The alkaloids of Banisteriopsis caapi, the plant source of the Amazonian hallucinogen Ayahuasca, stimulate adult neurogenesis in vitro

Banisteriopsis caapi is the basic ingredient of ayahuasca, a psychotropic plant tea used in the Amazon for ritual and medicinal purposes, and by interested individuals worldwide. Animal studies and recent clinical research suggests that B. caapi preparations show antidepressant activity, a therapeutic effect that has been linked to hippocampal neurogenesis. Here we report that harmine, tetrahydroharmine and harmaline, the three main alkaloids present in B. caapi, and the harmine metabolite harmol, stimulate adult neurogenesis in vitro. In neurospheres prepared from progenitor cells obtained from the subventricular and the subgranular zones of adult mice brains, all compounds stimulated neural stem cell proliferation, migration, and differentiation into adult neurons. These findings suggest that modulation of brain plasticity could be a major contribution to the antidepressant effects of ayahuasca. They also expand the potential application of B. caapi alkaloids to other brain disorders that may benefit from stimulation of endogenous neural precursor niches.

B. caapi β-carbolines induce proliferation and growth in neurosphere cultures. We analyzed next whether the treatment with test compounds modulated proliferation capacity in the neurosphere. For this purpose, we grew free floating neurospheres under proliferative conditions, i.e., in the presence of epidermal growth factor (EFG) and fibroblast growth factor (FGF) for 7 days, with the addition of saline (control) or each of the four tested β-carbolines. As shown in Fig. 2, addition of the alkaloids to the cultures markedly increased the rate of formation and the size of the neurospheres.
After 7 days of β-carboline treatment, the number of SVZ-derived neurospheres was significantly higher than in the vehicle-treated cultures (F(4,15) = 466.512, p < 0.001). Analogous results were observed for neurospheres derived from the SGZ (F(4,15) = 1226.445, p < 0.001). Regarding the size of the neurospheres, significant increases were observed for both SVZ-and SGZ-derived neurospheres after β-carboline treatment, as compared with non-treated cultures (SVZ: F(4,15) = 126.858, p < 0.001); SGZ: F(4,15) = 172.617, p < 0.001). In view of Figure 1. Effects of ayahuasca β-carboline alkaloids on stemness of cultured adult neurospheres. Representative Western blots and bar graphs showing expression levels of the precursor cell markers musashi-1, nestin and SOX-2 after treatment with each of the four alkaloids tested (1 µM). Values in bar graphs indicate mean ± SD of the quantification of at least three independent experiments corresponding to four different cellular pools. The left side of the image shows results for the subventricular zone (SVZ) of the brain. The right side of the image shows results for the subgranular zone of the hippocampus (SGZ). *p ≤ 0.05; **P ≤ 0.01; ***p ≤ 0.001 indicate significant results in the post-hoc pair-wise comparisons (Bonferroni) versus non-treated (basal) cultures.  these results, we next analyzed the expression of Ki67, a marker of dividing cells, and "proliferating cell nuclear antigen" (PCNA) (Fig. 3).
B. caapi β-carbolines increase neural stem cell migration. In addition to proliferation and differentiation, neurogenesis involves the migration of neural stem cells and their integration into functional circuits. We thus tested next the effects of the β-carbolines on precursor migration. We treated the neurospheres with each of the test compounds and we monitored migration of the newly formed cells using livescanning microscopy for 48-72 h. The results are presented in Supplementary Figure 1 (and in the Supporting Information Videos 1-5). As shown therein, β-carboline treatment resulted in significant increases in migration capacity (SVZ: F(4,15) = 349.872, p < 0.001); SGZ: F(4,15) = 1203.666, p < 0.001) in comparison with basal (control). Neural stem cells moved long distances out of the neurosphere body in the presence of the four tested compounds. On the contrary, cells in the control cultures remained close to the neurosphere core.

B. caapi β-carbolines induce differentiation of neural stem cells.
We investigated the capacity of the β-carbolines to promote cell differentiation into any of the three different cellular types that form the central nervous system: astrocytes, neurons and/or oligondendrocytes. We analyzed the formation of the different cell types using immunocytochemistry. Neurosphere cultures prepared from the SVZ and the SGZ were adhered to a substrate and cultured under differentiation conditions, that is, in the absence of growth factors and in the presence of 1% fetal bovine serum. The β-carbolines were added to the medium and, after 3 days in culture, we evaluated the expression of Tuj1-and MAP-2-positive cells by immunocytochemistry analysis (Fig. 4a) and we quantified these proteins by Western blot (Fig. 4b). After the three-day period, vehicle-treated cultures only showed a few positively-stained cells for Tuj-1 or MAP-2. On the contrary, in the β-carboline-treated cultures, the number of both Tuj1-(SVZ: F(4,15) = 16.295, p < 0.001); SGZ: F(4,15) = 45.997, p < 0.001) and MAP-2-positive cells (SVZ: F(4,15) = 41.631, p < 0.001); SGZ: F(4,15) = 22.951, p < 0.001) was significantly increased. These results indicate that the test compounds induce the differentiation of neural stem cells towards mature neurons.

Discussion
Here we showed that adult neural stem cell activity is regulated by harmine, THH, and harmaline, the most abundant alkaloids in B. caapi and ayahuasca, and by harmol, the main metabolite of harmine in humans 31 . Using an in vitro model of adult neurogenesis, we found that all four β-carbolines stimulated the proliferation and migration of progenitor cells and promoted their differentiation predominantly into neurons.
The four compounds tested effectively promoted proliferation, migration, and differentiation of progenitor cells obtained from the SVZ and the SGZ, the two main niches of adult neurogenesis in rodents. The β-carbolines increased the number and size of primary neurospheres, induced the loss of the neurospheres' undifferentiated state, and promoted subsequent cell migration and differentiation mainly into a neuronal phenotype, as indicated by the positive expression of β-III-tubulin and MAP2, but also into astroglial cells. Taken together, these three effects indicate that B. caapi alkaloids have the capacity to regulate the expansion and fate of stem cell populations.
Analysis of the proliferation stage showed that all four β-carbolines increased the number and size of neurospheres, the number of Ki-67-stained cells, and the amounts of Ki-67 and PCNA protein as measured by Western blot. Our results for the effect of harmine on proliferation are in line with a previous study showing harmine-induced increase in mitosis in cultured chick embryo cells 32 , and in human neural progenitor cells 33 . To our knowledge, the effects of harmaline and THH on neurogenesis have not been studied before. While harmaline is only present in small amounts in B. caapi, THH is the second most abundant β-carboline in the plant 4,29 . Additionally, THH shows more consistent plasma levels between individuals and studies than harmine, which is rapidly degraded to harmol when taken orally 29,34 . The latter, formed in vivo by O-demethylation of the parent compound 29 , showed proliferative effects of similar magnitude to those of harmine.
Our results showed that B. caapi β-carbolines promoted cellular migration and differentiation, suggesting that these alkaloids not only act as mitogens for neural stem cells, but also modulate cellular fate. The largest effects on migration were observed for harmaline and THH. Increased migration capacity is relevant in certain conditions such as brain injury, where stem cell niches are far from the damaged area [35][36][37] . All tested compounds also promoted cellular differentiation. Neural stem cells are known to differentiate into neurons, astrocytes, and oligodendrocytes 23,24 . The observed increases in Tuj-1 and MAP-2 protein expression indicated differentiation predominantly toward a neuronal phenotype. In the SVZ both proteins were equally expressed after each of the four treatments. However, in the SGZ harmine administration did not influence Tuj-1 levels, a marker of immature neurons, but significantly increased the expression of MAP-2, suggesting a larger impact on neuronal maturation.  All the above indicate that B. caapi β-carbolines facilitate neurogenesis at multiple levels. This capacity is of interest, since in pathological conditions the replacement of neurons may be optimized by acting simultaneously on various processes 38,39 . The effect of the β-carbolines on cellular proliferation and differentiation is not unique to these compounds, having been observed for endogenous molecules such as leukotriene B4 40 , BMPs 41 , the growth factors EGF/FGF2 42 , and NGF/BDNF/bFGF 43 , and the transcription factors Lmx1a and Lmx1b 44 . However, the fact that the β-carbolines also stimulated migration highlights the versatility of these exogenous compounds, as they can promote the three processes involved in full adult neurogenesis.
A likely possible explanation for the observed effects of β-carbolines in neurogenesis is the increase in monoamine levels caused by MAO inhibition. With this said, we must acknowledge that the magnitude of the neurogenic effects was similar for the four compounds, despite harmol and THH being inhibitors that are between a hundred and a thousand times weaker than harmine or harmaline 5 . Moreover, the role of monoamines in neurogenesis is not fully understood. Knocking out the 5-HT 1A receptor in mice impaired neurogenesis after fluoxetine but not after imipramine, indicating that neurogenesis was independent from elevated serotonin levels 45 . In another study, the authors reported the unexpected finding that serotonin depletion actually promoted hippocampal neurogenesis instead of decreasing it 46 . In a recent paper, harmine, but not the MAO inhibitor pargyline, stimulated proliferation of human neural progenitor cells in vitro 33 . Harmine effects were mediated through inhibition of the DYRK1A kinase rather than through MAO inhibition. This opens the possibility that the β-carbolines tested here regulated stem cell fate via DYRK1A or other alternative mechanisms. To our knowledge, the inhibitory effects of harmaline, tetrahydroharmine and harmol on DYRK1A has not been examined. Other potential molecular targets for the neurogenic effects of small molecules include the modulation of the GSK-3β/β-catenin pathway 47 , upregulation of the brain-derived neurotrophic factor 48 , increased levels of vascular endothelial growth factor 49 ; and glucocorticoid receptor activation 50 . Future studies should assess whether B. caapi β-carbolines interact with one or more of these pathways.
Our findings have relevant therapeutic implications. The association between neurogenesis and anti-depressant activity is well documented 45,51 . Enhanced hippocampal neurogenesis reduces depression-like behaviors in animals 51 . Furthermore, clinically effective antidepressants stimulate neurogenesis, independent of their chemical structure and mechanism of action. To cite a few examples, chronic treatment with the serotonin reuptake inhibitor fluoxetine increases neurogenesis in rats 52,53 , as does chronic treatment with the selective MAO-A inhibitor pirlindole 53 . The association between neurogenesis and antidepressant effect is not limited to rodents and pharmacological interventions. Electroconvulsive therapy in primates also stimulates proliferation of neural precursors in the hippocampus and their differentiation into neurons 54 . Hippocampal neurogenesis appears to be necessary for antidepressant action. Irradiation of the SGZ of the hippocampus in mice prevents the neurogenic and behavioral effects of fluoxetine and imipramine 45 .
In humans, two recent clinical studies have demonstrated rapid and long-lasting antidepressant effects after a single ayahuasca dose in patients who did not respond to conventional treatment 11,12 . The therapeutic potential of ayahuasca is an area of increasing research interest beyond depression 55 . Alterations in adult neurogenic niches have been associated with a number of pathologies affecting the central nervous system [56][57][58][59] . Stimulation of these niches is currently being investigated as a novel therapeutic strategy for neuropsychiatric disorders [60][61][62] . Regular ayahuasca use has been associated decreases in problematic alcohol, cocaine and opiate consumption, indicating anti-addiction properties for B. caapi preparations 63,64 . These potential anti-addictive properties are particularly relevant if we acknowledge the notorious difficulty of treating substance use disorders. Drug-dependent patients not only show functional deficits in reward processing and cognitive control, but also structural alterations in brain gray and white matter 65 .
Our study has a series of limitations that need to be acknowledged. Ayahuasca brews contain other active compounds that were not tested here. A popular version of ayahuasca in the USA and Europe contains DMT, a serotonergic psychedelic 9 . It is possible that DMT may have contributed to the antidepressant effects reported in clinical studies using ayahuasca 11,12 . This contribution could be due to both brain plasticity mediated by 5-HT 2A receptor activation 66 and to the profound psychological experiences induced by psychedelics 67 . While studying DMT in the neurogenesis model was not an objective of the present investigation, it could be assessed in a future study, comparing it with other 5-HT 2A agonists such as psilocybin or LSD. Although several animal studies have already shown that harmine improves behavioral measures of depression 17,18 , future studies could ideally test the four compounds assessed here for both in vivo neurogenesis and behavioral improvement. Finally, future research could also use positive controls to compare the potency of the B. caapi β-carbolines with that of other antidepressants, such as SSRIs and MAO inhibitors.
In conclusion, here we showed that the β-carboline alkaloids present in B. caapi, the plant source of the ayahuasca tea, promote neurogenesis in vitro by stimulating neural progenitor pool expansion, and by inducing cellular migration and differentiation into a neuronal phenotype. The stimulation of neurogenic niches in the adult brain may substantially contribute to the antidepressant effects reported for ayahuasca in recent clinical studies. The versatility and full neurogenic capacity of the B. caapi β-carbolines warrant further investigation of these compounds. Their ability to modulate brain plasticity indicates their therapeutic potential for a broad range of psychiatric and neurologic disorders.

Methods
Ethics. Animals used in this study were cared for following the animal experimental procedures previously approved by the "Ethics Committee for Animal Experimentation" of the Instituto de Investigaciones Biomédicas and carried out in accordance with the European Communities Council, directive 2010/63/EEC and National regulations, normative 53/2013. Scientific RepoRts | 7: 5309 | DOI:10.1038/s41598-017-05407-9 Drugs. Harmine and harmaline were obtained from commercial sources (Sigma-Aldrich). Harmol and tetrahydroharmine were obtained by synthesis following the procedures described below.

Synthesis of harmol and tetrahydroharmine. 1-Methyl-9H-pyrido[3,4-b]indol-7-ol hydrobromide (har-mol·HBr).
To a solution of harmine (151 mg, 0.71 mmol, Sigma-Aldrich) in acetic acid (3 mL) hydrobromic acid (3 mL, 47% aqueous solution) was added. The mixture was heated at 160 °C during 25 min in a microwave oven and then allowed to cool to room temperature. After few minutes at 25 °C, needles started to form and harmol·HBr·H 2 O (180 mg, 85% yield) was recovered by filtration as a yellow solid of mp 250-251 °C. 1 Methoxy-1-methyl-2,3,4,9-tetrahydro-1H-pyrido [3,4-b]indole hydrochloride (tetrahy-droharmine·HCl, THH·HCl). Sodium borohydride (680 mg, 18.00 mmol) was added in portions to a stirred solution of harmaline (1.54 g, 7.20 mmol, Sigma-Aldrich) in water (40 mL) at 0 °C and the mixture was acidified until pH 2 with aqueous HCl (2 M). After 75 min at room temperature, the mixture was made alkaline with NaOH (10% aqueous solution), extracted with dichloromethane (3 × 20 mL) and evaporated to dryness. Solid was dissolved in isopropyl alcohol, heated and then treated with an excess of concentrated HCl until a precipitate appeared. After 3 days at 4 °C, THH·HCl was collected by filtration (1.40 g, 77% yield) as a white solid of mp 201-202 °C. 1  Adult precursor isolation. Neural stem cells (NSCs) were isolated from the subgranular zone (SGZ) of the hippocampus and the subventricular zone (SVZ) of the lateral ventricle of adult C57BL/6 mice (3 months of age) and prepared following previously described methods 68 . A total number of 40 animals were used throughout the whole study, divided in four pools of 10 animals each. Every pool was individually used to perform all the experiments described in this section. Briefly, tissue was carefully dissected, dissociated in DMEM medium with glutamine, gentamicin and fungizone and then digested with 0.1% trypsin-EDTA + 0.1% DNAase + 0.01% hialuronidase for 15 min at 37 °C. The isolated stem cells were seeded into 12-well dishes at a density of ~40,000 cells per cm 2 in DMEM/F12 (1:1) containing 10 ng/mL epidermal growth factor (EGF), 10 ng/mL fibroblast growth factor (FGF) and N2 medium.

7-
Neurosphere cultures and treatments. NSCs were cultured under standard conditions and in the presence of growth factors (EGF and FGF) for a week, in 12-well dishes. After this time, small neural progenitor-enriched growing spheres known as neurospheres (NS) were formed. At this point, with all NS having the same stage and size, cultures were treated daily for 7 days with vehicle or 1 μM solutions of each of the four β-carbolines tested: harmol, harmine, harmaline and tetrahydroharmine (THH). The effective dose of compounds was chosen based on our previous studies on β-carboline analogs 69 and the 7-day period was selected as a standard time frame used previously by our own and other research groups to test the neurogenic potential of drugs 37,50,68,70 . The 1 μM concentrations chosen for the β-carbolines was lower than the 7.5 μM found to induce optimal proliferation in a previous in vitro study with harmine 33 but closer to the plasma concentrations of 0.52-1.52 μM reported for harmine and THH, respectively, in a recent study involving ayahuasca administration to healthy humans 71 . At these concentrations, none of the tested β-carbolines affected the viability of the cultured cells. Proliferation and growth analysis was assessed from these cultures and the number and diameter of NS were quantified using the Nikon Digital Sight, SD-L1 software. All the NS contained in ten wells per condition were counted. Some of these proliferating NS were used for immunoblotting analysis, while others were seeded onto poly-L-lysine precoated-6-well plates, and/or -coverslips and cultured in the presence of β-carboline alkaloids (1 μM) under proliferation conditions (medium containing 1% fetal bovine serum and without exogenous growth factors). Once neurospheres were differentiated (72 h), those grown on coated 6-well plates were used for immunoblotting and those on coverslips for immunocytochemical analysis. The remaining NS were cultured for 48-72 h under differentiation conditions onto μ-Slide 8-well plates (Ibidi, Martinsried, Germany) and used for migration assays.
Statistical analysis. Data were analyzed using a one-way ANOVA with treatment as factor (control, harmol, harmine, harmaline, THH). Significant results in the ANOVA (p < 0.05) were followed by post-hoc pair-wise test (Bonferroni). The SPSS statistical software package (version 20.0) for Windows (Chicago, IL, USA) was used for all statistical analyses.