Matrix origin determines how perlecan sugars steer stem cell fate

In a laboratory study of human infrapatellar fat pad-derived stem cells, removing heparan sulfate chains from the perlecan protein in cell-derived decellularized extracellular matrix did not produce the same biological effect in every matrix. Heparinase III treatment of matrix deposited by unmodified cells reduced the cells’ later capacity for cartilage formation, whereas the same treatment of matrix deposited by SV40 large T antigen-immortalized cells enhanced cartilage-related gene expression, matrix deposition and pellet growth. Only cells preconditioned on the enzyme-treated immortalized-cell matrix also formed mineralized nodules during osteogenic induction.

The results indicate that the cellular origin of the matrix can determine how perlecan-associated sugars influence stem-cell state and multilineage differentiation. The study focused on adult infrapatellar fat pad-derived stem cells, or IPFSCs, isolated from six de-identified young donors and pooled for the experiments. These cells have an inherent propensity for cartilage formation and can deposit extracellular matrix rich in components of the basement membrane.

The researchers generated two types of cell-derived extracellular matrix: dE from unmodified IPFSCs and s-dE from cells immortalized with the SV40 large T antigen. The dECM produced by the immortalized cells had previously been found to contain higher perlecan expression and to be more effective at supporting chondrogenic differentiation than dECM from unmodified cells. This comparison provided the basis for asking whether the functional contribution of perlecan’s sugar chains was also dependent on the cellular source of the matrix.

Perlecan is a multidomain proteoglycan that forms part of basement membranes and other pericellular matrices. Its N-terminal domain carries heparan sulfate glycosaminoglycan chains, which can store and present growth factors such as FGF-2 and BMP-2. The study used heparinase III, or HPA, to remove these chains selectively from matrix-resident perlecan while leaving the protein’s core intact.

HPA was applied to dECM substrates at 0.04 U/cm² for one hour at 37°C, while control substrates received buffer under the same conditions. Immunofluorescence supported the treatment: signal from the antibody 10E4, which recognizes intact heparan sulfate, declined markedly in treated matrices, while signal from 3G10, which recognizes the stub left after enzymatic cleavage, appeared. Strong perinuclear 10E4 staining remained in the cells, consistent with the enzyme reaching extracellular matrix but not newly synthesized intracellular chains.

Five experimental groups were compared: standard tissue-culture plastic, unmodified-cell matrix, heparinase-treated unmodified-cell matrix, immortalized-cell matrix and heparinase-treated immortalized-cell matrix. The cells were preconditioned on these surfaces for about 10 days before stemness markers and differentiation responses were measured. Most assays were conducted with at least three technical replicates, although the cells were pooled from six donors, so the experiment did not provide six independent donor-level estimates.

Gene expression was assessed by quantitative reverse-transcription PCR, surface markers by flow cytometry, and cartilage, fat and bone differentiation through 21-day induction protocols. Chondrogenic induction used three-dimensional pellets under 5% oxygen, adipogenic and osteogenic assays were conducted on plastic, and the resulting phenotypes were measured with molecular, histological and colorimetric readouts. Untreated dE and s-dE altered the expression of several stemness-associated genes relative to plastic.

Both matrices increased SOX2, NANOG and NES, with larger increases generally seen on s-dE, while MYC, NOV, BMI1 and KLF4 were lower. Heparinase treatment attenuated SOX2, NANOG and NES in both matrix conditions and increased NOV. Some responses were source-specific: MYC rose relative to dE after its chains were removed but fell relative to s-dE in the corresponding comparison, and KLF4 was selectively reduced in the treated immortalized-cell matrix condition. Flow cytometry also produced divergent marker profiles.

Heparinase treatment lowered SSEA4 positivity in the unmodified-cell model, but raised CD90, CD146 and SSEA4 positivity in the immortalized-cell model, which had the highest CD146 and SSEA4 percentages among all groups. CD73 positivity remained near 100% across the conditions. Both untreated dE and s-dE suppressed fat formation relative to plastic, producing lower expression of LPL, FABP4, CEBPA and PPARG and less lipid accumulation by Oil Red O staining.

The measured Oil Red O absorbance was 0.214 ± 0.005 for dE, 0.229 ± 0.004 for s-dE and 0.431 ± 0.008 for plastic. The treated groups accumulated more lipid droplets, with absorbance values of 0.342 ± 0.016 for dE(H) and 0.395 ± 0.007 for s-dE(H), and showed increased expression of all four adipogenic markers. Because this response occurred on both matrices, it appears less dependent on the source of the matrix than the cartilage response was.

The authors suggest that heparan sulfate chains may contribute to the anti-adipogenic influence of perlecan through their ability to sequester growth factors, although the experiment did not directly measure which signals accounted for the effect. Chondrogenic responses were strongly dependent on both matrix source and chain removal. After 21 days in three-dimensional pellet culture, cells preconditioned on untreated dE or s-dE expressed more SOX9, ACAN and COL2A1 than cells expanded on plastic and produced stronger Alcian blue and Safranin O staining, consistent with greater accumulation of sulfated glycosaminoglycans.

The untreated matrices also yielded larger pellets. Removing heparan sulfate from dE reduced SOX9, ACAN and COL2A1 relative to untreated dE and diminished the histological staining. In contrast, s-dE(H) produced the largest pellets and most prominent sulfated-glycosaminoglycan deposition of the matrix conditions.

Its day-21 expression values for SOX9, ACAN and COL2A1 were markedly higher than those of cells preconditioned on untreated s-dE: 11.713 ± 0.164, 14.310 ± 0.891 and 111,695.573 ± 5,567.585, respectively. The corresponding dE(H) values were 6.458 ± 0.075, 6.009 ± 0.166 and 38,001.631 ± 1,339.186. The osteogenic response was even more selective.

Alizarin Red S staining detected mineralized matrix exclusively in cells preconditioned on s-dE(H) after 21 days of induction, with an absorbance of 0.766 ± 0.029, whereas the other groups showed minimal mineralization. The molecular results did not follow a simple pattern: RUNX2, SP7, BGLAP and ALPL changed over time, and s-dE(H) did not show a corresponding increase in BGLAP despite its strong mineral deposition. The authors interpret this as a possible temporal separation between gene-expression peaks and matrix formation.

They also note that elevated CD146 and SSEA4 in s-dE(H) could be relevant because these markers have been associated with differentiation-competent progenitor populations, but the experiment did not establish that marker-positive cells caused the mineralization. The results therefore support a selective association between the treated immortalized-cell matrix and osteogenic output, while leaving its cellular mechanism unresolved. The authors propose that heparan sulfate chains from normal and immortalized-cell matrices differ in structure and signaling activity.

One possibility is that chains on unmodified-cell perlecan store FGF-2 in a way that supports a self-renewing state and cartilage formation, so their removal weakens those signals. Chains in s-dE may instead restrict receptor activation, with their cleavage revealing or enabling lineage-promoting interactions. BMP-2 binding, Wnt-related effects and changes in matrix mechanics could also contribute.

These remain hypotheses because the investigators did not directly characterize chain length, sulfation patterns, growth-factor release or receptor activation, and they did not block FGF or BMP signaling to establish causality. They also did not determine whether heparinase altered the stiffness or elasticity of the two matrices, meaning that biomechanical differences cannot be separated from biochemical effects in the observed differentiation outcomes. Taken together, the findings identify matrix-resident perlecan heparan sulfate as a context-dependent regulator of IPFSC behavior rather than a uniform brake or accelerator of differentiation.

Its removal consistently favored adipogenesis, but its effect on cartilage formation depended on whether the matrix came from unmodified or SV40LT-immortalized cells, and osteogenic mineralization occurred only after preconditioning on the treated immortalized-cell matrix. The findings provide a rationale for engineering decellularized matrices with defined glycosaminoglycan architecture, but they also show why the matrix must be evaluated as a complete biochemical and mechanical environment. Direct glycomic analysis, measurements of growth-factor availability, pathway-specific inhibition and controlled comparisons of substrate stiffness are needed before the molecular mechanism or translational potential can be established.

Because the work was performed entirely in vitro with pooled donor-derived cells, the results do not yet establish safety, efficacy or clinical benefit in people or in an animal model of tissue repair.

Subject of Research: Source-dependent effects of perlecan heparan sulfate on human stem-cell differentiation in decellularized extracellular matrix

Article Title: Source-dependent perlecan heparan sulfate chains in dECM modulate stem cell fate and multilineage differentiation

Article References: Wei, Z., Yu, Q., Shahriar, A., Xie, X., Chen, S., & Pei, M. (2026). Source-dependent perlecan heparan sulfate chains in dECM modulate stem cell fate and multilineage differentiation. Frontiers in Bioengineering and Biotechnology, 14, Article 1930107. https://doi.org/10.3389/fbioe.2026.1930107

Image Credits: AI Generated

DOI: 10.3389/fbioe.2026.1930107

Keywords: stem cells, decellularized extracellular matrix, perlecan, heparan sulfate, chondrogenesis, osteogenesis, adipogenesis, infrapatellar fat pad, heparinase III, tissue engineering, Source-dependent, heparan