Bioreactor-based stem cell therapy for liver fibrosis

Human UCMSCs were collected after obtaining written informed consent from women with uncomplicated pregnancies who underwent elective cesarean section at term in the obstetrics department of the first hospital of Lanzhou University. This study was performed in accordance with the declaration of Helsinki. All procedures were approved by the ethics committee. Fresh umbilical cords were soaked in sterile phosphate-buffered saline (PBS) (Biological Industries, Israel) and transferred to sterile operation tables. The arteries and veins in the tissue were removed and sheared into 1 × 1 mm tissue particles. Tissue particles were evenly spread on the bottom of a T75 cell culture bottle (Corning) and 15 ml Dulbecco's modified Eagle's medium/Ham's F12 nutrient medium (DMEM/F12) (GIBCO, USA) containing 10% fetal bovine serum (FBS) (GIBCO, USA) was added. When large numbers were at the bottom of the culture bottle, they were detached using 10% trypsin-digested medium (Biological Industries) and transferred to a new culture bottle. The culture medium was replaced with a serum-free medium (Biological Industries, Israel). UCMSCs were identified using cell surface markers (figure S1).

Rat hepatocytes BRL-3A were obtained from Procell Life & Technology (China) and cultured in DMEM supplemented with 10% FBS and 1% penicillin and streptomycin (GIBCO, USA). Cells were maintained at 37 °C and 5% CO₂ in a humidified incubator.

The MSC-serum-free medium (Biological Industries, Israel) was the basic medium used to evaluate UCMSC senescence. Rat liver fibrosis serum was mixed with MSC-specific serum-free medium (v:v) as intervention solution. A Senescence β-Galactosidase Staining Kit (Beyotime Biotechnology, China) was used to evaluate cell senescence, and the stained cells were senescence-positive cells. The operation was conducted in accordance with the manufacturer's instructions, including washing with PBS once, fixing at 25 °C for 15 min, and incubating with the dye solution at 37 °C overnight. Images were acquired using a microscope (CKX53; Olympus, Tokyo, Japan).

UCMSCs (4 × 10⁵ cells well⁻¹) were inoculated into 6-well plates with 2.5 ml of culture medium. Hepatocytes (6 × 10⁵ cells well⁻¹ or 1 × 10⁵ cells well⁻¹) were inoculated into 6-well transwell inserts (Corning Life Sciences, USA) containing 1.5 ml of culture medium. Liver fibrosis serum and MSC-specialized serum-free culture medium (Biological Industries, Israel) (5:95, v/v) were used in the co-culture system. After 12 h of cultivation, the conditioned medium was obtained after filtration with a 0.2 μm filter.

PLA discs (24 mm in diameter and 1 mm in thickness) suitable for a 6-well plate were manufactured. Gradient-density MSCs (0, 2 × 10⁴, 6 × 10⁴, 1 × 10⁵ cells well⁻¹) and hepatocyte BRL-3A (3 × 10⁵ cells well⁻¹) were implanted on the surface of the PLA discs. Culture medium (3 ml) was added to each well.

SEM was performed according to a general procedure. Briefly, samples fixed in electron microscope fixative buffer (Servicebio, China) were dehydrated with an ascending ethanol series (30%–100%), subjected to critical point drying (Quorum K850), and sputtered with a platinum–palladium Denton Vacuum (HITACHI MC1000). Field-emission SEM (HITACHI Regulus 8100; EHT = 15.00 kV, WD = 4 mm, detector in lens) was used for observation.

NH₄Cl powder (Sangon Biotech, China) was dissolved in DMEM/F12 basal medium to a concentration of 3 mmol l⁻¹ and filtered through a 0.2 μm filter. The culture medium was collected at 6, 12, and 24 h after the addition of the NH₄Cl medium to determine the remaining concentration. An ammonia assay kit (Nanjing Jiancheng Bioengineering Institute) was used for ammonia determination following the manufacturer's instructions. DMEM/F12 basic culture medium was used for the cell culture for the urea production and albumin secretion assays. We added 3 ml of the medium to a 6-well plate (Corning, Corning, NY, USA) and 40 ml of the medium to the bioreactor. After 12 h of culture, urea and albumin concentrations in the culture medium were measured. Urea and albumin assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) were used for urea and albumin determination, respectively, according to the manufacturer's instructions.

Total RNA was extracted from cultured cells using TRIzol reagent (TAKARA, Japan). One microgram of total RNA was reverse-transcribed using a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, USA). Real-time qPCR was conducted in a 20 µl final volume using 1 ng µl⁻¹ cDNA and forward and reverse primers (500 nM each) with TB Green Premix Ex Taq II (TAKARA, Japan) in a CFX Connect PCR System (Bio-Rad, USA). The primer sequences are shown in table S1.

Male Sprague-Dawley rats weighing approximately 130 ± 10 g were purchased from Charles River Laboratories (Beijing, China), carefully maintained under standard laboratory conditions, acclimated for 7 d before the experiments, and allowed free access to food and water. Hepatic fibrosis was induced in rats with CCl₄ (1 ml kg⁻¹, CCl₄: olive oil = 1:1 v/v) via intraperitoneal injection twice per week for 8 weeks. At the end of the modeling, the rats were quickly anesthetized and blood samples were collected by abdominal aortic puncture and centrifuged for 10 min at 1000 × g to acquire serum. Before preparing the induction medium, all liver fibrosis serum was mixed and stored at −80 °C.

A commercial 3D printer (Ultimaker S5, Utrecht, Netherlands) was used for the FDM. Specifically, commercial PLA filaments with a diameter of 2.85 ± 0.10 mm (Ultimaker, Utrecht, Netherlands) were used in this study. The glass transition temperature of the PLA filaments was 60 °C, and the melting temperature was approximately 145 °C–160 °C. The detailed mechanical and thermal properties of the commercial PLA are listed in tables S2 and S3. The digital bioreactor model used for 3D printing was constructed using the Autodesk Inventor 3D CAD software (Autodesk, San Rafael, CA, USA). The models were exported into the.stl format, imported into slicing software CURA (Ultimaker, Utrecht, Netherlands), and configured for the FDM printer with a 0.4 mm brass nozzle, a layer height of 0.1 mm, and an infill of 70%. The hot-end temperature was set to 210 °C, and the printing bed temperature was set to 60 °C. The printing speed was 60 mm s⁻¹, which enabled retraction and print cooling. Support structures were not required for the printing process. After FDM, the printed bioreactor was carefully cleared for cell culture.

The bioreactor sizes are shown in figure 5. The internal space of the bioreactor included an upper storage tank, a lower storage tank, and a vertical liquid flow channel. The total storage volume was 43.5 ml. The upper and lower storage tanks were undivided spaces, with a storage volume of 21.3 ml per tank. The liquid flow channel in the vertical section was divided into 12 channels of equal width by partitions, each with a length of 5.4 cm, a width of 0.5 cm and a cross-sectional area of 0.15 cm². The cells adhered to one side of the fluid flow channel with a total adhesion area of 32.4 cm². During the operation of the bioreactor, the culture medium was first injected into the lower storage tank and then into the upper storage tank through 12 regular liquid flow channels in the vertical section. The peristaltic pump provided power and silicone tubing provided a pathway for the circulation of the culture medium.

Satisfactory disinfection was achieved via UV irradiation and circulation of a 75% ethanol solution for 1 h. The sterilized bioreactor system was rinsed five times with sterile PBS (Biological Industries, Israel). During cell implantation, a vertical section of the bioreactor was laid flat to ensure cell sedimentation and adhesion to the sides of the flow channel. MSCs and hepatocytes were implanted into the bioreactor at cell densities of 0.15 × 10⁵ cm⁻² and 0.75 × 10⁵ cm⁻², respectively, with a culture medium volume of 10 ml. The bioreactor was then left at 37 °C and 5% CO₂ in a humidified incubator for 12 h. After cell adhesion, the culture medium was discarded and the bioreactor was rinsed three times with sterile PBS at a low flow rate (5 ml min⁻¹) for 30 s each. After emptying, 40 ml of diluted serum was injected into the bioreactor, and a specific fluid shear force was set based on the shear force calculation formula. The diluted serum was comprised of MSC-specialized serum-free culture medium (Biological Industries, Israel) mixed with rat serum (v/v). The shear rate at the cell adhesion surface was calculated with the equation σ = 6 μQ/(WH²), where Q is the flow rate, μ is the viscosity coefficient of the diluted serum (μ = 0.86 × 10^–2 dynes·s cm⁻²), W is the channel width, and H is the channel height. After 12 h of cyclic cultivation, the culture medium was obtained, and the therapeutic medium was obtained after filtration with a 0.2 μm filter. Throughout the process, the bioreactor was left at 37 °C and 5% CO₂ in a humidified incubator.

To freeze-dry the medium, they were removed from −80 °C storage and loaded into a vacuum freeze-dryer (Shanghai Jingxin Industrial Development, China). Samples were lyophilized in batches for 20 h following the manufacturer's protocol. Lyophilized samples were stored in a refrigerator at 4 °C and sealed with parafilm to protect them from moisture absorption. Immediately before use, aliquots were rehydrated to 20% of their original volume with sterile distilled water (Invitrogen, USA) and then briefly vortexed.

We mixed 500 μl of extraction solution containing internal standard (methanol acetonitrile in a volume ratio of 1:1, internal standard concentration of 20 mg l⁻¹) with 100 μl of liquid sample, then vortexed the mixture and sonicated it in an ice-water mixture for 10 min. The samples were incubated at −20 °C for 1 h and centrifuged at 12 000 × g for 15 min at 4 °C. Then, 500 μl of the supernatant was dried in a vacuum concentrator, and 160 μl of the extraction solution (acetonitrile:water in a volume ratio of 1:1) was added to the dried metabolites for redissolution. After vortexing for 30 s, the samples were sonicated in an ice water bath for 10 min. The solution was centrifuged at 10 000 × g and 4 °C for 15 min, which was then prepared for machine testing.

The LC-MS system used for metabolomic analysis consisted of a Xevo G2-XS QTOF high-resolution mass spectrometer (Waters, USA). The column used was a UPLC HSS T3 column (1.8 µm; 2.1 × 100 mm; Waters Acquity). In the positive ion mode, mobile phase A was a 0.1% formic acid aqueous solution, and mobile phase B was 0.1% formic acid in acetonitrile. In the negative ion mode, mobile phase A was a 0.1% formic acid aqueous solution, and the mobile phase was 0.1% formic acid/acetonitrile. In each data acquisition cycle, dual-channel data acquisition was simultaneously performed at both low and high collision energies. The low collision energy was 2 V, the high collision energy range was 10 ∼ 40 V, and the scanning frequency was 0.2 s for a mass spectrum. The parameters of the ESI ion source were as follows: capillary voltage: 2000 V (positive ion mode) or −1500 V (negative ion mode); cone voltage: 30 V; ion source temperature: 150 °C; desolvent gas temperature: 500 °C; backflush gas flow rate: 50 l h⁻¹; desolvent gas flow rate: 800 l h⁻¹. The raw data collected using MassLynx V4.2 were processed using Progenesis QI software for peak extraction, peak alignment, and other data processing operations based on the Progenesis QI software online METLIN database and Biomark's self-built library for identification. Theoretical fragment identification and mass deviation were within 100 ppm.

Proteins in the therapeutic medium were digested using sequencing-grade trypsin. Label-free mass spectrometry was performed using an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific, USA). The scan events were set as a full MS scan of m/z 250–1450 at a mass resolution of 120 000, followed by a CID MS/MS scan repeated for the 20 most intense ions selected from the previous full MS scan, with an isolation window. The normalized collision energy was set to 30% and the activation time was 50 ms. The second-stage linear ion trap fast mode was used for data acquisition with a 7000 automatic gain control, a maximum injection time of 35 ms, and a dynamic exclusion time of 18 s. The UniProt _HUMAN (2019.4.20 Download) database was used. The resulting MS/MS data were processed using MaxQuant version 1.5.2.8. Protein identification was set as follows: precursor ion mass tolerance: ±15 ppm; fragment ion mass tolerance: ±0.5 Da; max missed cleavages: 2; static modification: carboxyamidomethylation (57.021 Da) of Cys residues; dynamic modification: oxidation modification (+15.995 Da) of Met residues.

The liver fibrosis modeling was obtained as described in section 2.8. After establishment of the model, the concentrated treatment medium (1 ml kg⁻¹) was injected twice a week through the abdominal cavity or tail vein. The control group was injected with PBS using the same method. The health of rats was monitored regularly. All the animals received humane care in accordance with the criteria outlined in the 'Guide for the Care and Use of Laboratory Animals.' All procedures were approved by the ethics committee.

At the end of the treatment period, the livers were fixed in 4% paraformaldehyde (Solarbio, China) and embedded in paraffin for hematoxylin and eosin (HE), Sirius Red, Masson, immunohistochemical, and immunofluorescence staining. Antibodies against CD68 (28058-1-AP, Proteintech, USA), MPO (22225-1-AP, Proteintech, USA), CD3 (ab16669, abcam, USA), NK-P44 (GB11616, Servicebio, China), S100β (15146-1-AP, Proteintech, USA), α-SMA (GB111364, Servicebio, China), desmin (16520-1-AP, Proteintech, USA), γh2Ax (GB111841, Servicebio, China), Ki-67 (27309-1-AP, Proteintech, USA), CK-19 (ab52625, abcam, USA), and GGT (GB113421, Solarbio, China), as well as the HE stain kit (Solarbio, China), TUNEL apoptosis assay kit (Beyotime, China), and Masson's trichrome stain kit (Solarbio, China) were used according to the manufacturers' instructions. Sections were observed under a microscope (CKX53; Olympus, Tokyo, Japan). The positive cell count was based on five random fields of view and performed by two experienced pathologists. The positive area measurement was based on five random fields of view, and area recognition was performed using the ImageJ software.

Data were analyzed using SPSS Statistics 25.0 (IBM, USA) and Prism 6.0 software (GraphPad Software, USA). Data are presented as mean ± SD. The unpaired Student's t-test or Mann–Whitney U test was used to compare the means of the two groups. Specific statistical methods were selected based on the purpose of the study and the normality and homogeneity of the sample variance. Statistical significance was set at p < 0.05.

Omics analysis was performed using R version 3.6.1. Principal component analysis (PCA) was based on prcomp (R base function) 3.6.1, OPLS-DA was conducted based on ropls 1.6.2, and enrichment analysis was performed using clusterProfiler 4.4.4 and enrichplot 1.2.0. The t-test was used to calculate the differences in the significant p-values of each compound. Data with p < 0.05 and difference ratio >1 were selected for further analysis (VIP > 1 included in the metabolomic analysis).

Stem cell senescence induced by toxic microenvironments exacerbates inflammation through the SASP, which leads to deterioration of the pathological status of tissues [15]. To induce the potential therapeutic phenotype of UCMSCs while avoiding the adverse effects of cell senescence, the cells were treated with different concentrations of FS. The β-galactosidase staining results showed that UCMSCs were stable in normal medium, and 5% FS did not significantly induce senescence in UCMSCs within 12 h. At a serum concentration of 10%–100%, senescence and apoptosis of UCMSCs were widespread, and this effect occurred quickly (<12 h) (figure S2). Therefore, using 5% FS to intervene in UCMSCs was effective for inducing a therapeutic phenotype and avoiding cell senescence.

Next, we evaluated the response of UCMSCs to 5% FS (figure 1(b)). Transcriptome analysis showed that 5% FS induced the upregulation of 142 genes and downregulation of 249 genes in UCMSCs (figure 1(b)). The analysis of gene expression differences was based on p < 0.05 and fold change (FC) > 2.0. KEGG enrichment analysis showed that the upregulated genes were mainly enriched in NF-kappa B signaling pathway, cytokine-cytokine receptor interaction, TNF signaling pathway, and neuroactive ligand-receptor interaction (figure 1(c)). In addition, the downregulated genes were mainly enriched in TGF-β signaling pathway, cell adhesion molecule, hippo signaling pathway, and calcium signaling pathway (figure 1(d)).

Zoom In Zoom Out Reset image size

Based on the special response of UCMSCs to FS, we investigated whether this response mode had a therapeutic effect on liver fibrosis. Five-fold concentrated UCMSC-conditioned medium derived from 5% FS and 5% normal serum (5% NS) was injected into the body via the abdominal cavity or tail vein (figure 2(a)). Tissue evaluation showed that the four groups did not exhibit any significant therapeutic effects (figures 2(b)–(d)). Notably, the group of intraperitoneal injection of UCMSC-conditioned medium induced by 5% FS showed a trend of improvement in fibrosis degree. Considering the toxic effects of FS, the introduction of cells with metabolic and paracrine activities may have beneficial effects.

Zoom In Zoom Out Reset image size

To induce metabolic and paracrine activities, we introduced rat hepatocytes (BRL-3A cells) into the induction system (figure 3(a)). The conditioned medium derived from the co-culture system and 5% FS were used for the treatment of liver fibrosis in rats. Tissue evaluation showed that the intraperitoneal injection of conditioned medium derived from the co-culture system containing high-density (6 × 10⁵ cells well⁻¹) hepatocytes had a significant therapeutic effect (figures 3(b)–(d)). However, conditioned medium derived from a co-culture system containing low-density (1 × 10⁵ cells well⁻¹) hepatocytes and isolated high-density hepatocytes did not produce significant therapeutic effects. In addition, we found that a change in the injection mode of the conditioned medium led to a loss of the therapeutic effect. These results suggest that co-culture with hepatocytes plays a key role in the therapeutic effect of UCMSCs in response to 5% FS, and that the specific mode of administration is critical. Therefore, the further enhancement of hepatocyte activity is a key step in improving the therapeutic ability of co-culture systems.

Zoom In Zoom Out Reset image size

To further enhance hepatocyte function in the co-culture system and increase the production of therapeutic medium, we introduced a bioreactor system based on bionics and FDM. At this stage, the evaluation of cell adhesion and function on the surface of printing materials, optimization of the hepatocyte-UCMSC co-culture ratio, and generation of appropriate fluid shear force were gradually implemented.

The determination of cell adhesion and toxicity was a preliminary step. The implantation and functional determination of hepatocytes reflect the adhesion and toxicity of the PLA plane constructed using FDM. SEM revealed that the hepatocytes adhered well to the PLA surface and formed a cell layer (figure 4(a)). Ammonia scavenging and urea production suggested that the functioning of hepatocytes grown on the PLA surface was stronger than that of hepatocytes grown on the surface of 6-well plates (figures 4(b) and (c)). Analysis of the mRNA expression of key enzymes during ureagenesis showed that the PLA surface enhanced the expression of several key enzymes in hepatocytes, including ASL, coagulation factor II (F2), NAGS, OTC, and CPS1 (figure 4(d)). The albumin concentration in the culture medium and albumin mRNA expression in hepatocytes showed that the PLA surface did not have a promoting effect compared to the surface of the 6-well plate (figure 4(e)).

Zoom In Zoom Out Reset image size

The second step was to optimize the proportion of hepatocytes and UCMSCs in a co-culture system on the PLA surface. Hepatocytes were co-cultured with MSCs in gradient ratio (0, 15:1, 5:1 and 3:1) (figure 4(f)). A comprehensive analysis of ammonia clearance, urea production, and albumin secretion showed that hepatocytes mixed with UCMSCs at a ratio of 5:1 strongly promoted hepatocyte function (figures 4(g)–(i)). Next, a strong fluid shear force was applied to the cells that adhered to the PLA surface, with constant-speed shaking (40 rpm). SEM showed that the adhesion performance of the co-culture system was stronger, and the detachment of adherent cells in the fluid environment was avoided (figure 4(j)).

The third step involves the generation of an appropriate fluid shear force. Fluid is an important component of the liver microenvironment and an appropriate fluid shear force provides a specific microphysiological environment for hepatocytes and UCMSCs. To establish a stable fluid microenvironment, a novel bioreactor with regular flow channels, storage tanks, and gas valves was designed and fabricated. In this bioreactor, the cell adhesion surface with regular flow channels was the most critical (figures 5(a)–(d)). Twelve regular flow channels were included on the cell adhesion surface with cross-sectional dimensions of 0.5 × 0.3 cm. Hepatocytes and UCMSCs were co-cultured on one side of the fluid flow channel at a ratio of 5:1, and the ascending culture medium provided an appropriate fluid shear force for the cells (figure 5(e)). The culture medium was introduced into the lower storage tank at a constant flow rate and then into the upper storage tank through a regular flow channel (figures 5(e) and (f)). Different flow rates were used to determine the optimal fluid shear force. A comprehensive analysis of ammonia clearance, urea production, and albumin secretion showed that 0.6 dynes cm⁻² most strongly promoted hepatocyte function (figures 5(g)–(i)).

Zoom In Zoom Out Reset image size

In conclusion, bioreactors were designed and constructed to enhance the therapeutic effect of hepatocytes and UCMSCs co-culture system, including an appropriate co-culture ratio and fluid shear force. Next, 5% FS was introduced into the bioreactor to prepare the therapeutic medium.

5% FS was used as the circulating medium, with a circulation time of 12 h. Circulating 5% FS was concentrated five times by freeze-drying and redissolving the therapeutic medium. Therapeutic medium was intraperitoneally administered to treat liver fibrosis in rats (figure 6(a)). After four weeks of treatment, we analyzed the degree of liver fibrosis using Masson and Sirius Red staining and found that intraperitoneal injection of the therapeutic medium derived from 5% FS (5% FS-TM) significantly reduced collagen deposition (figures 6(b)–(d)). However, the direct use of 5% FS did not show a therapeutic effect. NS (5%) and its derived therapeutic medium (5% NS-TM) failed to alleviate liver fibrosis. These results suggest that intraperitoneal injection of 5% FS-TM is an effective method for alleviating liver fibrosis. Notably, this therapeutic effect did not rely on stem cell infusion but rather on changes in the secretion of stem cells in response to the FS environment in the bioreactor.

Zoom In Zoom Out Reset image size

In addition, we evaluated the therapeutic ability of these media via tail vein injection and the staining results showed that none of the media had a therapeutic effect (figure S3). These results emphasize the importance of using a therapeutic medium.

To further understand the response patterns and potential fibrosis regulatory mechanisms of the co-culture system in the bioreactor for FS, we used proteomics and metabolomics to identify differences in the medium before and after circulation.

A total of 522 proteins were identified in the treatment medium. The PCA results showed that the four media clustered separately with significant differences between them (figure S4(a)). Reactome pathway annotation revealed that 63.4% of these proteins are associated with the immune system (figure 7(a)). The protein expression heatmap showed significant differences between the pre-reaction and reaction media, as well as between normal and fibrosis serum (figure 7(b)). The analysis of protein expression differences was based on p< 0.05 and FC > 1.5. The results showed a significant difference in 156 proteins between serum with fibrosis and normal serum, including 61 upregulated and 95 downregulated proteins (figure 7(c)). Reactome pathway annotation showed that 33.3% of the differentially expressed proteins were enriched in the immune system (figure 7(d)). After circulation, 5% NS-TM had 50 upregulated and 57 downregulated proteins, and 30.8% of the differentially expressed proteins were enriched in the immune pathway (figures 7(e) and (f)). 5% FS-TM had 49 upregulated and 23 downregulated proteins after circulation; 40.2% of the differentially expressed proteins were enriched in the immune pathway (figures 7(g) and (h)). KEGG analysis showed that normal serum mainly induced reactions related to the Salmonella infection pathway and microRNAs in the cancer pathway in the bioreactor, whereas FS mainly induced the biosynthesis of amino acids and carbon metabolism pathway-related reactions (figures S4(b) and (c)). Owing to the significant therapeutic effect of 5% FS-TM, we further analyzed the common protein expression differences compared to those of the other three treatment media. Fourteen differentially expressed proteins were identified, of which 10 were upregulated and four were downregulated (figure 7(i)). Among the set of common differentially expressed proteins, NUTF2, NCL, COTL1, RAB1B, PFN1, moesin, PGK1, GAPDH, PDIA4, tenascin C, complement C6, complement C8, and CFHR4 were associated with immune regulation [16–27]. Therefore, the significant changes in protein levels in 5% FS-TM were mainly concentrated in immune regulatory proteins compared with those of other media.

Zoom In Zoom Out Reset image size

The analysis of metabolites in the medium was based on the LC-QTOF platform, with 4613 metabolites annotated (1846 annotated in negative ion mode, 2767 annotated in positive ion mode). The metabolite classification focused on lipids and lipid-like molecules, organic acids and derivatives, organic oxygen compounds, organoheterocyclic compounds, benzenoids, phenylpropanoids, and polyketides (figure S5(a)). The PCA results showed that the four media clustered separately with significant differences between them (figure S5(b)). The heatmap of metabolite expression shows a wide range of changes in the metabolites in the pre-reaction and reaction media (figure 8(a)). The following analysis of the differences in metabolite expression was based on p < 0.05 and FC > 1.5. Compared with normal serum, FS contained 1560 differential metabolites, including 1257 upregulated and 303 downregulated metabolites (figure S5(c)). The results of the KEGG enrichment analysis showed that primary bill acid metabolism was the most relevant pathway (figure 8(b)). Compared with 5% NS, 521 and 829 metabolites were upregulated and downregulated, respectively, in 5% NS-TM (figure S5(d)). KEGG enrichment analysis showed that glutathione metabolism was the most relevant pathway, consisting of six upregulated and six downregulated metabolites (figures 8(c) and (d)). A total of 552 metabolites were downregulated and 826 metabolites were upregulated in the 5% FS-TM group (figure S5(e)). KEGG enrichment analysis showed that glutathione metabolism was the most relevant pathway, consisting of six upregulated and four downregulated metabolites (figures 8(e) and (f)). The glutathione metabolic pathway is a common pattern, in which the bioreactor responds to both normal and fibrosis serum. Furthermore, common differential metabolites were identified between 5% FS-TM and the other three media, including 241 upregulated and 78 downregulated metabolites. Additionally, the results of KEGG enrichment analysis showed that arachidonic acid metabolism was the most relevant pathway for common differential metabolites (figures 8(g) and (h)). Arachidonic acid metabolic pathway is closely associated with immune regulation [28–30]. The unique enrichment of arachidonic acid-related metabolites in 5% FS-TM suggests an association between fibrotic regression mechanisms and immune regulation.

Zoom In Zoom Out Reset image size

Based on the results of proteomic and metabolomic analyses suggesting potential immune regulation by 5% FS-TM, we further analyzed the evolution of the immune state of the fibrotic liver induced by treatment. Macrophages, neutrophils, and natural killer (NK) cells are key immune subpopulations involved in liver fibrosis [31–34]. The immune state changes induced by the treatments were evaluated by determining the degree of infiltration of different immune cells (figure 9(a)). We found that the degree of infiltration of macrophages, neutrophils, and NK cells in the fibrotic liver was significantly increased and that 5% FS-TM significantly reversed this pathological process (figure 9(b)). The inflammatory status of the liver was evaluated based on mRNA expression levels of pro- and anti-inflammatory cytokines. 5% FS-TM reduced the expression levels of pro-inflammatory factors such as IL-6, TGF-β1, and TNF in the fibrotic liver but did not affect other cytokines (figure 9(c)).

Zoom In Zoom Out Reset image size

MMPs are another direct tool used by immune cells to influence fibrosis [35]. MMPs promote the remodeling of fibrotic tissue and are regulated by TIMPs. We further evaluated the mRNA expression levels of MMPs and TIMPs in different treatment groups. Most MMPs and TIMPs in the fibrotic liver remained stable except for the high expression levels of MMP13 and TIMP1 (figure 9(d)), and 5% FS-TM did not alter MMP expression levels. Therefore, the regression of fibrosis induced by 5% FS-TM could be partly explained by the improvement in the inflammatory environment, but not by tissue remodeling mediated by MMPs.

Immune infiltration and MMP secretion respond quickly to pathological changes; however, this response disappears after steady-state recovery. The results of our investigation after fibrosis relief did not detail the dynamic changes in the tissue remodeling process. Therefore, we further evaluated fibrotic liver status in the early treatment stages. Treatment with 5% FS-TM thrice did not significantly improve the degree of liver fibrosis (figure 10(a)). The results of the immune infiltration analysis showed changes in macrophage infiltration, including a decrease in the degree of infiltration and redistribution of the infiltrated areas (figures 10(b) and (c)). The main characteristic of macrophage redistribution was a decrease in infiltration in the parenchymal area, accompanied by an increase in infiltration in the fibrotic area (figure 10(c)). Neutrophils and NK cells showed decreased infiltration (figure 10(d)). In addition, the results of mRNA expression analysis showed a significant decrease in the expression of pro-inflammatory factors including IL-6, TGF-β1, and TNF, as well as a significant increase in the expression of anti-inflammatory factor IL-10 (figure S6(a)). Evaluation of MMPs and TIMPs showed a significant increase in MMP3 and MMP9 expression, and a decrease in TIMP1 and TIMP2 expression (figure S6(b)). These results suggest that 5% FS-TM induced immune redistribution, inflammation regression, and initiation of tissue remodeling processes; this evolution occurred during the initial treatment stages.

Zoom In Zoom Out Reset image size

HSC activation in response to an inflammatory environment is a key event in the progression of liver fibrosis [36–39]. The proliferation and activation of HSCs before and after treatment were evaluated based on α-SMA and desmin staining [40]. α-SMA staining results showed an increase in the number of activated HSCs in the fibrotic liver, whereas the number of activated HSCs significantly decreased after treatment with 5% FS-TM (figures S7(a) and (b)). Desmin is a marker of all HSCs. Desmin staining indicated significant proliferation of HSCs in the fibrotic liver, and 5% FS-TM treatment significantly reduced the number of HSCs (figures S7(c) and (d)).

Hepatocyte damage including apoptosis, DNA damage, and proliferation was evaluated histologically. TUNEL staining results showed an increase in hepatocyte apoptosis in the fibrotic liver and a significant decrease in the number of apoptotic cells after treatment with 5% FS-TM (figures S8(a) and (d)). γH2Ax is a specific marker of cellular DNA damage. γH2Ax staining showed an increase in hepatocyte DNA damage in fibrotic livers, and 5% FS-TM treatment reduced the degree of DNA damage (figures S8(b) and (d)). Ki-67 staining revealed a fibrosis-induced hepatocyte proliferation phenotype, which could be a cellular response to tissue damage and the inflammatory microenvironment (figures S8(c) and (d)). Owing to tissue remodeling and an improved microenvironment, 5% FS-TM reduced the degree of hepatocyte proliferation. In summary, fibrosis-induced stress in hepatocytes and treatment with 5% FS-TM improved this state.

Oval cells are key components of liver tissue remodeling and have been identified in rats via dual labeling with CK19 and GGT [41]. We evaluated the status of the oval cells in the fibrotic liver before and after treatment using dual-fluorescence staining. We observed the presence of CK19+/GGT+ oval cells in the bile duct tree region and a significant decrease in the number of oval cells, accompanied by a significant increase in the number of cholangiocytes in the fibrotic liver (figures S9(a)–(c)). After treatment, the number of oval cells recovered and the new bile ducts disappeared.