Bacterial culture and EV isolation were carried out as described previously [36]. Lactobacillus plantarum (KCTC 11401BP, Gen-Bank accession No. GQ336971) was cultured in MRS broth (MB Cell, CA, USA) for 18 h at 37℃ with gently shaking (150 rpm). When the optical density of the culture reached 1.0 at 600 nm, bacteria were pelleted at 10,000 ×g for 20 min and the supernatant was passed through a 0.22-µm bottle-top filter (Corning, NY, USA) to remove remaining cells. The filtrate was concentrated through a MasterFlex pump system (Cole-Parmer, IL, USA) using a 100 KDa Pellicon 2 Cassette filter membrane (Merck Millipore, MA, USA) and subsequently passed through a 0.22-µm bottle-top filter (Corning, NY, USA). The extracellular vesicles were obtained by ultracentrifugation at 150,000 ×g for 3 h at 4℃. Protein concentration was measured with BCA assay (Thermo Fisher Scientific, MA, USA) and the collected EV fractions were stored at −80℃ until use. The size of purified L-EVs was in the range of 20~100 nm in diameter, which was described recently [36].

HT22 cells were cultured as described previously [37,38]. In brief, HT22 cells were grown in Dulbecco's modified Eagle's medium (DMEM; LM001-05, Welgene, Gyeongan-si, Korea) containing 10% heat-inactivated fetal bovine serum (FBS; FB02-500, Serum Source International, Charlotte, NC, USA) and penicillin (20 units/ml)/ streptomycin (20 mg/ml) (LS020-02, Welgene) at 37℃ and 5% CO₂ supply conditions. At 70~80% confluency, HT22 cells were trypsinized and counted using a trypan blue (0.4%) staining method. They were plated at a density of 1.0×10⁵ cells/well on a 6-well plate (SPL Life Science, Pocheon-si, Gyeonggi-do), or 1.0×10⁶ cells on a 100-mm dish in DMEM media supplemented with 10% FBS and antibiotics. After 24 h of culture, cells were treated with glucocorticoid (GC; corticosterone, 400 ng/ml) or L-EV (20 µg/ml) in DMEM media supplemented with 1% FBS for 24 h. The dose of GC and L-EV was chosen based on previous studies [36,37].

Transfection of siRNA into the HT22 cells was performed as described previously [37,39]. HT22 cells were plated at a concentration of 1.0×10⁵ cells/well in DMEM supplemented with 10% FBS in a 6-well plate (SPL Life Science). After 24 h, the media was changed with DMEM containing 1% FBS and siRNA was transfected using Lipofectamine-2000 (13778-075; Invitrogen). To prepare the siRNA and Lipofectamine-2000 mixture, 20 µM siRNA (3 µl; final concentration, 50 pM) and Lipofectamine-2000 (9 µl; final concentration, 7.5 µl/well) were separately diluted in 150 µl of Opti-MEM® Medium (31985070, Gibco, Thermo Fisher Scientific, Paisley, Scotland, UK). Each mixture was diluted at 1:1 ratio, mixed slowly and then incubated for 5 min at room temperature. The siRNA and Lipofectamine-2000 mixture (250 µl/well) was carefully dripped onto HT22 cells while culturing in fresh DMEM containing 1% FBS with or without 20 µg/well of L-EVs. They were then incubated for 24 h. Sirt1-siRNA was purchased from Bioneer (#1452517).

Mice were treated with restraints as described previously [40,41]. Male C57BL/6J mice (7-weeks old) were purchased from Daehan BioLink, Inc. (Eumsung, Chungbuk, Korea). Mice were housed in pairs per cage allowing an ad libitum access to water and food at a temperature (23℃) and humidity (50~60%)-controlled room under a 12-h light/dark cycle (lights on at 07:00~19:00 h). Animals were handled in accordance with the animal care guidelines of Ewha Womans University and restraint treatment procedures in this study were approved by the Ewha Womans University Animal Care and Use Committee (IACUC 15-012). To deliver restraint, mice were placed in a 50-ml polypropylene tube carrying many punched holes for ventilation and were restrained for 2-h daily for 14 days. Control mice housed in pairs were maintained in home cages without disturbance.

L-EVs were administrated in mice in the following ways. L-EVs at a dose of 0.1 µg/kg each diluted in saline were intraperitoneally injected into a mouse at a volume of 100 µl 30 min prior to restraint treatment and after the 14 days of stress session, the daily L-EV injection was continued until the behavioral tests were completed (Exp. 1). L-EV treatment during the post-stress period was conducted in mice that were exposed to 2h×14d restraint treatment (Exp. 2). For the post-stress period, L-EVs were intraperitoneally injected at a volume of 100 µl containing increasingly higher doses; 0.1 µg/kg for the first 5 days, 0.18 µg/kg for the following 2 days, and 0.27 µg/kg for the final 7 days. Imipramine hydrochloride (IMI; I0899, Sigma) at a dose of 20 mg/kg was intraperitoneally injected at a volume of 100 µl for the indicated days.

Quantitative real-time PCR (qPCR) analysis was carried out as described previously [40,41]. Briefly, hippocampus tissue was obtained and ground using pellet pestles (Z359971, Sigma-Aldrich) in TRI-zol reagent (15596-018, Invitrogen). Harvested HT22 cells were homogenized with TRI-zol reagent. Total RNA was isolated from homogenates and eluted in RNase free water (129112, Qiagen, Hilden, Germany). Two µg of total RNA was treated with DNase I to remove genomic DNA contamination and then converted to cDNA using a reverse transcriptase system (Promega, Madison, WI, USA).

Four µl of 1/8 diluted cDNA, 10 µl of 2X iQTM SYBR Green Supermix (Bio-Rad Laboratories, Foster City, CA, USA), and 1 µl each of 5 pmol/µl forward and reverse primers were mixed to a total volume of 20 µl. The qPCR reaction was carried out using the CFX 96 Real-Time PCR System Detector (Bio-Rad Laboratories). The transcript expression levels were normalized relative to Gapdh and L32 levels.

The primers used in this study were: tBdnf (total form), forward 5′-TGGCTGACACTTTTGAGCAC-3′ and reverse 5′-GTTTGCGGCATCCAGGTAAT-3′; Bdnf1, forward 5′-CCTGCATCTGTTGGGGAGAC-3′ and reverse 5′-GCCTTGTCCGTGGACGTTTA-3′; Bdnf4, forward 5′-CAGAGCAGCTGCCTTGATGTT-3′ and reverse 5′-GCCTTGTCCGTGGACGTTTA-3′; Creb1, forward 5′-GAGGCAGCAAGAGAATGTCG-3′ and reverse 5′-CCAGTCCATTCTCCACCGTA-3′; Hdac2, forward 5′- GGGACAGGCTTGGTTGTTTC-3′ and reverse 5′-GAGCATCAGCAATGGCAAGT-3′; Nt3, forward 5′-TACTACGGCAACAGAGACG-3′ and reverse 5′-GTTGCCCACATAATCCTCC-3′; Nt4/5, forward 5′-AGCGTTGCCTAGGAATACAGC-3′ and reverse 5′-GGTCATGTTGGATGGGAGGTATC-3′; Sirt1, forward 5′-GATCCTTCAGTGTCATGGTTC-3′ and reverse 5′-ATGGCAAGTGGCTCATCA-3′; Trkb, forward 5′-AAGGACTTTCATCGGGAAGCTG-3′ and reverse 5′-TCGCCCTCCACACAGACAC-3′; Gapdh, forward 5′-AGAAGGTGGTGAAGCAGGCATC-3′ and reverse 5′-CGAAGGTGGAAGA GTGGGAGTTG-3′; L32, forward 5′-GCTGCCATCTGTTTTACGG-3′ and reverse 5′-TGACTGGTGCCTGATGAACT-3′

Western blot analysis was carried out as described previously [40]. HT22 cells were washed with 1X PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄) and harvested. Harvested HT22 cells or hippocampal tissue was homogenized in homogenization buffer (50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, and 0.1% sodium deoxycholate) containing a protease inhibitor cocktail (Roche) by sonicating on ice using an Epishear probe sonicator, with two rounds of 15-sec pulses and 30-sec rest intervals at 40% power outlet (Active Motif). The supernatant of homogenates was obtained after centrifugation at 13,000 ×g at 4℃ for 15 min.

The amount of protein was determined by the Bradford method (500-0006, Bio-Rad Laboratories). Twenty µg of tissue or cell sample was mixed with 6X gel loading buffer and boiled for 5 min. The proteins were resolved on SDS-PAGE and transferred onto PVDF membrane using Trans-Blot® SD semi-dry Electronic transfer cell and power supply system (1703848, Bio-Rad Laboratories). The blots were incubated with 5% skim milk or 1% bovine serum albumin (BSA) in 1X TBST (150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 0.1% Tween 20) to block a non-specific binding. Blots were incubated with a primary antibody in blocking solution followed by a secondary antibody in 1X TBST. Specific bands were visualized using Enhanced ChemiLuminescence (ECL)TM Western Blot Analysis System (RPN2109, Amersham, GE Healthcare, UK). Quantification of blot images was processed using the Image-Pro-Premier 6.0 software (MediaCybernetics, Rockville, MD, USA).

The primary antibodies used were: anti-proBDNF (ANT-006; 1:1,000, Alomone Labs, Hadassah Ein Kerem, Jerusalem, Israel), BDNF (ab108319; 1:2,000, Abcam, Cambridge, UK), and anti-β-actin (sc-47778; 1:1,000, Santa Cruz). The secondary antibodies used were: anti-mouse IgG-HRP (GTX213111-01; 1:1,000, Gene-Tex, Irvine, CA, USA) and anti-rabbit IgG-HRP (GTX213110-01; 1:1,000, GeneTex).

The behavioral tests were carried out as described previously [40,41]. Mice were acclimated in the testing room for at least 30 min prior to the start of each behavioral test. All behavioral tests were performed in the light cycle phase (9 a.m.~3 p.m.) and monitored with a computerized video tracking system (SMART; Panlab S.I., Barcelona, Spain) or a webcam recording system (HD Webcam C210, Logitech, Newark, CA, USA).

The sociability test was performed as described previously [40,42]. Briefly, the U-shaped two-choice field was prepared by partitioning an open field (40×40 cm²) with a wall (20-cm wide and 20-cm high). Circular grid cages (12 cm in diameter×33 cm height) were presented on each side of the U-shaped two-choice field. For habituation, a subject mouse was allowed to freely explore the U-shaped two-choice field with empty circular grid cages on each side for 5 min and returned to the home cage. After 10 min, a social target mouse was loaded to a circular grid cage at one side and the subject mouse was placed in the center of the U-shaped two-choice field where the subject mouse was able to see both grid cages. The subject mouse was allowed to explore both fields for 10 min while the trajectory spent in the fields was recorded by a video tracking system (SMART, Panlab S.I.) The field with a circular grid cage carrying a social target mouse and the field containing an empty circular grid cage were defined as the target field and non-target field, respectively. Social target mice with same age and sex as the subject mice were prepared. Social targets had never been exposed to subject mice from the acclimation stage.

The tail suspension test (TST) was carried out as described previously [40]. Mice were suspended individually by fixing their tails with adhesive tape to the ceiling of a shelve 50 cm above the bottom floor. The subject mouse was suspended for 6 min and the cumulative immobility time was measured. Behavioral performances were recorded with a webcam recording system (HD Webcam C210, Logitech) and subsequently analyzed.

The forced swim test was performed as described previously [40]. Mice were placed in a Plexiglas cylinder (15 cm in diameter×27 cm height) holding water at 24℃ with a depth of 15 cm. Mice were placed in the cylinder for 6 min and the cumulative immobility time was measured for the final 5 min. Immobility was defined as the time when a mouse was floating with all limbs motionless. The performance during the test was recorded using a webcam recording system and then analyzed.

Multiple comparisons were performed by one-way ANOVA followed by the Newman-Keuls post hoc test or two-way ANOVA followed by the Bonferroni post hoc test. All data are represented as mean±SEM, and statistical significance was accepted at the 5% level.

Administration of microbiota products in mice increased BDNF in hippocampus [13,43]. Treatment with probiotics mixture in aged rats increased BDNF levels in the hippocampus [44]. Administration of a probiotic formulation (Lactobacillus helveticus and Bifidobacterium longum mix) in mice rescued stressed-induced abnormal brain plasticity and reversed the reduction of neurogenesis and BDNF levels [45]. Therefore, we investigated whether treatment with L-EVs in neuronal cells could induce changes in the expression of BDNF and its related factors. The hippocampal cell line HT22 cells treated with glucocorticoid (GC; corticosterone, 400 ng/ml) had decreased expression of total Bdnf (tBdnf) and the BDNF splicing variants Bdnf1, Bdnf4, and Ngf compared with that of untreated control HT22 cells. In contrast, L-EV treatment (20 µg/ml) reversed GC-induced reduced expression of tBdnf, Bdnf1, Bdnf4, and Ngf (Fig. 1A). Western blot analysis confirmed that L-EV treatment blocked GC-induced reduced expression of proBDNF (Fig. 1B).

BDNF expression is regulated by transcription and/or epigenetic factors, including cAMP response element (CRE) binding protein (CREB1) [46], histone acetyltransferase 2 (HDAC2) [47], and Sirtuin1 (Sirt1) [48]. HDAC2 negatively regulates BDNF expression in stress-induced depression models [47], whereas CREB1 and Sirt1 enhance BDNF expression [47,48,49].

GC treatment in HT22 cells decreased Sirt1 expression, but not significantly Creb1 and Hdac2, whereas L-EV treatment increased the GC-induced reduced expression of Sirt1, but produced no significant change in Creb1 and Hdac2 expression (Fig. 1C). The siRNA-mediated knockdown of Sirt1 in HT22 cells reduced the expression of tBdnf, Bdnf1, Bdnf4, and Creb1, and under the suppression of Sirt1, L-EV treatment did not increase the expression of Bdnf4 and Creb1 (Fig. 1D and 1E). These results suggest that Sirt1 plays a role in L-EV-induced upregulation of Bdnf4 and Creb1.

Next, we examined whether L-EV treatment in mice affected BDNF expression in the brain and stress-induced depressive-like behaviors. Mice treated with restraint for 2-h daily for 14 days (CRST; 2h × 14d RST) exhibited depressive-like behaviors [37], and had reduced expression of BDNF in the hippocampus [50]. Consistent with these reports, mice treated with CRST had reduced expression of Bdnf1, Bdnf4, and Nt4/5 in the hippocampus compared to that of control mice (Fig. 2A and 2C). Conversely, mice treated with L-EVs prior to each restraint (RST+L-EV; Fig. 2A) had increased expression of Bdnf1, Bdnf4, and Nt4/5 compared to that of CRST mice (Fig. 2A and 2C). Western blot analysis also indicated that proBDNF levels were reduced after CRST treatment, whereas L-EV injection at each stress session blocked the stress-induced decreased expression of proBDNF (Fig. 2D).

The body weight of mice decreased during restraint treatment, but was not affected by L-EV treatment (Fig. 2B). Mice treated with CRST tended to have reduced expression of Sirt1 in the hippocampus, whereas L-EV treatment during the stress session blocked the stress-induced reduced expression of Sirt1 (Fig. 2E). In contrast, there was no change in Hdac2 and Creb1 expression. Behavioral tests indicated that mice treated with CRST exhibited increased immobility in the TST and FST (Fig. 2F and 2G). In contrast, L-EV treatment prior to each restraint (Fig. 2A) blocked stress-induced increased immobility in the TST and FST (Fig. 2F and 2G).

These results suggest that L-EV treatment in mice during the stress induction phase blocks stress-induced reduced expression of BDNF in the hippocampus and stress-induced depressive-like behaviors.

Next, we examined whether L-EV treatment in the post-stress phase produced antidepressant-like effects. Mice treated with restraint for 2-h daily for 14 days had reduced expression of tBdnf, Bdnf4, and Nt3, and tended to have reduced expression of Bdnf1 and Nt4/5 in the hippocampus compared to that of control mice on post-stress day 35 (p35) (Fig. 3A and 3C). Thus, stress-induced changes in the expression of these neurotrophic factors in the hippocampus were long-lasting. On the contrary, L-EV treatment in CRST-treated mice blocked the reduced expression of tBdnf, Bdnf1, Bdnf4, and Nt4/5 in the hippocampus (Fig. 3A and 3C). CRST-induced body weight changes occurred, whereas post-stress treatment with L-EVs or IMI produced no significant effect on body weight (Fig. 3B). Western blot analysis indicated that mice treated with CRST had reduced expression of proBDNF in the hippocampus, whereas post-stress treatment with L-EVs in CRST-treated mice suppressed stress-induced reduced expression of proBDNF and BDNF (Fig. 3D).

Mice treated with CRST showed reduced social interaction in the sociability test (Fig. 3E and 3F) and increased immobility time in the TST and FST (Fig. 3G and 3H). In contrast, post-stress treatment with L-EVs in CRST-treated mice reversed decreased sociability in the sociability test, and reduced stress-induced increased immobility in the TST and FST (Fig. 3E~H). These data indicated that antidepressant-like effects of L-EVs are comparable to those of imipramine (Fig. 3E~H).

Next, we examined whether antidepressant-like effects of L-EVs were sustained long after the termination of L-EV treatment (Fig. 3A). When reassessed on post-stress days 29–30, mice treated with CRST exhibited depressive-like behaviors in the sociability test (Fig. 3I and 3J) and in the TST and FST (Fig. 3K and 3L), whereas mice treated with CRST followed by L-EVs in the post-stress phase showed increased sociability and reduced immobility in the TST and FST, which were also comparable to those of mice treated with imipramine (Fig. 3I~L). These results suggest that L-EVs produce antidepressant-like effects in mice with stress-induced depression, and L-EV-induced antidepressant-like effects are stably maintained.

In the present study, we demonstrated that L-EV treatment in mice showing stress-induced depression increased BDNF expression in the hippocampus and produced antidepressant-like effects (Fig. 2 and 3). The results of antidepressant-like effects of L-EVs are consistent with the previous reports showing beneficial effects of Lactobacillus spp. on various mental disorders. For example, the supplement of Lactobacillus rhamnosus reduced stress-induced increase of corticosterone and anxiety- and depression-related behavior [8] and decreased stress-induced anxiety-like behavior and deficits in social interaction [51]. Treatment of Lactobacillus helveticus NS8 in hyperammonemia rats attenuated the level of inflammatory markers, restored cognitive function and improved anxiety-like behavior [52]. Furthermore, administration of Lactobacillus helveticus NS8 produced antidepressant effects in rats subjected to chronic restraint stress [10]. Oral ingestion of Lactobacillus paracasei in senescence accelerated mouse prone 8 (SAMP8) mice delayed age-related cognitive decline, possibly by preventing oxidation and inflammation [53]. Lactobacillus plantarum C29 treatment increased hippocampal BDNF and p-CREB expression and protected scopolamine-induced memory deficit in the Y-maze and Morris water maze tests [54]. Furthermore, supplementation of Lactobacillus plantarum prevented stress-induced depressive-like behaviors in mice [35], and attenuated anxiety-related behavior in zebrafish [34]. Thus, given that supplementation of Lactobacillus spp. produces beneficial effects on cognitive and emotional dysfunction in various experimental conditions, it is important to note that L-EVs are sufficient to produce antidepressant effects. The antidepressant-like effects of L-EVs on behaviors were comparable to those of IMI (Fig. 3), whereas L-EV effects on the expression of neurotropic factors and TrkB were only partially overlapped with those of IMI, raising the possibility that L-EVs might exert antidepressant-like effects in a way slightly different from that of IMI, and therefore L-EVs could provide something that IMI does not afford.

The results of EVs-induced increase of BDNF and other genes in HT22 cells (Fig. 1) indicate that EVs can induce genomic responses by directly acting on cells. Considering that intraperitoneally (i.p.) injected L-EVs in mice induced genomic responses in the brain that were similar to those observed in HT22 cells (Fig. 1, 2, 3), it is possible that i.p.-injected EVs directly acted on the brain. It is well known that i.p.-injected drugs are normally absorbed into the mesenteric veins that are carried, via the hepatic portal vein, to the liver, and then they enter into the systemic circulatory system via the heart. It might be possible that injected L-EVs in mice immediately enter into the circulatory system. Consistent with this speculation, recent studies reported that fluorescent lipophilic dyelabeled EVs, when treated orally (p.o.), intravenously (i.v.), subcutaneously (s.c.) or intraperitoneally in all cases, are accumulated in many organs in the body, including the brain, liver, and kidney, although accumulation levels vary among different regions. For examples, DiD dye-labeled lung cancer cell (LL-2 cell)-derived EVs i.p. injected in C57BL/6 mice were detected in the brain 24 h after treatment [55]. Staphylococcus aureus-derived EVs [56] or Helicobacter pylori-derived EVs (100 µg/100 µl) [57] treated in C57BL/6 mice were also found in the brain 24 h after treatment.

L-EV treatment in HT22 cells increased BDNF expression, and this L-EV-induced BDNF expression was in part mediated by Sirt1, but not HDAC2 (Fig. 1). Thus, our results are consistent with the notion that L-EVs in HT22 cells induced BDNF expression through a transcriptional mechanism. L-EV treatment in CRST-treated mice also increased BDNF and NT4/5 expression in the hippocampus (Fig. 3C and 3D) and these L-EV effects were sustained for over 21 days after termination of L-EV treatment (Fig. 2C and 3C). These results suggest that L-EVs exert persistent changes in the expression of neurotropic factors.

BDNF expression is reduced in the hippocampus of postmortem samples from major depressive disorder patients and in mice with stress-induced depression [58,59]. In contrast, depressive behavior is reversed by administration of recombinant BDNF in mice [60]. BDNF expression is regulated by epigenetic factors such as Sirt1, HDACs and MeCP2 in stress-induced depression models [61,62]. Sirt1 has a deacetylase activity [48]. Sirt1 indirectly regulates BDNF and CREB expression by decreasing miR-134 expression [48,63]. Sirt1 can promote axon development and dendritic arborization [63,64,65]. Sirt1 expression is conversely regulated by miRNAs or other factors [48,63,66]. Sirt1 is reduced in the hippocampus of mice displaying depressive-like behaviors induced by chronic unpredictable mild stress (CUMS), whereas its activation reversed depression-like behaviors [64]. In rats exposed to CUMS, depression-like behaviors were reversed by resveratrol, which increased the expression of Sirt1, CREB, and BDNF in the hippocampus, while decreasing miR-134 [67]. As demonstrated in the present study, Sirt1 was also reduced in the hippocampus of mice exposed to chronic restraint stress, whereas L-EV treatment increased Sirt1 and BDNF (Fig. 2 and 3). Sirt1 inhibition in HT22 cells reduced the expression of tBdnf, Bdnf1, Bdnf4, and Creb1. Sirt1 inhibition in L-EV-treated HT22 cells reduced the expression of Bdnf4 and Creb1, compared to that of L-EV-treated HT22 cells, suggesting that Sirt1 may have a role in L-EV-induced upregulation of Bdnf4 and Creb1 (Fig. 1D and 1E). However, L-EV-induced increase of tBdnf and Bdnf1 was not completely blocked by siRNA-Sirt1 (Fig. 1D and 1E). These results suggest that L-EVs might increase BDNF expression via Sirt1-dependent and Sirt1-independent mechanisms. MeCP2 is an epigenetic factor that can regulate BDNF expression [68,69]. It may be worthwhile to examine whether MeCP2 is involved in L-EV effects on BDNF expression in neuronal cells.

Which factor contained in L-EVs increases the expression of Sirt1 is not studied in the present study, but it is a challenging question that needs to be solved. EVs contain proteins, lipids, DNAs, RNAs, enzymes, bacterial metabolites, and toxins [70,71]. The genus Bacteroides secreted EVs containing glycosidases and proteases that digested glycans which was used for growth [72]. Bacteroides fragilis-derived EVs contained PSA, which prevented colitis by enhancing immunomodulatory effects [73]. EVs derived from Pseudomonas aeruginosa contained bacteria DNAs, which were transferred to lung epithelial cells and were amplified by PCR in the nuclear fraction of cells [74]. Thus, bacterial EVs that contain proteins, lipids, DNA/RNAs, or bacterial metabolites might have a potential to exert physiological effects on host cells. Regarding the dramatic changes in gene expression alteration and behavioral rescues by EVs (Fig. 1, 2, 3), the detailed mechanisms how bacteria-derived EVs affect brain function remain to be explored further.

In conclusion, L-EV treatment increases the expression of BDNF in hippocampal neurons and produces antidepressant-like effects. Thus, our results support the notion that L-EVs and their direct action on neuronal cells mediate antidepressant effects of Lactobacillus spp.