Introduction

The endogenous cannabinoid system consists of endocannabinoids (e.g., anandamide and 2-arachidonoylglycerol), the enzymes responsible for their synthesis and degradation, and cannabinoid receptors and transporters [1, 2]. Two types of cannabinoid receptors (CB1Rs and CB2Rs) have been identified. CB1Rs are highly expressed in the brain and functionally modulate presynaptic neurotransmitter release [1, 2]. In contrast, CB2Rs were initially identified in the spleen and regarded as peripheral cannabinoid receptors [3, 4]. This view has been challenged by recent findings that CB2R and its mRNA are expressed in the brains of rats and mice [5,6,7,8,9]. In addition, Western blot (WB) assays and immunohistochemistry (IHC) consistently detected CB2R signaling in multiple brain regions and neuronal phenotypes [5, 6, 10,11,12,13,14,15,16]. Quantitative real-time polymerase chain reaction (qRT-PCR) analysis has also identified CB2 mRNA and its isoforms in several regions of the central nervous system, including the retina [17, 18], cortex [8, 19,20,21], striatum [20, 21], hippocampus [21, 22], amygdala [19, 20], brainstem [5], and cerebellum [23]. In situ hybridization (ISH) assays revealed CB2R mRNA expression in several neuronal phenotypes, including glutamatergic neurons in the cortex and hippocampus [22, 24, 25] and dopaminergic (DA) neurons in the midbrain [9, 26,27,28,29]. Electrophysiological assays confirmed the presence of functional CB2Rs in brain glutamatergic neurons [22, 30,31,32], GABAergic neurons [33], and DA neurons [9, 22, 26, 29]. However, the specificity of the detected CB2R signals was questioned because CB2-knockout (CB2-KO) mice were not used as controls in many early studies [34]. Recent findings that CB2R antibody signals were detected in wild-type (WT) and CB2-KO mice support this skepticism [35, 36]. A C-terminal-deleted partial CB2-KO strain was used in those studies, and the expression of mutant or truncated CB2-like proteins may have partially contributed to the observed “non-specific” binding. Anti-rat or anti-human CB2R antibodies were used in those studies [35, 36], and species differences in antibody epitopes and CB2R structures [8, 28] may also confound interpretations of antibody signal specificity. Methodological limitations may be a valid reason for skepticism regarding CB2R expression in the brain. Therefore, determinations of CB2R antibody signal specificity are urgently needed to understand the presence and function of CB2Rs in the brain and the potential utility of CB2R ligands in the treatment of various neuropsychiatric disorders.

The present study used multiple approaches to investigate CB2R signal specificity. We used WB assays and four antibodies that targeted the receptor-deleted or -undeleted regions in CB2-KO mice. We used double-label fluorescent IHC assays and the same antibodies to examine CB2R immunostaining in midbrain DA neurons in different mouse genotypes. Midbrain DA neurons were chosen because functional CB2Rs were found in this region [7, 9]. We used qRT-PCR and three TaqMan probes to examine and compare CB2R gene (mRNA) expression in different mouse genotypes. All of these assays included multiple positive and negative controls, including immune peptides, CB2-rich spleen tissue, CB1-KO mice, and two strains of currently available CB2-KO mice (i.e., the N-terminal-deleted Deltagen strain and the C-terminal-deleted Zimmer strain) to determine the specificity of detected CB2R signals.

MATERIALS AND METHODS

Animals

Male WT, CB1-KO [37], and two strains of CB2-KO mice [4, 30] with C57BL/6J genetic backgrounds were bred at the National Institute on Drug Abuse (NIDA). Genotyping was performed in our laboratory prior to experimentation. All animals used in the present experiments were matched for age (8–14 weeks) and weight (25–35 g). Mice were housed individually in a climate-controlled animal colony room on a reversed light–dark cycle (lights on at 7:00 P.M., lights off at 7:00 A.M.) with free access to food and water. The animals were maintained in a facility fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. All experimental procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals of the U.S. National Research Council, and the Animal Care and Use Committee of the NIDA of the U.S. National Institutes of Health approved all procedures.

Western immunoblotting assays

Four CB2 antibodies were used. (1) The Abcam rat CB2 (rCB2) polyclonal antibody was purchased from Abcam (ab-3561, Abcam PLC, Cambridge, MA, USA) with epitope (amino acids 1–32) at the rCB2R N-terminal. The epitopes of rat and mouse CB2 (mCB2) receptors differ by 5 amino acid residues. (2) The Alomone rCB2 polyclonal antibody was purchased from Alomone (ACR-002, Alomone Labs, Jerusalem, Israel), which recognizes the third intracellular loop. The epitope of the Alomone antibody amino acids (228–242 amino acids) is identical between rCB2Rs and mCB2Rs. (3) The Mackie rCB2 antibody was provided by Dr. Ken Mackie at Indiana University. The epitope (amino acids 326–342) of the Mackie antibody differs by 3 amino acids between rCB2Rs and mCB2Rs. (4) A mCB2 polyclonal antibody (NIH5633, Baltimore, MD, USA) was custom designed. The epitope (amino acids 326–340) is located at the mCB2R C-terminal. The NIH5633 mCB2 antibody was produced by Genemed Synthesis, Inc. (San Antonio, TX, USA). Table 1 details the epitopes of these CB2 antibodies. The immune peptide for each individual antibody stated above was purchased from the same antibody provider and the amino acid sequences (epitopes) of the immune peptides are highlighted in Table 1.

All mice for WB assays were perfused transcardially with 0.9% saline under deep anesthesia to prevent contamination of brain tissue with CB2-rich immune cells in blood. Whole striatum and the spleen were dissected. Tissues were homogenized in RIPA lysis buffer (Cell Signaling Technology, Inc., Danvers, MA, USA), and the protein concentration of each sample was quantified using a Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA). A total of 30 μg spleen proteins and 50 μg striatal proteins were used for Western immunoblot assays. Membranes were incubated with a rabbit anti-CB2 antibody (Abcam, 1:2000; Alomone, 1:250; NIH5633, 1:1000; Mackie, 1:500) and mouse anti-β-actin (1:2500) (Sigma-Aldrich, St. Louis, MO, USA) and incubated with secondary antibodies, goat anti-mouse IgG for β-actin (IRDye 680CW), and goat anti-rabbit IgG for CB2 (IRDye 800CW) (LI-COR Bioneurosciences, Lincoln, NE, USA). Membranes were scanned using a LI-COR Odyssey Image System (LI-COR Biosciences, Lincoln, NE, USA). Band density was measured using the Image J software (http://rsb.info.nih.gov/ij/).

Table 1 The amino acid sequences of CB2Rs in rats, mice, and two strains of CB2-KO mice
Full size table

IHC assays

The IHC procedures were performed as reported previously [9]. Briefly, mice were deeply anesthetized and transcardially perfused with cold saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer. Brain tissues were transferred to 20% sucrose in phosphate buffer at 4 °C overnight. Coronal sections were cut at 10 μm on a cryostat (CM3050S, Leica Microsystems Nussloch GmbH, Nussloch, Germany). Tissue sections containing the ventral tegmental area (VTA) were blocked and floated in 5% bovine serum albumin and 0.5% Triton X-100 phosphate buffer for 2 h at room temperature. Dual-labeling IHC was performed using one of the above-listed CB2 antibodies (Abcam, 1:1000; Alomone, 1:250; NIH5633, 1:500; Mackie, 1:500) and an anti-tyrosine hydroxylase (anti-TH) monoclonal antibody (1:500; Millipore, Billerica, MA, USA). Sections were washed and incubated with a mixture of secondary antibodies, goat anti-rabbit Alexa 488 for CB2 receptors, and goat anti-mouse Alexa 568 for TH (1:500) in 5% bovine serum albumin and 0.5% Triton X-100 phosphate buffer for 2 h at room temperature. Sections were washed, mounted, and cover slipped. Fluorescent images were captured using a fluorescence microscope (Nikon Eclipse 80i) equipped with a digital camera (Nikon Instruments Inc., Melville, NY, USA). All images were captured under identical optical conditions. Densitometric analysis was used to quantify CB2 immunostaining density on individual VTA DA neurons and determine whether deletion of CB2Rs abolished or attenuated the expression of CB2Rs in VTA DA neurons [23]. Each CB2-positive DA neuron was outlined manually, and CB2 fluorescence intensity was measured using the Image J software. The background signal was defined as the mean background from 5 to 10 regions outside of DA neurons in each slice. The background signal was subtracted, and the ratio F/A was used to define the mean fluorescence of individual DA cells (F) normalized to total cellular surface (A). Quantification was performed on >100 cells from 2 to 5 animals of each strain.

qRT-PCR assays

Immune cells in blood contain a high density of CB2 receptors. Therefore, all mice used for qRT-PCR were perfused transcardially with 30–50 mL 0.9% saline under ketamine and xylazine anesthesia to prevent contamination of brain tissue. Brains and spleen were removed, and the prefrontal cortex, striatum, and midbrain were dissected. Total RNA was extracted using a RNeasy Mini Kit (QIAGEN, Valencia, CA, USA), according to the manufacturer’s instructions. The purity and integrity of each extract were determined using absorbance at 260 nm in an Eppendorf BioPhotometer Plus (Eppendorf AG, Hamburg, Germany).

qRT-PCR procedures for the detection of mCB2 mRNA were performed as reported previously [8, 9]. Briefly, three specific mCB2 probes were used: a mCB2A probe that recognizes the conjunction region of encoding exons 1 and 3 (aligns with base pair positions 89–182 of the cDNA NM_009924.2); a mCB2-Zimmer-ko probe (for the Zimmer CB2-KO strain) that targets a region close to the 3′ end of exon 3 (aligns with base pair positions 885-989 of NM_009924.2; the deleted exon 3 aligns with base pair positions 887–1227 of NM_009924.2); and a mCB2-Deltagen-ko probe (for the Deltagen CB2-KO strain) that targets a region near the 5′ end of exon 3 (aligns with base pair positions 336-409 of cDNA NM_009924.2; the deleted exon 3 aligns with base pair positions 259–593 of cDNA NM_009924.2). Mouse β-actin mRNA was used as an endogenous control. The specific base pair sequences of the MGB-Taqman probes and the primers used to detect mCB2 and β-actin mRNAs were reported previously. [9] All Taqman probes and primers were purchased from Applied Biosystems (Foster City, CA, USA).

qRT-PCR reactions were performed using a QIAGEN OneStep RT-PCR Kit (Catalog number 210212, QIAGEN Inc., Valencia, CA, USA). Each qRT-PCR assay was performed in triplicate in 96-well plates. The following thermal cycle conditions were used: reverse transcription at 50 °C for 30 min and an initial PCR activation step at 95 °C for 15 min followed by 40 PCR cycles of 94 °C for 1 min, 60 °C for 30 s, and 72 °C for 30 s. Final extension was performed at 72 °C for 10 min. qRT-PCR analyses of CB2-mRNA levels were performed using the 2−ΔΔCt method [38]. Data in the present study are presented as the fold-change in mCB2 gene expression normalized to the internal β-actin control gene and relative to normal cortex (control tissue) in WT mice. The cycle threshold (Ct) was defined as the number of cycles required for the fluorescent signal to cross the threshold (i.e., exceed the background level). ΔCt was determined as [mean of the triplicate Ct values for the mCB2 gene]−[mean of the triplicate Ct values for β-actin]. ΔΔCt represents the difference between the paired tissue samples as calculated by the formula ΔΔCt = [ΔCt of mCB2 in sample tissue−ΔCt of mCB2 in control tissue]. The N-fold differential expression of the mCB2 gene in spleen or brain tissues compared to control tissue (i.e., cortex in WT mice) is expressed as 2−ΔΔCt [8].

Data analyses

All data are presented as the means (±S.E.M.). One-way analysis of variance was used to analyze differences in the density of CB2 immunostaining between mouse strains. Individual group comparisons were performed using Student–Newman–Keuls post hoc test.

Results

CB2R gene structure, transcripts, and receptor proteins

The structures of the CB2R gene, transcripts (mRNA) and receptor proteins must be known in detail to determine whether a detected signal is mCB2R-specific. Figure 1 shows the mCB2R gene structure, transcripts, and predicted receptor expression in WT mice and two strains of partial CB2-KO mice. The mCB2 gene consists of three exons with two separate promoters (P1 and P2) [8, 13, 28], which encode two transcripts (mCB2A and mCB2B) using exon 1 or exon 2 as different 5′ untranslated region sequences. The mCB2A transcript contains exon 1 and exon 3, and the mCB2B transcript contains exon 2 and exon 3. The CB2R-encoding regions are located entirely on exon 3.

Fig. 1
figure 1

Diagrams illustrating the structures of the mCB2R gene (a), two transcripts (b), the gene-deleted regions, and the predicted CB2Rs in two strains of CB2R-KO mice (c)

Full size image

CB2-KO mice used in the present study

The ideal negative control for determining CB2R signal specificity would be the use of full CB2-KO mice. However, such mice are not currently available. Therefore, we used the two available partial CB2-KO strains, the Zimmer strain [4, 39] and the Deltagen strain (The Jackson Laboratory, Cnr2tm1Dgen/J) [30, 40]. The Zimmer strain has a C-terminal 131 amino acid deletion. This mutation eliminates part of the intracellular and extracellular third loops, trans-membrane regions 6 and 7, and the intracellular C-terminus region (Fig. 1). The Deltagen strain has an N-terminal 112 amino acid deletion. This mutation causes loss of part of the extracellular N-terminal (from amino acid residues 26 to 137), trans-membrane regions 1–3, and intracellular loops 1 and 2 (Fig. 1; Table 1).

CB2 receptor antibodies used in this study

Four different antibodies were used to examine CB2R expression using WB and IHC assays. Two strains of N-terminal and C-terminal CB2-KO mice were used as controls. Therefore, we chose different antibodies that targeted the receptor-deleted region or non-deleted regions. For example, we hypothesized that an antibody with an epitope at the N-terminal of CB2R (such as Abcam rCB2-Ab) should detect CB2R signals in WT and Zimmer strain CB2-KO mice when the C-terminal-deleted Zimmer strain of CB2-KO mice was used as a control, and CB2 antibodies with epitopes at the C-terminal (e.g., the Alomone rCB2-Ab, the NIH5633 mCB2-Ab, or the Mackie rCB2-Ab) should only detect a CB2R signal in WT mice and not in the Zimmer strain of CB2-KO mice. Similarly, antibodies with epitopes at the N-terminal or C-terminal of CB2Rs should detect a CB2R signal only in WT and not in Deltagen CB2-KO mice when the N-terminal-deleted Deltagen strain of CB2-KO mice was used as a control. Table 1 shows the amino acid sequences of CB2Rs in rats, WT mice and two strains of CB2-KO mice and the epitopes of the antibodies used in this study. There are 3–5 different amino acid residues (marked by the symbol asterisk (*) in Table 1) in the epitopes of the Abcam and Mackie rCB2 antibodies between rat and mouse CB2Rs. The epitope of the Alomone rCB2 antibody is identical in rat and mouse CB2Rs.

CB2-immunoreactive bands are detected in WT and CB2-KO mice

Figure 2 shows the WB results of the four different antibodies against rCB2 or mCB2 receptors. We found that all antibodies detected multiple bands in brain tissue (striatum) and one band in the spleen. We identified one band at ~40 kD that was likely CB2R-specific in the brain and spleen based on the predicted molecular weight of mCB2Rs. Antibody preabsorption with specific immune peptides blocked the immunoreactive bands (Fig. 2). The Abcam rCB2 antibody with an epitope at the N-terminal of CB2Rs (Fig. 1c) detected similar densities of CB2R bands in WT, CB1-KO, and Zimmer CB2-KO mice. The other three antibodies that targeted the C-terminal-deleted region (Fig. 1c) detected relatively lower CB2R band densities in the C-terminal-deleted Zimmer CB2-KO mice compared to WT mice.

Fig. 2
figure 2

Western blot results in WT, CB1-KO, and Zimmer CB2-KO mice using four different CB2 antibodies. a A diagram showing the binding sites of the four antibodies used in WB assays. b The Abcam rCB2-Ab detected an mCB2-like band at ~40 kD (mCB2 monomer, 39 kD) in the striatum of all three mouse genotypes, but the other three antibodies (Alomone, NIH5633, Mackie) detected relatively lower densities of an mCB2-like band in Zimmer CB2-KO mice. c CB2-immunoreactive bands detected in the spleen using the same antibodies. More bands were detected in the striatum than the spleen by the multiple antibodies. Preabsorption of the CB2 antibodies (Abcam, Alomone, NIH5633) with specific immune peptides blocked the immunoreactive bands

Full size image

Figure 3 shows similar findings in the N-terminal-deleted Deltagen strain CB2-KO mice. The Abcam rCB2 antibody with an epitope at the N-terminal-deleted region (Fig. 3a) detected a significantly lower density of CB2 band in the Deltagen CB2-KO mice, and the other two antibodies (NIH5633 and Alomone) with epitopes at the C-terminal of mCB2Rs (Fig. 3a) detected similar CB2R signal densities in WT and Deltagen CB2-KO mice (Fig. 3b). The downstream CB2 gene sequence from the gene-deleted region should not encode receptor protein because of the presence of an early stop codon, and this finding appears to directly undermine the signal specificity detected by these two antibodies.

Fig. 3
figure 3

Western blot results in WT, CB1-KO, and Deltagen CB2-KO mice using three different CB2 antibodies. a A diagram showing the binding sites of the three antibodies used in WB assays. b The Abcam rCB2 antibody detected a much lower density mCB2-immunoreactive band in the Deltagen CB2-KO mice than in WT mice, and the other two antibodies detected similar levels of the mCB2-like band in all three mouse genotypes. Ab antibody

Full size image

CB2R immunostaining is detected in midbrain DA neurons

Previous findings demonstrated functional CB2Rs in midbrain DA neurons [9, 41, 42]. Therefore, we used double-label fluorescence IHC assays to examine CB2 immunostaining in midbrain DA neurons in all four strains of WT, CB1-KO, and CB2-KO mice. Figure 4 shows the CB2 immunostaining of the Abcam rCB2-Ab (with an epitope at the extracellular N-terminal, Fig. 4b, c) and reveals significant CB2R immunostaining in TH-positive DA neurons in VTA in WT, CB1-KO, and Zimmer CB2-KO mice. However, the density was significantly lower in the N-terminal-deleted Deltagen CB2-KO mice compared to WT mice (Fig. 4a–d). Figures 5 and 6 show that the Alomone and NIH5633 antibodies with epitopes at the C-terminal of CB2Rs detected significantly lower densities of CB2R immunostaining in C-terminal-deleted Zimmer strain CB2-KO mice than in WT or CB1-KO mice. Both antibodies detected marked CB2R immunostaining in the Deltagen CB2-KO mice, but the mean density was significantly lower than that in WT mice. Notably, the Mackie rCB2 antibody detected little-to-no CB2R immunostaining in the VTA DA neurons in WT mice (Fig. 7).

Fig. 4
figure 4

mCB2 immunostaining in the midbrain DA neurons using the Abcam rCB2 antibody. a Representative IHC staining results illustrating that the Abcam rCB2-Ab detected relatively lower CB2 immunostaining in the Deltagen strain of CB2-KO mice than in the other strains of mice. b, c Diagrams showing the receptor-deleted regions and the binding site of the Abcam rCB2 antibody in both strains of partial CB2-KO mice. d The mean densities of mCB2 staining in DA neurons (over 200–400 DA neurons from 2 to 3 mice per genotype) illustrating a significant reduction in the Deltagen CB2-KO mice compared to the other strains. ***P < 0.001, compared to WT mice

Full size image
Fig. 5
figure 5

mCB2 immunostaining in midbrain DA neurons using the Alomone rCB2 antibody. a Representative IHC staining results illustrating that the Alomone rCB2 antibody detected significantly lower CB2 immunostaining in the Zimmer strain of CB2-KO mice. b, c Diagrams showing the receptor-deleted regions and the binding site of the Alomone rCB2 antibody in both strains of CB2-KO mice. d The mean densities of mCB2 staining in DA neurons (over 200–400 DA neurons from 2 to 3 mice per genotype) illustrating a significant reduction in mCB2 immunostaining in the Zimmer and Deltagen CB2-KO mice compared to the other strains. ***P < 0.001, compared to WT mice

Full size image
Fig. 6
figure 6

mCB2 immunostaining in the midbrain DA neurons using the NIH5633 mCB2 antibody. a Representative IHC staining images illustrating that the NIH5633 mCB2 antibody detected relatively lower CB2 immunostaining in the Zimmer strain of CB2-KO mice than in the other strains of mice. b, c Diagrams showing the receptor-deleted regions and binding site of the NIH5633 mCB2 antibody in both strains of CB2-KO mice. d The mean densities of mCB2 staining in DA neurons (over 200–400 DA neurons from 2 to 3 mice per genotype) illustrating a significant reduction in the Zimmer and Deltagen CB2-KO mice compared to the other strains. ***P < 0.001, compared to WT mice

Full size image
Fig. 7
figure 7

Representative mCB2R immunostaining using the Mackie rCB2 antibody illustrating non-significant CB2R immunostaining in VTA DA neurons in WT mice

Full size image

CB2 mRNA is detected in the brains of both strains of CB2-KO mice

We examined CB2 mRNA in the brain to determine whether the non-specific binding in the above assays was related to the presence of a CB2R gene that encoded CB2R proteins in these partial CB2-KO mice. We previously reported that the upstream sequence of the CB2R gene is present and detectable in the Zimmer strain of CB2-KO mice, but the designed gene-deleted region is not [9, 28]. Therefore, we further examined CB2 mRNA expression in the Deltagen CB2-KO mice to investigate whether any complete CB2R gene remained in this strain. Figure 8a shows the mCB2A transcript, CB2 gene-deleted regions in the Deltagen CB2-KO mice, and the target regions of three mCB2 TaqMan probes on mCB2 transcripts. The mCB2A probe targeted the upstream undeleted gene sequence, and the mCB2-Deltagen-ko probe targeted the deleted gene sequence in the Deltagen CB2-KO mice. The mCB2-Zimmer-ko probe targeted the downstream gene sequence in Deltagen CB2-KO mice. The mCB2A probe detected the same levels of mCB2 mRNA signal in the cortex and spleen in WT and Deltagen CB2-KO mice (Fig. 8b), and the mCB2-Deltagen-ko probe failed to detect any CB2 mRNA signal in Deltagen CB2-KO mice (Fig. 8c). The mCB2-Zimmer-ko probe detected similar levels of the downstream CB2 mRNA signals in WT and Deltagen CB2-KO mice (Fig. 8d). These findings suggest that the complete CB2 gene sequence is not present, but the gene sequences downstream from the gene-deleted region remain in Deltagen CB2-KO mice.

Fig. 8
figure 8

mCB2 mRNA expression in the Deltagen CB2-KO mice. a mCB2A transcripts, the CB2 gene-deleted regions in Deltagen CB2-KO mice, and the gene sequences detected by three different Taqman probes. b The mCB2A probe that targets the undeleted gene regions detected similar levels of mCB2-mRNA expression in the prefrontal cortex and spleen in WT and CB2-KO mice. c The mCB2-Deltagen-ko probe detected mCB2 mRNA only in WT, not in Deltagen CB2-KO, mice. d The mCB2-Zimmer-ko  probe that targets the gene-deleted sequence on exon 3 in Zimmer CB2-KO mice detected mCB2 mRNA in WT and Deltagen CB2-KO mice. NM_009924.2 is the cDNA code in the GenBank of National Center for Biotechnology Information (NCBI)

Full size image

Discussion

The present study demonstrated that multiple CB2R antibodies detected several CB2-like immunoreactive bands, including one at ~40 kD, in the mouse brain and spleen, as well as CB2R immunostaining in midbrain DA neurons in WT mice. However, we detected a relatively lower CB2R signal in the partial CB2-KO mice compared to WT controls using an antibody that targeted the receptor-deleted region. In contrast, we detected equivalent levels of CB2R signal in WT and CB2-KO mice using antibodies that targeted the predicted upstream or downstream receptor regions from the predicted receptor-deleted regions. qRT-PCR assays using three different probes that targeted the upstream, deleted, or downstream gene sequences in CB2-KO mice detected CB2 mRNA signals only at the upstream or downstream gene sequences and not in the designed gene-deleted region in the Deltagen CB2-KO mice. These findings suggest that antibody-based CB2R signals are not highly mCB2R-specific when the currently available partial CB2-KO mice are used as negative controls, which supports the urgent need for further development of full CB2-KO mice.

The presence of CB2Rs in the brain is controversial [34]. Extensive research over the past decade indicates that functional CB2Rs are expressed in glial cells and neurons in the brain[1, 43, 44], and evidence from qRT-PCR and ISH assays in our present study and previous reports strongly support this expression [9, 26, 28]. Electrophysiological evidence of brain CB2R modulation of neuronal activity in the cortex [31,32,33], hippocampus [22, 25, 30], striatum [29], and midbrain [9, 26] further support the presence of functional CB2Rs in the brain. We recently reported that systemic or local administration of the selective CB2R agonist JWH133 or GW405833 into the nucleus accumbens significantly inhibited DA release in this region and attenuated intravenous cocaine self-administration, cocaine-enhanced locomotion, and cocaine-enhanced extracellular DA in the nucleus accumbens. [7] The impact of CB2R agonism on cocaine’s effects was blocked by AM630, a selective CB2R antagonist, and was absent in Zimmer strain partial CB2-KO mice. The CB2R agonist JWH133 also inhibited cocaine-induced conditioned place preference and hyperactivity in rats [45] and cognitive and impulsive-like behavior in mice [20]. The CB2R antagonist AM630 produced enhanced anxiety [19] and anti-depressive effects [46]. Overexpression of CB2Rs in the brain decreased anxiety [47] and depression-like behaviors [46], and deletion of CB2 receptors increased food intake and body weight [48, 49] and caused schizophrenia-like effects [50]. Activation of CB2 receptors in the brain inhibited emesis in ferrets [5] and produced neuroprotective effects [23, 51,52,53]. Activation of CB2 receptors in the spinal cord or thalamus inhibited spontaneous and evoked neuronal responses to noxious stimuli [54,55,56,57]. Taken together, these data indicate that functional CB2Rs modulate a variety of neuronal activities and brain functions and strongly implicate brain CB2 receptors in behaviors that are reliant on the mesolimbic DA system. However, this conclusion is not fully supported by our present findings of the WB and IHC assays, which indicate that CB2R antibody signals may not be mCB2R-specific.

WB and IHC assays are commonly used for the identification of specific protein expression in the brain and the periphery. However, poor antibody specificity is often invoked to invalidate a conclusion that is fully supported by other studies. An ideal antibody should meet most or all of the following criteria [58,59,60]: (1) immunoblot bands in WB assays should match the correct molecular weight of a target protein; (2) different antibodies against different epitopes of the same proteins should yield the same results in WB and IHC assays; (3) an ideal antibody should produce consistent results in WB and IHC assays; (4) an immunizing peptide should block immunolabeling; (5) immunolabeling should be consistent with the results generated by other independent techniques, such as autoradiography, RT-PCR, ISH, electrophysiology, and in vivo behavioral assays; (6) immunolabeling should be observed in target protein-positive tissues (positive control); and (7) immunolabeling should be abolished in tissues from target gene-deleted animals (i.e., KO animals and negative controls). Our findings in the present study and other previous reports [7,8,9, 22, 26, 28] clearly meet most of the above criteria, except for the findings in the partial CB2-KO mice. The results in WB and IHC assays suggest that none of the tested CB2R antibodies are highly mCB2-specific. The four antibodies used exhibited only a certain degree of mCB2 specificity when the appropriate partial CB2-KO strain was used as a control.

The reasons for non-specific antibody binding in CB2-KO mice are not clear. There are at least three possibilities. First, the detected signals in this study are not mCB2R-specific or are totally non-specific. However, this explanation is not supported by the finding that certain antibodies consistently detected significantly lower (50–70% reduction) CB2-like signals in WB and IHC assays in CB2-KO mice compared to WT mice. A second possible reason for the apparent non-specific antibody binding may be related to the presence of mutant or truncated CB2R fragments in partial CB2-KO mice. As noted above, the gold standard for a negative control would be full CB2-KO mice, but these animals are not currently available. We used two CB2-KO mouse strains with partial CB2R gene deletion at the N-terminal or the C-terminal in this study as an alternative. We found that the designed gene-deleted sequence was absent, but the upstream and downstream CB2 gene sequences remained in the Deltagen strain CB2-KO mice and the Zimmer strain CB2-KO mice, as reported previously [9]. These findings suggest that the undeleted upstream or downstream gene sequences may encode mutant or truncated CB2R proteins or fragments in these partial CB2-KO mice, which may underlie the non-specific binding observed in the present study.

Few studies examined mutant CB2R expression in either strain of CB2-KO mice. It is generally believed that upstream CB2R fragments may be present in the Zimmer CB2-KO mice (Fig. 1), but these fragments may be unstable and degraded intracellularly. Our finding of the detection of multiple bands in brain tissues of the C-terminal-deleted Zimmer CB2-KO strain does not support this hypothesis and suggests the possible presence of CB2R fragments or their coagulations. Notably, multiple antibodies that targeted the downstream CB2R regions detected CB2-like signals in the N-terminal-deleted Deltagen CB2-KO mouse strain. However, the downstream CB2R fragments should not be expressed because of the presence of an early stop codon in the gene-replaced region (by neomycin). More studies are required to further address this issue. We recommend caution when using partial CB2-KO mice as negative controls for WB and IHC assays. It is likely that an antibody may detect CB2-like band(s) or immunostaining in partial CB2-KO mice, depending on the antibody epitope and the strain of the partial CB2-KO mice used.

A third possibility for the non-specific antibody binding may be related to the use of anti-rat CB2R polyclonal antibodies for the detection of mouse brain mCB2R expression. Table 1 shows that CB2Rs exhibit significant species differences in amino acid sequences. Therefore, we urge the use of species-specific CB2 antibodies to investigate CB2R expression in different species. For example, mouse CB2R antibodies with epitopes at the CB2R N-terminal should be used for investigation of CB2R signal specificity when N-terminal-deleted Deltagen CB2-KO mice are used as controls. Unfortunately, mCB2R-specific antibodies that targets the gene-deleted regions in both partial CB2-KO strains are not available. Therefore, we used an rCB2 antibody (Abcam) with epitope at the N-terminal (Fig. 1c) in the present study. This antibody has five different amino acid residues in its epitope between rCB2R and mCB2Rs (Table 1), and we presume that these differences may partially contribute to the relatively poor specificity observed in the present study. The use of the Abcam rCB2 and Mackie rCB2 antibodies in investigations of CB2R expression in the mouse is also problematic because the epitopes are significantly different between rCB2Rs and mCB2Rs (Fig. 1c; Table 1). We note that the Deltagen CB2-KO strain is not a full N-terminal KO because the N-terminal 1–25 amino acid residues are likely expressed and only amino acids 26–137 are likely deleted. The presence of the N-terminal fragment (residues 1–25) in the Deltagen CB2-KO strain may explain the “non-specific” binding observed in that strain. We note that the Alomone rCB2 and NIH5633 mCB2 antibodies should not detect CB2R signals in the C-terminal-deleted Zimmer strain CB2-KO mice because the epitope of the Alomone antibody is identical in rat and mouse CB2Rs. The non-specific binding detected by both of these antibodies may be related to the use of polyclonal, rather than monoclonal, CB2 antibodies.

In conclusion, none of the tested four antibodies were highly mCB2R-specific. However, the presence of non-specific binding is not surprising because poor specificity is a common problem in all antibody-based assays [58,59,60], not only in CB2R research. Unexpected findings (e.g., non-specific binding by polyclonal CB2 antibodies) in WB and IHC assays should not be used to invalidate a conclusion that is otherwise well supported by many other studies. It is likely that some antibodies detect CB2R-like signals in partial CB2-KO mice because mutant or truncated CB2R proteins may be expressed in these mice. Ideally, the use of anti-mouse monoclonal CB2R antibodies in combination with full CB2-KO mice as controls would determine CB2R antibody signal specificity. The findings in the present study do not invalidate the expression of CB2Rs in the brain but provide additional evidence to support our previous finding of the expression of functional CB2Rs in multiple neuronal phenotypes. Our findings do not suggest that partial CB2-KO mice are useless. In contrast, these mice are valid controls in receptor gene assays and various functional and behavioral assays.