Quorum Sensing Peptides Selectively Penetrate the Blood-Brain Barrier

Bacteria communicate with each other by the use of signaling molecules, a process called ‘quorum sensing’. One group of quorum sensing molecules includes the oligopeptides, which are mainly produced by Gram-positive bacteria. Recently, these quorum sensing peptides were found to biologically influence mammalian cells, promoting i.a. metastasis of cancer cells. Moreover, it was found that bacteria can influence different central nervous system related disorders as well, e.g. anxiety, depression and autism. Research currently focuses on the role of bacterial metabolites in this bacteria-brain interaction, with the role of the quorum sensing peptides not yet known. Here, three chemically diverse quorum sensing peptides were investigated for their brain influx (multiple time regression technique) and efflux properties in an in vivo mouse model (ICR-CD-1) to determine blood-brain transfer properties: PhrCACET1 demonstrated comparatively a very high initial influx into the mouse brain (Kin = 20.87 μl/(g×min)), while brain penetrabilities of BIP-2 and PhrANTH2 were found to be low (Kin = 2.68 μl/(g×min)) and very low (Kin = 0.18 μl/(g×min)), respectively. All three quorum sensing peptides were metabolically stable in plasma (in vitro) during the experimental time frame and no significant brain efflux was observed. Initial tissue distribution data showed remarkably high liver accumulation of BIP-2 as well. Our results thus support the potential role of some quorum sensing peptides in different neurological disorders, thereby enlarging our knowledge about the microbiome-brain axis.


Introduction
Bacteria communicate by means of chemicals (i.e. pheromones), produced and released by the bacteria and recognized by others. Once a threshold concentration of these molecules is reached, a coordinated change in bacterial behavior is initiated. This process of cell-to-cell communication is called 'quorum sensing'. Gram-negative and Gram-positive bacteria have phosphate dihydrate, sodium iodide, sodium dihydrogen phosphate monohydrate, sodium metabisulphite and Chloramine-T were obtained from Merck KGaA (Darmstadt, Germany). Calcium dichloride, D-glucose, formic acid (FA) and HEPES were purchased from Fluka (Diegem, Belgium) and dextran from AppliChem GmbH (Darmstadt, Germany). For the mobile phases, acetonitrile was obtained from Fisher Scientific (Erembodegem, Belgium) and water was purified using an Arium 611 Pro VF purification system (Sartorius, Göttingen, Germany) to laboratory-graded water (18.2 MΩ × cm). For the radiolabeling of the peptides, Iodo-Gen 1 coated tubes were purchased from Thermo Scientific (Erembodegem, Belgium) and the radioactive sodium iodide solution (Na 125 I) from Perkin Elmer (Zaventem, Belgium).

Ethics Statement
Female, Institute for Cancer Research, Caesarean Derived-1 (ICR-CD-1) mice (Harlan Laboratories, Venray, The Netherlands) of age 7-10 weeks and weighing 25-34 g, were used during the BBB-transport experiments. All animal experiments were performed in strict accordance with the Belgian legislation RD 31/12/2012 and the Ethical Committee principles of laboratory animal welfare; the protocol was approved by the Ethical Committee of Ghent University, Faculty of Veterinary Medicine (approval number 2012-157). All efforts were made to minimize suffering.
Human plasma was obtained from adult subjects who provided a written informed consent. Ethical Committee approval for the scientific use of these blood samples was not applied, according to the Belgian legislation 2013-03-19/03.

Peptide selection
The currently known quorum sensing peptides are continuously collected into the Quorumpeps database (http://quorumpeps.ugent.be) [26]. To select chemically diverse quorum sensing peptides for BBB-permeability investigations, we optimized the three-dimensional structure of these 231 peptides (status in August 2014) and calculated over 3000 descriptors for each peptide. After removal of the constant descriptors, and correction for molecular weight, a final dataset of 1468 descriptors was retained. Multivariate data-analysis on this resulting 231×1468 data-matrix was performed using Principal Component Analysis (PCA) with SIMCA-P+ 12.0 (Umetrics, Sweden) and different clusters identified [27]. Finally, three chemically diverse quorum sensing peptides were selected to investigate their brain permeability characteristics.

Peptide handling
The quorum sensing peptides were purchased at GL Biochem (Shangai, China) and the positive control dermorphin at Bachem (Bubendorf, Switzerland). The peptide purity was determined to be ! 90%, based on UPLC-PDA analyses [28]. Prior to experimental use, the peptides were dissolved in phosphate buffer (25 mM) at a concentration of 1 μmol/ml.

Peptide 125 I radiolabeling and purification
Dermorphin and BIP-2 were labeled using the Iodogen method [29,30]. Briefly, 0.1 μmol of the lyophilized peptide was dissolved in 100 μl of phosphate buffer (pH 7.4, 25 mM). A Iodo-Gen1 coated tube was previously rinsed with 1 ml of phosphate buffer. Subsequently, 50 μl of sodium iodide solution (1.1 μmol/ml) and 1 mCi of Na 125 I solution were transferred into this Iodo-Gen1 coated tube. The oxidation reaction was allowed to proceed for six minutes at room temperature, after which the iodonium solution was transferred to 50 μl of peptide solution (1 μmol/ml). The iodination reaction of the peptide was allowed to proceed another six minutes at room temperature. Next, the reaction mixture was analyzed by radio-HPLC and the eluting fractions determined for radioactive content (i.e. peptide concentration). The radio-HPLC apparatus consisted of a LaChrom Elite L-2130 pump with degasser (flow rate is 1 ml/min), a LaChrom Elite L-2300 column oven set at 30°C, a LaChrom Elite L-2400 UVdetector set at 215 nm (all Hitachi, Tokyo, Japan), a Rheodyne 7725i manual injector with 100 μl sample loop (Rheodyne, Rohnert Park, CA, USA), a Berthold LB500 HERM radioactivity detector (Berthold Technologies, Bad Wildbad, Germany) equipped with EZChrom Elite version 3.1.7 software for data acquisition (Scientific Software, Pleasanton, CA, USA) and a fraction collector FC 203 (Gilson International BV, Den Haag, The Netherlands). For separation, a Vydac Everest C 18 (250 × 4.6 mm, 5 μm particle size) column (Grace, Baltimore, MD, USA) was coupled to the HPLC system. Mixtures of water (0.1% FA m/V) and acetonitrile (0.1% FA m/V) were used to create appropriate gradients for separation of peptides and their iodinated forms. The mono-(and di-) iodinated peptide fractions were then concentrated (if necessary) by nitrogen drying and the appropriate peptide concentrations prepared using Lactated Ringer's solution containing 1% of BSA. The negative control, i.e. BSA, was iodinated using the same procedure and the iodinated protein isolated from free iodine using an argent filter.
Peptides phrANTH2 and phrCACET1 were iodinated using the Chloramine-T (CAT) method [31]: 50 μl of peptide solution (1 μmol/ml) was subsequently mixed with 20 μl of 4.5 mg/ml NaI in 100 mM phosphate buffer (phrANTH2) or 0.1% m/V formic acid in water (phrCACET1), 1 mCi of Na 125 I solution and 30 μl of a 4 mg/ml CAT in 100 mM phosphate buffer solution. For phrCACET1, 40 μl of 0.1% m/V formic acid in water was added before the CAT solution as well. The iodination reaction was continued for 120 (phrANTH2) or 40 (phrCACET1) seconds, after which 30 μl of sodium metabisulphite solution (8 mg/ml) was added to neutralize the oxidizing agent. Next, the reaction mixtures were analyzed by radio-HPLC using the described procedures and the eluting fractions determined for radioactivity amount. Again, nitrogen drying was performed on the mono-(and di-) iodinated fractions and the solutions prepared using Lactated Ringer's solution containing 1% of BSA.

Multiple time regression analysis
In order to determine whether the peptides could enter the brain, in vivo multiple time regression (MTR) analyses were performed. Therefore, ICR-CD-1 mice were anesthetized intraperitoneally using a 40% urethane solution (3 g/kg). Then, the jugular vein and carotid artery were isolated and 200 μl of the radiolabeled peptide solution, diluted to 30 000 cpm/μl using Lactated Ringer's solution containing 1% of BSA (LR/BSA), was injected into the jugular vein. At specified time points after injection (i.e. 1, 3, 5, 10, 12.5 and 15 min, with start and end in duplicate), blood was obtained from the carotid artery followed by decapitation of the mouse. The isolated brain was weighed and radioactivity measured in a gamma counter (Wallac Wizard automatic gamma counter, Perkin Elmer, Shelton, CT, USA), as well as from 50 μl serum, which was obtained by centrifuging the collected blood at 10 000 g for 15 min at 21°C. To evaluate the tissue distribution of the peptides during the BBB-experiments, seven other tissues, i.e. spleen, kidneys, lungs, heart, duodenum, muscles and liver, were collected immediately after decapitation of the mice at the last time point of 15 min. After weighing the tissues, the radioactivity was measured in a gamma counter.
The linear modeling of the multiple time regression analysis is based on the Gjedde-Patlak equation [32][33][34]: and where A m (T) is the amount of radioactivity in the brain at time T, C p (T) the amount of radioactivity in serum at time T, K in the brain influx rate constant, Θ the "exposure" time and V i the initial brain distribution volume. The exposure time represents the theoretical steadystate serum level of radiolabeled peptide at the serum concentration C p (t) and is defined as the integral of the serum radioactivity over time divided by the radioactivity at time T. During the multiple time regression experiments, peptides are intravenously injected and subsequently cleared by the organs. Therefore, the exposure time is used during the modeling of the brain influx of the peptides to account for the decreasing concentrations. The integral of radioactivity over time is represented by the area under the curve (AUC) [32,35,36].
Finally, the brain/serum ratios (μl/g) were plotted versus the exposure time and the slope of this relationship represents the unidirectional influx rate (K in ) from blood to brain and the intercept represents the initial brain volume of distribution (V i ).
If within the experimental time frame, there is a transition from unidirectional to net transfer, then the following expansion of the Gjedde-Patlak plot, a model of biphasic blood-brain transfer as derived from Wong et al [37], was used to fit the uptake: Where K 1 is the unidirectional clearance, K is the net clearance, Θ the exposure time, V g the tissue brain distribution volume and V 0 the vascular brain distribution volume, experimentally determined by the lowest V 0 obtained, i.e. from PhrANTH2, which was somewhat lower than the standard vascular marker iodinated BSA [38].' For the evaluation of the tissue distribution of the radiolabeled peptides 15 min after IVinjection, the percentage of the injected dose for each isolated tissue was calculated as follows: where A tissue and A injected are the measured activities of the isolated tissue and the activity of 200 μl of MTR solution, respectively, while w tissue is the weight of the selected tissue and w animal is the mass of the injected mouse. The results are the mean values of the duplicates.

Capillary depletion
We performed capillary depletion to determine whether the peptides, taken up by the brain, completely crossed the capillary wall into the tissue rather than just being trapped in the endothelium. The method of Triguero et al., as modified by Gutierrez et al., was used [39,40]. Briefly, ICR-CD-1 mice were first anesthetized intraperitoneally using a 40% urethane solution (3 g/kg). After isolation of the jugular vein, 200 μl of the iodinated peptide solution, diluted to 10 000 cpm/μl using LR/BSA, was injected in the jugular vein. Ten minutes after injection, blood was collected from the abdominal aorta and the brain was perfused manually with 20 ml of Lactated Ringer's buffer after clamping the aorta and severing the jugular veins. Subsequently, the brain was collected, weighed and the radioactivity measured in the gamma counter. Then, the brain was homogenized with 0.7 ml of ice-cold capillary buffer (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM CaCl 2 , 1 mM MgSO 4 , 1 mM NaH 2 PO 4 and 10 mM D-glucose adjusted to pH 7.4) in a pyrex glass tube and mixed with 1.7 ml of 26% ice-cold dextran solution in capillary buffer. The resulting solution was weighed and centrifuged in a swinging bucket rotor at 5400 g for 30 min at 4°C, after measuring the radioactivity in the gamma counter. Pellet (capillaries) and supernatant (parenchyma and fat tissues) were collected, weighed and measured in a gamma counter. After centrifuging the obtained blood (10 000 g, 21°C, 15 min), the radioactivity of 50 μl serum was measured in a gamma counter as well.
Compartmental distribution was calculated as follows: where CD tissue represents the ratio of the activity of the capillaries or parenchyma and the activity of serum for the fraction of radiolabeled peptide in the capillaries and parenchyma, respectively.

Brain-to-blood transport
We quantified the amount of peptide exported out of the brain as previously described [41]. ICR-CD-1 mice were anesthetized intraperitoneally using a 40% urethane solution (3 g/kg). Then, the skin of the skull was removed and a hole was made into the lateral ventricle using a 22 G needle marked with tape at 2 mm at the following coordinates: 1 mm lateral and 0.34 mm posterior to the bregma. The anesthetized mice received an intracerebroventricular (ICV) injection of 1 μl of the diluted iodinated peptide solution using LR/BSA (25 000 cpm/μl) by pumping the peptide solution at a speed of 360 μl/h for 10 s using a syringe pump (KDS100, KR analytical, Cheshire, UK). At specified time points after ICV-injection (i.e. 1, 3, 5, 10, 12.5 and 15 min), blood was collected from the abdominal aorta and subsequently the mouse was decapitated. Then, the whole brain was collected, weighed and measured in a gamma counter, as well as from 50 μl of serum, which was obtained by centrifuging the collected blood at 10 000 g during 15 min at 21°C. The efflux half-life was calculated from the linear regression of the natural logarithm of the residual radioactivity in brain versus time as follows: where k out is defined as the efflux rate constant calculated as the negative value of the slope of the linear regression, applying first order kinetics.

In vitro human plasma stability
In vitro metabolic stability of the quorum sensing peptides was determined in human plasma using previously described procedures [42]. In brief, 100 μl of non-radiolabeled peptide (1 mg/ml) was incubated in 400 μl of Krebs-Henseleit buffer pH 7.4 and 500 μl of plasma at 37°C while shaking. At predetermined time intervals (i.e. 0, 30 and 120 minutes), 100 μl aliquots were immediately transferred into microtubes containing 100 μl of 1% (V/V) trifluoroacetic acid solution in water. The enzyme reaction was further stopped by heating the solution at 95°C for 5 minutes. Next, the samples were cooled for 30 minutes in ice and subsequently centrifuged to precipitate the denatured proteins; the supernatant was analyzed using UPLC-PDA/MS.

Brain influx and tissue distribution
Two of the three investigated quorum sensing peptides showed a statistically significant influx into the mouse brain. It is also clear that dermorphin, PhrCACET1 and BIP-2 showed a biphasic brain influx model [46], characterized by a rapid initial uptake followed by a plateauing equilibrium. In Fig 2, the ratio of the brain and serum radioactivity is plotted versus the exposure time; the quantitative influx parameters of the molecules are summarized in Table 1. The data were fitted using a simple linear regression model (Eq 1) for BSA and PhrANTH2 and a biphasic model (Eq 2) for dermorphin, PhrCACET1 and BIP-2.
The positive control peptide, dermorphin, clearly showed an influx with a rate constant of 5.26 μl/(gxmin). It is well established that BSA, the negative control protein, shows a very small, almost negligible brain influx [47], which was confirmed here (K in of 0.12 μl/(g×min)). Both controls thus indicated a good performance of the brain influx experiments and their obtained K in values can be used to benchmark the influx results of the quorum sensing peptides. Peptide PhrANTH2 showed a very small influx into the brain, comparable to BSA: a K in of 0.18 μl/(g×min) was obtained, i.e. statistically not significantly higher than the K in of BSA. Peptide PhrCACET1 showed a very high initial influx into the brain with a K in of 20.87 μl/ (g×min); the high V g value indicates a rapid influx as well. BIP-2 showed a small initial influx into the brain with a K in of 2.68 μl/(g×min), which is statistically significantly higher than the K in of BSA and comparable to dermorphin (not significantly different), but lower than the K in of PhrCACET1.
These initial brain influx results thus correspond well with the overall classification conclusions [48]: PhrCACET1 showed the highest brain influx, followed by BIP-2 and dermorphin.
The results of the capillary depletion study (Fig 3) at 10 min after injection validate the high brain influx of peptide PhrCACET1: the absolute amount of peptide in the brain (i.e. absolute y-axis values of Fig 3) is much higher for this peptide compared to the others. The relative amount of peptide that is effectively transferred over the capillaries into the brain showed a higher brain parenchyma (77%-85%) versus a lower capillary retention (15%-23%) for the three peptides (Table 1).
In Fig 4, the relative concentrations in the different tissues, obtained from MTR experiments at the last time points (i.e. at 15 minutes post-injection) are graphically presented. BSA showed a clear liver distribution, while dermorphin was mainly distributed in the duodenum, followed by the liver, 15 min after i.v. injection. The high duodenum accumulation of this control peptide is explained by the presence of peripheral mu opioid receptors for which dermorphin Table 1. Overview of the multiple time regression results using both the linear as the biphasic model (mean ± 1 standard error).
Peptide ( 125 I-) K in (μL/g × min) Brain distribution volume, V 0 (μL/g)  The amount of peptide that is effectively transferred into the brain shows a higher brain parenchyma versus a relatively low capillary distribution for the three peptides. The capillary and parenchyma fraction (%) are indicated above the histogram.
doi:10.1371/journal.pone.0142071.g003 possesses high affinity [49]. PhrCACET1 showed a higher distribution in the kidneys and duodenum compared to the other investigated tissues, while BIP-2 was strongly accumulated in the liver.

Brain-to-blood transport
The natural logarithm of the measured residual radioactivity in the brain was plotted versus the time, to calculate the efflux constant k out . No significant efflux was observed during the experimental time of 15 minutes for the three quorum sensing peptides (k out BIP-2 = -0.06 ± 0.02; k out PhrANTH2 = -0.05 ± 0.07; k out PhrCACET1 = -0.02 ± 0.03).

In vitro human plasma stability
In order to correctly interpret the BBB permeability results, we investigated the stability of these quorum sensing peptides in human plasma. All three peptides were found to be stable during the experimental set-up time (i.e. maximum 15 minutes) although quite different metabolization kinetics were observed: a half-life value of 56.4, 23.6 and 320.7 minutes was obtained for PhrANTH2, PhrCACET1 and BIP-2, respectively.

Discussion
The quorum sensing peptide PhrCACET1 (SYPGWSW) very efficiently crossed the BBB, with a measured clearance by the brain that was higher than that of dermorphin (positive control), which generally shows a low but significant clearance at about 0.05% of the plasma flow to the brain [48]. The brain clearance of BIP-2 (GLWEDLLYNINRYAHYIT) was similar to dermorphin, but still higher than the plasma marker BSA. In contrast, PhrANTH2 (SKDYN) shows no significant transport across the BBB in this in vivo mouse model. To evaluate these transport results, we compared them with the net clearances of other investigated peptides, obtained using the same technique, for which the data is summarized in the Brainpeps1 database [50].
The classification system of BBB influx data, using the BBB in response proposed by Stalmans et al. [48], indeed indicated a 'very high influx' for the quorum sensing peptide PhrCACET1. A 'low influx' classification was obtained for both BIP-2 and dermorphin, while BSA and PhrANTH2 are classified as 'very low influx' compounds. The brain influx of PhrCACET1 and BIP-2 showed a biphasic behavior: after a steep increase of the brain to serum activity ratio, the curves reached a plateau. After this plateau phase, a reduction can not be ruled out. For PhrCACET1, this behavior can be explained by a number of mechanisms, including metabolisation in serum or brain tissue, leading to metabolites with different BBB transport properties as PhrCACET1 has the lowest in vitro human plasma stability of the three peptides. Accumulation of the peptide in the kidneys and duodenum can also contribute to this behavior. For BIP-2, the high liver accumulation can influence the brain influx behavior as metabolism of BIP-2 in human plasma was limited during the experimental time frame and thus was not judged to be the primary cause of its observed BBBkinetic behavior.
The three quorum sensing peptides mainly differ from each other in lipophilicity (principal component 2 in PCA plot) and size/compactness (principal component 1 in PCA plot): lipophilicity is determined by e.g. the clogP value, the number of hydrogen bonds (nHBonds) and the polarity of the molecule (TPSA), while size is evaluated by i.a. the WHIM size index and the Wiener W index [27]. As BBB permeability generally increases with lipophilicity, i.e. PhrANTH2 < BIP-2 < PhrCACET1, our findings are consistent with current physicochemical explanations of BBB transport characteristics of peptides [51]. When we analyze the clustering results of all Quorumpeps peptides (n = 231, Fig 1) [26,27] and select the peptides with a comparable lipophilicity as PhrCACET1 (i.e. the peptide with the highest BBB permeability characteristics), we mainly find quorum sensing peptides that are synthesized by Bacillus subtilis. These peptides thus can, theoretically based on this in silico clustering, cross the blood-brain barrier and exert their biological effect in the brain. Bacillus subtilis, originally described as a soil organism, is also found in the human gut [52], so these PhrCACET1 lipophilicity-related quorum sensing peptides can be available for the brain once they have reached the blood circulation.
Some quorum sensing peptides, when present in the blood, thus can reach the brain tissue by penetrating the BBB barrier and exert local CNS-effects. BIP-2 is synthesized by the Grampositive bacterium Streptococcus pneumoniae, a commensal of the nasopharynx in the majority of healthy children. Other commensal viridans streptococcal species, such as Streptococcus mitis, which are genetically highly related to Streptococcus pneumoniae, have commensal nasopharynx and oropharynx properties as well. Next to their commensal behavior, a variety of infectious complications can be assigned to these pathogens, including meningitis, endocarditis, bacteremia and septicemia [44,53]. Based on this bacterial presence, the quorum sensing peptides produced by these genetically related bacteria are expected to be present in the systemic circulation, presented to the brain, followed by BBB-penetration and subsequently biological activity. PhrCACET1 was found in Clostridium acetobutylicum [54], but a high genetic homology with other pathogenic Clostridium species, i.e. Clostridium botulinum, Clostridium perfringens, Clostridium difficile and Clostridium tetani [55,56], indicates that comparable quorum sensing peptides can probably also be synthesized in the human body by these bacteria.
Our findings corroborate well with recent clinical findings. The predominant presence of Clostridium spp. in the human gut was associated with autism in children [57]. As the quorum sensing peptide PhrCACET1 rapidly crosses the BBB, the neurological effect of this peptide remains to be investigated as well, as these results could potentially explain the mechanisms by which commensal flora trigger autism; the intestinal absorption capability of selected quorum sensing peptides was recently demonstrated [11]. Next to autism, recent studies have indicated that microbiota have dramatic effects on other central nervous dysfunctions as well. For example, a shift in the gut microbial composition was shown in hepatic encephalopathic patients, indicating the possible involvement of the microbiota in the pathophysiology of this disease [58]. Moreover, the human microbiome also plays a role in depression, anxiety and stress [59]. With depression, a decrease in mood disorders was observed with probiotics containing Lactobacilli and Bifidobacteria [60]. The same beneficial effect of these probiotics is observed on anxiety as well; in contrary, infection with Campylobacter jejuni elevated anxiety-like behavior [60,19]. An altered gut microbiota composition was also associated with stress: a decreased abundance of i.a. Bacteroides and an increased abundance of i.a. Clostridium was observed after stress exposure [61]. Immune-mediated neuro-psychiatric disorders may be influenced by the microbiota as well, including multiple sclerosis (MS), neuromyelitis optica and Guillain-Barré syndrome [60]. These recent findings, together with our results of BBB-penetration of quorum sensing peptides into the brain, thus excite the research on the influence of these signaling molecules on the development of central nervous system (CNS) disorders.