Rakkyo (Allium chinense)-Derived Fructan Stimulates Collagen and Hyaluronan Synthesis in Human Dermal Fibroblasts

TL;DR

A fructan from rakkyo (Allium chinense) boosts collagen and hyaluronan production in human skin cells, suggesting potential for skin health.

Contribution

Demonstrates that rakkyo-derived fructan directly stimulates extracellular matrix synthesis in dermal fibroblasts.

Findings

Rakkyo fructan increased viable cell numbers and upregulated COL1A1 and HAS2 gene expression.

Treatment led to 3.8-fold and 1.3-fold increases in type I collagen and hyaluronic acid production.

Lysyl oxidase mRNA was upregulated, indicating enhanced collagen cross-linking.

Abstract

Background: Fructans are fructose-based polysaccharides with diverse biological activities; however, their direct activity on skin cells remains unresolved. This study investigated the biological activity of fructan extracted from rakkyo (Allium chinense) (RF) and examined its effects on extracellular matrix (ECM) metabolism, particularly collagen and hyaluronan synthesis, in human dermal fibroblasts. Methods: RF was prepared from fresh rakkyo bulbs by aqueous extraction, alkaline clarification, and membrane filtration. The average molecular weight and structural characteristics of RF were analyzed using size-exclusion chromatography and 13C NMR spectroscopy. Normal human dermal fibroblasts (NHDFs) were treated with RF by culturing cells in RF-supplemented medium (0.1–1.0 mg/mL). Cell viability and viable cell number were evaluated using the thiazolyl blue tetrazolium bromide and trypan blue exclusion assays, respectively. Expression of ECM-related genes was analyzed by qRT-PCR, and collagen and hyaluronan production were quantified by Sirius Red staining and ELISA. Results: RF had an average molecular weight of approximately 11,500 Da and consisted of nearly equal proportions of inulin- and levan-type fructans. RF (≤1 mg/mL) increased the number of viable cells and markedly upregulated collagen, type I, alpha 1 (COL1A1) and hyaluronic acid synthase 2 (HAS2) expression while downregulating Hyal1 expression. After 9 days of treatment, the cumulative production of type I collagen and hyaluronic acid increased by 3.8- and 1.3-fold, respectively, as compared with controls. Upregulation of lysyl oxidase (LOX) mRNA suggested enhanced collagen cross-linking, whereas MMP-1 showed only modest induction. Conclusions: Rakkyo-derived fructan directly stimulates collagen and hyaluronan synthesis in dermal fibroblasts, likely through regulation of ECM-related genes. These results suggest that rakkyo-derived fructan modulates ECM-related readouts in NHDFs under controlled in vitro conditions. Further validation in more complex skin models and in vivo studies is necessary.

Figures

Full text

1. Introduction

Dermal fibroblasts play a central role in maintaining skin structure through the synthesis and remodeling of extracellular matrix components, particularly collagen and hyaluronic acid. These matrix components are critical determinants of dermal integrity, hydration, and tissue homeostasis and are therefore widely used as functional readouts of fibroblast activity. Accordingly, evaluating extracellular matrix-related responses provides a biologically relevant framework for assessing the potential effects of bioactive compounds on dermal fibroblast function. Although extracellular matrix components are widely used as functional readouts of dermal fibroblast activity, the effects of plant-derived fructans on collagen and hyaluronic acid-related responses in these cells have not been systematically examined. Accordingly, extracellular matrix-related outcomes were selected as primary endpoints in the present study.

The structure and function of skin are influenced by intrinsic and extrinsic factors, including aging, ultraviolet (UV) radiation, and air pollution [1]. Among these, UV exposure is the primary cause of skin aging (photoaging), characterized by a reduction and degeneration of extracellular matrix (ECM) components, such as collagen (COL) and elastin, leading to wrinkles and sagging [2]. Hyaluronic acid (HA) is a linear polysaccharide composed of repeating disaccharide units of D-glucuronic acid and N-acetyl-D-glucosamine. HA occupies the interstitial space between collagen and elastin fibers in the dermis, playing a crucial role in maintaining skin hydration and turgor. COL is a fibrous protein that constitutes the extracellular matrix of most connective tissues in mammals [3,4], with type I, II, and III accounting for 80–90% of total body COL [5]. As skin turnover slows with age, the barrier function weakens and moisture retention decreases, leading to dry skin. Thus, promoting ECM synthesis while suppressing its degradation is critical for preserving healthy, youthful skin. Recent studies have shown that topical or oral administration of COL and HA can improve skin elasticity and hydration [6,7,8,9]. Given the central role of extracellular matrix components in maintaining skin structure and function, increasing attention has been directed toward bioactive compounds capable of modulating ECM synthesis and metabolism.

Polysaccharides are ubiquitous biopolymers with diverse physiological activities. In particular, mucopolysaccharides such as HA and chondroitin sulfate are important for maintaining skin hydration and elasticity, and are widely used in cosmetic and medical applications [6,7,8,9]. Other natural polysaccharides, including β-glucans derived from mushrooms and yeast, have demonstrated immunomodulatory and skin barrier-enhancing effects [10], whereas glucosamine and fucoidan derived from seaweed have been investigated for their anti-aging properties [11].

Among these natural polysaccharides, fructans represent a distinct class of fructose-based polymers whose biological activities have been extensively studied in systemic contexts but remain poorly characterized in skin cells. Fructans, a group of fructose-based polysaccharides, are primarily classified into two main structural types: inulin-type and levan-type [12]. Inulin-type fructans consist of a linear polymer formed by β-(2⟶1)-linked fructose units, and occur abundantly in plants such as chicory and Jerusalem artichoke. Levan-type polysaccharides are formed by β-(2⟶6)-linked fructose chains, often exhibit a branched structure, and are typically produced by bacteria such as Bacillus subtilis. Fructans exhibit a variety of physiological functions, with their prebiotic activity being the most extensively studied [13]. Fructans are fructose-based polysaccharides that are broadly classified into inulin-type and levan-type structures. Fructans are widely recognized for their prebiotic properties; however, their direct effects on dermal cells remain poorly characterized. Because orally ingested fructans are largely metabolized by the gut microbiota.

Allium chinense is commonly known as rakkyo in Japan and is widely consumed as a pickled vegetable. During processing, large quantities of byproducts, such as undersized bulbs and brine residues, are generated and often discarded, despite containing potentially valuable components, including sugars and polysaccharides. Rakkyo is presumed to contain fructans, primarily of the inulin type. In the present study, RF was defined as fructan extracted and purified from fresh rakkyo. We investigated the direct action of rakkyo-derived fructan (RF) on human dermal fibroblasts as a key cell type regulating extracellular matrix homeostasis. This approach also provides a rationale for the potential topical use of RF as a naturally derived cosmetic ingredient.

Recent evidence has highlighted the importance of the gut-skin axis, whereby the composition of gut microbiota influences skin health. Dietary fructans may alleviate symptoms of atopic dermatitis and improve skin barrier integrity [14]. However, despite these systemic benefits, the specific mechanisms by which fructans directly affect skin cells remain unresolved, and further research is required in this area. It should be noted that fructose is a naturally occurring monosaccharide present in fruits and honey and serves as an energy source in human nutrition. Importantly, adverse health outcomes are most consistently reported in the context of excessive intake of free fructose or fructose-containing added sugars, rather than typical dietary exposure. Therefore, in discussing skin-related effects, it is critical to distinguish free fructose from polymerized fructans, which are non-digestible dietary fibers. In contrast, fructose, the monomeric unit of fructans, has been associated with adverse effects on skin aging. For example, high concentrations of fructose induce senescence markers (p16, p21, p53) in human dermal fibroblasts through the formation of advanced glycation end-products, inhibition of cell proliferation, and delay of wound closure in an artificial wound model [15]. Chronic fructose consumption in rats has also been linked to excessive COL cross-linking in skin and bone, suggesting reduced skin elasticity and accelerated aging [16]. These findings suggest that while free fructose may impair dermal function, polymerized fructans may exert distinct, potentially beneficial effects.

Rakkyo is known to be rich in fructans, which have been extensively studied for their systemic and gastrointestinal effects. However, to our knowledge, no previous studies have examined the direct effects of rakkyo-derived fructans on dermal fibroblasts or extracellular matrix metabolism, and no preliminary studies addressing this specific cellular context are available prior to this work. This knowledge gap provided the rationale for selecting rakkyo-derived fructan as the focus of the present study. By investigating extracellular matrix-related responses, which represent a core functional output of dermal fibroblasts, the present study provides the first in vitro evaluation of the direct cellular effects of rakkyo-derived fructan in a skin-relevant context. Specifically, we examined the influence of RF on ECM metabolism, focusing on the biosynthesis and metabolism of HA and COL, and the expression of related gene clusters. The average molecular weight and structural characteristics of RF were also determined. Through these investigations, we aimed to clarify the direct action of fructans on skin cells and facilitate their potential as naturally derived ingredients for skin care applications.

2. Materials and Methods

2.1. Materials

RF were provided by Minorikanpo Co., Ltd. (Kitamorokata, Miyazaki, Japan). Normal human dermal fibroblasts (NHDFs) were purchased from Kurabo (Osaka, Japan). Thiazolyl blue tetrazolium bromide (MTT) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Primers were purchased from Invitrogen (Carlsbad, CA, USA). RNAiso Plus, PrimeScript™ RT Reagent Kit and TB Green™ Premix ExTaq were purchased from Takara Bio Inc. (Kusatsu, Shiga, Japan). Dulbecco’s Modified Eagle Medium (DMEM) and Qubit^®^ RNA Broad Range Assay Kits were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Fetal bovine serum (FBS) was purchased from Biowest (Nuaillé, France). Human Collagen Type I ELISA Kit was purchased from ACEL Inc. (Sagamihara, Kanagawa, Japan). Quantikine Hyaluronan ELISA Kit was purchased from R&D Systems, Inc. (Minneapolis, MN, USA). Levan was purchased from Megazyme (Bray, Ireland). All other chemicals and solvents were analytical grade and purchased from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan).

2.2. Preparation of RF

Rakkyo (Allium chinense) harvested in 2021 in Miyazaki Prefecture, Japan, was used as the raw material. A total of 1700 g of rakkyo material was divided into three batches. After the edible portion was separated, the remaining material was homogenized with twice its weight of distilled water for at least 10 min, and the liquid fraction was recovered by centrifugation and filtration. During the initial extraction process, approximately 260 g of insoluble residue was generated. The recovered extract (approximately 280 g; estimated based on the recovery process) was adjusted to pH 10 by adding a 10% calcium hydroxide solution (prepared from guaranteed reagent grade calcium hydroxide; FUJIFILM Wako Pure Chemical Corporation) and stirred for 60 min, after which precipitates were removed by centrifugation and filtration. The clarified fraction was transferred into cellulose dialysis tubing (nominal molecular weight cut-off: 10,000 Da) and purified at 4 °C through five cycles of 8 h dialysis. Dialysis against distilled water with complete dialysate replacement in each cycle was used to remove low-molecular-weight components, including inorganic salts/ions originating from the clarification step. Thus, this repeated dialysis process serves as a practical deionization/neutralization step prior to freeze-drying. Finally, the dialyzed fraction was freeze-dried to yield 170 g of purified fructan (RF), corresponding to an approximate yield of 10% based on the starting rakkyo material (Figure 1).

RF samples were dissolved in distilled water and transferred to HPLC vials. Size-exclusion chromatography was performed using an HPLC system (Alliance 2695; Waters, Milford, MA, USA), equipped with a UV detector (Waters 2487; Waters) set at 190 nm. Separation was achieved on a YMC-Pack Diol-300 column (5 µm, 300 × 8.0 mm; YMC Co., Ltd., Kyoto, Japan) using a mobile phase of 100 mmol/L phosphate buffer and acetonitrile (8:2 (v/v)), pH 6.8, at a flow rate of 0.5 mL/min and a column temperature of 40 °C. The molecular weight (MW) distribution of RF was estimated by the external standard method using reference compounds fructose (MW = 180 Da) and dextrans with MW of 5000, 12,000, and 25,000 Da.

The sample was dissolved in deuterium oxide (D_2_O) for nuclear magnetic resonance (NMR) measurements, and one drop of acetone-d6 was added as an internal standard. ^13^C NMR spectra were recorded at 100 MHz on an AVANCE NEO 400 (Bruker Corporation, Billerica, MA, USA) spectrometer at room temperature. NMR chemical shifts were recorded in ppm with the solvent signal acting as the internal reference (δC 206.7 ppm for acetone-d6).

2.3. Cell Culture and RF Treatment

NHDFs were cultured in DMEM supplemented with 10% FBS. The culture medium was replaced every 24–48 h, and cells were passaged at 80% confluence by standard trypsinization. Cells between passages 2 and 10 were used for all experiments. RF was dissolved in sterile distilled water to prepare a stock solution and was subsequently diluted with culture medium to the indicated final concentrations (0–10 mg/mL). RF-containing medium was freshly prepared before each experiment and added directly to NHDFs. Cells receiving medium only served as the untreated control group. In selected experiments, vitamin C (L-ascorbic acid phosphate magnesium salt hydrate) was added at the time of medium replacement as a reference compound. Vitamin C was prepared in sterile distilled water and diluted with culture medium to a final concentration of 100 µM. Whereas cells receiving medium only served as the untreated control group. Vitamin C (ascorbic acid) was included as a reference (positive-control) compound in selected experiments to benchmark ECM-related responses (e.g., collagen-related outcomes) in dermal fibroblasts [17,18].

2.4. Cell Viability and Cell Number Assays

NHDFs treated with RF as indicated were incubated with MTT for 3 h at 37 °C and lysed in a 0.04-mol/L HCl/isopropyl alcohol solution. The absorbance at 570 nm was measured using a spectrophotometer (SpectraMax^®^ iD5, Molecular Devices, Sunnyvale, CA, USA). The results are expressed as a percentage of the control (control = 100%).

Cell viability (%) was calculated as Viability (%) = [(ABS_sample − ABS_blank)/(ABS_control − ABS_blank)] × 100, where ABS_blank represents wells containing medium and MTT without cells. The MTT assay was used to define a non-cytotoxic working concentration range of RF. Because discrete concentrations were tested to determine working doses rather than to construct a full dose–response curve, an IC_50_ value was not calculated.

The trypan blue exclusion assay was performed according to the manufacturer’s protocol. NHDFs were pretreated with trypsin and then suspended and stained with an equal volume of trypan blue dye. The cells were counted using a hemocytometer and a light microscope ECLIPSE Ts2 (Nikon, Tokyo, Japan). The number of viable cells is expressed as cells per well.

2.5. Quantification of Gene Expression

Total RNA was extracted from NHDFs using RNAiso Plus according to the manufacturer’s instructions. Complementary DNA (cDNA) were transcribed from mRNAs using the PrimeScript™ RT Reagent Kit with oligo(dT) and random primers. The cDNA was used as templates for qRT-PCR in a LightCycler 96 (Roche Diagnostics, Tokyo, Japan) with TB Green™ Premix ExTaq, using the following amplification settings: 40 cycles at 95 °C for 5 s and 60 °C for 30 s. Gene expression levels were normalized to GAPDH expression as the internal control, and relative expression was calculated using the 2^−∆∆CT^ method (control = 1.0). Primer sequences are listed in Table 1.

2.6. Quantification of HA and COL

The amounts of hyaluronan (HA) and type I collagen in the culture medium were quantified using commercially available ELISA kits according to the manufacturers’ instructions: a Collagen Type I ELISA Kit (Cat. #EC1-E205; ACEL Inc., Sagamihara, Kanagawa, Japan) and a Quantikine Hyaluronan ELISA Kit (Cat. #DHYAL0; R&D Systems, Inc., Minneapolis, MN, USA). Samples were assayed neat or after appropriate dilution to fall within the linear range of the standard curve. After RF treatment, culture supernatants were collected and centrifuged at 12,000× g for 10 min to remove cellular debris. The clarified supernatants were stored at −80 °C until analysis. Standard curves were generated using serial dilutions of the provided standards, and sample concentrations were calculated from the calibration curves according to the manufacturers’ protocols. The absorbance at 450 nm was measured using a spectrophotometer (SpectraMax^®^ iD5, Molecular Devices, Sunnyvale, CA, USA). The concentrations of HA and type I collagen in the culture supernatants were quantified by ELISA and are expressed as ng/well.

Sirius Red staining was performed using a modified version of a previously described method [19]. NHDFs were washed with PBS(−) and fixed in Bouin’s solution for 1 h. After rinsing under running water for 15 min, the samples were stained with a 1% (v/v) Sirius Red solution for 1 h. The staining solution was removed, and the cells were washed five times with 0.01 mol/L HCl to remove unbound dye. After air-drying, the bound dye was eluted with 0.1 mol/L NaOH. The resulting solution was centrifuged, and the absorbance (540 nm) of the supernatant was measured using the SpectraMax^®^ iD5 spectrophotometer. The results of Sirius Red staining are expressed as a percentage of the control (control = 100%) based on absorbance measurements.

2.7. Statistical Analysis

All experiments were performed with at least three independent biological replicates. Data are expressed as mean ± standard deviation (S.D.). Relative values are presented as fold changes or percentage of control, as indicated in the figure legends. Statistical comparisons among three or more groups were performed using either Dunnett’s or Tukey’s post hoc test, as appropriate, with JMP^®^ Pro software (version 17.2.0; SAS Institute Inc., Cary, NC, USA).

3. Results

3.1. Procedure for Extracting RF from Rakkyo

The MW distribution of RF was measured by size-exclusion chromatography. The average MW of RF was estimated to be approximately 11,500 Da (Figure 2).

3.2. Effect of RF on NHDFs Viability and Viable Cell Number

Cell viability and cytotoxicity were assessed using the MTT assay after 24 h of RF treatment whereas the viable cell number was evaluated separately by counting viable cells at multiple time points using the trypan blue exclusion assay. RF treatment resulted in a concentration-dependent decrease in cell viability after 24 h. At RF concentrations of 2.5 mg/mL or higher, cell viability decreased to 65.3% or less of the control, representing a statistically significant reduction compared with untreated cells (Figure 3a). Based on this dose-dependent cytotoxicity profile, 1 mg/mL was selected as the maximum RF concentration for subsequent experiments. Accordingly, all subsequent assays were conducted at ≤1 mg/mL, i.e., within the non-cytotoxic working range defined by the 24 h MTT assay. Moreover, consistent with the absence of overt cytotoxicity at 1 mg/mL, viable cell numbers increased over time at this concentration in the trypan blue exclusion assay (Figure 3b). After nine days of RF treatment, the number of viable cells in the 1 mg/mL RF group increased 1.4-fold relative to the control (Figure 3b).

3.3. Effects of RF on Hyaluronic Acid HA Synthesis

The effects of RF on HA synthesis and metabolism were examined. After 9 days of RF treatment, HAS2 mRNA expression was significantly increased by 1.4-fold and 1.9-fold in the 0.1 and 1 mg/mL RF groups, respectively, whereas HAS1 mRNA expression showed no significant change. In contrast, Hyal1 mRNA expression was significantly reduced to 0.6-fold and 0.4-fold of control levels in the 0.1 and 1 mg/mL RF-treated groups, respectively, while Hyal2 expression remained unchanged (Figure 4a).

Consistent with these gene expression changes, the cumulative HA content in the culture supernatant was significantly increased by approximately 1.3-fold in the 1 mg/mL RF group by day 9 (Figure 4b).

3.4. Effects of RF on Collagen COL Synthesis

Next, the effects of RF on COL synthesis and metabolism were evaluated. COL1A1 mRNA expression was significantly increased in a dose- and time-dependent manner following RF treatment. On day 3, COL1A1 expression increased by 1.5-fold and 2.1-fold in the 0.1 and 1 mg/mL RF groups, respectively. The 1 mg/mL RF group continued to exhibit elevated expression levels over time (2.2-fold on day 6 and 2.8-fold on day 9), whereas the 0.1 mg/mL RF group reached a 2.1-fold increase by day 9.

LOX mRNA expression was also significantly upregulated, showing 1.3-fold and 1.7-fold increases in the 0.1 and 1 mg/mL RF groups, respectively, on day 3, and a 1.3-fold increase in the 1 mg/mL RF group on day 6. In contrast, MMP-1 mRNA expression was only mildly induced, with a 1.3-fold increase observed in the 1 mg/mL RF group on day 9 (Figure 5a).

Consistent with these transcriptional changes, the cumulative levels of type I collagen in the culture medium were significantly increased in the 1 mg/mL RF group, reaching 3.1-fold and 3.8-fold of control levels on days 6 and 9, respectively (Figure 5b). In addition, cell-associated (insoluble) collagen content was also elevated, increasing to 116% and 134% of control levels in the 0.1 and 1 mg/mL RF groups, respectively, by day 9 (Figure 5c).

Sirius Red staining provides a semi-quantitative assessment of total collagen deposition whereas ELISA specifically quantifies soluble collagen released into the culture supernatant. These methodological differences should be considered when interpreting the results obtained using each assay.

3.5. Effect of RF Chemical Structure on COL Biosynthesis

Further analyses were performed to compare the effects of RF (1 mg/mL) with those of pure levan and a levan–inulin mixture. As shown in Figure 6a, cell-associated collagen levels were significantly higher in the RF-treated group than in either the levan or levan + inulin group on day 3, indicating a stronger effect of RF on collagen accumulation under these conditions.

To characterize the structural features of RF, ^13^C NMR spectroscopy was performed. The spectra revealed characteristic signals corresponding to levan-type fructans (80.7 ppm) and inulin-type fructans (81.5 ppm). The comparable peak intensities of these signals indicated that RF contains levan- and inulin-type fructans in approximately equal proportions. NMR analysis confirmed that RF possesses a branched fructan structure, providing structural characterization of the purified compound (Figure 6b).

4. Discussion

In this report, we investigated the effects of fructan extracted from rakkyo on the synthesis and metabolism of HA and COL, key components of the extracellular matrix (ECM), in NHDFs. Ultraviolet radiation, aging, and oxidative stress reduce the production and promote the degradation of ECM components, such as HA and COL, by fibroblasts. This imbalance leads to typical skin-aging phenomena, including wrinkles, sagging, and dryness. As HA and COL are essential for maintaining skin elasticity, hydration, and structural integrity, preserving ECM homeostasis is vital for preventing or delaying skin aging. The novelty of this study lies in demonstrating a direct, cell-based effect of a purified rakkyo-derived fructan on dermal fibroblasts, showing coordinated modulation of both HA- and collagen-related pathways in vitro.

In the cell viability assay (MTT assay), RF exhibited concentration-dependent cytotoxicity, with no significant toxicity observed at treatments below 1 mg/mL. Because RF at concentrations of ≥2.5 mg/mL reduced cell viability, concentrations above this threshold were considered cytotoxic and were not used for subsequent functional assays. Therefore, 1.0 mg/mL was chosen as the maximum non-cytotoxic concentration in the experiments. At this concentration, fibroblast numbers increased by day 9 (Figure 3), suggesting that RF may increase viable fibroblast numbers at suitable doses. However, further studies are required to determine whether RF functions directly as a growth factor or whether the observed effect reflects improved cell viability and metabolic activity within a specific concentration range.

Because cytotoxicity emerged at higher RF concentrations (≥2.5 mg/mL), we restricted all functional analyses to ≤1 mg/mL. The basis of the high-dose toxicity was not determined in the present study and may involve nonspecific stress at high polysaccharide concentrations (e.g., osmotic/viscosity-related effects) and/or residual impurities. In the present study, we did not directly quantify residual inorganic ions (e.g., Ca^2+^; ICP-MS) or endotoxin levels (e.g., LAL assay). Therefore, future work should include rigorous quality assessments such as elemental analysis for residual inorganic ions and endotoxin testing to exclude impurity-driven responses.

HA is synthesized by three transmembrane enzymes, hyaluronic acid synthases 1, 2, and 3 (HAS1, HAS2, and HAS3) [20,21], which exhibit distinct expression patterns and can produce HA with different molecular weight distributions [22,23]. Among them, HAS1 exhibits lower enzymatic activity than HAS2 or HAS3 [24,25]. Although HAS3 is another isoform involved in hyaluronan synthesis, it was not examined in the present study; HAS2 is the predominant isoform responsible for hyaluronan production in dermal fibroblasts whereas HAS3 expression is reported to be relatively low in these cells [26]. In addition, because we did not assess the molecular-weight distribution of HA, potential HAS3-related differences in HA size and associated biological outcomes should be addressed in future studies. In tissue, HA is degraded extracellularly by hyaluronidase (Hyal), reactive oxygen species, and intracellularly by lysosomal enzymes [27]. Six hyaluronidases have been identified in humans [28]. On the cell surface, Hyal2 binds to HA, mediates its internalization, and the resulting fragments are further degraded to tetrasaccharides by Hyal1 [29]. In this study, RF treatment increased HAS2 mRNA expression by day 9, whereas downregulation of the mRNA level of the degradative enzyme Hyal1 was observed (Figure 4a). These results suggest that RF simultaneously upregulates HA synthesis and suppresses degradation, leading to enhanced HA accumulation within the ECM. Consistent with these molecular changes, the cumulative HA level in the culture supernatant increased in the 1 mg/mL RF group (Figure 4b). These extracellular measurements represent cumulative outcomes over time and may not directly mirror messenger ribonucleic acid levels at a single time point. These findings indicate that RF promotes HA retention through the bidirectional, coordinated regulation of its synthesis and degradation pathways. Although the magnitude of the increase in hyaluronic acid content was modest, the observed 1.3-fold elevation was statistically significant. Given that HA represents a tightly regulated final extracellular matrix product in dermal fibroblasts. Even moderate increases in HA levels may therefore have functional relevance in the context of dermal matrix homeostasis

Type I COL is primarily a heterotrimer composed of two α1 chains and one α2 chain, and is characterized by its triple-helical structure and repeating sequence [Gly-X-Y]n motif, where proline and hydroxyproline typically occupy the X and Y positions, respectively [5]. Lysyl oxidase (LOX) catalyzes the oxidative deamination of lysine residues in collagen, resulting in cross-link formation and increased tissue strength. In contrast, metalloproteinases play a key role in the degradation of COL in the skin. For example, UV-damaged collagen molecules are cleaved into smaller fragments by matrix metalloproteinase-1 (MMP-1) [2,30]. RF treatment upregulated COL1A1 mRNA expression from an early stage (day 3), and LOX mRNA expression was also slightly upregulated by day 9 (Figure 5a). A mild induction of MMP-1 mRNA was also observed. These findings suggest that RF may influence collagen maturation, potentially by modulating the expression of COL1A1 and LOX. It should be noted that in the present study, we evaluated LOX expression only at the mRNA level. Enzymatic activity of LOX and the formation of mature collagen cross-links, such as pyridinoline, were not directly assessed. Therefore, although the upregulation of LOX mRNA suggests potential enhancement of collagen maturation, further studies are required to confirm its functional relevance at the protein and activity levels. The mild induction of MMP-1 observed in this study may suggest a potential involvement in extracellular matrix remodeling; however, functional assays will be required to determine whether this response reflects controlled matrix turnover. Similar observations have been reported by Erginer et al. [31], who found that levan derived from Halomonas and its sulfonated derivative markedly increased COL1A1 expression, supporting the notion that certain fructans can enhance collagen synthesis.

It should be noted that changes in messenger ribonucleic acid expression and extracellular protein accumulation are not necessarily synchronous. Transcriptional responses can occur rapidly, whereas measurable increases in secreted proteins in the culture supernatant and collagen deposition reflect downstream processes (translation, secretion, processing, and accumulation) over time. Therefore, the enzyme-linked immunosorbent assay and Sirius Red staining results should be interpreted as cumulative extracellular readouts rather than as direct, immediate correlates of messenger ribonucleic acid levels at a single time point.

Consistent with these transcriptional effects, type I COL levels in the culture medium increased (Figure 5b), and the amount of cell-associated collagen rose to a maximum by day 9 (Figure 5c). At 1 mg/mL, the magnitude of collagen accumulation induced by RF was comparable to, and in some measures exceeded, that induced by the positive control, vitamin C. Vitamin C is widely used as a reference compound in dermal fibroblast studies because it has been shown to stimulate collagen synthesis in human skin fibroblasts and enhance extracellular matrix production in vitro. For example, vitamin C treatment increased both collagen mRNA expression and total collagen output in cultured fibroblasts [17,18]. These results indicate that RF promotes both COL secretion and matrix deposition, in agreement with the observed increase in COL1A1 and LOX expression, and is thought to support RF-mediated ECM homeostasis. The difference in the rate of increase between ELISA and Sirius Red staining likely reflects differences in specificity, as ELISA quantifies type I COL selectively using antibodies, whereas Sirius Red binds both type I and type III COL [32,33]. Taken together, the combined use of Sirius Red staining and ELISA provides complementary readouts of deposited and soluble collagen in this in vitro system. These results demonstrate that RF enhances ECM composition not only at the gene expression level but also through functional increases in protein production, secretion, and deposition.

In comparative analyses, RF induced greater increases in cell-associated COL content than the levan group or inulin + levan group (Figure 6a). Structural characterization revealed that RF contained both inulin- and levan-type fructans (Figure 6b), suggesting that the specific biological activity of RF may arise from its unique branched fructan structure. Although the structure of rakkyo-derived fructans has not been reported previously, fructans from related species such as onion (Allium cepa) and garlic (Allium sativum) have been shown to possess inulin-type β(2→1) linkages [34,35]. Our data suggest that rakkyo fructan contains both inulin- and levan-type fructans in almost equal proportions, forming a unique branched structure that may underlie its enhanced biological activity. Structure–activity relationships have been reported for fructans, where branching, degree of polymerization, and molecular-weight distribution can affect physicochemical properties (e.g., hydration/viscosity) and thereby influence interactions with the cell surface and the pericellular matrix. Such architectural differences (e.g., β(2→1)-linked inulin-type vs. β(2→6)-linked levan-type, and their mixtures) may contribute to differences in biological potency among plant-derived fructans. In the present study, our nuclear magnetic resonance (NMR) spectroscopy and size-exclusion chromatography (SEC) profiles support that RF is a mixture of inulin- and levan-type fructans with an apparent average molecular weight in the ~10 kDa range; however, we did not determine a rigorous molecular-weight distribution nor linkage frequencies. Therefore, while RF enhanced ECM-related outputs in NHDFs in our system, detailed structure–activity relationships should be addressed in future work using advanced analyses such as size-exclusion chromatography coupled with multi-angle light scattering (SEC–MALS), high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC–PAD), and linkage analysis coupled with mass spectrometry (MS). However, in the present study, we did not directly test structure–function relationships; therefore, the contribution of specific structural features to these cellular responses remains speculative and warrants future investigation.

The ability of RF to stimulate HA and COL synthesis highlights its potential as a bioactive material for improving ECM quality and skin health. HA contributes to skin hydration, maintaining elasticity, and the cushioning properties of the ECM, whereas COL provides structural support and tensile strength to the ECM. Therefore, compounds that can upregulate both components simultaneously are promising candidates for anti-aging and skin regeneration applications. Similar effects have been reported for other polysaccharides, such as marine-derived sulfated polysaccharides and plant extracts. For example, extracts derived from Ulva species have been reported to promote the production of COL and HA in fibroblasts [36]. These findings provide conceptual support for the RF effect in this study. Additionally, a negative feedback mechanism has been reported in which COL fragments inhibit HA synthesis during aging [37]. Thus, agents like RF that enhance ECM synthesis and counteract such feedback inhibition may help restore ECM homeostasis.

Based on the above description, RF has potential applications as an ECM regulatory factor capable of simultaneously controlling HA and COL levels within fibroblasts, thereby extending its scope beyond conventional reports. Although the precise molecular mechanisms by which RF induces HAS2, COL1A1, LOX, and other genes remain unresolved, available evidence suggests the possible involvement of the transforming growth factor-β (TGF-β) signaling pathway. TGF-β is a central regulator of ECM homeostasis and has been widely reported to induce the production of collagen, fibronectin, proteoglycans, hyaluronic acid, and other components [38]. Reports indicate that TGF-β increases HAS gene expression and promotes HA production [38]. Moreover, HA can further activate fibroblasts via CD44-mediated positive feedback, amplifying TGF-β1 signaling [39]. Therefore, it is conceivable that RF may stimulate or activate TGF-β or its receptor-mediated Smad2/3 pathway, leading to upregulation of HAS2 and COL1A1 expression.

The observed modest increase in MMP-1 mRNA expression can be interpreted as part of ECM “remodeling”, facilitating the turnover of older COL during matrix renewal. Controlled MMP activity is essential for balanced ECM dynamics, with excessive suppression likely promoting sclerotic and pathological fibrosis, whereas moderate induction supports physiological remodeling. From the perspective of COL fragment-mediated feedback [37], moderate MMP activation by RF may play a role in promoting ECM remodeling without triggering inhibitory feedback loops.

This study has several limitations. First, HAS3 is another isoform involved in hyaluronan synthesis, but it was not examined in this study; HAS2 is the predominant isoform responsible for hyaluronan production in dermal fibroblasts whereas HAS3 expression is reported to be relatively low in these cells. Second, although potential signaling pathways such as TGF-β are discussed in the context of the existing literature, these pathways were not experimentally examined here. Further studies will be required to clarify the involvement of additional HAS isoforms and upstream signaling mechanisms in RF-mediated regulation of extracellular matrix metabolism. Third, this study demonstrates that rakkyo-derived fructan (RF) modulates extracellular matrix–related parameters in normal human dermal fibroblasts under controlled in vitro conditions. However, because the present work was limited to a single cell type in vitro, the relevance of these findings to skin aging, regeneration, or cosmetic applications remains to be established in more complex skin models and in vivo/clinical studies. Finally, this study is that we did not evaluate bioavailability, skin penetration, or formulation. RF is a high-molecular-weight polysaccharide with an average molecular weight of approximately 11.5 kDa, which is expected to exhibit limited permeability across the stratum corneum. Therefore, although our results demonstrate ECM-related responses in dermal fibroblasts in vitro, translating these findings to topical cosmetic applications will require further research. Such research should include molecular weight optimization (e.g., fractionation or controlled depolymerization), formulation stability assessment, and validation in more complex skin models (e.g., reconstructed human skin), which can better capture barrier properties.

5. Conclusions

In conclusion, rakkyo-derived fructan (RF) enhanced extracellular matrix (ECM)-related outputs in normal human dermal fibroblasts (NHDFs) under controlled in vitro conditions. RF increased the expression of ECM-associated genes (including COL1A1 and HAS2) and modulated hyaluronan metabolism-related genes (e.g., Hyal1). Consistent with these transcriptional changes, RF increased hyaluronan and type I collagen levels in the culture supernatant (ELISA) and promoted collagen deposition in NHDFs (Sirius Red assay).

These findings are based on an in vitro fibroblast model, and the physiological relevance remains to be established. Future studies should evaluate RF in vivo (animal models and/or clinical studies) and further investigate molecular mechanisms (e.g., receptor involvement and downstream signaling). In addition, more rigorous structural characterization (including molecular-weight distribution and linkage features) will be important to clarify structure–activity relationships and to optimize potential skin-relevant applications.