Maximizing stem cell yield from limited adipose tissue: A novel two-step digestion approach and future artificial intelligence-integrated perspectives

Abstract

Adipose tissue-derived stromal vascular fraction (SVF) is a valuable source of regenerative cells for various clinical applications. However, obtaining a sufficient number of cells from patients with limited adipose reserves, particularly from the pediatric population, is a challenge. A study investigated a two-step enzymatic digestion approach to maximize stem cell yield from small adipose tissue samples. Their work demonstrated that the second collagenase digestion of the residual adipose tissue (typically discarded after conventional isolation) yielded a considerable additional population of viable regenerative cells (SVF2). Although SVF1 contained higher absolute cell numbers, SVF2 exhibited superior plating efficiency and higher colony-forming units per 1000 mononucleated cells. This simple modification substantially improved the regenerative cell yield from limited adipose tissue sources. This review offers a critical evaluation of the study methodology and propose future directions, including the integration of artificial intelligence to optimize digestion parameters, and the establishment of standardized potency assays to facilitate its translation into broader clinical practice.

Key Words: Stromal vascular fraction; Adipose-derived stem cells; Enzymatic digestion; Cell yield; Regenerative medicine; Pediatric applications; Artificial intelligence

Core Tip: The residual adipose tissue typically discarded after conventional enzymatic digestion contains viable and functionally-potent regenerative cells. The second digestion step yielded stromal vascular fraction 2, which exhibited high clonogenic potential. Although promising, the study’s generalizability is limited by its small sample size and lack of donor diversity. This review critically evaluates the methodology, discusses the impact of donor variability and enzyme lots on reproducibility, and proposes an artificial-intelligence-driven framework for personalizing digestion protocols. These insights aim to accelerate the clinical adoption and optimization of this valuable technique.

INTRODUCTION

Adipose tissue has emerged as the preferred source of regenerative cells for therapeutic applications due to its abundance, accessibility, and abundance of multipotent stromal cells[1-4]. The stromal vascular fraction (SVF), a heterogeneous cell population isolated from the adipose tissue, contains adipose-derived stem cells, endothelial progenitors, pericytes, and immunomodulatory cells that collectively contribute to tissue regeneration, angiogenesis, and anti-inflammatory effects[5,6]. The clinical applications of SVF have expanded across diverse fields, including wound healing, cosmetic surgery, orthopedic repair, and the treatment of degenerative diseases[7,8]. However, a persistent challenge is the efficient isolation of sufficient regenerative cells from limited tissue volumes, particularly in pediatric patients or cachectic individuals[9].

The original article by Mushtaq et al[10] addresses a critical gap by investigating whether residual adipose tissue, conventionally discarded after the first enzymatic digestion, contains viable cells that can be harvested through a second digestion. The authors provided compelling evidence that this simple modification substantially enhanced total stem cell yield. We appreciate the opportunity to offer a critical perspective on this innovative work, discussing both its strengths and methodological limitations, and proposing future avenues for optimization and clinical translation.

METHODOLOGICAL SUMMARY AND STRENGTHS

This study used a sequential two-step collagenase type IV digestion of adipose tissue samples from healthy donors. Following the standard first digestion to obtain SVF1, the remaining undigested tissue was subjected to a second digestion to yield SVF2. The key findings are noteworthy. Both SVF1 and SVF2 exhibited high cell viability (> 95%). SVF2 contributed to 38%-40% of the total cell yield, representing a clinically-meaningful increase. SVF2 demonstrated superior functional properties, with a substantially higher number of colony-forming units per 1000 mononucleated cells, suggesting enrichment of clonogenic progenitors. Flow cytometry confirmed the presence of mesenchymal stem cell markers (CD44, CD73, and CD146) with comparable expression profiles[11]. Although not statistically significant, the expression of CD73 and CD146 tended to be higher in SVF2 cells, indicating their enrichment in pericyte-like cells with enhanced angiogenic and regenerative potential[12]. A detailed comparison of the two methods is presented in Table 1, which highlights the additional yield, improved plating efficiency, and higher clonogenic potential of the two-step approach while noting comparable viability and surface marker expression[13-23].

Table 1 Comparison of conventional single-step and two-step digestion methods for adipose tissue-derived stromal vascular fraction isolation.

Feature	Conventional single-step digestion (SVF1 only)	Two-step digestion (SVF1 + SVF2)
Total cell yield (per sample)	1.09 × 106 ± 2.07 × 105 cells	Additional 38%-40% increase per sample (SVF2 contributes 4.19 × 105 ± 1.56 × 105 cells)
Cell viability	97.22% ± 1.88%	96.84% ± 3.39% (not statistically different, P = 0.83)
Total mononuclear cells	4.11 × 105 ± 1.62 × 105 cells	SVF2 adds 1.99 × 105 ± 1.05 × 105 cells (P = 0.03)
Mononuclear cell percentage	38.71% ± 16.11%	45.46% ± 13.54% (not significantly different, P = 0.49)
Plating efficiency	0.29 ± 0.11	0.72 ± 0.34 (P = 0.03) - higher clonogenic capacity per seeded cell
CFU per 1000 mononuclear cells	8.23 ± 3.47	15.75 ± 5.50 (P = 0.03) - higher progenitor density
CD44+ cells	42.85% ± 23.21%	32.38% ± 23.55% (comparable, no significant difference)
CD73+ cells	10.86% ± 5.79%	11.41% ± 9.13% (comparable)
CD146+ cells	29.09% ± 16.23%	27.68% ± 15.21% (comparable)
Gene expression (VEGF, SDF, NOS3, CD14, MMP3, IL4R, P16, P53)	Reference expression levels	Similar transcriptional profiles; no statistically significant differences (P > 0.05)
Histological findings	After first digestion: Partial ECM degradation, moderate adipocyte disruption, some cells still embedded	After second digestion: Substantial structural breakdown, complete loss of adipocyte integrity, release of tightly bound regenerative cells
Clinical relevance	Discards residual tissue; may yield insufficient cells for pediatric or low-adipose-reserve patients	Maximizes cell harvest from small samples; particularly valuable for pediatric and low-reserve patients

SVF: Stromal vascular fraction; CFU: Colony-forming unit; VEGF: Vascular endothelial growth factor; SDF: Stromal cell-derived factor; MMP3: Matrix metalloproteinase 3; IL4R: Interleukin 4 receptor; ECM: Extracellular matrix.

Gene expression analysis indicated similar transcriptional profiles between fractions for key functional genes, including vascular endothelial growth factor (angiogenesis), stromal cell-derived factor (homing), NOS3 (mobilization), matrix metalloproteinase 3 (matrix remodeling), interleukin 4 receptor (immunomodulation), and the senescence markers, P16 and P53. The comparable expression of tumor suppressor genes between the fractions was particularly reassuring from a safety perspective, indicating that extended digestion did not induce aberrant proliferative profiles. Histological evaluation indicated a morphological correlation with functional findings, demonstrating progressive tissue dissociation across digestion stages. The undigested adipose tissue showed intact adipocytes, with regenerative cells embedded within the perivascular and interstitial zones. After the first digestion, partial matrix degradation and cellular release were observed, whereas after the second digestion, the tissues exhibited substantial structural breakdown with the release of tightly-bound stromal cells. These findings support the concept that distinct cellular niches within adipose tissue were progressively accessed during enzymatic digestion, with perivascular and matrix-anchored progenitors requiring extensive digestion for effective liberation[24].

This study has several significant clinical implications. In autologous cell therapy, the number of transplanted regenerative cells is often correlated with therapeutic outcome[25]. For pediatric patients or those with limited adipose reserves, the inability to harvest sufficient cells may preclude treatment or necessitate repeated invasive tissue collection. The two-step digestion approach offers a practical solution for maximizing the cell yield from a single tissue sample without increasing donor-site morbidity. Furthermore, the enhanced clonogenic potential of SVF2 suggests that it may be particularly valuable for applications that require rapid cell expansion or robust tissue regeneration[26,27].

A CRITICAL EXAMINATION OF METHODOLOGICAL LIMITATIONS

Despite its strengths, the study by Mushtaq et al[10] had several methodological limitations that warrant consideration and should be addressed in future research. First, the sample size and donor diversity were limited. The experiments were conducted on a small cohort, and although the authors aimed to address pediatric needs, donor samples were obtained primarily from healthy adults. Validation in a larger and more diverse cohort that includes pediatric samples with varied body mass index and tissues from different anatomical depots is essential to establish the generalizability of the two-step protocol, as donor variability is known to substantially affect SVF yield and quality[11].

Second, the study lacked critical optimization of the digestion parameters. The authors selected sequential 30-minute digestions but did not compare it with other collagenase types or enzyme combinations. As summarized in Table 2[13-23], different collagenase types vary substantially in yield, viability, surface marker preservation, and safety profile [e.g., type I offers the highest cell yield but has batch-to-batch variation, while recombinant good manufacturing practice (GMP)-grade collagenase ensures clinical safety at a higher cost]. These differences underscore the necessity of careful enzyme selection and protocol optimization for reproducible SVF isolation. Furthermore, the authors used sequential 30-minutes digestions without comparing them to a single 60-minutes digestion or explore other time points. This raises a fundamental question: Is the benefit derived from the “two-step” process itself, or simply from a longer total digestion time? The study did not clarify whether an intermediate centrifugation step was necessary to recover cells released early or if prolonged single digestion could achieve a similar yield. Furthermore, critical parameters such as collagenase activity can vary substantially between lots. The study did not address how lot-to-lot variability might affect the reproducibility of the SVF2 yield, a crucial consideration for GMP compliance.

Table 2 Comparison of collagenase types, digestion time, advantages, and disadvantages for isolating adipose-derived stem cells from lipoaspirate.

Enzyme type	Typical concentration	Digestion time (37 °C)	Advantages	Disadvantages	Ref.
Collagenase type I	0.075%-0.2% (w/v) or 0.1-0.2 U/mL (activity-based)	30-90 minutes	Highest cell yield; optimal for loose lipoaspirate; rapid digestion; good viability	Batch-to-batch variation; animal origin; risk of over-digestion (> 60-90 minutes)	[13-15]
Collagenase type II	0.1%-0.2% (w/v) or 0.15-0.25 U/mL (activity-based)	45-90 minutes	Effective for fibrous or bloody lipoaspirate	Longer digestion; slightly lower yield compared to type I; viability may decrease (> 60 minutes)	[16,17]
Collagenase type IV	0.1%-0.2% (w/v) or 0.15-0.2 U/mL (activity-based)	45-75 minutes	Preserves surface markers (e.g., CD44, CD73); lower tryptic activity	Lower digestion efficiency (60%-80% yield of type I); longer incubation; higher cost	[18,19]
Collagenase I + dispase/thermolysin	Collagenase I: 0.075% + dispase: 0.1-0.2 U/mL	30-45 minutes	Potential for higher single-cell yield; gentler release from ECM; basis for many GMP-grade blends (e.g., liberase)	Higher cost; slight risk of over-digestion; requires pre-testing for optimal ratio	[20,21]
Recombinant/GMP-grade collagenase	As per manufacturer (activity-based)	40-60 minutes (needs calibration)	Animal-free; consistent activity; high viability; clinically safe; low endotoxin	Expensive; requires pre-testing for optimal time and concentration	[22,23]

ECM: Extracellular matrix; GMP: Good manufacturing practices.

Third, the characterization of SVF2 cells, although thorough, is incomplete. The Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy proposed minimum criteria for defining multipotent mesenchymal stromal cells, which include trilineage (adipogenic, osteogenic, and chondrogenic) differentiation capacity in vitro[28-31]. However, Mushtaq et al[10] did not include these differentiation assays. Demonstrating that SVF2 cells retain this multipotency is crucial to confirm their equivalence or superiority to SVF1 as true stem/progenitor cells.

To aid in selecting the most appropriate enzyme for sequential digestion, Table 3 provides a five-star rating evaluation of different collagenase types based on cell yield, viability, surface marker retention, and clinical safety. This practical guide complements the detailed comparison in Table 2 and can help researchers balance yield, safety, and regulatory requirements when designing GMP-compliant protocols.

Table 3 Comparison of different collagenase types in liposuction-related cell isolation.

Collagenase type	Cell yield	Cell viability	Surface marker retention	Clinical safety	Recommended rating
Collagenase type I	★★★★★	★★★★☆	★★★★	★★	★★★★★
Collagenase type II	★★★☆	★★★	★★★	★★	★★☆
Collagenase type IV	★★★	★★★★	★★★★★	★★	★★★
Collagenase I + dispase/thermolysin	★★★★★	★★★★	★★★★	★★	★★★★
Recombinant/GMP-grade collagenase	★★★★☆	★★★★★	★★★★★	★★★★★	★★★★★

This table uses a five-star rating system to evaluate and compare different performance in each aspect. Higher star ratings indicating better performance. GMP: Good manufacturing practices.

TRANSLATIONAL CHALLENGES AND FUTURE DIRECTIONS: AN AI-INTEGRATED APPROACH

From a translational perspective, the two-step digestion approach aligns with broader efforts to optimize the cell therapy manufacturing processes. Standardization of isolation protocols and establishment of release criteria are essential for clinical advancement[32,33]. The authors’ comprehensive characterization approach, including viability, clonogenicity, surface marker expression, and gene expression profiling, provides a framework for quality assessment that could inform future protocol standardization efforts (Figure 1).

Figure 1 Illustrates a proposed artificial intelligence integrated workflow to personalize and standardize the twostep digestion process. A flowchart showing two parallel paths: Left side - conventional two-step digestion with fixed 30 minutes + 30 minutes; right side - artificial intelligence (AI)-optimized workflow. The AI-optimized path includes data collection from donors (age, body mass index, medical history), tissue imaging (density, color), and enzyme lot characteristics. These data feed into a machine learning model that predicts optimal digestion time for step 1 and step 2, then outputs a personalized protocol. The final step for both paths is cell harvest and quality assessment, with feedback loop to retrain the AI model. BMI: Body mass index; AI: Artificial intelligence.

To advance this technique for routine clinical application, we propose a roadmap centered on personalization and standardization[34,35]. One particularly promising avenue is the integration of artificial intelligence (AI)[36,37]. AI algorithms can be trained on large datasets that incorporate donor variables (age, body mass index, medical history), tissue characteristics (density, color, source depot), and collagenase lot parameters to predict the optimal two-step digestion protocol. This “digital twin” of the digestion process could, for instance, recommend personalized digestion times for each step to maximize viable cell yield for an individual patient, moving beyond the one-size-fits-all, 30-minute + 30-minute model. This concept aligns with broader efforts to introduce intelligent automation into cell therapy manufacturing[33].

Furthermore, establishing robust potency assays that reliably predict the in vivo efficacy of SVF preparations is imperative. Although colony-forming unit assays and marker expression are informative, functional assays, such as in vitro tubule formation for angiogenic potential or T-cell proliferation assays for immunomodulatory capacity, could serve as more meaningful release criteria[38-40]. Finally, clinical trials comparing SVF1 alone vs combined SVF1 + SVF2 for specific indications (e.g., osteoarthritis and wound healing) are needed to provide definitive evidence of the superior therapeutic benefits. The development of closed-system GMP-compliant devices capable of performing sequential digestion would also facilitate clinical adoption[41,42].

CONCLUSION

Mushtaq et al[10] made a strong contribution to the field by introducing a two-step enzymatic digestion process that substantially increased the regenerative cell yield from adipose tissue. Although we applaud this innovation, our critical analysis highlights key areas for further investigation, including validation across diverse populations, optimization of digestion parameters, and comprehensive cellular characterization. By addressing these limitations and embracing forward-looking solutions such as AI-driven protocol personalization, this promising technique can be robustly optimized for clinical applications, ultimately improving outcomes for patients requiring cell-based therapies from limited tissue sources.