Chotosan ameliorates cognitive and emotional deficits in an animal model of type 2 diabetes: possible involvement of cholinergic and VEGF/PDGF mechanisms in the brain

Male 6-week-old C57BLKS/J-db/db mice and their age-matched non-diabetic m/m littermates were purchased from Japan SLC Inc. (Hamamatsu, Japan). The animals were housed in groups of 6 – 9 mice/cage (24 × 17 × 12 cm) in a laboratory animal room, which are maintained at 25 ± 1°C with 65 ± 5% humidity, on a 12 h light/dark cycle (lights on: 07:30-19:30) for 1 week before the start of the experiments. The animals were given food and water ad libitum. The study was conducted according to the experimental schedule depicted in Figure 1. After 1 week of acclimatization, blood samples were collected from the tail vein to measure serum glucose level. The db/db mice were randomly divided into 5 groups and then received daily administration of the test drugs, except for one group that was used for anxiolytic drug treatment. The drug administration was continued during the experimental period. All animal research procedures used in the present study were in accordance with the Guiding Principles for the Care and Use of Animals (NIH Publications No. 80-23, revised in 1996). The present study was also approved by the Institutional Animal Use and Care Committee of the University of Toyama (approval No.: S-2009 INM-1).

Test drug administration was performed using a feeding needle once daily at around 9 a.m. during the experimental period. The m/m and db/db vehicle (control) groups were given water perorally (p.o.), while the other groups were orally administered CTS daily at doses of 375 and 750 mg/kg body weight or injected intraperitoneally with THA (2.5 mg/kg) (Sigma-Aldrich Japan, Tokyo, Japan) dissolved in physiological saline, except when stated otherwise.

CTS used in this study was purchased from Tsumura Co. (Tokyo, Japan) and was the same lot (Lot #202004-7010) as used in previous studies[13, 20]. CTS was extracted from a mixture of 3.0 parts Uncariae Uncis cum Ramulus (hooks and branch of Uncaria rhynchophylla MIQUEL), 3.0 parts Aurantii Nobilis pericarpium (peel of Citrus unshiu MARKOVICH), 3.0 parts Pinelliae tube r (tuber of Pinellia ternate BREITENBACH), 3.0 parts Ophiopogonis tuber (root of Ophiopogon japonicus KER-GAWLER), 3.0 parts Hoelen (sclerotium of Poria cocos WOLF), 2.0 parts Ginseng radix (root of Panax ginseng C.A. MEYER), 2.0 parts Saphoshnikoviae radix (root and rhizome of Saposhnikovia divaricata SCHISCHKIN), 2.0 parts Chrysanthemi flos (flower of Chrysanthemum morifolium RAMATULLE), 1.0 part Glycyrrhizae radix (root of Glycyrrhiza uralensis FISHER), 1.0 part Zingiberis rhizome (rhizome of Zingiber officinale ROSCOE), and 5.0 parts Gypsum fibrosum (CaSO₄ 2H₂O). The yield of CTS extract was 16.1%. To identify the chemical constituents of CTS, 3 dimensional high-performance liquid chromatography (3D-HPLC) analysis was conducted as previously described[11, 13]. Briefly, Chotosan (2.5 g, Tsumura, Tokyo, Japan) was filtered and then subjected to HPLC analysis. HPLC equipment was controlled with an SLC-10A (Shimadzu, Kyoto, Japan) using a TSKGELODS-80TS column (4.6 × 250 mm), eluting with solvents (A) 0.05 M AcONH₄ (pH 3.6) and (B) CH₃CN. A linear gradient of 100% A and 0% B changing over 60 min to 0% A and 100% B was used. The flow rate was controlled with an LC-10 AD pump at 1.0 ml/min. The eluent from the column was monitored and was processed using an SPD-M10A diode array. For chemical profiling of CTS, liquid chromatography-mass spectrometry (LC-MS) analysis was performed with a Shimadzu LC-IT-TOF mass spectrometer equipped with an electrospray ionization (ESI) interface. The ESI parameters were as follows: source voltage +4.5 kV, capillary temperature 200°C, and nebulizer gas 1.5 l/min. The mass spectrometer was operated in positive ion mode scanning from m/z 200 to 2000. A Waters Atlantis T₃ column (2.1 mm i.d. × 150 mm, 3 μm) was used and the column temperature was maintained at 40°C. The mobile phase was a binary eluent of (A) 5 mM ammonium acetate solution and (B) CH₃CN under the following gradient conditions: 0-30 min linear gradient from 10% to 100% B, 30-40 min isocratic at 100% B. The flow rate was 0.15 ml/min. Mass spectrometry data obtained from the extract have been listed in MassBank database[23] and stored together with the pharmacological information on the extract in the Wakan-Yaku Database system (WakanDB ID: LCMS: Chotosan/11000001http://​wakandb.​u-toyama.​ac.​jp/​wiki/​LCMS:​Chotosan/​11000001), Institute of Natural Medicine, University of Toyama. Voucher specimen (CTS: No. 2020047010) obtained from Tsumura Co. Ltd. has been deposited at our institute.

Before and after completing the behavioral studies, serum samples were collected from tail vein of each animal group under pentobarbital (50 mg/kg, i.p.) anesthesia. The serum glucose levels were measured using a commercially available kit (Glucose CII-Test, Wako Pure Chem., Osaka, Japan) as previously described[5].

Western blotting was performed by a slightly modified version of a method described previously[5, 20, 21, 26]. Briefly, hippocampal tissues were dissected from the brain and proteins were extracted by homogenization in protein lysis buffer (50 mM Tris (pH 7.4), 150 mM NaCl, 0.5% sodium deoxycholate, 1%(v/v) NP-40, 0.1%(v/v) sodium dodecyl sulfate (SDS), 150 mM NaF, 8.12 μg/ml aprotinin, 2 mM sodium orthovanadate, 10 μg/ml leupeptin, and 2 mM phenylmethylsulfonyl fluoride using TissueLyser® (Qiagen, Osaka, Japan). Lysate samples were centrifuged at 10,000 rpm (4°C) for 5 min, and the supernatants were centrifuged for protein concentration assay. The protein concentration was determined using a BCA™ protein assay kit (Thermo Scientific, Rockford, IL, U.S.A.) and a microplate reader (Sunrise Classic; TECAN Japan, Kawasaki, Japan). Total protein (15 μg) prepared from each sample was electrophoresed on 7.5% SDS-polyacrylamide gel and then electro-blotted onto a polyvinylidene difluoride membrane (Clear Blot Membrane-p; ATTO). The membranes were incubated in a 5% non-fat milk-containing wash buffer (50 mM Tris (pH 7.5), 150 mM sodium chloride, and 0.1% Tween 20) for 1 h at room temperature. They were then probed with a specific antibody of interest at 4°C for 24 h. After the membranes were rinsed in TBS-T, the blots were incubated with bovine anti-goat IgG secondary antibodies linked with horseradish peroxidase (Santa Cruz Biotechnology, CA, USA) or anti-rabbit or anti-mouse secondary antibodies linked with horseradish peroxidase (DakoCytomation EnVision + System-HRP-labeled Polymer) (Dako Cytomation, Inc., Carpinteria, CA, USA) according to the manufacturer’s instructions. The immune complexes were detected by the enhanced chemiluminescence method (Immobilon® Western Chemiluminescent HRP Substrate) (Millipore, Temecula, CA, USA) and imaged using Lumino Image Analyzer LAS-4000 (Fujifilm, Tokyo, Japan). Band images were analyzed by VH-H1A5 software (Keyence, Osaka, Japan). The quantity of immunoreactive bands was normalized by comparing with their expression levels in treatment-naïve control mice. The dilution ratios and commercial sources of the antibodies used were as follows: anti-choline acetyltransferase (ChAT) goat polyclonal antibody (AB-144P; 1:3000 dilution) (Millipore, CA, USA); anti-β-actin mouse monoclonal antibody (ab8226; 1:10,000 dilution) and anti-PDGFR-α rabbit polyclonal antibody (ab61219; 1:500 dilution) (Abcam, Tokyo, Japan); anti-Akt (pan) (C67E7) rabbit monoclonal antibody (#4691; 1:1000 dilution), anti-PDGFR-β (28E1) rabbit monoclonal antibody (#3169; 1:500 dilution), anti-phospho-Akt (Ser473) rabbit polyclonal antibody (#4060; 1:500 dilution), anti-phospho-PKCα/βII (Thr638/641) rabbit polyclonal antibody (#9375; 1:1000 dilution), anti-PKCα (C67E7) rabbit polyclonal antibody (#2056; 1:1000 dilution) (Cell Signaling Technology, Danvers, MA, USA); anti-muscarinic ACh receptor rabbit polyclonal antibodies (1:500 dilution) (M1: sc-9106; M3: sc-9108; M5: sc-9106), anti-PDGF-A (E-10) mouse monoclonal antibody (sc-9974; 1:1000 dilution), anti-PDGF-B (H-55) rabbit polyclonal antibody (sc-7878; 1:1000 dilution), and anti-VEGF (A-20) rabbit polyclonal antibody (sc-152; 1:1000 dilution), (Santa Cruz Biotechnology, Santa Cruz, CA, USA); and anti-VEGFR2 rabbit polyclonal antibody (Ab-951; 1:500 dilution) (Signalway Antibody, USA).

CTS administration-induced changes in expression levels of ChAT in the medial septum and VEGF in the retia of m/m and db/db mice were also examined by immunohistochemical analysis[5]. Mice were fixed by intracardiac perfusion with 4% paraformaldehyde in phosphate-buffered saline (PBS) under anesthesia. Brain and retinal tissues were post-fixed with 4% paraformaldehyde overnight at 4°C. A series of 5 μm sections were obtained by using a sliding microtome (IVS-410, SAKURA Finetek Japan, Tokyo, Japan). The paraffin-embedded specimens were deparaffinized in xylene and dehydrated with ethanol. Endogenous peroxidase was blocked with 0.1% hydrogen peroxide-methanol for 30 minutes at room temperature. Washed with Tris-buffered saline (TBS), the specimens were incubated in a microwave oven (250 W, 4 s on and 3 s off, MI-77, Azumaya, Tokyo, Japan) in target retrieval solution (Dako, Denmark) for 15 minutes and then washed with distilled water and TBS. Nonspecific binding was blocked by treatment with a special blocking reagent (Dako, Denmark) for 15 minutes. For staining of ChAT positive cells in the medial septum, the specimens were challenged with 1:400 dilution of goat anti-ChAT polyclonal antibody (AB-144P: Millipore, CA, USA) and then incubated in a moist box at 4°C overnight. Washed with TBS, the specimens were incubated with a peroxidase-conjugated bovine anti-goat IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA). After three washes in TBS, a reaction product was detected with 3,3’-diaminobenzidine tetrahydrochloride (0.25 mg/ml) and hydrogen peroxide solution (0.01%). Counter-stained with hematoxylin, the sections were rinsed, dehydrated and covered. Also included in each staining run were negative controls in which the primary antibody was omitted. The images were captured with a microscope (AX-80, Olympus, Japan). For staining of VEGF positive cells in the retina, the specimens were challenged with 1:400 dilution of anti-rabbit VEGF (A-20) polyclonal antibody (sc-152: Santa Cruz Biotechnology, Santa Cruz, CA, USA) in target retrieval solution (Dako, Denmark), incubated in a microwave oven (250 W, 4 s on and 3 s off, MI-77, Azumaya, Tokyo, Japan) for 15 minutes, and then incubated in a moist box at 4°C overnight. After washing the specimens with TBS, they were incubated with Alexa Fluor®488-conjugated goat anti-rabbit IgG antibodies (A11008; 1:400 dilution, Invitrogen, Tokyo, Japan) for 30 min at room temperature. Fluorescent images were captured using a fluorescent microscope (AX-80, Olympus, Tokyo, Japan).

Statistical analysis of the data was conducted according to Curran-Everett and Benos[27]. All data are expressed as the mean ± S.D. The behavioral data were analyzed using paired Student’s t-test or one-way repeated measure ANOVA or two-way repeated measure ANOVA followed by the Tukey test for multiple comparisons among different groups. The neurochemical data were compared using unpaired Student’s t-test or one-way ANOVA followed by the Tukey test for multiple comparisons. Differences of P < 0.05 were considered significant.

In this study, three different types of behavioral tasks, namely, ORT, MYT, and WMT, were used to elucidate non-spatial short-term memory[13, 21], novelty-related spatial working memory[25, 26], and hippocampus-dependent spatial learning and reference memory[12, 33], respectively, in diabetic db/db mice and non-diabetic m/m mice. The results revealed that, compared with non-diabetic control mice, diabetic mice exhibited clearly impaired cognitive learning and memory performance in these behavioral tasks. Moreover, it was observed that diabetic mice showed reduced motor activity, that is, locomotor activity in the MYT and swimming speed in the WMT. These findings are consistent with previous reports that db/db mice are hypolocomotive[5, 34] and have cognitive deficits[5, 6, 35]. Therefore, it is likely that learning and memory deficits of db/db mice observed in this study were caused by reduced motor activity relevant to elevated body weight of this animal group. However, this possibility seems to have been ruled out by the data obtained in our present and previous studies[5]. Indeed, our data indicated that daily administrations of THA and Kampo medicines ameliorated cognitive learning and memory performance of db/db mice without affecting their locomotor and swimming ability. Considering the fact that CTS and THA had no effect on serum glucose levels or body weights in db/db mice, these drugs likely ameliorate and/or prevent diabetes-induced cognitive deficits via a mechanism independent of anti-hyperglycemia and/or anti-obesity.

The present study also demonstrated that db/db mice exhibited anxiety-like behavior in the EPM test, suggesting that they are more susceptible to phobia-driven anxiety than the control m/m strain. This finding is in accord with the work of Dinel et al.[6]. They reported that db/db mice had neuropsychiatric symptoms such as cognitive deficits and anxiety-like behaviors and that the symptoms were associated with increased inflammatory cytokines and reduced expression of brain-derived neurotrophic factor (BDNF) in the hippocampus. Interestingly, in our present study, the anxiety-related behaviors of db/db mice were significantly attenuated in the db/db groups that had received an acute injection of the anxiolytic drug diazepam and daily administration of CTS. Moreover, it was of interest to note that, in contrast to CTS, THA had no effect on the emotional performance of db/db mice. These results suggest that CTS has a beneficial effect on neuropsychiatric symptoms of diabetic animals and allow us to speculate on possible involvement of GABAergic mechanisms in the anxiolytic-like action of CTS in db/db mice. However, this possibility can be ruled out for a couple of reasons. First, it is generally recognized that facilitation of central GABAergic systems by drugs such as diazepam impairs learning and memory performance[36, 37]. Secondly, in our previous study using SAMP8, a senescence-accelerated mouse model, we demonstrated that repeated administration of CTS ameliorated not only cognitive deficits but also a reduced level of anxiety observed in this animal model[38]. Thirdly, there is a possibility that the anxiolytic-like effect of CTS observed in db/db mice is mediated by interaction of CTS with BNDF and VEGF/PDGF systems in the brain since these systems are reportedly implicated in attenuation of anxiety behavior in rodents[17, 39]. However, considering the present and previous data obtained from THA-treated db/db animals, this possibility seems unlikely. THA administration failed to affect anxiety behavior of db/db animals, but it significantly revered the reduced levels of BDNF[5] and VEGF/PDGF in the brain. Therefore, the anxiolytic-like action of CTS may be mediated by neuronal mechanism(s) independent of central cholinergic and VEGF/PDGF systems. Further investigations are needed to clarify the mechanism(s) underlying the anxiolytic-like effect of CTS in db/db mice.

We next elucidated the effects of CTS and THA administration on phosphorylation of Akt and PKCα as indices of impaired signaling and activation of PKC related to hyperglycemia-induced complications in db/db mice[40‐42]. In this study, Akt phosphorylation and PKCα/βII autophosphorylation in the hippocampi of db/db mice were significantly down-regulated and up-regulated, respectively, compared with those in the control m/m mice and these alterations caused by diabetes were reversed by CTS and THA administrations. Evidence indicates that the Akt activation involves tyrosine kinase receptors, such as receptors for insulin, or certain growth factors, such as vascular endothelial growth factor, and that it is triggered by Akt phosphorylation via phosphatidylinositol 3 kinase and phosphoinositide-dependent protein kinases. An Akt activation mechanism is reduced in the CNS of diabetes models including db/db mice[43, 44], although there is a report with conflicting findings[45]. Moreover, recently it was reported that phosphatidylinositol 3-kinase cascade including Akt phosphorylation is involved in neuroprotective influence of cholinergic drugs on glutamate-induced neurotoxicity[46], indicating that PI3K-Akt signaling pathway plays an important role in the cholinergic mechanism[47]. On the other hand, it is generally believed that intracellular PKC is activated by the diabetes-induced accumulation of its co-factor, diacylglycerol, inside the cells[48] and that, once activated, PKC undergoes autophosphorylation via translocation from the cytosol to the plasma membrane and other subcellular compartments[41]. Therefore, our data are consistent with these findings. Considering the effects of THA, our findings suggest that CTS and THA can improve dysfunctions of the Akt and PKC signaling systems in the CNS of diabetic db/db animals via facilitation of central cholinergic systems and that these effects are independent of their effects on serum glucose levels.

Interestingly, this study demonstrated that, compared with the vehicle-treated non-diabetic control group, the vehicle-treated db/db group exhibited down-regulation of the VEGF/PDGF signaling systems in the hippocampus that were reversible by CTS and THA administration. VEGF is a hypoxia-inducible secreted protein that interacts with receptor tyrosine kinases such as VEGFR2 on endothelial cells and promotes angiogenesis. PDGF as well as VEGF plays an important role in angiogenesis that depends on endothelial cell invasion and proliferation and requires pericytes coverage of vascular sprouts for vessel stabilization. VEGF and PDGF coordinate these processes on endothelial cells and vascular smooth muscle cells, respectively[49, 50]. Moreover, these growth factors are implicated in the adverse vascular effects of hyperglycemia-related complications[15, 51] and their functions in the peripheral tissues are up-regulated. Therefore, down-regulation of VEGF/VEGFR and PDGF/PDGFR systems found in this study was in contrast to these previous findings. Recent evidence indicates that the signaling pathways of VEGF and PDGF receptors also involve Akt activation via phosphatidylinositol-3 kinase/Akt pathways and thereby exhibit neuroprotective activities[52, 53] or mediate protective influence of cholinergic drugs on ischemic cell damage[54, 55]. Therefore, it is likely that alterations of expression levels of VEGF/PDGF and their receptors in the hippocampus are relevant to the aforementioned decrease of phosphorylated Akt in the db/db mice.

In the central nervous system, VEGF and VEGFR2 are expressed not only in vascular endothelial cells but also in other cells such as neurons and neural progenitor cells[56] and are involved in brain functions including enhancement of neurogenesis through the direct activation of neural progenitor cells[57], amelioration of cognitive deficits via the promotion of neurogenesis, and protection of endothelial cells and neurons during brain ischemia in adult rats[57, 58]. There is also evidence that PDGF-A and -B and their receptors (PDGFRα and PDGFRβ) play a role not only in the proliferation, migration, and differentiation of oligodendrocytes[59] but also in neurite outgrowth[60] and neuroprotection via phosphatidylinositol 3-kinase, a mitogen-activated kinase kinase pathway[30]. These signaling mechanisms are important in the long-term potentiation of learning and memory, a biological index of memory formation[61]. We previously reported using a senescence-accelerated mouse model that aging induces dysfunction of the VEGF/VEGFR2 and PDGF/PDGFR signaling systems in the brain and that reversal of impaired VEGF/VEGFR2 and PDGF/PDGFR signaling systems is a part of the mechanism(s) underlying the ameliorative effects of CTS on spatial and non-spatial cognitive deficits caused by aging[20]. Taken together, the present results suggest that the CTS-induced reversal of expression level changes of VEGF/VEGFR2 and PDGF-B/PDGFRβ in db/db mice contributes to the improvement of cognitive performance by CTS administration.

It is of interest to note that the effects of CTS and THA on VEGF/PDGF systems in the brain of db/db animals were different from those on the retina in db/db mice. VEGF/PDGF has been implicated as a major contributor to the development of diabetic complications such as diabetic retinopathy[32, 62]. Elevated expression of VEGF and its receptor has also been demonstrated in diabetic retinas[63]. These roles of VEGF/PDGF in diabetic complications raise the possibility that CTS administration may exacerbate diabetic retinopathy. However, this possibility seems to be little since, although VEGF was highly stained in the endothelial cells and microvessel regions of db/db mice compared with that in the control m/m mice are consistent with these previous reports, CTS or THA treatment had no effect on the retinal VEGF expression in the db/db mice. The reason for the different susceptibility of VEGF/PDGF systems to CTS and THA treatment between the brain and retinal tissues is unclear. However, there may be a difference in the mechanism to regulate expression of these factors between the central nervous system and the peripheral tissues.

To obtain a better understanding of the mechanism by which CTS ameliorates cognitive deficits in db/db diabetic animals, we next elucidated the effect of CTS and THA on the central cholinergic systems since these systems play an important role in cognitive performance and their hypofunction is closely related to the progression of memory deficits[64, 65]. Indeed, evidence indicates that muscarinic receptors such as M₁, M₃, and M₅ subtypes have an important role in cognitive function in rodents[13, 66‐68]. In this study, we found that the expression levels of cholinergic marker proteins in the hippocampus, namely, choline acetyltransferase (ChAT) and muscarinic M₁, M₃, and M₅ receptors, were down-regulated in the vehicle-treated db/db mice compared with those in the age-matched vehicle-treated m/m group and that CTS treatment, as well as THA treatment, reversed the downregulated expression of these proteins in db/db mice. These findings are consistent with our previous study that db/db mice had dysfunction of the central cholinergic systems[5]. In the previous study, we demonstrated that the dysfunction likely occurs in an aging-dependent manner and is accelerated in db/db mice because no differences in expression levels of these marker proteins were observed when compared between young (7-week-old) db/db mice and age-matched m/m mice. Taking these findings together, one of the plausible explanations for CTS- and THA-induced amelioration of cognitive deficits of db/db mice is that these drugs in part protect the central cholinergic systems from aging-induced degeneration of cholinergic neurons that is accelerated in db/db animals. This idea is supported by the present immunohistochemical study that the medial septum in db/db mice had fewer ChAT-immunopositive cells providing projections to the hippocampus than that in the age-matched control m/m mice and that the decrease of the cells was prevented in the db/db groups treated with CTS and THA. The reversal of downregulated expression of the cholinergic marker proteins in CTS- and THA-treated db/db mice seems to be consistent with the findings reported by Kakinuma et al.[54, 55]. They demonstrated using cardiomyocytes that an increase in acetylcholine level by an acetylcholinesterase inhibitor or vagal nerve stimulation directly transduces cell survival signal through muscarinic receptors, activates the PI3K/Akt/HIF-1α/VEGF pathway, and leads to protection of the cell and increase of ChAT expression.

It is of interest that the effects of CTS on cognitive function and central cholinergic systems in db/db mice are quite similar to those of THA. From these findings, one may infer that chemical constituents and/or their metabolites of CTS have a potential to inhibit the activity of acetylcholinesterase in the brain; however this possibility seems to be little, since, in our previous study, administration of THA but not CTS reduced ex vivo activity of acetylcholinesterase in the brain[13]. Our findings suggest that the mechanism of action of CTS may differ from that of THA. Therefore, the exact mechanism underlying CTS-induced modulation of central cholinergic systems in db/db mice requires further investigations.