Paracrine Effects Of Adipose-Derived Stromal/Stem Cells And Stromal Vascular Fraction In An In Vitro Fibrogenesis Model Of Human Vocal Fold Scarring

Background: Vocal folds (VF) scarring leads to severe dysphonia which negatively impacts daily life of patients. Current therapeutic options are limited due in large part to the high complexity of the microstructure of the VF. Innovative therapies derived from adipose tissue such as stromal vascular fraction (SVF) or adipose derived stromal/ stem cells (ASC) are currently being evaluated in this indication and paracrine anti-fibrotic effects are considered as predominant mechanisms. Methods: The paracrine anti-fibrotic effects of SVF and ASC from healthy donors were tested in an innovative in vitro fibrogenesis model employing human VF fiboblasts (hVFF) and the principles of macromolecular crowding (MMC). Biosynthesis of collogen and alpha-smooth-muscle actin (αSMA) expression in hVFF were quantified after five days of indirect coculture with ASC or SVF using silver stain, western blot and RT-qPCR analysis. Results: Fibrogenesis was promoted by addition of transforming growth factor beta 1 (TGFβ1) combined with MMC characterized by an enhanced deposition of fibrillar collagens and the acquisition of a myofibroblast phenotype (overexpression of αSMA). Adipose-derived therapies led to a reduction in the αSMA expression and the collagen content was lower in hVFF co-cultivated with SVF. Discussion: ASC and SVF promoted significant prevention of fibrosis in an in vitro fibrogenesis model through paracrine mechanisms, supporting further development of adipose-derived cellular therapies in VF scarring.


Introduction
Vocal fold (VF) scar is one of the most challenging benign laryngeal pathologies [1]. Depending on severity and extent, it can result in a range of causes both physical and psychological disability, especially for persons with a high vocal demand, such as teachers, singers or call center agents.
VF microstructure is complex [2], particularly due to its foliated organization allowing optimal vibration. The proportions and the organization of the extracellular matrix (ECM) components largely determine the mechanical properties of the VF. VF scarring is most often the result of laryngeal microsurgery, but it may also be found congenitally (sulcus vocalis) oracquired following a trauma or chronic inflammatory phenomena. Thus, the normal microstructure is replaced by fibrous tissue, with an excessive and disorganized ECM inducing vibration disorder [1]. Compared to normal fibroblasts, VF scar tissue fibroblasts produce a larger amount of collagen, proliferate more rapidly and overexpress α-smooth-muscle actin (αSMA), also known as actin alpha 2 (ACTA2), a myofibroblast differentiation marker [3,4].
Other more experimental treatments such as administration of basic Fibroblast Growth Factor (bFGF) [9, 10] or Hepatocyte Growth Factor (HGF) [11][12][13][14] have shown encouraging effects in vitro or in vivo on animals. However, only one phase I/II clinical trial assessed the safety and effectiveness of intracordal injection of a recombinant human HGF in patients with vocal fold scar or sulcus [15]. The results of this first-in-man clinical trial indicated no serious adverse events and encouraging clinical results. Thus, the identification of innovative strategies able to improve the mechanical vibrational properties of VF remains an important clinical challenge.
Mesenchymal stromal cells (MSC), which were first described in bone marrow to support hematopoiesis [16], are multipotent stem cells that can be isolated from a variety of tissues but adipose tissue remains a very attractive source due to its abundance and accessibility. Autologous adipose-derived Stromal Vascular Fraction (SVF) is a heterogeneous population of cells obtained in a few hours after enzymatic digestion of adipose tissue that does not require an expansion step. SVF is also an advantageous source of adipose derived stromal cells (ASC) because of its fluidity and ease to inject. Besides, SVF brings the synergistic effect of its various cellular subpopulations, including ASC but also endothelial progenitor cells combining vasculogenic, anti-fibrotic and anti-inflammatory properties. The exact mechanism of action of ASC and SVF are not fully elucidated but it is now admitted that the major effect rests on paracrine activity through a large type of secreted molecules [17][18][19].
To facilitate research on this field, Graupp et al. described an in vitro fibrogenesis model using human vocal fold fibroblasts (hVFF) and the principles of macromolecular crowding (MMC) [20]. MMC promotes fibrogenesis after the addition of transforming growth factor-beta 1 (TGF-β1) to a medium containing inert macromolecules. The aim of this experimental study was to assess the in vitro paracrine potential of SVF cells and ASC on scarred hVFF in this innovative model.

Materials And Methods Donors
ASC and SVF cells from healthy donors (n = 4) were obtained from adipose tissue surgical residues following liposuction for aesthetic purposes. All patients provided informed consent for the scientific use of surgical residues.
Cell culture Immortalized human vocal fold fibroblasts (hVFF) were obtained from the lab of Prof. Susan Thibeault (University of Wisconsin, Madison, USA) [21]. Dulbecco´s modified Eagle´s medium (DMEM) (Life Technologies, Waltham, Massachusetts) supplemented with 10% fetal calf serum (FCS; Sigma-Aldrich, Vienna, Austria) and 100 µg/mL Normocin (Invivogen, San Diego, California, USA) was used as standard medium (SM). Adipose tissue collection was conducted under conscious sedation with local anesthesia. SVF was obtained in a Cell Therapy Unit using the automated processing Celution 800/CRS system (Cytori Therapeutics Inc., San Diego, California, USA) as previously described [22]. Trypsin-EDTA and passed to extend the culture. For co-culture experiments, hVFF from passages 13-20 were seeded into cell culture plates (15.000 cells/cm 2 ). The following day, SM was switched to FCS-free SM for another 24 h. Cells were then allocated to different treatment groups. For the control group DMEM enriched with 0.5% FCS, 100 µg/mL Normocin and 100 nM of L-ascorbic acid 2phosphate was used [23]. Myofibroblast differentiation was induced by adding of 5 ng/mL TGF-β1 [24] combined with "macromolecular crowding" (MMC) by adding a mixture of 37.5 mg/mL 70 kDa Ficoll (Fc) with 25 mg/mL 400 kDa Fc (both Sigma-Aldrich) [25,26]. To assess possible anti-fibrotic effects, three conditions were tested in co-culture with hVFF: (i) recombinant HGF (50 ng/mL, Sigma-Aldrich), (ii) paracrine effect of thawed SVF and (iii) paracrine effect of thawed ASC from passages 3-5. After 5 days of incubation, samples were further processed.
Reverse Transcription-qPCR (RT-qPCR) RNA isolation of hVFF was performed using the QIAZOL Lysis Reagent (Qiagen, Hilden, Germany) and total mRNA was isolated with the miRNeasy Mini Kit (Qiagen) according to the manufacturer´s instructions. Purified RNA was eluated in RNAse-free water and concentration was determined using the NanoDrop 2000c spectrophotometer (Thermo Scientific). Reverse transcription (RT), as well as RT quantitative PCR (RT-qPCR), was performed as previously described [27]. Primer sequences are provided in Table 1. Relative quantification of all mRNAs of interest was performed based on the 2 −ΔΔC T method [28].
Pepsin digestion, sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and silver stain Supernatants and cell layers from hVFF seeded in 24-well-plates were harvested separately and pepsinized as previously described [29]. Briefly, 50 µl of a pepsin stock solution (1 mg/mL dissolved in 1N HCL) were added to 500 µL of supernatants, while a pepsin digestion solution (25% pepsin stock solution, 0.005% Triton X-100 in ddH 2 O) was added to cell layers. All samples were incubated for 2 h followed by neutralization with 1 N NaOH. Gels were subsequently stained with the SilverQuest™ Silver Staining kit (Thermo Scientific) according to the manufacturer´s protocol. Gel images were acquired using Quantity One software (Bio-Rad) and densitometric analysis of bands was performed using Image Lab Software, Version 5.2.

Western blot
Proteins were extracted from hVFF cell layers seeded in 6-well-plates and subjected to SDS-PAGE using 4-12% Criterion XT Bis-Tris Gels (Bio-Rad). SDS-PAGE was run at 200 V for 60 min, followed by electroblotting of proteins (90 min at 0.5 A and 4 °C) onto Nitrocellulose membranes (Bio-Rad).
Subsequently, blots were incubated with secondary antibody (Goat Anti-Mouse 1:5000 for ACTA2; Goat Anti-Rabbit 1:5000 for GAPDH, both abcam, Cambridge, UK). Detection was conducted using the SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific) and the ChemiDoc Touch Imaging System (bio-rad), densitometric analysis was performed using Image Lab Software, Version

Validation of the MMC/TGF-β1 model
We first validated the myofibroblast differentiation induced by TGF-β1 combined with MMC (MMC/TGF-β1) through the assessment of ACTA2 and Collagen Iα1 (COL1A1) contents in hVFF. Results from RT-qPCR analysis revealed that mRNA expression of ACTA2 and COL1A1 were significantly up-regulated under MMC/TGF-β1 condition (p = 0.019 and p = 0.002 respectively) ( Fig. 1A and 2A) compared to the negative condition (untreated fibroblasts). Western blot and silver stain analysis performed on hVFF protein lysates were in line with the previous data: we observed a significant elevation of the ACTA2/GADPH ratio (p = 0.024) and of fibrillar collagen contents (p = 0.0004) in the MMC/TGF-β1 condition compared to the untreated condition ( Fig. 1B and 1C). Altogether, these data validated the myofibroblast phenotype acquisition of immortalized hVFF after MMC/TGF-β1 treatment.

Co-culture
The expression of the myofibroblast marker ACTA2 and collagen content of hVFF co-cultured with ASC or SVF or treated with HGF for 5 days were assessed. The results revealed a reduction in the mRNA level of ACTA2 on hVFF when they were co-cultured with ASC and SVF (p = 0.043 and p = 0.028 respectively) (Fig. 1A). Western blot analysis confirmed these results highlighting a tendency to reduce the level of ACTA2 in hVFF co-cultured with both ASC and SVF (p = 0.061 and p = 0.083 respectively) (Fig. 1B and 1C). Figure 2 represents the expression of mRNA of COL1A1 and fibrillar collagen in protein lysates of hVFF. We observed no significant reduction in the COL1A1 mRNA expression ( Fig. 2A) in both ASC and SVF co-culture conditions whereas silver stain analysis revealed a significant reduction of fibrillar collagen content in hVFF after co-culture with SVF (p = 0.006) (Fig. 2B and 2C). Of note, no significant reduction of ACTA2 or collagen levels were obtained when hVFF were treated with HGF ( Figs. 1 and 2).

Discussion
Vocal fold scarring remains one of the most difficult conditions to treat among benign VF diseases and adipose-derived cell therapies represent hope for these patients. The therapeutic potential of MSC has largely been attributed to paracrine activity [30,31]. In this study, we demonstrated that ASC and SVF promoted significant prevention of fibrosis in an in vitro model through paracrine mechanisms.
Previously, a similar study explored whether ASC have therapeutic potential for treating vocal fold scarring through the effects of HGF on scar fibroblasts in an indirect co-culture model [4]. They showed that co-culture with ASC led to a significant decrease of ACTA2 expression in fibroblasts comforting our data. However, they also demonstrated a reduction in collagen production in the ASC co-culture conditions whereas, in our study, we did not observe significant effect of ASC on the fibrillar collagen content.
Compared with ASC, SVF can be real time obtained in a sufficient quantity without in vitro culture, reducing drastically the delay of preparation. Yasuda et al. reported that injection of SVF had a protective effect in an acute kidney injury in vivo model by secreting renoprotective molecules, such as HGF and VEGF [32]. We can hypothesize that in the present study, SVF and ASC reduced the ACTA2 and collagen expression in hVFF partly by secreting those growth factors transferred through the co-culture system. Only conditions with SVF showed a reduction in both ACTA2 and collagen contents which might be attributed to the higher secretion of soluble mediators by SVF. In line with this, the secretome of ASC and SVF has been compared recently and revealed that SVF released a greater variety of cytokines or soluble protein at a significantly higher amounts as compared with ASC [33].
Several in vivo studies on animals reported significant improvements in scarred vocal folds after injection of ASC (immediately after injury or at a longer interval) in both histological and macroscopic aspects of the vocal folds [34] with an excellent tolerance of the procedure. A first clinical trial of SVF injection which enrolled 8 patients was recently reported by our group [35] suggesting a clear improvement in the majority of the vocal parameters obtained at 12 months. However, the etiology of vocal fold scarring is very heterogenous so a better understanding of the mechanisms underlying the efficacy of ASC and SVF is mandatory. In addition, identification of in vitro potency assay able to anticipate the anti-fibrotic potential of the cell therapy product could allow a better selection of patients who can benefit from these innovative therapies.

Ethics approval and consent to participate
All patients provided informed consent for the scientific use of surgical residues.