Support for Natural Small-Molecule Phenols as Anxiolytics

Natural small-molecule phenols (NSMPs) share some bioactivities. The anxiolytic activity of NSMPs is attracting attention in the scientific community. This paper provides data supporting the hypothesis that NSMPs are generally anxiolytic. The anxiolytic activities of seven simple phenols, including phloroglucinol, eugenol, protocatechuic aldehyde, vanillin, thymol, ferulic acid, and caffeic acid, were assayed with the elevated plus maze (EPM) test in mice. The oral doses were 5, 10 and 20 mg/kg, except for phloroglucinol for which the doses were 2.5, 5 and 10 mg/kg. All tested phenols had anxiolytic activity in mice. The phenolic hydroxyl group in 4-hydroxycinnamic acid (4-OH CA) was essential for the anxiolytic activity in the EPM test in mice and rats compared to 4-chlorocinnamic acid (4-Cl CA). The in vivo spike recording of rats’ hippocampal neurons also showed significant differences between 4-OH CA and 4-Cl CA. Behavioral and neuronal spike recording results converged to indicate the hippocampal CA1 region might be a part of the anxiolytic pathways of 4-OH CA. Therefore, our study provides further experimental data supporting NSMPs sharing anxiolytic activity, which may have general implications for phytotherapy because small phenols occur extensively in herbal medicines.


Rat Surgery
On the day of surgery, rats were anesthetized using pentobarbital sodium at 50 mg/kg, intraperitoneally (i.p.), and restrained in a stereotactic apparatus (Narishige Co., Tokyo, Japan). The subcutaneous layer of tissue was removed to expose the skull. The implantation position (4.5 mm posterior to bregma, 3.5 mm laterally) for electrode arrays (16 channels) (Plexon Inc., Hong Kong, China) was set according the rat brain atlas drawn by the Paxinos and Watson, and then a hole at this position was drilled in the skull. A microdrive was positioned and the electrode array was lowered through the drilled hole into the CA1 of the left dorsal hippocampus (−2.5 mm relative to the brain surface, CA1 region). The gaps between the electrodes and hole were filled with softened paraffin and the microdrive was secured with dental cement. After completing the surgery, antibiotic penicillin (75,000 U) was i.p. administered for 3-5 days to prevent possible infections. One week after surgery, the EPM test was performed simultaneously with multi-channel in vivo extracellular recordings in the hippocampus CA1, 30 min after the rats received 4-OH CA (14 mg/kg), 4-Cl CA (14 mg/kg), or vehicle.

In Vivo Electrophysiological Recording
Methods were similar to those described previously [16]. In brief, neuronal spikes, filtered at 7-400 kHz and digitized at 40 kHz, were recorded during the entire experimental process using the multichannel acquisition processor system (OmniPlex, Plexon Inc., Dallas, TX, USA.). Spike units were isolated using the Plexon Offline Sorter Software 1.0. The actual recording position was marked by passing a 10-s 20 mA current through two selected electrodes. Figure 3A shows an illustration of the spike unit used for quantification in the in vivo electrophysiological recording. A successful 5-min recording was used to spike counting and frequency calculation.

Statistical Analysis
Data are expressed as means ± SEM. Behavioral data were analyzed by one-way/two-way analysis of variance (ANOVA) followed by Dunnett's t-test. Chi-square tests were used to analyze electrophysiological data. p < 0.05 was considered significant. Statistical software was the SPSS 18.0 (SPSS Inc., Chicago, IL, USA).

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by comparing the activity of 4-hydroxycinnamic acid (4-OH CA) and 4-chlorocinnamic acid (4-Cl CA) ( Figure 1) in the EPM test in mice and rats at a comparable molar dosage. The experiment data are shown in Figure 2 (A and B for mice; C and D for rats).

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by comparing the activity of 4-hydroxycinnamic acid (4-OH CA) and 4-chlorocinnamic acid (4-Cl CA) ( Figure 1) in the EPM test in mice and rats at a comparable molar dosage. The experiment data are shown in Figure 2 (A and B for mice; C and D for rats).

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of smallmolecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by vehicle as revealed by Dunnett's t-test. Eugenol and thymol were tested during the same time course so that they shared the same vehicle control, as were ferulic acid and caffeic acid.

Anxiolytic Differences in 4-OH CA and 4-Cl CA
Summarizing the above results and through reviewing literature on the bioactivities of small-molecule phenols, we focused on the role of the phenol hydroxyl (OH) group in their anxiolytic activities. The essential anxiolytic activity of the phenol moiety in NSMPs was demonstrated by comparing the activity of 4-hydroxycinnamic acid (4-OH CA) and 4-chlorocinnamic acid (4-Cl CA) (Figure 1) in the EPM test in mice and rats at a comparable molar dosage. The experiment data are shown in Figure 2 (A and B for mice; C and D for rats).
In the mouse EPM test, one-way ANOVA revealed significant differences among treatment groups for OT (F(7,79) = 3.744, p < 0.01; Figure 2A) and OE (F(7,79) = 5.154, p < 0.001; Figure 2B In the mouse EPM test, one-way ANOVA revealed significant differences among treatment groups for OT (F(7,79) = 3.744, p < 0.01; Figure 2A) and OE (F(7,79) = 5.154, p < 0.001; Figure 2B). The post hoc tests (Dunnett's t-test) revealed that mice spent more time in the open arms and went into the open arms more frequently after administration of 4-OH CA (20 and 40 mg/kg), compared to mice in the vehicle group. Conversely, 4-Cl CA had no anxiolytic effects at corresponding doses. One trial of two-way ANOVA between the 4-OH CA and 4-Cl CA groups (treatment × doses) revealed significant drug treatment effects, as shown by the overall significant differences in OT (F(1,16) = 5.504, p < 0.05; Figure 2A) and OE (F(1,16) = 6.192, p < 0.05; Figure 2B). No significant differences existed in the effects of the doses.
To verify the above experimental results, the anxiolytic effects of 4-OH CA and 4-Cl CA were assayed using the rat's EPM ( Figure 2C,D). The results revealed that 4-OH CA significantly increased OT (F(7,64) = 7.688, p < 0.001; Figure 2C) and OE (F(7,64) = 8.976, p < 0.001; Figure 2D), indicating significant differences between the groups. For rats, the post hoc tests revealed that OT and OE in the diazepam (1 mg/kg) and 4-OH CA (7 and 14 mg/kg) groups were significantly higher than those in the vehicle group. Two-way ANOVA the between 4-OH CA and 4-Cl CA groups (treatment × doses) revealed significant drug treatment effects, as shown by the overall significant differences in OT (F(1,16) = 24.326, p < 0.001; Figure 2C) and OE (F(1,16) = 29.869, p < 0.001; Figure 2D). No significant differences were observed in the dose effects. The results in rats showed that diazepam and 4-OH CA were anxiolytic in the EPM test, consistent with the results from the mouse experiments.

Electrophysiological Differences in 4-OH CA and 4-Cl CA
The differences in 4-OH CA and 4-Cl CA in neural activity were demonstrated by their effects on neuronal spikes in the rat hippocampus. After the rats recovered from the implantation of the electrode array into the hippocampal A1 region, the anxiety-related behaviors and spikes of the hippocampal neurons were simultaneously monitored ( Figure 3A). As shown in Figure 3C, 4-OH CA significantly increased OT compared to the vehicle and 4-Cl CA (14 mg/kg) groups, 30 min after intragastric administration (F(2,15) = 9.463, p < 0.01), indicating the implantation surgery did not affect the brain structures (hippocampal A1 region) in response to 4-OH CA. From the four rats under in vivo spike recording, 17 neurons with typical spikes ( Figure 3B) were selected to assay their responses to 4-OH CA (14 mg/kg) and 4-Cl CA (14 mg/kg). The responses were defined as an increase in spike frequency (at least double those of the vehicle group) or a decrease in spike frequency (50% of vehicle or lower). The data in Figure 3D are the quantitative judgment of response or non-response, showing that the difference between groups (4-OH CA and 4-Cl CA) was significant (p < 0.01, λ 2 -test). The data in Figure 3E-G are the spike responses from one hippocampal neuron after administration of vehicle, 4-OH CA, or 4-Cl CA, respectively, demonstrating 4-OH CA increased spike responses, but 4-Cl CA did not. The data in Figure 3G,H are the spike responses from another hippocampal neuron after administration of vehicle, 4-OH CA, or 4-Cl CA, respectively, demonstrating 4-OH CA decreased spike responses, but 4-Cl CA did not.  To verify the above experimental results, the anxiolytic effects of 4-OH CA and 4-Cl CA were assayed using the rat's EPM ( Figure 2C,D). The results revealed that 4-OH CA significantly increased OT (F(7,64) = 7.688, p < 0.001; Figure 2C) and OE (F(7,64) = 8.976, p < 0.001; Figure 2D), indicating significant differences between the groups. For rats, the post hoc tests revealed that OT and OE in the diazepam (1 mg/kg) and 4-OH CA (7 and 14 mg/kg) groups were significantly higher than those in the vehicle group. Two-way ANOVA the between 4-OH CA and 4-Cl CA groups (treatment × doses) revealed significant drug treatment effects, as shown by the overall significant differences in OT (F(1,16) = 24.326, p < 0.001; Figure 2C) and OE (F(1,16) = 29.869, p < 0.001; Figure 2D). No significant differences were observed in the dose effects. The results in rats showed that diazepam and 4-OH CA were anxiolytic in the EPM test, consistent with the results from the mouse experiments.

Electrophysiological Differences in 4-OH CA and 4-Cl CA
The differences in 4-OH CA and 4-Cl CA in neural activity were demonstrated by their effects on neuronal spikes in the rat hippocampus. After the rats recovered from the implantation of the electrode array into the hippocampal A1 region, the anxiety-related behaviors and spikes of the hippocampal neurons were simultaneously monitored ( Figure 3A). As shown in Figure 3C, 4-OH CA significantly increased OT compared to the vehicle and 4-Cl CA (14 mg/kg) groups, 30 min after intragastric administration (F(2,15) = 9.463, p < 0.01), indicating the implantation surgery did not affect the brain structures (hippocampal A1 region) in response to 4-OH CA. From the four rats under in vivo spike recording, 17 neurons with typical spikes ( Figure 3B) were selected to assay their responses to 4-OH CA (14 mg/kg) and 4-Cl CA (14 mg/kg). The responses were defined as an increase in spike frequency (at least double those of the vehicle group) or a decrease in spike frequency (50% of vehicle or lower). The data in Figure 3D are the quantitative judgment of response or non-response, showing that the difference between groups (4-OH CA and 4-Cl CA) was significant (p < 0.01, λ 2 -test). The data in Figure 3E-G are the spike responses from one hippocampal neuron after administration of vehicle, 4-OH CA, or 4-Cl CA, respectively, demonstrating 4-OH CA increased spike responses, but 4-Cl CA did not. The data in Figure 3G,H are the spike responses from another hippocampal neuron after administration of vehicle, 4-OH CA, or 4-Cl CA, respectively, demonstrating 4-OH CA decreased spike responses, but 4-Cl CA did not.

Discussion
The EPM is a classical model to assess anxiety-like behaviors in mice [17]. Using this test, the anxiolytic activity of phloroglucinol, eugenol, protocatechuic aldehyde, vanillin, ferulic acid, and thymol, were revealed for the first time, demonstrated by significant increases in the OT or OE compared to the vehicle control. The anxiolytic activities of caffeic acid [18] and 4-hydroxycinnamic acid [19] were consistent with previously-reported studies. Furthermore, the dominative role of the phenol moiety was proven by the substitution of an OH group in 4-hydroxycinnamic acid by chlorine atom (changing to 4-chlorocinnamic acid) leading to the loss of anxiolytic activity, which was observed at comparable dosages. In addition, a significant difference between 4-OH CA and 4-Cl CA was found in their effects on neuronal spikes in the rat hippocampal CA1 region, indicating that influencing the neuronal spikes in CA1 might be one of the anxiolytic pathways of 4-OH CA.
The above behavioral findings directed us to perform a systematic literature review. NSMP is specifically defined here as a natural compound whose molecular weight is less than 350 Dalton, and contains only carbon (C), hydrogen (H), oxygen (O), and at least one OH in the structure. Using SciFinder Scholar, we discovered 54 anxiolytic NSMPs as of June 2016 (a full list of NSMPs is not shown in this paper; part of them can be found be one of our review [20]). The anxiolytic activity of a single phenol (carbolic acid) has not been documented, perhaps due to its strong toxicity [21]. However, the reported anxiolytic NSMPs include those phenols whose molecular weight is as low as possible, such as orcinol [14], paeonol [6], gallic acid [22], carvacrol [8], and those in this study (Table 1). Phloroglucinol contains three phenolic hydroxyl groups and has no other substitutions on the phenyl skeleton. The structural diversity, other than phenol moiety of these NSMPs, demonstrated the key role of the phenol group. As shown in Table 1, and as with the 54 anxiolytic NSMPs identified from SciFinder database, these phenols have a narrow therapeutic window (several to several tens of mg/kg), indicating their efficacy may be strictly controlled by the chemical properties of phenol-OH. In other words, the moiety of phenol may act as a carrier of pharmacological signals. Another interesting finding suggests those phenols with relatively higher molecular weight, like kaempferol and quercetin, should be metabolized to smaller phenols, like (p-hydroxyphenyl acetic acid and 3,4-dihydroxyphenylacetic acid, for the expression of their anxiolytic effects [9]. Thus far, our hypothesis that NSMPs are anxiolytic is reasonable. Additionally, we hypothesize that small phenols are anxiolytic, and not only in natural products, because some synthesized small-molecule phenols, such as propofol and isobutylparaben, are also anxiolytic [23,24]. To determine the anxiolytic activities of NSMPs, their overlaps of pharmacological mechanisms should be checked. Several clarified mechanisms contribute to the effects of clinical anxiolytic drugs, such as their actions on reversing the imbalance of neurotransmitters, neuroendocrine dysfunction, or dysimmuneneuropathy [25]. In our previous research, we demonstrated that orcinol can penetrate the blood-brain barrier [14], indicating the possibility that NSMPs act on the central nervous system. For the anxiolytic mechanism, numerous references report that NSMPs have effects on neurotransmitters and responding receptors. GABA A receptor-mediated signaling may be the most probably NSMP target. The anxiolytic effects of effusol [26], honokiol [27], sinapic acid [28,29], obovatol [30], 4-hyroxybenzaldehyde [31], 6-hydroxyflavone [32], chrysin [33], apigenin [34], wogonin [35], p-coumaricacid [19], ellagic acid [36], and baicalein [37,38] were proven to be related to the GABA A receptor or its signaling cascade. The GABA A receptor may be one of the pharmacological targets of NMSPs. In addition to GABA signaling, other monoamine neurotransmitters may be involved. As anxiolytics, myricetin [39], epicatechin [40], 4-hyroxybenzyl alcohol [31], gallic acid [41], and cannabidiol [42], were reported to act through the serotonin pathway. Caffeic acid was found to be related to the adrenergic receptor [18], and danshensu worked through dopamine signaling [10]. Until now, no papers on NSMPs have been published revealing the anxiolytic mechanism beyond the monoamine neurotransmitter. Numerous brain regions have been identified associated with the expression of anxiety-like behaviors, including the hippocampus, amygdala, prefrontal cortex, and so on [43]. Judging from spike frequency, hippocampal neurons were inhibited or enhanced, or not affected after administration of 4-OH CA. From limitations of this study, we cannot draw a conclusion on which kind of response is related the anxiolytic effects of 4-OH CA. Results in Figure 3 show that hippocampus CA1 region may be involved in the anxiolytic mechanism of 4-OH CA, and possibly for other NSMPs. However, because of the response diversity of neuronal spikes and multiple hierarchical brain structures involved in anxiety modulation, it is difficult to say that the molecular targets (GABA A receptor and/or others) are located in hippocampal CA1 region.
Other bioactivities of NSMPs also support the idea of considering NSMPs as a whole. The typical anxiolytic benzodiazepines, such as diazepam, show sedative and muscle-relaxant effects at relatively high doses [44,45]. By searching the SciFinder database as of June 2016, we chose 43 NSMPs with sedative activities, and 73 NSMPs with muscle-relaxant/spasmolytic activities (Figure 4). A full list of these NSMPs is not provided in this paper. Dehydroeffusol [11,46], apigenin [34,47], chrysin [33,48], kaempferol [9,49,50], luteolin [51][52][53], quercetin [9,54,55], and ellagic acid [36,56,57] are anxiolytic, sedatives, and muscle-relaxants, similar to the pharmacological profile of diazepam. NSMPs have other common bioactive properties, as mentioned previously. Although likely to be disputed, we would like to introduce a new word, "phenolism", to pharmacologically cover these properties, including, but not limited to, anti-oxidation, anti-microorganism, anti-spasm, sedation, neuroprotection, and anti-anxiety activities. More cycles are provided in Figure 4, and maximum overlap is expected. 3 show that hippocampus CA1 region may be involved in the anxiolytic mechanism of 4-OH CA, and possibly for other NSMPs. However, because of the response diversity of neuronal spikes and multiple hierarchical brain structures involved in anxiety modulation, it is difficult to say that the molecular targets (GABAA receptor and/or others) are located in hippocampal CA1 region.

Conclusions
NSMPs extensively exist in herbal medicines, vegetables, fruits, or plant-source food. However, systemic studies are lacking on the effects of NSMPs on human anxiety when receiving a sufficient dose. Our hypothesis, although expecting more supportive evidence, may support the reasonable use of herbal medicines and new anxiolytic drug development. For example, we could attempt to discover new phenol-rich plants or their extractions to treat anxiety, or to control the everyday intake of NSMPs in food to modulate mood. Furthermore, our hypothesis may be helpful for understanding the pharmacological mechanisms of herbal medicine traditionally used to treat anxiety.