L-type calcium channel antagonist isradipine age-dependently decreases plaque associated dystrophic neurites in 5XFAD mouse model

Abstract

Epidemiological studies suggest that L-type calcium channel (LTCC) antagonists may reduce the incidence of age-associated neurodegenerative diseases including Alzheimer’s disease (AD). However, the neuroprotective mechanism of LTCC antagonists is unknown. Amyloid-β (Aβ) pathology disrupts intracellular calcium signaling, which regulates lysosomes and microglial responses. Neurons near Aβ plaques develop dystrophic neurites, which are abnormal swellings that accumulate lysosomes. Further, microglia accumulate around Aβ plaques and secrete inflammatory cytokines. We hypothesized that antagonism of LTCCs with isradipine would reduce Aβ plaque-associated dystrophic neurites and inflammatory microglia in the 5XFAD mouse model by restoring normal intracellular calcium regulation. To test this hypothesis, we treated 6- and 9-month-old 5XFAD mice with isradipine and tested behavior, examined Aβ plaques, microglia, and dystrophic neurites. We found that isradipine treatment age-dependently reduces dystrophic neurites and leads to trending decreases in Aβ but does not modulate plaque associated microglia regardless of age. Our findings provide insight into how antagonizing LTCCs alters specific cell types in the Aβ plaque environment, providing valuable information for potential treatment targets in future AD studies.

1. Introduction

1.1. Calcium dysregulation in Alzheimer’s disease

Aging is the main risk factor for Alzheimer’s disease (AD), the most common form of dementia (“2022 Alzheimer’s disease facts and figures, ” 2022). Intracellular calcium (Ca²⁺) dysregulation is a shared feature between aging and AD (Berridge, 2010; Khachaturian, 1987). The two key pathological features of AD, amyloid β (Aβ) aggregates and tau neurofibrillary tangles are linked to intracellular Ca²⁺ dysregulation. Neurons, astrocytes, and microglia show increased Ca²⁺ dysregulation near plaques (Brawek et al., 2014; Busche et al., 2008; Kuchibhotla et al., 2008). This Ca²⁺ dysregulation can in turn lead to defective neuroplasticity (Kuchibhotla et al., 2008) and induction of additional amyloid and tau pathology (Pierrot et al., 2006) contributing to cognitive deficits in AD (Bezprozvanny and Mattson, 2008). Around Aβ plaques, microglia display Ca²⁺ dysregulation (Brawek et al., 2014) and take on a neurodegenerative phenotype, characterized by toxic gain-of function (e.g., proinflammatory cytokine production) and loss-of-function (e.g., reduced ability to remove Aβ); these functions are regulated by intracellular Ca²⁺ (Bolmont et al., 2008; Brawek and Garaschuk, 2013; Hickman et al., 2008). Dystrophic neurites are also found in the vicinity of Aβ plaques which are characterized as bulbous protrusions from neurons that accumulate dysfunctional autophagic vesicles and lysosomes (Gowrishankar et al., 2015). The accumulated autophagic vesicles and lysosomes are impacted by intracellular Ca²⁺ dysregulation which diminishes the ability to properly degrade debris, destabilizes lysosomal membrane and possibly leads to increase in Aβ production in the lysosomes (McBrayer and Nixon, 2013). Compounding Aβ-induced Ca²⁺ dysregulation, Ca²⁺ can increase Aβ production from amyloid precursor protein (APP) (Querfurth et al., 1997; Querfurth and Selkoe, 1994) and accelerate formation of Aβ fibrils (Isaacs et al., 2006). This feedforward cycle of intracellular Ca²⁺ dysregulation and Aβ is a major driver of AD pathology (LaFerla, 2002), therefore by intervening or stopping the cycle this could abrogate AD pathology.

1.2. L-type calcium channels in AD

There are several potential sources of Ca²⁺ dysregulation in amyloid pathology. Aβ can create pores within the plasma membrane allowing Ca²⁺ to flow into the cell (Arispe et al., 1993; Strodel et al., 2010) or can modulate the activity of other channels on the on the plasma membrane including Ca²⁺ ATPases, sodium/Ca²⁺ exchanger, NMDA receptors, or G-protein coupled receptors in addition to organelle resident channels such as the ryanodine receptors or the mitochondrial calcium uniporter (Stutzmann, 2007). Of note are long lasting-type calcium channels (LTCCs). LTCCs are present on multiple brain cell types including neurons and glia and regulate Ca²⁺ influx directly and indirectly by triggering release of intracellular Ca²⁺ stores (Catterall, 2011; Hopp, 2021). Aβ increases the expression (Daschil et al., 2014; Kim and Rhim, 2011) and activity (Lopez et al., 2008) of LTCCs on neurons and other cells. Aβ induces neurotoxicity via LTCCs (Weiss et al., 1994) and disrupts normal synaptic activity via LTCC (Gholami Pourbadie et al., 2016). Further, Aβ induces LTCC-dependent microglial production of inflammatory mediators (Sanz et al., 2012). Epidemiological studies suggest that LTCC antagonists could protect against the development of AD pathology or clinical dementia (Forette et al., 2002; Hoffman et al., 2009; Yasar et al., 2005). Despite promising evidence from epidemiological and preclinical studies, interventional clinical trials with LTCC antagonists have shown mixed results, with some studies demonstrating clinical benefit and others showing either minimal or no benefit in AD and dementia patients (Lawlor et al., 2018; López-Arrieta and Birks, 2000; Morich et al., 1996; Tollefson, 1990). To better understand the treatment potential of LTCCs antagonists in AD, which may be cell specific, we tested the hypothesis that antagonism of LTCCs with isradipine would reduce Aβ plaque associated dystrophic neurites and inflammatory microglia in the 5XFAD mouse model of AD by restoring normal intracellular Ca²⁺ regulation.

2. Materials and methods

2.1. Experimental design

5XFAD mice and wild-type (WT) littermates, both sexes, were aged to 6 or 9 months and then treated with the LTCC antagonist isradipine at a dose of 3 mg/kg/day, delivered by subcutaneously implanted extended-release pellets over the course of thirty days. 6-month group had 16 WT: vehicle: 8 (males 2, females 6); isradipine 8 (males 3, females 5) and 24 5XFAD: vehicle 11 (males 6, females 5); isradipine 13 (males 7, females 6). 9-month group had 18 WT: 9 vehicle (males 5, females 4); 9 isradipine (males 4, females 5) and 22 5XFAD: vehicle 10 (males 5, females 5); 11 isradipine (males 6, females 5). Several isradipine delivery methods were evaluated prior to the selection of subcutaneous extended-release pellets (Section 2.3). Mice underwent behavioral testing over the course of 9 days starting on day 20 of treatment. Behavior included habituation, open field (OF), novel object recognition (NOR), and Morris water maze (MWM). After the final day of behavior mice rested for 24 h prior to euthanasia and tissue processing (Fig. 1A and B).

Fig. 1. Experimental Design.

A. Mice at either 6- or 9-month-old of age were implanted with subcutaneous pellets to deliver isradipine at 3 mg/kg/day for a total of 30 days. On day 20 of treatment mice underwent behavior: OF, NOR and MWM. Mice had a 24-h rest period prior to sacrifice B. For tissue collection brain was hemisected, left hemisphere was postfixed for IHC. IHC looked at two different co-stains: DAPI, LAMP1 and D54D2 or Iba1 and ThioS. Right hemisphere was used for RNA and protein isolation. Blood and right cerebellum were collected for LC/MS/MS analysis.

2.2. Mice

The University of Texas Health San Antonio Institutional Animal Care and Use Committee approved all mouse experiments following national and institutional guidelines. Mice were grouped housed in 14-h light and 10-h dark cycle in ventilated cages with ad libitum access to food and water. Experimental mice were generated by breeding 5XFAD hemizygous mice (Jackson Labs, # 034848-JAX) with C57BL/6 J mice (Jackson Labs, #000664) providing hemizygous 5XFAD mice and WT littermates on a congenic C57BL/6 J background. Mice were genotyped by polymerase chain reaction and gel electrophoresis following Jackson Labs protocol. 5XFAD transgenic mice express APP with three AD mutations and presenilin 1 with two AD mutations. 5XFADs develop amyloid plaque pathology as early as two months of age (Oakley et al., 2006) and Ca²⁺ dysregulation is evident as early as 4 months of age (Kaczorowski et al., 2011; Rosales Jubal et al., 2021; Yao et al., 2020). Starting at 8–12 weeks of age, mice were handled for 1 min once per week to reduce stress during behavior.

2.3. Isradipine treatment

Isradipine (CAS #75695–93-1, purchased from Sigma and BOC Sciences), which crosses the blood brain barrier (Uchida et al., 1997), was delivered to mice at 3 mg/kg/day. This dose was chosen using the allometric scaling method (Nair and Jacob, 2016) to replicate doses typically used in humans for the treatment of hypertension (Carrara et al., 1994) and to replicate doses that were reported to yield comparable blood and brain levels in mice (Anekonda et al., 2011; Guzman et al., 2018). Three isradipine delivery methods were evaluated: subcutaneous osmotic minipumps, isradipine milled into mouse chow, and subcutaneous extended-release pellets. Data from the pellet-treated mice is reported in the main text for downstream analyses; any data from minipump- and chow-treated mice can be found in supplemental materials (Figs. S1 and S2).

2.3.1. Osmotic minipumps

Isradipine delivery with thirty-day release Alzet osmotic minipumps (model 2004, 0.25 μL/h) was attempted in accordance with previously published reports (Guzman et al., 2018) with some modifications. We were unable to use the previously reported vehicle (50% DMSO, 15% PEG 300, 35% DI H₂O) due to the insolubility (<1 mg/L at 37 °C) of isradipine in any solution that contains water. Our reformulated vehicle (50% DMSO, 50% PEG 300) readily solubilized isradipine. Minipumps were filled with vehicle or isradipine to deliver 3 mg/kg/day, calculated based on an average body weight of 32 g for males and 24 g for females. For surgery, mice were anesthetized with isoflurane at 5% with oxygen delivered at 1–2 L per minute in an induction chamber. After no toe pinch reflex, mice were moved to a heated pad with nose cone delivering isoflurane at 1.5%. The incision site was sterilized using betadine and 70% ethanol alternating three times followed by subcutaneous injection of 0.1 mL of 0.5% lidocaine at the incision site, where the pump was inserted subcutaneously between shoulder blades. Following closure of the incision site, mice were treated with post-operative analgesia (0.1 mL of 0.005 mg/mL buprenorphine subcutaneously) and antibiotic ointment was applied to suture site. Mice were monitored in heated recovery cage until awake and ambulatory. After surgery, mice had ad libitum access to Tylenol dissolved in water (32 mg/mL) for 3 days and were checked daily for signs of distress. When mice were euthanized, we found crystalized isradipine on the exterior of the pump (Fig. S1A) and isradipine was not detectable in blood or brain except for 2 mice (Fig. S5A).

2.3.2. Isradipine chow

Isradipine was milled into mouse chow at 18 parts per million (Purina 5K54, Research Diets Inc. New Brunswick, NJ) to deliver 3 mg/kg/day in mice consuming approximately 5 g/food/day. Food pellets were analyzed by LC/MS/MS to confirm uniformity of isradipine distribution (see section 2.4.1). Food available ad libitum throughout the study was monitored to confirm equivalent consumption in control mice (data not shown). However, isradipine was not detectable in the blood or brain, and contamination was found in vehicle (Figs. S5B and C). Lack of detection may be due to the extensive first pass metabolism and low oral bioavailability of isradipine (Tse and Jaffe, 1987).

2.3.3. Extended-release pellets

Isradipine delivery with extended-release subcutaneous pellets was tested in accordance with previously published reports (Anekonda et al., 2011) with some modifications. Pellets were formulated to release vehicle or isradipine at a dose of 3 mg/kg/day for a total of 30 days (Innovative Research of America, Florida, USA) and extended-release pellets were analyzed by LC/MS/MS to confirm isradipine content (see section 2.4.1). Pellets were subcutaneously implanted between the shoulder blades following procedure used for minipump implantation (Section 2.3.1); staples were used instead of sutures and removed after 10 days or when the incision site healed. Like the previous two methods, isradipine was not detectable in the brain (Fig. S5D). However, behavioral analyses revealed significant treatment-dependent change in velocity during MWM (Section 3.1) in both WT and 5XFAD mice treated with isradipine, suggestive of a pharmacodynamic effect. To more closely examine the pharmacodynamic effect of pellet-delivered isradipine, we moved forward with additional characterization of pellet delivery of isradipine. Twelve-month C57s (average weight 30 g) were implanted with isradipine pellets and sacrificed at day 16 (corresponding with a pre-behavior timepoint), day 23 (corresponding with a time point during behavioral testing), and day 28 (corresponding with completion of behavioral testing). LC/MS/MS revealed the C57 mice had detectable levels of isradipine at day 16, 23, and 28 in plasma and day 16 in brain, showing drug was on board prior to behavior (Figs. S5E and F).

2.4. High performance liquid chromatography-tandem mass spectrometry (LC/MS/MS)

HPLC-MS was used to evaluate isradipine levels in brain, blood, chow, and subcutaneous pellets. Isradipine standard was purchased from Sigma. Methanol and Milli-Q water were used for preparation of all solutions. The HPLC system consisted of a Shimadzu SIL 20 A HT autosampler, LC-20AD pumps (2), and an AB Sciex API 3200 tandem mass spectrometer with turbo ion spray in positive mode. The HPLC analytical column was a C18 Excel 3 ACE (3 × 75 mm, 3 μm) purchased from MacMod (Chadds Ford, PA) and was maintained at 25 °C during the chromatographic run using a Shimadzu CT-20 A column oven. Mobile phase A consisted of 99.9% H₂O with 0.1% formic acid. Mobile phase B consisted of 99.9% acetonitrile with 0.1% formic acid. The flow rate of the mobile phase was 0.4 mL/min. Isradipine was eluted with the following gradient: 0–1 min, 10% B; 1–5 min, linear 10–99% B; 5–8 min, 99% B; 8–8.1 min, step gradient 99–10% B; 8.1–11 min, 10% B to equilibrate the column before the next injection. The isradipine transition was at m/z 372 → 312.2 Da and the internal standard D5-Fentanyl transition was at m/z 342 → 188 Da. Isradipine and internal standard super stock solutions were prepared in methanol at a concentration of 1 mg/mL and stored in aliquots at −80 °C. Working stock solutions were prepared each day from the super stock solutions at concentrations of 100 and 10 μg/mL to spike the calibrators.

2.4.1. Isradipine quantification in food, extended release pellets, whole blood and brain

Brain tissue samples were weighed and homogenized in a 10x volume of 75% MeOH. Calibrator samples were prepared by spiking 10 mg of ground extended release pellet, 100 mg of food pellets, 200 μL of blank whole blood samples, or 100 μL of homogenized blank brain tissue. Spikes were prepared at concentrations of 0, 1, 10, 50, 100 PPM for extended release pellet samples, 0, 5, 10, 20, 50, and 100 PPM for food pellet samples, and 0, 1, 5, 10, 50, 100, 500, 1000 ng/mL for blank whole blood and brain tissue. Isradipine was quantified by mixing 10 μL of 1 μg/mL Internal Standard (D5-Fentanyl) into 100 μL (for brain tissue) or 200 μL (for whole blood) of calibrator and unknown samples followed by 2 mL of mobile phase B; extended release pellet and food samples were quantified with 1 mL of mobile phase B. The samples were vortexed vigorously and shaken for 30 min, centrifuged at 3200 g for 30 min, and the supernatant transferred to clean tubes and dried under a gentle stream of nitrogen. The final samples were resuspended in 0.1 mL of a 1:1 solution of acetonitrile:H₂O/0.1% formic acid, transferred to autosampler vials, and then 10 μL (food and pellets) or 15 μL (blood and brain) of the final sample were injected into the LC/MS/MS. The peak area ratio (isradipine/D5-isradipine) for each unknown sample was compared against a linear regression of calibrator ratios to quantify isradipine. When detectable, the concentration of isradipine was expressed as ng/mL in whole blood and pg/mg in brain tissue.

2.5. Behavior

On the day prior to behavioral testing, mice acclimated in behavioral testing room for 1 h in their home cages; the same procedure was followed for all behavioral testing. All testing took place during the light phase with standard room lighting. An overhead camera connected to Noldus Ethovision 14 or 15 recorded behavior. All behavioral apparatuses were cleaned with 70% ethanol between mice.

2.5.1. Open field

The open field task assessed gross locomotor function and hyperactivity. Mice were placed in the center of a testing arena surrounded by a solid white barrier (dimensions 42 cm × 20.3 cm × 20.3 cm) and explored for 10 min before returning to home cage. Velocity and total distance moved was recorded for analysis.

2.5.2. Novel object recognition (NOR)

NOR assessed recognition memory. NOR used the same arena as OF, therefore OF served as acclimation for NOR. For NOR training phase, two identical objects (plastic tea light candles diameter 3.8 cm height 5 cm) were placed at one end of the cage approximately 3.8 cm from the wall of cage and 13.3 cm distance from the other object. Mice entered the arena opposite of the objects and explored for 10 min before returning to home cage. 24 h later, recognition memory was tested by replacing one familiar object with a novel object (Rubik’s cube 3.2 cm × 3.2 cm × 3.2 cm). Object exploration was defined as the mouse nose overlapping with the object or within the object perimeter (3.8 cm) with nose pointing toward the object. Data collected include velocity, distance traveled, and total exploration time during both training and testing. Discrimination Index (DI) was calculated to evaluate recognition memory. Mice that did not explore objects >20 s on day 1 and day 2 were excluded from analysis.

2.5.3. Morris water maze (MWM)

MWM assessed spatial learning and memory using a 1.2 m diameter pool surrounded by multiple distal visual cues. Water was dyed opaque white with 200 mL of nontoxic paint and hidden platform with a diameter of 10.2 cm was submerged just under the surface of the water. The platform remained in the same position for each training trial. Each mouse performed two blocks of three training trials per day for a total of six trials per day for five consecutive days. During training and probe trials mice were released into the pool at one of five locations which were varied so mice could not take the same path to the platform more than twice per day. After the mouse located the hidden platform, it remained on the platform for 30 s before retrieval with a retrieval scoop. Mice that were unable to locate the platform within 60 s were directed to the platform using the retrieval scoop and required to remain on the platform for 30 s prior to removal. After completion of the final training trial, the mice had a 2-h break prior to a probe trial. During probe trial, platform was removed from the maze and mice swam for 60 s. After completion of the probe trial, the mice underwent three visible platform trials wherein the platform had a red object placed in the center was moved between 3 new locations to control for any group- or drug-related differences in visual acuity. Mice that failed to locate the visible platform within 60 s were eliminated from all data analysis, including downstream tissue analysis. Velocity, distance traveled, and latency to platform were analyzed for training and visible trials. Passes over platform location were analyzed for the probe trial.

2.6. Euthanasia and tissue processing

Mice were euthanized with isoflurane overdose prior to cervical dislocation following American Veterinary Medical Association guidelines. Trunk blood was collected into K₂EDTA tubes (365974, Becton, Dickson and Company) for downstream analysis via LC/MS/MS (Section 2.4). The brain was dissected and hemisected. The left hemisphere was post-fixed in 4% PFA for 24 h at 4 °C then submerged in 15% glycerol in 1X PBS and stored at 4 °C until used for immunohistochemistry (see section 2.7). The right cerebellum was flash frozen for LC/MS/MS analysis while the remaining right hemisphere was minced and either flash frozen in liquid nitrogen for protein or submerged in RNAlater (cat AM7021, Invitrogen) for RNA and stored at −80 °C.

2.7. Immunohistochemistry (IHC)

Hemibrains were sliced using a vibratome (VT1000S V2.2, Leica Biosystems) at a thickness of 40 μm. IHC staining was performed on free floating slices. Briefly, slices were washed in 500 μL of 1X PBS three times followed by permeabilization in 0.5% Triton in 1X PBS for 1 h, then blocked in 10% NGS (PCN5000, Life technologies) diluted in 0.5% Triton for 1 h. Primary antibodies were diluted in 50% blocking buffer (rat anti-LAMP1, dilution 1:300, 1D4B-S, Developmental Studies Hybridoma Bank; rabbit anti-D54D2, dilution 1:1000, 8243 S, Cell Signaling; or rabbit anti-Iba1, dilution 1:1000, 01919741, WACO) and slices were incubated overnight at 4 °C with gentle agitation. The following day, slices were washed with 1X PBS three times followed by incubation with secondary antibodies diluted in 25% blocking buffer goat anti-rat AlexaFluor 555 (cat A21434, Invitrogen), goat anti-rabbit Alexafluor488 (cat A11034, Invitrogen), or goat anti-rabbit AlexaFluor 555 (cat A21428, Invitrogen) for 90 min. Slices were washed as before and after the final wash mounted onto slides and dried in the dark overnight. The following day, slices were stained with DAPI (cat 66248, Thermo Scientific diluted 1:10000 in DI H₂O) or ThioflavinS (ThioS) (T1892–25G, Sigma). For ThioS staining slices incubated in freshly prepared 1% ThioS in DI H₂O for 8 min, washed in 70% ethanol for 4 min, and washed with DI H₂O for 2 min, all steps were performed protected from light, with gentle agitation at room temperature. After DAPI or ThioS staining slides dried in the dark overnight prior to being cover slipped with mounting media (cat P36970, Invitrogen), dried overnight, sealed with nail polish, and stored at 4 °C.

2.8. Microscopy

Images were acquired using Cytation 5 (Software: Gen5 Image+ 3.11, Agilent) with 20X objectives and with DAPI (Filter 1225100, LED 1225000), GFP (Filter 1225101, LED 1225001) and RFP (Filter 1225103, LED 1225003) cubes. Images were analyzed and quantified using ImageJ (National Institutes of Health) (Section 2.8.1).

2.8.1. Image quantification

ImageJ was used to quantify microscope images. Whole brain area was determined by generating a whole section mask using automatic thresholding (Default) and particle analysis. For all staining, 16-bit images underwent background subtraction, thresholding (Otus), and particle analysis (see Supplemental Text File for Macros). One brain slice was analyzed per mouse for both stains, average bregma −2.15 mm. To analyze plaque coverage by microglia, three plaque cores were selected per mouse above the center of hippocampus in the cortex (brain region: parietal association cortex or visual cortex), a region of interest (ROI) was defined as a 25 μm radius from center of plaque core using macro (Thomas and Gehrig, 2020), and data on the percent area of coverage of both Iba1 and ThioS measured.

2.9. RNA isolation and gene expression analysis

RNAlater was removed and tissue was homogenized in Trizol Reagent (cat 15596026, Ambion). RNA was isolated following previously described procedure (Rio et al., 2010) with minor modifications. The pellet of RNA was purified using a kit (Aurum Total RNA mini-Kit, 732–6820, Bio-Rad) following the manufacturer’s procedure. RNA purity and quantity was quantified using Cytation 5 with Take3 plate and stored in −80 °C. cDNA was prepared using iScript cDNA synthesis kit (cat 1708891, Bio-Rad) and following the manufacturer’s procedure and stored at −20 °C. Gene expression was quantified in triplicate using SYBR green (cat #K0364, Thermo Scientific) for quantitative polymerase chain reaction (qPCR) on a Bio-Rad CFX96 Real-Time PCR detection system. Primers were obtained from Sigma (Supplement Table 2) and validated to have an efficiency between 80% and 120%. Actin was used as a reference gene. Data were analyzed by correcting for primer efficiency and calculating corrected Ct prior to calculating ΔCt (see supplement). All data is represented as relative gene expression of WT vehicle (fold change compared to WT vehicle and relative to Actb).

2.10. Protein isolation and ELISA

Soluble and insoluble protein fractions were collected from a subset of both 6- and 9-month-old mice (6 mice per treatment/genotype balanced by sex). Brain tissue was Dounce homogenized in ice cold TBS extraction buffer (4μL/1 mg of tissue) with 20 strokes, ultracentrifuged at 100,000g for 1 h at 8 °C. The resulting supernatant was stored in −80 °C as the soluble fraction. The remaining pellet was Dounce homogenized in 5 M guanidine HCl in 50mMTris at pH 8.00 rotated for 2 h at room temperature, and then ultracentrifuged at 20,800g 8 °C for 20 min. Supernatant was collected and stored in −80 °C for use as insoluble fraction. Protein was quantified by Bradford Coomassie Protein Assay (Thermo Scientific, 1856209) and Aβ_1–42 was quantified by ELISA (KHB3441, Invitrogen) following the manufacturer’s protocol. 0.2 μg of total insoluble protein was loaded, while 39.5 μg of total soluble protein was loaded for each ELISA sample. ELISA data is reported as pg Aβ_1–42 per mg of total protein.

2.11. Statistics

GraphPad Prism Software version 9.1.1 was used for all statistical analysis. Data followed a parametric distribution and represented as mean with standard error of the mean (SEM). Tests used include unpaired student’s t-test, 2-way ANOVA, and repeated measure 3-way ANOVA with Tukey Post-Hoc correction for multiple comparisons.

3. Results

3.1. Blockade of LTCCs alters behavior but does not improve learning and memory in 5xFAD mice

After 20 days of treatment with the isradipine or vehicle, we characterized the behavioral phenotype of WT and 5XFAD mice at 6 and 9 months of age. First, we examined the velocity and the distance the mice traveled during the open field task to evaluate hyperactivity and gross motor function; analysis revealed no effect of genotype or treatment on either parameter at either age (Fig. 2A and B, Figs. S3A and E). Next, we measured spatial recognition memory using NOR. Analysis revealed no effect of genotype or treatment at either age on recognition memory as measured by the NOR discrimination index (Fig. 2C and D). We also monitored velocity during NOR and found a trending main effect of genotype in 9-month-old mice indicative of hyperactivity in 5XFADs compared to WTs; this hyperactivity was absent in 6-month-old mice (Figs. S3B and F). Finally, we examined spatial learning and memory in the MWM (Fig. 3A). To evaluate swim speed, we averaged the velocity across all training trials and found that 6- month-old isradipine-treated mice, regardless of genotype, had a significantly increased in velocity (Fig. 3B). We then measured the path length that 6-month-old mice took to locate the hidden platform to evaluate spatial learning, with longer path lengths indicative of a spatial learning deficit (Fig. 3D). The learning curve revealed mild effect of isradipine treatment on learning in 6-month-old mice and showed a significant day and block interaction with genotype in 6-month-old 5XFAD mice (Fig. 3D). On average, over all trials, 5XFADs took paths that were 17.63 cm longer than WTs to locate the hidden platform. Notably, isradipine treatment decreased velocity in the 9-month-old mice regardless of genotype (Fig. 3C). Path length taken to the platform during learning showed 9-month-old 5XFADs took longer paths to locate the platform, showing a main genotypic effect but no effect of isradipine treatment (Fig. 3E). With respect to spatial memory, there was no effect of genotype or treatment on performance in the probe trial in 6- or 9-month-old mice (Figs. S3D and H). Overall, these data indicate isradipine affected behavior, suggesting brain penetrance, but had no effect on spatial memory or spatial learning.

Fig. 2. Isradipine treatment does not change open field or novel object recognition behavior.

A and B. Total distance traveled in OF for 6- and 9-month-old mice, respectively. 6-month group showed a trend for a main effect of treatment. (6-month, n = 8–13, treatment F (1,36) = 3.04, p = 0.088; 9 month, n = 9–11). C and D. 6- or 9-month-old mice, respectively, show no genotypic deficit or treatment effect via DI in NOR. (2-way ANOVA (genotype x treatment) used for B, C, E and F).

Fig. 3. Isradipine treatment alters velocity during Morris water maze but does not improve learning.

A. Left to right: Schematic of MWM, representative trace of WT, representative trace of 5XFAD. Both traces are of 9-month-old vehicle treated mice from Day 3 Block 2 trial 3. B and C. Velocity in MWM averaged over all trials for both 6- and 9-month-old revealed main effect of treatment (6 month, n = 8–13, treatment F (1,36) = 6.50, p = 0.015; 9 month, n = 9–11, treatment F (1,35) = 0.5, p = 0.050). D and E. Learning curves for 6- and 9-month-old mice over 5 days, 2 blocks per day revealed learning deficit between genotypes and a trend for a treatment effect in 6-month group. (Repeated measures 3-way ANOVA (day and block x genotype x treatment), 6-month-old treatment, F (1,36) = 4.11, p = 0.050, day and block × genotype interaction, F (9,324) = 2.01, p = 0.038; 9-month-old genotype, F (1,35) = 6.22, p = 0.018).

3.2. Blockade of LTCCs age-dependently reduced dystrophic neurite and Aβ pathology in 5xFAD mice

To test whether LTCC antagonism altered the neurotoxicity of Aβ to nearby neurites, we analyzed hemibrain sections stained with lysosomal associated membrane protein 1 (LAMP1) and D54D2 (Fig. 4A and B). LAMP1 is found in late endosomes, autolysosomes and lysosomes and is commonly used to visualize dystrophic neurites which are accumulations of vesicles that create bulbous protrusions in neurons. D54D2 stains Aβ_1–37, Aβ_1–38, Aβ_1–39, Aβ_1–40, and Aβ_1–42 allowing visualization of diffuse plaques and halos around plaque cores. 6-month-old 5XFAD isradipine treated mice showed no change in D54D2+ plaques or LAMP1+ particles per μm² of total brain area or in total area coverage compared to vehicle-treated controls (Fig. 4C and D, Figs. S4A and B). In contrast, 9-month-old isradipine treated 5XFAD mice displayed a significant decrease in LAMP1+ particles per μm² and area coverage compared to vehicle-treated controls (Fig. 4E). 9-month-old 5XFADs treated with isradipine also displayed a trending decrease in both Aβ plaque area and Aβ particles per μm² (Fig. 4F, Fig. S4F). Additionally, we performed an ELISA to measure soluble and insoluble Aβ_1–42. We found that isradipine treated 9-month-old 5XFAD mice had an insignificant decrease in the soluble Aβ_1–42 while 6-month-old mice showed no change (Fig. 4G and I). Isradipine treatment had no effect on insoluble Aβ_1–42 in either 6- or 9-month-old 5XFADs (Fig. 4H and J). Overall, these data suggest that there is an age-dependent effect of isradipine treatment on dystrophic neurite and Aβ pathology, since 9-month-old but not 6-month-old 5XFADs treated with isradipine show decreased LAMP1 and trending decreases in Aβ as measured by both IHC and ELISA.

Fig. 4. Isradipine treatment age-dependently decreases LAMP1 and causes a trending decrease in Aβ plaque area.

A. Representative images of 6- and 9-month-old 5XFAD vehicle and isradipine treated mice. LAMP1, in red, labels lysosomes and other endo-lysosomal vesicles accumulated in dystrophic neurites, D54D2, in green, labels diffuse plaques and plaque halos, and DAPI, in blue, stains nuclei. B. Zoomed in image of dystrophic neurite from 9-month isradipine mouse in A. C. Quantification of 6-month-old 5XFAD isradipine treated (n = 10; 7♂, 3♀) vs. vehicle (n = 9; 6♂, 3♀) mice showed no change in proportion of LAMP1+ area of total brain area. D. Quantification of 6-month-old 5XFAD isradipine treated vs. vehicle mice showed no change in the proportion of D54D2+ plaque area of total brain area E. Quantification of 9-month-old 5XFAD isradipine (n = 9; 6♂, 3♀) vs. vehicle (n = 9; 5♂, 4♀) treated mice revealed a significant decrease in proportion of LAMP1+ area of total brain area (t = 2.37, df = 16, p = 0.031). F. Plaque staining in 9-month-old 5XFADs revealed a trend for a decrease in the proportion of D54D2+ plaque area of total brain area (t = 1.93, df = 16, p = 0.072). G and H. Quantification of 6-month 5XFAD soluble and insoluble Aβ_1–42 protein with ELISA showed no change in quantities of Aβ_1–42. I and J. Quantification of 9-month-old 5XFAD soluble and insoluble Aβ_1–42 protein with ELISA showed an insignificant decrease in the soluble Aβ_1–42 and no change in the insoluble Aβ_1–42. (Unpaired t-test for B-I).

3.3. Isradipine treatment does not alter microglia plaque coverage or phenotype

Given the role of microglia in barrier formation around Aβ plaques thus restricting dystrophic neurite growth (Condello et al., 2015), we evaluated whether microglia coverage of plaque cores was modulated by isradipine treatment. We visualized microglia with Iba1 and labeled plaque cores with ThioS (Fig. 5A). Regardless of age, treatment with isradipine did not decrease plaque core labeling ThioS (Fig. 5B and C, Figs. S4E and F) or alter microglia area coverage in 25 μm radius from center of plaque core (Fig. 5D, E and F). Additionally, we evaluated the role of both LTCCs and microglia phenotype during Aβ pathology by performing qPCR against proinflammatory cytokines Tnfa and Il1b, microglia phagocytic marker Cd68, and the LTCC subunits Cacna1c and Cacna1d. 5xFADs at both 6 and 9 months of age show expected increases in Cd68 and Il1b gene expression and trending increases in Tnfa (Supplemental Table 1). Of note, treatment did not affect expression of Cacna1c or Cacna1d in 6- or 9-month-old mice (Supplemental Table 1), suggesting there was not a compensatory upregulation of LTCC gene expression due to isradipine treatment. Overall, these data indicate isradipine did not alter the microglia plaque associated phenotype in either age of mouse.

Fig. 5. Isradipine treatment did not alter plaque cores or microglia coverage of plaques.

A. Representative images of 6- and 9-month-old 5XFAD vehicle and isradipine treated mice (n’s same as Fig. 4). Iba1 labels microglia in red and ThioS marks plaque cores (fibular Aβ in beta-pleated sheets) in green. B and C. Treatment with isradipine in 5XFADs did not change plaque core particle per μm² of total brain area in 6- or 9-month-old mice, respectively. D. Representative image and schematic of the region of interest used to quantify microglia coverage of plaque cores. E and F. Microglia coverage of plaque cores did not change based on treatment in 6- or 9-month-old 5XFAD mice. (Unpaired t-test for B,C, and F).

4. Discussion

The study herein tested whether isradipine, an LTCC antagonist, could improve Aβ plaque associated phenotypes of dystrophic neurites and inflammatory microglia in the 5XFAD mouse model by restoring normal intracellular Ca²⁺ regulation at two different ages. In the present study, we found that in the 5XFAD mouse model of Aβ pathology, isradipine treatment age-dependently decreased the area and number of neuritic dystrophies labeled with LAMP1 and led to trending reductions in Aβ but did not alter plaque associated microglia phenotype or improve cognition. Dystrophic neurites are a common pathological feature found in the vicinity of Aβ plaques in human AD brains and animal models of AD (Dickson et al., 1999; Sharoar et al., 2019) and the size of neuritic dystrophies can be modulated by cell intrinsic processes such as lysosomal proteolysis (Lee et al., 2011) or cell extrinsic processes such as microglia barrier formation (Condello et al., 2015). These and our findings together suggest that antagonism of LTCCs with isradipine may impact lysosomal processing and maturation pathways via modulation of neuronal rather than microglial LTCCs.

Normally, lysosomal maturation takes places as the late endosomes or lysosome undergo retrograde transport from neuronal axon to the cell soma, where the proper machinery allows for the degradation of debris (Maday et al., 2012). Several steps in this process are disrupted in AD and may contribute to the pathogenesis and progression of Aβ pathology (Rubinsztein et al., 2005; Yue, 2006). Therefore, our observation that antagonism of LTCCs with isradipine decreases the area of dystrophic neurites suggests that isradipine impacts a step in the lysosomal processing pathway. Antagonism of LTCCs may effectively decrease intracellular Ca²⁺ dysregulation in the vicinity of Aβ plaques, thus allowing for the proper translocation, acidification, and maturation of lysosomes. Indeed, these processes are all partially regulated by Ca²⁺: extracellular Ca²⁺ influx triggers inositol triphosphate (IP₃) and subsequent release of Ca²⁺ from the endoplasmic reticulum Ca²⁺ stores, and this Ca²⁺ in turn acts as a signaling molecule for many different organelles and pathways, including lysosomes. Multiple studies have shown evidence of antagonizing Ca²⁺ channels to modulate lysosome activity, most postulating a role for alteration of ER intracellular Ca²⁺ stores or improvement in fusion and acidification in lysosomes (Liu et al., 2020; Mu et al., 2008; Park et al., 2014; Siddiqi et al., 2019).

The decrease in LAMP1 we observed in 9-month-old 5XFAD mice may be related to the trending decrease in Aβ that we observed via IHC and ELISA in the same group of mice. Current literature suggests that dystrophic neurites, and specifically the endo-lysosomal vesicles they contain, are sites for cleavage of APP into Aβ and contribute to plaque deposition due to improper acidification and activation of enzymes necessary for degradation (Lee et al., 2022; Xiao et al., 2015). Therefore, the decreases in both Aβ and LAMP1 that we observe in isradipine-treated 9-month-old 5XFAD mice suggests that LTCC antagonism restores effective lysosomal degradation. In alignment with our findings, previously reported work using subcutaneous pellets to deliver isradipine to 3XTg-AD mice, which display both Aβ and tau aggregates, demonstrated a trending reduction Aβ plaque load and insoluble Aβ_1–42 (Copenhaver et al., 2011). In contrast to our findings, treating 5XFAD mice with the Cav1.2-targeting dihydropyridine LTCC antagonist nimodipine did not lead to a reduction in dystrophic neurites or Aβ, despite using a higher dose of 33–51 mg/kg/day (Sadleir et al., 2022). These contrasting findings may be due to different properties of different dihydropyridine LTCC antagonists. Others have noted divergent effects of different dihydropyridines on Aβ production and Aβ transcytosis in vitro (Bachmeier et al., 2011; Paris et al., 2011). In these studies, isradipine had no effect on Aβ transcytosis (Bachmeier et al., 2011) and increased Aβ_1–42 production in APP-expressing CHO cells (Paris et al., 2011). In vivo, the dihydropyridine LTCC antagonists nilvadipine, nitrendipine, and cilnidipine enhanced Aβ clearance in WT mice (Bachmeier et al., 2011) and acute nilvadipine and nitrendipine reduced soluble Aβ_1–42 brain levels and increased plasma Aβ_1–42 in the APP/PS1 animal model of Aβ pathology (Paris et al., 2011). Chronic nilvadipine treatment in chow also reduced Aβ plaque burden (Paris et al., 2011). These data suggest that a subset dihydropyridine LTCC antagonists, but not isradipine, enhance Aβ clearance and/or reduce Aβ production in normal cells. Taken together with our data, this suggests that the trending reduction in Aβ that we observe following isradipine treatment is not due to enhanced Aβ clearance across the blood-brain barrier or reduced baseline Aβ production. Notably, Aβ modulatory effects of another dihydropyridine LTCC antagonist, nimodipine, is not related to blockade of Ca²⁺ influx (Facchinetti et al., 2006) suggesting the effects of isradipine we observe may not be due to LTCC antagonism but to off-target modulation of other pathways.

Our data presented here do not support a role for microglia in our observed reduction in dystrophic neurites and Aβ following isradipine treatment in 9-month-old 5XFAD mice. Enhancement of microglia barrier formation decreases the area of Aβ and LAMP1+ dystrophic neurites (Condello et al., 2015) but microglia can also exert detrimental effects during AD pathology such as secretion of neurotoxic pro-inflammatory cytokines (Hansen et al., 2018; Streit, 2004). Microglia express both Cav1.2 and Cav1.3 and LTCC antagonists directly modulate microglia function in vivo and in vitro (Colton et al., 1994; Espinosa-Parrilla et al., 2015; Hegg et al., 2000; Hopp et al., 2015; Silei et al., 1999; Wang et al., 2019). Notably, plaque-associated microglia display altered intracellular Ca²⁺ fluctuations (Brawek et al., 2014). Based on these data, we had predicted that the LTCC antagonist isradipine would modulate the phenotype of plaque-associated microglia, thus mediating the decrease in Aβ plaques and dystrophic neurites that we observed in 5XFAD mice treated with isradipine. However, we did not observe changes in microglia plaque coverage or inflammatory phenotype following isradipine treatment. This aligns with a report that the LTCC antagonist nifedipine did not alter phagocytic capacity of microglia (Espinosa--Parrilla et al., 2015), but contrasts with our previous report that LTCC antagonism with nimodipine reduced microglial pro-inflammatory phenotype and enhanced spatial learning in a rat model of chronic neuroinflammation (Hopp et al., 2015) and another report that LTCC antagonism with nifedipine decreased in TNFα and NO release during neuroinflammation (Espinosa-Parrilla et al., 2015). In the present study, it is possible that brain levels of isradipine were not sufficient to act on microglial LTCCs and alter their phenotype. In vitro, 30 μM of nimodipine (Li et al., 2009) or 10 μM of nifedipine (Espinosa-Parrilla et al., 2015) was necessary to inhibit microglial production of pro-inflammatory mediators following stimulation with lipopolysaccharide. The levels of isradipine found in the brains of our mice were less than 1 ng/mL, which would be picomolar concentration; these data suggest that high brain exposure levels would be necessary to target microglia in vivo with dihydropyridine LTCC antagonists with similar potency. Additionally, even higher concentrations of dihydropyridines may be required to target microglial LTCCs, since microglia express more Cacna1d (Cav1.3) than Cacna1c (Cav1.2) (Zhang et al., 2014) and current dihydropyridines are selective for Cav1.2 over Cav1.3. Notably, isradipine targets Cav1.3 with higher affinity than other dihydropyridines (Koschak et al., 2001; Lipscombe et al., 2004). Development of selective Cav1.3 antagonists is an area of active research for development of neuroprotective agents for the treatment of Parkinson’s disease (PD) (Kang et al., 2012).

In PD mouse models, isradipine targets dopaminergic neuron Cav1.3 to relieve mitochondria oxidative stress (Guzman et al., 2018) and attenuate substantia nigra neurodegeneration (Ilijic et al., 2011). We were unable to replicate these studies’ use of subcutaneous osmotic minipumps to deliver isradipine to the brain (Section 2.3.1, Fig. S1A, Fig. S3A). However, when we performed a time course to evaluate isradipine blood and brain levels in C57 mice treated with extended release pellets, we detected isradipine in the blood at day 16 (72.28 ± 0 ng/mL), 23 (45.36 ± 22.53 ng/mL), and 28 (29.93 ± 9.44 ng/mL) and in the brain at day 16 (1.303 ± 0.54 ng/mg) in brain, showing drug was on board during behavioral testing (Figs. S5E and F; see Fig. 1 for timeline). The blood concentrations were similar to those previously published (33 ± 7 ng/mL) from 60 day release pellets in female 3xTg-AD mice, but the brain levels we detected were less than the previously reported 47 ± 1 ng/mL (Anekonda et al., 2011). Notably, subcutaneous pellets were the only isradipine delivery method that led to any behavior changes in 5XFAD and WT littermates with a main effect of treatment. However, at the time of euthanasia isradipine was undetectable in brain and blood of 5XFAD and WT littermates regardless of drug delivery (osmotic pump, food and subcutaneous pellet, Figs. S5A–D). Isradipine does cross the blood-brain barrier (Uchida et al., 1997) but is subject to extensive first pass metabolism and has low oral bioavailability (Tse and Jaffe, 1987). This could account for the lack of isradipine detected in the blood and brain of mice treated with the isradipine milled into chow (Section 2.3.2, Figs. S5B and C). One previous report found that high concentrations of isradipine in chow was required to yield detectable levels of isradipine in rodents (Tse et al., 1989) which aligns with recent studies delivering nimodipine or nilvadipine via chow (Paris et al., 2011; Sadleir et al., 2022). These oral dihydropyridine doses administered to rodents would be supratherapeutic to typical doses delivered to humans for the treatment of hypertension (Carrara et al., 1994), while we aimed to deliver a dose comparable to normal human dosing regimens based on allometric scaling (Nair and Jacob, 2016) and previous reported rodent plasma levels. However, the level of brain exposure in human patients treated with oral isradipine is not known therefore it difficult to know if there is similar brain exposure in rodents and this may in part account for the lack of efficacy in human trials of LTCC antagonists for the treatment of neurodegenerative disorders from preclinical research. Indeed, even the mice we treated with isradipine subcutaneous pellets, which exhibited isradipine-dependent changes in behavior regardless of genotype, showed levels of isradipine below the limits of detection (<1 ng/mL) upon LC/MS/MS analysis. Overall, our challenges with achieving adequate blood and brain exposure of isradipine in rodents demonstrates a potential challenge with using isradipine and other dihydropyridine LTCC antagonists for the treatment of AD and other age-associated neurodegenerative disorders.

Despite reductions in dystrophic neurites and Aβ in 9-month-old isradipine-treated 5XFAD mice, we did not see any improvements in behavioral deficits in these mice. The 5XFAD mouse model shows age-dependent increases in amyloid plaque pathology, beginning at 2 months of age, closely followed by behavioral deficits as early 4 months of age (Forner et al., 2021; Oakley et al., 2006; Oblak et al., 2021). Age-dependent increases in hyperactivity is one of the most robust phenotypes in the 5XFAD mouse model (Forner et al., 2021; Oblak et al., 2021). We observed 9-month-old 5XFADs exhibited hyperactivity on both day 1 and day 2 of NOR, suggesting a lack of habituation to their environments (Figs. S3E and F). Habituation to a novel environment is a basic form of learning behavior found across species that is characterized by a decrease in locomotor activity (e.g., velocity, distance traveled) over time spent in a novel environment from beginning to end of a session, or between two separate visits to the same environment from one session to another (Leussis and Bolivar, 2006; Reverchon et al., 2020; Vianna et al., 2000). Habituation behavior requires an intact hippocampus, suggesting the lack of habituation in 5XFADs may be a sign of defective hippocampal-dependent memory. Previous reports also note recognition memory deficits in the 5XFAD mouse model, but we were unable to replicate these findings (Fig. 2C and D) similar to what other groups have reported (Wei et al., 2016). We did not observe deficits in spatial memory in the MWM probe trial in either 6- or 9-month--old 5XFAD mice (Figs. S3D and H). We did observe deficits in spatial learning in the MWM in 9-month-old 5XFADs (Fig. 3F). This aligns with previously published literature showing deficits in spatial learning in 5XFAD mice (Flanigan et al., 2014). Overall, these data demonstrate the difficulty of modeling the memory deficits characteristic of human AD in animal models, which has been noted by others (Oblak et al., 2021; Sukoff Rizzo et al., 2020; Vitek et al., 2020). We did not observe any significant effect of isradipine on 5XFAD behavioral deficits, with only a minor main effect in 6-month group MWM (Fig. 3D). However, we found that isradipine increased velocity in 6-month-old mice and decreased in 9-month-old mice during MWM, regardless of genotype (Fig. 3B and C). These data provide evidence that isradipine can alter behavior and impact brain physiology despite very low (<1 ng/mL) blood and brain levels.

Notably, the beneficial effects of isradipine treatment on amyloid and lysosome pathology was only observed in 9-month-old mice and not 6-month-old mice. These data are in contrast to data suggesting that the LTCC antagonist nimodipine is only effective early in amyloid pathology the Tg2576 model (Ishii et al., 2019). However, the timepoint at which isradipine is protective in our study aligns with the emergence of previously reported enhancement the Ca²⁺-sensitive afterhyperpolarization in CA1 hippocampal neurons from 5XFAD mice (Kaczorowski et al., 2011) which is linked to both LTCC hyperactivity and β-secretase mediated Aβ production (Disterhoft and Oh, 2006). At younger ages, 5XFAD CA1 neurons exhibit a substantially reduced afterhyperpolarization (Ghoweri et al., 2020; Yao et al., 2020). Overall, these data demonstrate that age-related changes in Ca²⁺ signaling during amyloid pathology may alter the outcome of different pharmacological interventions targeting Ca²⁺ dysregulation.

5. Conclusion

Overall, our data suggest that the reduction of plaque associated LAMP1+ neuritic dystrophies observed in 5XFAD mice after treatment with isradipine could be due to direct action on neuronal LTCCs rather than microglia LTCCs. Epidemiological studies provide evidence for action of dihydropyridines in AD and other neurodegenerative diseases but the actual mechanism behind these improvements is still under investigation. Cell-type specific conditional genetic removal of LTCCs on neurons, microglia, or other cell types could directly assess the contributions of different cell types and individual cellular mechanisms to elucidate the neuroprotective effect pharmacological antagonism of LTCCs with isradipine. We predict that antagonism of LTCCs modulates intracellular Ca²⁺ thereby mediating improvements in autophagic or lysosomal processing by fusion and trafficking. Further investigation is needed to decipher the specific lysosomal or autophagy pathways isradipine targets. Overall, this study provides insight into how antagonizing LTCCs with isradipine improves plaque associated phenotypes and contributes to honing in on potential treatments and therapeutic targets for AD.

Abbreviations:

AD

Alzheimer’s Disease

LTCC

Long lasting type calcium channel

CCB

Calcium channel blocker

Aβ

amyloid beta

Ca²⁺

calcium

WT

Wild type

OF

open field

NOR

novel object recognition

MWM

Morris water maze

LC/MS/MS

liquid chromatography tandem mass spectrometry

LAMP1

lysosomal associated membrane protein

qPCR

quantitative polymerase chain reaction

PD

Parkinson’s Disease

APP

Amyloid Precursor Protein

Data availability

Data will be made available on request.