Development of a monoclonal antibody specific for a calpain-generated Δ48 kDa calcineurin fragment, a marker of distressed astrocytes

Abstract

Background:

Calcineurin (CN) is a Ca²⁺/calmodulin-dependent protein phosphatase. In healthy tissue, CN exists mainly as a full-length (~60 kDa) highly-regulated protein phosphatase involved in essential cellular functions. However, in diseased or injured tissue, CN is proteolytically converted to a constitutively active fragment that has been causatively-linked to numerous pathophysiologic processes. These calpain-cleaved CN fragments (ΔCN) appear at high levels in human brain at early stages of cognitive decline associated with Alzheimer’s disease (AD).

New method:

We developed a monoclonal antibody to ΔCN, using an immunizing peptide corresponding to the C-terminal end of the ΔCN fragment.

Results:

We obtained a mouse monoclonal antibody, designated 26A6, that selectively detects ΔCN in Western analysis of calpain-cleaved recombinant human CN. Using this antibody, we screened both pathological and normal human brain sections provided by the University of Kentucky’s Alzheimer’s Disease Research Center. 26A6 showed low reactivity towards normal brain tissue, but detected astrocytes both surrounding AD amyloid plaques and throughout AD brain tissue. In brain tissue with infarcts, there was considerable concentration of 26A6-positive astrocytes within/around infarcts, suggesting a link with anoxic/ischemia pathways.

Comparison with existing method:

The results obtained with the new monoclonal are similar to those obtained with a polyclonal we had previously developed. However, the monoclonal is an abundant tool available to the dementia research community.

Conclusions:

The new monoclonal 26A6 antibody is highly selective for the ΔCN proteolytic fragment and labels a subset of astrocytes, and could be a useful tool for marking insidious brain pathology and identifying novel astrocyte phenotypes.

1. Introduction

Calcineurin (CN) is a Ca²⁺/calmodulin-dependent protein phosphatase ubiquitously expressed in most mammalian tissues, but found at especially high levels in brain (Aramburu et al., 2000; Nguyen and Di Giovanni, 2008; Vihma et al., 2008; Baumgartel and Mansuy, 2012) where it links increases in intracellular Ca²⁺ to a broad range of processes. In neurons, CN is a key regulator of synaptic plasticity through dephosphorylation of numerous substrates associated with the neuronal cytoskeleton, neurotransmitter receptor trafficking, and Ca²⁺ channels (Malinow et al., 1988; Lisman, 1989; Mulkey et al., 1993; Wang and Kelly, 1996; Foster and Norris, 1997). In glial cells, CN regulates immune inflammatory signaling and glutamate regulation through interactions with the nuclear factor of activated T cells (NFAT), and other transcription factors (Norris et al., 2005; Fernandez et al., 2007; Canellada et al., 2008; Perez-Ortiz et al., 2008; Sama et al., 2008; Nagamoto-Combs and Combs, 2010; Shiratori et al., 2010; Furman and Norris, 2014; Sompol et al., 2017).

Ca²⁺ dysregulation, common to brain aging, Alzheimer’s disease (AD) and other AD-related dementias (Gibson and Peterson, 1987; Landfield, 1987; Khachaturian, 1989; Disterhoft et al., 1994; Foster and Norris, 1997; Norris et al., 2006; Thibault et al., 2007) is thought to disrupt cellular signaling and increase vulnerability to disease through many pathways, including those regulated by CN (Reese and Taglialatela, 2011; Sompol and Norris, 2018). Indeed, many studies have shown that expression and/or activity levels of CN are elevated in human brain tissue at early and late stages of AD (e.g. Liu et al., 2005; Abdul et al., 2009; Wu et al., 2010, Mohmmad Abdul et al., 2011; Pleiss et al., 2016), or in experimental models of AD and ADRDs (e.g. Norris et al., 2005; Dineley et al., 2007; Reese et al., 2008; Dineley, Kayed et al., 2010; Sompol et al., 2017; Sompol et al., 2023). Furthermore, blocking CN signaling through pharmacologic or genetic approaches ameliorates numerous pathophysiologic and cognitive phenotypes related to AD (Reese and Taglialatela, 2011; Sompol and Norris, 2018; Kraner and Norris, 2018).

In addition to regulation by Ca²⁺/calmodulin, CN may undergo aberrant Ca²⁺-dependent proteolytic activation in diseased and/or injurious states (Norris, 2014; Schultz et al., 2021). Proteolysis is mediated by calpains and occurs primarily near the carboxyl terminus of the CN A catalytic subunit (Wu et al., 2004) leading to the generation of several CN fragments (ΔCNs). Unlike full-length CN (~60 kDa), ΔCNs exhibit constitutive activity, even in the absence of Ca²⁺ due to the removal or disruption of a critical C terminus autoinhibitory domain (AID). CN fragments, with molecular masses of 45, 48, and 57 kDa, have been observed in human AD tissue (Liu F, Grundke-Iqbal et al., 2005, Wu et al., 2010; Pleiss et al., 2016) and may appear very early in the disease process (Mohmmad Abdul et al., 2011). In neural cells, overexpression of similar ΔCNs can recapitulate AD-like phenotypes including Ca²⁺ dysregulation (Norris et al., 2010), synapse dysfunction or degeneration (Winder et al., 1998; Wu et al., 2010; Pleiss et al., 2016; Hopp et al., 2018), upregulation of inflammatory cytokines and/or inflammatory transcriptional signatures (Norris et al., 2005; Fernandez et al., 2007; Sama et al., 2008) and cognitive decline (Mansuy et al., 1998).

Exposure of the proteolyzed site in the CN C-terminus domain provides an opportunity to develop novel antibody reagents that selectively detect pathological forms of CN (i.e. ΔCN) associated with disease. We previously developed a rabbit polyclonal antibody that selectively detects ΔCN in postmortem human brain tissue expressing amyloid deposits or infarcted cerebral vessels (Pleiss et al., 2016). Labeling was most prominent in reactive astrocytes in the vicinity of frank pathology, but considerable heterogeneity was noted across astrocytes, suggesting the existence of unique and possibly functionally distinct astrocyte subtypes (Pleiss et al., 2016). As a next step toward developing a consistent and readily available research tool, we have developed a monoclonal antibody (26A6) to ΔCN using the same peptide antigen-based strategy for polyclonal antibody generation (Pleiss et al., 2016). Here, we show immunofluorescence and immunohistochemical labeling of human brain tissue with the 26A6 antibody. Similar to the polyclonal antibody, 26A6 selectively identifies a ~48 kDa calpain-generated ΔCN fragment and strongly labels subsets of reactive astrocytes in multiple human cases of known pathology.

2. Material and methods

2.1. Ethics statement

Mice used for generation of hybridomas were handled at GenScript (GenScript, Piscataway, NJ) under a protocol approved by the University of Kentucky Institutional Animal Care and Use Committees. All animal procedures were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Human brain sections were obtained from the University of Kentucky Alzheimer’s Disease Research Center (UK-ADRC) Tissue Repository. All human samples were collected with informed consent and in accordance with procedures under a protocol approved by the University of Kentucky Institutional Review Board.

2.2. Immunization and development of hybridoma lines

The immunization and development of hybridoma lines was carried out under contract with GenScript (GenScript, Piscataway, NJ), a company with an established track record of successful antibody development. They initially immunized 3 Balb/c and 3 C57Bl/6 mice with the peptide antigen CGGGESVLTLK, based on the Lys 424 calpain-dependent cleavage site (Wu et al., 2004; Pleiss et al., 2016), as a KLH conjugate. After 3 boosts, they sent us serum to test in a Western blot, and our initial screens indicated that 2 of the mice had substantial selective reactivity towards the proteolyzed CN band. These mice were selected for further boosting and fusion with a Sp2/0 myeloma partner. The resulting hybridomas were screened first in ELISA assays to the immunizing peptide and secondly in Western blot assays to the uncleaved and cleaved CN, to confirm that they selectively bound the cleaved CN. A positive clone was identified—26A6. This clone was developed/subcloned further and characterized as to isotype. 26A6 is an IgG2b with a kappa light chain.

This hybridoma line was shipped to us and cultured as directed. The culture medium is high glucose DMEM supplemented with 10% heat-inactivated fetal bovine serum, 2 mM glutamine, 1 mM sodium pyruvate, 100 units/ml penicillin, and 300 ug/ml streptomycin. The cultures are maintained in a 6% CO₂ cell culture incubator. Culture medium is collected and either used directly at dilutions ranging from 1:2–1:5 or the antibodies from the culture supernatants purified on Protein G as described below.

2.3. ELISA assays

ELISA assays were carried out at GenScript. The immunizing peptide itself was bound to plates at a concentration of 1 μg/ml, 100 μl per well and supernatants from the hybridomas were used to bind to the peptide. An anti-IgG-HRP conjugated secondary antibody was used to detect the primary 26A6 antibody.

2.4. Calpain cleavage of CN in vitro and characterization of 26A6 binding to fragments by Western analysis

To assess the preference of the 26A6 antibody for the cleaved form of CN relative to the full-length form, cleavage of CN with calpain was carried out in vitro and the ability of 26A6 to bind full-length versus cleaved fragments assessed in Western blots. All cleavage reactions were carried out in CN dilution buffer (4 mM CaCl₂, 5 mM DTT and 50 mM Tris, pH 7.5). Purified human CN was from Enzo Life Sciences (Cat. # BML-SE163, Farmingdale, NY). This CN is a recombinant human CN with a 60 kDa Aα chain and a 15 kDa CN B chain in a buffer containing 50 mM Tris, pH 7.5, 100 mM NaCl, 6 mM MgCl₂, 5 mM DTT, 0.025% NP40, and 0.5 mM CaCl₂. Endogenous mouse brain CN was prepared from the cytosolic fraction of wild type mouse brain extracts, desalted on Zeba columns (Cat. # 89882, Thermofisher, Waltham, MA) to remove protease inhibitors and exchange into CN dilution buffer. Both purified human and desalted endogenous mouse CN are stable when stored at − 80. Purified calmodulin (Cat. # BML-SE325, Enzo Life Sciences, Farmingdale, NY) was dissolved into CN dilution buffer at a concentration of 1 μg/μl, then aliquoted and stored at − 80. Purified human calpain 1 (Cat. # C6108, Sigma, St. Louis, MO) was received in a buffer of 5 mM Beta mercaptoethanol, 1 mM EDTA, 1 mM EGTA, 30% Glycerol, 20 mM Imidazole hydrochloride at a concentration of 1 unit/μl. Purified recombinant rat calpain 2 was kindly provided by Dr. Dorothy Croall from the University of Maine. It was in a buffer of 50 mM MOPS, pH 7.5, 1 mM EDTA, 1 mM EGTA and 0.5 mM DTT. The cleavage reactions were carried out by combining 5 ul of human or endogenous mouse CN with 5 ul of calmodulin (or buffer control) and 5 ul of 1:10 diluted calpain 1 or 1:5 diluted calpain 2. The calpain 1 reactions were heated at 37 C for 15 min. The calpain 2 reactions were heated at 28 C for 15 min. After calpain treatment, samples were quenched in 2X gel loading buffer and heated at 65 C for 30 min prior to gel and Western blot analysis.

Following denaturation, full-length or proteolyzed protein samples were resolved on 4–20% gradient gels (Criterion gels #5671095, Bio-Rad, Hercules, Ca) and transferred to PVDF membranes. As a positive control in Westerns, we used a commercially-available rabbit polyclonal antibody to the amino terminus of CN (1:2500, Cat. No. 07–1492, Millipore, Burlington, MA) that detects both the full-length (60 kDa) and proteolyzed fragments (48 kDa, 45 kDa). The rabbit primary is detected by a 680RD anti-rabbit secondary (P/N 926–68071, LI-COR, Lincoln, NE) and the 26A6 mouse monoclonal is detected by a 790 anti-mouse light chain secondary (Cat. No.115–655–174, JacksonImmunoResearch, West Grove, PA). It is important to use a secondary directed at the mouse antibody light chain, since the 50 kDa antibody heavy chain in samples can lead to interference with data assessment. Westerns were imaged on an Odyssey Scanner (LI-COR, Lincoln, NE) and the quantification was performed using Image Studio 2.1 Software (LI-COR; RRID: SCR_013715). Data is shown with both the green (800) and red (680) channels together, resulting in a yellow signal where there is overlapping detection, or individual channels are shown in black and white.

2.5. Purification of antibodies

The 26A6 antibody was purified from antibody supernatants using Protein G agarose (Cat. No. 20398, Pierce BioTech, Waltham, MA). Cell culture supernatants were diluted 1:1 with phosphate buffered saline, pH 7.2, and run over columns containing the resin. The resin was washed with 10 column volumes of phosphate buffered saline, and eluted with a buffer containing 0.1 M glycine, pH 2.8. The eluate was collected into fractions with 1/10 vol of Tris, pH 8.8 buffer. Immediately, A₂₈₀ readings were taken on a spectrophotometer (DU 730, Beckman Coulter) and the peak fractions pooled. The samples were dialyzed against cold phosphate buffered saline, pH 7.2 overnight at 4 °C. The column was regenerated by washing with 10 volumes of the glycine elution buffer and re-equilibrated with phosphate buffered saline containing 0.05% sodium azide for storage at 4 °C. The column could be re-used multiple times. The next morning following dialysis, a final protein reading was taken, samples were taken for gel analysis of purity, and the antibodies aliquoted and stored at − 80 °C.

2.6. Human brain sections preparation and staining

Human brain sections were obtained from the University of Kentucky Alzheimer’s Disease Research Center (UK-ADRC) Tissue Repository, with samples provided as numbered specimens. A total of 7 different disease cases, either AD or CVD, were examined and 5 different age-matched controls. All subjects were participants in the UK-ADRC Autopsy program, and the postmortem time interval for all subjects used in these studies is approximately 3 hrs. (Nelson et al., 2007; Abdul et al., 2009; Mohmmad Abdul et al., 2011). These samples were formaldehyde-fixed, paraffin-embedded sections cut into ~ 8 μm thickness and mounted on glass microscope slides. Slides were baked in a 40 °C oven overnight, then deparaffinized using SafeClear (Fisher Scientific, Hampton, NH), and rehydrated through a series of alcohols and finally into PBS. Antigen retrieval was carried out in pressurized decloaking chamber (BioCare Medical, Concord, CA), according to manufacturer’s directions using the Borg antigen retrieval solution (BioCare Medical, Concord, CA). Following antigen retrieval, slides were washed 3x in PBS. For sections designated for staining with the β-amyloid antibody downstream, the sections were treated with 70% formic acid for 5 mins, then washed. All sections were treated with 3% H₂O₂ in methanol to block endogenous peroxide activity. Slides were washed in PBS and sections were either stained for immunofluorescence or immunohistochemistry.

For immunofluorescence, sections were incubated in blocking solution (PBS with 3% bovine serum albumin, 5% goat serum, and 0.05% Triton × 100) for 30 min, and incubated overnight at room temperature in primary antibody in that same blocking solution. Primary antibodies used were: rabbit anti-GFAP (1:200; catalog #12389, Cell Signaling Technology, Danviers, MA) and mouse 26A6 hybridoma supernatants diluted 1:5 or 10 ug/ml of purified 26A6 antibody. Sometimes 75 μg/ml immunizing peptide was included with the 26A6 antibody as a negative control. Following washes in PBS, primary antibodies were incubated with 594-conjugated goat anti-rabbit (1:200, Invitrogen, Waltham, MA), biotinylated anti mouse (Tyramide Superboost Amplification kit, Invitrogen, Waltham, MA), and DAPI (1:1000, ThermoFisher, Waltham, MA). Slides were washed in PBS and the biotinylated anti-mouse antibody signal was amplified using the 488-conjugated tyramide reagent according to the manufacturer’s directions in the Tyramide Superboost Amplification kit. Alternatively, sometimes sections were stained with 26A6 and rabbit antibody to amyloid β 1–42 (1:1000, ab201060, Abcam, Cambridge, UK). The sections were then counterstained with the mouse 488 Tyramide Superboost kit and 594-conjugated goat anti-rabbit. Subsequently, sections were stained with rat monoclonal antibody to GFAP (1:200, 13–0300, Invitrogen) and counterstained with 405-conjugated goat anti-rat antibody (1:200, ab175671, Abcam, Cambridge, UK). Slides were washed and mounted in EverBrite True Black Hardset Mounting Medium (23017, Biotium, Fremont, CA). Fluorescent confocal microscope images were taken using a Nikon Eclipse Ti microscope (Minato City, Tokoyo, Japan).

For immunohistochemistry, sections were blocked in PBS with 3% bovine serum albumin, 5% goat serum, 5% horse serum, and 0.05% Triton × 100. Incubation of sections with 26A6 antibody and rabbit antibody to amyloid β 1–42 (1:1000, Abcam, ab201060) was done as described above. Secondary antibodies and avidin-biotin complex (ABC) amplifications used were from VectaStain Elite ABC-HRP kits (mouse PK-6102 and rabbit PK-6101, Vector Laboratories, Newark, CA), according to manufacturer’s directions. The mouse 26A6 antibody was detected with the brown DAB substrate and the rabbit amyloid β 1–42 antibody was detected with the blue SG substrate (Vector Labs). Following dehydration in alcohols and SafeClear, the sections were mounted in DPX mounting medium (100504–938, VWR, Radnor, PA) and imaged on a Nikon Eclipse 90i.

2.7. Quantification and statistical analysis of HRP-labeled sections

After immunohistochemistry staining, ten field of views (FOV) of gray matter from each human brain sample were randomly acquired using a 60x objective (Nikon Eclipse 90i). Areas with and without Aβ plaques were captured from AD samples. For image analysis, the entire set of images were converted to 8 bit, then background subtraction, and thresholding were homogeneously performed using FIJI software. Number of positive staining, area, and sum intensity were measured and statistically processed using GraphPad Prism.

3. Results

3.1. A monoclonal antibody, 26A6, shows greater reactivity to 48 kDa CN fragment (Δ48 CN)

Full-length CN is susceptible to cleavage by the endogenous Ca²⁺-activated protease, calpain, at several sites (Wang et al., 1989; Wu et al., 2004) and Fig. 1A). Some of the resulting fragments, a 48 kDa fragment and a 45 kDa fragment, lack an autoinhibitory domain (AID), causing these fragments to become constitutively active (see Fig. 1A). Our previous work indicated that the 48 kDa CN fragment (Δ48 CN) in particular, was associated with postmortem human AD brain samples (Mohmmad Abdul et al., 2011; Pleiss et al., 2016), and therefore we immunized mice with the same peptide antigen CGGGESVLTLK, we used previously to generate a polyclonal rabbit antibody (Pleiss, Sompol et al., 2016). This immunizing peptide corresponds to the C-terminus of Δ48 CN (Fig. 1A).

Fig. 1.

Mouse monoclonal antibody 26A6, preferentially binds 48 kDa ΔCN fragment. A, Cartoons of full-length CN (CN) and 48 kDa and 45 kDa calpain-generated CN cleavage fragments are shown, depicting location of catalytic domain, CN B binding domain (CN B), Ca²⁺ activated calmodulin binding domain, and autoinhibitory domain (AID). The site of calpain cleavage was previously determined by mass spectrometry analysis (Wu, Tomizawa et al., 2004). The immunizing peptide is shown, with C-terminal sequence from 48 kDa shown in red. B, Results of 26A6 in ELISA to immunizing peptide, and isotype of monoclonal antibody are shown. C, Pure human CN (Hu CN) or endogenous mouse CN (Ms CN) were cleaved by pure human calpain 1 or 2 (clpn 1/2) in the presence or absence of calmodulin (calmod). The left panel shows digestion with calpain 1, and the right panel shows digestion with calpain 2. The top blot shows the detection of CN with the rabbit N-terminal CN antibody in the red channel (which detects all CN forms), and the detection with the mouse monoclonal 26A6 in the green channel. The overlap of detection by the red and green channels gives rise to a bright yellow band only for the 48 kDa form of CN. A black and white of only the green channel is also shown, to more clearly demonstrate the bands specifically detected by 26A6 (26A6 BW). The 26A6 antibody detects the 48 kD band of the pure human CN to a much higher degree than a faint reactivity to the full-length CN. In addition, the 26A6 antibody detects the endogenous mouse brain CN cleaved with calpain 2, as indicated by the black arrow in the black and white panel.

We obtained a number of hybridomas that were positive in the initial ELISA screen to the peptide antigen (total of ~64 wells), including monoclonal antibody 26A6 (Fig. 1B). However, secondary screens in Western blots to proteolyzed CN revealed only one antibody that selectively detected Δ48 CN in Western analysis as well (Fig. 1C). Mouse monoclonal antibody, 26A6, bound the 48 kDa calpain-cleaved CN with a high degree of specificity, showing very little reactivity towards either the full-length CN or the Δ45 CN (Fig. 1C). There did seem to be some preference for CN cleaved by calpain 2 over calpain 1, as the endogenous mouse brain CN Δ48 CN was detected when cleaved by the former and not the latter, as was seen most clearly in the 26A6 black and white exposure (Fig. 1C). Having developed an antibody reagent that selectively detects the Δ48 CN in Westerns, we were now positioned to address the issue of what cells expressed this fragment in vivo and with what pathologies these cells were associated using immunostaining approaches.

3.2. The 26A6 monoclonal detects Δ48 CN to a higher degree in astrocytes of high pathology human brain tissue

Our initial labeling experiments utilized human brain sections from patients with known pathologies (see Table 1). Patients had dementia with underlying Alzheimer’s disease (AD) and/or cerebrovascular disease (CVD), or were clinically normal. Initially, we co-stained samples with the 26A6 monoclonal and an antibody to glial fibrillary acid protein (GFAP), a well-known marker of astrocytes. As shown for cases R5329 and R1251 in Fig. 2A, there was considerable co-labeling of cells with these two markers, seen best as the yellow overlap between the two fluorescent labels in the overlay. These data indicated that many of the 26A6 labeled, and therefore Δ48 CN-containing, cells were astrocytes. The low magnification images showed a robust scattering of 26A6-labeled cells in both pathological samples (Fig. 2A), while there was no discernable labeling with 26A6 in the normal brain specimen, case 5434 (Fig. 2B). The higher magnification images (Fig. 2A, lower panel and Fig. 2C) showed that the 26A6 label was centrally located in the cell body rather than penetrating out into the astrocytic processes the way that GFAP labeling does. In addition, there was considerable heterogeneity in the 26A6 labeling. This was best shown in Fig. 2C. Only a subset of astrocytes labeled with 26A6, and there were a few cells that stained with 26A6 and not GFAP. Primarily, however, it seemed 26A6 labeled a subset of astrocytes in brain tissue with pathologies.

Table 1.

A summary of the cases used in this report. Cases 5329, 1251, and 5434 are shown in Fig. 2. Case R5200 is shown in Figs. 3 and 4. Cases R5434, R5356, 5500, 5501 and 5450 (Normal) and R1259, R5329, R5332, R1251, R5200, and R5409 (Demented) were analyzed and quantified in Fig. 5. Case R5444, containing a large infarct, is shown in Fig. 6.

Case No.	Age at Death	Sex	APOE	Last Clinical Index	Notes	ADCERAD	Consensus Dx
R5444	86	M	e3/e3	MCI	1) AD - low level 2) CVD, moderate with large infarct 3) ARTAG 4) Mild Vermian Atrophy, Cerebellum	A = CERAD possible	CVD+ARTAG
R5434	75	M	e3/e4	NORMAL	1) ARTAG, widespread 2) CVD - mild 3) NO AD Pathology	No	ARTAG
R5500	89	F	e3/e3	NORMAL	1) Normal 2) Low AD Pathology 3) LB pathology, Olf. Bulb only 4) CVD - Mild	No	Not done yet
R5501	83	M	e3/e4	NORMAL	1) Normal 2) Low AD Pathology 3) CVD - Mild	No	Not done yet
R5450	83	F	e3/e3	NORMAL	1) PART - Definite 2) CVD - mild 3) LATE - Stage 2	No	Normal+LATE
R5356	94	M	e2/e3	NORMAL	Low AD Pathology	NO	Normal
R5409	99	F	e3/e3	DEMENTED	AD-High Level	YES	AD + HS + ARTAG + CVD
R5200	91	F	e3/e3	DEMENTED	AD	YES	AD + CVD + HS
R1251	82	M	e3/e4	DEMENTED	Multifocal CVD	C = Definition AD	AD + CVD
R5332	83	M	e3/e3	DEMENTED	AD	C = Definition AD	AD + CVD
R5329	86	F	e3/e4	DEMENTED	High AD Pathology	C = Definition AD	AD + DLB + CVD
R1259	87	F	e3/e4	DEMENTED	AD	C = Definition AD	AD + CVD

APOE (apolipoprotein E) gene, ADCERAD (Semiquantitative measure of neuritic plaques), MCI (mild cognitive impairment), AD (Alzheimer’s Disease), CVD (cerebrovascular disease), ARTAG (aging-related tau astrogliopathy), LB (Lewy body), PART (primary age-related tauopathy), LATE (limbic predominant age-related TDP-43 encephalopathy).

Fig. 2.

Mouse monoclonal antibody 26A6 shows higher levels of labeling in brain tissue with pathologies. A, Examples of human brain with known pathologies that have been labeled with 26A6 antibody (green), GFAP (red) and DAPI nuclear label (blue) indicated that a large number of astrocytes co-labeled with 26A6 and GFAP in the diseased tissue (yellow in the overlay, indicated by single white arrowheads). The top set of panels showed examples at lower magnification and the bottom panels showed examples with higher magnification. B, An example of a normal human brain section did not label with 26A6 antibody. C, There was heterogeneity of 26A6 labeling of astrocytes—some astrocytes labeled with GFAP and 26A6 (yellow in overlays, single arrowheads), others labeled for 26A6, but not GFAP (indicated by double arrowheads in middle panels), while others were positive for GFAP, but did not label strongly for 26A6 (double arrows, in lower panels). Taken together, these data show that 26A6 robustly labels a subset of astrocytes in tissue with pathologies.

3.3. The immunizing peptide displaces 26A6 labeling, indicating specificity of 26A6 binding to Δ48CN

In order to confirm that the labeling detected by the 26A6 monoclonal antibody in immunostaining was truly due to the presence of Δ48CN in the tissue, we carried out a competition analysis using the immunizing peptide to displace 26A6 labeling in the tissue of a particularly high pathology case (Fig. 3). As was clearly shown, the cellular label by 26A6 was displaced in the presence of the peptide, while there was strong labeling in the control tissue.

Fig. 3.

Mouse monoclonal antibody 26A6 labeling is displaced by the immunizing peptide as a competitor. Using a case that had especially robust staining with 26A6 antibody, we carried out a competition assay to provide evidence that the 26A6 antibody was selectively labeling Δ48 kDa CN in the human brain sections. 26A6 antibody was divided into 2 matching aliquots, and immunizing peptide (75 ug/ml) was added to one half. These were incubated at room temperature for 30 min, and then these aliquots were used for overnight staining of human brain sections, in the same manner as for Fig. 2. The competitor displaced almost all the 26A6 signal (with peptide, green 26A6 panel).

3.4. Δ48 CN is found both around plaques and distributed throughout brain in AD tissue

A particularly high pathology AD case, R5200, was examined in detail by immunofluorescent and colorimetric immunohistochemical staining techniques. As shown in Fig. 4A, brain sections were probed with GFAP, 26A6 and an antibody to amyloid β 1–42. The amyloid antibody detected clear plaques, and in the bottom row of Fig. 4A, a blood vessel in cross-section. Around these amyloid-stained features, 26A6 and GFAP co-labeled astrocytes, best observed as a teal color in the overlay. It was noteworthy that only a subset of astrocytes contain the Δ48 CN, as there were more cells that express GFAP alone than both 26A6 and GFAP. Nevertheless, a high level of primarily astrocytes expressing Δ48 CN were present in AD tissue. To look at this tissue more broadly, this same case was stained for both 26A6 and amyloid β 1–42 using sensitive colorimetric immunohistochemical techniques in Fig. 4B. Again there was a robust peppering of labeled 26A6 cells in brown, as well as both amyloid-labeled plaques and blood vessels in blue. It is noteworthy that there was prominent 26A6-staining around plaques, but not all plaques are surrounded by 26A6 labeled cells. In addition, there were plaque-free regions of tissue that still have robust 26A6 labeling (e.g. see Fig. 4B). In Fig. 5, we carried out a quantitative assessment Δ48 CN distribution from several AD cases. Although there was case-to-case variablity, on the whole there was far greater concentration of Δ48 CN in AD cases compared to neglible amounts in non-demented controls. In general, there was robust expression of Δ48 CN both around plaques and throughout the AD tissue.

Fig. 4.

There is robust expression of Δ48 CN around amyloid plaques and throughout AD tissue A, A particularly high pathology AD case was selected for analysis with amyloid β 1–42 antibody (red channel), 26A6 (green channel) and GFAP (blue channel). There were numerous amyloid-labeled plaques and a blood vessel in cross section. Surrounding these features, there were a number of GFAP and 26A6 co-labeled cells, best seen as teal cells in the overlay and indicated by white arrowheads. There were a few green cells that did not label with GFAP, but the majority of the 26A6-labeled cells were GFAP-positive astrocytes. B, To extend the analysis of this tissue, sensitive colorimetric IHC was also carried out using DAB-labeling of 26A6 containing cells in brown and SG-labeling of β-amyloid in blue. Again, there was robust peppering of Δ48 CN containing cells around the plaques and blood vessels (BV), but there is also expression of Δ48 CN in areas devoid of plaques and blood vessels in the near vicinity. Quantification in Fig. 5.

Fig. 5.

Quantitative assessment of of Δ48 CN around amyloid plaques and throughout AD tissue from multipe cases. Multiple Alzheimer’s Disease (AD) cases (see Table 1, cases R5409, R5200, R5332, R5329, R1259) and non-demented controls (ND, R5434, R5356, 5500, 5501 and 5450) were co-stained for 26A6 and amyloid β plaques, as shown in Fig. 4B for case R5200. The average from 10 fields of view (FOV) from each case was were randomly acquired. Areas with and without Abeta plaques were captured from AD samples. For image analysis, the entire set of images were converted to 8 bit, then background subtraction, and thresholding were homogeneously performed using FIJI software. Number of positive staining, area, and sum intensity were measured and statistically processed using GraphPad Prism. Although there was some variability in AD cases, the level of 26A6 staining was considerably higher than ND controls. The difference between regions with and without plaques was not significant, indicating that Δ48 CN is expressed throughout AD tissue.

3.5. Δ48 CN is enriched in regions of infarcts within brain tissue from subjects with cerebrovascular disease

Since previous work with the polyclonal antibody to ΔCN had shown especially robust labeling of cerebral infarcts (Pleiss et al., 2016), we examined a case from a subject with known infarcts to determine how the monoclonal would react to this same type of feature (Fig. 6). Within the expanse of the infarct, there was considerable 26A6 staining, and many of these 26A6 positive cells co-labeled with GFAP (seen as yellow in overlay). In regions outside the infarct, there were fewer 26A6-positive cells, although these also co-labeled with GFAP. Taken together, these data indicated that there is an association of ΔCN with and around infarcts.

Fig. 6.

Δ48 CN Ab labels infarcts. The SMTG region of case 5444, a case with cerebrovascular disease (CVD) that was known to have large infarcts, was used to analyze the reactivity of the 26A6 antibody to this type of feature. The 26A6 antibody labels a large number of cells intensely within an infarct, and many of these co-stain with GFAP, best seen in the overlay (yellow, examples indicated by arrowheads), although there are also 26A6-positive cells that do not stain with GFAP (examples indicated by double arrowheads). In regions away from the infarct, there are not nearly as many 26A6 labeled cells, although the ones that are present also co-label with GFAP.

4. Discussion

4.1. The 26A6 antibody is a unique tool for assessing the presence of proteolyzed CN in disease

Most of the commercially available antibodies to CN are targeted against the N-terminus or the C-terminus of the protein. Both have their advantages and disadvantages. Antibodies directed towards the N-terminus are able to detect both the full-length CN and all forms of CN cleaved by calpain (Wang et al., 1989; Wu et al., 2004; Wu et al., 2007), including the constitutively active 48 kDa and 45 kDa forms. While these N-terminal antibodies can distinguish between CN forms in Western analysis because of the molecular weight differences on the gels, they cannot distinguish between them in immunostaining, and therefore the utility of these antibodies for this purpose is limited. Antibodies directed to the C-terminus of CN (e.g., Abcam ab71149, Sigma C1956 or Cell Signaling 2614) are even more limited since they cannot detect proteolyzed CN even in Westerns, and thus are only useful for detecting full-length CN. The use of such antibodies may not reveal contributions of the proteolyzed form during brain disease processes (Billingsley et al., 1994; Sidoli et al., 2021). Our lab previously produced a rabbit polyclonal antibody to the Δ48 CN and obtained data quite similar to that seen in this paper (Pleiss et al., 2016), but like all polyclonal antibodies, there was a limited supply. To overcome this problem, we opted to use the same peptide-base immunization strategy used to make the polyclonal to create a monoclonal—the 26A6 monoclonal.

The most attractive feature of the 26A6 monoclonal is that it, like the previous rabbit polyclonal, selectively detects the Δ48 CN (Pleiss et al., 2016) and Fig. 1). Proteolysis of CN occurs both under in vitro conditions and also in intact nervous tissue after cellular insult (Wu et al., 2004; Huang et al., 2005; Shioda et al., 2006; Shioda et al., 2007; Mohmmad Abdul et al., 2011; Rosenkranz et al., 2012). Although this proteolysis has been widely documented, only the advent of the antibodies directed to the 48 kDa C-terminus has allowed us to probe where these proteolysis products accumulate. Comparison of the immunological staining results obtained with the rabbit polyclonal and this new 26A6 monoclonal are strikingly similar. Both antibodies preferentially stain brain tissue from subjects with AD or CVD/infarcts; both antibodies stain robustly around amyloid plaques but also throughout AD tissue; both antibodies preferentially stain in and around infarcts. Although both antibodies stain a small number of non-astrocytic cells, they primarily stain a subset of astrocytes. It’s not clear why 26A6 apparently exhibits low levels of neuronal labeling, as neurons can also show extensive Ca²⁺ dysregulation with AD, and CN is known to become hyperactive in experimental models of AD (e.g. Dineley et al., 2007, Hudry et al., 2012, Hopp et al., 2018). Moreover, calpains have been shown to cleave CN in primary neurons (Wu et al., 2004). Several possibilities for differential labeling of astrocytes and neurons in the present study include: (1) 26A6 selectively detects the 48 kDa species of CN and may have missed other calpain-mediated fragments (e.g. 57 and 45 kDa) that may occur at similar or higher levels in neurons. (2) Calpains may have greater access to CN in astrocytes in vivo, or CN proteolysis may be associated with pathological processes that are unique to astrocytes. (3) Ca²⁺ dysregulation and CN proteolysis may emerge at different disease stages in astrocytes and neurons. For instance, assessment of human cases at the MCI stage of AD may have revealed similar or higher levels of Δ CN48 labeling in neurons. Finally, significant CN proteolysis may be happening in neurons, but it is restricted to dendritic spines (and other micro compartments), where CN and calpains are known to interact, and, as such, may be below the detection methods used here.

4.2. Δ48 CN is likely an endogenous marker of distressed astrocytes

A broad body of work indicates that the Δ48 CN is a marker of distressed astrocytes. Both the Δ48 CN and the Δ45 CN lack the autoinhibitory domain (AID) of the CN A chain, and therefore are constitutively active phosphatases that are no longer regulated by calcium (Wang et al., 1989; Wu et al., 2004; Wu et al., 2007; Kraner and Norris, 2018; Sompol and Norris, 2018). It is noteworthy that the Δ48 CN is only produced in the presence of calmodulin, whether added to the pure human CN samples or present in the endogenous mouse brain extracts (Fig. 1). In the absence of calmodulin, the calpain-cleavage generates the Δ45 CN (Fig. 1). These observations are consistent with those observed previously—that cleavage into the Δ48 CN requires calmodulin (Wang et al., 1989; Wu et al., 2004). Work from others has shown that the Δ48 CN is found in brain extracts from AD tissue (Mohmmad Abdul et al., 2011), and the 26A6 monoclonal is very selective for this form.

High levels of CN proteolysis can dramatically alter cell function and viability (Wang et al., 1999; Wu et al., 2004; Mohmmad Abdul et al., 2011). In cultured astrocytes, expression of the ΔCN leads to the production and release of a variety of cytokines and chemokines linked to neuroinflammation and glial activation (Sama et al., 2008). Astrocytes infected with ΔCN overexpressing viruses also take on a reactive, hypertrophied morphology (Norris et al., 2005). Similar to effects in neurons (Wu et al., 2010; Hopp et al., 2018), expression of ΔCN in astrocytes of hippocampus has been shown to reduce synaptic strength (Pleiss et al., 2016), suggesting that ΔCN in astrocytes leads to impaired neurologic function. However, others have suggested that astrocytic ΔCN could have some beneficial effects in brain, as well (Fernandez et al., 2012). Clearly, additional work is required to assess the nuances of ΔCN signaling in astrocytes associated with neurodenenerative conditions, and the 26A6 antibody could be a useful tool for these investigations.

4.3. Distressed astrocytes and different brain pathologies

The pattern of Δ48 CN-expressing astrocytes was somewhat different in different brain pathologies. In AD, amyloid-positive plaques and blood vessels were robustly surrounded by these astrocytes, but there was also a peppering of labeled astrocytes throughout brain tissue in these cases (Fig. 4). In addition, there was variability in the degree to which different AD-positive cases expressed Δ48 CN-positive astrocytes (Fig. 5). Tissue from control subjects had negligible labeling (Fig. 1B and Fig. 5). These data are similar to what we observed previously with the rabbit polyclonal antibody (Pleiss et al., 2016) and may account for the elevated levels of CN and CN-dependent signaling mediators, such as NFATs, reported in previous data sets for both mouse models of AD and related dementias, as well as human disease cases (Wu et al., 2004; Reese et al., 2008; Abdul et al., 2009; Mohmmad Abdul et al., 2011; Serrano-Perez et al., 2011; Neria et al., 2013; Caraveo, Auluck et al., 2014).

Previous studies have shown that CN proteolytic fragments appear with both stroke/global ischemia (Shioda, Moriguchi et al., 2007, Rosenkranz et al., 2012) as well as microinfarcts (Pleiss et al., 2016). Consistent with these observations, the present study observed extensive co-labeling of high cerebral pathology cases with 26A6, especially in GFAP-positive cells within infarcted regions (Fig. 6). However, labeling was highly variable and numerous GFAP-negative/Δ48 CN positive cells were also prominent. The high level of Δ48 CN expression with infarcts may contribute to high levels of morbidity and co-morbidity in dementia and suggests a link between Δ48 CN and anoxic/ischemic processes (Kraner and Norris, 2018; Price et al., 2018). These data were somewhat similar to those found previously with a rabbit polyclonal antibody to Δ48 CN (Pleiss et al., 2016; Kraner and Norris, 2018), except the polyclonal antibody appeared to label areas within the infarct as well as the apparent scar-forming astrocytes on the periphery of the infarct. These difference could be due to recognition of both 45 and 48 kDa species of ΔCN by the polyclonal antibody. Interestingly, the polyclonal antibody also tended to label more neurons in diseased tissue than 26A6.

Increasingly, it is recognized that infarcts and microinfarcts contribute to vascular dementia, either by themselves or in conjunction with other neurodegenerative diseases such as AD (O’Brien et al., 2003; Kalaria et al., 2012; van Norden et al., 2012; Raz, Daugherty et al., 2015; Nelson et al., 2016; Vemuri and Knopman, 2016; Wilcock et al., 2016; Corriveau et al., 2017; Price et al., 2018; Shih et al., 2018). Work from our group and others has shown that hyperactive CN signaling arising from amyloid pathology, cerebrovascular pathology, and other sources exacerbates neurodegenerative processes and hastens cognitive decline (Reese and Taglialatela, 2011; Furman and Norris, 2014; Kraner and Norris, 2018; Sompol and Norris, 2018; Price et al., 2021). Moreover, mounting evidence gathered from preclinical models suggests that the inhibition of CN signaling, either through genetic or pharmacologic approaches, ameliorates numerous pathophysiologic and cognitive phenotypes of AD and ADRDs (e.g. Dineley et al., 2007; Dineley et al., 2010; Wu et al., 2010; Hudry et al., 2012; Furman et al., 2012; Sompol et al., 2017; Radhakrishnan et al., 2021; Sompol et al., 2021; Sompol et al., 2023). This work is consistent with recent epidemiologic studies showing that use of FDA approved CN inhibitors is associated with a reduced prevalence of dementia in human populations (Taglialatela et al., 2015; Silva et al., 2023). Thus, tools for identifying the appearance of proteolytic forms of CN (e.g. the monoclonal antibody 26A6) could be very important for determining which pathologies may benefit from CN-inhibiting strategies.

5. Conclusions

The new monoclonal 26A6 is highly selective for the Δ48 kDa CN proteolytic fragment and labels a subset of astrocytes and, possibly other cell types, under pathological conditions. In brain tissue with infarcts, there was considerable concentration of 26A6-positive astrocytes within/around infarcts, suggesting a link with anoxic/ischemic pathways. Cell specific labeling of Δ48 CN with 26A6 opens up the possibility of using newer spatial profiling technologies to discover how these cells differ from those which do not express this protein, and perhaps confirm the effects of cell Ca²⁺ changes on a cell-by-cell basis.

Abbreviations:

AD

Alzheimer’s Disease

N-terminus

amino-terminus

AID

Autoinhibitory Domain

CN

calcineurin

ΔCN

calcium (Ca2 +) calpain-cleaved calcineurin

C-terminus

carboxy-terminus

CVD

cerebrovascular disease

EDTA

ethylenediaminetetraacetic acid

EGTA

ethylene glycol tetraacetic acid

GFAP

glial fibrillary acid protein

MOPS

3-(N-morpholino) propanesulfonic acid

ND

Non-demented

NFAT

Nuclear Factor of Activated T Cells

PBS

phosphate buffered saline

SMTG

superior and middle temporal gyri

Data availability

Data will be made available on request.