Development of an iPSC-based screening platform identifying enhancers of chondrogenesis

Abstract

Objective

There is currently no long-term treatment for the repair of damaged cartilage and osteoarthritis (OA). Induced pluripotent stem cells (iPSCs) are an ideal cell source for screening platforms due to their ability to self-renew and differentiate to cell types that would otherwise require invasive surgeries to obtain, such as chondrocytes and mesenchymal stromal cells (MSCs). Here, we developed an iPSC-based screening platform and tested previously described pro-chondrogenic small molecule compounds, to determine their potential to identify hits.

Design

iPSC derived chondroprogenitors (iCPs) and neural crest cell (NCC) derived MSCs (iNCC-MSCs) were generated, and their chondrogenic potential was confirmed. The iPSC derived cells and a primary bone marrow derived MSC (BM-MSC) line were cultured as pellets and treated with different concentrations of small molecule compounds, in the presence of chondrogenic inducing growth factors, over 14 days at 2 ​% O₂. Glycosaminoglycan (GAG) synthesis was quantified by a 1,9- dimethylmethylene blue (DMMB) assay.

Results

After 14 days of chondrogenesis, forskolin, baicalin and sesamin enhanced GAG synthesis in the iCPs, and forskolin enhanced GAG synthesis in the iNCC-MSCs, while no small molecule compounds enhanced GAG synthesis in the BM-MSCs.

Conclusion

Our findings further demonstrate how the small molecules pro-chondrogenic effects are dependent on the screening platform conditions, including the cell type, molecule concentration, 3D culture, hypoxia, and the inclusion of additional growth factors. The iPSC-based screening platform developed has the potential to identify disease modifying OA drugs (DMOADs) in novel compound screening libraries.

1. Introduction

Osteoarthritis (OA) is a major debilitating disease estimated to affect 595 million people globally, with increasing prevalence due to an ageing population [1]. It is characterized by the progressive degradation and loss of articular cartilage, inflammation of the synovial tissue, damage to the subchondral bone, and the formation of osteophytes [2]. The presence of partial or full thickness focal cartilage defects increases the risk of further cartilage damage and OA [3]. There is an unmet clinical need for cellular and pharmacological treatments for cartilage repair and OA. A number of disease modifying OA drugs (DMOADs) are described to have pro-chondrogenic and/or chondroprotective properties, for example Kartogenin and LNA043 [4,5]. Some molecules such as Sprifermin and SM04690 have shown promising results in clinical trials [6,7], but so far none have been approved for clinical use in treating OA or cartilage defects.

The development of methods to generate human induced pluripotent stem cells (iPSCs) has had a transformative impact on medical science by providing accessible and powerful tools for disease modelling, drug screening and for devising new cell-based therapies. The ability to generate disease-specific iPSC lines from patients and to study their differentiated progeny has opened many opportunities for understanding the cellular pathology of monogenic diseases and more complex disorders. The same cellular models may be used to understand drug responses and, because of the ease of scale up, may be used as information-rich screening tools for discovery of candidate molecules for therapeutic development. Applications of iPSCs in drug screening may also involve the generation of organoids, recapitulating the complexity of whole tissues and reducing the need for animal experimentation in research. Finally, the use of iPSCs as an alternative to primary cells in the development of cellular therapeutics provides new opportunities as these cells are self-renewing and provide non-depletable stocks, removing the heterogeneity associated with primary cells. iPSCs can be induced toward a chondroprogenitor (iCP) or mesenchymal stromal cell (iMSC) intermediate, that can generate hyaline-like articular cartilage with therapeutic potential [8,9]. However, challenges associated with iPSC chondrogenic differentiation methods include cell heterogeneity, loss of chondrogenic capacity during culture, and possible differentiation toward fibrous and hypertrophic cartilage [10,11].

Screening platforms using iPSC-derived cells have been developed to test DMOADs [12,13]. Murine iPSC-derived chondrogenic pellets cultured in 96-well plates were challenged with interleukin 1 alpha (IL-1α), treated with candidate OA drugs, and glycosaminoglycan (GAG) release in the media was measured by a dimethymethylene blue (DMMB) assay. The nuclear factor kappa B (NF-κB) inhibitor SC-514 had the strongest chondroprotective effects [13]. Organ-on-a-chip models can better mimic cartilage and the OA environment. Lin et al. (2019) generated iPSC derived osteochondral tissue chips, induced an OA disease phenotype with IL-1β, and treated the tissue chips with Celecoxib which reduced expression of inflammatory and catabolic genes [12].

Our objective in this work was to develop a screening platform involving iPSCs to facilitate early identification of molecules as targets for cartilage disease and OA. Following directed differentiation protocols, human iPSCs were differentiated toward iPSC derived chondroprogenitors (iCPs) and iPSC derived neural crest cell derived-MSCs (iNCC-MSCs) for inclusion in the screening platform. Primary mesenchymal stromal cells are often used to screen for DMOADs [5], therefore human bone marrow derived-MSCs (BM-MSCs) were included as a control cell line. All three cell types were treated with small molecule compounds described to have pro-chondrogenic properties, to determine their potential to identify hits. Secondly, any hits identified may contribute toward the development of more optimal iPSC chondrogenic differentiation protocols. A schematic of the identification and selection of small molecule compounds is described in Fig. 1. Following a PubMed literature review, an exploration of chemicals predicted as upstream regulators of chondrogenesis using Qiagen's Ingenuity Pathway Analysis software (QIAGEN Inc., https://digitalinsights.qiagen.com/IPA), and an exploration of commercially available screening compound libraries, 99 small molecules of interest were identified (Supplementary File 1). These molecules were described to have one or more roles in promoting chondrogenic differentiation, cartilage extracellular matrix (ECM) synthesis, cell proliferation, anti-inflammation and cartilage protection. 41 molecules were excluded because they were not described to have pro-chondrogenic properties, or the results were limited. 12 molecules were additionally excluded because they are either not commercially available, not a chemical compound, or too toxic. 18 molecules were selected for inclusion in the screening platform, aiming to include both synthetic and naturally derived compounds, that ranged in their mechanism of action and working concentrations, and were investigated in different cell types/animal models (Table 1). The iCPs, iNCC-MSCs, and BM-MSCs were cultured as chondrogenic pellets in hypoxia (2 ​% O₂), and treated with small molecules in the presence of chondrogenic inducing growth factors, over 14 days. 3 small molecule compounds were found to enhance GAG synthesis in iPSC derived cells, while no hits were identified from BM-MSCs.

Fig. 1.

Schematic describing the identification and selection of small molecule compounds for inclusion in the screening platform.

Table 1.

List of molecules included in the screening platform.

Small Molecule	Mechanism of action	In Vitro study (cell type)	In Vivo study (animal model)	Reference
XAV939	Tankyrase inhibitor	Mouse ACs, mouse limb-bud mesenchymal cells, human BM-MSCs	Mouse DMM OA model	[14]
TD-198946	Induced RUNX1 expression	ATDC5 cells, C3H10T1/2 ​cells	Mouse OA model (ligament transection and meniscectomey)	[15]
BNTA	Induced SOD3 expression	Human OA chondrocytes and cartilage explants	Rat ACLT OA model	[16]
RCGD 423	Activated gp130 signalling	Healthy and OA human ACs	Rat partial meniscectomy model and rat osteo-chondral defect model	[17]
Ellipticine	Induced SOX9 expression	Human ADSCs	Rat type II collagenase induced OA model	[18]
17-DMAG	HSP90 and p21 inhibitor	Human SSCs and ACs, mouse limb-bud chondrocytes	Mouse auricular cartilage injury model	[19]
DAPT	Gamma gamma secretase inhibitor	Mouse limb bud MPCs		[20]
Sesamin		Human AF-MSCs		[21]
	Activated BMP2 and ERα signalling	Human BM-MSCs	Mouse osteoporotic bone fracture model	[22]
Salidroside	Activated TGFβ/Smad3 signalling	Rat ACs	Rat articular cartilage defect	[23]
SP600125	JNK inhibitor	ATDC5 cells		[24]
CGS 21680	Adenosine A2A receptor agonist		Mouse obesity induced OA model and rat post-traumatic OA model (ACL rupture)	[25]
SM04690 (Lorecivivint)	Wnt pathway inhibitor	Human BM-MSCs, ATDC5 cells and primary mouse calvaria cells	Rat ACLT and partial medial meniscectomy OA model	[26]
	Inhibited CLK2 and DYRK1A activity	Human MSC, chondrocytes and synovial fibroblasts	Rat MIA induced OA model	[27]
GSK1016790A	TRPV4 agonist	ATDC5 cells		[28]
		Mouse Col2A1-GFP iCPs		[29]
FK506 (Tacrolimus)	Calcineurin inhibitor	ATDC5 cells		[30]
	Activated Smad signalling pathways	Human SSCs		[31]
Fluvastatin		Human OA chondrocytes	Rabbit ACLT OA model	[32]
	Induced BMP2 expression and inhibited RhoA-ROCK signalling	Human ADSCs		[33]
Curcumin	NF-κB inhibitor	Rat AC and BM-MSC cocultures	Rat MIA induced OA model	[34]
	Promoted SIRT1 expression	Rat chondrocytes	Rat ACLT OA model	[35]
Forskolin	Adenylyl cyclase activator	Human BM-MSC		[36]
		Human PSC derived ecto-mesenchymal cells and nasal chondrocytes		[37]
Baicalin	Activated HIF1-α expression	Mouse chondrocytes		[38]
	Activated TGFβ/Smad3 signalling.	ATDC5	Mouse DMM OA model	[39]

Abbreviations: OA, osteoarthritis. DMM, destabilization of the medial meniscus. ACLT, anterior cruciate ligament transection. ACs, articular chondrocytes. BM-MSCs, bone-marrow derived mesenchymal stromal cells. SSCs, synovial stromal cells. ADSCs, adipose derived stromal cells. MPCs, mesenchymal progenitor cells. AF-MSCs, amniotic fluid derived mesenchymal stromal cells. PSC, pluripotent stem cell. RUNX1, runt-related transcription factor 1. SOD3, superoxide dismutase 3. GP130, glycoprotein 130. SOX9, SRY-box transcription factor 9. HSP90, heat shock protein 90. BMP2, bone morphogenetic protein 2. ERα, estrogen receptor alpha. TGFβ3/Smad3, transforming growth factor beta 3/SMAD family member 3. JNK, c-Jun N-terminal kinase. CLK2, CDC-like kinase 2. DYRK1, dual-specificity tyrosine phosphorylation-regulated kinase 1A. MIA, monosodium idoacetate. TRPV4, transient receptor potential vanilloid 4. RhoA, ras homolog family member A. ROCK, Rho associated coiled-coil containing protein kinase. IL-1β, interleukin 1 beta. NF-kB, nuclear factor kappa B. SIRT1, sirtuin 1. HIF1-α, hypoxia inducible factor 1 subunit alpha.

2. Materials and methods

2.1. Cell culture

iPSCs (https://hpscreg.eu/cell-line/UMNi001-A) were generously provided by Professor James R. Dutton (University of Minnesota), and cultured on Geltrex (Gibco) coated 6-well tissue culture plates in Essential 8 (E8) Flex Medium (Gibco), at 37 ​°C, 5 ​% CO₂. The cells were passaged within 5 days of culture when the colonies reached 80 ​% confluency, with Gentle Cell Dissociation Reagent (Stem Cell Technologies), and passaged as aggregates onto pre-coated 6-well plates. BM-MSCs were cultured as a monolayer in MSC culture medium containing 10 ​% Fetal Bovine Serum (FBS, HyClone), 1 ​% Penicillin/Streptomycin (Gibco), and MEM alpha GlutaMAX (Gibco), supplemented with 1 ​ng/ml basic fibroblast growth factor (bFGF, Peprotech), at 37 ​°C, 5 ​% CO₂. BM-MSCs were passaged at 70 ​% confluency with 0.25 ​% Trypsin-EDTA solution (Gibco), and seeded at 2.8 ​× ​10³ ​cells/cm² in tissue culture flasks in MSC culture medium.

2.2. iCP differentiation

iPSCs were induced toward iCPs (N ​= ​3 individual experiments) and cultured following a previously published method [40]. The chondrogenic potential of iCPs was confirmed at passage (P) 1. The cells were pelleted in sterile screw cap microcentrifuge tubes (Sarstedt) at 3.5 ​× ​10⁵ ​cells/pellet, in chondrogenic induction medium described by Adkar et al. (2019) [40] supplemented with 10 ​ng/ml transforming growth factor beta 3 (TGFβ3, Peprotech) to induce chondrogenesis. The caps were loosened to allow for gas exchange, and incubated under hypoxic conditions at 37 ​°C, 5 ​% CO₂, 2 ​% O₂ for 28 days, changing the media every 2–3 days.

2.3. iNCC-MSC differentiation

iPSCs were differentiated toward iNCCs and subsequently iNCC-MSCs (N ​= ​3 individual experiments) following previously published methods with slight modifications [[41], [42], [43]]. iNCC-MSCs were seeded at a density of 2.8 ​× ​10³ ​cells/cm² on 6 ​μg/ml fibronectin (Sigma) coated tissue culture flasks, in MSC culture media supplemented with 20 ​ng/ml bFGF, and passaged at 70 ​% confluency with Trypsin-EDTA up to P4. iNCC-MSCs at P3 were assessed for their chondrogenic differentiation capacity. The cells were pelleted in sterile screw cap microcentrifuge tubes at 3 - 3.5 ​× ​10⁵ ​cells/pellet, in incomplete chondrogenic medium (ICM) containing DMEM High Glucose (DMEM HG, Sigma), 1 ​% ITS ​+ ​Premix (6.25 ​μg/ml insulin, 6.25 ​μg/ml transferrin, 6.25 ​ng/ml selenous acid, 5.35 ​μg/ml linoleic acid, 1.25 ​mg/ml bovine serum albumin. Corning), 1 ​% Penicillin Streptomycin, 50 ​μg/ml l-ascorbic acid 2-phosphate (AA2P, Sigma), 40 ​μg/ml l-proline (Sigma), 100 ​nM Dexamethasone (Sigma), and 1 ​mM Sodium Pyruvate (Gibco). The caps were loosened to allow for gas exchange, and incubated under hypoxic conditions at 37 ​°C, 5 ​% CO₂, 2 ​% O₂. On day 0, chondrogenesis was induced with 40 ​ng/ml platelet derived growth factor BB (PDGF BB, Protientech), on day 2, 5, and 8 the media was changed to ICM supplemented with 40 ​ng/ml PDGF BB and 10 ​ng/ml TGFβ3. From day 10–28 the media was changed to ICM containing 10 ​ng/ml TGFβ3 and 50 ​ng/ml bone morphogenetic protein 4 (BMP4, Proteintech), every 2–3 days.

2.4. Reverse transcription and quantitative polymerase chain reaction

RNA was isolated using the Purelink RNA Mini kit (Invitrogen) and genomic DNA was removed with Purelink DNase (Invitrogen). 28-day chondrogenic pellets were snap frozen on dry ice and homogenized with a Pellet Pestle Motor (Fisherbrand) prior to RNA isolation. RNA concentration and purity was measured with a NanoDrop 2000 spectrophotometer. RNA was reverse transcribed to complementary DNA (cDNA) using the Superscript VILO cDNA Synthesis kit (Invitrogen). The qPCR was set up in MicroAmp Fast Optical 96-well plates (Applied Biosystems) using Fast SYBR Green master mix (Applied Biosystems) and 5 ​ng cDNA per reaction. Samples were run in technical duplicates or triplicates on a StepOnePlus Real-Time PCR System (Applied Biosystems) or QuantStudio5 Real-Time PCR System (Applied Biosystems). Target genes were normalized to Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH), and the Comparative Ct method (2^−ΔΔCt) was used to calculate the Log2 fold change in gene expression relative to iPSCs. Primer sequences for GAPDH, pluripotency genes POU class 5 homeobox 1 (OCT4) and Nanog homeobox, and cartilage related genes SRY box transcription factor 9, Aggrecan (ACAN), and Collagen (COL) type II alpha 1 chain (COL2A1), -type I alpha 1 chain (COL1A1), -type X alpha 1 chain (COL10A1) are included in Supplementary File 2, Table S1.

2.5. Glycosaminoglycan and DNA quantification of chondrogenic pellets

GAG synthesis was quantified by a DMMB assay as previously described [44]. DNA content was measured by Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen) according to the manufacturer's instruction. Samples were analysed in technical triplicates.

2.6. Histology

28-day chondrogenic pellets were fixed in 10 ​% Neutral Buffer Formalin, embedded in paraffin wax, sectioned at 5 ​μm thickness, and rehydrated as previously described [44]. Cartilage proteoglycans were visualised by staining with 0.1 ​% Safranin O and Fast Green FCF. Immunofluorescence staining was performed by incubating sections in Tris-EDTA buffer (10 ​mM Tris, 1 ​mM EDTA, 0.05 ​% Tween 20, pH 9) for 20 ​min at 95 ​°C followed by 25 ​μg/ml Proteinase K (Sigma) for 10 ​min at 37 ​°C. The sections were blocked in 10 ​% normal goat serum (Cell Signalling Technology) for 1 ​h at room temperature. Sections were incubated with primary antibodies from Abcam and Invitrogen against Collagen Type I (COL I, ab138492, 1:500), COL II (ab185430, 1:100) COL X (ab58632, 1:100), ACAN (MA3-16888, 1:50), and Lubricin (PRG4, ab28484, 1:200) overnight at 4 ​°C, followed by Alexa Fluor 488 and 555 conjugated secondary antibodies from Cell Signalling Technology (4412 and 4409 respectively, 1:1000) for 1 ​h at RT, and mounted using Fluoroshield mounting medium with DAPI (Abcam). The sections were imaged with an Olympus FV3000 confocal laser scanning microscope.

2.7. Screening assay

Following a previous study with modifications [45], a small molecule screening platform was developed using a JANUS automated liquid handling workstation (PerkinElmer), during the first two weeks of chondrogenesis in pellet culture, measuring GAG synthesis as the readout method. Small molecule compounds were prepared in dimethyl sulfoxide (DMSO) and included in the assay at 3 final concentrations (Supplementary File 2, Table S2), based on the literature included in Table 1. On day −1, P2 iCPs and P2 BM-MSCs were pelleted in 96-well v-bottom plates (Sarstedt) at 0.5 ​× ​10⁵ ​cells/well/100 ​μl ICM, and P4 iNCC-MSCs at 1 ​× ​10⁵ ​cells/well/100 ​μl ICM. The plates were incubated overnight at 37 ​°C, 5 ​% CO₂, 2 ​% O₂. On day 0 chondrogenesis was induced by adding 100 ​μl ICM containing growth factors and small molecules, growth factors and DMSO (baseline control), or DMSO alone (negative control, 0.1 ​%–0.2 ​% DMSO/well), to achieve the desired final concentrations of small molecules and growth factors (10 ​ng/ml TGFβ3 for iCPs and BM-MSCs, or 10 ​ng/ml TGFβ3, 50 ​ng/ml BMP4 and 40 ​ng/ml PDGF BB for iNCC-MSCs). The outer wells were filled with ICM and not included in the assay. The plates were cultured under hypoxic conditions at 37 ​°C, 5 ​% CO₂, 2 ​% O₂ for 14 days. Half media changes were performed on day 2, 5, 8 and 11. On day 14 the assay was ended and a DMMB assay was performed to measure GAG synthesis. The screening experiment was repeated three times to obtain N ​= ​3 biological replicates per cell line.

2.8. Quantification of glycosaminoglycans after 14-day small molecule screening assay

The pellets were digested in 100 ​μl of media containing 100 ​μg/ml papain (Sigma) per well. The plates were sealed with adhesive covers (Applied Biosystems) and incubated at 60 ​°C for 4 ​h. After digestion the plates were left to cool and briefly spun down. A DMMB assay was run in 384-well microplates (Greiner), with 25 ​μl standard ​+ ​75 ​μl DMMB stock solution in triplicate wells per standard, and 15 ​μl pellet digest ​+ ​10 ​μl dilution buffer ​+ ​75 ​μl DMMB stock solution in duplicate wells per sample. The plates were read with a PerkinElmer Victor X3 plate reader at 595 ​nm. GAG concentration from the pellet digest was calculated from the standard curve and plotted as μg GAG/pellet digest. For BM-MSCs, after digestion the plates were diluted 1:1 with dilution buffer and then used for the DMMB assay, for the samples to be read on the standard curve.

2.9. Statistical analysis

GraphPad prism software version 10 was used for statistical analysis, including one-way ANOVA with Bonferroni's multiple comparisons test, and Welch's unpaired two-tailed t-test, where indicated.

The quality of the screening platform was assessed by calculating the Z-factor (Z’) using the formula: $Z^{\prime }=1-\frac{\left({3SD}_{+}+{3SD}_{-}\right)}{\left|{Ave}_{+}-{Ave}_{-}\right|}$

3. Results

3.1. Chondrogenic potential of iPSCs

ICPs and iNCC-MSCs were generated following directed differentiation protocols. Both cell types were confirmed to have chondrogenic potential, producing cartilage proteoglycans and ECM proteins, and cartilage related genes were upregulated (Fig. 2). The iNCC-MSC derived chondrogenic pellets were observably larger in size, with high GAG/DNA ratios (38.99±3.14 ​μg GAG/μg DNA, Fig. 2b). COL X protein expression was observed in both cell types (Fig. 2c), and COL10A1 gene expression was significantly upregulated in iNCC-MSC derived chondrogenic pellets (Log2 Fold Change relative to iPSCs ​= ​12.4±1.79, P ​= ​0.0184, Fig. 2e), suggesting the chondrocytes are hypertrophic, and the iCP derived chondrogenic pellets highly expressed COL I (Fig. 2 c and d. Log2 Fold Change relative to iPSCs ​= ​5.3±0.14, P ​= ​0.0005), a marker of fibrous cartilage [46]. The iNCC-MSCs were also characterised for expression of mesenchymal surface markers, and their adipogenic and osteogenic potential (data not shown).

Fig. 2.

iCP and iNCC-MSC Chondrogenic Differentiation. (a) Representative images of Safranin O/Fast Green FCF stained chondrogenic and negative control pellet sections. Scale bar ​= ​500 ​μm. Insets displayed are magnified images of the areas indicated by the arrowheads. (b) GAG quantification relative to DNA of tested chondrogenic pellets. (c) Representative images of immunofluorescence staining for cartilage matrix proteins. Images are 4 and 9 20X images stitched together for iCP and iNCC-MSC derived pellets respectively. iCP scale bar ​= ​200 ​μm, iNCC-MSC scale bar ​= ​500 ​μm. (d) and (e) mRNA expression of pluripotency and cartilage related genes in tested iCP and iNCC-MSC derived chondrogenic pellets relative to undifferentiated iPSCs. P-values indicate statistical significance by Welch's unpaired two-tailed t-test. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.2. Efficiency of the screening platform

Pre-screening experiments confirmed the assay was running efficiently (Fig. 3), by calculating the Z′, a measurement of assay quality [47]. In this platform, a Z’ > 0.5 was considered an excellent assay, and a Z’ > 0.3 was moderate. The Z′ results confirmed the assay developed can identify molecules enhancing GAG synthesis in BM-MSCs and iCPs, at 0.5 ​× ​10⁵ ​cells/well (Fig. 3b). The culture conditions for iNCC-MSCs require optimisation as an acceptable Z’ could not be obtained (Fig. 3b). GAG synthesis could not be detected in iNCC-MSCs seeded at 0.5 ​× ​10⁵ ​cells/well or 1 ​× ​10⁵ ​cells/well, and cultured in the presence of PDGF BB, TGFβ3, and BMP4 over 14 days (Fig. 3c). Nevertheless, the iNCC-MSCs were still included in the screening platform, to determine whether any small molecule compound could enhance GAG synthesis.

Fig. 3.

Pre-screening experiment using the JANUS liquid handling workstation. 14 day chondrogenesis of BM-MSCs, iCPs, and iNCC-MSCs (a) Chondroitin 6 sulphate standard curve of the DMMB assay run in 384-well plates. (b) Quantification of GAG (μg) per pellet digest, seeding density 0.5 ​× ​10⁵ ​cells/well. (c) Quantification of GAG (μg) per pellet digest after 14 days of chondrogenesis of iNCC-MSCs at different seeding densities, induced with 40 ​ng/ml PDGF BB ​+ ​10 ​ng/ml TGFβ3+ 50 ​ng/ml BMP4.

3.3. Identification of hits after 14 ​Days of chondrogenesis

The 14-day small molecule screening platform included BM-MSCs, iCPs, and iNCC-MSCs, cultured as pellets at 2 ​% O₂ in the presence of chondrogenic inducing growth factors (Fig. 4a and b). To determine significant GAG production, all samples were compared against the baseline and negative controls using a one-way ANOVA. Statistical significance is only marked in samples where the GAG content was greater than the baseline control.

Fig. 4.

Small Molecule Screening Platform. (a) 96-well plate layout of small molecule compounds and their final concentrations. The final concentrations of growth factors per cell line, and the conditions for the baseline control and negative control are described in bullet points. (b) Simplified overview of the small molecule screening experiment, Created with BioRender.com.

No small molecule enhanced GAG synthesis of BM-MSCs any further than the baseline control (12.64±0.09 ​μg GAG/pellet digest). Several molecules, particularly at higher concentrations, had negative effects where GAG synthesis was inhibited or reduced (Fig. 5a). In the iCPs, all three concentrations of forskolin (50 ​μM, 25 ​μM, and 10 ​μM) significantly enhanced GAG synthesis compared to the baseline control (50 ​μM forskolin: 6.68±0.38 ​μg GAG/pellet digest P ​= ​0.0053, 25 ​μM forskolin: 6.52±0.29 ​μg GAG/pellet digest, P ​= ​0.0092, 10 ​μM forskolin: 7.2±0.1 ​μg GAG/pellet digest, P ​= ​0.0009, versus the baseline control: 2.22±0.25 ​μg GAG/pellet digest, Fig. 5b). GAG quantities in the wells cultured with 10 ​μM sesamin, and 25 ​μM and 10 ​μM baicalin were over double the amount of the baseline control (4.79±0.41, 4.52±0.39, and 5.06±0.67 ​μg GAG/pellet digest respectively, Fig. 5b). From this analysis, forskolin, sesamin, and baicalin were regarded as hits. Similar to the BM-MSCs, a number of small molecules inhibited or reduced GAG production in the iCPs. Forskolin promoted GAG synthesis in the iNCC-MSCs with the highest GAG quantities produced at 50 ​μM forskolin (2.91±0.59 ​μg GAG/pellet digest, P ​< ​0.0001, Fig. 5c).

Fig. 5.

Small molecule screening platform results. GAG quantification (μg) per pellet digest of (a) BM-MSCs, (b) iCPs, and (c) iNCC-MSCs after 14 days of chondrogenesis. P-values indicate statistical significance by one way ANOVA and Bonferroni's multiple comparison test.

4. Discussion

A small molecule screening platform was developed and successfully identified hits in iPSC derived cells. Forskolin, sesamin, and baicalin enhanced GAG synthesis in iCPs. Forskolin promoted GAG synthesis in iNCC-MSCs, in the presence of chondrogenic growth factors. However, the iNCC-MSC screening platform proved suboptimal, as GAG synthesis was not detectable after 14 days with growth factors alone. Consequently, the assay cannot reliably identify hit compounds using iNCC-MSCs, and no conclusions can be drawn regarding potential hits. The assay did not identify hits in BM-MSCs, and none of the small molecules enhanced GAG synthesis compared to TGFβ3 alone for this cell type. The culture conditions for iNCC-MSCs could be further optimized to provide additional improvements for the screening assay. For example, iNCC-MSC derived chondrogenic pellets may be cultured with growth factors alone for at least 10 days to induce chondrogenic commitment, before commencing small molecule treatment. However, this approach would increase costs and lengthen the assay time. In addition, the three compounds identified as hits in iCPs could enhance GAG synthesis at 10 ​μM concentrations, therefore future iPSC-based screening of naturally derived compound libraries could be run at a single concentration of 10 ​μM to reduce costs and complexity.

Many of the molecules included in the screening platform were studied in vitro without additional chondrogenic growth factors such as TGFβ3 or BMP4. Treating the cells with a combination of small molecule compounds and growth factors in 3D and hypoxic conditions may have a negative effect, particularly at higher concentrations. For example, SM04690, Ellipticine, Fluvastatin, and 17-DMAG were all reported to promote chondrogenesis in primary cells without the use of chondrogenic inducing growth factors [18,19,26,33]. Whereas in this study, treating the cells with these small molecules in the presence of TGFβ3 impaired GAG synthesis. TD-198946 and TGFβ3 enhanced chondrogenesis of human synovial stromal cells compared to TGFβ3 alone, but only at specific concentrations of 1 ​nM or less [48], and TD-198946 enhanced chondrogenesis of iNCC-MSCs when it was added on day 10 of chondrogenic differentiation [49]. Small molecule compounds can exhibit cytotoxicity and off-target effects. The inclusion of several compounds in the screening platform led to pronounced inhibition of GAG synthesis, suggesting these compounds may impair cell viability rather than modulate the target pathway. Small molecule cytotoxicity should be evaluated in future screening platforms. The observed results echo previous reports, that small molecule activity and their effects are dependent on the culture conditions [50], including the cell type, molecule concentration, and length of treatment.

There is growing interest to explore the pharmaceutical properties of natural products for clinical therapies [45] and the three compounds indicated as hits in this study, Forskolin, sesamin, and baicalin, are all naturally derived small molecule compounds isolated from medicinal plants. Forskolin is isolated from the Coleus Forskohlii plant and is known as an adenylyl cyclase activator. Forskolin enhanced chondrogenesis of hPSC derived ectomesenchymal cells and suppressed COL10A1 expression in hPSC derived cells and nasal chondrocytes, reducing hypertrophy [37]. Continuous exposure of forskolin in BM-MSCs inhibited chondrogenesis, while intermittent exposure enhanced differentiation [36]. These findings align with our observed results in the screening platform and demonstrate how the small molecule effects are context-dependant. Sesamin, found in Sesamin indicum (sesame) seeds and oil, promoted chondrogenesis in amniotic fluid derived-MSCs [21], BM-MSCs, and fracture healing in a mouse osteoporotic model [22]. A mechanism of action was suggested through activation of BMP2 and estrogen receptor alpha (ERα) signalling [22]. Baicalin, isolated mainly from the Scutellaria plant species, promoted ECM synthesis in mouse chondrocytes and ATDC5 cells through activation of hypoxia inducible factor 1 subunit alpha (HIF-1α) and TGFβ/Smad signalling pathways, and demonstrated chondroprotective effects [38,39].

The pro-chondrogenic effects of forskolin, sesamin, and baicalin will need to be validated in additional cell lines, over longer culture periods, and expression of cartilage ECM components should be assessed by qPCR and immunofluorescence staining. It would be recommended to investigate if combinations of these small molecules can induce or rescue chondrogenesis in the absence of growth factors and enhance the production of hyaline-like articular cartilage without hypertrophic differentiation. Promoting chondrogenesis with small molecule compounds instead of growth factors could reduce the overall cost of running differentiation experiments, as purchasing growth factors are considerably more expensive. In vivo studies are warranted to evaluate the pharmaceutical applications of forskolin, sesamin and baicalin for cartilage repair, and assess their toxic effects. Systemic administration, direct delivery to the joint, and pre-treating cell-based implants with these compounds to enhance ECM quality, should be compared.

To further optimise the screening platform, including a fluorescent reporter line could allow for more high throughput analysis. For example, COL2, SOX9, and COL10 fluorescent reporter lines have been utilised in high throughput screening assays [15,17,18]. Strong expression of the hypertrophic marker COL X was observed in iCP and iNCC-MSC derived chondrogenic pellets from this study. Hypertrophic chondrocytes play a role in endochondral ossification [46], and their therapeutic potential for the long-term repair of articular cartilage is limited in contrast to articular chondrocytes that produce stable hyaline cartilage. Further assessment of matrix quality could include developing a dual fluorescent reporter line to measure and compare COL2 and COL10 expression profiles across treatment groups, identifying molecules enhancing hyaline cartilage production and inhibiting hypertrophy in the same platform. Leading to the development of more relevant disease modifying drugs and cell therapies.

In conclusion, this study demonstrated that iPSC-based small molecule screening platforms can successfully identify potential DMOADs, and the small molecule compounds identified in this study may contribute toward enhanced derivation of articular cartilage from iPSCs.

Author contributions

AOB, MX, GS, AMM, MM and FB contributed toward the conception and design of the study, and interpretation of the data. EOC contributed toward the screening experiment design and provided technical support. AOB carried out experimental work, data collection and analysis. MX characterised the iPSCs. AM assisted in the characterisation of the iNCC-MSCs. JRD provided the iPSCs. AOB and FB wrote, edited, and critically reviewed the manuscript. All authors contributed to the article and approved the submitted version.

Role of the funding source

This work was funded by the Celtic Advanced Life Science Innovation Network, an Ireland Wales programme part funded by the European Regional Development Fund through the Welsh Government (Grant no: 80885). The funders were not involved in the design of the study, data collection, data analysis, interpretation of the data, or in the writing of the manuscript.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Frank Barry reports a relationship with Orbsen Therapeutics Ltd that includes: board membership and equity or stocks. Frank Barry reports a relationship with BioRecell that includes: consulting or advisory. Frank Barry reports a relationship with University Health Network Schroeder Arthritis Institute that includes: consulting or advisory. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Handling Editor: Professor H Madry

Appendix A. Supplementary data

The following are the Supplementary data to this article: