Cocaine Induces Inflammatory Gut Milieu by Compromising the Mucosal Barrier Integrity and Altering the Gut Microbiota Colonization

Cocaine use disorder (CUD), a major health crisis, has traditionally been considered a complication of the CNS; however, it is also closely associated with malnourishment and deteriorating gut health. In light of emerging studies on the potential role of gut microbiota in neurological disorders, we sought to understand the causal association between CUD and gut dysbiosis. Using a comprehensive approach, we confirmed that cocaine administration in mice resulted in alterations of the gut microbiota. Furthermore, cocaine-mediated gut dysbiosis was associated with upregulation of proinflammatory mediators including NF-κB and IL-1β. In vivo and in vitro analyses confirmed that cocaine altered gut-barrier composition of the tight junction proteins while also impairing epithelial permeability by potentially involving the MAPK/ERK1/2 signaling. Taken together, our findings unravel a causal link between CUD, gut-barrier dysfunction and dysbiosis and set a stage for future development of supplemental strategies for the management of CUD-associated gut complications.


Results
Cocaine induces gut dysbiosis. To achieve detailed insight into the causal association of cocaine exposure and gut dysbiosis we first used 16S rRNA sequencing to analyze the microbiota from the fecal droppings and the colon of mice that were administered cocaine (20 mg/kg; i.p./day) for 7 consecutive days. We found 5061 OTUs that were common between the cocaine-exposed and control groups and, 9371 and 12420 OTUs that were specific to the saline controls and cocaine-exposed groups respectively (Fig. 1A). We further clustered the OTUs between the fecal droppings and the colon, and found 2362 OTUs that were common between the fecal droppings and colon irrespective of cocaine-or saline-exposure; and 11911 and 8926 OTUs that were specific for cocaine-exposed fecal droppings and colon, respectively (Fig. 1B). Species accumulation reached saturation (Fig. 1C), richness and evenness (Fig. 1D). Bacterial alpha diversity analyzed as observed species, or the Chao1 index remained unchanged between the cocaine-exposed and the saline administered mice (Fig. e,f). We next analyzed for shifts in beta diversity and observed alterations in bacterial composition based on the weighted UniFrac ANISOM diversity metric (R = 0.592, p = 0.001) thereby suggesting microbiota dissimilarities between cocaine-versus saline-administered mice. To better understand the effect of cocaine on the gut microbial composition, we next studied the effect of cocaine exposure on the taxonomic composition at the genus level. Cocaine exposure decreased colonization in the colon of four genera, i.e., Mucispirillum, Butyricicoccus, Pseudoflavonifractor and unclassified Ruminococcaceae with a concomitant increase in colonization of other genera such as Barnesiella, and unclassified members of Porphyromonadaceae, Bacteriodales, and Proteobacteria ( Fig. 2A-C, Table 1). Fecal droppings of the cocaine-exposed mice also demonstrated decreased abundance of Lachnospiracea incertae sedis, Pseudoflavonifractor and Streptophyta and increased abundance of Turicibacter, Alistipes, Odoribacter and unclassified members of Porphyromonadaceae and Proteobacteria, (Fig. 1A-C, Table 2) thereby suggesting alterations of gut microbial homeostasis. Furthermore, it was also found that bacterial population in the colon and fecal droppings of cocaine-administered mice clustered separately from saline controls, based on the phylogenetic distance UPGMA weighted UniFrac (Fig. 1D) and principal coordinate analysis (Fig. 1E).
We next analyzed the effect of cocaine exposure at the species level and found depletion of unclassified species of Mucispirillum, Butyricoccus and Ruminococcaceae with a concomitant upregulation of unclassified species of Porphyromonadaceae, Bacterroidales, Barnisiella, Proteobacteria and Alistipes among others in the colon (Supplementary Table S1). Fecal droppings of cocaine-exposed mice demonstrated a similar dysregulation of microbial homeostasis involving downregulation of several bacterial species including unclassified Lachnospiracea incertae sedis, Streptophyta and uncultured Lactobacillus species (Supplementary Table 2). Cocaine-induced alterations in Lachnospiracea, Butyricoccus, Ruminococcaceae and other butyrate-producing bacteria suggested disturbances in gut energy balance as well as butyrate-mediated neural and immune functions.
Cocaine administration alters gut homeostasis. Next, we sought to examine whether cocaine exposure could affect gut homeostasis. We first assessed phosphorylation of ERK1/2 kinase that lies upstream of both CDX-2 and NF-κB -transcription factors that are critical for regulating gut homeostasis. While the levels of total ERK1/2 kinase remained unchanged in cocaine-administered mice, its phosphorylation was significantly upregulated in both the colon and ileum of cocaine-exposed mice compared with the control group (Fig. 3A,B). As shown in Fig. 3C,D, there was significant upregulation of CDX2 expression both in the colon and ileum of cocaine administered mice. A concomitant increase in the expression of both total and phosphorylated NF-κB was also observed in the colon of cocaine administered mice (Fig. 4A,B), thereby underscoring the role of cocaine in the modulation of gut homeostasis. The next step was to determine whether cocaine exposure resulted in the induction of an inflammatory environment in the gut by analyzing the gene expression of specific chemokines and cytokines. Total RNA from the colon was subjected to qPCR analysis using a Qiagen mouse (common) cytokines and chemokines PCR array panel (PAMM 150, Qiagen). As shown in Fig. 4C,D, there was increased expression of several chemokines and cytokines and also downregulation of some others in the colon of cocaine administered mice compared to the control group. Further validation of these findings was also done using qPCR analysis for chemokines such as CCL-2, CCL-7, CXCL-10, CCL-11 as well as for cytokines IL-18 and IL-1β. As shown in Fig. 4E,F, cocaine exposure was found to induce a pro-inflammatory gut milieu. Notably, additional analysis using sera of either saline or cocaine-administered mice further demonstrated increases in known pro-inflammatory mediators such as Lipopolysaccharide Binding protein (LBP) and soluble CD14 ligand (sCD14) (Supplementary Fig. S1).
Cocaine administration alters mucosal epithelial barrier composition and integrity. Gut barrier dysfunction is often attributed to tight junction dysregulation, which includes the claudin family of proteins. Based on the premise that expression of claudins is altered in an inflammatory milieu, we next sought to assess for the barrier integrity by measuring the trans-epithelial electrical resistance (TEER) in the widely used in vitro model of intestinal epithelial cells -polarized monolayers of Caco-2 cells 9-11 . Confluent cell monolayers on transwell filter support (0.04 µM pore size) were either exposed to cocaine (10 µM) for 24 h or left unexposed followed by TEER measurements. As shown in Fig. 5A, cocaine exposure resulted in a significant reduction of TEER in Caco-2 cells. Complementary analysis of the paracellular permeability using apicobasal passes of a FITC-dextran dye (4 kDa) in these cells further demonstrated a significant increase in permeability compared with control cells (Fig. 5B).
We next assessed intestinal permeability in mice that were administered cocaine (20 mg/kg, i.p.) or saline (n = 9/group) for seven days and subsequently administered FITC Dextran (4 kDa) on the final day of cocaine injection followed by collection of blood after 4 h for detection of FITC by flourometry. As shown in Fig. 5c, cocaine increased intestinal permeability as FITC-dextran concentration in blood was higher in mice administered with cocaine compared with mice administered with the vehicle (p = 0.08).  Alpha diversity metrics in cocaine-or saline-administered mice. Wild-type mice (C57BL/6) were administered cocaine (i.p, 20 mg/kg) or saline for seven consecutive days followed by euthanasia within 1 hr of the last injection. Fecal droppings and distal colon fecal matter were collected for 16S-rRNA sequencing. As shown in (A,B), cocaine administration altered several Operational Taxonomic Units (OTU) both the colon and the fecal droppings. Metrics for species accumulation curve (C), rank abundance (D), Observed species (E) and Chao1 (F) phylotype diversity are shown. n = 10/group. www.nature.com/scientificreports www.nature.com/scientificreports/ We next examined the expression and cellular distribution of the important claudin proteins (claudin-1, 2, 3 and 7) in the intestinal epithelium. Interestingly, the expression of these tight junction proteins was found to be upregulated in the colon of cocaine administered versus saline control mice (Fig. 6A). Notably, the expression upgma_weighted_unifrac D.  www.nature.com/scientificreports www.nature.com/scientificreports/ of the pore-forming claudin-2 was upregulated the most with up to 4-fold increase compared with controls. Analysis using immunofluorescence staining and imaging further validated alterations in the expression and cellular distribution of these proteins (Fig. 6B).

S = saline C = cocaine c = colon d = droppings
Cocaine administration also modulated claudin expression in 3d-culture of colonic crypts potentially by inducing ERK1/2 signaling. We next sought to assess if alterations in the expression of claudin proteins can be recapitulated in 3d-culture of the colon crypt organoids that were either exposed to cocaine or left unexposed in the presence or absence of the ERK1/2 inhibitor U0126. As shown in Fig. 7A, exposure of colon crypt organoids to cocaine modulated expression of claudins and similar to the changes in vivo in cocaine-treated mice (versus control), and these changes in claudin expression were accompanied with a significant increase in expression of phosphorylated ERK1/2. There are reports, including ours, demonstrating MAPK/ERK-mediated regulation of claudin expression and localization [12][13][14][15] . We thus next sought to explore whether inhibition of ERK phosphorylation could restore cocaine-mediated disruption of claudin expression. To address this, colon crypt organoids were pretreated with the specific inhibitor of MAP/ERK1/2 kinase inhibitor -U0126 (10 µM), followed by exposure of the crypt cultures to cocaine and assessed for the expression of claudins. Pretreatment of crypt organoids with U0126 ameliorated cocaine-mediated expression of claudin especially claudin-2, compared to the control crypt organoids. In the same samples, the cocaine induced claudin-1, −3, & −7 expression was not affected by inhibiting ERK-activation (Fig. 7A,B). Taken together, our data suggested potential role of ERK1/2 signaling in cocaine induced modulation of specific claudin expression and barrier deregulation.

Discussion
Cocaine use disorder (CUD) is estimated to affect up to 22.5 million people worldwide 3 resulting in marked changes in behavior and lifestyle emanating from its psychoactive and addictive effects. While most studies focus on the effects of cocaine on the brain, the current study was undertaken to assess the effects of cocaine on dysregulation of gut homeostasis including alterations in microbiota diversity, epithelial barrier function as well as innate immunity. Previous studies have demonstrated that depletion of gut microbiota by antibiotics in mice resulted in enhanced sensitivity to the reward and sensitizing properties of cocaine in the context of cocaine-mediated  www.nature.com/scientificreports www.nature.com/scientificreports/ behavioral plasticity 8 . The current study provides insights into the mechanism(s) of cocaine-induced dysregulation of gut-barrier function, microbial colonization, and inflammation by directly examining the effect of cocaine administered by the intraperitoneal route on the intestinal microbial colonization, in both the fecal droppings and the colon. Most notably, our findings suggest that cocaine administration specifically depletes the Mucispirillum, Ruminococcaceae, Lachnospiracea, Pseudoflavonifractor and Butrycicoccus bacteria, that are key producers of short-chain fatty acids (SCFA) and related metabolites and that play critical roles in maintaining mucosal epithelial and immune homeostasis. Importantly, SCFA constitute a significant source of energy for the resident microbiota in the colon, and thus by inference depletion of SCFA producers could, in turn, contribute to dysbiosis 16 . In this regard, Kiraly et al. demonstrated that supplementation of SCFA reverses cocaine-mediated behavioral effects in antibiotic-treated mice, thereby underscoring the critical role of bacterial metabolites in modulating behavioral changes observed in these animals 8 . Since SCFA are well-known histone deacetylase inhibitors (HDACs) 17 , and also since previous reports have shown that administration of HDAC inhibitors in rodents resulted in reduced behavioral responses to cocaine 18,19 , it is plausible to envision that cocaine-mediated behavioral changes are regulated by a mechanism(s) involving alterations in gut microbiota. Intriguingly, similar to injection of cocaine, a recent study on rats chronically exposed to volatile cocaine also demonstrated similar alterations in the gut microbiota 20 .
Cocaine has been shown to activate NF-κB signaling pathway resulting in upregulated expression of pro-inflammatory cytokines and adhesion molecules in various cell types [21][22][23][24][25][26] . Furthermore, phosphorylated ERK1/2 that lies upstream of NF-κB is also known to activate CDX-2, a key transcription factor expressed in the gut and that modulates the expression of several genes involved in cellular proliferation, differentiation, and inflammation including integral proteins of the barrier [27][28][29][30] . The current study demonstrated that i.p. cocaine administration induced an inflammatory environment in the gut as evidenced by increased expression of multiple www.nature.com/scientificreports www.nature.com/scientificreports/ pro-inflammatory cytokines (IL-18, IL-1β) and chemokines (CCL-2, CCL-7, CXCL-10, CCL-11) and furthermore, this involved increased activation of the transcription factors NF-κB and CDX-2. Understanding the contribution of cocaine-induced gut inflammation in mediating neuroinflammation and, its role in the development of drug addiction warrants further investigation.
We also observed that cocaine upregulated the expression of claudin-2 in the colon. Claudin-2 has been shown to form paracellular channels that are permeable to cations and water [31][32][33] . Interestingly, upregulation of claudin-2 has also been shown to increase tight junction permeability to sodium ions and water during enteric pathogen www.nature.com/scientificreports www.nature.com/scientificreports/ infection 34 . Our data suggest that cocaine increased the expression of claudin-2, which in turn, contributed to increased epithelial permeability. Further studies in models of claudin genetic ablation are required to ascertain the contribution of claudins to cocaine induced barrier dysregulation and microbial translocation.
An intriguing potential mechanism that could provide a plausible explanation underlying cocaine-mediated dysregulation of gut homeostasis was the breach of the gut-barrier integrity, key player involved in regulation of intestinal inflammation. Recent studies have suggested a feedback mechanism between the gut-barrier function and inflammatory cytokines such as IL-13, TNF-α, IL-1β, and IFN-γ [35][36][37][38][39] , thereby implying a positive association between gut-barrier dysregulation, dysbiosis, and inflammation. Our in vitro and in vivo analyses support such a postulate and demonstrate that cocaine administration could impact the gut-barrier composition and integrity, leading in turn, induction of gut dysbiosis and inflammatory milieu. It is worth noting that deregulation of claudin protein expression and its altered cellular distribution has emerged as a critical factor in the disruption of mucosal barrier integrity 31,34,40,41 . Furthermore, the potential non-canonical role of these proteins in regulating mucosal epithelial and immune homeostasis including Notch-signaling have also been described [42][43][44] . Our data demonstrating that cocaine-induced alterations in claudin proteins were rescued by inhibition of ERK1/2 signaling further supports the possibility of a non-canonical role and warrants future in-depth investigation.
While our finding suggest that cocaine can disrupt both the microbiota and compromise gut barrier integrity, we cannot conclusively suggest whether cocaine impact on the microbiome leads to permeability changes or that cocaine effects on the gut epithelium lead to altered microbiota. Additional studies using germ-free or antibiotic-treated mice are warranted to help resolve this issue.
In summary, the overall findings herein demonstrate gut homeostasis dysregulation as a severe health concern related to abuse of cocaine and identify gut-barrier dysregulation, altered gut microbiota colonization, and inflammation as contributing factors, potentially acting in an inter-dependent manner. The possibility that these cocaine-mediated changes in gut homeostasis could subsequently contribute to neuroinflammation is highly plausible. Understanding cocaine-induced gastrointestinal tract dysregulation thus appears to be critical in light of the emerging role of the gut in modulating behavior especially the addictive behaviors.
Mice and drug treatments. All animal procedures were performed in strict accordance with the protocols approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Centre and the National Institutes of Health. Eight to ten weeks old male mice (C57BL/6N) were purchased from Charles River Laboratories (Wilmington, MA, USA). They were housed under conditions of constant temperature and humidity on a 12-h light, 12-h dark cycle, with lights on at 0700hrs. Food and water were available ad libitum. Mice were randomly divided into two groups administered either Cocaine (C5775, Sigma-Aldrich, 20 mg/kg, i.p.) or saline once a day for 7 days with 2 cages per group (n = 5/cage), The cocaine dose of 20 mg/kg, given once daily, has been shown to induce cocaine's rewarding effects in C57BL/6 mice as reflected by the development of  www.nature.com/scientificreports www.nature.com/scientificreports/ conditioned place preference in these mice 45 . Several studies from our group have utilized this dose over a 7-day period and observed significant molecular changes [46][47][48] . We thus used a cocaine dose (20 mg/kg; i.p./day) known to produce robust molecular and behavioral changes to examine the relationship between cocaine exposure and operational taxonomic units (OTU) in the gut. Mice were sacrificed by isoflurane anesthesia 1 hour post the final drug injection for gut removal. Droppings and colon fecal matter were collected for microbial DNA isolation. Mice droppings and fecal samples were immediately frozen on dry ice and then stored at −80 °C. Colon or ileum tissues were collected and used for extraction of protein and total RNA. mRNA and protein levels of pro-inflammatory cytokines and signaling proteins were assessed by quantitative RT-PCR and western blotting, respectively. DNA isolation, 16S rRNA sequencing, and analysis. Bacterial DNA was extracted from fecal matter using the PowerSoil DNA isolation kit (MO Bio, Carlsbad, CA, USA Cat # 12888-100) according to manufacturer's protocol. DNA samples were stored at −20 °C (or −80 °C for longer storage) until amplification. Polymerase chain reaction (PCR) of the variable 3 and 4 (V3 and V4) of the 16S rRNA gene was performed using 515F (5′-GTGYCAGCMGCCGCGGTAA-3′) and 806R (5′-GGACTACNVGGGTWTCTAAT-3′) primers 49,50 . Thermocycling conditions were 98 °C for 10 s, 58 °C for 30 s, 72 °C for 45 s for 35 cycles and 72 °C for 10 min. DNA was sequenced using the Illumina MiSeq platform at LC Sciences (Huston, Texas). Briefly, Operational Taxonomic Units (OTUs) were clustered at 97% sequence similarity using cd-hit. Taxonomy was assigned using the Ribosomal Database Project (RDP) classifier v.2.1 against the Greengenes reference database (13_8). Quantitative Insights Into Microbial Ecology (QIIME) was used for subsequent analysis of within-and betweencommunity diversity (alpha and beta diversity). The relationship between 7-days cocaine administration (20 mg/ kg) and microbiota was explored using Principal Coordinate Analysis (PCoA) on unweighted UniFrac phylogenetic distances between communities. The dissimilarity distance between the cocaine-exposed mice group and the saline control group was tested using ANOSIM statistical method.
RNA isolation and quantitative polymerase chain reaction (qPCR) from gut tissue. qPCR was performed as previously described 51  Western blotting. Colon tissues or crypts were lysed in RIPA buffer supplemented with a protease inhibitor cocktail (ThermoFisher Scientific, 78430) followed by ultrasonication for 15 sec, at 80% amplitude. Western blotting was performed as previously described 51 . Lysates were cleared by centrifugation at 12000 g (10 min; 4 °C). Using Thermo Scientific BCA kit (Cat# 23225), protein concentration was quantified by the BCA method. Equal amounts of protein (10-20 µg) were electrophoresed in a sodium dodecyl sulfate-polyacrylamide gel under reducing conditions. Following transfer, PVDF membranes (Millipore, IPVH00010) were blocked with 5% nonfat dry milk for 60 mins at room temperature. Next, membranes were probed with primary antibodies overnight at 4 °C, washed with TBS-T, then incubated with appropriate HRP-conjugated secondary antibodies and developed with SuperSignal West Dura or Femto substrate. Densitometric analyses were done using NIH ImageJ software (ImageJ v1.44, NIH). Protein amounts were normalized to β-actin.
Trans-epithelial electrical resistance (TEER) assay. Caco-2 cells were cultured on polycarbonate transwell filter supports (0.04 µM pore size), exposed to cocaine (10 µM) and trans-epithelial electrical resistance (TEER) was measured as previously described 41,52 . Caco-2 cells form cell monolayers with characteristics of mature enterocytes and have been widely used as an in vitro model of intestinal epithelial cells [9][10][11]53,54 . Results were expressed relative to the initial TEER value and presented as mean plus or minus standard error for three independent experiments. FITC dextran permeability flux assay. Caco-2 cells were cultured on transwell filter support to confluency and exposed to cocaine 10 µM) for 24 h. Five microliters of FITC conjugated-dextran (120 mg/ml, 4 kDa, Cat # FD4-1G, Sigma) was added to the apical compartment of the transwell chamber system for 24 hours. The fluorescence intensity in the basal compartment was measured by fluorometry (excitation and emission wavelength of 485 nm and 520 nm, respectively).