Transient Gain of Function of Cannabinoid CB₁ Receptors in the Control of Frontocortical Glucose Consumption in a Rat Model of Type-1 Diabetes

Abstract

Here we aimed to unify some previous controversial reports on changes in both cannabinoid CB₁ receptor (CB₁R) expression and glucose metabolism in the forebrain of rodent models of diabetes. We determined how glucose metabolism and its modulation by CB₁R ligands evolve in the frontal cortex of young adult male Wistar rats, in the first 8 weeks of streptozotocin-induced type-1 diabetes (T1D). We report that frontocortical CB₁R protein density was biphasically altered in the first month of T1D, which was accompanied with a reduction of resting glucose uptake ex vivo in acute frontocortical slices that was normalized after eight weeks in T1D. This early reduction of glucose uptake in slices was also restored by ex vivo treatment with both the non-selective CB₁R agonists, WIN55212–2 (500 nM) and the CB₁R-selective agonist, ACEA (3 µM) while it was exacerbated by the CB₁R-selective antagonist, O-2050 (500 nM). These results suggest a gain-of-function for the cerebrocortical CB₁Rs in the control of glucose uptake in diabetes. Although insulin and IGF-1 receptor protein densities remained unaffected, phosphorylated GSKα and GSKβ levels showed different profiles 2 and 8 weeks after T1D induction in the frontal cortex. Altogether, the biphasic response in frontocortical CB₁R density within a month after T1D induction resolves previous controversial reports on forebrain CB₁R levels in T1D rodent models. Furthermore, this study also hints that cannabinoids may be useful to alleviate impaired glucoregulation in the diabetic cortex.

1. Introduction

Marihuana has long been known to modulate carbohydrate metabolism in animal models (de Pasquale et a., 1978) and man (Benowitz et al., 1976). One of the major targets of marihuana’s Δ⁹-THC in the body is the cannabinoid CB₁ receptor (CB₁R), the foremost metabotropic receptor of the endocannabinoid system, and is widely expressed in most mammalian organs and tissues, including the CNS, the pancreas, the liver and the skeletal muscle (Katona and Freund, 2012; Solymosi and Köfalvi, 2017). The CB₁R is a major regulator of the body’s energy homeostasis, via three major targets: neural circuits, neurohumoral communication and cellular metabolism (Piazza et al., 2017; Ruiz de Azua et al., 2019).

Overactivation of the endocannabinoid system may lead to systemic insulin resistance (Bowles et al., 2015; Jourdan et al., 2017; Sidibeh et al., 2017), which is a primary pathomechanism of diabetes. One theory suggests that the CB₁R inhibits insulin receptor signaling via physical complexing, that is, heteromerization (Dalton and Howlett, 2012; Kim et al., 2012). Indeed, we found in the rat nucleus accumbens that the CB₁R co-immunoprecipitates with the insulin receptor, and its activation prevents insulin from stimulation of [³H]deoxyglucose uptake (Pinheiro et al., 2016). Interestingly, the interaction between insulin and endocannabinoids is even more intricate. For instance, insulin stimulates endocannabinoid release in the ventral tegmental area (Labuèbe et al., 2013), while the CB₁R stimulates insulin secretion in the endocrine pancreas (Bermúdez-Silva et al., 2016). This heralds the possibility that diabetes could be associated with alterations in CB₁R signaling.

Our particular area of interest is the cerebral cortex because of the widely accepted but little understood link between systemic metabolic diseases and neurodegeneration (Duarte, 2015; Rehni et al., 2018). It has been speculated that cerebral dysmetabolism can contribute to the onset of brain disorders (de Ceballos and Köfalvi, 2017; Zilberter and Zilberter, 2017), prompting the interest to identify new pharmacological targets to mitigate brain dysmetabolism. Recently we reported that untreated type-1 diabetes (T1D) triggered a reduction both in CB₁R density and in glucose uptake in hippocampal and frontocortical slices of mice. Furthermore, the genetic deletion of the CB₁R also caused similar glucose dysmetabolism, but strikingly, no further impairment was observed in the diabetic CB₁R knockout animals (Moura et al., 2019). This strongly suggests that CB₁R dysfunction may mediate some deleterious effects of T1D on brain glucose metabolism.

We now used ex vivo frontocortical tissue from control and diabetic rats to compare glucose metabolism, CB₁R density, and the effect of CB₁R ligands at 2, 4 and 8 weeks after a single i.p. injection with streptozotocin (STZ) or vehicle. We found that all the investigated parameters significantly change according to the time-course of the disease and that acute ex vivo CB₁R stimulation may prevent early dysregulation of glucose metabolism in the frontocortical slices.

2. Materials and Methods

2.1. Animals and diabetes models

All studies were conducted in accordance with the principles and procedures outlined as “3Rs” in EU guidelines (86/609/EEC), FELASA, and the National Centre for the 3Rs (the ARRIVE; Kilkenny et al., 2010), and were approved by the Animal Care Committee of the Center for Neuroscience and Cell Biology of the University of Coimbra, Portugal. We also applied the ARRIVE guidelines for the design and execution of in vitro experiments (see below), as well as for data management and interpretation (McGrath et al., 2010).

Eighty two male Wistar Han rats (10-week-old) were purchased from Charles River (France) and housed in a temperature and humidity-controlled environment with 12 h light on/off cycles and ad libitum access to food and water. All efforts were made to minimize the number of animals used and to minimize their stress and discomfort.

Fifty two rats were randomly assigned to sham- and STZ-injected groups. These groups were further randomly stratified into 3 subgroups, which were kept alive for 2, 4 and 8 weeks post-injection. At 12 weeks of age, type-1 diabetes mellitus (T1D) was induced with one single intraperitoneal (i.p.) injection of 60 mg/kg streptozotocin (STZ) (Szkudelski, 2001), after four hours of food deprivation. STZ was kept at −20°C until use, and its solution was freshly prepared in citrate buffer (10 mM; pH 4.5) less than 10 min before i.p. injection. Control rats were injected with the vehicle, citrate buffer, and maintained under the same conditions as the treated ones. Both body weight and blood glucose levels were measured both before and 3 days after STZ injection as well as on the day of sacrifice (i.e. 2, 4 and 8 weeks post-STZ injection) from tail vein blood with the glucose oxidase method, using a glucometer and reactive test stripes (Elite-Bayer SA, Portugal) (SFig. 1). Rats with blood glucose levels higher than 250 mg/dL were considered diabetic (de Morais et al., 2016).

We also had access to freshly isolated cortex for Western blotting and real-time PCR from 5-month-old male Goto-Kakizaki rats (Taconic, Lille Skensved, Denmark) – a model for type 2 diabetes mellitus (T2D), which spontaneously develops insulin resistance without obesity. For control, we used age-matched Wistar-Hannover-Galas rats (also from Taconic).

2.2. In vitro tandem [³H]DG and [¹⁴C]-glucose uptake in brain slices

This assay was carried out now on frontocortical slices as before on hippocampal and accumbal slices (Lemos et al., 2012; Pinheiro et al., 2016), with slight modifications after optimization. Sections 1.2 and 2.1 of the Supplemental File contain the detailed description of the optimization process of the in vitro tandem [³H]DG/[³H]MG and [¹⁴C]-glucose uptake in cerebrocortical slices. The optimization process determined that [³H]DG rather than [³H]MG is the suitable “stable” analogue for a 30-min-long glucose uptake assay in cortical slices, after 60 min recovery. The following optimized methods correspond to the experiments presented here under Sections 3.1. and 3.2. of the Results.

Before decapitation with a stainless steel guillotine, the rats were deeply anesthetized with 2-bromo-2-chloro-1,1,1-trifluoroethane (halothane; 5%, 1 L/min flow rate in an induction chamber), thus showed no reaction to tail pinch and handling while still breathing. The brains were quickly removed and the cortices were dissected in ice-cold Krebs-HEPES assay medium of the following composition (in mM): NaCl 132, KCl 3, KH₂PO₄ 1.2, MgSO₄ 1.2, CaCl₂ 2.5, NaHCO₃ 25, glucose 3, HEPES 10 (pH 7.4). Subsequently, 400 µm-thick coronal cortical slices were cut with the help of a McIllvain tissue chopper (the Mickle Laboratory Engineering Co. Ltd), then the slices were collected in a 50 mL pregassed (by 5% CO₂, 95% O₂) and prewarmed (to 37 °C) Krebs-HEPES assay medium (aka. recovery bath). Each holding chamber contained six submerged baskets, which are 15 mm tall and 10 mm wide, and have a 80 µm-pore nylon mesh bottom. This setup allows freely transferring and batch incubating slices up to ~5 mg protein/well in a 50 mL bath.

The pool of slices from each rat was divided into four holding chambers. In the same experiment, frontocortical slices from both sham and diabetic rats were co-incubated in the same holding chambers (although, in separate baskets) to increase statistical power. After 55 min incubation (recovery period) under gentle gassing at 37 °C, chamber 3 received the synthetic cannabinoid agonist, WIN55212–2 (final concentration: 500 nM) or when noted, the synthetic CB₁R-selective agonist, ACEA (3 µM), chamber 4 received the CB₁R-selective neutral antagonist, O-2050 (500 nM), and chambers 1 and 2 received the vehicle DMSO (0.1%). The concentrations of the cannabinoids were selected based on a previous review (Solymosi and Köfalvi, 2017). Four min later, chamber 2 received 263 µL of a 3 M KCl solution to create the depolarized condition (final K⁺ concentration: 20 mM), while the rest of the chambers received an equivalent amount of NaCl serving as osmotic control. One min later, [³H]DG (1 nM) and [¹⁴C]₆-glucose (50 nM) were bath-applied in the chambers (for more information on the radiochemicals, see Supplemental File). Thirty min later, the baskets with the slices were transferred to large volumes of ice-cold assay medium for extensive washing, and then the slices were removed to 1 mL of NaOH (0.5 M) to determine the uptake of the tracers and protein quantities. Original bath samples were also counted to measure the precise concentrations of the two radiotracers. For calculations, see Supplemental File.

2.3. Western blotting

The remaining frontocortical slices from each rat were used for total membrane extraction: The tissue was homogenized at 4 °C in sucrose (0.32 M) – HEPES (15 mM) buffer (pH: 7.4). The homogenate was centrifuged at 3,000 g for 10 min at 4 °C. The supernatant was then resuspended in a solution of 50 mM Tris and 10 mM MgCl₂ (pH 7.4), centrifuged at 28,000 g for 20 min at 4 °C. From the resulting pellet, proteins were extracted with RIPA buffer (50 mM Tris/HCl [pH 8.0], 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 2 mM EDTA, proteases inhibitor cocktail, phosphatase inhibitor cocktail and 1 mM dithiothreitol) and these protein lysates were denatured at 95 °C, for 5 min, in sample buffer (0.125 mM Tris [pH 6.8], 2% w/v SDS, 100 mM dithiothreitol, 10% glycerol and bromophenol blue). Thirty µg of total protein was resolved on 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% (w/v) fat-free dry milk in Tris-buffered saline containing 0.1% (v/v) Tween 20, for 1 h at room temperature.

After blocking and washing, membranes were incubated overnight at 4°C with the primary antibodies against the different proteins studied (for their suppliers, see 2.6. Materials): IRβ (1:1000), IGF-1R (1:1000), p-GSK3α (1:1000), GSK3α (1:1000), p-GSK3β (1:1000), GSK3β (1:1000), CB₁R (1:1000), and β-actin (1:1000). After incubation, membranes were washed and incubated for 1 h at room temperature, with alkaline phosphatase-conjugated anti-rabbit, anti-guinea pig or anti-mouse antibody (Santa Cruz Biotechnology; all at 1:5000). The membranes were exposed to ECF reagent followed by scanning for blue excited fluorescence on the VersaDoc (Bio-Rad Laboratories). The generated signals were analyzed using the Image-Quant TL software, and the background for each individual band was sampled adjacent to each bar. Density values were normalized to β-actin and these normalized values of the diabetic rats were again normalized to each WT sham normalized values in the same membrane.

2.4. Real-time PCR

Rat cortex samples were freshly isolated and immediately frozen in liquid nitrogen. Total RNA was extracted from these tissue samples with MagNA Lyser Instrument and MagNA Pure Compact RNA Isolation kit (Roche, Portugal) according to the manufacturer’s instructions. Reverse transcription for first-strand cDNA synthesis from each sample was performed using random hexamer primer with the Transcriptor First Strand cDNA Synthesis kit (Roche) according to manufacturer’s instructions. Resulting cDNAs were used as template for real-time PCR, which was carried out on LightCycler instrument using the FastStart DNA Master SYBR Green I kit (Roche) and the primers (Tib MolBiol, Germany) listed in Table 1.

Table 1.

Primers used in the real-time PCR analysis.

	Accession number	Primer sequence	Expected product size (bp)
CB1R (Hansson et al., 2007)	NM_012784	F 5’- AGA CCT CCT CTA CGT GGG CTC G −3’  R 5’- GTA CAG CGA TGG CCA GCT GCT G −3’	314
β-actin (Peinnequin et al., 2004)	V01217	F 5’- AAG TCC CTC ACC CTC CCA AAA G −3’  R 5’- AAG CAA TGC TGT CAC CTT CCC −3’	97

Quantification of mRNA in the samples was carried out on the basis of standard curves run simultaneously. The cDNA standards for the calibration curve were generated by conventional PCR amplification (Table 2). The PCR products were run in a 3% agarose gel electrophoresis to verify fragment size and the absence of other contaminant fragments, quantified by absorbance at 260 nm, and serially diluted to produce the standard curve (100 to 108 copies/µL). Each real-time PCR reaction was run in triplicate and contained 2 µL of cDNA template, 0.3 µM of each primer, and 3 or 3.5 mM MgCl₂ (see Table 2), in a reaction volume of 20 µL. Cycling parameters are described in Table 2. The analysis of melting curves ensured that only a single product was amplified. The expression of mRNA was calculated relative to β-actin mRNA expression.

Table 2.

Cycling parameters for real-time and conventional PCR analysis.

Real-time PCR
	MgCl2 (mM)	Initial denaturation	Amplification				Melting curve analysis
			Denaturation	Annealing	Extension	Cycles
CB1R	3	95 °C, 10 m	95 °C, 10 s	62 °C, 10 s	72 °C, 14 s	40
β-actin	3.5	95 °C, 10 m	95 °C, 10 s	61 °C, 5 s	72 °C, 4 s	45
Conventional PCR
	MgCl2 (mM)	Initial denaturation	Amplification				Elongation
			Denaturation	Annealing	Extension	Cycles
CB1R	1.5	95 °C, 5 m	95 °C, 1 m	59 °C, 1 m	72 °C, 1 m	45	72 °C, 4 m
β-actin	1.5	95 °C, 5 m	95 °C, 1 m	56 °C, 1 m	72 °C, 1 m	45	72 °C, 4 m

2.5. Data handling

For data presentation and manipulation, we followed the rules of Curtis et al. (2015). All data represent the mean and S.E.M. of n ≥ 5 individual observations (animals). Means were tested for normality with the Kolmogorov-Smirnov test. Statistical significance was determined by one-way ANOVA of repeated measures followed by Bonferroni’s post-hoc test or two-tailed Student’s t-test (except where noted otherwise), and P < 0.05 was accepted for significant difference, according to Curtis et al. (2015). Tests were performed and figures were prepared using the GraphPad Prism 5.0 software package.

2.6. Materials

The antibodies against phospho-(p)-(Ser21) glycogen synthase kinase α (GSK3α), total and p(Ser9)-GSK3β, total and p(Ser473)-Akt, were purchased from Cell Signaling Technologies (Danvers, MA, USA). The antibodies against the β chain of the insulin receptor (IR) and the insulin-like growth factor receptor (IGF-1R) were purchased from Santa Cruz Biotechnology (Santa Cruz, California, USA). The antibody against total GSK3α was obtained from UpState (Lake Placid, NY) and the antibody against β-actin was bought from BioLegend (San Diego, CA). The guinea-pig anti-CB₁R antibody was obtained from Frontier Institute co., ltd. (Hokkaido, Japan), and we have tested its selectivity before (Bitencourt et al., 2015). The chemifluorescence (ECF) reagent was acquired from GE Healthcare (Chalfont St. Giles, UK). 2-deoxy-2-(3-(methyl-3-nitrosoureido)-D-glucopyranose (streptozotocin or STZ) and other unspecified (in)organic reagents were from Merck (formerly, Calbiochem, Sigma-Aldrich, Merck-Millipore Corporation; Darmstadt, Germany).

3. Results

3.1. Glucose uptake in frontocortical slices of diabetic rats

The subjects of the following assay are rats in which diabetes was elicited with 1 single i.p. injection of STZ at 12 weeks of age, together with their sham-injected controls. SFig. 1 documents that the STZ-injected rats became in fact diabetic. The sham rats showed steadily increasing body weight, while the STZ-injected rats showed a reverse tendency. Blood glucose levels were highly elevated in the diabetic animals.

The following assay was carried out in pairwise manner, that is, randomly chosen slices of one sham and one diabetic animal were simultaneously incubated in four common holding chambers (besides an extra chamber on ice to measure external non-specific [³H] and [¹⁴C] labeling). The age of the sham animals (i.e. the fact that they were sacrificed at 14, 16 and 20 weeks of age) did not significantly affect glucose uptake under either resting conditions or high-K⁺ depolarization (Fig. 1), although there was a nearly significant tendency (P = 0.05–0.1) for higher glucose uptake under high-K⁺-depolarization in slices from diabetic rats, relative to their own control (Fig. 1D). The rationale behind evaluating resting and depolarized glucose consumption is to compare two extreme conditions. Under prolonged depolarization, the brain slices are submitted to an extreme energetic stress similar to an epileptic activity in vivo which can unveil additional impairments. In contrast, the basal uptake probably correlates better with what is seen in human subjects resting under a [¹⁸F]-FDG-PET scan, even if this is speculative. In the sham animals, only depolarization affected significantly glucose uptake, while the synthetic CB₁R ligands, – as we previously published in the hippocampus (e.g. Lemos et al., 2012) were devoid of effect on glucose uptake in the resting slice.

Fig. 1.

STZ-induced insulinopenia impairs resting glucose uptake in a CB₁R-dependent fashion. (A) Two-weeks after T1D induction, resting glucose uptake was significantly lower than in the sham animals (n=14; P < 0.01), which was further decreased by CB₁R blockade with the neutral CB₁R antagonist, O-2050 (500 nM), although non-significantly (P = 0.058 vs. STZ control). This reduction was recovered to the sham control value in the presence of the synthetic cannabinoid, WIN55212–2 (500 nM) (P > 0.05 vs. sham control). (B) After 4 weeks with T1D, resting glucose uptake was no longer different from the sham value (P > 0.05), and hence, WIN55212–2 failed to stimulate glucose uptake further, but O-2050 uncovered a significant reduction as compared to either the sham or the respective diabetic DMSO control value (n=8; P < 0.05). (C) After 8 weeks with T1D, the metabolic disturbances fully recovered, and neither the activation nor the blockade of the CB₁R affected glucose uptake (n=8; P > 0.05). (D) The difference in the amplitudes of high K⁺-stimulated glucose uptake normalized to the appropriate sham or STZ control was tendentiously at the border of significance between sham and diabetic animals. Note that all four chambers had identical osmotic condition (achieved with extra NaCl or KCl) and 0.1% DMSO as a vehicle. All bars and symbols represent mean + or ± S.E.M.; *P < 0.05, **P < 0.01, ***P < 0.001 vs. the respective control (sham or STZ), and ^$P < 0.05, ^$ $P < 0.01 vs. sham control from the same experiment in pairwise arrangement, as assessed with Repeated Measures ANOVA following by Bonferroni’s post-hoc test for selected pairs of data sets.

Two weeks post-STZ injection, rat cortical slices exhibited a 15.2 ± 3.9% decrease in resting glucose uptake (diabetic DMSO control; n = 14, P < 0.01; as assessed with repeated measures ANOVA on the raw data) (Fig. 1A), which was normalized by treatment of the slices with the non-selective cannabinoid agonist, WIN55212–2 (500 nM; P > 0.05 vs. DMSO). To certify that the effect of WIN55212–2 is CB₁R-dependent, we have also tested the CB₁R-selective agonist, ACEA (3 µM) (Hillard et al., 1999) in a new batch of 2 × 6 rats. ACEA had no significant effect on ex vivo cortical glucose uptake in the sham-injected rats (95.2 ± 11.6% of sham DMSO control, n = 6, P > 0.05 with repeated measures ANOVA) (figure not shown). Two weeks after STZ injection, ex vivo glucose uptake in the diabetic cortex amounted to 83.2 ± 6.8% of sham DMSO control (P < 0.05), which was mitigated by ACEA (93.2 ± 9.8% of sham DMSO control, n = 6, P > 0.05) (figure not shown). Additionally, the CB₁R-selective neutral antagonist, O-2050 (500 nM) exacerbated the impairment of glucose uptake to −24.8 ± 7.3% compared to sham DMSO (P < 0.05), and this further decrease was almost significantly different from the respective STZ DMSO control (P = 0.06) (Fig. 1A). Four weeks post-STZ injection, glucose uptake was no longer statistically different from the sham DMSO value, and thus, WIN55212–2 also failed to further modulate glucose uptake in the STZ group (Fig. 1B). Intriguingly, acute CB₁R blockade significantly reduced glucose uptake by 22.7 ± 6.8% (n = 6, P < 0.05 vs. both sham and STZ DMSO controls). Altogether, these data suggest that the CB₁R gains a transient role in cortical glucoregulation under insulinopenia. When the slices from the T1D rats were challenged with high K⁺ 2 and 4 weeks post-injection, they tended to take up almost significantly more glucose than the sham slices (Fig. 1D), but this tendency was lost after 8 weeks in T1D. In fact, 8 weeks post-STZ-injection, neither basal nor high-K⁺-stimulated glucose uptake was different from the sham slices, and neither WIN55212–2 nor O-2050 (Fig. 1C), nor ACEA (3 µM; n = 6; figure not shown) affected glucose uptake in the two cohorts.

3.2. Dissipative glucose metabolism in frontocortical slices from diabetic rats

In the above set of experiments, we simultaneously measured the accumulation of [³H]DG and the [¹⁴C] label from [¹⁴C]₆-glucose. Since the latter tracer is metabolized much faster than the former, the difference in [³H] vs. [¹⁴C] contents roughly measures the dissipative metabolism of glucose in the slice (for details see Lemos et al. (2012) and Supplemental File). Dissipative metabolism was not statistically different among the three groups of sham animals. As expected, high-K⁺-depolarization strongly augmented glucose metabolism in all groups of animals as gauged by the strong reduction in [¹⁴C] content as compared to the [³H] content, and there was a tendency too for a greater metabolic rate in the diabetic rats (P > 0.05) (Fig. 2). Furthermore, CB₁R activation and blockade had no significant effect on the metabolism of glucose in the slices from the 6 groups of rats (Fig. 2), although WIN55212–2 almost significantly stimulated glucose metabolism after 8 weeks in diabetes (Fig. 2C; P > 0.05).

Fig. 2.

Dissipative glucose metabolism is not affected by chronic insulinopenia or acutely by CB₁R ligands or the two combined in the rat frontal cortex. (A) Two-weeks, (B) four weeks and (C) eight weeks after T1D induction, resting glucose uptake was similar between the two cohorts (n=14, 8 and 8, respectively; P < 0.05), and only responded significantly to high-K⁺ treatment (P < 0.001). (D) The amplitude of response to depolarization was also similar between the two groups, although there was a non-significant tendency (n=14, 8 and 8, respectively; P > 0.05) for greater [¹⁴C] release in the slices prepared from the diabetic rats. All bars and symbols represent mean +/± S.E.M.; ***P < 0.001 vs. the respective sham or STZ control, from the same experiment in pairwise arrangement, as assessed with Repeated Measures ANOVA following by Bonferroni’s post-hoc test for selected pairs of data sets.

These data suggest that chronic insulinopenia affects basal glucose uptake in the frontal cortex. We then asked if resting and high-K⁺-depolarization-evoked glucose uptake and metabolism are subject to modulation by either insulin or cannabinoids, but more importantly, by the combination of insulin with cannabinoid agonists and antagonist, in cortical slices of ad libitum fed, 16-hour fasted (i.e. acutely insulinopenic) and in T1D rats. The outcome of that assay is presented in the Supplemental File (Section 2.2. and Supplemental Table 1), as we judged that those data did not add groundbreaking information to the present study.

3.3. CB₁R density and expression in the diabetic cortex

Our above functional data suggest a gain-of-function for the CB₁R in the early phase of T1D. Previous studies assessing CB₁R density in the brain of diabetic rats and mice have found contrasting results, depending on the duration of T1D after STZ-injection until sacrifice and the brain area in question (Duarte et al., 2007; Díaz-Asensio et al., 2008; de Morais et al., 2016; Moura et al., 2019). We now evaluated CB₁R immunoreactivity by Western blotting, analyzing the selective band at ~52 kDa (Bitencourt et al., 2015) in prefrontocortical homogenates 2, 4 and 8 weeks after STZ injection. We found a significant increase (24.5 ± 9.0%) in CB₁R density after 2 weeks of diabetes (n = 6, P < 0.05; Fig. 3AB) as compared to the respective sham animals. Two weeks later, this increase in CB₁R density was surprisingly inverted into an overt 30.5 ± 5.8% reduction (n = 6, P < 0.01; Fig. 3AB). By the end of the 8-week period with diabetes, there was no significant alteration in CB₁R density (n = 6, P > 0.05; Fig. 3AB). We also had access to frontocortical tissue of eight 5-month-old type-2 diabetic Goto-Kakizaki (GK) rats and ten age-matched controls, in which we observed a significant, on average 15.0% reduction in CB₁R density at 20 and 30 µg protein loads (P < 0.05; Fig. 3C). Finally, we divided our remaining frontocortical tissues from all the cohorts so that the material from the 2-week- and 8-week-diabetic animals was further analyzed by Western-blotting, while the 4-week-diabetic as well as the GK rat tissue was subjected to PCR analysis for CB₁R expression. The real-time PCR analysis normalized to the constitutive expression of β-actin mRNA showed that CB₁R mRNA expression was not statistically different either in the 4-week-diabetic or in the GK rats, as compared to their respective controls (P > 0.05; Fig. 3D).

Fig. 3.

The cortical density rather than the expression of the CB₁R is affected by T1D and T2D. (A) representative blots showing the 53 kDa CB₁R immunoreactivity and the 42 kDa β-actin immunoreactivity at 30 µg protein load, in diabetic rats at 2, 4 and 8 weeks after STZ-injection, and in the corresponding sham-injected animals. Panel (B) shows the mean ± S.E.M. of CB₁R immunoreactivities normalized to β-actin (CB₁R densities), and expressed to the same normalized data obtained in the respective sham animals. (C) CB₁R densities (mean ± S.E.M) are significantly lower in the type-2 diabetic GK rats (n = 10) than in their controls (n = 8). (D) Bar graphs summarizing average CB₁R expression + S.E.M. in the 16 week-old sham and T1D rats as well as in the 20-week old control and GK rats. *P < 0.05 and **P < 0.01 vs. the respective sham control; n.s., not significant.

3.4. Insulin receptor and IGF-1 receptor densities and patterns of GSK3 phosphorylation

Insulin receptor and IGF-1 receptor immunoreactivities at the expected 95 kDa molecular weight remained unaffected by 2 or 8 weeks with T1D (Fig. 4AB). Glycogen synthase kinase 3α and β isoforms are downstream targets of both the CB₁R (Solymosi and Köfalvi, 2017) and the insulin/IGF-1 receptors (Beurel et al., 2015). We now measured the phospho-Ser21 vs. total GSK3α immunoreactivity at 51 kDa and the phospho-Ser9 vs. total GSK3β immunoreactivity at 47 kDa and observed that the former decreased by 17.4 ± 4.1% (n = 6, P < 0.01) 2 weeks after STZ injection, but it was no longer significantly different from sham levels after 8 weeks with diabetes (Fig. 4CD), while GSK3β phosphorylation was initially at the sham level, but at the end of the 8-week period, it significantly increased by 19.5 ± 7.2% (n = 6, P < 0.05; Fig. 4CD).

Fig. 4.

STZ-induced diabetes alters the activation of the insulin signaling pathway while leaving insulin and IGF-1R densities unaffected. (A,C) Representative blots and (B,D) the respective bar graphs showing that insulin and IGF-1 receptor densities remain unaffected after 2 and 8 weeks with T1D (as normalized to the β-actin levels). Furthermore, 2 weeks with diabetes decreases phospho(Ser21)GSK3α density while leaving phospho(Ser9)GSK3β density unaffected, but after 8 weeks with T1D, the former returns to control why the latter significantly increases. All bars represent mean + S.E.M of 6 rats; *P < 0.05, **P < 0.01 vs. sham ratios taken as 100%.

4. Discussion

We here report a transient reduction of basal glucose uptake and its recovery in a CB₁R-dependent fashion, in the frontal cortex of rats in the first weeks after T1D induction. We also found biphasic changes in CB₁R receptor protein levels and GSK3 signaling. These changes appear counter-regulatory and sequential. Namely, after 2 weeks with T1D, the CB₁R showed an increased density and a gain-of-function, as inferred from the stimulatory effect of the cannabinoid agonist, WIN55212–2. After 4 weeks with T1D, the reduced CB₁R density seemed to be compensated by increased endocannabinoid signaling to recover basal glucose uptake rates, as deduced from the inhibitory effect of the CB₁R-selective neutral antagonist which was unseen in sham rats. Notably, this is the first report of an involvement of CB₁Rs in glucose uptake in acute brain slices, linked to our previous observation that CB₁Rs inhibit mitochondrial intermediary metabolism in astrocytes and neurons (Duarte et al., 2012). Finally, by the end of the 8^th week with T1D, the endocannabinoid-CB₁R axis was no longer involved in glucose uptake, and CB₁R levels returned to normal.

One may note that WIN55212–2 is a non-selective agonist of both the cannabinoid CB₁ and CB₂ receptors (Solymosi and Köfalvi, 2017). Even though the CB₂R has been occasionally shown to modulate neuronal functions (e.g. Andó et al., 2012), the CB₁R is believed to have predominant roles in marihuana’s psychoactivity as well as in brain physiology and pathology (Katona and Freund, 2012; Araque et al., 2017; Fernández-Ruiz 2019). The use of the CB₁R-selective agonist, ACEA and the CB₁R-selective antagonist, O-2050 also modulated glucose uptake, underpinning the involvement of the CB₁R in our model.

Although insulin and IGF-1R levels remained constant at least 2 and 8 weeks after T1D induction, phospho-GSK3α levels were lower at the first time-point. We previously reported that CB₁R-selective ligands control GSK3α phosphorylation in acute hippocampal slices (Lemos et al., 2012). Here we found that two weeks after T1D induction, GSK3α phosphorylation at Ser21 was significantly decreased, which could have contributed to a compensatory increase in CB₁R density. In contrast, in the above paper we found no direct control by cannabinoids on GSK3β phosphorylation at Ser9, and this is in agreement with the present findings: while GSK3α phosphorylation and CB₁R functioning both became “normal” again after 8 weeks with T1D, phospho-(Ser9)-GSKβ levels were independently increased. Clearly, the dynamic interactions among CB₁R, insulin and their common downstream targets are complex and the present data only provides evidence for their involvement in the control of glucose metabolism in the frontal cortex at the early phase of T1D.

While most complications in T1D patients arise from the recurrent insulin administration as well as from unawareness of hypoglycemia rather than from chronic insulinopenia and hyperglycemia, even in the absence of microvascular complications, lower spontaneous activity is detected in the default mode network, especially in the frontal cortex of T1D patients (Xia et al., 2018). Brain glucose uptake is also smaller in both T1D patients and STZ-injected diabetic rats under hyperglycemic clamp, though there was no alteration of intermediary glucose metabolism in rats (Hwang et al., 2018). Among T1D patients, hypoglycemia-unaware patients exhibit further compromised cerebral glucose uptake at euglycemia and hypoglycemia (Cranston et al., 2001) despite the preserved transport of glucose through the blood-brain barrier (Duarte, 2015). Taken that impaired cerebral energy metabolism is associated with accelerated brain aging (de Ceballos and Köfalvi, 2017; Zilberter and Zilberter, 2017), it is evident that these complications of T1D have deleterious effects on brain microstructure (Yoon et al., 2018), leading to cognitive decline (Ryan et al., 2016).

Similarly to human T1D patients but to a much greater extent, rodent models also exhibit neuroglycopenia up to −50% of control in brain areas such as the frontal cortex, without significant alterations in cerebral blood flow (Jakobsen et al., 1987; Mooradian and Morin, 1991). Since we could recapitulate the reduced glucose uptake in resting slices in the present study, and in both resting and high-K⁺-stimulated brain slices of diabetic mice, we can safely conclude that glucose uptake is impaired in brain cells in T1D. This cerebral dysmetabolism in rodent T1D models is associated with impaired acquisition and retention of memory, impaired spatial, working and reference memories, depressive behaviour and cognitive deficits (Baydas et al., 2003; de Morais et al., 2016; Lin et al., 2018), and at the cellular and molecular level, with monoaminergic dysbalance (de Morais et al., 2016; Lin et al., 2018), altered neural cell adhesion molecule expression (Baydas et al., 2003), hampered hippocampal neurogenesis, impaired synaptic plasticity, neuroinflammation, tau hyperphosphorylation and oxidative stress (Stranahan et al., 2008; Duarte, 2015; de Morais et al., 2016; Elahi et al., 2016). Whether apoptosis occurs in the brain in response to systemic STZ injection is under debate (e.g. Guven et al., 2009; vs. Hao et al., 2019), but transient changes in neural proliferation and apoptosis, which are tightly controlled by the CB₁R (Rodrigues et al., 2019), could certainly contribute to dysmetabolism.

Insulin and IGF-1 receptors are widely distributed in the mammalian brain (Duarte et al., 2012). Cerebral insulin receptors play a major role in the regulation of the energy metabolism of the whole body (Brüning et al., 2000; Varela and Horvath, 2012). Insulin receptor activation can also stimulate local and global rates of cerebral glucose metabolism, but only a few studies have documented these findings in vivo in humans (Bingham et al., 2002) and rats (McNay et al., 2010). In acute cortical slices, we could not demonstrate an impact of insulin in glucose uptake and metabolism (see Supplemental Table 1), which is in accordance with a previous study in cortical slices (Abdul-Ghani et al., 2007), but is in contrast to studies using neuronal and astrocytic cell cultures (Clarke et al.,1984; Werner et al., 1989; Kum et al., 1992; Benomar et al., 2006). Knowing that insulin is both de novo synthesized and taken up from the circulation in the brain (Santos et al., 1999; Banks, 2004; Molnár et al., 2014), it is possible that a fraction of basal glucose uptake in the acute brain slices (but not in cell cultures) comes from intrinsic insulin action, which would occlude or mask effects of exogenously added insulin.

If insulin indeed stimulates cerebral glucose uptake, it could explain the consistent reports of neuroglycopenia in both T1D animal models and T1D patients, as one would expect reduced insulin levels in the brain parenchyma under T1D. But somewhat counterintuitively, the brain increases insulin production in T1D (Havrankova et al., 1978). This could probably be a key reason for the lack of changes in cerebral insulin (and IGF-1) receptor densities seen in T1D models (Havrankova et al., 1978; Packold et al., 1979). Therefore, it is less likely that an impaired insulin signaling causes cerebral hypometabolism in T1D. Instead, our previous paper (Moura et al., 2019) reveals that the T1D CB₁R KO mice had no further impairment in glucose uptake and metabolism as compared to their healthy controls. This suggests that an impairment in CB₁R signaling is downstream to insufficient insulin signaling in the cascade leading to neuroglycopenia.

The negative role of cerebral and peripheral CB₁Rs in the genesis of metabolic syndrome, insulin resistance, obesity and T2D is well-established (Di Marzo et al., 2011; Piazza et al., 2017; Ruiz de Azua et al., 2019). Yet, surprisingly little is known about the involvement of the endocannabinoid system in the development of T1D. Interestingly, a landmark study suggests that the overactivation of CB₁Rs intrinsic to β-cells can heteromerize and inhibit local insulin receptors, leading to insulin resistance, apoptosis and consequently, T1D (Kim et al., 2012). Previous studies already addressed the effect of STZ-induced diabetes on cerebral CB₁R expression/density in rats with various outcomes. For instance, Díaz-Asensio et al. (2008) reported that 4 weeks after injecting intravenously 6–8-week-old Sprague–Dawley rats with STZ, CB₁R density remained unaffected in the cortex and hippocampus, but increased in the striatum and the hypothalamus. In 12-week-old Wistar rats, 4-week after STZ injection, CB₁R density increased in the hippocampus, while CB₁R expression, i.e. mRNA levels, decreased, suggesting either an accelerated translation or a post-translational modification or both (Duarte et al., 2007).

Notably, the findings by de Morais and colleagues (2016) are similar to our previous and present findings. They found that 4 weeks after an intraperitoneal STZ injection, hippocampal CB₁R density significantly increased in young adult Wistar rats, while in the prefrontal cortex, it became ~30% smaller (although P > 0.05). It is known that chronic hyperglycemia is sufficient to downregulate neuronal CB₁R expression in neurons (Zhang et al., 2007), but there is certainly more to it, since our diabetic rats had lower CB₁R densities only once out of the three time-points. Perhaps changes in insulin signaling, high extra- and low intracellular glucose levels, glycosylation, neuroinflammation, apoptosis and gliosis all contribute to oscillating (sub)cellular changes in CB₁R levels in T1D.

5. Concluding remarks

We previously reported in mice (Moura et al., 2019) and now we found in the rat that the CB₁R is involved in cerebral hypometabolism in T1D. The present data suggest that targeting cerebral CB₁Rs in T1D could be an interesting strategy to correct neuroglycopenia. We previously proposed a set of testable hypotheses about the mechanisms underlying the influence of CB₁Rs on hippocampal and cortical glucose uptake (Moura et al., 2019). It would require a sizable effort to explore these hypotheses in animal models, including the in vivo injections of the diabetic animals with cannabinoids both systemically and locally in the brain, which is certainly necessary to elucidate the intricate interaction among insulin and endocannabinoid signaling in brain energy metabolism. But even without knowing the exact molecular mechanisms, preclinical tests targeting the CB₁R have already been prompted by a handful of animal studies to alleviate peripheral and central symptoms of T1D (Weiss et al., 2006, 2008; Barutta et al., 2010; Vera et al., 2012).

Highlights.

• Frontocortical glucose uptake is smaller early after type-1 diabetes (T1D) induction

• The recovery from neuroglycopenia in T1D is cannabinoid CB₁ receptor-dependent

• T1D induction is followed by a biphasic change in CB₁R levels and insulin signaling

Abbreviations:

[³H]DG

[³H]-2-deoxy-D-glucose

2-AG

2-arachidonoyl-glycerol

3Rs

Replacement, Refinement and Reduction of Animals in Research

ACEA

arachidonyl-2’-chloroethylamide

ARRIVE

Animals in Research: Reporting In Vivo Experiments

CB₁R(s) and CB₂R(s)

cannabinoid CB₁ and CB₂ receptor(s)

FELASA

Federation for Laboratory Animal Science Associations

GSKα/β

glycogen synthase kinase α/β

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

IGF-1R

insulin-like growth factor-1 receptor(s)

i.p.

intraperitoneal

[³H]MG

[³H]-3-O-methyl-glucose

RIPA

radioimmunoprecipitation assay

SDS-PAGE

sodium dodecyl sulphate-polyacrylamide gel electrophoresis

STZ

streptozotocin

T1/2D

type-1/2 diabetes