Materials

A hydro-alcoholic glycerine solution of R. nigrum buds (1.5%) was kindly provided by Biokyma S.r.l. (Anghiari, Arezzo, Italy). The RNeasy® Mini Kit was obtained from Qiagen S.p.A. The iScript cDNA synthesis kit was purchased from Bio-rad s.r.l. Fluocycle® II SYBR® was purchased from Euroclone s.p.a. (Milan, Italy). TNF-α was purchased from Sigma Aldrich (Milan, Italy). High-performance liquid chromatography (HPLC)-grade water (18 mΩ) was prepared by a Mill-Ω⁵⁰ purification system (Millipore Corp., Bedford, MA, USA). All the reagents and materials were obtained from commercial sources with a high grade of purity.

Isolation and culture of human GMSCs

GMSCs were obtained after processing de-keratinized gingival tissues previously collected from four healthy female patients (average age 35.5 years) undergoing clinical crown lengthening procedures. The protocol received approval from the ethical committee of the University Hospital of Pisa (Pisa, Italy; protocol no. 32835/2016) and informed consent was obtained from the included patients.

The tissues were processed as previously reported with a few modifications [32]. Briefly, after surgical removal, discharged gingival specimens were de-epithelialized and placed in sterile phosphate-buffered saline (PBS) with 100 U/mL penicillin and 100 μg/mL streptomycin (Sigma-Aldrich, Milan, Italy) at 4 °C. The tissues were minced into 1–2 mm² fragments and digested in Dulbecco’s modified Eagle’s medium (DMEM)-F12 containing dispase (1 mg/mL; Sigma-Aldrich) and collagenase IV (2 mg/mL; Sigma-Aldrich) at 37 °C for 30 min. Then, the suspension was discarded, and the remained tissues were digested in the same solution for 90 min at 37 °C. The solution was then filtered with a 70-μm cell strainer (Sigma-Aldrich) and seeded with DMEM-F12 containing 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL streptomycin, and 200 mM l-glutamine in a 25-cm² tissue culture flask. At 24 h after isolation, the non-adherent cells were washed with PBS and replaced with fresh medium (passage 0).

Cell cultures

The isolated GMSCs were maintained in growth medium (DMEM-F12 containing 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 200 mM l-glutamine) and incubated at 37 °C in 5% CO₂ and 95% air. The medium was changed to remove non-adherent cells every 3 to 4 days and the cells were used at passages 0 to 6.

Human dermal fibroblasts cells (HuDe; purchased from the Istituto Zooprofilattico Sperimentale, Brescia, Italy) were maintained in DMEM-F12 containing 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 200 mM l-glutamine, and incubated at 37 °C in 5% CO₂ and 95% air. The medium was changed to remove non-adherent cells every 3 to 4 days and the cells were used at passages 0 to 20.

The immortalized human microvascular endothelial cell line (HMEC-1) was obtained from ATCC and maintained in a culture medium of MCDB131 (without l-glutamine), 10 ng/mL epidermal growth factor (EGF), 10 μg/mL hydrocortisone, 10 mM l-glutamine, and 10% FBS. Cells were maintained in a 37 °C humidified incubator with 5% CO_2. The medium was changed to remove non-adherent cells every 3 to 4 days and the cells were used at passages 10 to 18.

Colony-forming unit-fibroblast (CFU-F) assay and doubling time

For CFU-F assay, the GMSCs were plated into a six-well plate at densities of 500 or 1000 cell/well and maintained in growth medium. After 14 days, the cells were fixed with 4% paraformaldehyde (Sigma-Aldrich) and stained with 1% crystal violet. Groups of ≥ 50 cells were scored as a CFU-F colony. The numbers of colonies were statistically evaluated.

The assessment of the GMSC doubling time was performed as previously described [33]. Cells of different passages were seeded onto 12-well plates at 10³ cells/well in growth media and the cells in each well were scored at 24, 48, 72, and 96 h. Population doubling was calculated using: number of divisions = log2 (number of cells at subculture/number of cells seeded).

PCR and real-time RT-PCR analysis

The gene expression of GMSCs and HMECs were quantified by performing a real-time reverse transcription (RT)-PCR analysis. Briefly, GMSCs (5.0 × 10³ cell/cm²) or HMECs (1.0 × 10⁴ cell/cm²) were treated with glycerine-ethanol solution or RBE (50 μg/mL) in the absence or the presence of TNF-α (1 ng/mL or 10 ng/mL) for the indicated times. The cells were collected, and the total RNA was extracted using the Rneasy® Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The purity of the RNA samples was determined by measuring the absorbance at 260/280 nm. cDNA synthesis was then performed with 500 ng RNA using the i-Script cDNA synthesis kit following the manufacturer’s instructions. Real-time RT-PCR reactions consisted of 25 μL Fluocycle® II SYBR®, 1.5 μL of both 10 μM forward and reverse primers, 3 μL cDNA, and 19 μL H₂O. All reactions were performed for 40 cycles using the following temperature profiles: 98 °C for 30 s; 55 °C for 30 s; and 72 °C for 3 s. The primers used were designed to span intron/exon boundaries and β-actin was used as the housekeeping gene. The mRNA levels for each sample were normalized against β-actin mRNA levels, and the relative expression was calculated using a Ct value. PCR specificity was determined by both melting curve analysis and gel electrophoresis.

Immunofluorescence analysis of stem cell markers

GMSCs were seeded in an eight-chamber slide at a density of 5 × 10³ cells/chamber and incubated overnight. The next day, the cells were fixed with 2% paraformaldehyde (Sigma-Aldrich), and then permeabilized in 1% Triton- X100 with PBS for 3 min. GMSCs were stained with anti-CD90 (Millipore, cat. no. CBL415) and Alexa-Fluor-488-conjugated goat anti-mouse. Non-specific binding was controlled by omitting the primary antibody or by substituting the same concentration of non-specific isotype immunoglobulin (data not shown). Nuclei were stained with DAPI (1:5000 dilution; Sigma-Aldrich). The specimens were mounted in VECTASHIELD_ Mounting Medium with DAPI (Vector laboratories) and then observed under a fluorescence microscope (Axiovert 200; Zeiss).

Western blot analysis

The protein expression of the surface marker CD90 was detected in GMSCs by performing a Western blot analysis. Cells were lysed for 60 min at 4 °C by the addition of 200 μL RIPA buffer. Equal amounts of the cell extracts (35 μg of proteins) were diluted in a Laemmli solution, resolved by SDS-PAGE (15%), and then transferred to PVDF membranes and probed overnight at 4 °C with primary antibody anti-CD90 (diluted 1:100; Millipore, cat no. CBL415) or β-actin (diluted 1:1000; MAB1501, Merck KGaA, Darmstadt, Germany). The primary antibodies were then detected using anti-rabbit IgG light chains conjugated to peroxidase (diluted 1:5000; 12–348; Millipore). The peroxidase was detected using a chemiluminescent substrate (ECL, Perkin Elmer), and the images were acquired by photographic film or by LAS4010 (GE Health Care Europe, Uppsala, Sweden). Immunoreactive bands were quantified by performing a densitometric analysis with Image J Software (version 1.41; Bethesda, MD, USA).

Mineralization assay

GMSCs were seeded (3 × 10³ cells/cm²) in growth media. After 24 h the media was replaced with osteogenic-induction medium (Euroclone) containing dexamethasone (10⁻⁸ M), l-glutamine, ascorbate (50 μg/mL), penicillin/streptomycin (2 mM), and β-glycerophosphate (2 mM). As controls, GMSCs were cultured in growth medium. The medium was changed every 3 days and the mineralization was quantified after 15 or 21 days of treatment. The rate of mineralization was quantified using Alizarin Red staining as previously reported [34] with a few modifications. Briefly, cells were washed with PBS, fixed (4% formaldehyde in PBS) for 15 min, and then cells were stained with Alizarin Red S (1:100 in distilled water, adjusted to pH 4.2, and filtered), washed (five times) in 50% ethanol, and air dried. For quantification, cells were destained overnight in 10% (w/v) cetylpyridinium chloride at room temperature with continuous agitation and the absorbance (562 nm) was read using a spectrophotometer (Victor Wallac 2, Perkin Elmer).

Cell viability assays

GMSCs and HMEC-1 were seeded in 96-well microplates (3.0 × 10³ cells/well) and treated with different concentrations of RBE (100 ng/mL to 100 μg/mL) or TNF-α (1–100 ng/mL) alone or in combination for 24, 48, or 72 h. Following the treatment period, cell proliferation was determined using an MTS assay (CellTiter 96 AQueous One Solution Cell Proliferation Assay kit; Promega) according to the manufacturer’s instructions. The absorbance of formazan at 490 nm was measured in a colorimetric assay with an automated plate reader (Victor Wallac 2, Perkin Elmer).

The cytotoxicity was assessed by a neutral red assay. Briefly, cells treated as above were incubated for 1 h with 100 μL of neutral red dye (33 μg/mL) dissolved in serum-free medium. At the end of the incubation period, cells were washed twice with the warm PBS solution. The cell-incorporated dye was solubilized in 100 μL of neutral red assay solubilization solution (50% EtOH and 1% AcOH). The cultures were allowed to stand for 10 min at room temperature with gentle shaking. The absorbance was measured at a wavelength of 540 nm. The background absorbance of multiwell plates at 690 nm was subtracted from the measurement at 540 nm.

Analysis of apoptosis

For apoptosis measurements, GMSCs and HMEC-1 (5.0 × 10³ cell/cm²) were treated with or without TNF-α (100 ng/mL) for 72 h. The percentages of living, apoptotic, and dead cells were then quantified and analysed by the Muse™ Cell Analyzer (Merck KGaA, Darmstadt, Germany). The live, early apoptotic, and late apoptotic/dead cells were discriminated between using staining with Annexin V and 7-aminoactinomycin D (7-AAD).

The protein-protein interaction networks (PPIN)

Network analysis was performed using the STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) website (http://string-db.org/). The co-mentions, co-expression, and associations were set as the evidence for functional links with a medium confidence score of 0.4.

Cytokine release

GMSCs were treated with glycerine-ethanol solution (control) or RBE (50 μg/mL) in the absence or the presence of TNF-α (1 ng/mL or 10 ng/mL) for 24 h. The amounts of cytokines presented in the culture medium (interleukin (IL)-6, IL-10, and transforming growth factor (TGF)-β) or expressed on the cell membrane (cyclooxygenase (COX)-2) was measured using enzyme-linked immunosorbent assay (ELISA) kits (Thermo Fisher Scientific, Rodano, Milan, Italy, Enzo Lifescience) following the manufacturers’ instructions.

Wound healing analysis

GMSCs (5.0 × 10³ cell/cm²) were treated with glycerine-ethanol solution or RBE (50 μg/mL) in the absence or the presence of TNF-α (1 ng/mL or 10 ng/mL) for 24 h. Media from GMSCs were collected, and conditioned media in the proportion of 20% + 80% (20% GMSC conditioned media + 80% HMEC-1 culture media) were applied onto HMEC-1.

HMEC-1 were seeded in 96-well plates and grown to 90% confluence. A scratch was then made through the cell layer using a sterile micropipette tip. After washing with PBS, cells were treated with condition media from GMSCs treated as above. The images of the wounded area were captured immediately after the scratch (t₀) and 8 h later (t₈) to monitor cell regeneration into the wounded area. Photographs were then taken at 10× magnification on an inverted microscope. The wound-healing abilities were quantified by measuring both the average gap width and the percentage of gap closed. The data were analysed with Image J software.

HPLC-PDA/UV-ESI-MS/MS analysis

For the HPLC-photodiode array (PDA)/UV-electrospray ionization (ESI) - tandem mass spectrometry (MS/MS) (HPLC-PDA/UV-ESI-MS/MS) study, 500 μL of a hydro-alcoholic glycerine solution of R. nigrum buds was added to 1.5 mL of methanol and the mixture was first centrifuged, then filtered and injected into the LC-MS system. Qualitative HPLC-PDA/UV-ESI-MS/MS analysis was performed using a Surveyor LC pump, a Surveyor autosampler, coupled with a Surveyor PDA detector, and a LCQ Advantage ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA) equipped with Xcalibur 3.1 software. Analysis was performed using a 4.6 × 250 mm, 4-μm, Synergi POLAR-RP 80A column (Phenomenex, Castel Maggiore, Bologna, Italy). The eluent was a mixture of methanol (solvent A) and a 0.1% aqueous solution of formic acid (solvent B). The solvent gradient was as follows: 0–5 min, 5–50%; 5–25 min, 50% A isocratic mode; 25–85 min, 50–100% A. Elution was performed at a flow rate of 0.8 mL/min with a splitting system of 2:8 to MS detector (160 μL/min) and PDA detector (640 μL/min), respectively. The volume of the injected methanol solution was 20 μL. Analyses were performed with an ESI interface in the negative mode. The ionization parameters used were as follows: capillary temperature, 270 °C; capillary voltage, −16.0 V; tube lens offset, −5 V; sheath gas flow rate, 60.00 arbitrary units; auxiliary gas flow rate, 3.00 arbitrary units; spray voltage, 4.50 kV; scan range of m/z 150–1200 [35]. N₂ was used as the sheath and auxiliary gas. PDA data were recorded in the 200–600 nm range, with preferential channels 254, 280, and 325 nm as the detection wavelengths.

Statistical analysis

The GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA) was used for data analysis and graphical presentations. All data are presented as the mean ± SEM. Statistical analysis were performed by a one-way analysis of variance (ANOVA) with Bonferroni’s corrected t test for post-hoc pair-wise comparisons. P < 0.05 was considered as statistically significant.

Characterization of the isolated GMSCs

GMSCs were isolated from gingival tissues from four healthy female subjects (average age 35.5 years). The cells were maintained in culture for 14–21 days and characterized for their phenotypic and genotypic profile. The primary cultures of single-cell suspensions exhibited a fibroblast-like spindle cell shape and formed MSC-like colonies after 10–14 days of culture at a low density (Fig. 1a). The GMSCs were analysed for their ability to form CFU-Fs, and they produced a concentration-dependent increase in colony formation of seeded GMSCs (Fig. 1b), confirming that the gingival tissue-derived cells were clonogenic.

The isolated GMSCs exhibited long-term proliferation capacity exceeding 12 passages in culture, and the doubling-time at passage 4 was 35.7 ± 3.7 h in accordance with literature data [33, 36].

The presence of specific GMSC markers were evaluated by PCR analysis (Fig. 1c); the specific MSC-associated surface markers CD90 and CD105 were highly expressed. Conversely, the hematopoietic stem cell markers CD34 and CD45 were not expressed (Fig. 1c). The expression of the specific GMSC marker CD90 was verified by immunofluorescence (Fig. 1d) and Western blot analysis (Fig. 1e, f). The CD90 protein expression was significantly higher in isolated GMSCs with respect to the fibroblast cell line (HuDe) (25.4 ± 3.2% vs. GMSC, P ≤ 0.001; Fig. 1e, f). These results demonstrated that the isolated cells shared a similar expression pattern with that of MSCs derived from other tissues, such as from bone marrow and dental pulp [36, 37].

GMSCs are characterized by their ability to differentiate; thus, the osteogenic potential of the isolated cells was examined. In the presence of osteogenic factors, the gingival cells formed mineralized nodules or aggregates (Fig. 1g, h). Calcium mineralization was confirmed by Alizarin Red S staining that showed a significant increase in osteogenic differentiation with respect to cells cultured in growth medium (143.5% ± 11.0 vs. control, P ≤ 0.01; Fig. 1h). Since these cells showed all the characteristics of human MSCs, including phenotype, self-renewal ability, and differentiation potential, these cells were designated as the GMSC population.

TNF-α effects on GMSCs and HMEC-1 cells

The effects of a wide range of TNF-α concentrations (0.1 ng/L to 100 ng/mL) on GMSC proliferation and viability were evaluated after 48 or 72 h of cell treatment (Fig. 2a, b and Additional file 1: Figure S1). The results showed that TNF-α was able to slightly increase the GMSC proliferation at a low concentration (5–10 ng/ml) after 48 h. This effect was lost after 72 h of treatment. Conversely, a high TNF-α concentration (100 ng/ml) slightly decreased the GMSC proliferation after 48 h and become significant after 72 h of treatment (P ≤ 0.05). Moreover, a neutral red assay demonstrated that the cytokine significantly affects the numbers of living cells only after 72 h of cell treatment, with a higher effect when the 100 ng/mL concentration was used (Additional file 1: Figure S1). These data revealed that, under our experimental conditions, TNF-α presented paradoxical effects on GMSCs dependent on cell exposure time and cytokine concentration.

Neovascularization is necessary during all the phases of wound healing. This occurs through endothelial cell activation, proliferation, and migration [38, 39]. In this process, a pivotal role is played by the MSCs that orchestrate the endothelial function through the release of different cytokines and growth factors. In this respect, the effects of a wide range of TNF-α concentrations (0.1 ng/mL to 100 ng/mL) on HMEC-1 proliferation were evaluated after 48 or 72 h of cell treatment (Fig. 2c, d). The results showed that TNF-α (100 ng/mL) slightly decreased the HMEC-1 proliferation after 48 h, with this effect becoming significant after 72 h of treatment (Fig. 2d), in accord with literature data [40]. The inflammatory cytokine did not show any positive effects on endothelial cells, but a high TNF-α concentration for a long period of exposure could decrease the cell proliferation in accord with the effects exerted on GMSCs.

The mechanism underlying the cell proliferation decrease was then investigated by evaluating the TNF-α (100 ng/mL) apoptotic activity on GMSCs and HMEC-1 cells (Fig. 2e and Additional file 1: Figure S2). The 72 h of treatment with a high TNF-α concentration was able to induce a significant phosphatidylserine externalization in the absence of 7-AAD staining (P ≤ 0.05), thus denoting the signs of the early phase of apoptosis in both GMSCs and endothelial cells (Fig. 2e).

Modulation of TNF-α on GMSC cytokine and growth factor release

The GMSC secretome controls and modulates the activity of several other cell types. Classic growth factors and cytokines, such as TGF-β, IL-10, IL-6, and COX2, serve as paracrine control molecules secreted or packaged into extracellular vesicles, or exosomes, by GMSCs [41, 42]. A computational STRING analysis was performed to investigate the functional interaction between NF-kB, TNF-α, and the different cytokines released by the mesenchymal cells. Based on the criteria set, a network of protein-protein interactions (PPI) that linked together the NF-kB and the different cytokines/growth factors were obtained (Fig. 3a).

Hereafter, TNF-α was used at the two concentrations we previously demonstrated to exert opposite effects on GMSC proliferation (Fig. 2). Challenging GMSCs with TNF-α at 10 or 100 ng/mL caused a concentration-dependent increase in pro-inflammatory cytokine release (IL-6 and COX2; Fig. 3b, c) and a concentration-dependent decrease of the anti-inflammatory cytokine IL-10 levels (Fig. 3d). The growth factor TGF-β, which is a well-known immunosuppressive cytokine [43], presented a hermetic concentration-response course: TNF-α at 10 ng/ml concentration increased significantly the release of the factor, whereas, the higher cytokine concentration caused a slight decrease (69.5 ± 5.5, control; 97.6 ± 9.8, TNF-α 10 ng/mL; 57.1 ± 2.9, TNF-α 100 ng/mL; Fig. 3e) of the same factor.

TNF-α modulation of GMSC and endothelial cell interplay

The GMSC secretome inhibits inflammatory responses, promotes endothelial and fibroblast activities, and facilitates the proliferation and differentiation of different cells [44]. To investigate the effects of the GMSC secretome on the endothelial cell functions, HMEC-1 were treated with the conditioned medium (CM) derived by the GMSCs (Fig. 4a). In particular, GMSCs were treated for 24 h with a low (10 ng/ml) and high (100 ng/ml) concentration of TNF-α; the collected medium was applied to HMEC-1 and, then, the endothelial cell proliferation and motility were analysed.

The CM derived by the GMSCs without treatment (control CM) was able to promote endothelial cell proliferation after 24 h (Fig. 4b) and, more efficiently, after 48 h of cell treatment (120.6 ± 2.6% vs control; Fig. 4b). Furthermore, control CM was able to significantly increase the endothelial cell motility (Fig. 4d–f). These data demonstrate that the GMSCs, per se, were able to exert a trophic activity on the endothelial cells in accordance with the reported trophic properties of GMSCs [18].

Next, the effects of TNF-α on the GMSC-HMEC-1 interplay were evaluated. A high concentration of TNF-α (100 ng/mL) was able to modify the GMSC secretome, producing a significant negative effect on HMEC-1 proliferation (120.6 ± 2.6%, CM; 107.5 ± 4.4%, TNF-α; Fig. 4b, c) and motility (56.4 ± 0.8%, CM; 49.7 ± 2.0%, TNF-α; Fig. 4e, f). Conversely, no significant effects on HMEC-1 well-being or functionality were observed with a low dose of the cytokine, despite a positive trend in the promotion of endothelial motility being evident (Fig. 4). These results demonstrate that cytokines such as TNF-α could turn off the beneficial and trophic effects of GMSCs, negatively affecting their interaction with surrounding cells dependent on time and concentration.

Chemical composition of the bud preparation

The chemical composition of the RBE hydro-alcoholic glycerine solution was investigated by means of HPLC-PDA/UV-ESI-MS/MS techniques. The LC-PDA/UV chromatogram, acquired at 325 nm, is shown in Fig. 5. Compounds 1–14 (in Fig. 5a) were identified by comparison with their elution orders, ESI-MS/MS data, and PDA/UV absorbance (Table 2) with data reported in the literature. All identified molecules belong to the phenols class, including phenolic acid derivatives, such as chlorogenic acids (1, 3, and 4 in Fig. 5a), p-coumaroylquinic acids (2, 5, and 7 in Fig. 5a), and flavonol glycosides, such as myricetin 3-O-rutinoside (8 in Fig. 5a), myricetin 3-O-glucoside and/or myricetin 3-O-galactoside (9 in Fig. 5a), quercetin 3-O-rutinoside (10 in Fig. 5a), quercetin 3-O-glucoside and/or quercetin 3-O-galactoside (11 in Fig. 5a), kaempferol 3-O-rutinoside (12 in Fig. 5a), kaempferol 3-O-glucoside (13 in Fig. 5a), and isorhamnetin glucoside and/or isorhamnetin galactoside (14 in Fig. 5a).

Peak	Compound	tR (min)	M	[M − H]−	MS/MS base peak (m/z)	MS/MS ions (m/z)	λmax (nm)
	Phenolic acids
1	Caffeoylquinic acid	17.7	354	353	179	191, 173, 135	257, 325
2	3-p-Coumaroylquinic acid	25.2	338	337	163	293, 203, 191,173	313
3	Caffeoylquinic acid	29.6	354	353	179	191, 173, 155, 135	257, 329
4	Caffeoylquinic acid	32.3	354	353	191	179, 173, 135	257, 329
5	4-p-Coumaroylquinic acid	41.6	338	337	173	293, 191, 173, 163, 155	313
7	5-p-Coumaroylquinic acid	45.8	338	337	191	293, 207, 173, 163	314
	Flavonol glycosides
6	Kaempferol di-hexoside	44.4	610	609	447	573, 489, 285, 179	270, 356
8	Myricetin 3-O-rutinoside	55.1	626	625	317	607, 479, 271, 243, 203, 179	270, 357
9	Myricetin 3-O-glucoside and/or myricetin 3-O-galactoside	55.9	480	479	317	461, 271, 179, 151	269, 358
10	Quercetin 3-O-rutinoside (rutin)	61.1	610	609	301	591, 447, 271, 243, 179	265, 356
11	Quercetin 3-O-glucoside (isoquercitrin) and/or quercetin 3-O-galactoside (hyperoside)	62.3	464	463	301	445, 273, 179	257, 356
12	Kaempferol 3-O-rutinoside	66.3	594	593	285	447, 257, 229, 179	270, 350
13	Kaempferol 3-O-glucoside (astragalin)	67.1	448	447	285	257, 229, 179	268, 349
14	Isorhamnetin glucoside and/or isorhamnetin galactoside	67.8	478	477	315	285, 271, 243, 229, 179	271, 356

Compound numbers correspond with peak numbers in Fig. 5

Compounds 1, 3, and 4 in Fig. 5a are three isomers showing a similar mass spectra consisting of the same parent ion [M − H]⁻ at m/z 353 and different fragment ions at m/z 191, due to the loss of the caffeoyl moiety, 179, generated by the loss of the quinic unit, 173, corresponding to a product ion [quinic acid−H − H₂O]⁻, and 135, due to a [caffeic acid−H − CO₂]⁻ ion. These data are in agreement with those reported for these compounds by Clifford et al. [45]. The presence of 3-O-caffeoylquinic acid (neochlorogenic acid), 4-O-caffeoylquinic acid (cryptochlorogenic acid), and 5-O-caffeoylquinic acid (chlorogenic acid) in RBE buds was recently reported [46, 47]. Peaks 2, 5, and 7 in Fig. 5a (λ_max 313–314) displayed the same full mass spectra with deprotonated molecules [M − H]⁻ at m/z 337. MS/MS spectra have a similar fragmentation pathway, showing fragment ions at m/z 173, due to the loss of a p-coumaric acid molecule ([M − 164]⁻), and two other diagnostic ions at m/z 191 and 163, in accord with the fragmentation profile of p-coumaroylquinic acids [48]. These finding are in agreement with the study of Ieri et al. who reported the presence of p-coumaroylquinic acid isomers in commercial bud preparations of blackcurrant [47]. It is possible to discriminate between each of the three isomers on the basis of the base peak value generated in the MS/MS spectra, according to Clifford et al. [48]. Indeed, base peaks at m/z 163, 173, and 191 were generated by fragmentation of p-coumaroylquinic acid parent ions carrying the esterification at position 3, 4, and 5, respectively. Thus, compounds 2, 5, and 7 could be identified as 3-p-coumaroylquinic acid, 4-p-coumaroylquinic acid, and 5-p-coumaroylquinic acid, respectively.

Peaks 8–14 in Fig. 5a were assigned to flavonoid glycosides showing characteristic UV spectra, with two strong absorption peaks at 255–270 and 349–358 nm, typical of a flavonol structure. Aglycon portions were represented by myricetin, kaempferol, quercetin, and isorhamnetin as deduced by fragment ions in the MS/MS spectra at m/z 317, 285, 301, and 315, respectively. Parent ions ([M − H]⁻) of flavonoid monoglycosides at m/z 479 (9 in Fig. 5a), 463 (11 in Fig. 5a), 447 (13 in Fig. 5a), and 477 (14 in Fig. 5a) displayed product ions generated by the loss of one hexose moiety ([M − 162]⁻) due to the cleavage of O-sugar bond. MS/MS spectra of compounds 8 and 10 in Fig. 5a showed product ions generated by the loss of one rhamnose moiety ([M − 146]⁻) and one hexose moiety ([M − 162]⁻) attributable to the presence of the disaccharide rutinose linked to the myricetin and quercetin aglycons, respectively. Thus, compounds 8 and 10 were identified as myricetin 3-O-rutinoside and quercetin 3-O-rutinoside, respectively. The detection of all these bioactive constituents was in agreement with results previously reported [46, 47].

The MS/MS experiment for peak 6 in Fig. 5a ([M − H]⁻ at m/z 609) provided product ions at m/z 447 and 285, generated by the subsequent losses of two hexose moieties, not identifiable through spectral data; thus compound 6 was supposed to be a kaempferol di-hexoside by PDA/UV and MS/MS data [49].

R. nigrum anti-inflammatory effects on GMSCs and HMEC cells

The anti-inflammatory properties of RBE [31], and the effects of some of its constituents on inflammatory gene transcription such as NF-kB and TNF-α [50–52], have been demonstrated. Inflammation is a pivotal process that may positively or negatively affect regenerative processes and the different cell types involved [53]. Thus, the ability of the RBE to modulate NF-kB and TNF-α gene transcription was evaluated in both GMSCs and HMECs (Fig. 5b).

Challenging the GMSCs with the RBE (50 μg/mL) for 24 h produced a significant decrease in NF-kB (0.60 ± 0.09 fold with respect to control, P ≤ 0.001) gene expression, and only a slight decrease in TNF-α transcription. In HMEC-1 cells treated with RBE (50 μg/mL) for 48 h, the NF-kB expression was significantly decreased (0.54 ± 0.09 fold with respect to control, P ≤ 0.001) as well as the TNF-α expression (0.62 ± 0.10 fold with respect to control, P ≤ 0.001), which is one of the pro-inflammatory cytokines mainly affecting the endothelial cell fate [54]. Taken together, these results demonstrated that RBE has anti-inflammatory effects on GMSCs and HMEC-1 cells, negatively modulating pro-inflammatory pathways.

Effects of R. nigrum on GMSC and HMEC-1 cell proliferation and well-being

To investigate the putative effects of RBE on stem cell proliferation/viability, isolated GMSCs were treated with a wide range of the extract concentrations (100 ng/mL to 100 μg/mL) for 48 and 72 h. The results (Fig. 6a, b) showed that higher concentrations of RBE slightly affected the GMSC proliferation after 48 h of cell treatment and the effects become significant only after 72 h. The proliferative effects were dose-dependent, with a maximum increase of 123.3 ± 2.5% (50 μg/mL) of cell growth. Moreover, a neutral red assay demonstrated that the extract significantly affected the numbers of living cells only after 72 h of cell treatment, with the higher effect when the 50 μg/mL concentration was used (Additional file 1: Figure S1). These data show the ability of RBE to affect the proliferation of the GMSCs, increasing the cell number in a dose- and time-dependent manner. The induction of MSC proliferation and the maintenance of stemness features represents a pivotal mechanism for improving the regenerative properties of the MSCs [55].

The role of the transcription factors that regulate self-renewal and differentiation is well known in embryonic stem cells [56]. Among these, Oct4 (octamer-binding transcription factor 4) and SOX2 (sex determining region Y (SRY)-box 2) work together to support each other’s expression and that of other self-renewal genes repressing differentiation genes [57, 58]; thus, the effects of RBE on the expression of Oct4 and SOX2 stemness genes were investigated (Fig. 6c). In parallel, the expression of mammalian target of rapamycin (mTOR) that was involved in MSC differentiation was quantified.

GMSC treatment with RBE (50 μg/mL) did not significantly affect mTOR expression (Fig. 6c). Conversely, it was able to significantly increase the gene expression of the stemness markers Oct4 and SOX2 (1.8 ± 0.3 and 1.7 ± 0.1 fold change, respectively). Taken together, these results demonstrated the positive effect of the RBE on GMSC well-being, promoting stemness maintenance.

To investigate the effects of RBE alone on the proliferation/viability of endothelial cells (HMEC-1), cells were treated with different concentrations of RBE (100 ng/mL to 100 μg/mL) for 48 or 72 h. The results (Fig. 6d–f) showed that it affected the HMEC-1 proliferation at higher concentrations after 24 h of cell treatment in a concentration-dependent manner, with an EC₅₀ value of 34.5 μg/mL. This effect become more evident after 48 and 72 h, with an increase in both the EC₅₀ values (27.4 and 21.0 μg/mL, respectively) and the maximum effects (147.7 ± 7.6% and 162.0 ± 1.7%, respectively). These data show the ability of RBE to affect the proliferation of HMEC-1 cells. Interestingly, the endothelial cells showed a higher sensitivity to the extract exposure per se with respect to the GMSCs (Fig. 6a, b).

R. nigrum decrease the negative effects of TNF-α in GMSCs and HMEC-1 cells

To fully investigate the effects of TNF-α on GMSC trophic activity, the RBE was used as anti-inflammatory agent since it had been demonstrated to interfere with the inflammatory pathways in GMSCs and HMEC-1 cells, promoting their well-being. RBE cell treatment for 72 h in the presence of 100 ng/ml TNF-α completely counteracted the cytokine effects on GMSC proliferation, restoring the cell growth rate as demonstrated by MTS and neutral red assays (Fig. 7a, b and Additional file 1: Figure S2).

Next, the effects of RBE on TNF-α-induced blockade of HMEC-1 cell proliferation and viability were evaluated (Fig. 7c, d). The extract per se increased the HMEC-1 viability; interestingly, RBE was able to completely counteract the negative effects of 100 ng/mL TNF-α, restoring the cell growth rate. Taken together, these results suggest that RBE exerts a cyto-protective effect in an experimental model of inflammation counteracting the negative effects exerted by a high TNF-α concentration, both in human GMSCs and, more markedly, in endothelial cells.

Restoration of GMSC cytokine and growth factor release by R. nigrum

First, the putative effects of RBE alone on the GMSC secretome were investigated (Fig. 8). Isolated GMSCs were treated with RBE (50 μg/mL) for 24 h and quantitative analysis of the released cytokines was performed. The unstimulated GMSCs secreted low levels of anti-inflammatory molecules (TGF-β and IL-10), and RBE alone was not able to alter their expression pattern (Fig. 8). Thus, RBE per se was not able to alter the trophic function of GMSCs.

Then, GMSCs were treated with an inflammatory stimulus (TNF-α 100 ng/mL) in the presence of RBE. The presence of RBE was able to significantly reduce the release of IL-6 and the expression of COX2, with levels comparable with that obtained with a low TNF-α (10 ng/mL) stimulation. RBE significantly increased IL-10 release and expression. Finally, RBE was able to only partially increase TGF-β. Furthermore, in a radar plot, a shift of the cytokine production versus the basal condition was shown (Fig. 8i). These results highlight that RBE is able to modulate the inflammatory process. Furthermore, the extract did not completely block the cytokine release but only restored the levels necessary to mediate the positive effect of a controlled inflammation.

GMSC and endothelial cell interplay is modulated by R. nigrum extract

Considering that RBE promotes anti-inflammatory actions through the modulation of the GMSC secretome, the extract effects on GMSC-HMEC-1 cell interplay was evaluated by the CM method (Fig. 9). The extract alone did not modify the HMEC-1 cell proliferation (Fig. 9a, b) in accord with the absence of the GMSC secretome composition; interestingly, RBE was able to completely counteract the negative effects of TNF-α (100 ng/mL) on HMEC-1 cells, restoring the levels of cell growth rate triggered by GMSC control CM. GMSC-CM and TNF-α (100 ng/mL) produced comparable effects on HMEC-1 motility to that reported above (Fig. 4d–f). Furthermore, the treatment of GMSCs with the extract alone produced a CM that did not modify endothelial cell motility (Fig. 9c–e), in agreement with the lack of effects on GMSC cytokine release (Fig. 8). Surprisingly, in the presence of a highly inflammatory microenvironment, RBE was able to significantly counteract the negative effects of the high concentration of TNF-α. These results suggest that the use of natural agents, such as RBE, are able to counteract the effects of a high TNF-α concentration, restoring the GMSC secretome similarly to that obtained in controlled inflammation (a low TNF-α dose) which could effectively be used to modulate the GMSC response during the regenerative process. In particular, these results demonstrate a necessity for controlling the inflammatory priming of GMSCs during the regenerative process both in vivo and in vitro when GMSCs are used in engineering grafts.