Peganum harmala enhanced GLP-1 and restored insulin signaling to alleviate AlCl3-induced Alzheimer-like pathology model

Peganum harmala (P. harmala) is a folk medicinal herb used in the Sinai Peninsula (Egypt) as a remedy for central disorders. The main constituents, harmine and harmaline, have displayed therapeutic efficacy against Alzheimer’s disease (AD); however, the P. harmala potential on sensitizing central insulin to combat AD remains to be clarified. An AD-like rat model was induced by aluminum chloride (AlCl3; 50 mg/kg/day for six consecutive weeks; i.p), whereas a methanolic standardized P. harmala seed extract (187.5 mg/kg; p.o) was given to AD rats starting 2 weeks post AlCl3 exposure. Two additional groups of rats were administered either the vehicle to serve as the normal control or the vehicle + P. harmala seed extract to serve as the P. harmala control group. P. harmala enhanced cognition appraised by Y-maze and Morris water maze tests and improved histopathological structures altered by AlCl3. Additionally, it heightened the hippocampal contents of glucagon-like peptide (GLP)-1 and insulin, but abated insulin receptor substrate-1 phosphorylation at serine 307 (pS307-IRS-1). Besides, P. harmala increased phosphorylated Akt at serine 473 (pS473-Akt) and glucose transporter type (GLUT)4. The extract also curtailed the hippocampal content of beta amyloid (Aβ)42, glycogen synthase (GSK)-3β and phosphorylated tau. It also enhanced Nrf2, while reduced lipid peroxides and replenished glutathione. In conclusion, combating insulin resistance by P. harmala is a novel machinery in attenuating the insidious progression of AD by enhancing both insulin and GLP-1 trajectories in the hippocampus favoring GLUT4 production.

Insulin is conveyed from the periphery through a blood-brain barrier (BBB) transporter system to different brain regions 1 . Insulin receptor (IR) is densely expressed in the hippocampus, hypothalamus, and cerebral cortex 2 , where the de novo synthesis of insulin takes place 1,3 . In the brain, insulin signaling plays a crucial role in cognition, where insulin/IR interaction activates signal transduction cascades that augment synaptic plasticity mechanisms 4 . Besides the involvement of insulin and its trajectory in memory consolidation, the glucagon-like peptide (GLP)-1, another peptide hormone, plays a crucial role in maintaining the physiology of the central nervous system 5 . This hormone can also pass BBB and interacts with its widely expressed receptor in different brain regions 6 . GLP-1 has documented its capacity to enhance learning and memory, as well as to mediate neuroprotective effects among its pleiotropic pharmacological potentials 7 . Previous studies revealed that GLP-1 prompts the insulin signaling cascade to be one of its neuroprotective mechanisms 7,8 . Over the past decade, hippocampal insulin resistance has been considered a major incentive for Alzheimer's disease (AD) pathogenesis 9,10 . Currently, it has widely been recognized that AD may be a brain type of diabetes mellitus (DM) or "type 3 diabetes", combining insulin deficiency with insulin resistance 11 . An early investigation probed by Steen et al. 12 demonstrated an irregular neuronal insulin and insulin growth factor (IGF)-I/-II signaling mechanisms in AD brains. In augmentation, the amyloid beta (Aβ) protein aberrantly interferes with the insulin signaling cascade and acts as a nexus between insulin resistance and cognitive impairment in AD 13,14 . This link, hence, highlights the mutual causality between insulin resistance and Aβ oligomerization 15 that consequently fosters mitochondrial dysfunction to induce oxidative damage 16 and initiates membrane-dependent Aβ oligomerization mediated lipid peroxidation 17 . Oxidative stress promotes serine phosphorylation of insulin receptor substrate-1 (IRS-1) at 307, thereby progressing insulin resistance metabolic dysfunction in the brain 18 . On the other hand, GLP-1 receptor agonists reduced AD pathologic markers and improved memory in the absence of diabetes experimentally 19 ; recently, a randomized double-blind placebo-controlled trial reported that long-term treatment with dulaglutide, a long-acting GLP-1 receptor agonist, reduced the hazard of substantive cognitive impairment in type 2 diabetic patients 20 .
Peganum harmala (P. harmala) L. is a traditional plant that belongs to the family of Zygophyllaceae and it is popularly known as Harmal or Haramlaan in North Africa 21,22 , but it is also well-adapted to other dry zones, including the Middle East, India, Iran, and Mongolia 23,24 . In Egypt, it has been used for centuries in folk medicine predominantly in the Sinai Peninsula, where the plant seeds were consumed as an analgesic and the smoke emitted from its burning leaves was inhaled for the relief of headaches and CNS disorders 22 . Phytochemical studies have established that P. harmala has an abundance of β-carboline alkaloids, mainly greater in the seeds 23,25 . These include harmine, harmaline, harmalol, norharmane, and harmane 25 ; allowing P. harmala to have a wide array of therapeutic activities 24 . Studies have confirmed that harmine and harmaline retain an acetylcholinesterase inhibitory effect and an antioxidant activity, to afford a potential in AD treatment 25,26 . Moreover, harmine and harmaline exhibited an antidiabetic action, when tested in a streptozotocin rat model, an effect that was attributed to enhancing insulin sensitivity 24,27 .
In view of the aforementioned data, this study investigated the potential modulatory effect of a standardized methanolic P. harmala seed extract against a rat model of AD underscoring its impact on alleviating insulin resistance with relation to Aβ, tau and oxidative stress. Experimental design. Fifty-two rats were randomly allocated into 4 groups (n = 13 each). Rats in group I received saline (i.p) and served as the normal control group, whereas those in group II provided the AD model in which rats were injected with AlCl 3 dissolved in saline (50 mg/kg/day; i.p) for 6 consecutive weeks 29 . In group III, normal rats received daily i.p saline for 2 weeks followed by daily doses of P. harmala per se (187.5 mg/kg; p.o) for 4 weeks 25 , while in group IV, AlCl 3 -exposed rats were treated with the oral doses of P. harmala 30 min after each neurotoxin injection for a period of 4 weeks, starting from the 15th day of AlCl 3 injection in rats that showed declined cognition 29 .
Behavioral studies. Six days prior to the end of the trial, rats were subjected to behavioral tests starting with the least stressful Y-maze test on day 37, followed by Morris Water Maze (MWM) test that was carried out for 5 successive days as follows; training on days 38 to 41 and on the 42 nd day, the probe trial was performed ( Fig. 1).
Y-maze task. The test is a simple and quick measure of spatial memory that depends upon the natural propensity of rats to exhibit "spontaneous alternation behavior", in which all three arms in the Y-maze are entered successively without repeated entries 30 . The Y-maze apparatus consisted of 3 identical painted wooden arms (40 cm long, 35 cm high, 12 cm wide) designated as A, B, and C, with an angle of 120° between them. Each rat was placed in the center of the maze and was allowed to navigate freely through the maze for 5 min without cues or reinforcements, such as food or water. The number of entries/arm was recorded manually and the arm entry was considered valid when the hind paws of the animal are completely inside the arm. The total number of arm entries and the number of spontaneous alternations were counted to calculate the spontaneous alternation percentage (SAP) according to the following Eq. 31  MWM. This test is particularly sensitive to hippocampal-dependent spatial memory in rats 32 . The test was carried out following the Morris 1984 protocol 33 using a circular pool (150 cm diameter, 60 cm high), filled to a depth of 40 cm with water (26 ± 1 ο C). The pool was divided into 4 virtual quadrants of equal sizes and a circular  www.nature.com/scientificreports/ transparent movable escape platform (10 cm diameter) that was submerged 1 cm below the water surface in the middle of one quadrant (i.e. the target quadrant). For the 4 subsequent days of the training session (acquisition phase), rats were exposed to 4 daily trials to find the hidden platform within 120 s. The time taken to find the platform was recorded for each trial (i.e. escape latency) and the animal was left on the platform for 10 s; however, rats that failed to allocate the platform were guided and placed on the platform for 30 s with a given latency of 120 s. On the 5 th day of the MWM paradigm (day 42 of experimentation), the probe test was carried out, where the platform was removed and rats were allowed to explore the maze for 120 s, in which the percent quadrant time (Q) 34 was recorded using the following equation: Q = (time spent in target quadrant (s)/120) × 100. Twenty-four hours after the completion of behavioral testing, animals were euthanized using a high dose of thiopental (100 mg/kg) and brains were carefully harvested. The hippocampi were gently dissected out to be used for the determination of the biochemical parameters.
Histopathological examination. Moreover, the whole brains of the last 3 rats/group were fixed in 10% neutral buffered formalin for 72 h followed by paraffin embedding for histopathological examination. The formalin fixed brains of each group were trimmed, dehydrated in serial grades of ethanol, cleared in xylene, and embedded in Paraplast tissue embedding media. Sagittal brain Sects. (4 µm thickness) were obtained using a rotatory microtome and were stained with Harris Hematoxylin and Eosin (H&E) staining as a general tissue structure examination. Likewise, other sections were subjected to Nissl staining using Toluidine blue to assess neuronal survival in hippocampal Cornu Ammonis 1 (CA1) and Dentate gyrus (DG) regions. Six random nonoverlapping fields were analyzed blindly for counting intact neurons by using full HD microscopic camera operated by Leica application Suite for tissue sections analysis (Leica Microsystems GmbH, Wetzlar, Germany). All histopathological procedures and evaluation were performed by an external investigator in a blinded manner.
Statistical analysis. All data sets were represented as mean ± standard deviation (SD). GraphPad Prism software 6.0 (GraphPad Software, CA, USA) was used to perform statistical analysis. Comparisons between means were carried out using one-way Analysis of Variance (ANOVA) followed by Tukey's Multiple Comparisons test. Escape latency results from the MWM training sessions were analyzed using two-way ANOVA and followed by Tukey's Multiple Comparisons test. Statistical significance was achieved when p < 0.05.

Results
The administration of P. harmala to normal rats showed no substantial changes compared with their control counterparts for all measured parameters, with the exception of hippocampal GLUT4 content. Therefore, comparisons were only made with reference to the control group.  www.nature.com/scientificreports/ Standardization of P. harmala extract. As illustrated in Fig. 2, the (a) harmine and harmaline in P. harmala extract, quantified through HPLC analysis, showed a retention time of 4.667 and 9.108 min, respectively, which was similar to those of the two standards; viz., (b) harmine and (c) harmaline. The amount of harmine and harmaline present in P. harmala extract was estimated to be about 14 and 21% (w/w), respectively.

P. harmala alleviates the AlCl 3 -induced learning and memory deficits.
The results of the Y-maze test (Fig. 3) confirmed the spatial memory deficit and the decreased general activity in the AlCl 3 -treated rats, as evidenced by the marked inhibition in the (a) SAP and (b) total arm entries by 48 and 68%, respectively, as compared to the control group. On the contrary, post-treatment with P. harmala in AlCl 3 -exposed rats improved spatial memory and restored both measured parameters. Moreover, using the MWM test, the insulted rats showed a learning shortfall indicated by the prolongation of the (c) escape latency across the 4 training days in the acquisition phase compared to the control. Conversely, rats administered P. harmala 2 weeks after initiating AlCl 3 injection, showed shortness in the escape latency, as compared with the AlCl 3 group signifying improved learning. Additionally, the MWM probe test confirmed the cognitive deficit in AlCl 3 -treated rats to show compromised spatial memory that was expressed as a salient decline in (d) Q (63%) compared to the control group. However, P. harmala effectively ameliorated the AlCl 3 -induced memory impairment, restoring the time spent in the target quadrant to its value in normal control rats.  Section of (a) normal hippocampal CA1 area, shows intact pyramidal neurons (black arrow), while section of (b) AlCl 3 group displays marked neuronal loss with varied areas of darkly pyknotic, shrunken degenerated neurons (blue arrow) with mild edema and vacuolation of neutrophils (dashed arrow). Section of (c) Ph group presents intact histological structures similar to normal rats and photomicrograph of (d) AlCl 3 + Ph group reveals fewer degenerated shrunken pyknotic pyramidal neurons (blue arrow) combined with many intact cells and milder records of glial cell infiltrates (arrow head). In the hippocampal DG area, section of (e) the control group receiving saline or (g) P. harmala shows normal morphological features of hippocampal layers, including granule cells with intact nuclear details (black arrow) along with an unaltered hilar region. The section of (f) AlCl 3  www.nature.com/scientificreports/

Discussion
The current report provides novel insights into the anti-amnestic effects of the methanolic extract of P. harmala seed in a sporadic AD model. The enhanced cognition with reduced Aβ42/tau pathology in AlCl 3 -injected rats coincided with an attenuated hippocampal insulin resistance. P. harmala activated the Akt trajectory to inhibit GSK3β and enhance GLUT4 via two influential molecules, namely, GLP-1 and insulin that mediated pS307-IRS-1 reduction (Fig. 9). The standardized extract also activated Nrf2 to enhance the brain antioxidant capacity that receded oxidative stress to additionally improve insulin sensitivity. Intriguingly, our findings firstly validated that the insulin-signaling dysregulation in the hippocampi of rats exposed to AlCl 3 could be ameliorated after the administration of P. harmala seed extract for 4 weeks. P. harmala commenced its central insulin sensitizing effect by enhancing the hippocampal content of insulin, results that further support a similar effect in the periphery. In a previous work, P. harmala seeds disclosed their hypoglycemic effect via triggering insulin secretion from pancreatic β-cells in a streptozotocin diabetic model 24 .
The enhancing effect of P. harmala on central insulin, recorded in our work, can be owed to the elevated hippocampal content of GLP-1, a hormone that is known to improve insulin signaling 35 . Our finding underpins the results of Gu et al. 36 , who reported that upon screening of plants for their possible dipeptidyl peptidase IV (DPP-4) enzyme inhibitory activity, harmine, a main active constituent of P. harmala, showed an antidiabetic DPP-4 inhibitory effect that in turn increased the endogenous levels of GLP-1 36,37 . It is worth mentioning that both GLP-1 and its receptor are decreased in AD human brain and in experimental AD models, a fact that coincide with the present findings in the current model 38 .
Following the increased hippocampal insulin and GLP-1 contents, P. harmala thrusted the insulin trajectory forward and extended its insulin sensitizing effect by downregulating the protein expression of pS307IRS-1, a leading instigator of brain insulin resistance 39 . Previously, it was reported that the decrement in IRS-1 serine phosphorylation allows the activation of the PI3K cascade 9 ; this verity is confirmed herein, where the inhibition of pS307IRS-1 by P. harmala was allied by an increase in the hippocampal protein expression of the phosphorylated/activated pS473Akt, a downstream of the PI3K molecule. Although the central insulin sensitizing effect of P. harmala has not been reported before, a study conducted in 2016 27 has shown that P. harmala extract mitigated palmitic acid-induced insulin resistance in muscle cells by modulating the p-IRS/p-Akt hub. Moreover, our results are compatible with an earlier study conducted by Naresh et al. 40 , which recounted that treatment of L6 skeletal muscle cells with 4-hydroxypipecolic acid, isolated from P. harmala seeds, stimulated GLUT4 translocation through the PI3K-dependent insulin signaling pathway.
In further support of our current results, the aptitude of P. harmala to increase GLP-1 adds to the activation of insulin cascade, since GLP-1 agonists, such as liraglutide and exendin-4, have activated the IRS-1/Akt trajectory www.nature.com/scientificreports/ in mouse models of AD 41,42 . Notably, GLP-1 mediates its effect by stimulating adenylyl cyclase to consequently trigger the PI3K/Akt signaling pathway 43 . This fact was further verified in other studies; Gupta et al. 44 in their in vitro study revealed that treatment with exendin-4 has phosphorylated elements of the insulin-signaling hub, including Akt1. Moreover, another study showed that GLP-1 improves glucose metabolism through activating Akt 45 . The modulation of p-IRS-1/p-Akt signaling by insulin and GLP-1 was coupled by an increase in the hippocampal content of GLUT4 to form the last step of the pathway and to highlight the beneficial role of P. harmala in improving cerebral insulin sensitivity that was reflected also on the refining of memory function. This notion was documented in a recent study stating that GLUT4 is one of the foremost molecules responsible for hippocampal memory and insulin sensitivity 46 . In addition to its effect on GLUT4, increased levels of GLP-1 play a critical role in improving cognition. An earlier study on GLP-1 receptor knockout mice has revealed memory impairment and cognitive dysfunction in the MWM test 47 to coincide with the present data in AD rats. These findings, hence, emphasize the role of the active insulin cascade along with GLP-1 in the P. harmala-induced upturn in memory and cognition.
A crosstalk between insulin cascade and AD manifestations has been highlighted earlier, where it was identified that Aβ, a surrogate marker of AD pathology, competitively inhibits the binding of insulin to its neuronal receptors 48,49 and interferes with tyrosine phosphorylation/activation of IRS-1 50 , while permitting its serine phosphorylation/inhibition to hinder GLUT 4 translocation 15 . Moreover, post-mortem studies 51,52 have displayed decreased levels of insulin in different AD brain regions; besides, brain insulin deficiency was reported to impede memory, synaptic transmission, IRS-1 tyrosine phosphorylation/activation, and the activation of Akt 53 , events that contribute to AD pathology 54 . In the present work, the treatment with P. harmala for 4 weeks, not only activated the insulin trajectory, but it curbed also the bolstered hippocampal content of Aβ42, as well as the protein expression of the phosphorylated/activated tau, the two main molecules related to AD pathology. Hence, these effects support the folk notion about the effectiveness of P. harmala in easing CNS disorders and offer another mechanism to alleviate AD-like pathology besides augmenting acetylcholine 25,26 . The aptitude of P. harmala to turn on the insulin cascade can explain in part the decreased levels of Aβ42 and p-tau, where increased insulin was reported to foster Aβ clearance and prevent Aβ plaque formation 15,55 . Furthermore, the elevated insulin level triggers the tau phosphatase to deter tau pathology 56 .
Not only the increased insulin, but also GLP-1 plays a central role in lowering both Aβ42 and p-tau as mentioned before 57,58 . These authors displayed that different GLP-1 receptor agonists have inhibited Aβ, tau www.nature.com/scientificreports/ hyperphosphorylation and neuronal damage. Besides, using DPP-4 inhibitors 59,60 or the GLP-1 analogue Liraglutide in AD animal models succeeded to attenuate the amyloid load and improve cognitive abilities and memory 61 . Additionally, the P. harmala-induced pS473-Akt plays a role too, since earlier investigations confirmed that pS473-Akt bestows neuronal protection against AD 62,63 . A fundamental effect of the pS473-Akt is to phosphorylate one of its arsenal downstream molecules glycogen synthase kinase (GSK)-3β at its serine 9 residue 64 , a fact that was recorded here in the P. harmala treated group. The aberrant insulin signaling augments Aβ neuropathology via activating GSK3β 65 , which interferes with the two main steps responsible for amyloidogenesis 65,66 and is considered the main kinase responsible for tau phosphorylation, a step that precedes the emergence of neurofibrillary tangles (NFT) 67 . Mutually, Aβ upregulates GSK3β activity by inhibiting its phosphorylation as evidenced from an earlier in vitro study 67 and here. Hence, the inhibition of the hippocampal GSK-3β adds to the mechanism of the current extract in dwindling the AD pathomolecules. The activation of Akt/ GSK-3β pathway is partly attributed to the increased GLP-1, where its analogs have restored this trajectory and attenuated memory deficits in amyloid precursor peptide/presenilin 1 mice 58,68 .
Although harmine has been proven to effectively enhance spatial learning and memory, as shown here and hitherto 69 , Aβ was not lowered by harmine, but rather by quinazoline derivatives 70 . On the contrary, harmine is known as a potent inhibitor of dual-specificity tyrosine phosphorylation-regulated kinase 1A (DYRK1A), which plays a central role in tau hyperphosphorylation 71,72 to explain partially the current inhibition of p-tau by P. harmala.
From another prospective, Aβ is tied to insulin resistance via inciting oxidative stress that serves as a midpoint mechanism in this trajectory. In this context, high binding protein metals tend to interact with Aβ aggregates 73 to produce reactive oxygen species (ROS) 74 , where hydrogen peroxide (H 2 O 2 ) promotes IRS-1 phosphorylation at serine 307 18 , hence adversely afflicting insulin signaling and aggravating insulin resistance. Meanwhile, as depicted from the well-recognized oxidative stress theory of AD, the neuronal lipid membrane is penetrated by Aβ oligomers to initiate and propagate lipid peroxidation by acting as sulfur free radicals, evoking thus neurotoxicity and synapse deterioration 17 . This fact is considered as an important factor in initiating neurodegeneration in both AD patients 75 and AlCl 3 exposed animals 76 to validate the present Aβ-mediated cognitive decline associated by insulin resistance and oxidative stress. The latter was documented by the increased lipid www.nature.com/scientificreports/ peroxidation with the depletion of the defense molecule GSH and the suppression of the antioxidant transcription factor Nrf2. In our study, P. harmala displayed an antioxidant capacity through bolstering GSH and Nrf2 to attenuate MDA formation. Such results are supported by Li et al. (2018) who stated that harmine and harmaline reversed scopolamine-induced oxidative damage in C57BL/6 mice 77 . Of note, activating Nrf2 has been previously reported to abrogate Aβ production through blunting β-secretase expression 78 and control GSH synthesis being an upstream regulator of glutathione synthetase 79 . Since active GSK3β causes the proteasomal degradation of Nrf2 80 , hence, P. harmala-mediated Akt activation/phosphorylation could be one culprit for GSK3β deactivation to enhance Nrf2, improve cognition, and reduce insulin resistance. To this end, the therapeutic benefit of P. harmala extract against AD-like pathology resides in its central ability to ameliorate hippocampal insulin resistance and oxidative stress through the crossed interaction of insulin/ PI3K/Akt/GLUT4, GLP-1/Akt/GLUT4 and Akt/pGSk3β/Nrf2 cues. The activated pathways acted on tackling Aβ aggregation, tau phosphorylation mainly by inactivating/phosphorylating GSK3β. Accordingly, the novel findings of this study propose a compelling therapeutic mechanism for P. harmala, beyond its traditional use, in the treatment of AD, which needs further clinical proof.