Drugs

The seven herbs (Radix Bupleuri, Pericarpium Citri Reticulatae, Ligusticum Wallichii, Radix Paeoniae Alba, Rhizoma Cyperi, Fructus Aurantii, and Glycyrrhiza uralensis Fisch; the ratio of which is 4:4:3:3:3:3:1) constituting BSS were purchased from the Department of Pharmacy of Xiangya Hospital of Central South University and authenticated by the herbal medicine botanist Professor Hu, Z.H., Nanjing University of Chinese Medicine. The original herbs were kept at the Institute of Common Disease and Depression, Nanjing University of Chinese Medicine, China. Meranzin hydrate (MH) (C₁₅H₁₈O₅, CID, 5070783; Lot, 5673-37-0) supplied by DIAO (Chengdu, China) was greater than 98% pure. Albiflorin (C₂₃H₂₈O₁₁, CID, 24868421; Lot, MUST-19041601), FA (C₁₀H₁₀O₄, CID, 445858; Lot, MUST-19032928), Naringin (C₂₇H₃₂O₁₄, CID, 442428; Lot, MUST-18050808), Hesperidin (C₂₈H₃₄O₁₅, CID, 10621; Lot, MUST-19030701), Glycyrrhizic acid (C₄₂H₆₂O₁₆, CID, 14982; Lot, MUST-19052407), Saikosaponin A (C₄₂H₆₈O₁₃, CID, 167928; Lot, MUST-19101401), Nobiletin (C₂₁H₂₂O₈, CID, 72344; Lot, MUST-19041210), Tangeretin (C₂₀H₂₀O₇, CID, 68077; Lot, MUST-18012910) and Glycyrrhetinic acid (C₃₄H₅₀O₇, CID, 10114; Lot, MUST-19032207) with ≥ 98% purity were obtained from Chengdu Must Bio-technology Co., Ltd. (Chengdu, China). HPLC/MS grade methanol, acetonitrile, and formic acid (CNW, USA) were used for the analysis. Deionized water purified by the Milli-Q system (Millipore, Billerica, MA, USA) was available. [D-Lys3]-GHRP-6 was received from Abcam Co, Cambridge, England, UK. DMSO, phenylephrine hydrochloride (PE), acetylcholine (Ach) chloride, WAY100635, NBQX, SB207266 and NG-nitro-L-arginine methyl ester (L-NAME) were produced by Sigma-Aldrich Co. St. Louis, MO, USA. During cell culture procedures, DMEM (Hyclone, USA), pancreatin digest (Hyclone, USA), fetal bovine serum (FBS) (Gibco, USA) and penicillin-streptomycin solution (Beyotime, Shanghai) were used.

Preparation of BSS, the BPSDS

The extraction of BSS was performed as described in our previous study [13]. Briefly, the seven herbs were mixed, soaked in 10-fold volumes of water for 30 min, refluxed twice for 30 min, and filtered to obtain the combined extract. Whereafter the blended liquid was concentrated, frozen rapidly, and dried in a freeze-dryer (Labconco, USA). The yield of lyophilized powder (BSS for short) was 27.80%.

BPSDS was prepared as follows. After 45 minutes of intragastrically administering 30 g·kg^–1 BSS or an equal dose of normal saline to rats [6], blood was collected from the accessory aorta of anesthetized rats and processed to obtain BSS post-dose rat serum (PRS) or blank serum (BS) by centrifuging at 3,500 rpm for 10 min. Afterwards, 200 µL of purified water was added to 1 mL of CMS or BS, mixed for 2 min, and placed in boiling water for 5 min to precipitate the protein. Then we homogenized the sample by grinding at 30 Hz for 2 min, centrifuged at 12,000 r min^–1 for 15 min at 4℃. The supernatant was transferred to another clean vial, forming BPSDS as one of the herbal ex vivo proxies. Another proxy was the previously mentioned representative herbal ACs.

Experimental animals

Sprague-Dawley rats (male, 220 ± 20 g) were purchased from Nanjing Qingzilan Technology Co., Ltd. All animal welfare and experimental procedures carried out were consistent with the guide for the Care and Use of Laboratory Animals approved by the Institutional Animal Care and Use committee at Nanjing University of Chinese Medicine (a 12 h light/dark cycle, 23 ± 2℃, and 60 ± 5% relative humidity). After acclimatizing to the surroundings for seven days, the rats were randomly divided into groups.

The rats (n = 6) got the intragastrical administration of BSS (30 g·kg^–1) and other drugs before narcosis through intraperitoneal injection of 3% pentobarbital sodium (30 mL·kg^–1). Then, rats were euthanized by drawing blood from the abdominal aorta. All animal experimental procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals approved by the Institutional Animal Care and Use Committee at Nanjing University of Chinese Medicine (Resolution No. 201912A008).

Qualitative and quantitative determination of 10 ACs in BSS and serum

A chromatographic separation was conducted on the Waters ACQUITY BEH C18 column (2.1 × 100 mm, 1.7 μm) maintained at 35℃. The mobile phase comprised 0.1% formic aqueous solution (A) and acetonitrile solution (B) at a gradient program as follows: 0.0–1.0 min, 10% B; 1.0–3.5 min, 10–20% B; 3.5–5.5 min, 20–26% B; 5.5–21.0 min, 26–95% B; and 21.0–25.0 min, 95% B. The flow rate was set at 0.4 mL/min, and the injection volume was 2 μL.

MS data were acquired using the Waters Xexo G2-XS QTOF system (Waters, USA) equipped with an electrospray ionization (ESI) source. The ionization voltage was set at 1 kV. Other parameters were set as follows: source offset, 50 L/h; source temperature, 120℃; desolvation temperature, 450℃; and desolvation gas flow rate, 800 L/h. Analyte detection was carried out using MRM in a positive mode. The precursor-to-product ion transitions, acquisition duration, con voltage (Con), and collision energy were shown in Tables S1 and S2. Relevantextract ion flow chromatograms of 10 ACs in mixed reference solution (A) and BSS decoction sample (B) and internal standard (IS) were shown in Figure S1.

Calculating compound doses

Various doses of BSS (32.58, 65.16, and 130.30 mg/mL) were designed as ex vivo doses for isolated aortic baths or H-ECs. The present and previous 30 g/kg [13] single in vivo BSS doses were multiplied by the coefficient of 0.25 to determine the polyphenol doses for rats (0.25 kg) as 7.5 g. Next, the above total dose (7.5 g) following complete absorption (100% bioavailability) was divided by the blood volume (16 mL for rats, respectively) to obtain 468.75 mg/mL of blood, similar 100% bioavailability. The 468.75 mg/mL blood was multiplied by the yield of lyophilized powder (27.80%) to acquire an in vitro dose of 130.30 mg/mL crude extract in blood. Thus, the ex vivo dose (mg/L) of BSS at 32.58 mg/mL and 65.16 mg/mL equaled ~25% or ~50% bioavailability. And according to the above calculation, the L-, M-, and H-doses of BPSDS in rats were designated as 7.5, 15 and 30 g/kg. Our ACs each ex vivo and in vivo doses equated their contents in PRS and BSS.

NO, eNOS and ROS in ECs

Sprague Dawley rat aortic ECs were provided by Procell Life Science & Technology Co., Ltd. and cultured in DMEM supplemented with 10% FBS + 1% penicillin/streptomycin. The medium was replaced every two to three days. We performed the cell passage when the confluence of cells reached about 80% and digested the cells using trypsin. Rat aortic endothelial cells (RAECs) were cultivated in 96-well plates, washed with phosphate-buffered saline (PBS) three times, and incubated at 37°C for 24 h. Then ECs were stimulated as a cellular model of oxidative stress by incubating with 0.5 mM H₂O₂ for 12 h. BPSDS, 10 ACs, alone or in different mixes whose dose equals the determined content in serum, as shown in Table 1, were added into the culture medium 45 mins before adding H₂O₂. The cell supernatant was collected, and the levels of eNOS and NO were detected using colorimetry and ELISA kits, while the ROS level was detected by flow cytometry. Relevant results are shown in Tables 1 and 2.

Establishment of multiple cell models

Rat brain microvascular endothelial cells (BMECs): Cells were isolated from postnatal day 7–14 (P7–P14) Sprague-Dawley (SD) rat pups of SPF grade, without distinction of sex. Rat hippocampal microglia (HM): Cells were obtained from postnatal day 1 (P1) Sprague-Dawley (SD) rat pups of SPF grade, without distinction of sex. Rat hippocampal neurons (HN): Cells were prepared from postnatal day 1 (P1) Sprague-Dawley (SD) rat pups of SPF grade, without distinction of sex. Rat HN, HM and BMECs were isolated, purified and cultured according to the method of neurovascular unit (NVU). After seeding neurons into six-well culture plates for 3–5 days, purified astrocytes were seeded to the outside of the well inserts, and then the well inserts were placed into the corresponding six-well culture plates. After 1 day of co-culture, purified BMECs were seeded to the inner side of the well inserts, and the NVU was successfully established. Specific markers such as neuron-specific enolase, glial fibrillary acidic protein and platelet adhesion molecule-1 (CD31) were used to identify HN, HM and BMEC. After treatment with 200 μM corticosterone for 18 h, MH treatment was given for 45 min to establish the cell model. The levels of 5-HT (5-hydroxytryptamine), BDNF and IL-6 (interleukin-6) were measured using the appropriate kits.

10⁵ gastric smooth muscle cells and 400 μL complete medium were added to sterile centrifuge tubes and incubated at 37°C for 24 h, followed by a 45-min MH treatment. Ach at a final concentration of 0.1 μM was then added to each tube for 10 min for detection. MCP-1 (monocyte chemoattractant protein-1), ET-1 (endothelin-1), and IL-6 levels were measured using the appropriate kits.

Forced swimming test (FST)

In the previous study [3, 8], we have evaluated the antidepressant effect through FST, which was described by Porsolt et al. [14]. The rats were randomly divided into group with 8 rats in each group. Rats were individually forced to swim twice at 24 h intervals in a vertical glass cylinder (40 cm high, 18 cm in diameter) filled with water (25 ± 2℃) up to 30 cm. On the first day, each rat was subjected to a pre-swimming for 15 min to induce the AFS model. The immobility time in 5 min was measured by putting the rats into the same glass cylinder individually, 30 min after administration. All conductors were arranged from 9:00 a.m. to 11:00 a.m. The water was changed between the two trials. The rats were considered immobile when they showed behavior of floating motionlessly or merely making actions to prevent drowning through the TopScan Version 3.0 software.

Changes in velocity-time integrals (VTI) by BPSDS vs. absorbed ACs

The left anterior descending coronary artery (LAD) diastolic flow VTI were achieved by analyzing and detecting the diastolic phase of the Doppler signals and parameters using image analysis software to evaluate CF. The Vevo 2100 system (Visual Sonics, Toronto, Canada) was used to perform all ultrasound exams to establish their in vivo bioactive link and to further study BSS-induced vasodilation. Each rat was given 3% inhalant isoflurane for anesthesia, and maintained with 1.0–1.25% isoflurane in oxygen-enriched air (30% oxygen/70% air) with spontaneous ventilation. The hair of the rats’ chest walls was carefully removed, and the chest wall was smeared with ultrasonically coupled fluid for ultrasound imaging examination. VTI is often used as a typical indicator of CF or CFMD.

Changes in vascular tension by BPSDS vs. absorbed ACs

The thoracic aortas were quickly separated out from the anesthetized rats and then placed in a precooled dissecting tray filled with Krebs-Henseleit (K-H) buffer [the composition (mM) of which is NaCl 119, KCl 4.7, CaCl₂ 2.5, KH₂PO₄ 1.2, MgSO₄ 1.2, NaHCO₃ 25, glucose•H₂O 11, and EDTA-2Na 0.5, pH 7.4] at 37℃. The vascular rings, latch onto which the tissue and fatty tissue were removed, were cut into small pieces 3–4 mm in length. Soon afterwards, the vascular rings were individually fixed by two stainless steel hooks and placed in a 20 mL K-H buffer, continuously oxygenated with a gas mixture (95% O₂ and 5% CO₂) inside and heated by water outside. One hook was attached to the bottom of the bath, and the other was connected to the isometric transducer (model MLT0420, AD Instruments, Sydney, Australia). Then we acclimatized the rings to the in vitro environment preloaded with 1,000 mg for 1 h, until the contractility was stable. The contractility was recorded and analyzed in a monitor using Power Lab software (https://www.adinstruments.com/products/powerlab-daq-hardware). The K-H buffer was replaced to wash every 15 min during stabilization. The vascular activity was verified by stimulating the rings to obtain a reproducible shrinkage reaction with 60 mM KCl. Whereafter, we induced ＞ 80% vasodilation of rings constricted with 1 μM phenylephrine three times to bear out the function of the endothelium. After equilibrating for 30 min, we contracted rings by using phenylephrine again to curve the concentration relaxation curve of the ex vivo proxies, BSS, BPSDS, and ACs, alone or in combination. The doses of the components were consistent with their contents in serum.

The effects on gastrointestinal motility in vivo

Motility was estimated by determining GE and IT. Firstly, each rat was forced to swim for 15 min to induce acute stress before gastrointestinal motility tests, and intragastrically administered with 1 mL Evans blue 30 min after drug administration (dissolved in 0.9% NaCl with 0.5% methylcellulose at the concentration of 50 mg/mL). The stomachs of euthanized rats were cut from the esophageal sphincter to the pyloric sphincter. The absorbance at 565 nm was determined with the Infinite M 200 Pro as previously reported by De Winter et al. [15].

The absorbance of Evans from rats in the control group was obtained immediately as a standard stomach control. The GE was acquired as follows:

Where:

“A565 reference” is the abbreviation of the absorbance of the reference stomach at 565 nm; “A565 sample” is the abbreviation of the absorbance of the sample stomach at 565 nm.

The IT was determined by measuring the distance travelled by the Evans blue from the pylorus and the total length of the small intestine. The IT was acquired as follows:

Mechanism involved in AGEV polypharmacology

Fluoxetine (FLX), Ach, and mosapride (MOS) were added, respectively, into RAECs at 10 μM as relevant positive controls. Besides, to verify the relevant signaling path, the MF (MH + FA), HS (hesperidin + Saikosaponin A), and HST (HS + tangeretin) were chosen to separately emphatically explore the influence of suppressant, demonstrating their uniqueness and making way for the discovery of lead compounds. The 100 μM [D-Lys3]-GHRP-6 (the inhibitor of growth hormone secretagogue receptor, GHSR), 100 μM L-NAME (the inhibitor of eNOS), 100 μM PP2 (the inhibitor of Src), 100 μM WAY100635 (the inhibitor of 5-HT1A), 100 μM NBQX (the inhibitor of AMPA), and 100 μM SB-207266 (the inhibitor of 5-HT4) were added separately into RAECs 30 min before administration. In addition, the rats were injected with saline (1 mL/kg, s.c.) or relevant antagonists at a dose of 0.5 mg/kg 30 min before administration.

Statistical analysis

Data were shown as the mean ± SD and analyzed using GraphPad Prism 8.0 software (GraphPad Software, San Diego, California, USA, www.graphpad.com). For comparing multiple groups, one-way ANOVAs followed by Q tests were used. Comparisons between two groups were made using Student’s t-tests. Statistical significance was assumed when p < 0.05.

To quantitatively compare different readouts from different tissues, organs and experimental levels, several normalization formula were used in this study. Model raw normalization increment (mrni) was used to indicate the deviation degree of the Vehicle (V)/model group from the Sham (S) group, with different formulas applied according to whether the V value was lower or higher than the S value. Drug raw normalization increment (drni) was used to evaluate the relative change produced by drug treatment on the basis of the V/model group. Phytochemical raw normalization increment (prni) was used to evaluate the relative effect and contribution of phytochemical/monomer treatment under the model background. On this basis, the CR was further calculated to estimate the relative contribution of a monomer or absorbed constituent to its parent herb/formula, and the rate of deviation (RD) was used to show the percentage difference between prediction and observation. These indices were introduced to reduce the fluctuation-related bias of raw values and to make comparisons among endpoints with different change directions and different magnitudes more homogeneous.

CR of monomer to parent herbs is calculated as follows:

RD of prediction to observation is calculated as follows:

mrni formula is calculated as follows:

drni formula is calculated as follows:

prni formula is calculated as follows:

Temporal trajectories of five disease clusters

The ancient, multifunctional Shugan-guided BAPevive strategy [5]—after eliminating invalid compounds and correcting all orthodox database errors through normalization calculation—activates shared Ghr-driven SRNs in the Gan axis, enabling accurate tailoring of BSS’s five sets of multifunctional AC tools within half a year, unlike previous BAP-alone methods that took three years to identify 8 out of 10 ACs, some of which were invalid [3, 4]. The identified AGEV ACs not only anchor the 24-h post-AFS depressed rat phenotype (B₂) but also uncover additional temporal trajectories—including 1-h B₁C and 24-h GE disorders—that constitute the full brain₁-coronary-brain₂-gastro-enteric (B₁CB₂GE) five-disease trajectory, previously overlooked by researchers. Furthermore, we used normalization-calculated herbal multifunctional ACs to anchor multimorbidity cluster dimensions, where the G mrni increased to 15.31—the highest among B₁CB₂GE clusters—corroborating epidemiological findings that poor appetite is a refractory root cause of depression-induced cardiac events. Thus, four graphs representing popular and orthodox two-dimensional models of five diseases (with B₁ and B₂ depressive behaviors grouped into one) were holistically integrated into a single, three-dimensional representation of B₁CB₂GE trajectories. The mrni ranking was: G 15.31 > B₂ 1.17 > B₁ 0.85 > C 0.49 > E 0.32 (Figure 1). Interestingly, for the first time to our knowledge, AFS rats and mice—including data from nine prior non-random, repeatable post-AFS AGE-3 disease rat [3, 4, 6–11, 16, 17] experiments and newly observed time-dependent depressive and coronary disorders—have been tailored as multimorbidity model frameworks homogeneously anchored by multifunctional ACs. From epidemiological surveys, we concluded that antidepressants (e.g., SSRIs, esketamine) increased the incidence of GE-hypomotility by 67%, along with heightened risks of hypertension, treatment discontinuation, and cardiac events [18–22]. These adverse effects were significantly reduced by the Shugan-like BSS and its five sets of ACs (Table 3).

Selected statistically positive time points (in h) and their post-AFS rankings—based on integrated non-random epidemiological incidence—were used as criteria to identify multimorbidity clusters. These included primary immobility (0.083–1 h; ranked 1, 3, and 4), along with E+ CFMD positivity from 1–3 h (other indices were negative), and secondary immobility at 24 h (ranked 2), which was the only time point showing GE hypomotility (Figure 1). The peak values of parallel immobility and isolated increases in E+ CF observed at 0.083–1 h post-AFS (ranked 1, 3, and 4) are considered primary BC clusters. In contrast, immobility at 24 h (ranked 2) and GE hypomotility observed only at 24 h are designated as secondary multimorbidity clusters, forming the temporal 1–24 h B₁C-B₂GE multimorbidity trajectories. During 13 years at different cities and universities, including the present (Figure 1), research groups have performed multiple repeatability tests of 24 h BGE [3, 6–11, 16, 17] (Figure 1) to 1–24 h B₁C-B₂GE cluster trajectories under nearly similar conditions of modeling, constituting stable non-random multimorbidity clusters traits in AFS rats, and found results highly consistent with epidemiological clustering of depression, cardiac events, and poor appetite, with roughly temporal relation from multimorbidity patients [21, 23].

The Ghr-driven SRN

0.083, 1, 12 and 24 h post-AFS rat’s peak values synced with peak values of blood biochemicals of three in the immobility (CRH > MDA > IL-6, P < 0.05–0.001 vs. S), nine in coronary disorder (ROS > H₂O₂ > TMAO > SOD > MDA > NO > eNOS > ET-1 > MEP-1, P < 0.05–0.0001), eight in hypomotility (Ghr > 5-HT > DA > CRH > ACTH > TNF-α > IL-6 > IL-1β, P < 0.05–0.001) and three in reduced jejunum stripe tension (BDNF > CORT > NA) (Figure 2). The eight peaks ranked mrni of 0.083 h B-immobility 1.67 > 24 h B₂’s 1.17 > 1 h B₁’s 0.85 > 24 h ex vivo jejunum tension 0.79 > 1 h C’s 0.49 > 24 h G’s 0.36 (for refractory root plus butyrate 14.95 as 15.31) > 24 h E’s 0.32 > reduced psychopathology 0.28. Of these, (i) the highest immobility peak in 0.083 h post-AFS in rats heterogeneously synced with the increases in human plasma catecholamine post-multiple 6-s sprint interspersed with 30-s recovery [24], whereas lacking catecholamine evidence in the brain decays seconds post-stress [25]. 0.083 → 1 h → 24 h of ↓ butyrate + → +++ via dysbacteriosis is inversely related to immobility, and GE hypomotility – → + (Figure 1); (ii) the 1 and 24 h time points were selected as two inherent cornerstones of multimorbidity because they were associated with nine and three plasma biochemical peaks, syncing these from hippocampus, aorta and G-antrum (Figure 2B). BDNF immunoreactive bands differed across tissues, with a predominant band at ~28–35 kDa in aorta and ~14–15 kDa in gastric antrum. These signals were therefore quantified as the major tissue-specific BDNF-immunoreactive bands, partly similar to these past findings [6, 9]. This tissue-specific observation was interpreted within the B₁CB₂GE multimorbidity framework, whose depression-cardiovascular component is consistent with epidemiological evidence linking emotional triggers, depression, myocardial infarction, coronary heart disease, and broader multimorbidity [26–28].

Display full size

Figure 2. The synchronized profiles of all and part 20 plasma biochemical molecular levels [2 sets of values, plasma molecules (raw data and rni), one set/other organ]. (A) Ghrelin (Ghr), BDNF, TrkB, IL-6, TNF-α and MDA in hippocampus; (B) Ghr, 5-HT, BDNF, NA, DA, CRH, ACTH, CORT, ROS, H₂O₂, TMAO, SOD, MDA, NO, eNOS, ET-1, TNF-α, MCP-1, and IL-6 in plasma; (C) NO, BDNF immunoreactive band at ~28–35 kDa (putative proBDNF), TrkB, ET-1, and eNOS in aorta; (D) Ghr, BDNF immunoreactive band at ~14–15 kDa (putative mature BDNF), and TrkB in gastric antrum. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001, ^#### p < 0.0001 vs. Sham; ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001 vs. V-24 h. V: Vehicle; 5-HT: 5-hydroxytryptamine; BDNF: brain-derived neurotrophic factor; eNOS: endothelial nitric oxide; ET-1: endothelin-1; IL-6: interleukin-6; MCP-1: monocyte chemoattractant protein-1.

Through five functional ACs, the Shugan-like therapy captured both the dynamic features and quantification of post-AFS 1–24 h B₁C-B₂GE five-disease cluster dimensions, along with associated plasma biochemical molecule levels in the rat Gan axis, representing the dual or multidirectional post-AFS interactions of BCGE [29] (Figure 1). This serves as a paradigm for a multimorbidity modeling mechanism framework, uniquely positioned among 549 publications (9 involving three-disease models and 540 involving two-disease models). The elaborately tailored, BSS-derived multifunctional AC tools and their corresponding multi-disease targets function like a tenon-and-mortise mechanism—repeatable and precise—akin to how two ACs derived from the phytomedicine Fructus aurantii unlock antidepressant and prokinetic functions through modulation of 5-HT₃ and Ghr.

Ex vivo-to-in vivo extrapolation of multifunctional ACs as BSS/BPSDS proxies

We used the established three-step ex vivo-to-in vivo extrapolation framework for AC(s) as BSS proxies of BCGE multifunctions, including vasodilation, entero-gastro prokinetic and antidepressant activities (Figure 3A–C). Across the multi-cell models, the selected AC(s) showed highly comparable CRs to BPSDS/BSS in ex vivo and AFS rats in vivo, with 1st/3rd-vs.-2nd step differences all kept within a low range (Figure 3D), indicating that the inserted rapid ECs step could stably represent the multifunctional contribution profiles of multicellular systems and shared AFS rats (Figure 3A–D).

Display full size

Figure 3. The paragraph of ex vivo to in vivo extrapolation of AC(s) as BSS proxies for vasodilation/entero-gastro prokinetic/antidepressant [CEGB (Coronary-Entero-Gastro-brain) multi-function] in 1st multicells (standard), 2nd ECs (rapid evive), and 3rd multicells (validation). The workflow proceeds from left to right, showing ex vivo-to-in vivo extrapolation. In the top left part of Figure 3 is the image of the Cell Painting assay protocol of neurovascular unit (microvascular ECs, hippocampal microglia, hippocampal neurons), Ach-treated gastric smooth muscle cells, and ECs. In the right part of the image is the comparison of the contributions of ex vivo and in vivo administration. Up to down of (A) 1st standard evive → (B) 2nd rapid evive → (C) 3rd validation of evive, including each disease’s contribution, accuracy range, and differences between those 3 steps. (D) The contribution comparisons of ex/in vivo extrapolation (evive) of ACs as BPSDS/BSS proxies for multifunctions in 1st standard or 3rd validated multicells + AFS rats vs. 2nd ECs + AFS rats, i.e., 1st, 2nd and 3rd systems each and contribution comparison of 1st/3rd vs. 2nd. The differences are obtained by subtracting each step minused and taking the absolute value. 4 sets of AC(s) were used in both 2nd and 3rd step multifuntional ACs. HS: Saikosaponin A + hesperidin; HST: HS + tangeretin; FA: ferulic acid; MF: MH + FA; Hippo: hippocampal; MV: microvascular; ECs: endothelial cells; Ach: acetylcholine; AFS: acute forced swimming; BSS: Bupleurum chinense Shugan-San; MH: meranzin hydrate.

BPSDS eliminated ex vivo interferences and served as an interference-free herbal proxy

To further optimize herbal ex vivo proxies, we compared direct BSS addition, post-dose serum, and deproteinized BPSDS in the same multicells/ECs-AFS rat framework (Figure 4A–L). Direct addition of BSS caused aortic inactivation and increased EC apoptosis (Figure 4E and F), whereas serum addition still caused organ-bath foam (Figure 4G and H). In contrast, 100°C water-bath deproteinization for 5 min generated BPSDS that preserved multifunction while removing interference from crude preparation, as shown by the absence of aortic inactivation, EC apoptosis, and bath foam (Figure 4D, Figure 4I–K). Quantitatively, BPSDS reduced the interference-associated readouts of inactivation, foam, ROS, apoptosis, IL-6, and TNF-α by 36.64–47.45% (Figure 4L).

Display full size

Figure 4. Graphic comparison of ex vivo-to-in vivo extrapolation of ACs as multifunctional contributors to BPSDS/BSS in multicells/ECs-AFS rat. (A) Left BSS of directly addition into ex vivo system as the first popular ex vivo herbal agent method; (B, C) post-dose herb extract of serum as the proxy of herbal ex vivo agents of purported interference-free when added into ex vivo system; (D, I) 100°C water bath for 5 min to obtain BPSDS (BSS postdose supernatant of deproteinized serum) without interferences of aortic inactivation, EC apoptosis (J) and foam in organ bath (K), while adding BSS caused aortic inactivation (E), ↑ EC apoptosis (F) and adding serum caused ↑ foam (G, H); (L) quantitative interference eliminated by BPSDS was evaluated by change percentage of inactivation, foam, ROS, apoptosis, IL-6, TNF-α (36.64–47.45%). AFS: acute forced swimming; BSS: Bupleurum chinense Shugan-San; ECs: endothelial cells; IL-6: interleukin-6.

Three new normalization algorithms corrected the bias of orthodox model, drug, and phytochemical evaluation methods

We next compared three pairs of distinct new-old algorithms. Methodologically, mrni, drni, and prni served different purposes: mrni normalized model deviation across different endpoints, drni quantified drug-induced correction relative to the model background, and prni estimated phytochemical contribution after subtracting the S-V bias. Together, these indices reduced raw-value distortion and improved cross-endpoint comparability in the present study, namely mrni as new formula 1, drni as new formula 2, and prni as new formula 3, vs. the orthodox or popular calculations that separately handled S and V or directly compared ACs with herbs (Figure 5A–J). These new formulae were used to eliminate calculation-related biases.

Display full size

Figure 5. The difference comparison of three pairs of distinct new old algorithms. Model raw normalization increment (mrni) as new formula 1 vs. old or popular V; drugs responses of (Drug – V)/(V – S) as new formula 2 (drni) vs. old or popular [(Drug – V)/V], and phytochemical action contribution of [(ACs – V) – (V – S)]/[(BSS – V) – (V – S)] as new formula 3 (prni) vs. ACs/herbs. (A–J) S and V are not calculated by formula 1 and only S was used as subject of comparison and orthodox statistics following compared with V, without two-in-one as new value, without showing true relationship and resultant modeling value bias; 59.2% was FA vasodilator strength from old formula (Drug – V)/V, FA’s 93.4% from new formula 2, resultant their bias was 34.2% whose 28.75 mm (VTI, E+ CFMD or coronary flow) was absent as S. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001, ^#### p < 0.0001 vs. S; ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001 vs. V. V: Vehicle; S: Sham; BSS: Bupleurum chinense Shugan-San; FA: ferulic acid.

Ghr/Src-PI3K/Akt-eNOS signaling

As a representative vasoactive AC, FA induced dose/E+ vasodilation in aortic rings, which can be abolished by D-Lys, ODQ, LY294002, PP2 and L-NA, indicating positive involvement of Ghr (Figure 6A). In H-ECs, H₂O₂ increased ROS and reduced NO, both reversed by FA; the inhibitory effect of these FA-induced reversions was mainly caused by D-Lys (Figure 6B). Western blot showed that Ghr, p-Src, p-PI3K, p-Akt and p-eNOS were decreased in H-ECs vs. S and can be restored by FA, while these restored proteins were again abolished by D-Lys, PP2, LY294002, ODQ and L-NA (Figure 6C). Under oxidoreductase intervention, the FA-restored phosphorylated proteins ranked as PEG-catalase > PEG-SOD > MnTMPyP, showing NO-bioavailability activation’s anti-oxidative sensitivity (Figure 6D).

Display full size

Figure 6. The endothelium-dependent (E+) vasodilation induced by escalating dose FA in isolated aortic rings and H₂O₂-treated endothelial cells (H-ECs) via a parallel dual upstream ghrelin (Ghr)/Src-PI3K/Akt-eNOS, an anti-oxidative sensitivity, and subsequently activating the bioavailability of nitric oxide (NO) signaling pathway. This pathway was examined using six inhibitors [Ghr receptor’s D-Lys 25 μmol/L, Src kinase’s PP2 10 μmol/L, Akt’s wortmannin 0.1 μmol/L, PI3-kinase LY294002 30 μmol/L, endothelial NO synthase L-NAME 100 μmol/L, soluble guanylyl cyclase (ODQ) 1 μmol/L] and three artificial oxidoreductases, i.e., poly-ethyleneglycol-superoxide dismutase (SOD) (PEG-SOD) as the membrane permeant analogue of catalase (PEG-catalase) 500 μmol/L, PEG-SOD 100 units/mL and cell-permeable SOD mimetic (MnTMPyP) 100 μmol/L. (A) E+ vasodilation induced by escalating dose of FA in endothelium-intact isolated aortic rings and was abolished when pretreatment of D-Lys, ODQ, LY294002, PP2, and L-NA, respectively, identifying positive regulation of Ghr, Src, PI3K, Akt, and eNOS. (B) NO was detected by the NO kit. The generation of ROS in H-ECs was increased by 200 ng/mL FA. 0.5 mmol/L hydrogen peroxide-loaded ECs were stimulated with FA, and the ethidium bromide fluorescence was monitored over 10 min using a confocal microscope. H₂O₂ reduced levels of NO via impaired endothelium in ECs, and reversed by FA, while these reversions were markedly abolished when pretreatment of D-Lys, PP2, LY294002, ODQ, and L-NA for NO level and of MnTMPyP for ROS level in H-ECs. (C) Western blot analyses showed that the levels of Ghr, p-Src, p-PI3K, p-Akt, and p-eNOS were ↓ in H-ECs compared with S and reversed by FA. The latter ↑ levels of Ghr, p-Src, p-PI3K, p-Akt and p-eNOS were all markedly abolished by pretreatment of D-Lys, PP2, LY294002, ODQ and L-NA. D-Lys (Ghr) and PP2 (Src) without interaction, abolished middle/downstream inhibitors of roles and not abolished by them except for D-Lys with stronger mutual abolitions with L-NA, embodying a parallel dual upstream of Ghr/Src-PI3K/Akt-eNOS pathway as a regulation mechanism of FA-rising vasodilation. (D) Modeling, FA dose, or the changes in the levels of Ghr and phosphorylated pathway protein are the same or nearly the same as those in part (C). V’s p-Src, p-PI3K, p-Akt, and p-eNOS levels were lower than S, while FA’s these were higher than V. We compared FA-rose phosphorylated pathway protein levels in the presence and absence of three artificial oxidoreductases, such as ROS, etc., evidencing damaging E+ vasodilator pharmacology, resultant ranking when present were PEG-catalase (↑ p-Src/PI3K/Akt/eNOS levels) > PEG-SOD (↑ p-Src/PI3K/eNOS levels) > MnTMPyP (↑ p-PI3K level), as direct evidence of anti-oxidative sensitivity of activating bioavailability of NO pathway. ^& p < 0.05, ^&& p < 0.01, ^&&& p < 0.001, ^&&&& p < 0.0001 vs. S; ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001 vs. H₂O₂. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001, ^#### p < 0.0001 vs. FA. ECs: endothelial cells; eNOS: endothelial nitric oxide; S: Sham; FA: ferulic acid.

D-Lys identified Ghr as the shared pluripotent abolisher of multifunctional AC anchoring

Among several inhibitors tested against MH- or different sets of ACs-induced multifunction pharmacodynamics, the Ghr antagonist D-Lys showed the broadest and strongest abolishing effect in vivo across four BCGE-like functions (Figure 7A). Accordingly, D-Lys abolished 17.17–39.09% of the BCGE-4 functions induced by the strongest 3 sets of 2–3 AC(s), with rankings relatively similar to ex vivo results. In H-EC-based mimic ex vivo systems, D-Lys also abolished cross BCGE-like, ROS, and NO readouts, ranking top 1 in six units, including 3 NO and 2 ROS indices (Figure 7B and C). After in vivo pretreatment, D-Lys mainly abolished MH-contributed BGE 3-functions (Figure 7D).

Display full size

Figure 7. Multifunction comparisons between several doses of ontological disease pathogenesis or function inhibitors using pretreatment of abolishing prokinetic-based pluripotent MH or 4 sets of multi-ACs induced multifunction pharmacodynamic in vivo and mimic ex vivo. (A) Ghrelin (Ghr) inhibitor D-Lys: abolishing 4 sets of 2–3 ACs-induced C (inhibitors D-Lys, L-NA) BGE (plus WAY100635, NBQX and SB207266) multifunctions in vivo: among above ontological in vivo bioactivity of 5 pairs inhibitors and 4 functions (BCGE-like ↑ immobility, ↑ coronary disorder, ↓ G-emptying, ↓ enteric transit), D-Lys is pluripotent (4 rather than 1 functions were effective) and the strongest (three were top 1, one as top 2 inserted the middle of 2 depressive inhibitors) while other 4 inhibitors (2 shared depression inhibitors, one vasoactive, one hypomotility) only half of depressive inhibitor was top 1. (B) The inhibition rate to in vivo BCGE multifunction-induced by the strongest 3 sets of 2–3 ACs when pretreatment of Ghr antagonist D-Lys, which abolished ACs-contributed 17.17–39.09% BCGE-4 function as top 1 (brain-coronary-gut of three disorder of: one item D-Lys top 1 and two top 2 > other 4 inhibitors ranking), relatively similar to their ex vivo conclusion. (C) D-Lys, when pretreated, abolished 4 sets of 2–3 ACs-induced mimic ex vivo index strength of multifunction of BCGE (inhibitors of: C = D-Lys, L-NA and PP2; E = C + SB207266; parallel B and G = E+ WAY100635 and NBQX, above sum × 2 = ROS and NO of parallel 38 indices in H-ECs), as BCGE-like (depressive 2 inhibitors, one prokinetic inhibitor with GE-hypomotility and vasodilator L-NAME), antioxidant and pluripotent Ghr inhibitor of 6 function ex vivo model. And thereinto D-Lys was a shared protagonist (its 6 units, including 3 NO and 2 ROS indices as top 1). (D) The pretreatment of D-Lys mainly abolished MH-contributed in vivo BGE 3-functions whose another increase in E+ CFMD was invalid, which was replaced by effective ferulic acid or HS in Table 2. ^$ p < 0.05, ^$$ p < 0.01, ^$$$ p < 0.001, ^$$$$ p < 0.0001 vs. MH or HS or HST or HS; ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001 vs. V-1 h or V-24 h. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001, ^#### p < 0.0001 vs. Sham (S). CFMD: coronary flow-mediated dilation; GE: gastric emptying; MH: meranzin hydrate.

Time-dependent dendritic psychopathology after AFS and its reversal by Shugan-like therapy

AFS induced time-dependent dendritic psychopathology in dentate gyrus (DG) neurons, as shown by representative apical dendritic segments (Figure 8A) and quantitative analysis (Figure 8B). Total dendritic spine density was reduced at 12 and 24 h post-AFS, whereas total dendritic length mainly decreased at 12 h; in contrast, the number of intersections remained unchanged (Figure 8B). After intervention, BSS, MH, and FLX all improved the AFS-induced reductions in dendritic spine density and dendritic length (Figure 8C), together with restoration of GluA1, postsynaptic density protein 95 (PSD95), and BDNF in the DG (Figure 8D). HE staining of the heart and transmission electron microscopy of hippocampal synapses further showed that Shugan-like therapy alleviated structural injury in both myocardium and hippocampal synaptic ultrastructure after AFS (Figure 8E).

Display full size

Figure 8. Changes of total dendritic spine density and dendritic length after time-dependent acute forced swimming (AFS). (A) Representative examples of apical dendritic segments of DG pyramidal neurons selected for quantitative analysis of dendritic spines from the different experimental groups of animals. Scale bar represents 50 μm. (B) Total dendritic spine density. (C) Effect of BSS (30 g/kg), MH (6.81 mg/kg), and Fluo (10 mg/kg) on changes induced by AFS on total dendritic spine density, dendritic length and expression of PSD95, GluA1 and BDNF protein in DG. Representative examples of apical dendritic segments of DG pyramidal neurons selected for quantitative analysis of dendritic spines from the different experimental groups of animals. Scale bar represents 50 μm. Total dendritic spine density. Total dendritic length. (D) Western blot: expression of phospho-AMPA receptor 1 (GluA1), PSD95 and BDNF protein. The expression of GluA1, PSD95 and BDNF was quantified using Image Pro Plus software. (E) S group and effect of BSS (30 g/kg), MH (6.81 mg/kg) and Fluo (10 mg/kg) on changes induced by AFS. HE-staining of heart, Bar = 100 μm. Representative transmission electron microscopy images revealing the ultrastructure of synapses in the hippocampus (12,000×, scale: 1,000 nm). ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001 vs. BSS. ^# p < 0.05, ^## p < 0.01, ^### p < 0.001, ^#### p < 0.0001 vs. S. One-way ANOVAs followed by Q tests were used. Data are expressed as the mean ± S.E.M of the number of dendritic spines/10 μm of dendritic segment (n = 4–5 animals, 15–25 dendritic segments, 600–1,400 μm total length). BDNF: brain-derived neurotrophic factor; BSS: Bupleurum chinense Shugan-San; DG: dentate gyrus; MH: meranzin hydrate; PSD95: postsynaptic density protein 95; S: Sham.