Serum Proteomic Analysis of Cannabis Use Disorder in Male Patients

Cannabis use has been growing recently and it is legally consumed in many countries. Cannabis has a variety of phytochemicals including cannabinoids, which might impair the peripheral systems responses affecting inflammatory and immunological pathways. However, the exact signaling pathways that induce these effects need further understanding. The objective of this study is to investigate the serum proteomic profiling in patients diagnosed with cannabis use disorder (CUD) as compared with healthy control subjects. The novelty of our study is to highlight the differentially changes proteins in the serum of CUD patients. Certain proteins can be targeted in the future to attenuate the toxicological effects of cannabis. Blood samples were collected from 20 male individuals: 10 healthy controls and 10 CUD patients. An untargeted proteomic technique employing two-dimensional difference in gel electrophoresis coupled with mass spectrometry was employed in this study to assess the differentially expressed proteins. The proteomic analysis identified a total of 121 proteins that showed significant changes in protein expression between CUD patients (experimental group) and healthy individuals (control group). For instance, the serum expression of inactive tyrosine protein kinase PEAK1 and tumor necrosis factor alpha-induced protein 3 were increased in CUD group. In contrast, the serum expression of transthyretin and serotransferrin were reduced in CUD group. Among these proteins, 55 proteins were significantly upregulated and 66 proteins significantly downregulated in CUD patients as compared with healthy control group. Ingenuity pathway analysis (IPA) found that these differentially expressed proteins are linked to p38MAPK, interleukin 12 complex, nuclear factor-κB, and other signaling pathways. Our work indicates that the differentially expressed serum proteins between CUD and control groups are correlated to liver X receptor/retinoid X receptor (RXR), farnesoid X receptor/RXR activation, and acute phase response signaling.


Introduction
Cannabis sativa L. and Cannabis indica L. contain a variety of secondary metabolites. Cannabis plants species differ based on many factors, including the quantity of cannabinoids. Some of them are psychoactive and induce hallucinating effects such as delta-9tetrahydrocannabinol (THC) while the others are non-psychoactive such as cannabidiol (CBD) [1]. The complexity of cannabis makes its use censorious because cannabis users may develop unexpected side effects, including central nervous system (CNS) side effects due to certain chemical ingredients. Therefore, exposure to cannabis means that the users will expose to several cannabinoids (~60) that are associated with pharmacological effects. In addition, the duration of exposure is a critical factor that significantly affects the quantity of these cannabinoids in the body.
Despite the presence of multiple compounds, cannabinoids family are the most abundant phytochemicals present in the cannabis plants [1]. Psychoactive effects resulted from cannabis exposure have been linked to THC [2]. This compound can modulate the processing of visual and auditory hallucination effects [3]. However, regulatory agencies have approved few cannabinoids to be used for certain indications. For instance, the U.S. Food and Drug Administration (FDA) has approved products containing CBD for seizures associated with Dravet syndrome and Lennox-Gastaut syndrome in one-year-old and older patients reported previously in clinical studies [4,5]. Moreover, FDA-approved synthetic products containing THC for the treatment of vomiting and nausea caused by chemotherapy treatments in patients who lack the response to conventional antiemetic treatments [6]. In addition, they can be prescribed to manage anorexia-associated with weight loss in patients diagnosed with acquired immunodeficiency syndrome [7].
Cannabis use disorder (CUD) is widespread across numerous countries [8]. The hallucination effects of cannabis use leads to drug abuse [9]. Governments and regulatory agencies set guidelines and policies to minimize the undesirable effects of cannabis [10,11]. As some countries have legalized the use of marijuana, smoking products containing marijuana are legally marketed nowadays.
Studies have documented toxicological effects in different models exposed to cannabis ingredients [12][13][14][15][16]. A recent study reported that high-grade atrioventricular block was developed in a young male following chronic exposure to marijuana [14]. Moreover, a recent case series study concluded that vaping cannabis oil was associated with acute respiratory depression [15]. It is important to consider that tachycardia and neurotoxicity were reported after acute inhalation of cannabis in humans [16]. Fivefold increase in blood carboxyhemoglobin levels were found in subjects who smoked marijuana for at least five years as compared with those who smoke tobacco cigarettes [13]. This study also noted that the burden of tar and carbon monoxide in the respiratory system has increased in marijuana smokers as compared with those who smoked a similar quantity of tobacco. On the other hand, CBD showed the ability to regulate immunological responses using in vivo and in vitro assays [17]. Moreover, CBD exhibits antioxidant and anti-inflammatory properties [18][19][20]. These anti-inflammatory effects were also found in non-psychoactive cannabinoids [20].
Several reports have determined the serum proteomic profiling of humans exposed to amphetamine analogs [21,22]. A recent study from our group identified differentially expressed proteins in the serum of individuals with amphetamine use disorder compared with a healthy control group [23]. Moreover, prior clinical proteomic studies utilized serum samples to determine the levels of proteins in patients who had developed neurodegenerative diseases [24,25], neurodevelopmental disorders [26][27][28], major depressive disorder, and bipolar disorders [29,30]. In the present study, we investigate the expression changes of CUD patients' serum proteins compared with healthy controls, using an untargeted proteomic approach employing two-dimensional (2D) alteration in gel electrophoresis (2D-DIGE) coupled with mass spectroscopy (MS).

Demographic Information
Demographic and clinical information of all participants included in our study has been collected. This information includes marital and employment status, age, gender, history of cannabis use disorder, and route of cannabis administration (Table 1). Table 1. Clinical and demographic information of the CUD and control groups recruited in the present study. HIV: Human immunodeficiency virus; HPC: Hepatitis C virus; TB: Tuberculosis.

CUD Group Control Group
Number of patients 10

Identification of Differentially Expressed Proteins and 2D-DIGE Analysis
The current study assessed the difference in protein expression among 10 cannabisexposed individuals and 10 controls (20 samples from 10 gels) using 2D-DIGE analysis technique before statistical analysis is performed with Progenesis software. Fluorescent protein profiles of a 2D-DIGE of control samples labelled with Cy3 are presented in Figure 1A. The CUD samples were labeled with Cy5 ( Figure 1B), pooled internal control labeled with Cy2 ( Figure 1C), and overlap of 2D-DIGE gels of samples labeled with Cy3/Cy5 ( Figure 1D). A total of 1700 spots were identified on the gels, 156 were significantly different (ANOVA, p ≤ 0.05; fold-change ≥ 1.5) between the CUD and control groups ( Figure 2). For alignment and further analysis, the spot patterns were reproducible across all 10 gels. The internal standard Cy2-labeled was included to perform normalization among the whole gels set in addition to the quantitative of the protein levels differential analysis. A total of 156 spots displayed a statistical significance among the two groups. These spots were manually excised from the preparative gel and underwent protein identification using MS.

Figure 2.
Fluorescence labeled (CyDyes)-2D-DIGE numbered spots indicate those proteins that were identified to be differentially abundant (defined as fold-change >1.5, p < 0.05) between the two groups (controls and CUD). These were successfully identified with matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry (MS). MW, protein molecular weight; pI, isoelectric point.  Figure 3. Differential expression of statistically significant protein spots from controls and cannabis-exposed samples.

Principal Component Analysis
To determine and visualize the CUD and control subjects' samples, the principal component analysis of the Progenesis SameSpots software was used. The analysis was made on all 121 spots that exhibited statistically significant changes in abundance identified by MS. The analysis shows that the two groups clustered distinctly based on different proteins with score of 64% ( Figure 4). . Differential expression of statistically significant protein spots from controls and cannabisexposed samples.

Principal Component Analysis
To determine and visualize the CUD and control subjects' samples, the principal component analysis of the Progenesis SameSpots software was used. The analysis was made on all 121 spots that exhibited statistically significant changes in abundance identified by MS. The analysis shows that the two groups clustered distinctly based on different proteins with score of 64% ( Figure 4).

Protein-Protein Interaction Networks
Using Ingenuity Pathway Analysis (IPA), the protein-protein interaction analysis was completed for all 121 regulated proteins. The analysis demonstrated that 35 proteins interacted directly/indirectly via protein networks ( Figure 5A). The software calculates the best fit score obtained from the input data set of proteins and the biological functions database in order to generate a protein-protein interactions network. The generated network is favorably enriched for proteins with extensive and specific interactions. The interacting proteins are characterized as nodes and their biological relationships as a line. Based on the resulted data, four interaction networks were recognized for the proteins exhibiting variance expression profiles. The highest scoring network (score = 52) ( Figure  5, Supplementary Figure S1) incorporated 25 proteins. The proposed highest interaction network pathway was related to free radical scavenging, cellular compromise, and inflammatory response. Alone the top pathways are presented ( Figure 5A). Canonical pathways that enriched in current dataset are presented in Figure 5B. The canonical pathways are sorted down to decreasing log (p-value) of enrichment. The most interesting enriched canonical pathways included liver X receptors/retinoid X receptor (LXR/RXR) activation (11% overlap, p-value: 3.7 × 10 −16 ), farnesoid X receptors/retinoid X receptor activation (10.7% overlap, p-value: 5.78 × 10 −16 ), acute phase response signaling (7.4% overlap, pvalue: 7.15 × 10 −14 ), atherosclerosis signaling (5.6% overlap, p-value: 3.87 × 10 −7 ), and production of nitric oxide (NO) and reactive oxygen species (ROS) in macrophages (4.3% overlap, p-value: 4.24 × 10 −7 ). More details about the identified canonical pathways are shown in supplementary files (Supplementary Figure S1).

Protein-Protein Interaction Networks
Using Ingenuity Pathway Analysis (IPA), the protein-protein interaction analysis was completed for all 121 regulated proteins. The analysis demonstrated that 35 proteins interacted directly/indirectly via protein networks ( Figure 5A). The software calculates the best fit score obtained from the input data set of proteins and the biological functions database in order to generate a protein-protein interactions network. The generated network is favorably enriched for proteins with extensive and specific interactions. The interacting proteins are characterized as nodes and their biological relationships as a line. Based on the resulted data, four interaction networks were recognized for the proteins exhibiting variance expression profiles. The highest scoring network (score = 52) ( Figure 5, Supplementary Figure S1) incorporated 25 proteins. The proposed highest interaction network pathway was related to free radical scavenging, cellular compromise, and inflammatory response. Alone the top pathways are presented ( Figure 5A). Canonical pathways that enriched in current dataset are presented in Figure 5B. The canonical pathways are sorted down to decreasing log (p-value) of enrichment. The most interesting enriched canonical pathways included liver X receptors/retinoid X receptor (LXR/RXR) activation (11% overlap, p-value: 3.7 × 10 −16 ), farnesoid X receptors/retinoid X receptor activation (10.7% overlap, p-value: 5.78 × 10 −16 ), acute phase response signaling (7.4% overlap, p-value: 7.15 × 10 −14 ), atherosclerosis signaling (5.6% overlap, p-value: 3.87 × 10 −7 ), and production of nitric oxide (NO) and reactive oxygen species (ROS) in macrophages (4.3% overlap, p-value

Subcellular and Functional Characterization of the Differentially Expressed Proteins
Following MS analysis, all 121 identified proteins between the CUD and control samples were subjected to the PANTHER classification system (http://www.pantherdb. org, accessed on 1 February 2021). The classification was performed according to their molecular function ( Figure 6A), biological process ( Figure 6B), and cellular component ( Figure 6C). The main functional categories recognized were binding proteins (47%), catalytic activity (30%), and molecular function regulatory proteins (21%). Further, the identified proteins were located in the organelle region (35%), extracellular space (24%), followed by cytoplasmic and cytoskeletal regions, and each of these two account for (19%). The majority of the identified protein was involved in cellular process, metabolic process and biological regulations.
Following MS analysis, all 121 identified proteins between the CUD and control samples were subjected to the PANTHER classification system (http://www.pantherdb.org, accessed on 1 February 2021). The classification was performed according to their molecular function ( Figure 6A), biological process ( Figure 6B), and cellular component ( Figure  6C). The main functional categories recognized were binding proteins (47%), catalytic activity (30%), and molecular function regulatory proteins (21%). Further, the identified proteins were located in the organelle region (35%), extracellular space (24%), followed by cytoplasmic and cytoskeletal regions, and each of these two account for (19%). The majority of the identified protein was involved in cellular process, metabolic process and biological regulations.

Immunoblotting Confirmation of Changes in Selected Proteins
Immunoblot assay confirmed the expression of the selected proteins that were differentially abundant by 2D-DIGE analysis (Figure 7). The proteins selected for confirmation were serotransferrin and retinol-binding protein 4. Immunoblots revealed that the serum protein expression of serotransferrin and retinol-binding protein 4 were decreased and increased, respectively, in CUD group as compared with control group (p ≤ 0.05). To normalize the immunoblot data, β-actin was used in the present study as a housekeeping protein ( Figure 7A,B).

Immunoblotting Confirmation of Changes in Selected Proteins
Immunoblot assay confirmed the expression of the selected proteins that were differentially abundant by 2D-DIGE analysis (Figure 7). The proteins selected for confirmation were serotransferrin and retinol-binding protein 4. Immunoblots revealed that the serum protein expression of serotransferrin and retinol-binding protein 4 were decreased and increased, respectively, in CUD group as compared with control group (p ≤ 0.05). To normalize the immunoblot data, β-actin was used in the present study as a housekeeping protein ( Figure 7A,B).

Ingenuity Pathway Analysis
3.1.1. LXR/RXR Activation IPA analysis showed that LXR/RXR is activated in humans chronically exposed to cannabis. Our prior proteomic study showed that LXR/RXR activation is observed in humans exposed to amphetamine for chronic period of time [23]. This indicates that LXR/RXR activation is highly sensitive following exposure to amphetamine and cannabinoids. A microarray study performed IPA analysis showing that THC could induce alterations on the genes, including the LXR/RXR gene, highly affected by lipopolysaccharide in BV-2 microglia cells [31]. This may suggest that LXR/RXR is highly sensitive to THC. Importantly, LXR/RXR activation is linked to signaling pathways, including apolipoproteins such as apolipoprotein AI, for cholesterol metabolism [32]. In the present study, we reported that the serum expression of apolipoprotein AI was increased in patients diagnosed with CUD as compared to healthy control group. This suggests that cannabis-modulated apolipoprotein AI and LXR/RXR may be involved in the metabolism of cholesterol. In addition, LXR activation plays a crucial role in the inhibition of inflammatory responses [33], indicating that LXR/RXR is one of the pathways mediated by cannabinoids to inhibit the formation of inflammatory reactions. For instance, LXR/RXR, PPARα/RXRα, and STAT3 signaling pathways are essential pathways to inhibit the inflammatory reactions [34]. A prior study demonstrated that LXR could play a significant role in induction protective effects against immunological responses-induced by Mycobacterium tuberculosis in mice [35]. Therefore, LXR/RXR activation is an efficient therapeutic

LXR/RXR Activation
IPA analysis showed that LXR/RXR is activated in humans chronically exposed to cannabis. Our prior proteomic study showed that LXR/RXR activation is observed in humans exposed to amphetamine for chronic period of time [23]. This indicates that LXR/RXR activation is highly sensitive following exposure to amphetamine and cannabinoids. A microarray study performed IPA analysis showing that THC could induce alterations on the genes, including the LXR/RXR gene, highly affected by lipopolysaccharide in BV-2 microglia cells [31]. This may suggest that LXR/RXR is highly sensitive to THC. Importantly, LXR/RXR activation is linked to signaling pathways, including apolipoproteins such as apolipoprotein AI, for cholesterol metabolism [32]. In the present study, we reported that the serum expression of apolipoprotein AI was increased in patients diagnosed with CUD as compared to healthy control group. This suggests that cannabis-modulated apolipoprotein AI and LXR/RXR may be involved in the metabolism of cholesterol. In addition, LXR activation plays a crucial role in the inhibition of inflammatory responses [33], indicating that LXR/RXR is one of the pathways mediated by cannabinoids to inhibit the formation of inflammatory reactions. For instance, LXR/RXR, PPARα/RXRα, and STAT3 signaling pathways are essential pathways to inhibit the inflammatory reactions [34]. A prior study demonstrated that LXR could play a significant role in induction protective effects against immunological responses-induced by Mycobacterium tuberculosis in mice [35]. Therefore, LXR/RXR activation is an efficient therapeutic target to modulate cholesterol metabolism, transport and absorption, inflammatory responses, and immunological reactions. Studies are warranted to explore the beneficial effects of targeting LXR/RXR by cannabinoids to modulate the inflammation, immunological reactions, and cholesterol transport. Further research may investigate the effects of cannabinoids on the diseases through acting on LXR/RXR.

FXR/RXR Activation
IPA analysis showed that FXR/RXR is activated in humans chronically exposed to cannabis. This is in agreement with our previous work showing that FXR/RXR activation is documented in patients diagnosed with amphetamine use disorder [23]. Therefore, FXR/RXR activation can play a critical role in the toxicological effects of abused drugs such as amphetamine and cannabis. The FXR is a bile acid binding site and has a role in the metabolism of lipids and glucose [36]. Our findings reported upregulatory effects on the serum expression of apolipoprotein AI in CUD patients as compared to healthy control group. A previous work highlighted that FXR activation might be a potential strategy for the treatment of hypertriglyceridemia and type 2 diabetes mellitus (T2DM) [37]. This study demonstrated that activation of FXR was associated with reduce plasma concentrations of triglyceride, fasting glucose, and insulin in T2DM rat models. In addition, treatment with a FXR agonist, chenodeoxycholic acid, could reverse the reduction of FXR expression in the liver of T2DM rat models. These findings were supported by another study reporting that hyperlipidemia and hyperglycemia were improved following activation of FXR in diabetic mice models [38]. It is highly recommended to explore the role of cannabinoids on modulating hyperglycemia and hyperlipidemia through modulating FXR/RXR pathways in humans. Further work should study the effects of cannabinoids on the diseases through acting on FXR/RXR.

Acute Phase Response Signaling
In our IPA analysis, we found that acute phase response signaling is stimulated in cannabis users. The acute phase proteins, including haptoglobin, alpha-1-antitrypsin, and complement factors, are proteins that are changed in response to the inflammatory cytokines [39,40]. Acute phase responses are correlated to various diseases, including immunological diseases [41]. The acute phase response signaling was the most interesting enriched canonical pathway involved in patients with amphetamine use disorder as shown in our previous proteomic study [23]. Interestingly, THC exposure was found to increase the mortality in mice infected with Legionella pneumophila at least in part by altering the acute phase responses of proinflammatory cytokines, an effect was not observed with cannabinol and cannabidiol as well as a synthetic cannabinoid, CP 55,940 [42]. This suggests that psychoactive cannabinoids might be more likely to modulate the acute phase responses of inflammatory biomarkers. However, cannabidiol and its synthetic analogs have been reported to exert anti-inflammatory and antioxidant effects [18]. CBD is a negative allosteric modulator of cannabinoid receptor 1 (CB1) [43]; moreover, CBD was reported to behave inverse agonist properties to CB2 receptor indicating that CBD-mediated anti-inflammatory effects through modulating CB1 and 2 receptors [44]. The anti-inflammatory properties were also documented with psychoactive cannabinoids [20]. Therefore, it is critical to elucidate the role of psychoactive and non-psychoactive compounds in modulating acute phase proteins.

Atherosclerosis Signaling
In our study, atherosclerosis signaling has been found to be modulated in CUD patients. Atherosclerosis signaling was one of the top canonical pathways that are involved in protein-protein interactions in amphetamine use disorder patients [23]. Note that apolipoprotein AI and inflammatory pathways interact with atherosclerosis signaling [45]. A previous review work discussed that THC might attenuate the plaques, generated from atherosclerosis, through modulating CB2 receptors [46]. Additionally, activation of CB1 re-ceptors in the brain might be a therapeutic strategy to prevent ischemic stroke. Importantly, 2-arachidonoylglycerol (2-AG) and palmitoylethanolamide (PEA) are endocannabinoids that were found to attenuate the acute complication of atherosclerosis such as myocardial ischaemia in isolated rat hearts [47]. These effects were determined by measuring the activities of cardiac creatine kinase (CK) and lactate dehydrogenase. The beneficial effects of 2-AG and PEA on myocardial ischemia were abolished following exposure to a CB2 receptor antagonist, SR144528, indicating that the endocannabinoids might play a vital role in preventing the acute complications of atherosclerosis through acting on CB2 receptors. The significant role of CB2 receptors in preventing the myocardial ischemia was further supported by a study showing that a CB1/2 receptors agonist (WIN55212) was able to reduce the infraction size, an effect abolished with a selective CB2 receptor antagonist (AM630) but not with a selective CB1 antagonist (AM251). Moreover, treatment with a low dose of THC could reduce the progression of atherosclerosis in apolipoprotein E knockout mouse model [48]. This effect was abolished following exposure to a CB2 receptor antagonist. However, smoking cannabis with uncontrolled quantity of cannabinoids may induce adverse effects on the cardiovascular system [49]. These findings provide information about the potential therapeutic values of using cannabinoids for the treatments of atherosclerosis-related diseases.

Production of Nitric Oxide (NO) and Reactive Oxygen Species (ROS) in Macrophages
In our IPA analysis, we reported that NO and ROS production is significantly changed in CUD patients. Studies found that both ROS pathways/production are modulated following exposure to abused drugs [18,50,51]. These data are in agreement with our proteomic work demonstrating that production NO and ROS in macrophages was one of the top canonical pathways that are involved in protein-protein interactions in amphetamine use disorder patients [23]. Importantly, both CB1 and CB2 receptors are key proteins in regulating the productions of ROS and inflammatory cytokines by macrophages [52]. Interestingly, it was shown that activation of CB1 receptor mediated proinflammatory responses by macrophages via increased ROS production in part by inducing p38-mitogen-activated protein kinase phosphorylation [52]. This effect was attenuated through activating Ras-related protein 1, an effect mediated by CB2 receptor pathway. It is found that CBD exhibited potential antioxidant effects through direct and indirect pathways [18]. These pathways include modulating proteins such as CB1 and 2 receptors, antioxidant enzymes, adenosine A 2A receptors, and other proteins. Therefore, it is recommended to test the effectiveness of CBD against many diseases associated with oxidative stress. CBD has better safety profile against psychoactive cannabinoids such as THC. Additionally, both THC and CBD could induce neuroprotective effects due to their potential antioxidant properties [53]. Alternatively, a synthetic cannabinoid exposure showed the ability to attenuate the production of NO in chondrocytes treated with IL-1 [54]. This study was supported by another study showing that a synthetic cannabinoid inhibited lipopolysaccharide-induced NO release in macrophages, an effect mediated by CB2 receptor pathway [55]. Cannabichromene, a cannabinoid TRPA1, reduced NO production by macrophages and attenuated colitis of murine [56]. Moreover, THC and CBD attenuated NO production in macrophages exposed to lipopolysaccharide, and the study found that THC was more potent than CBD in reducing the NO production [57]. Furthermore, THC exhibited ability to attenuate the gene expression of inducible NO synthase enzyme through modulating nuclear factor-κB (NF-κB) pathway in macrophages treated with lipopolysaccharide [58]. Taken together, cannabinoids might be potential compounds in modulating NO and ROS production by macrophages in various diseases.

Selected Proteins
Our work provides clinical understanding about the serum proteomic profiling in patients diagnosed with CUD. We found that there are significant alterations in the serum proteins expression and these proteins have been found to be essential in inflammations, protein binding, acute phase reactions, metabolic pathways, and other pathways. Moreover, these proteins have been found to be involved in oxidative stress, thyroid diseases, Alzheimer's diseases, and lipid disorders. Our data indicate that cellular processes and cellular anatomy are highly affected by cannabis. For this study discussion, we selected proteins that either highly significant altered (p values) or changed at different locations in CUD patients as compared with control group.

Albumin
Our study investigated the serum expression of proteins that are highly involved in drug binding, including albumin [59]. Our work revealed that the serum expression of albumin was decreased in CUD patient as compared with control group. As albumin occupies high amount human serum proteins, the expression level of this protein is critical in patients who have developed other diseases and taken certain drugs. An important note that certain antipsychotic, antihypertensive, antiepileptic, antidepressant, antibiotic, and other classes of drugs have high protein binding properties and have narrow therapeutic windows [60]. Therefore, comorbidity of CUD with other diseases/disorders may result in toxicological effects of drugs that were used to treat these diseases or disorders. It is recommended here that CUD patients should be carefully monitored when they take other medicines. Drug-drug interaction between cannabinoids and other classes of drugs has been previously reported [61,62].

Haptoglobin
Regarding the serum haptoglobin level, our study reported a controversial result regarding the circulatory serum levels of haptoglobin in the CUD patients as compared with the control group. It is noteworthy that there is a correlation between haptoglobin and the inflammation process [63,64]. Cannabinoids were found to exert ant-inflammatory effects in animal models [65,66]. Haptoglobin was found to exert a protective effect against oxidative stress induced by an increase in the level of the hemoglobin in pre-clinical models [67]. Moreover, a proportional correlation was observed between haptoglobin and the levels of inflammatory cytokines [68]. In our study, we demonstrated that haptoglobin serum expression was upregulated in some locations and downregulated in other locations in the CUD group as compared with the control group. This differential expression of haptoglobin may result from post-translational modifications, cleavage by enzymes, or different protein species presence. Importantly, THC has been reported to induce oxidative stress, an effect associated with decreased antioxidant parameters [69]. However, CBD was found to produce antioxidant effect in neuronal cells [70,71]. Notably, exposure to C. sativa for 30 days resulted in a reduction in the total antioxidant capacity, an effect associated with an increase in the levels reactive oxygen species in male albino rats [72]. More research is required to explore the role of serum haptoglobin level in humans exposed to cannabis and its applications in medical sciences.

Apolipoprotein A-I
Our study found that serum circulatory levels of apolipoprotein A-I is highly abundant in CUD patients as compared with control group. Note that cholesterol levels and lipid metabolism are highly regulated by apolipoprotein A-I. A study reported that Apolipoprotein A-I interacted with high density lipoprotein particles [73]. Gene therapy using apolipoprotein A-I was shown to induce protective effects against lipid disorders [74]. A prior study found that cannabis exposure is associated with weight loss and reduced body mass index in humans [75]. Importantly, it was also found that cannabinoids could induce anorexia in part through acting on cannabinoid receptors [76]. Therefore, cannabis use may lead to an improvement in lipid metabolism and anorexia effects. These effects provide hope to develop novel therapeutic agents from C. sativa for the treatment of the lipid diseases. This is in an agreement with previous studies showing that cannabinoids improved heart diseases [77] suggesting the involvement of apolipoprotein A-I in this effect. Alternatively, prior studies found that apolipoprotein A-I had ability to reduce the beta amyloid accumulation [78,79]. This effect may lead to beneficial consequences against Alzheimer's disease. It is critical to mention here that THC and CBD might be potential compounds for prevention and treatment Alzheimer's disease symptoms [80]. Moreover, low levels of apolipoprotein A-I in the serum was observed in schizophrenic patients [81]. CBD may have therapeutic effects against schizophrenia [82]. However, studies found that THC exposure was associated with psychosis and schizophrenia [83,84]. The pharmacological effects of cannabis constituents against schizophrenia and Alzheimer's disease should be further investigated.

Type I and Type II Keratins
Keratin is an essential component that is involved in the epithelial lining. Keratins have protective functions and provide structure to the epithelium [85]. The keratin is a protein that is a fibrous structure and localized in nails, hair, epithelial cells of the skin outer layer, and others [85]. Our study revealed that type I and type II keratins are highly abundant proteins in the serum of CUD compared with control group. Studies found that keratins are essential proteins in cell growth, differentiation, and proliferation [86][87][88]. In addition, they provide mechanical integrity as protection against external stress [89]. They also have cycloprotection properties against non-mechanical stresses [90]. These proteins have an additional role in the digestive system [91]. It is critical to figure out which constituents (cannabinoids vs. non-cannabinoids) in C. sativa are responsible for increasing both types (type I and type II) of keratins in the circulatory system in individuals who are chronically exposed to cannabis.

Serotransferrin
Serotransferrin is a critical protein to transport the iron from absorption sites or heme degradation to tissues for storage or utilization [92]. In our current study, we showed a downregulation of the serotransferrin serum expression in the of CUD patients as compared with control group. Additionally, serotransferrin expression was decreased in the urine of cannabis users [93], suggesting that cannabis users have downregulation in serotransferrin in the serum and urine. It is important to consider that a previous proteomic study found that serotransferrin was decreased in the lungs of smokers compared with control group [94]. However, the serum serotransferrin was found to be reduced in amphetamine use disorder patients [23]. This suggests that chronic exposure to abused drugs induces dysregulation in iron and heme balance.

Transthyretin
Transthyretin is an important protein to transport the thyroid hormone, thyroxine, and retinol-binding protein [95]. In this study, we found that there is an upregulation of the serum transthyretin expression in the of CUD patients as compared with control group. In addition, previous studies found that the accumulation or mutations of transthyretin was associated with amyloid diseases such as senile familial amyloid polyneuropathy, systemic amyloidosis, and familial amyloid cardiomyopathy [95]. However, transthyretin showed ability to bind to beta amyloid attenuating beta amyloid aggregation [96,97], which has beneficial consequences against Alzheimer's disease. Therefore, it is recommended to explore the role of cannabis constituents in modulating transthyretin as a potential biomarker that is involved in many diseases.

Tumor Necrosis Factor Alpha-Induced Protein 3
Tumor necrosis factor alpha-induced protein 3 (TNFAIP3) was found to regulate the activity of NF-κB, especially through the receptor of TNF-alpha [27]. Importantly, NF-κB was found to be increased in the states of inflammation and activated immune cells [98]. Moreover, TNFAIP3 was found to be a negative feedback mechanism for NF-κB activation [99]. Note that a reduction in the expression of TNFAIP3 is a predictor for inflammation and increased NF-κB expression [100]. For instance, it has been found that the gene expression of TNFAIP3 was decreased in peripheral blood mononuclear cells of rheumatoid arthritis patients as compared with healthy control [101]. Moreover, the gene expression of TNFAIP3 was also reduced in peripheral blood mononuclear cells of patients with psoriasis vulgaris [102]. This study found that the gene expression level of TNFAIP3 was negatively linked to the severity of the disease. However, increased TNFAIP3 expression was linked to low survival rate of esophageal squamous cell carcinoma in a noncancerous esophageal cell line [103]. We reported here that TNFAIP3 was less abundant protein in the circulatory serum of CUD patients as compared with control group.

Inactive Tyrosine Protein Kinase PEAK1
Inactive tyrosine protein kinase PEAK1 is an encoded gene and involved in the cellular response inside the cells following the activation of tyrosine kinase receptor [104]. We here found that inactive tyrosine protein kinase PEAK1 is less abundant in CUD patients as compared with healthy controls. Importantly, increased PEAK1 has been linked to the progression and metastasis of breast cancer [104,105]. We suggest that further investigation of the modulatory role of cannabinoids in the progression of breast cancer is required. PEAK1 has been found to regulate the responses of transforming growth Factor β in breast cancer models [104]. Moreover, overexpression of inactive tyrosine protein kinase PEAK1 can modulate anus kinase-2 and extracellular signal-regulated kinase-1/2, which was involved metastasis of tumors in lung cancers [106]. Previous and our findings suggest that future directions may explore the potential role of cannabis plants in cancer research focusing on the PEAK1 pathways.

Ethical Approval and Participate Consent
The IRB committees at the Eradah Complex for Mental Health (Riyadh, Saudi Arabia) and College of Medicine-King Saud University reviewed and approved all procedures and protocols. This study was performed according to the rules of the Declaration of Helsinki 1975 and later amendments. The written consents were acquired from all individuals involved in the study. This study was conducted at the Proteomics Unit, Obesity Research Center, College of Medicine and King Khalid University Hospital, Medical City, King Saud University, Riyadh, Saudi Arabia.

Study Design and Selection Criteria
Twenty male subjects were involved in two different groups in this study: CUD and healthy control. Ten subjects diagnosed with CUD (age of 30.4 ± 4.36 years) were enrolled at the Eradah Complex for Mental Health (Riyadh, Saudi Arabia) and compared with a control group containing 10 healthy individuals (age of 24.7 ± 3.63 years). To perform the power analysis and determine the least possible number of required biological replicates, we used Progenesis SameSpots software (Nonlinear Dynamics, Newcastle, UK). The diagnosis for CUD was performed according to the Diagnostic and Statistical Manual of Mental Disorders guidelines (DSM-5) [107]. The included participants have no history of blood disorders, diabetes mellitus, obesity, psychosis, renal diseases, or any other infectious diseases. Clinical and demographic information is shown in Table 1. The control group demographic information and gel scanning are used from our previous study [23] with respect of experimental procedures and timing. The CUD group tested positive for cannabinoids only without any detection of other abused drugs at the time of blood collection. Blood samples were collected, centrifuged for ten minutes at 1000× g. The resulting serum samples were aliquoted and stored at −80 • C for proteomic analysis.

Serum Protein Extraction
Proteins were extracted from the serum samples via centrifugation (5 min, 12,000× g) as described previously [23]. The depletion of high-abundance serum proteins (i.e., albu-min, IgG) was achieved using Depletion SpinTrap for Albumin and IgG (GE Healthcare, Chicago, IL, USA) following the manufacturer's instructions. Further, the remaining proteins were extracted by the TCA/acetone method [108]. The depleted samples were mixed with ice-cold acetone containing 10% w/v TCA (1:4), and the mixture was vortexed for 15 s to ensure uniform mixing. Next, the mixture was incubated overnight at −20 • C for protein precipitation. After incubation, the tubes were centrifuged for 15 min at 4 • C at a speed of 12,000× g, and the pellet was solubilized in labeling buffer (7 M of urea, 2 M of thiourea, 30 mM of Tris-HCl, 4% CHAPS, pH 8.5). After that, the concentration of protein samples was determined in triplicate employing the 2D-Quant Kit (GE Healthcare, Chicago, IL, USA).

Fluorescence Labeling of Samples with CyDyes and 2-Dimensional Difference in Gel Electrophoresis (2D-DIGE)
Fifty micrograms of protein from each sample of both the CUD and control groups was labeled with 400 pmol of Cy3 and Cy5 dyes. Then, the internal standard was prepared by mixing an equal amount of all samples after pooling and labelling with Cy2. A dye swapping strategy was employed during labelling in order to avoid any dye-specific bias (Supplementary Table S1). 1st-dimension analytical GE followed by 2nd-dimension sodium dodecyl sulfate (SDS)-polyacrylamide GE (SDS-PAGE) were implemented on 12.5% fixed gels as described in previous studies [23,109]. Further, the 2D-DIGE gels were scanned using the Typhoon 9400 scanner (GE Healthcare, Chicago, IL, USA) where specific excitation/emission wavelengths were used (488/520 nm) for Cy2, (532/580 nm) for Cy3, and (633/670 nm) for Cy5.

Statistical Analysis
Progenesis SameSpots software (v2.0, Nonlinear Dynamics, Newcastle, UK) was used to analyze the 2D-DIGE gel images. The image analysis was done by an automated spot detection and comparison method between the samples of CUD and control groups. Although the automatic analysis was completed to detect all the spots across all the 10 gels, each selected spot was manually edited and verified wherever necessary. The differentially expressed spots were identified by normalized volumes. The normalized volume of each spot on each gel was calculated from Cy3 (or Cy5) to Cy2 spot volume ratio using the software. To generate normal distributed data, log transformation of the spot volumes was done by the software. To calculate statistically significant differences between the two groups, one-way ANOVA was used and p < 0.05 was considered statistically significant. A cut-off ratio ≥1.5-fold was considered significant. A pre-filtration and manual check have been done on all spots before testing the statistical differences. In statistical analysis, the normalized spot volumes were applied instead of intensities of the spots. Any spots fulfil the above statistical criteria was analyzed by MS.

Protein Identification with Mass Spectrometry
Coomassie-stained gel spots from a preparatory gel were washed then digested according to methods described previously [23,109,110]. To describe briefly, total protein (1 mg) was obtained from a pool of equal protein amounts of the 20 serum samples (10 CUD and 10 control). This sample was denatured in lysis buffer and then mixed in a rehydration buffer. Then, the proteins samples were separated by first and second dimensions with the same conditions in the DIGE section. Then, the gels were fixed in 40% (v/v) ethanol containing 10% (v/v) acetic acid (overnight) and then washed (3×, 30 min each, ddH 2 O). The gels were incubated (1 h, 34% (v/v) CH3OH containing 17% (w/v) ammonium sulphate and 3%(v/v) phosphoric acid) prior to the addition of 0.5 g/L Coomassie G-250. After 5 days, the stained gels were briefly rinsed with Milli-Q water and stored until the spots could be picked and identified by MS. Digestion was performed by adding 15 µL of (20 ng ice-cold trypsin solution in 25 mM NH 4 HCO 3 , 5 mL CH 3 CN, 5 mL distilled water) and incubated 20 min at 4 • C, and digestion continued overnight at 37 • C. To extract the peptides, 1 µL of 1% Trifluoracetic acid was added on the gel pieces and placed in vortex incubator for mass spectrometric analysis (1 h, 400 rpm, 25 • C).

Network Pathway and Functional Analysis
The IPA Software program (Version: 42012434, Ingenuity Systems, Redwood City, CA, USA, http://www.ingenuity.com, accessed on 2 February 2021) was used to analyze the identified proteins and to annotate them with related functions and pathways. The annotations involved overlaying the proteins with their most significant networks and biochemical pathways based on previous publications on the proteins. The identified proteins were classified into different categories according to their biological process, cellular components, and molecular function using protein analysis through evolutionary relationships (PANTHER) classification system (http://www.pantherdb.org, accessed 1 February 2021).

Immunoblotting
Immunoblotting assay was performed in the current study to further confirm the findings of the proteomic study. Two differential abundance proteins with statistically significant were chosen and determined by immunoblotting. Primary monoclonal antibodies against transferrin (mouse, cat # SC-365871), retinol-binding protein (RBP, mouse, cat # SC-69795), and β-actin (goat, N-18, cat # SC-1616) were bought from Santa Cruz Biotechnology (Santa Cruz, TX, USA). One-dimensional discontinuous slab gel electrophoresis (12% sodium dodecyl sulfate (SDS)-polyacrylamide gel) was used to separate an equal amount of protein from each sample (50 µg). A mini trans-blot electrotransfer cell (BioRad, California, CA, USA) was employed to transfer proteins from the run gels to an Immobilon-P, polyvinylidene difluoride (PVDF) transfer membrane (Millipore, Massachusetts, MA, USA) To test the efficiency of the transfer, the membranes were stained with Ponceau-S. Subsequently, the membranes were blocked with tris-buffered saline (TBS)-containing 5% fat-free milk (FFM), for one hour at room temperature, and then the membranes were rinsed three times with TBS-T in 10 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20 buffer. After rinsing, the membranes were incubated with the selected primary antibodies at dilution of (1:200) using a blocking buffer. Membranes were then incubated with the matched immunoglobulin G (IgG)-horseradish peroxidase (HRP)-conjugated secondary antibody, and the enhanced chemiluminescence (ECL, Thermo Fisher Scientific, Massachusetts, MA, USA) was used to detect the immunoreactive bands. These bands were visualized by scanning with Sapphire Biomolecular Imager (Azure Bio systems, Dublin, OH, USA) and digitalized via the image analysis software Sapphire Capture system (Azure Biosystems, Dublin, OH, USA).

Conclusions
Our findings provide clinical insight about the potential effects of cannabis abuse on the circulatory protein expression. These proteins are highly involved in different applications and diseases/disorders. The present study highlighted that drug discovery research can further investigate the effects of cannabis ingredients on immunological and inflammatory responses as well as the diseases involved, e.g., atherosclerosis. This will prove a novel direction to discover and develop potential compounds for the prevention or treatments of these diseases. These researches might target acute phase proteins, NO and ROS pathways, atherosclerosis signaling, or LXR/RXR and FXR/RXR pathways. One of applications obtained from the current work is the drug-drug interaction since our present study showed that the serum expression of albumin, which is a major serum protein binding, is decreased in CUD patients. Our study also found that cannabis abuse might modulate several diseases and disorders. However, the quantity of cannabis inside the humans were not controlled; therefore, this hypothesis needs more investigations with controlled doses. In addition, cannabis include cannabinoids and non-cannabinoids where the cannabinoids are either psychoactive or non-psychoactive. Thus, future studies should further investigate our findings using a specific cannabis constituent. This will provide a clear understanding about the responsible compound for a specific effect. A limitation in our study is the age variation in both groups. More research is required to exclude any age variation and also investigate the serum proteomic profiling of female CUD patients.

Supplementary Materials:
The following are available online, Figure S1: Pathways and canonical pathways identified in the IPA functional analysis, Figure S2: Gel images, Figure S3: Example of the full western blots (not truncated), Table S1: Experimental design, Table S2: Mass spectrometry list of significant differentially abundant proteins, Table S3: List of 25 proteins with accession numbers depicted in IPA network pathway, and Table S4: List of proteins with accession numbers for the top 5 canonical pathways.