Letter to the Editor: Advancing induced pluripotent stem cell-derived mesenchymal stem cells - insights from comparative differentiation protocols and emerging innovations

Abstract

Induced pluripotent stem cell-derived mesenchymal stem cells (iMSCs) depict a transformative tool in regenerative medicine, offering scalable, patient-specific alternatives to primary mesenchymal stem cells. The recent study by Ababneh et al published in the World Journal of Stem Cells provides a comprehensive comparison of five differentiation protocols, highlighting protocol-dependent variations in iMSC functionality. This article discusses the key findings, their implications for standardization, and recent developments in iMSC generation, including artificial intelligence-guided protocols and clinical advancements including NCT02943829 and NCT03839238 trials. These insights underscore the need for optimized methods to enhance therapeutic efficacy.

Key Words: Induced pluripotent stem cells; Mesenchymal stem cells; Differentiation protocols; Induced pluripotent stem cell-derived mesenchymal stem cells; Regenerative medicine

Core Tip: The study by Ababneh et al underscores the impact of differentiation protocols on induced pluripotent stem cell-derived mesenchymal stem cell (iMSC) functionality, revealing variations in osteogenesis, adipogenesis, proliferation, and migration. Recent advancements, including artificial intelligence-guided optimization and CRISPR-edited iMSCs, enhance standardization and safety. Future research should address donor diversity, scalability, and engraftment challenges through multi-donor studies, automated bioreactors, and advanced gene editing to utilize therapeutic potential of iMSCs in regenerative medicine.

TO THE EDITOR

Mesenchymal stem cells (MSCs) are pivotal in regenerative medicine due to their multipotency, immunomodulatory properties, and role in tissue repair[1,2]. However, limitations such as invasive sourcing, donor variability, and limited expansion potential hinder their clinical utility[3]. Induced pluripotent stem cells (iPSCs), reprogrammed from somatic cells, offer an unlimited source for generating iPSC-derived MSCs (iMSCs), which retain MSC characteristics while overcoming these constraints[4,5]. We read with great interest the study by Ababneh et al[6], published in the World Journal of Stem Cells, which evaluates five distinct differentiation protocols for iMSC generation, providing critical insights into how methodological variations influence phenotypic and functional outcomes. This work aligns with ongoing efforts to standardize iMSC production for reproducible therapeutic applications[7].

Ababneh et al[6] conducted a rigorous comparative analysis of iMSCs derived from four iPSC lines using five protocols: SB431542-based (iMSC1 and iMSC3), iMatrix-free (iMSC2), growth factor (GF)-supplemented (iMSC4), and embryoid body (EB) formation with retinoic acid (EB-iMSC). All iMSCs met the 2006 International Society for Cell & Gene Therapy (ISCT) criteria, including plastic adherence, expression of CD105, CD73, CD90, and CD44, and absence of pluripotency markers (NANOG, TRA-1-60) and hematopoietic markers[8].

Functional assessments revealed protocol-specific differences. EB-iMSC and iMSC2 showed enhanced osteogenesis (elevated runt-related transcription factor 2 expression; P ≤ 0.01 and P ≤ 0.0001, respectively), while adipogenesis was reduced in iMSC2 (adipsin; P ≤ 0.01) and EB-iMSC (adipsin and peroxisome proliferator-activated receptor gamma; P ≤ 0.0001 and P ≤ 0.01) compared to bone marrow MSCs. MSC4 exhibited superior proliferation (MTT assay; P values: 0.01-0.001), with most iMSCs showing enhanced colony-forming units over bone marrow MSCs (e.g., iMSC3: P ≤ 0.0001). iMSC3 and iMSC4 had reduced mitochondrial membrane potential (P ≤ 0.001), but all iMSCs displayed lower reactive oxygen species levels (P values: 0.001-0.0001) and senescence (P values: 0.01-0.0001, except iMSC1). Migration was impaired in iMSC4 (P ≤ 0.001 at 24 hours; P ≤ 0.0001 at 48 hours). These variations emphasize that while all protocols yield functional iMSCs, selection should align with intended applications - e.g., EB-iMSC for osteogenesis in bone repair.

Implications and significance

The findings of Ababneh et al[6] highlight the heterogeneity in iMSC quality arising from differentiation strategies, echoing prior studies on protocol-dependent outcomes[9,10]. For instance, EB-based methods enhance intercellular signaling but may limit adipogenesis due to retinoic acid’s inhibitory effects[6,11]. This underscores the need for multi-parameter evaluations beyond ISCT criteria, incorporating functional potency assays as per the updated 2025 ISCT guidelines[12].

Clinically, optimized iMSCs could address limitations of primary MSCs, such as senescence and variability[13]. The study’s use of unified culture conditions strengthens its validity, providing a benchmark for future protocol refinements[1]. However, limitations include reliance on two donor-derived iPSC lines, suggesting the need for broader donor diversity to capture inter-individual variability[6,14]. The limited donor diversity may underestimate inter-individual variability in iMSC differentiation, which is critical for clinical batch consistency. A key limitation of comparative protocol studies that use only a small number of iPSC donors is that they may underestimate inter-individual variability arising from donor genetics and epigenetic memory. Limited donor diversity can lead to inflated estimates of reproducibility and may mask interactions between iPSC origin and differentiation conditions that affect potency or safety. Future comparative studies should adopt multi-donor designs or consortia-based sample sharing to ensure clinical generalizability[15,16].

Recent developments

Since the publication of Ababneh et al[6], advancements in iMSC differentiation have accelerated, driven by precision engineering and clinical translation. Sequential GF protocols, building on iMSC4-like approaches, have improved efficiency: For example, Tsukamoto et al[17] reported a canine iPSC-to-MSC pipeline using GFs (e.g., fibroblast GF 2, bone morphogenetic protein 4) for enhanced mesodermal induction, achieving > 90% MSC marker expression in suspension cultures.

Artificial intelligence (AI)-guided differentiation represents a breakthrough: Cerneckis et al[5] integrated machine learning to optimize protocols, predicting optimal GF combinations for iMSC yield, reducing variability by 30% compared to traditional methods. CRISPR/Cas9 editing has also enhanced iMSC safety; a 2025 review by Zhao and Liu[18] details gene-edited iMSCs with knocked-out tumorigenic risks, facilitating Good Manufacturing Practice (GMP)-compliant production.

Clinical trials have surged: As of December 2024, 115 human pluripotent stem cell-derived trials are registered, with 28 involving iMSCs[19]. Some prominent examples include NCT02943829 trial that achieved 53% complete response by day 100 in steroid-refractory acute graft-vs-host disease (8/15 patients). The overall response was 87%[20]. The NCT03839238 trial Used MAG200 (iMSC-like allogeneic MSCs) and achieved 40% Western Ontario and McMaster Universities Osteoarthritis Index pain reduction at 12 months[21]. Food and Drug Administration approvals in 2023-2025 include iPSC-MSC therapies for retinal degeneration, with iMSCs differentiating into retinal cells in heterogeneous populations[22]. Human iMSCs have shown promise in skin regeneration, promoting wound healing via paracrine effects[23]. Table 1 summarizes select recent protocols, illustrating innovations beyond those in Ababneh et al[6]. Based on Ababneh et al[6] and 2025 ISCT guidelines, we propose a three-tier potency assay in Table 2.

Table 1 Summary of recent induced pluripotent stem cell-to-mesenchymal stem cells differentiation protocols (2024-2025).

Protocol type	Key features	Advantages	Limitations	Proposed solution	Ref.
Sequential GF induction	Stepwise use of FGF2, BMP4, PDGF, EGF; 2-3 weeks duration	High yield (> 85% MSCs); scalable	Donor variability; GF cost	Use AI-optimised minimal GF cocktails and pooled GMP-grade GF suppliers	[18]
AI-guided optimization	Machine learning predicts media/GF combos; integrates CRISPR	Reduced heterogeneity; 20%-30% efficiency gain	Requires computational resources	Share trained models and datasets via consortia to lower barrier and validate generalisability	[5]
Suspension-based	ROCK inhibitor + StemMACS media; feeder-free	GMP-compatible; stimulates angiogenesis	Lower adipogenic potential	Combine 3D microcarrier conditions or post-differentiation niche conditioning to restore adipogenic capacity	[24]
Mesodermal-focused	Activin A + CHIR99021 for primitive streak; FGF2 for MSCs	Enhanced skin regeneration applications	Protocol-specific senescence	Combine small-molecule-driven mesodermal induction (e.g., CHIR99021/Wnt activation) with minimal essential growth factor sets to reduce cost and variability	[25]
Gene-edited hybrid	CRISPR for pluripotency knockdown; GF supplementation	Tumor-safe; tailored potency	Ethical/regulatory hurdles	Adopt a Quality-by-Design framework for gene-edited iMSCs, including off-target genome analysis, transcriptomic comparability to parental iMSCs, and long-term culture stability testing	[6]

GF: Growth factor; FGF2: Fibroblast growth factor 2; BMP4: Bone morphogenetic protein 4; PDGF: Platelet-derived growth factor; EGF: Epidermal growth factor; MSC: Mesenchymal stem cell; AI: Artificial intelligence; GMP: Good Manufacturing Practice; iMSCs: Induced pluripotent stem cell-derived mesenchymal stem cells.

Table 2 A three-tier potency assay.

Tier	Assay category	Specific metrics (thresholds)
Phenotypic core	Surface + lineage transcription	CD73/CD90/CD105 ≥ 95%
		RUNX2 ≥ 2.5-fold vs BM-MSC (qPCR)
		PPARG ≤ 0.5-fold
Functional potency	Proliferation, migration, stress resilience	CFU-F ≥ 50/103 cells
		Transwell migration index ≥ 0.7 (24 hours)
		ROS ≤ 60% BM-MSC baseline (DCFDA)
Translational predictors	3D microenvironment + AI modeling	3D spheroid invasion ≥ 70% (72 hours, Matrigel)
		AI-predicted batch consistency (R2 ≥ 0.85 via scRNA-seq input)

RUNX2: Runt-related transcription factor 2; BM-MSC: Bone marrow mesenchymal stem cell; qPCR: Quantitative polymerase chain reaction; PPARG: Peroxisome proliferator-activated receptor gamma; CFU-F: Colony forming units-fibroblast; ROS: Reactive oxygen species; DCFDA: 2’,7’-dichlorofluorescin diacetate; AI: Artificial intelligence; scRNA-seq: Single-cell RNA sequencing.

Future directions

Despite the progress highlighted in Ababneh et al[6] and recent advancements, several research gaps persist in the field of iMSCs, hindering full clinical translation. One major gap is the limited donor diversity in comparative studies, as seen in Ababneh et al[6], which relied on only two donor-derived iPSC lines, potentially overlooking inter-individual epigenetic and genetic variability that affects differentiation efficiency and functionality. This heterogeneity contributes to inconsistencies in MSC batches and iPSC reprogramming, leading to epigenetic instability and tumorigenic risks such as teratoma formation. Additionally, current protocols often exhibit inefficient differentiation, with iMSCs sometimes aberrantly forming fibroblasts instead of targeted lineages in disease microenvironments, exacerbating issues like fibrosis in liver injury models. Translational challenges include low engraftment rates (< 10% retention post-transplantation) due to shear stress and cell death, as well as scalability issues in GMP-compliant manufacturing, where processes struggle to meet clinical trial demands. Furthermore, organoid models derived from iMSCs lack vascularization and immune components, limiting their predictive value for in vivo outcomes. Safety concerns, including genomic mutations and reactivation of pluripotency genes, remain prominent, alongside a lack of standardized potency assays aligned with the 2025 ISCT guidelines.

Key translational gaps remain in scalability like GMP-compatible expansion or bioreactor workflows, robust potency assays, and long-term safety monitoring. Recent advances illustrate pathways to address these gaps like automated/robotic platforms, AI-guided optimization, and targeted gene engineering (CRISPR). By integrating these approaches in a Quality-by-Design pipeline may accelerate iMSC translation while meeting regulatory expectations.

To address these gaps, future research should prioritize multi-donor studies using diverse iPSC lines to capture biological variability and enhance reproducibility, potentially through international consortia for shared biobanks. Integration of CRISPR/Cas9 for precise gene editing could mitigate safety risks by correcting epigenetic instabilities and knocking out tumorigenic factors, while minimizing off-target effects through advanced delivery systems. Scalability can be improved via automated bioreactors and Quality-by-Design frameworks, incorporating process analytical technology for real-time monitoring of critical quality attributes during expansion and differentiation. AI and machine learning should be leveraged to predict differentiation pathways from single-cell sequencing data, optimizing protocols and reducing heterogeneity by 20%-30% as demonstrated in recent models. For translational gaps, expanded preclinical in vivo studies in large animal models (e.g., rhesus macaques) and bioengineered scaffolds could enhance engraftment and microenvironment integration, paving the way for more phase 3 trials like those for iMSC-based osteoarthritis treatments. Ultimately, interdisciplinary approaches combining 3D bioprinting, organoids, and personalized medicine will standardize iMSC production, accelerating their application in regenerative therapies in future.

Conclusion

Ababneh et al[6] provides a foundational comparison of iMSC protocols, revealing the critical role of methodological choices in functionality. Recent innovations, including AI and gene editing, promise to standardize and enhance iMSCs for clinical use. Future research should integrate these with multi-donor studies to accelerate translation, ultimately realizing iMSCs’ potential in regenerative medicine.