Introduction

Diabetes mellitus (DM) has become a major health threat worldwide, with ample evidence indicating DM as an increased risk of cognitive decline and dysfunction [1,2,3]. DM patients often exhibit reduced psychomotor efficiency [4], cognitive flexibility and rapid information-processing dysfunctions [5]. Moreover, the impact of hyperglycemia on brain and its functionalities is evidently illustrated in Alzheimer-type neuropathological changes among diabetes individuals [6]. Additionally, spatial learning impairment is reportedly associated with the impairment of central nervous system in STZ-induced diabetic rat model [7, 8].

In DM, excessive advanced glycation end products (AGEs) are accumulated [9]. The receptor for advanced glycation end products (RAGE) belongs to the superfamily of immunoglobulins and functions as a central signal transduction receptor of AGEs. The interactions between excessive AGEs and RAGE can deteriorate body functions by interfering with activities of glycated macromolecules and enzymes, leading to inflammatory responses [10]. Moreover, AGEs can impair synapse and memory functionality and efficiently exacerbate memory deterioration [11]. DM reportedly results in release of inflammatory cytokines, including tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) in the plasma and brain tissue [12, 13], all of which contribute to the accelerated neurodegeneration [14,15,16]. The elevated levels of IL-1β and TNF-α can trigger and maintain numerous deleterious changes in the brain, with the synaptic architectures and functionalities significantly compromised [17,18,19,20]. Growth-associated Protein-43 (GAP-43) is located on the growth cones of axons and is involved in structural synaptic plasticity in the nervous system [21] and is downregulated in diabetic rats [22]. Synaptophysin (SYN) is a vesicle-associated protein which is rapidly recruited to presynaptic terminals in response to presynaptic neuronal activity [23]. SYN is downregulated in the hippocampus in rats with STZ-induced diabetes [24, 25].

Luteolin has been investigated for its antioxidant, anti-inflammatory and neuroprotective activities [26, 27]. In addition, luteolin can reportedly suppress the release of inflammatory cytokines and abate microglial activities, thereby improving the spatial memory in aged mice [28]. The findings in our prior experiment revealed that luteolin ameliorated the memory loss in diabetic rats [29]. Nevertheless, there is a paucity of information regarding the protective effect of luteolin on diabetes-associated neuronal impairment and memory deficit. Herein, we focused on investigating whether luteolin could suppress the inflammatory cytokines and thus increase synapse-associated proteins in STZ-induced diabetic rat models.

Materials and Methods

Animals and Diabetes Models

Male Sprague-Dawley rats (aged 8–10 weeks) (Branch of National Breeder Center of Rodents, Shanghai, China) were housed in polypropylene cages and maintained at 23 ± 1 °C under a 12-h light/dark cycle with ad libitum access to water and rodent chow. All experiments were performed in conformance with Chinese Council on Animal Care Guidelines and approved by the Institutional Animal Care Committee. Rats were randomized into four groups (n = 8 each): blank vehicle (control) group, STZ-induced diabetic group (DM), and luteolin DM groups (i.e. 50 mg/kg, LT50 group and 100 mg/kg, LT100 group). Diabetes mellitus was induced by intraperitoneal injection of STZ (60 mg/kg, Sigma-Aldrich Co. LLC, St. Louis, MO, USA) dissolved in the vehicle citrate buffer (pH 4.4, 0.1 M). Three days after STZ injection, rats presenting with fasting plasma glucose levels of over 11.1 mM (200 mg/dl) and signs of polyuria, polyphagia and polydipsia were recruited for the experimental protocols, with further intraperitoneal injection of STZ to induced the non-conformant rats.

Treatment Conditions

With the rat model established, luteolin (Shaanxi Sciphar Biotechnology Co., Ltd., Shanxi, China) dissolved in 0.5% (w/v) sodium carboxymethylcellulose (CMC-Na) solution was administered at daily doses of 50 mg/kg and 100 mg/kg by gavages in the corresponding diabetic rat groups till the end of eight weeks [29], in parallel with daily gavage of CMC-Na in the DM and control groups. In addition, body weight and fasting plasma glucose levels were measured prior and subsequent to the luteolin treatments. The treatment lasted for 8 weeks commencing from DM model establishment. Thereafter, animals were scored for learning and memory task in Morris water maze for 5 consecutive days. Subsequently, all animals underwent deep anesthesia and blood collection via femoral vein for serum separation. After sacrifice, brain tissues were rapidly removed and the hippocampus isolated.

Morris Water Maze Test

The Morris water maze (MWM) test was performed [30] with minor modifications. With the rat diabetic model established after STZ injection for 3 days, rats were given luteolin for 8 consecutive weeks, followed by the MWM. The test was carried out in a circular pool (diameter, 120 cm; height, 60 cm) filled with water. Poured milk into the pool to make it opaque and controlled the water temperature at 24 ± 1 °C. The pool was divided into four quadrants, in which an escape platform was hidden in the third quadrant. The rats received four consecutive daily training trials. Each trial lasted until the rats found a platform within 90 s. If a rat did not find the escape platform within 90 s, it was given a latency score of 90 s. A probe trial was performed on the fifth day, the rats were placed and released opposite the site where the platform had been located. The probe trial consisted of a 90 s free swim in the pool without the platform. The percentage of time spent in the target quadrant and number of crossing to target platform indicates the degree of memory consolidation. Thereafter, the animals underwent sodium pentobarbital (50 mg/kg, i.p.) anesthesia and blood collection from the femoral vein, followed by rat sacrifice and isolations of the rat brain and the hippocampus, which were to be stored at − 80 °C until subsequent procedures.

Nissl Staining

Subsequent to the behavioral test, rats were subjected to sodium pentobarbital anesthesia (50 mg/kg, i.p.), followed by heart exposure and transcardial perfusion with ice-cold normal saline via the left ventricle for blood clearance and tissue fixation by 4% paraformaldehyde. The brain tissues were removed and immersed in 4% paraformaldehyde for 72 h for fixation, followed by cryoprotection in 15% sucrose solution for 24 h and 30% sucrose solution for 12 h. For Nissl staining, rat brain tissues were coronally prepared at the thickness of 30 µm using frozen sectioning technique (model: E, Thermo, USA). The thick sections in each group were stained with Cresol Blue, soaked in 0.1% cresol violet staining solution, dehydrated by 95% or 100% anhydrous ethanol and washed in dimethylbenzene. The population of positive cells in the hippocampal CA1 area was microscopically counted by an investigator blinded to the protocol. All procedures were in triplicate, with the mean calculated from bilateral fields. Data were represented as cell count per mm2.

ELISA

Blood samples were centrifuged at 12,000 rpm for 10 min at 4 °C to obtain serum. The concentrations of AGEs, IL-1β and TNF-α concentration in the plasma were examined using corresponding rat enzyme-linked immunosorbent (ELISA) kits [31] (Westang Bio-tech Co. Ltd., Shanghai, China) according to the manufactures’ instructions.

Reverse Transcriptase-PCR (RT-PCR)

For mRNA quantification, total RNA was extracted with RNAqueous kit (Applied Biosystems, Foster City, CA, USA). The cDNA was synthesized using High Capacity RNA-to-cDNA kit (Applied Biosystems) as indicated by the manufactures’ instructions. The housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as an internal control for normalization in parallel with each gene examined. Amplified products were separated by electrophoresis on a 1% agarose gel, followed by visualization under a UV trans-illuminator and photography. To verify reproducibility, each brain sample was analyzed in triplicate for each gene (IL-1β, 5′-TTGACGGACCCCAAAAGATG-3′(F); IL-1β, 5′-AGAAGGTGCTCATGTCCTCA-3′(R); TNF-α, 5′-GGGAGTAGACAAGGTACAAC-3′(R); TNF-α, 5′-TCTCATCAGTTCTATGGCCC-3′(F); GAPDH, 5′-CCTTCCGTGTTCCTACCC-3′(F); GAPDH, 5′-CAACCTGGTCCTCAGTGTAG-3′(R)). The values for the target gene expression were normalized to GADPH and quantified relative to the expression in control samples. The products were analyzed by densitometry using the Quantity One 1-D analysis software (BioRad, Hercules, CA, USA), with quantification performed relative to GADPH (Table 1).

Table 1 Sequences and annealing temperatures of the oligo primers
Full size table

Western Blot Analysis

For protein extraction, frozen hippocampus was homogenized in ice-cold extraction buffer (20 mM Tris–HCl buffer, pH 7.6, 150 mM NaCl, 2 mM EDTA·2Na, 50 mM sodium fluoride, 1 mM sodium vanadate, 1% Nonidet P-40, 1% sodium deoxy cholate, 0.1% SDS, 1 mg/ml aprotinin and 1 mg/ml leupeptin). Homogenates were centrifuged at 12,000 rpm for 10 min at 4 °C to obtain supernatants, which was determined for protein concentrations by a Pierce BCA Protein Assay Kit (Beyotime Institute of Biotechnology, Shanghai, China). Equivalent proteins (20 µg) were separated by 10% SDS-polyacrylamide gel electrophoresis and transferred onto a nitrocellulose membrane (Biosharp Biotech CO. Ltd., Hefei, Anhui, China). The membrane was blocked with 5% nonfat milk in a washing buffer [Tris-buffered saline containing 0.05% (v/v) Tween 20] at room temperature (r/t) for 2 h, followed by incubation overnight with the primary antibodies (RAGE, 1:1000, #4679S, SYN, 1:1000, #6855S, and GAP-43, 1:1000, #5307S, Cell Signaling Technology, Inc., Danvers, MA, USA). With each membrane twice rinsed for 15 min, the rabbit polyclonal antibodies (Santa Cruz Biotechnology, CA, USA) against the active RAGE, SYN and GAP-43 were incubated and visualized by 3,3′,5,5′-tetramethylbenzidine-stabilized substrate (Promega, Madison, WI, USA). β-actin (1:1000; Beijing Zhongshan Golden Bridge Biotechnology Co. Ltd., Beijing, China) served as an internal control for the cytosolic extracts. The membranes were exposed to BCIP/NBT alkaline phosphatase color developing reagent (Beyotime) for 15 min. Comparison between treatment groups was performed by determination of protein/β-actin ratio of the immunoreactive area by densitometry. Protein bands were quantified by densitometry using Scion Image analysis software (Scion Corp., Frederick, MD, USA). The OD values were normalized using β-actin as an internal control.

Statistical Analysis

All statistical analyses were performed by the SPSS 16.0 (SPSS Inc., Chicago, IL, USA). Data were expressed as the mean ± SEM (n = 3 each group). Statistical significance was set at P < 0.05. The intergroup differences were analyzed using one-way ANOVA followed by the Turkey’s tests.

Results

Effect of Luteolin on Body Weight and Fasting Plasma Glucose Levels in Rats

Compared with the control rats, diabetic rats exhibited significant increase in fasting plasma glucose levels (P < 0.01, Fig. 1a), as well as a marked decline in body weight (P < 0.05, P < 0.01, Fig. 1b). At the end of luteolin treatment, the fasting plasma glucose levels were significantly reduced in both treatment groups versus DM group (all P < 0.001; Fig. 1a). The body weight in LT100 group was markedly increased versus DM group (P < 0.05; Fig. 1b).

Fig. 1
figure 1

Fasting plasma glucose levels (a) and body weight (b) in the groups. Data are expressed as mean ± S.E.M., n = 8. *P < 0.05, **P < 0.01 versus the control group. #P < 0.05, ##P < 0.01, ###P < 0.001 versus the DM group

Full size image

Effect of Luteolin on MWM Test in Rats

To examine whether luteolin could attenuate the diabetes-induced memory impairments, we tested the spatial learning and memory with the MWM test. At day 1, there was no significant difference in mean escape latency among groups (Fig. 2a), and at day 2, significant differences were evident in escape latency between the diabetic and the control rats (P < 0.001). From the third day, the escape latency was significantly decreased in groups LT50 (P < 0.01, P < 0.001) and LT100 (P < 0.01, P < 0.001) as compared to the DM group.

Fig. 2
figure 2

Effects of luteolin on escape latency (a) and number of crossings (b) in MWM test, and the percentage of time spent in target quadrant (c). Data are expressed as mean ± S.E.M., n = 8. *P < 0.05, ***P < 0.001 versus the control group. #P < 0.05, ##P < 0.01, ###P < 0.001 versus the DM group

Full size image

At day 5, in the probe trial of the MWM test, the rats showed a significant difference in behavioral performance. The numbers of crossings in both the arena (P < 0.05, Fig. 2b) and the time spent in target quadrant (P < 0.05, Fig. 2c) were significantly reduced in DM group versus the control group, reflecting memory impairment. The number of crossings in the arena (P < 0.05, Fig. 2b) and the time spent in the target quadrant (P < 0.05, Fig. 2c) were significantly increased with luteolin (100 mg/kg) treatment.

Nissl Staining in the Hippocampus

Following Nissl staining of the hippocampus in each group, the hippocampal CA1 subfields were observed for microphotography. The number of surviving neurons was significantly lower in the hippocampus in DM group than in the control group (P < 0.001, Fig. 3b). Administrations of luteolin (50 and 100 mg/kg) significantly increased the amount of surviving neurons versus the DM group (P < 0.01; P < 0.001).

Fig. 3
figure 3

Representative Nissl staining in the hippocampal CA1 subfield (a) The population of intact cells in the hippocampal CA1 subfield were counted as shown respectively in (b) Data are expressed as mean ± S.E.M., n = 3. ***P < 0.001 as compared to the control group; ##P < 0.01, ###P < 0.001 as compared to the DM group

Full size image

Effect of Luteolin on the Expression of AGEs and RAGE in Diabetic Rats

AGEs levels were significantly increased in the serum (P < 0.001; Fig. 4a) in DM group versus the control group. However, luteolin treatment (50 and 100 mg/kg) resulted in significant decrease in the levels of AGEs (P < 0.01; P < 0.001) versus the DM group.

Fig. 4
figure 4

Effects of administration of luteolin on the serum level of AGEs (a, n = 8) and RAGE proteins in the hippocampus (b, c, n = 3) in rats. Data are expressed as mean ± S.E.M., *P < 0.05, ***P < 0.001 versus the control group. #P < 0.05, ##P < 0.01, ###P < 0.001 versus the DM group

Full size image

Diabetic rats exhibited significantly increased RAGE expression versus the control rats (P < 0.05, Fig. 4c) in the hippocampus. Also, there was a marked decline in the RAGE expression of LT100 group versus DM group (P < 0.05).

Effect of Luteolin on Inflammatory Response in the Hippocampus and Serum

The expressions of IL-1β and TNF-α mRNA in the hippocampus were determined to assess the effects of luteolin on the levels of inflammatory cytokines. As shown in Fig. 5, the expressions of IL-1β (P < 0.05, Fig. 5b) and TNF-α (P < 0.05, Fig. 5d) mRNA were significantly upregulated in the hippocampus in diabetic group versus the control group. Treatments with luteolin (50 and 100 mg/kg) significantly downregulated the TNF-α mRNA expression in hippocampus (P < 0.05; P < 0.01; Fig. 5d) in diabetic rats versus DM rats. In addition, treatment with luteolin (100 mg/kg) significantly downregulated the IL-1β mRNA expression in hippocampus (P < 0.05; Fig. 5b) in diabetic rats versus DM group.

Fig. 5
figure 5

Effects of administration of luteolin on inflammatory cytokine expression in rat hippocampus. a The IL-1β level in the hippocampus. b The TNF-α level in the hippocampus. Data are expressed as mean ± S.E.M., n = 3. *P < 0.05 versus the control group. #P < 0.05, ##P < 0.01 versus the DM group

Full size image

DM is characterized by excessive peripheral inflammation. Thus, we investigated whether luteolin could have any effect on the expressions of serum IL-1β and TNF-α in diabetic rats. By ELISA, expressions of IL-1β (P < 0.05, Fig. 6a) and TNF-α (P < 0.01, Fig. 6b) were significantly upregulated in diabetic rat serum versus the control group. Luteolin administrations (50 and 100 mg/kg) significantly downregulated the TNF-α expression (P < 0.05; P < 0.01; Fig. 6b) and the IL-1β expression in LT100 group (P < 0.05; Fig. 6a) was significantly downregulated, with no significant difference in LT50 group (P > 0.05; Fig. 6a) versus the DM rats.

Fig. 6
figure 6

Effects of administration of luteolin on the serum level of IL-1β and TNF-α in rats. a The level of IL-1β in serum. b The level of TNF-α in serum. Data are expressed as mean ± S.E.M., n = 8. *P < 0.05, **P < 0.01 versus the control group. #P < 0.05, ##P < 0.01 versus the DM group

Full size image

Effect of Luteolin on the Expression of Synapse-Associated Proteins

To further determine the neuroprotective effect of luteolin, we assessed variations of two representative synapse-associated proteins, GAP-43 and SYN, in the luteolin-treated diabetic rats.

Western blotting analysis showed that GAP-43 (P < 0.01, Fig. 7b) and SYN (P < 0.05, Fig. 7c) levels in hippocampus in the diabetic rats were significantly decreased versus control rats. Treatments with luteolin (50 and 100 mg/kg) significantly upregulated the GAP-43 expression in the hippocampus (P < 0.05; P < 0.05; Fig. 7b) in the diabetic rats versus DM group, and the SYN expression in the hippocampus was significantly upregulated with luteolin treatment (100 mg/kg) (P < 0.05; Fig. 7c) versus the DM group.

Fig. 7
figure 7

Effects of administration of luteolin on the expression of synapse-associated proteins in rat hippocampus. Data are expressed as mean ± S.E.M., n = 3. *P < 0.05, **P < 0.01 versus the control group. #P < 0.05 versus the DM group

Full size image

Discussion

DM has become a serious health risk worldwide. In the present study experiment with rat model of DM, we demonstrated that luteolin could not only attenuate inflammatory responses by downregulating AGEs and RAGE, but also improve cognitive behaviors by upregulating the synapse-associated proteins in the hippocampus in STZ-induced diabetic rat model.

In this study, we affirmed that the diabetic rats exhibited increased escape latency and the abridged duration in target quadrant in Morris water maze test, indicating impairment to the abilities of spatial learning and memory. We also demonstrated that long-term treatment with luteolin at high doses significantly reversed these behavioral changes, validating the protective potential of luteolin against diabetes-induced cognitive impairment. Our assessments of blood glucose and body weight in each group revealed that long-term luteolin treatment at both low and high doses significantly decreased blood glucose in the STZ-induced diabetic rats. Luteolin at high dose also increased body weight in STZ-induced rat model.

Increasing evidence shows that oxidative stress under diabetes mellitus can result in the excessive activation of reactive oxygen species (ROS), and the elevated of ROS may affect the downstream RAGE/AGE pathway [32]. Activate RAGE/AGE pathway increases inflammation factor, leading to the production of ROS. In addition, the elevated ROS could activate the inflammatory signal, further to promote the expression of inflammatory mediators such as TNF-α and IL-6 [33, 34]. Furthermore, upregulating the level of inflammatory factor IL-1β may damage cognitive function [35, 36]. Based on our previous study [29], STZ induced neuronal damage, cognitive dysfunction and aggravated oxidative stress. Luteolin could improve neuronal damage and cognitive function, and alleviate oxidative stress in diabetic rats.

Diabetes hyperglycemia reportedly enhances the production of AGEs [37,38,39] and RAGE [40], with AGEs being important pathogenetic mediators of almost all diabetic complications, including diabetic encephalopathy [41]. In our present study, the levels of AGEs in serum and the RAGE levels in hippocampus were significantly increased in diabetic rats, whereas luteolin administrations markedly decreased the AGEs and the RAGE levels, which was also supported by previous reports [42, 43]. Increased IL-1β and TNF-α expression in the serum and the hippocampus in the STZ-induced DM rats. The luteolin treatment significantly inhibited the IL-1β and TNF-α expression in the serum and the hippocampus in diabetic rats, thereby we postulated that luteolin could benefit diabetic rats via inhibition of the chronic STZ-induced inflammatory response and amelioration of cognitive impairment. Nissl staining of the hippocampus CA1 in the diabetic rats also confirmed that luteolin could evidently increase the CA1 neuron population.

GAP-43 is a crucial in axonal elongation, synaptic plasticity and nerve sprouting during neuronal development and maturation in adult rats [44, 45]. SYN levels are reportedly reduced in type 2 diabetic patients compared with control individuals [46]. Herein, we demonstrated that the levels of proteins GAP-43 and SYN in the hippocampus were decreased in the diabetic rats. More importantly, we validated that luteolin treatment significantly reversed the STZ-induced decline in both GAP-43 and SYN levels. The potential of luteolin in improving cognition may be associated with its ability to increase GAP-43 and SYN protein levels in the hippocampus in the STZ-induced rat model.

In summary, we demonstrated that luteolin exerts protective effects against the cognitive decline in rat model of STZ-induced diabetes mellitus. Luteoin might inhibit inflammatory responses either directly via the suppression of inflammatory cytokines or indirectly via the downregulation of AGEs/RAGE in STZ-induced diabetic rats, and ultimately the upregulation of synapse-associated proteins in the rat hippocampus. These behavioral and biochemical improvements may shed light on novel regimens with luteolin for diabetes-induced cognition decline.