocytosis diverges markedly between 3xTg and WT cells.

Table 2.

Exocytotic modes and foot kinetics of single exocytotic events amperometrically recorded from WT and 3xTg CCs at different ages stimulated with ACh (100 µM) for 1 min

Age	Mouse type	Event, cells, cultures	Flickering (%)	Spikes with foot (%)	Ifoot (pA)	Qfoot (pC)	Tfoot (ms)
2 m	WT	1925, 41, 4	4.02 ± 0.98	38.75 ± 1.79	8.89 ± 0.38	0.09 ± 0.007	8.64 ± 0.44
	3xTg	913, 35, 7	20.20 ± 1.68***	36.62 ± 3.33	6.87 ± 0.67***	0.08 ± 0.010	11.64 ± 1.25
6 m	WT	1152, 33, 4	6.23 ± 1.25	29.01 ± 2.56##	9.92 ± 0.73	0.09 ± 0.010	7.97 ± 0.62
	3xTg	1046,34, 7	18.46 ± 1.94***	41.62 ± 2.88**	6.65 ± 0.56***	0.05 ± 0.005**	7.18 ± 0.54##
12 m	WT	1230, 40, 4	8.37 ± 0.81###	30.42 ± 1.63#	9.24 ± 0.34	0.06 ± 0.007###	5.41 ± 0.38###
	3xTg	1234, 30, 5	20.00 ± 1.63***	31.23 ± 1.51	6.91 ± 0.40***	0.07 ± 0.008	8.48 ± 0.75***

Data are presented as means ± SEM of the number of spikes, cells and cultures indicated

^**p ≤ 0.01; ***p ≤ 0.001 compared to WT. Data was compared using the nonparametric Mann–Whitney test for non-normally distributed data and using the parametric t-Student analysis for normally distributed data

^#p ≤ 0.05; ##p ≤ 0.01; ###p ≤ 0.001 compared to their respective mouse type (WT or 3xTg) at 2 m age. Non-normally distributed data was compared using the nonparametric Kruskal–Wallis test, with Dunns post-test and normally distributed data using the parametric one-way ANOVA analysis of variance, with the Tukey comparison test. Only cells with a minimum of 15 spikes were considered for the statistical analysis

Ion channel dysregulation in 3xTg chromaffin cells

ACh from the splanchnic nerve triggers the release of catecholamines into the bloodstream. Upon activation of nicotinic acetylcholine receptors (nAChRs), voltage-gated sodium channels open, allowing the influx of Na⁺ into the CCs. This depolarization of the CC membrane subsequently activates voltage-gated calcium channels, permitting calcium entry into the cell. The increase in intracellular calcium induces the exocytosis of catecholamines and activates voltage- and calcium-gated potassium channels, which restore the membrane potential to its resting state, thereby resetting the cell for a new activation cycle [18, 70]. As previously observed, there are significant alterations in vesicle release dynamics in 3xTg CCs compared to WT CCs. Since vesicle release depends on the activation cycle of CCs, we further investigated the ion channel currents involved in this process.

The properties of acetylcholine currents (I_ACh), sodium currents (I_Na), calcium currents (I_Ca), voltage-dependent potassium currents (I_K(v)), and calcium-dependent potassium currents (I_K(Ca)) were explored in WT and 3xTg CCs voltage-clamped at − 80 mV. To record I_ACh, cells were stimulated with 250 ms pulses of an extracellular solution containing 100 µM ACh (Fig. 2a). Typical recordings of whole-cell inward I_ACh are displayed in Fig. 2b; the current activates very fast and gradually inactivates during the ACh pulse. Averaged I_ACh peak is plotted in Fig. 2c. The most notable difference between WT and 3xTg cells was observed at 2 m, with a 31.7% decrease in I_ACh in the latter compared to the former. Additionally, the current in the 12-month WT cells decreased by about 30% when compared to the 2 m WT cells. In contrast, the I_ACh levels in the 12 m 3xTg cells remained the same as those in the 2 m 3xTg cells. These results indicate early impairments in nAChR function in the 3xTg model. At 6 m, I_ACh amplitudes increased significantly in 3xTg CCs compared to 2 m 3xTg CCs, reflecting a partial compensatory response. However, I_ACh levels remained lower than WT levels, indicating persistent deficits in cholinergic signaling. By 12 m, there is no significant difference between the two groups, suggesting convergence of nAChR function over time.

Fig. 2.

Ion channel currents in the chromaffin cell activation cycle are altered in 3xTg mice. a Scheme of the experimental approach: a recording pipette was placed in a chromaffin cell (blue), and a puffing pipette was located nearby. b Representative nicotinic acetylcholine receptor (nAChR) current with a 250 ms ACh pulse from WT and 3xTg mice at the different ages studied (scale bar: 250 pA; 500 ms). c Averaged nAChR current (I_ACh) in CCs of WT and 3xTg mice from different ages. d Top left: Voltage-clamp protocol for voltage-gated sodium channels. Square depolarizing 10 ms pulses were applied from − 60 to + 60 mV in 10 mV increments from a holding potential of − 80 mV at 15 s intervals. Average current–voltage (I–V) relationship for voltage-gated sodium channels in CCs from WT and 3xTg mice at different ages. e Top: Representative peak sodium currents (I_Na) at − 30 mV from WT and 3xTg mice at the different ages studied (scale bar: 500 pA; 5 ms). Bottom: Averaged peak I_Na in CCs of WT and 3xTg mice from different ages. I_Na peak was usually reached at approximately − 30 mV. f Top left: Voltage-clamp protocol for voltage-gated calcium channels. Square depolarizing 50 ms pulses were applied from − 50 to + 50 mV in 10 mV increments from a holding potential of − 80 mV at 10 s intervals. Average I–V relationship for voltage-gated calcium channels in CCs from WT and 3xTg mice at different ages. g Top: Representative peak calcium currents (I_Ca) at 0 mV from WT and 3xTg mice at the different ages studied (scale bar: 100 pA; 25 ms). Bottom: Averaged peak I_Ca in CCs of WT and 3xTg mice from different ages. I_Ca peak was usually reached at 0 mV. h Top left: Voltage-clamp protocol for calcium-dependent potassium channels. Square depolarizing 400 ms pulses were applied from − 40 to + 150 mV in 10 mV increments from a holding potential of − 80 mV at 20 s intervals. To isolate the calcium-dependent component of these currents, an external solution containing 2 mM Ca²⁺ was applied (2 Ca²⁺). Average I-V relationship for calcium-dependent potassium channels in CCs from WT and 3xTg mice at different ages. i Top: Representative peak calcium-dependent potassium currents (I_K(Ca)) at + 60 mV from WT and 3xTg mice at the different ages studied (scale bar: 2 nA; 100 ms). Bottom: Averaged peak I_K(Ca) in CCs of WT and 3xTg mice from different ages. I_K(Ca) peak was usually reached at + 60 mV. j Top left: Voltage-clamp protocol for voltage-dependent potassium channels. Square depolarizing 400 ms pulses were applied from − 40 to + 150 mV in 10 mV increments from a holding potential of − 80 mV at 20 s intervals. To eliminate the calcium-dependent component, an external solution with 0 mM Ca²⁺ was applied (0 Ca²⁺). Average I–V relationship for voltage-dependent potassium channels in CCs from WT and 3xTg mice at different ages. k Top: Representative peak voltage-dependent potassium currents (I_K(V)) at + 150 mV from WT and 3xTg mice at the different ages studied (scale bar: 2 nA; 100 ms). Bottom: Averaged peak I_K(V) in CCs of WT and 3xTg mice from different ages. I_K(V) peak was reached at + 150 mV. Data are presented as mean ± SEM. Each dot represents a different cell from at least three different mice. Statistical significance *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001 vs. WT; #p ≤ 0.05; ##p ≤ 0.01; ###p ≤ 0.001 vs. 2 m of the same genotype

The intensity vs. voltage (I–V) curves were generated using the specified extracellular and intracellular solutions outlined in the Materials and Methods section to record whole-cell voltage-dependent currents. I_Na was generated by 10 ms depolarizing pulses from − 60 to + 60 mV at 10 mV steps (Fig. 2d, top left). I_Na currents were smaller in 3xTg cells than WT cells, even at the pre-disease stage (Fig. 2d, e). Upon examining the superimposed averaged I–V curves obtained from WT and 3xTg CCs, it is evident that I_Na amplitudes are smaller in 3xTg compared to WT, especially at negative voltages (Fig. 2d). These results suggest a potential impairment in the depolarization phase of the activation cycle in 3xTg CCs.

I_Ca was recorded during a 50 ms depolarizing pulse at different voltages steps (Fig. 2f, top left), in the presence of an extracellular solution containing 2 mM Ca²⁺. The initial fast-inactivating component corresponds to I_Na, followed by a very slow-inactivating I_Ca. The averaged I_Ca peak values are presented in Fig. 2g. A modest decrease in I_Ca was observed in 2 m 3xTg cells compared to their WT counterparts, with I_Ca reduced from 267.8 ± 22.3 pA in 2 m WT cells to 204.7 ± 22.2 pA in 2 m 3xTg cells (p = 0.041). Interestingly, I_Ca in 3xTg cells increased as the mice aged while remaining stable in WT cells. By 6 m, I_Ca became more pronounced in 3xTg cells compared to WT. However, by 12 months, the peak I_Ca reached similar levels in both cell types, averaging around 330 pA. Averaged I–V curves obtained from WT and 3xTg CCs at different voltages are depicted in Fig. 2f. The reduction of I_Ca at 2 m in 3xTg may reflect an impairment in calcium entry into CCs at a pre-disease stage, while the increase observed at 6 m suggests a compensatory mechanism in 3xTg CCs. However, at longer times, I_Ca levels are comparable, highlighting a convergence in calcium channel function over time.

Figure 2h shows the Ca²⁺-dependent component of I_K, which was studied using 400 ms pulses in a solution containing 2 mM Ca²⁺ (top left). The outward I_K(Ca) exhibited notable age-dependent changes. Averaged peak I_K(Ca) values, measured at approximately + 60 mV, revealed a significant increase in 3xTg cells starting at 6 m compared to WT cells (Fig. 2i). This increase in I_K(Ca) was also evident at 2 m upon examining the superimposed averaged I-V curves of WT and 3xTg CCs (Fig. 2h). In addition to I_K(Ca), the voltage-dependent component of I_K was recorded using an extracellular solution devoid of calcium to isolate this current (Fig. 2j, top left). In young mice, the outward I_K(V) was comparable between WT and 3xTg cells (Fig. 2k). However, by 12 m, I_K(V) was significantly larger in 3xTg cells, particularly at test voltages exceeding + 80 mV (Fig. 2j, k). Quantitative analysis revealed that while peak I_K(V) values were approximately 3.0 nA in both WT and 3xTg cells at 2 m and 6 m, by 12 m, I_K(V) increased from 3.0 nA in WT cells to 5.0 nA in 3xTg cells (Fig. 2k). The increase in I_K may suggest enhanced repolarization mechanisms in 3xTg CCs, potentially linked to dysregulated excitability.

The data demonstrate that ion channel dysfunction in 3xTg CCs leads to an age-dependent disruption of catecholamine secretion, beginning with early deficits in I_ACh, I_Na, and I_Ca currents that reduce vesicle release and progressing to compensatory upregulation of I_K(Ca) and I_K(V) that helps stabilize the membrane potential yet fails to normalize exocytotic kinetics. Overall, these findings underscore a progressive pathological remodeling in 3xTg cells compared to WT, with compensatory mechanisms emerging over time but ultimately insufficient to restore normal neurotransmission.

Alterations in cell excitability in 3xTg chromaffin cells

Changes in ion currents can affect cell excitability and influence the generation of action potentials. This was investigated in current-clamped conditions. Figure 3a displays representative resting membrane potential (RMP) traces in single CCs from WT and 3xTg mice. There is a slight depolarization in the RMP of 3xTg CCs at 2 m and 6 m compared to WT CCs (Fig. 3b). However, this difference returns to normal values at 12 m (Fig. 3b). CCs fire spontaneous action potentials (sAP) in resting conditions. The frequency of sAPs was higher in 3xTg cells. However, the high variability in cell behaviors resulted in a lack of statistical significance despite observed differences on a quantitative basis (Fig. 3c). CCs of the adrenal medulla are physiologically stimulated by the splanchnic nerve, which releases ACh. Therefore, we subsequently examined the changes in cell excitability and AP generation following prolonged ACh stimulation (to replicate the stimulation used in the amperometry experiments, a 1-min ACh pulse was employed). Figure 3d shows representative membrane potential recordings upon an ACh pulse in WT and 3xTg CCs. ACh-evoked depolarization was significantly smaller in 2 m and 6 m 3xTg cells than WT cells (Fig. 3e). By 12 months, ACh-evoked depolarization is restored and becomes comparable to WT cells (Fig. 3e). The number of evoked action potentials (eAPs) across different ages remained similar, likely due to the great variability among individual cells, which reflects population heterogeneity (Fig. 3f). Even within the same age group, individual cells exhibited significantly different responses to ACh, potentially masking subtle age-dependent trends when data were averaged across the population. These findings indicate that while 3xTg CCs initially exhibit significantly reduced ACh-evoked depolarization compared to WT, compensatory adaptations restore depolarization by 12 months, with the overall number of evoked action potentials remaining similar across ages, likely due to inherent cellular variability. This adaptive response supports neurotransmission by restoring depolarization and maintaining similar levels of evoked action potentials over time.

Fig. 3.

Early alterations in chromaffin cell excitability in 3xTg mice are compensated over time. a Representative resting membrane potential recordings in WT (top) and 3xTg (bottom) chromaffin cells at 2, 6 and 12 months of age (scale bar: 60 s). The dashed line represents each genotype's resting membrane potential (RMP) at 2 m. b Averaged RMP in CCs of WT and 3xTg mice from different ages. c Averaged spontaneous action potential firing (sAP) in CCs of WT and 3xTg mice from different ages. d Representative membrane potential recordings in WT (top) and 3xTg (bottom) chromaffin cells at 2, 6 and 12 months of age with a 1-min ACh (100 µM) pulse (scale bar: 10 s). The dashed line represents each genotype's ACh evoked depolarization at 2 m. e Averaged membrane potential depolarization from its resting state to the maximum depolarization after ACh was applied in CCs of WT and 3xTg mice from different ages. f Averaged action potential firing evoked during the 1-min pulse of ACh (eAP) in CCs of WT and 3xTg mice from different ages. Data are presented as mean ± SEM. Each dot represents a different cell from at least three different mice. Statistical significance *p ≤ 0.05; **p ≤ 0.01 vs. WT; ##p ≤ 0.01 vs. 2 m of the same genotype

Alterations in the actin cytoskeleton in 3xTg chromaffin cells

The actin cytoskeleton plays a key role in regulating exocytosis by facilitating the transport and maturation of secretory vesicles and ensuring their passage to the plasma membrane [49, 88]. In chromaffin cells, F-actin forms a dense network beneath the membrane, actively participating in vesicle dynamics. Disruption of this network has been shown to increase exocytosis, highlighting its functional relevance [30, 84]. Based on this, we investigated whether the exocytosis alterations observed in 3xTg mice could be linked to changes in the actin cytoskeleton. To this end, we conducted an immunofluorescence study on cultured CCs from WT and 3xTg mice at 2, 6, and 12 months of age. The cells were stained with phalloidin (which binds to the F-actin cytoskeleton) and DAPI (which stains the cell nucleus), and images were acquired using confocal microscopy (Fig. 4a, b). We first explored potential alterations in the actin cytoskeleton density during aging and Alzheimer's disease progression. The actin cytoskeleton remained stable across all ages in WT cells. However, 3xTg cells exhibited a significant reduction in F-actin intensity starting at 6 months of age (Fig. 4c, left; p < 0.001). The distribution of F-actin also differed significantly. In WT mice, F-actin consistently formed a well-defined cortical ring beneath the plasma membrane at all ages. In contrast, chromaffin cells from 3xTg mice progressively lose the cortical ring, resulting in a more homogeneously distributed F-actin throughout the cytosol. This homogeneous distribution is very apparent in 3xTg mice of 12 m of age. Quantitative analysis of the cortical/cytosolic phalloidin intensity ratio confirmed a significant decrease in 3xTg cells at 12 m (Fig. 4c, right; p = 0.001) compared to WT controls. Additionally, this ratio was significantly lower in older 3xTg mice when compared to younger animals of the same strain (Fig. 4c, right; p < 0.001).

Fig. 4.

Progressive disruption of the actin cytoskeleton in 3xTg chromaffin cells. a Scheme of the experimental approach: chromaffin cells were stained to visualize the actin cytoskeleton (red filaments) using Phalloidin 546 and nuclei (blue) with DAPI. b Pseudocolor confocal images showing WT CCs (top) and 3xTg CCs (bottom) at different ages stained with Phalloidin 546 to label F-actin (red) and DAPI to label the nucleus (blue) (scale bar: 5 µm). c Left. Averaged F-actin intensity was measured in arbitrary units (AU) in CCs of WT and 3xTg mice of different ages. Right. Averaged cortical/cytosolic phalloidin intensity ratio in CCs of WT and 3xTg mice from different ages. Data are presented as mean ± SEM. Each dot represents a different cell from 4 different mice for each genotype and age. Statistical significance *p ≤ 0.05; ***p ≤ 0.001 vs. WT; ###p ≤ 0.001 vs. 2 m of the same genotype

These results indicate that both the density and spatial organization of the actin cytoskeleton are progressively disrupted in 3xTg chromaffin cells, particularly at later stages of Alzheimer's disease. Such alterations may underline the exocytosis deficits previously observed in this model.

Neurotransmitter levels are altered in 3xTg mice

Due to the increased release of catecholamines observed in 3xTg CCs with Alzheimer’s disease progression (Fig. 1c, d), we questioned whether this increase affected the release of different catecholamines and/or other neurotransmitters equally. To address this question, we measured adrenaline (ADR), noradrenaline (NA), dopamine (DA), and serotonin (SER) levels using high-performance liquid chromatography-mass spectrometry (HPLC–MS) (Fig. 5a). These measurements were performed on plasma samples—since CCs release catecholamines into the bloodstream—and in the prefrontal cortex and hippocampus, regions widely studied for their alterations in Alzheimer’s disease. We found no difference in ADR levels across all ages and mice strains (Fig. 5b). However, a significant increase in NA levels was observed in 12 m 3xTg mice (Fig. 5c; p = 0.001). Although this increase appeared to begin at 6 m, the variability in the data for these animals prevented statistical significance from being reached. Regarding DA levels, significant differences were observed starting at 6 m, with lower plasma DA levels in 3xTg mice compared to WT mice at 6 and 12 months (Fig. 5d). SER levels also showed significant differences between these two ages, being lower in 3xTg mice than in WT values. These differences were significant at 6 m (Fig. 5e; p = 0.001) and at 12 m (Fig. 5e; p = 0.029). These findings reveal differential changes in catecholamine and neurotransmitter levels in 3xTg mice, characterized by increased noradrenaline and decreased dopamine and serotonin with disease progression. Such alterations may contribute to the pathophysiology of Alzheimer’s disease and could play a role in modulating stress responses, mood regulation, and synaptic dysfunction commonly observed in the disease.

Fig. 5.

Altered plasma catecholamine and serotonin levels in 3xTg mice during Alzheimer’s disease progression. a Schematic representation of the experimental approach: blood was collected from WT and 3xTg mice. After plasma separation, neurotransmitter levels were measured using high-performance liquid chromatography-mass spectrometry (HPLC–MS). b Adrenaline levels in plasma from WT and 3xTg mice at different ages. c Noradrenaline levels in plasma from WT and 3xTg mice at different ages. d Dopamine levels in plasma from WT and 3xTg mice at different ages. e Serotonin levels in plasma from WT and 3xTg mice at different ages. Data are presented as mean ± SEM. Each dot represents an individual plasma sample, with at least three different mice per genotype and age group. Statistical significance: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001 vs. WT; #p ≤ 0.05; ##p ≤ 0.01; ###p ≤ 0.001 vs. 2-month-old mice of the same genotype

In addition to plasma, neurotransmitter levels were measured in brain tissue from all experimental groups, specifically in the hippocampus and prefrontal cortex (Supplementary Fig. 2). As in the plasma, we found no changes in ADR levels either in the prefrontal cortex (Supplementary Fig. 2b) or the hippocampus (Supplementary Fig. 2g). However, we found a significant increase in NA levels in transgenic mice at both 6 m and 12 m (Supplementary Fig. 2c, h). Surprisingly, DA levels were also higher in 12-month-old 3xTg mice compared to controls (Supplementary Fig. 2d, i). This contrasts with the reduction observed in the plasma. Conversely, and similar to plasma findings, SER levels in the hippocampus and prefrontal cortex are lower in 3xTg mice at 6 m and 12 m (Supplementary Fig. 2e, j).

These findings indicate region-specific neurotransmitter alterations in AD. While ADR levels remained unchanged, the differential patterns observed for NA, DA, and SER across plasma and brain tissue suggest distinct regulatory mechanisms, reinforcing the complexity of neurotransmitter dysregulation in the disease progression.

Dysregulation of the anxiety state in 3xTg mice

Lastly, we wanted to investigate whether all the alterations found in the CCs of 3xTg mice could translate into changes in the animal's behavior. Since catecholamine release from the adrenal medulla is directly related to the fight-or-flight response, we conducted various tests to assess the animals'anxiety levels.

The open field test [54] allows us to analyze a wide range of behavioral parameters in mice by placing them in a novel box, which they tend to explore. In our study, we used this test to evaluate the anxiety levels of WT and 3xTg animals at the three studied ages. To do so, we measured the time the animals spent exploring the area closest to the walls of the box (an innate thigmotactic behavior, meaning they stay close to the walls of the open field to feel safer) compared to the time they spent in the inner zone of the box, where they had no contact with the walls. As shown in Fig. 6a–f, this anxiety-related behavior is altered in the transgenic 3xTg model, even before the onset of cognitive impairment pathology. Specifically, 3xTg mice covered a significantly greater distance in the inner zone of the box (Fig. 6b, c) as early as two months of age, maintaining this difference across all three studied age groups. This observation happened with no major changes in the distance the mice covered close to the walls (Fig. 6d). 3xTg animals spent more time in the inner zone of the box (Fig. 6e), indicating that their expected thigmotactic behavior is altered. As anticipated, the number of entries into the inner zone was also greater in 3xTg mice across all three ages (Fig. 6f).

Fig. 6.

Altered anxiety- and depressive-like behaviors in 3xTg mice during Alzheimer’s disease progression. a Schematic representation of the open field test (OFT), used to assess anxiety-like behavior by measuring thigmotactic behavior and exploration tendencies. b Representative tracking plots (left) and heat maps (right) of a 12-month-old (12 m) WT and 3xTg mice in the OFT. c Quantification of the total distance traveled (cm) in the inner zone of the open field arena across different ages. d Distance traveled (cm) along the walls of the open field. e Time spent (s) in the inner zone of the arena. f Number of entries into the inner zone. g Schematic representation of the elevated plus maze (EPM), consisting of two open and two closed arms, used to assess anxiety-related behavior. h Time spent in the open arms of the elevated plus maze by 12 m WT and 3xTg mice. i Time spent in the closed arms of the maze for 12 m WT and 3xTg mice. j Time spent at the edges of the open arms, an indicator of increased exploratory behavior and potential disinhibition of fear responses. k Schematic of the tail suspension test (TST), used to evaluate depressive-like behavior by measuring immobility time. l Percentage of immobility time in the TST across different ages. Data are presented as mean ± SEM. Each dot represents an individual mouse. Statistical significance: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001 vs. WT; #p ≤ 0.05 vs. 2 m of the same genotype

To confirm the observed alterations in the anxiety-related behavior of transgenic mice, we additionally performed the elevated plus maze test [96] on WT and 3xTg mice at 12 months of age. Briefly, the maze features an elevated cross-shaped structure with two open and two enclosed arms (Fig. 6g). The maze is positioned 40 cm above the ground. Mice are placed in the center of the maze and allowed to freely explore for 10 min. Typically, animals explore novel environments but avoid elevated, open areas due to predation risk. We observed that 3xTg mice spent more time in the open arms of the maze than control animals (Fig. 6h) without changes in the time spent in the closed arms (Fig. 6i). The time spent exploring the edges of the open arms of the maze was also greater in 3xTg mice than in WT mice at 12 m (Fig. 6j). Increased-edge exploration could indicate a disinhibition of fear responses, altered decision-making, or a deficit in perceiving environmental threats. These results, along with those from the open field test showing that 3xTg mice spend more time in the center, further support the idea that these animals display reduced anxiety or modified risk assessment. Thus, anxiety-like behavior appears to be altered (likely decreased) in 3xTg mice.

To evaluate potential depressive-like behaviors related to the progression of Alzheimer’s pathology, we performed the tail suspension test (TST). This test was developed by Steru et al. [104] as an alternative to the Forced Swim Test. The test is based on the innate escape response of mice when suspended by their tails. Initially, the animals actively try to escape, but after repeated unsuccessful attempts, they become immobile (Fig. 6k). The duration of this immobility is considered an indicator of behavioral despair, with longer immobility times reflecting greater depressive-like behavior. As shown in Fig. 6l, starting from six months of age, 3xTg mice exhibited a significantly greater percentage of immobility time—about twice that of their respective controls—indicating that Alzheimer’s mice displayed depressive-like behaviors from six months onward, which were not seen in WT mice.

Taken together, our behavioral analysis revealed significant alterations in anxiety- and depressive-like behaviors in 3xTg mice. These findings suggest that catecholaminergic dysfunction in 3xTg mice is accompanied by distinct behavioral alterations, reinforcing the link between neurochemical imbalances and neuropsychiatric symptoms observed in Alzheimer’s disease.

Discussion

This study uncovers a previously overlooked aspect of Alzheimer’s disease pathology: the early and progressive dysfunction of adrenal chromaffin cells and its potential impact on systemic catecholaminergic signaling. While AD research has mainly focused on the CNS, our findings emphasize the role of the peripheral nervous system in disease-related changes. Using electrophysiological, amperometric, and neurochemical analyses, we show that CC impairments occur before cognitive symptoms and worsen over time. These changes are associated with significant neurotransmitter imbalances in plasma and brain regions, indicating broader neurochemical dysregulation. We discuss the implications of these findings for understanding AD-related neuropsychiatric symptoms and autonomic dysfunction, as well as their potential for identifying new biomarkers and therapeutic targets.

We observed a pronounced impairment of the quantal release of catecholamines in 2 m 3xTg mice, which are cognitively normal at this age (Supplementary Fig. 1 [2, 6, 92];). This suggests exocytotic dysfunction at a pre-symptomatic stage, aligning with the current concept of AD as a continuum beginning with silent pathology [1, 36]. We confirmed the presence of brain pathology in our transgenic mouse model, detecting Aβ plaques and phosphorylated tau accumulation in the hippocampus (Supplementary Fig. 3a, b). In contrast, Aβ was undetectable in adrenal CCs, though phosphorylated tau was elevated (Supplementary Fig. 3c–f), suggesting that adrenal changes are tau-driven rather than Aβ-dependent.

Our data indicate that peripheral alterations occur before disease onset, as shown by exocytosis impairment in 2 m 3xTg mice with normal cognition (Fig. 1b, c). Altered ion channel activity may underlie the exocytotic impairment. During stressful conflicts, the release of catecholamines at the sympatho-adrenal axis is triggered by ACh-elicited action potentials at both the sympathetic ganglia and the adrenal medullary CCs [18, 25, 32]. ACh depolarizes the CC to open voltage-gated Na⁺ and Ca²⁺ channels, allowing Ca²⁺ entry and the explosive release of adrenaline and noradrenaline [33, 34]. Early-stage deficits in I_ACh, I_Na, and I_Ca currents (Fig. 2b–g) are consistent with a lower vesicle release and catecholamine secretion (Fig. 1b–d). At 6-month, these parameters normalize, possibly due to a partial recovery in nicotinic and calcium currents (Fig. 2c, g). The upregulation of I_K(Ca) may stabilize the membrane potential, manage the prolonged calcium entry, and support repetitive vesicle release cycles, though release kinetics remain abnormal (Fig. 1 and Tables 1 and 2). I_K(Ca) and I_K(V) significantly increase in 12 m 3xTg compared with WT CCs (Fig. 2h–k), potentially stabilizing membrane potential and facilitating cell repolarization during sustained stimulation. This supports functional voltage-gated channels and quick resets for depolarization during prolonged activity, contributing to the higher vesicle release in 3xTg CCs (Fig. 1c) without directly impacting the exocytotic machinery. The lack of change in I_ACh implies that cholinergic stimulation is likely not a major contributor to the enhanced exocytosis in 3xTg cells at this stage. Although I_Na is lower than in WT CCs and I_Ca remains unchanged, both currents significantly increase with disease progression in 3xTg CCs (See teal columns of Fig. 2e, g), indicating compensatory mechanisms to support neurotransmission with aging. While I_Na remains relatively stable in WT CCs, I_Ca shows a tendency to increase with age. This contrast highlights the pathological remodeling in 3xTg cells, but also points to aging-related adaptations in ion channel function. Interestingly, although the increase in I_Ca at 12 months did not reach statistical significance in WT chromaffin cells, a clear trend was observed. This is consistent with our unpublished observations in aging mouse models, where a significant increase in I_Ca occurs with age. Such changes may reflect compensatory mechanisms to support catecholamine release but could also interact with age-related alterations in excitability, ultimately affecting stimulus-secretion coupling. These findings suggest that aging alone may influence calcium dynamics in chromaffin cells, even in the absence of pathology.

Beyond ion channel dysfunction and its compensatory mechanisms, it is important to consider how these alterations affect cholinergic signaling. Changes in ACh-evoked depolarization and action potential firing provide further insight into the functional consequences of these adaptations in 3xTg CCs. The ACh-evoked depolarization and the triggering of AP bursts were transient and tended to fade during the 1-min pulse (Fig. 3d), likely due to the desensitization of nAChRs and/or ion channel inactivation during prolonged stimulation. At 2 and 6 months, ACh-evoked depolarization was significantly smaller in 3xTg cells than WT cells (Fig. 3e). This can be attributed to reductions in I_ACh, I_Na, and I_Ca currents in 3xTg CCs at early stages. By 12 months, ACh-evoked depolarization is restored and comparable to WT cells (Fig. 3e), aligning with increased vesicle release and catecholamine secretion at this stage. The number of evoked action potentials across various age groups stayed relatively consistent, probably because of the high variability seen in individual cells (Fig. 3f). Altogether, these findings suggest that early impairments in excitability—reflected in changes in resting membrane potential and depolarizing responses—may weaken stimulus-secretion coupling in 3xTg CCs, contributing to peripheral catecholaminergic dysfunction with potential relevance to AD-associated symptoms.

Additionally, the significant disruption in actin cytoskeleton organization observed during disease progression (Fig. 4) may also contribute to the vesicle secretion alterations in CCs. The actin cytoskeleton plays a crucial role in vesicle trafficking, docking, and exocytosis in CCs, and its dysregulation has been implicated in various neurodegenerative diseases, including AD. Actin filaments are known to form a dynamic cortical barrier that regulates the movement of secretory vesicles toward the plasma membrane, facilitating neurotransmitter release [74, 80, 89–91, 101, 109, 115]. Additionally, actin remodeling is tightly controlled by signaling pathways that modulate synaptic plasticity and vesicle dynamics [31].

Previous studies have shown that tau pathology disrupts actin organization, as hyperphosphorylated tau binds to actin filaments, leading to cytoskeletal instability and impaired intracellular trafficking [5, 41]. The marked increase in hyperphosphorylated tau observed in the adrenal medulla of 3xTg mice (Supplementary Fig. 3e, f) may contribute to the actin cytoskeleton alterations detected in these cells (Fig. 4). Given that actin remodeling is essential for vesicle fusion and release, tau-related cytoskeletal changes likely contribute to secretory dysfunction in 3xTg CCs. Similar findings have been reported in neuronal models, where actin dysregulation leads to defective synaptic vesicle recycling and reduced neurotransmitter release [8, 46, 79, 93].

Consistently, our immunohistochemical analysis showed that cytoskeletal alterations at 12 months coincided with pTau accumulation in 3xTg CCs. In contrast, WT mice displayed only weak pTau immunoreactivity and no amyloid pathology. These findings reinforce the link between tau pathology and cytoskeletal disruption in CCs during disease progression. Disruptions in neurotransmitter release at the cellular level can have widespread consequences on neural circuit function, potentially contributing to the neurotransmitter imbalances observed in various brain regions and systemic circulation in 3xTg mice (Fig. 5 and Supplementary Fig. 2). Consistently, we found significant alterations in neurotransmitter levels in the hippocampus, cortex, and plasma of these mice, further supporting the idea that vesicle trafficking defects may play a key role in the pathophysiology of the disease. While ADR levels remained unchanged (Fig. 5b and Supplementary Fig. 2b, g), NA levels were significantly increased in both brain and plasma (Fig. 5c and Supplementary Fig. 2c, h), suggesting a systemic adrenergic dysregulation. This observation is consistent with previous reports indicating that noradrenergic dysfunction, particularly the loss of neurons in the locus coeruleus (LC), correlates with cognitive impairment and neuroinflammation in AD [52]. Although human studies consistently report reduced NA levels in AD brains due to LC degeneration [40, 59, 75, 77, 116], we found elevated NA in aged 3xTg mice. However, many AD mouse models lack significant LC degeneration, possibly explaining divergent NA levels. Indeed, reports vary widely, with some models showing decreased, stable, or increased NA levels [42]. One possibility is that residual LC neurons compensate for early NA loss, temporarily maintaining levels. With disease progression, this compensation may fail, leading to depletion. Alternatively, overcompensation may transiently elevate NA before eventual decline [43]. This dynamic may explain discrepancies across studies.

DA is a key neurotransmitter regulating reward, movement, and cognition [20]. In the 3xTg mouse model, we showed a significant increase in DA in the hippocampus and prefrontal cortex, accompanied by a decrease in plasma DA levels (Fig. 5d and Supplementary Fig. 2d, i). This contrasts with human AD studies, which often report dopaminergic loss in these regions due to degeneration of substantia nigra and ventral tegmental area projections [12, 47, 106]. Similar to NA, this increase may reflect an early-stage compensatory response to neuronal dysfunction, potentially involving upregulation of DA synthesis or receptor sensitivity. However, as the disease progresses, these compensatory mechanisms may become insufficient, leading to the dopaminergic deficits observed in advanced AD stages. The concomitant decrease in plasma DA levels suggests systemic dysregulation of dopaminergic signaling, which could be attributed to altered peripheral metabolism or impaired dopamine transport across the blood–brain barrier [20, 85, 111]. These studies provide insight into the complex role of dopamine in AD pathology and underscore the importance of considering both central and peripheral alterations in dopaminergic systems when investigating disease mechanisms and potential therapeutic approaches.

SER was consistently reduced in plasma and brain tissue, suggesting widespread serotonergic dysregulation. Reduced SER has been implicated in mood disturbances and cognitive decline in AD [78, 103], and its systemic decrease may exacerbate disease symptoms. Together, these findings highlight the intricate neurotransmitter imbalances in AD, emphasizing the need for further studies to explore their mechanistic underpinnings and therapeutic implications.

Importantly, our findings suggest that these peripheral catecholaminergic alterations are not merely epiphenomena but may have direct functional consequences on behavior. Catecholamines, particularly NA and DA, play key roles in modulating arousal, stress responses, and anxiety-related behaviors. The early and persistent impairments in chromaffin cell exocytosis and the resulting plasma catecholamine imbalances in 3xTg mice may therefore contribute to the altered behavioral phenotypes observed in our anxiety and depression paradigms.

Given the central role of catecholamines in regulating stress responses, we investigated anxiety- and depression-related behaviors in 3xTg mice to determine the functional impact of these neurochemical imbalances. Anxiety-like behavior was altered in 3xTg mice from an early age. In the open field test, WT mice displayed the expected thigmotactic behavior, spending most of their time near the walls and occasionally exploring the center of the arena. However, 3xTg mice showed a significant increase in time spent in the inner zone of the box, as well as a greater number of entries into this area, even at two months of age (Fig. 6a–f). This suggests an altered risk assessment and exploratory drive, which persisted across all ages studied. These findings were further supported by results from the elevated plus maze test, where 3xTg mice spent more time in the open arms and at the edges of the maze, behaviors typically avoided due to their association with increased vulnerability to predators (Fig. 6g–j). This disinhibited behavior has been described in other AD mouse models [66, 102] and is reminiscent of impulsivity, reduced threat perception, and socially inappropriate actions observed in AD patients [22].

Mechanistically, the observed increase in peripheral NA, alongside reduced DA levels in plasma, could underlie the disinhibited and risk-prone behaviors seen in 3xTg mice. Noradrenaline is known to heighten arousal and promote anxiety-like responses [10], yet paradoxically, dysregulated noradrenergic signaling, especially when coupled with impaired feedback from the central nervous system, can result in abnormal threat assessment and impulsivity [3, 95, 107] as reflected by increased exploration of open and exposed areas in both the open field and elevated plus maze tests (Fig. 6a–j). Furthermore, the reduction in peripheral dopamine may exacerbate deficits in motivational and reward-related behaviors [20], potentially contributing to the altered exploration and increased behavioral activation observed.

Peripheral catecholamines are increasingly recognized as modulators of CNS function via the vagus nerve, the hypothalamic–pituitary–adrenal (HPA) axis, or even direct transport across the blood–brain barrier [50, 58, 108, 110]. Thus, the early and progressive dysfunction in adrenal chromaffin cells and the resulting systemic catecholaminergic imbalance may not only reflect but also actively contribute to the neuropsychiatric symptoms of AD, such as anxiety and depression, by disrupting central neurotransmitter dynamics and stress responses.

The persistent serotonergic deficit detected in both plasma and brain tissue also aligns with the increased immobility time in the tail suspension test, a marker of depression-like behavior (Fig. 6k–l). Serotonin’s role in mood regulation is well established, and systemic reductions may impact both central and peripheral circuits involved in affective processing.

Interpreting anxiety-related behaviors in 3xTg mice remains complex, as the literature shows conflicting results. Many studies use commercially available C57BL/6 mice as controls, whereas our study employs WT mice with the same genetic background as 3xTg, providing more accurate comparisons and increasing reliability.

In addition to anxiety-related alterations, 3xTg mice exhibited depressive-like behaviors in the tail suspension test. From six months of age, these mice displayed significantly increased immobility times, indicating higher behavioral despair than WT controls (Fig. 6k, l). This supports previous reports of depressive-like phenotypes in both Aβ- and tauopathy-based AD models [64, 97]. These behaviors parallel the high prevalence of depression in AD patients and highlight the relevance of 3xTg mice for studying AD-associated neuropsychiatric symptoms [22].

Taken together, our behavioral analysis indicates that neurochemical dysfunction in 3xTg mice is accompanied by significant alterations in anxiety- and depression-related behaviors, highlighting a potential connection between catecholaminergic dysregulation and neuropsychiatric symptoms in AD.

We acknowledge that this study was conducted solely in the 3xTg AD mouse model, without human data. The model was used to investigate peripheral changes over time and gain insight into non-CNS aspects of AD. While the animal model provided experimental control, this remains a limitation. Future studies should validate our findings in human samples to confirm their translational relevance.

Our findings also challenge the traditional CNS-centric paradigm of Alzheimer’s disease by demonstrating that early peripheral catecholaminergic dysfunction occurs prior to cognitive decline and central pathology in the 3xTg mouse model. This suggests that peripheral neurosecretory alterations may play a more integral role in AD pathogenesis than previously recognized. To translate these findings and assess their clinical relevance, longitudinal studies in at-risk populations (e.g., individuals with familial AD mutations, mild cognitive impairment, or early amyloid/tau pathology) could assess plasma catecholamine profiles, adrenal function, and chromaffin cell biomarkers over time. If early peripheral dysfunction is validated in humans, it could act as a minimally invasive biomarker for early diagnosis or risk stratification and may also be a target for preventive interventions.

Additionally, although this study focused exclusively on male 3xTg mice to ensure consistency across age and experimental conditions, future research must investigate whether similar alterations occur in females. Exploring potential sex-specific differences in adrenal medulla function, neurotransmitter regulation, and behavioral phenotypes will be essential to enhance the translational relevance of these findings.

In conclusion, we show that chromaffin cells from 3xTg mice undergo profound dysfunction affecting vesicle dynamics, ion channels, cytoskeletal integrity, and neurotransmitter release. These changes arise early and worsen with age, reflecting AD’s progressive nature. Neurochemical imbalances correlate with altered anxiety- and depression-like behaviors, suggesting a connection between peripheral catecholaminergic dysfunction and neuropsychiatric symptoms in AD. Our findings propose that the adrenal medulla and peripheral catecholaminergic system play a larger role in AD pathophysiology than previously recognized. These alterations may contribute to systemic stress dysregulation and exacerbate disease progression. Understanding these peripheral changes is crucial—they may offer new avenues for biomarker discovery and therapeutic intervention targeting the autonomic nervous system.

Author contributions

CN and LG designed the experiments. CN and IC performed and analyzed electrophysiology and amperometry studies. AMB analyzed electrophysiology recordings. AM performed and analyzed immunoblotting and behavioral studies. AW and ARN performed and analyzed HPLC studies. CN generated the figures. The first draft of the manuscript was written by CN. LG, AGM, RdP, CN and AGG contributed to critical discussion and edited the manuscript. All authors commented on the manuscript. All authors read and approved the final manuscript.

Funding

Spanish Ministry of Economy and Competitiveness, SAF2016-78892-R, SAF2016-78892-R, Ministerio de Ciencia e Innovación, PID2020-117127RB-I00, PID2020-117127RB-I00.

Availability of data and materials

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Experiments were conducted according to the recommendation of the Ethics Committee from Universidad Autónoma de Madrid on the use of animals for laboratory experimentation, in accordance with the code of ethics and guidelines established by the European Community Directive (2010/63/EU) and Spanish legislation (RD53/2013). All efforts were made to avoid animal suffering and to use the minimum number of animals allowed by the experimental protocol and the statistical power of group data. Mice were housed under controlled temperature, a 12:12 h light cycle, and food and water were provided ad libitum.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.