# Fileset

[ChemBioChem - 2026 - Kato - CD44‐Binding Peptide‐Functionalized Antibiofouling Polymer Surface for High‐Performance (1).pdf](https://mdr.nims.go.jp/filesets/7dc284a2-4e2f-45a9-853b-1ccfa1bdd6fe/download)

## Creator

Moe Kato, Tadashi Nakaji, Kazuaki Matsumura, [Chiaki Yoshikawa](https://orcid.org/0000-0002-6589-387X), Yuki Usui, Takahiro Kishioka, Taito Nishino

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[CD44‐Binding Peptide‐Functionalized Antibiofouling Polymer Surface for High‐Performance Separation of Human Mesenchymal Stromal Cells](https://mdr.nims.go.jp/datasets/9f482ab0-2acb-4c50-8b4e-949b6aa3102d)

## Fulltext

CD44‐Binding Peptide‐Functionalized Antibiofouling Polymer Surface for High‐Performance Separation of Human Mesenchymal Stromal CellsCD44-Binding Peptide-Functionalized AntibiofoulingPolymer Surface for High-Performance Separationof Human Mesenchymal Stromal CellsMoe Kato1 | Tadashi Nakaji-Hirabayashi1,2,3,4 | Kazuaki Matsumura5 | Chiaki Yoshikawa4 | Yuki Usui6 |Takahiro Kishioka6 | Taito Nishino71Graduate School of Innovative Life Science, University of Toyama, Toyama, Japan | 2Faculty of Engineering, Academic Assembly, University of Toyama,Toyama, Japan | 3Graduate School of Science and Engineering, University of Toyama, Toyama, Japan | 4Research Center for Macromolecules andBiomaterials, National Institute for Materials Science (NIMS), Tsukuba, Japan | 5School of Materials Science, Japan Advanced Institute of Science andTechnology, Nomi, Japan | 6Materials Research Laboratories, Nissan Chemical Corporation, Toyama, Japan | 7Biological Research Laboratories, NissanChemical Corporation, Shiraoka, JapanCorrespondence: Tadashi Nakaji-Hirabayashi (nakaji@eng.u-toyama.ac.jp)Received: 29 October 2025 | Revised: 23 December 2025 | Accepted: 5 January 2026Keywords: antifouling surface | cell separation | human mesenchymal stem cell | label-free purification | peptide-functionalized polymer | regenerativemedicine | zwitterionic polymerABSTRACTCell transplantation therapy is a promising strategy for next-generation regenerative medicine. However, its clinical application islimited by the absence of safe and label-free methods for obtaining pure populations of functional cells. This study reports apeptide-functionalized zwitterionic polymer system designed for the selective and noninvasive isolation of human mesenchymalstem cells (hMSCs) from heterogeneous cell populations. The substrate was modified with poly(CMBMAm-co-PGMAn-co-MPTMS1) (PCmPnM1), which was synthesized via free radical polymerization, and subsequently conjugated with a CD44-bindingpeptide (CD44BP). Among the compositions examined, the PC7P2M1–CD44BP surface exhibited superior antifouling properties,effectively suppressing nonspecific protein adsorption and adhesion of nontarget cells while selectively capturing hMSCs. A col-umn packed with PC7P2M1–CD44BP–modified silica microparticles successfully isolated high-purity hMSCs from mixed cell sus-pensions within 35 min. Flow cytometry confirmed that the cells eluted later exhibited higher CD44 and CD105 expression,indicating separation based on antigen expression. The isolated hMSCs maintained proliferation and differentiation capacitiesequivalent to those of the preseparated cells, demonstrating that this device preserved cell functionality. This peptide–polymerhybrid column provides a simple and clinically adaptable platform for safe, label-free purification of hMSCs intended for celltransplantation therapy.1 | IntroductionCell transplantation therapy is anticipated to become the corner-stone of next-generation medical interventions, and numerousstudies are actively progressing toward its practical implementa-tion. However, this therapy requires overcoming several chal-lenges, including the establishment of a stable cell supply [1],precise control of cellular behavior [2], and safety assurance [1].Among these, the development of techniques for the selective iso-lation and purification of target cells is particularly critical.The candidate sources of transplantable cells include autologousand allogeneic cells [3]. However, regardless of the source, thecollected cell populations consisted of multiple cell types, makingthe isolation of specific target cells difficult. Several studies havereported that undesired cells are associated with carcinogenesisThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, providedthe original work is properly cited.© 2026 The Author(s). ChemBioChem published by Wiley-VCH GmbH.ChemBioChem, 2026; 27:e202500822 1 of 16https://doi.org/10.1002/cbic.202500822ChemBioChemRESEARCH ARTICLEhttps://orcid.org/0000-0002-7562-3419mailto:nakaji@eng.u-toyama.ac.jphttp://creativecommons.org/licenses/by/4.0/https://doi.org/10.1002/cbic.202500822https://doi.org/10.1002/cbic.202500822http://crossmark.crossref.org/dialog/?doi=10.1002%2Fcbic.202500822&domain=pdf&date_stamp=2026-02-12and reduced differentiation efficiency. Moreover, the inclusion ofundifferentiated cells within differentiated cell populations inc-reases the risk of carcinogenesis, highlighting the essential needfor selective cell isolation, not only during tissue engineering butalso during transplantation. In this context, the development ofreliable techniques for the isolation and recovery of specific celltypes is crucial [4], and continued technological advancementsare indispensable for the successful realization of cell transplan-tation therapy [5].Currently, techniques such as fluorescence-activated cell sorting(FACS) [6], magnetic-activated cell sorting (MACS) [7], and den-sity gradient centrifugation [8] are being widely used to isolatespecific cell types. Although these methods are well-establishedand commonly used in research, challenges remain regardingtheir safety and procedural complexity when applied to the sep-aration of transplantable cell sources. For instance, FACS caninduce cytotoxicity owing to residual fluorescent labeling agentson the cell surface and the mechanical stress exerted by fluid flow[9]. Similarly, MACS is associated with residual labeling agents,and the removal of the magnetic beads is highly complex. In thecase of density gradient centrifugation, concerns persist regard-ing residual solvents, such as Ficoll and Percoll. Ficoll is cytotoxicand unsuitable for clinical use, whereas Percoll, although lesstoxic, requires caution when used in clinical applications.Therefore, although the existing separation methods offer highaccuracy and versatility, their suitability for clinical-grade cellseparation remains limited.Considering these limitations, we aimed to develop a devicecapable of separating specific cell types from heterogeneous cellpopulations with minimal invasiveness and a simple operation.Considering practical applications, the device was designed forconvenient operation on a clean bench, eliminating the needfor large-scale equipment. Our research group previously demon-strated that coating surfaces with zwitterionic polymers effec-tively inhibited nonspecific proteins and cell adhesion [10].Recent studies have highlighted the advantages of zwitterionicpolymer-based surface coatings, particularly their long-term sta-bility and enhanced antibiofouling performance. For instance,self-renewing zwitterionic silicone hybrid coatings have beenreported to provide multimodal bacterial resistance and robustantibiofouling properties [11]. These advancements underscorethe growing relevance of zwitterionic materials in preventingundesired bioadhesion and support the rationale for employingsuch materials in the development of practical cell separationplatforms. Based on this foundation, various functional materialshave been developed for this purpose. Additionally, we havedeveloped advanced device designs that enable the immobiliza-tion of proteins and oligopeptides onto substrates while preserv-ing their biological activity, thereby promoting the adhesion andfunctionality of target cells [12].Based on these insights, we developed a device capable of selec-tively capturing specific cell types. Human mesenchymal stromal/stem cells (hMSCs) were selected as the target cell type becausethey are among the most promising cells for clinical applications.Using phage display technology, our research group previouslyidentified a peptide sequence, QQGWFPGAG (CD44-binding pep-tide, CD44BP), that selectively binds to the CD44 cell membranereceptor antigen in hMSCs [13]. By employing this peptide tointeract with antigens on the surface of target cells, we hypothe-sized that highly precise and noninvasive cell separation could beachieved. Furthermore, to minimize the nonspecific adhesion ofunintended cells and proteins [4], which is often observed in sep-aration methods utilizing molecular interactions, we introducedzwitterionic polymers to suppress random adsorption.In this study, we developed microparticle carriers capable ofselectively capturing hMSCs by incorporating a carboxymethylbetaine (CMBMA) polymer [10], which suppresses nonspecificprotein and cell adhesion, and introducing CD44BP into the poly-mer side chains. Subsequently, we constructed a cell separationcolumn packed with silica microparticles coated with a CD44BP-functionalized antibiofouling polymer [13]. The selective separa-tion efficiency of hMSCs using this column was comprehensivelyevaluated along with the maintenance of hMSC functionalityafter separation.2 | Materials and Methods2.1 | Synthesis of Ternary Copolymer:Poly(carboxymethylbetaine-co-propargylmethacrylate-co-3-methacryloyloxypropyltrimethoxysilane)Amixed solvent of ethanol (EtOH) and N, N-dimethylformamide(DMF), dehydrated using 3 Å molecular sieves, was preparedby dissolving carboxy-N, N-dimethyl-N-(2 0-methacryloylox-yethyl) methanaminium inner salt (CMBMA; Osaka OrganicChemical Industry Co., Ltd., Osaka, Japan), propargyl meth-acrylate (PGMA; Hydrus Chemical Inc., Tokyo, Japan),and 2,2 0-azobisisobutyronitrile (AIBN; FUJIFILM Wako PureChemical Co., Ltd., Osaka, Japan) at a concentration of 2.19 ×10−2 M. The total monomer concentration for the polymerizationreaction was fixed at 0.5 M. To remove dissolved oxygen, themixture was purged with nitrogen gas for 20 min. Subsequently,3-(trimethoxysilyl) propylmethacrylate (MPTMS;TokyoChemicalIndustry Co., Ltd., Tokyo, Japan) was added, followed by anadditional 10-min nitrogen purge. The react ion vessel was thensealed and maintained at 70°C for 8 h (Scheme 1). To clarify thedesign rationale for the copolymer, the roles of each monomerare summarized as follows: CMBMA possesses a zwitterionicstructure and exhibits strong antibiofouling properties by sup-pressing nonspecific proteins and cell adhesion. PGMA, whichcontains a terminal alkyne group, serves as a reactive site for theintroduction of the CD44-binding peptide (CD44BP) into thepolymer side chains. In this study, CD44BP modified with anazide group was used, and the peptide was covalently immobi-lized through a click reaction between the azide and alkynegroups, forming a stable 1,2,3-triazole linkage under mild andefficient reaction conditions. MPTMS acts as a silane couplingmonomer, enabling the covalent attachment of the copolymer tothe silica substrate through hydrolyzable alkoxysilane groups.By combining these three monomers, the copolymer simulta-neously achieved antifouling capability, click-chemistry-basedpeptide immobilization, and strong anchoring to the substrate,thereby creating a functional surface well suited for the selectiveisolation of hMSCs. Table 1 summarizes the polymerizationconditions for each composition ratio and molecular weightof the synthesized polymers, which were determined usinggel permeation chromatography (GPC). The polymers obtainedwith various composition ratios, poly(CMBMAm-co-PGMAn-co-MPTMS1), are abbreviated as PCmPnM1.2 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License2.2 | Synthesis of CD44-Binding PeptideAzidohomoalanine (Aha)-terminated CD44BP was synthesizedvia solid-phase peptide synthesis using an Rotary Shaker N-500(Kokusan Chemical Co., Ltd.). Fmoc-NH-SAL-PEG resin(0.21mmol/g; Watanabe Chemical Industries Ltd., Hiroshima,Japan) was washed with DMF and methanol, followed by swellingin a solvent containing 25% dimethyl sulfoxide (DMSO) in DMFfor 30min. Subsequently, the resin-bound amino acid Fmoc-protecting groups were removed by washing withDMF and stirringthe resin in 20% piperidine/DMF for 30min. The resin wasthen washed again with DMF and methanol and resuspendedin 25% DMSO/DMF. Fmoc-amino acids (0.63mmol; WatanabeChemical Industries Ltd.) was coupled to the resin by the addit-ion of 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholiniumchloride (1.05mmol) and 4-methylmorpholine (0.42mmol), fol-lowed by stirring to promote coupling. Peptide cleavage fromthe resin was achieved by stirring the resin in a solvent mixturecontaining 9.5mL of trifluoroacetic acid (TFA), 0.25mL of triiso-propylsilane, and 0.25mL of water per 1 g of Fmoc-NH-SAL-PEGresin for 3 h. After cleavage, the peptide was precipitated by coolingwith diethyl ether and collected by centrifugation (2380 × g,30min, −4°C) to remove residual ether. The peptide was redis-solved in water, and the resin was removed by filtration. The filtratewas flash-frozen in liquid nitrogen and lyophilized using a smallfreeze dryer (FDS-1000, Tokyo Rikakikai Co., Ltd., Tokyo, Japan).Amino acid sequences of the synthesized peptides are summa-rized in Table 2. To evaluate the effect of linker length on thecell capture efficiency, three CD44BPs with different linkerlengths were synthesized. Additionally, to confirm that cellularcapture was specifically mediated by CD44BP, inhibition experi-ments were performed using freely available (nonimmobilized)CD44BP. Therefore, a CD44BP variant without Aha was synthe-sized for comparison.TABLE 1 | Reaction conditions and characterization of poly(CMBMAm-co-PGMAn-co-MPTMS1) (PCmPnM1).aSampleFeeding ratioa(CMBMA : PGMA : MPTMS)Solvent(EtOH/DMF) Mw (Mw/Mn)bPC9P0M1 9 : 0 : 1 10/0 3.29 × 104 (3.81)PC8P1M1 8 : 1 : 1 9.5/0.5 2.53 × 104 (3.28)PC7P2M1 7 : 2 : 1 7.5/2.5 2.29 × 104 (3.15)PC6P3M1 6 : 3 : 1 7.5/2.5 2.00 × 104 (2.95)PC5P4M1 5 : 4 : 1 7.5/2.5 1.66 × 104 (3.26)aThe composition of MPTMS was fixed at 10%. This ratio was optimized for achieving adequate surface coverage and enhancing the antibiofouling performance ofCMBMA [10].bThe average molecular weight and polydispersity index of PCmPnM1 were determined by GPC. Solvent: methanol/DMF mixture (3:1). Column: Wako Beads G-50.Standard: PMMA.SCHEME 1 | Synthesis of a functional copolymer composed of CMBMA (antifouling unit), PGMA (alkyne unit for click-mediated peptideconjugation), and MPTMS (silane monomer for anchoring to silica substrates).TABLE 2 | Amino acid sequences of CD44-binding peptides.Peptide Amino acid sequenceCD44BP-Aha(L0) Gln–Gln–Gly–Trp–Phe–Pro–AhaCD44BP-Aha(L3) Gln–Gln–Gly–Trp–Phe–Pro–Gly–Ala–Gly–AhaCD44BP-Aha(L14) Gln–Gln–Gly–Trp–Phe–Pro–(Gly–Ala)7–AhaCD44BP(L3) Gln–Gln–Gly–Trp–Phe–Pro–Gly–Ala–GlyChemBioChem, 2026 3 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License2.3 | Construction of Surface Modified WithTernary Copolymer Immobilized With CD44BPGlass substrates (26 × 20mm; ASONE Co., Ltd., Tokyo, Japan)and silicon wafers (20 × 20mm; Matsuzaki Seisakusyo Co.,Ltd., Shimane, Japan) were immersed in a piranha solution (7:3mixture of sulfuric acid and hydrogen peroxide) for 30min. Thesubstrates were then rinsed thoroughly with distilled water andmethanol and dried under an air stream. The cleaned substrateswere subsequently immersed in a 1% w/v solution of PCmPnM1 atroom temperature for 24 h to modify the polymer surface via asilane coupling reaction. After the modification, the substrateswere removed from the solution, rinsed with ethanol and meth-anol, and dried.The PCmPnM1-modified substrates were then immersed in anaqueous solution containing CD44BP-Aha (9.32× 10−4 mol/mL),copper sulfate (9.32× 10−5 mol/mL), and sodium L-ascorbate(9.32× 10−5 mol/mL), and reacted at room temperature for 24 h.This step immobilized CD44BP onto the polymer surface througha copper(I)-catalyzed azide–alkyne cycloaddition “click” reactionbetween the alkyne groups of the polymer side chains and theazide terminus of the peptide (Scheme S1). After the reaction,the modified substrates were thoroughly washed with distilledwater and dried under a stream of nitrogen gas.The surface modification of nonporous silica microspheres (250–350 μm; Fuji Manufacturing Co., Ltd., Tokyo, Japan) used in thecell separation experiments was performed as follows. The silicamicrospheres were immersed in the piranha solution for 30minin a glass centrifuge tube and then thoroughly rinsed with distilledwater until the pH reached 7.0. Subsequently, the solution wasreplaced with dehydrated ethanol and the microspheres were agi-tated at room temperature for 24 h in a 2w/v% PCmPnM1 solution.After washing with ethanol and distilled water, CD44BP-Aha solu-tion (1mg/mL) was added, and the suspension was agitated atroom temperature for 24 h. After the reaction, the particles wererinsed thoroughly with distilled water to obtain PCmPnM1–CD44BP–modified silica microspheres (Scheme S2).2.4 | Characterization of PCmPnM1The molecular weight of PCmPnM1, prepared as a 10% (w/v) solu-tion, was determined using GPC (Agilent 1260 Infinity; AgilentTechnologies Japan, Ltd., Tokyo, Japan). Wakopak Wakobeads-G-50 columns (FUJIFILM Wako Pure Chemical Co., Ltd.) wereused for chromatographic separation. The eluent consisted of amixture of DMF) and methanol (1:3, v/v) with a flow rate of1.0mL/min. The polymers were detected using a refractive-indexdetector (Agilent Technologies, Japan, Ltd.). poly(methyl methac-rylate) (PMMA; Standard M-75, Showa Denko K.K., Tokyo, Japan)was used as the calibration standard. Data analysis was conductedusing an SIC μ7 Plus system (SYSTEM INSTRUMENTS Co., Ltd.,Tokyo, Japan).2.5 | Characterization of CD44BPCircular dichroism (CD) spectra of the synthesized peptides wereobtained using a circular dichroism spectropolarimeter (J-820;JASCO Corporation, Tokyo, Japan). Measurements were per-formed with a sensitivity of 100 mdeg over a wavelength rangeof 300–180 nm, at a scanning speed of 100 nm/min, with data accu-mulated over eight scans at 37°C. A quartz cell with a 0.2 cm pathlength of used. Data analysis was conducted using the JASCOSpectrum Manager software (JASCO Corporation, Tokyo, Japan).The mean residue ellipticity ([θ], deg·cm2·dmol−1) was calculatedaccording to Equation (1), where θ represents ellipticity (mdeg), l isthe path length (cm), c is the peptide concentration (mol/L), and nis the number of residues in the peptide [14].θ½ �= θ=10lcn (1)Liquid chromatography–mass spectrometry (LC–MS) of thesynthesized peptides was performed using a ShimadzuProminenceThermo LXQ system. The mobile phase consisted ofacetonitrile: water (1:1, v/v) containing 0.1% TFA. The peptidesolution (1mg/mL) was eluted at a flow rate of 1.0mL/minusing a GL Science Inert Sustain C18 column (5 μm). Data anal-ysis was performed using Xcalibur software (Thermo FisherScientific Inc.).2.6 | Wettability of Ternary Polymer- andCD44BP-Carried Ternary Polymer-ModifiedSurfacesTo evaluate the wettability of the PCmPnM1– and PCmPnM1–CD44BP-modified surfaces, contact angle measurements wereconducted using a contact angle meter (DMs-401; KyowaInterface Science Co., Ltd., Saitama, Japan). A 1.0 μL dropletof distilled water was placed on the substrate at room tempera-ture, and the contact angle was measured after 30 s using the θ/2method. Ten measurements were performed for each modifiedsurface and the mean value was recorded as the representativestatic contact angle. In addition, dynamic contact angle measure-ments were performed to assess the time-dependent changes insurface wettability. A 1.0 μL water droplet was placed on the sur-face, and the contact angle was recorded every 5 s for 10min.Dynamic contact angle profiles were obtained from ten indepen-dent measurements for each sample [15].2.7 | Characterization of Ternary Polymer- andCD44BP-Carried Ternary Polymer-Modified SurfaceEach modified surface was characterized using X-ray photo-electron spectroscopy (XPS) with an ESCALAB 250Xi instru-ment (Thermo Fisher Scientific K.K., Waltham, MA, USA).Measurement conditions included a sample angle of 0°, anX-ray incidence angle of 32°, and a detector reflection angleof 90°, with a 200 μm X-ray spot size. The binding energiesof the detected elements were as follows: C1s, 280–300 eV;O1s, 527–545 eV; N1s, 392–410 eV; and Si2p, 95–110 eV.Spectral analysis was conducted using Advantage software(version 5.1; Thermo Fisher Scientific Inc.).Attenuated total reflection Fourier transform infrared (ATR–FTIR) spectroscopy was performed using an infrared spectropho-tometer (Nicolet iS5 FT-IR; Thermo Fisher Scientific, Inc.).Silicon wafers were used as the baseline reference, and absorp-tion spectra were recorded in the range of 400–4000 cm−1 with512 scans. Data analysis was conducted using the OMNIC soft-ware (Thermo Fisher Scientific Inc.).4 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseThe film thickness on the substrate surface in the dry state wasdetermined using an automatic ellipsometer (MARY-102, FiveLab Co., Ltd., Saitama, Japan) based on the ellipsometry principle.A 632.8 nm He–Ne laser was employed as the light source. For thepolymer layer measurements, the refractive index of PMMA wasset to 1.489, and that of the peptide layer was set to 1.459.The surfaces of the polymer-modified and peptide-functionalizedsubstrates were examined using a tabletop scanning electronmicroscope (SEM) (Miniscope TM4000Plus II, Hitachi High-Tech Corp., Tokyo, Japan; 2021). The acceleration voltage wasset to 10 kV. The SEM images were acquired at two magnifica-tions, 100× and 5000×, to evaluate the overall surface uniformityand fine microstructural features.The amount of peptide loaded onto the substrate was quantifiedusing the bicinchoninic acid (BCA) assay. The dry weights ofthe polymer-modified substrates (26mm× 38mm) and peptide-carrying polymer-modified substrates (26mm× 38mm) were mea-sured. After pretreatment with Dulbecco’s phosphate-bufferedsaline (PBS, without Ca2+/Mg2+), 400 μL of BCA reagent wasapplied to each substrate and incubated for 2 h at 37°C. The abs-orbance at 570 nm was measured using a microplate reader(Multiskan JX; Thermo Fisher Scientific Inc., USA). A calibrationcurve was constructed using standard albumin (BSA) standardsolutions.2.8 | Antibiofouling Property of Ternary Polymer-and CD44BP-Carried Ternary Polymer-ModifiedSurfacesThe amount of adsorbed BSA was quantified using the bicincho-ninic acid assay to evaluate the inhibition of protein adsorption oneach modified surface. Each polymer-modified and peptide-loadedsubstrate (26mm× 38mm) was pretreated with PBS and subse-quently incubated with a 2 wt% BSA solution in PBS at room tem-perature for 2 h. After incubation, the BSA solution was removedand each substrate was washed ten times with PBS to eliminateunbound proteins. A silicon frame cleaned and cut to match thedimensions of the substrate was placed at both ends of the modifiedsubstrate. A cleaned glass slide was then placed on top using a sili-con frame as the spacer. The gap between the modified and glasssubstrates was filled with BCA reagent, which served as the colori-metric indicator, and the assembly was incubated at 37°C for 2 h.The absorbance of the resulting solution was measured at 570 nmusing a microplate reader (Multiskan JX; Thermo Fisher ScientificInc.). A calibration curve was generated using standard BSA solu-tions of known concentrations.2.9 | Selectivity of Cell Capture on CD44BP-Carried Ternary Polymer-Modified SurfacesTo evaluate the cell capture selectivity of the PCmPnM1-andPCmPnM1–CD44BP-modified surfaces, hMSCs, which expressthe CD44 surface antigen, and human embryonic kidney 293(HEK293) cells, which lack CD44 expression, were seeded ontoeach modified surface.Frozen human mesenchymal stem cells (hMSCs; MSC-R50, CellNo. : HMS0047; RIKEN BioResource Research Center, Japan,passage 2) and HEK293 cells (Research Resource Identifier(RRID): CVCL_0045, passage 157) were thawed and seeded onto10 cm polystyrene culture dishes, then cultured for 3 daysin growth medium containing 10% fetal bovine serum (FBS),100 U/mL penicillin, and 100 μg/mL streptomycin in Dulbecco’sModified Eagle Medium (DMEM, high glucose) at 37°C and 5%CO2. Mouse fibroblast cells (NIH3T3; RRID: CVCL_0594, passage148) were maintained under similar conditions using MinimumEssentialMedium supplementedwith 10%FBS, 100U/mLpenicil-lin, and 100 μg/mL streptomycin at 37°C and 5% CO2.Upon reach-ing 70%–80% confluence, the cells were detached using 0.25%trypsin–1mM EDTA solution (Nacalai Tesque, Kyoto, Japan)and used for subsequent experiments.Thecollected single-cell suspensionswere seededontounmodifiedglass, PCmPnM1-modified, or PCmPnM1–CD44BP-modified glasssubstrates (26mm × 20mm) at a density of 2.0 × 104 cells/cm2.All the substrates were sterilized by immersion in 70% ethanol,air-dried, and handled aseptically on a clean bench. After 24 hof incubation at 37°C and 5% CO2, the cells were rinsed withPBS to remove nonadherent cells. Cell adhesion was observedusing a phase-contrast microscope (IX71; Olympus Corporation,Tokyo, Japan), and the number of adherent cells was quantifiedby staining the nuclei with Hoechst 33 342.2.10 | Construction of Cell-Separating ColumnThe silica microspheres functionalized with PC7P2M1–CD44BP-modified surfaces were sterilized by immersion in 70% ethanol.Within a clean bench, the ethanol was completely replaced withPBS, and the microspheres were packed into a syringe equippedwith a 100 μm filter (Cell Fraction Filter Filcon Syringe 100 μm;AS ONE Corporation, Osaka, Japan) to construct the PC7P2M1–CD44BP-modified column. The column was densely packed withsilica microspheres by continuously passing PBS overnightthrough a peristaltic pump (Quantitative Liquid Feed Pump;Tokyo Rikakikai Co., Ltd., Tokyo, Japan) (Figure S1).2.11 | Fluorescent Labeling of Cells for CellSeparation AnalysisTo differentiate between HEK293 cells and hMSCs, fluorescentstaining was performed using the CellTracker dye (ThermoFisher Scientific Inc., USA). HEK293 cells were labeled withCellTracker Green CMFDA dye and hMSCs were labeled withCellTracker Orange CMRA dye. Both dyes were dissolved inDMSO to prepare 1.0 mM stock solutions, which were subse-quently diluted in serum-free medium to final concentrationsof 3.0 μM (CellTracker Green) and 2.0 μM (CellTracker Orange).After culturing the cells to �70%–80% confluence, the HEK293and hMSC monolayers were rinsed with PBS, and the respectivestaining solutions were added. Cells were incubated at 37°C and5% CO2 for 30 min. Following incubation, the staining solutionswere removed and the cells were recovered by adding DMEMsupplemented with FBS and incubating under the sameconditions.2.12 | Specific Cell SeparationThe PC7P2M1–CD44BP-modified column was first equilibratedwith PBS, after which 200mL of serum-free DMEM was passedChemBioChem, 2026 5 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensethrough the column to replace the buffer (Figure S1). PrelabeledHEK293 and hMSC suspensions were prepared, each containing1.0 × 106 cells. A 1mL aliquot of the mixed cell suspension wasloaded onto the column, followed by perfusion with 50mL ofserum-free DMEM. The eluate, which immediately flowed outfrom the bottom of the column, was collected in 1mL fractions.After collecting 50mL of the medium, 100mL of 5 mM EDTA inPBS was introduced into the column to detach weakly adherentcells, which were subsequently recovered. The number of cells ineach fraction was determined using a hemocytometer under afluorescence microscope. Additionally, the cells recovered withEDTA/PBS were subjected to a second round of separation byreloading onto a new column.Furthermore, to verify that the cell-separation capability of thePC7P2M1–CD44BP-modified column originates from the specificinteraction between hMSCs and immobilized CD44BP on the sil-ica particle surface, a competitive inhibition experiment was per-formed by preincubating cells with free CD44BP prior to columnloading. This experiment demonstrated that the precontact ofhMSCs with free CD44BP attenuated their interaction withthe PC7P2M1–CD44BP-modified column, thereby decreasing theseparation efficiency. As described above, a PC7P2M1–CD44BP-modified column was prepared and suspensions of CellTracker-labeled HEK293 cells and hMSCs (each 1.0 × 106 cells in 1 mL)were supplemented with 700 nmol/mL free CD44BP. The mixedcell suspension (1 mL) was loaded onto the column, followed byperfusion with 30 mL of a medium containing the same concen-tration of CD44BP. The eluate was collected in 1mL fractions.Finally, 20 mL of the medium and 5mM EDTA in PBS werepassed through the column to recover the residual cells. Thenumber of cells in each fraction was determined as describedpreviously.2.13 | Analysis of Separated Cells Using FlowCytometryCells separated using the developed PC7P2M1–CD44BP-modifiedcolumn were analyzed for the quantitative expression of CD44and CD105 surface antigens by flow cytometry. The eluate col-lected from the column was fractionated into 10 mL portions andcentrifuged. The resulting cell pellets were washed with phos-phate buffered saline (PBS) and fixed in 4% paraformaldehyde(PFA) for 20 min. After removing the PFA solution, nonspe-cific protein adsorption was blocked using a blocking buffer.Subsequently, the cells were incubated in blocking buffer con-taining CD105 antibody (1:150; CD105 (Endoglin) MonoclonalAntibody (SN6), PE; eBioscience, Thermo Fisher Scientific Inc.)and CD44 antibody (1:200; Anti-Human/Mouse CD44 FITC;eBioscience, Thermo Fisher Scientific Inc.) for 2 h. After thor-oughly washing with PBS containing 0.05% Tween-20, thecell suspension was adjusted to a final concentration of 2 ×105 cells/mL (1 mL per sample).Flow cytometric analysis was performed using a FACS Canto IIflow cytometer (BD Biosciences, Franklin Lakes, NJ, USA).Side and forward scatter signals were used to identify cell pop-ulations, with unstained HEK293 cells and hMSC serving asnegative controls. The fluorescence intensity settings wereestablished using positively stained cell populations (CD44and CD105). The fluorescence profiles of the separated cellpopulations were analyzed using FACSDiva 6.1 software (BDBiosciences).2.14 | Evaluation of Multipotency of SeparatedCellsTo evaluate the proliferative and differentiation potential of hMSCsafter separation, the cells were subjected to proliferation assays andinduced to differentiate into osteoblasts and adipocytes. Theirbehavior was compared with that of hMSCs before separation.Cells collected from the PC7P2M1–CD44BP-modified column andconventionally cultured hMSCs were seeded at a density of1.0 × 104 cells/cm2 and cultured at 37°C in a humidified atmo-sphere containing 5% CO2. The proliferative capacity wasassessed by counting the number of viable cells using a hemocy-tometer on days 1, 2, 3, and 4 after seeding.For differentiation analysis, both separated and conventionallycultured hMSCs were seeded at a density of 2.0× 104 cells/cm2and cultured under the same conditions until they reached conflu-ence. The medium was then replaced with the osteogenic oradipogenic differentiation inductionmediumcompositions shownin (Table 3). After 20 days of culture, osteogenic differentiationwasqualitatively assessed by alkaline phosphatase staining (AlkalinePhosphatase Live Stain, Invitrogen, Thermo Fisher ScientificInc.), and adipogenic differentiation was evaluated by Oil RedO staining for lipid droplet formation (FUJIFILM Wako PureChemical Co. Ltd.).2.15 | Statistical AnalysisStatistical analyses of the contact angle, protein adsorption, andcell adhesion were performed using JMP Pro 15 software (SASInstitute Inc., Tokyo, Japan). One-way analysis of variance(ANOVA) was conducted to determine the statistical significanceamong groups. Tukey’s honest significant difference test was sub-sequently applied for multiple comparisons, with the significancelevel set at p< 0.001.TABLE 3 | Composition of media for osteogenic and adipogenicdifferentiation.Osteogenic differentiation ConcentrationFBS 10%Penicillin–streptomycin 1% 1%Dexamethasone 100 nML-Ascorbic acid 2-phosphate trisodium salt 50 μMβ-Glycerophosphate disodium salt n-hydrate 10 mMAdipogenic differentiationFBS 10%PC/SM 1%Dexamethasone 1 μMIndomethacin 0.2 mM3-Isobutyl-1-methylxanthine 0.5 mMInsulin 10 μg/mL6 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License3 | Results and Discussion3.1 | Characterization of Ternary PolymerIn this study, the composition of MPTMS was fixed at 10% for allpolymer formulations (PCmPnM1), and polymers with molecularweights ranging from �20,000 to 40,000 were successfully syn-thesized. Our research group has previously demonstrated twokey findings: (1) a surface coating with high coverage can beachieved when the MPTMS content is maintained at 10% andthe molecular weight is between 20,000 and 40,000, and (2)increasing the MPTMS content or producing polymers of highermolecular weight leads to gelation during synthesis, as reportedin reference [10]. Therefore, these conditions were consideredoptimal in the present study.The combination of the three monomers used in this studywas intentionally selected to integrate their antifouling pro-perties (CMBMA), peptide conjugation capability (PGMA), andsubstrate-anchoring functionality (MPTMS), which togetherenabled the formation of a stable and selective surface for hMSCcapture. We attempted to determine the copolymer composition ofvarious PCmPnM1 samples using 1H NMR spectroscopy. However,the characteristic proton peaks of each monomer unit overlapped,making an accurate analysis impossible. Consequently, the compo-sitional variations between the CMBMA and PGMA units wereevaluated indirectly by preparing polymer-modified surfaces andanalyzing the changes in the elemental composition using XPS(discussed later). The molecular weights obtained from the GPCanalysis ranged from �20,000 to 40,000 for all polymers. A slightdecrease in the molecular weight was observed with increasingPGMA content, which is likely attributable to the hydrophobicnature of the PGMA monomers, limiting their copolymerizationefficiency with the hydrophilic monomers CMBMA and MPTMS.A comparison of the feed and obtained monomer ratios revealedthat the PGMA content in the final copolymer was slightly lowerthan that in the initial feed, suggesting incomplete incorporation ofPGMA into the polymer structure.3.2 | Characterization of the CD44 BindingPeptideThe synthesized CD44BP–Aha and CD44BP peptides were char-acterized by CD spectroscopy (Figure 1). Upon comparison ofCD44BP(L0)–Aha, CD44BP(L3)–Aha, and CD44BP(L14)–Aha(Figure 1A) with CD44BP(L3) (Figure 1B), distinct negativeCotton peaks were observed around 216 nm, indicating the pres-ence of β-strand and β-sheet conformations in these peptides.Additionally, CD44BP(L3) exhibited a negative Cotton peak near200 nm, whereas in CD44BP(L14)-Aha, the negative Cotton peaknear 215 nm disappeared, leaving only a feature around 200 nm.Based on these characteristic spectral features and their relativeintensities, it can be inferred that CD44BP(L0)–Aha, which lacksa linker sequence, tends to adopt β-strand and β-sheet conforma-tions, whereas the introduction of a linker sequence promotes thetransition toward random coil structures.The molecular weights of the synthesized CD44BP peptides werefurther evaluated by LC–MS (Figure S2). The theoretical molecularweights of the peptides were as follows: CD44BP(L0)–Aha,887.57 g/mol; CD44BP(L3)–Aha, 1073.14 g/mol; CD44BP(L14)–Aha, 1784.48 g/mol; and CD44BP, 947.39 g/mol. The measuredmolecular weights of all the peptides closely matched their theoret-ical values, confirming their successful synthesis. Additionally, pep-tide purity was calculated from the peak areas in the LC–MSspectra, which revealed purities greater than 95% for all samples.These findings confirmed that highly pure peptides were success-fully obtained.3.3 | Surface Characterization of theCD44BP-Carried Ternary Polymer SurfaceFor the surfaces modified with PCmPnM1, the CMBMA contentwas evaluated by elemental analysis using XPS based on thecarbon-to-nitrogen (C/N) ratio. Denoting the proportion ofCMBMA as “a” and the combined proportion of PGMA andMPTMS as “b”, Equation (2) was applied to PC9P0M1, whileEquation (3) was used for the other copolymers.CN=120a+ 120b14a(2)CN=120a+ 204b14a(3)The CMBMA contents determined from the C/N ratios of the vari-ous PCmPnM1-modified surfaces are shown in Figure 2A. TheCMBMA content decreased as the initial feed ratio of CMBMAdecreased. However, the obtained values deviate slightly fromFIGURE 1 | Circular dichroism (CD) spectra of CD44BP peptides. (A) CD spectra of CD44BP(L0)–Aha (solid line), CD44BP(L3)–Aha (dashed line),and CD44BP(L14)–Aha (dotted line). (B) CD spectrum of CD44BP(L3) without Aha modification.ChemBioChem, 2026 7 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensethe theoretically expected CMBMA compositions. This deviationcan be attributed to the influence of carbon originating from themethoxy groups in MPTMS, as the polymer attachment to the glasssubstrate was achieved via silane coupling through MPTMS.Consequently, fluctuations in the carbon content of MPTMS likelyaffected the C/N ratio. Although precise quantification of theCMBMA content and copolymer composition was not possiblefrom these measurements, the results clearly demonstrated the suc-cessful fabrication of polymer-coated surfaces with varying ratios ofCMBMA and PGMA. Therefore, these polymer variations enabledfurther exploration of the optimal surface conditions for selectivecell capture. In addition to these observations, it is important toacknowledge the methodological limitations associated with deter-mining copolymer composition. First, 1H NMR analysis could notaccurately quantify the amounts of CMBMA, PGMA, and MPTMSbecause the characteristic peaks of each monomer unit stronglyoverlapped, preventing precise peak assignment. Second, composi-tional estimation based on XPS also has intrinsic constraints; thecarbon signals derived from the methoxy groups of MPTMS con-tribute to the overall C/N ratio, causing systematic deviations fromthe theoretical feed composition and preventing accurate quantifi-cation of PGMA, which serves as the peptide anchoring point.Despite these quantitative limitations, this issue did not underminethe interpretation of the overall polymer behavior. Importantly,multiple independent surface analyses, including XPS elementalcomposition, water contact angle measurements, ellipsometric filmthickness, and ATR–FTIR spectra, revealed trends that were con-sistent with the expected variations resulting from the initial feedFIGURE 2 | Surface characterization of PCmPnM1 and PCmPnM1–CD44BP coatings. (A) Ratio of CMBMA incorporated into PCmPnM1 copolymersimmobilized on the surface. (B) Static water contact angles measured on surfaces modified with PCmPnM1 (dark gray) and PCmPnM1–CD44BP(L3) (lightgray). (C) Static water contact angles measured on PC8P1M1, PC7P2M1, and PC6P3M1 surfaces functionalized with CD44BP peptides containing differentlinker lengths (L0, L3, L14). (D) Film thicknesses of surfaces coated with PCmPnM1 (black bars) and PCmPnM1–CD44BP(L3) (gray bars). (E) IR spectra ofPC7P2M1-modified surfaces before (black line) and after (gray line) CD44BP(L3) immobilization. Data presentation and statistics: Data in panels (A–D)are presented as mean ± SD (n = 3 for A and D; n = 10 for B and C). Statistical significance was evaluated using one-way ANOVA followed by Tukey’sHSD test (p < 0.001). Bars labeled with different letters indicate statistically significant differences among samples.8 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licenseratios of CMBMA and PGMA. These results strongly supported theidea that adjusting the feed ratio effectively modulates the surfaceproperties and peptide-anchoring capability of the copolymer.Therefore, the selection of PC7P2M1 as the optimal formulationis justified, even though NMR- and XPS-based quantification can-not provide exact compositional values.Next, the wettability of the PCmPnM1- and PCmPnM1–CD44BP-modified surfaces was assessed using static water contact anglemeasurements (Figure 2B). On the PCmPnM1-modified surfaces,the contact angle increased as the CMBMA fraction decreasedand the PGMA fraction increased, reflecting a reduction insurface hydration associated with a lower zwitterionic contentand a higher proportion of alkyne-bearing PGMA units. UponCD44BP(L3) conjugation, the surfaces became more hydrophilic,with the most pronounced decrease in the contact angle observedfor PC7P2M1–CD44BP and PC6P3M1–CD44BP. This increasein wettability is consistent with the intrinsic hydrophilicity ofthe peptide, which includes polar residues such as glutamine.Notably, the extent of the decrease in the contact angle dependedon the copolymer composition; PC8P1M1–CD44BP exhibited onlya minor change, whereas PC7P2M1–CD44BP exhibited a markedenhancement in wettability. Although a higher PGMA fractionprovides more alkyne handles for click conjugation, excessivePGMA content (>30%) may promote hydrophobic associationand dense packing of the alkyne side chains, which could reducethe accessibility of reactive groups and limit effective peptidepresentation. Collectively, these results identified PC7P2M1 ashaving the optimal balance between surface hydration and con-jugation efficiency for CD44BP attachment. To better capture thedynamic interfacial changes after droplet deposition, the time-dependent contact angles were measured (Figure S3). Allpolymer- and peptide-modified surfaces exhibited a monotonicdecrease in contact angle over time, characterized by an initialrapid drop, followed by a slower relaxation toward a quasi-equilibrium value. This behavior supports hydration/relaxationprocesses within the grafted layer (e.g., water uptake and chainrearrangement), in addition to macroscopic spreading, andqualitatively corroborates the wettability trends obtained fromthe static measurements. Subsequently, the wettability of thesurfaces functionalized with CD44BP(L0), CD44BP(L3), andCD44BP(L14) was evaluated using PC8P1M1, PC7P2M1, andPC6P3M1 (Figure 2C). Although differences in linker length wereexpected to modulate surface wettability, no significant differen-ces in the contact angles were detected among the three CD44BPvariants on any copolymer background. Therefore, the impact oflinker length was further considered, primarily in the context ofcell capture and adhesion performance.The film thicknesses of various PCmPnM1- and PCmPnM1–CD44BP(L3)-modified surfaces were evaluated using spectroscopicellipsometry (Figure 2D). The film thickness of the PC9P0M1-modified surface was measured to be 5.67± 0.83 nm. Our resea-rch group previously evaluated the film thicknesses of similarcopolymer-modified surfaces [10], demonstrating that coating withPC9P0M1 of �10,000 g/mol in average molecular weight produceda film thickness of approximately 3 nm. Because the copolymersused in this study possessed molecular weights ranging from�20,000 to 30,000 g/mol, the observed increase in the average filmthickness was consistent with expectations based on higher molec-ular weights. Among the PCmPnM1-modified surfaces, those coatedwith PC9P0M1, PC8P1M1, or PC7P2M1 showed no significantdifferences in film thickness. However, a noticeable reduction inthe film thickness was observed for PC6P3M1 and PC5P4M1 asthe PGMA content increased. This decrease can be attributed tothe lower average molecular weights of these copolymers andthe denser molecular packing induced by the hydrophobic interac-tions among the alkyne groups at higher PGMA compositions.Furthermore, the film thickness increased upon CD44BP(L3) mod-ification compared with the corresponding copolymer surfacesalone. The PC7P2M1-modified surface exhibited the most pro-nounced increase in film thickness following CD44BP(L3) conju-gation. This trend correlated well with the enhanced hydrophilicityobserved upon peptide attachment, suggesting successful and uni-form immobilization of CD44BP(L3) on the copolymer surface.The ATR–FTIR spectra in the range of 500–4000 cm−1 for thePC7P2M1- and PC7P2M1–CD44BP(L3)-modified surfaces areshown in Figure 2E, whereas the spectra for the other PCmPnM1-and PCmPnM1–CD44BP(L3)-modified surfaces are presented inFigure S4. Absorption bands were observed at �1200 and1700 cm−1 for all copolymer-modified surfaces, correspondingto the stretching vibrations of the O─C═O bonds in the meth-acrylate backbone and CMBMA side chains. In contrast, thespectra of the PCmPnM1–CD44BP(L3)-modified surfaces exhib-ited additional absorptions near 1100 cm−1, attributable to aro-matic C─H stretching vibrations of peptide amino acids, andaround 1650 cm−1, corresponding to the C═O stretching vibra-tions (amide I band) of the peptide bonds. These characteristicpeaks confirmed the successful immobilization of CD44BP onthe copolymer-coated surface. Furthermore, absorption peaksassigned to alkynes appeared between 2100 and 2300 cm−1 onPC6P3M1-modified surfaces. Although the intensities of thesepeaks decreased after CD44BP(L3) conjugation, they remainedclearly detectable, suggesting that the residual alkyne groups per-sisted at higher PGMA contents [16, 17]. This observation impliesthat the alkynyl side chains tend to aggregate within the polymermatrix owing to hydrophobic interactions [18–20], a finding con-sistent with the film thickness measurements described earlier.Scanning electron microscopy (SEM) was performed to examinewhether the polymer coating (PCmPnM1) and subsequent peptideconjugation (PCmPnM1–CD44BP) induced visually discerniblechanges in surface morphology (Figure S5). At a low magnifica-tion (100×; Figure S5A), the bright region at the top of eachimage corresponds to the cross-sectional edge of the glass sub-strate, confirming that the imaging plane and focus were prop-erly aligned, even in the absence of prominent surface structures.Such wide-field images were intentionally acquired because,for smooth surfaces, high-magnification images alone can app-ear similar and may obscure subtle differences in the overall flat-ness. All the substrates exhibited a uniformly featureless andsmooth appearance with no observable cracks, pores, or micron-scale aggregates. Moreover, neither local contrast variations norexposed substrate areas were detected on the PCmPnM1 orPCmPnM1–CD44BP surfaces, indicating the absence of increa-sed roughness or uncoated defective regions. Consistent withthe low-magnification observations, high-magnification images(5000×; Figure S5B) revealed no particulate deposits, phase-separated domains, or roughened structures after PCmPnM1 coat-ing or CD44BP conjugation. For all the copolymer compositionsexamined (PC8P1M1, PC7P2M1, and PC6P3M1), the correspondingPCmPnM1 and PCmPnM1–CD44BP surfaces displayed similarlysmooth morphologies, forming continuous defect-free thin films.ChemBioChem, 2026 9 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseThese SEM observations were in good agreement with thefilm thickness data obtained by spectroscopic ellipsometry(Figure 2D). All the PCmPnM1 and PCmPnM1–CD44BP coatingsexhibited film thicknesses below 10 nm, and the maximumincrease in thickness upon the conversion of PCmPnM1 toPCmPnM1–CD44BP was�2 nm. For such sub-10-nm ultrathinfilms, pronounced micro- to submicron-scale topographical cha-nges are not expected, which is consistent with the smooth,defect-free surface morphology observed by SEM. Therefore,the differences in wettability and cellular responses were primar-ily attributed to changes in surface chemistry, namely, copolymercomposition and peptide functionalization, rather than unin-tended variations in morphology or topography.The loading of CD44BP onto various PCmPnM1-coated surfaceswas evaluated using the microBCA method with silica particleswhose surfaces were modified with PCmPnM1 (Table 4). Theresults showed a general increase in peptide loading with incr-easing PGMA content. However, the amount of CD44BP(L3)immobilized on PC6P3M1-modified surfaces did not increase sig-nificantly. This may be attributed to the difficulty in introducingCD44BP(L3) when copolymers with high PGMA content allowalkynyl side chains to pack densely through hydrophobic inter-actions. This interpretation is consistent with the previouscontact angle, film thickness, and ATR–FTIR results, whichcollectively suggest that alkynyl groups are localized withinthe copolymer film owing to hydrophobic aggregation.Based on the measured CD44BP(L3) loading, the peptide densitywas theoretically sufficient to capture cells expressing CD44 anti-gens. Table 4 summarizes the peptide amount per unit area of thesilica particles and the corresponding specific surface area per pep-tide molecule. Given that the diameter of hMSCs is �10–15 μm, itwas estimated that each silica particle carries about 5× 107CD44BP molecules, based on its cross-sectional area. Althoughthe receptor densities vary, membrane receptors are generallypresent at �1× 106 molecules per cell [21]. Therefore, theCD44BP density on the silica particle surface was considered suffi-cient for effective interactions with CD44 antigens on the cellmembrane.3.4 | Suppression of Nonspecific Adsorption andSelective Cell Adhesion on CD44BP-FunctionalizedTernary Polymer SurfacesThe nonspecific adsorption of proteins on PCmPnM1- andPCmPnM1–CD44BP(L3)-modified surfaces was evaluated(Figure 3A). On all PCmPnM1-modified surfaces, the amount ofnonspecific protein adsorption was markedly reduced compared tothat on the bare glass surface, indicating that the copolymer contain-ing CMBMA effectively suppressed undesired protein adsorption.Moreover, a trend of increased protein adsorption was observed asthe CMBMA content decreased. This can be attributed to a reductionin the CMBMA fraction, which is responsible for the antibiofoulingproperties, combined with the increased hydrophobicity caused bythe higher PGMA content.Conversely, on PCmPnM1–CD44BP(L3)-modified surfaces, a rel-ative increase in nonspecific protein adsorption was observed,particularly on the PC6P3M1–CD44BP(L3) surface. This is pre-sumed to result from the lower actual loading of CD44BP(L3)compared to its theoretical value, leaving unreacted PGMA sidechains exposed on the surface, which may promote nonspecificprotein binding.Cell adhesion was evaluated on PCmPnM1- and PCmPnM1–CD44BP(L3)-modified surfaces. NIH3T3 cells, HEK293 cells,and hMSCs were seeded onto each surface, and after 24 h, boththe adhesive morphology and number of adherent cells were ana-lyzed (Figure 3B–E).On the PCmPnM1-modified surfaces, the cell adhesion of NIH3T3,HEK293, and hMSCs wasmarkedly suppressed, as shown in photo-graphs (b)–(e) of Figure 3C–E. These findings are consistent withthe protein adsorption results, indicating that the suppressionarises from the bioinert nature of the CMBMA units. Conversely,on PCmPnM1–CD44BP(L3)-modified surfaces, where bioactive pep-tides were introduced, the adhesion of NIH3T3 and HEK293 cellsremained low, whereas hMSCs exhibited a distinct attachment.Notably, on the PC8P1M1–CD44BP(L3) surface, the number of cap-tured cells was lower than that on the other CD44BP(L3)-carryingsurfaces. This reduction in cell captures was attributed to lowerCD44BP loading and higher CMBMA content of the copolymer.Although hMSC adhesion was not significantly different betw-een PC7P2M1 and CD44BP(L3) and PC6P3M1–CD44BP(L3), thereduced antibiofouling effect associated with the unreacted prop-argyl groups in the PGMA side chains suggests that the PC7P2M1–CD44BP(L3) surface offers optimal conditions for selective hMSCcapture. Collectively, these results demonstrated that the fabricatedPCmPnM1–CD44BP(L3) surface selectively captured hMSCs, withPC7P2M1 identified as the optimal copolymer composition forselective cell adhesion.The adhesion and selective capture of hMSCs were examinedusing CD44BP peptides with different linker lengths. HEK293cells and hMSCs were cultured on their respective surfaces,and their adhesion behavior was observed under a phase-contrastmicroscope 24 h after seeding (Figure 4).Quantitative analysis (Figure 4A) showed that HEK293 celladhesion was effectively suppressed, whereas hMSCs were selecti-vely captured on CD44BP-modified substrates. Among all the sam-ples, the PC7P2M1-modified substrate exhibited the highest degreeof hMSC adhesion when functionalized with CD44BP(L3).Conversely, the PC6P3M1-modified substrate showed reducedselectivity, as HEK293 cell adhesion was also observed. Thisreduced selectivity was attributed to the lower CMBMA content,which diminished the antibiofouling properties of the surface.Meanwhile, the PC8P1M1-modified surface, containing a higherproportion of CMBMA, exhibited strong suppression of cell adh-esion overall but captured fewer hMSCs owing to its lowerCD44BP loading.TABLE 4 | Amount of CD44BP(L3) bound to PCmPnM1 on the glasssurface.CopolymertypeAmount ofCD44BP(L3)(pmol/cm2)aOccupied area ofCD44BP (nm2/molecule)PC8P1M1 17 ± 6.2 10PC7P2M1 50 ± 5.4 3.3PC6P3M1 53 ± 6.2 3.2aThe amount of CD44BP bound to the copolymer surface was determinedusing the microBCA assay. Silica microparticles modified with PCmPnM1 wereemployed in this analysis.10 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseThe results obtained from the PC7P2M1- and PC6P3M1-modifiedsubstrates suggest that the amino acid sequence QQGWFP,which specifically interacts with hMSCs, is less exposed on sur-faces lacking a linker (L0). Conversely, when the linker wasexcessively long (L14), the hydrophobic residues within thelinker likely caused the hydrophilic glutamine residues at theCD44BP terminus to become embedded within the polymermatrix, thereby reducing the exposure of the bioactive site.Based on these findings, CD44BP(L3) was identified as theoptimal configuration for bioactive peptide presentation, andPC7P2M1–CD44BP(L3)-modified surfaces were selected for usein the subsequent cell-selective capture device fabrication.Furthermore, the PC7P2M1–CD44BP(L3)-modified surface, whichdemonstrated the highest selectivity for hMSC capture, exhibitedapproximately 50% reduction in hMSC adhesion compared tothe unmodified glass surface. Possible explanations for this reduc-tion include (1) insufficient peptide loading; (2) partial embeddingof immobilized CD44BP(L3) within the polymer chains, resultingin a lower effective concentration of accessible peptides comparedto total loading; and (3) heterogeneity within the seeded cell pop-ulation, including cells with reduced CD44 antigen expression.Regarding point (1), as discussed earlier regarding the antigendensity on the cell surface, CD44BP loading on the PC7P2M1–CD44BP-modified surface (approximately 50 pmol/cm2) was esti-mated to be sufficient for cell capture; therefore, it is unlikely toaccount for the observed reduction in adhesion. For point (2), nomeasurement technique is currently available for directly evalu-ating the structural state of CD44BP within the copolymermatrix; thus, this factor cannot be conclusively confirmed. If aportion of the immobilized CD44BP were buried within thecopolymer and rendered nonfunctional, increasing the total pep-tide loading would have led to an increase in cell adhesion.However, the number of adhered hMSCs on the PC6P3M1–CD44BP-modified surface, which possessed a higher peptideloading ratio, was nearly identical to that on the PC7P2M1–CD44BP-modified surface, with only approximately half of theseeded cells adhering. This observation suggests that the maxi-mum feasible amount of CD44BP had already been immobilizedon the PC7P2M1–CD44BP-modified surface and, consequently,that peptide loading was not responsible for the reduction inadhesion. Therefore, point (3) was considered the most plausibleexplanation for the decreased number of adhered cells. Previousstudies have reported that CD44 antigen expression diminishesin hMSCs during differentiation or due to degradation overextended culture periods [22–24]. Consequently, it is likely thatcells exhibiting reduced CD44 expression were not captured bythe PC7P2M1–CD44BP-modified surface, resulting in fewerFIGURE 3 | Evaluation of nonspecific protein adsorption and cell adhesion on copolymer- and CD44BP(L3)-modified surfaces. (A) Amount ofnonspecifically adsorbed proteins on unmodified glass and on surfaces modified with PCmPnM1 without peptide (black bars) or PCmPnM1–CD44BP(L3) (gray bars). Data are presented as mean ± SD (n= 10). Statistical significance was evaluated using one-way ANOVA followed byTukey’s HSD test (p< 0.001). (B) Quantitative analysis of cell adhesion on glass, on surfaces modified with PCmPnM1 without peptide (a), and on surfacesmodified with PCmPnM1–CD44BP(L3) (b) (n= 3). NIH3T3 (light gray), HEK293 (gray), and hMSCs (black) were seeded at 2.0 × 104 cells/cm2 and culturedfor 1 day. Data are shown as mean ± SD. Statistical significance among all samples was assessed using one-way ANOVA followed by Tukey’s HSD test(p< 0.001). Bars labeled with different letters represent statistically distinct groups. (C–E) Representative microscopy images of (C) NIH3T3 cells,(D) HEK293 cells, and hMSCs. (E) adhered to surfaces coated with various copolymers: (a) glass, (b) PC9P0M1, (c) PC8P1M1, (d) PC7P2M1,(e) PC6P3M1, (f ) PC8P1M1–CD44BP(L3), (g) PC7P2M1–CD44BP(L3), and (h) PC6P3M1–CD44BP(L3). Cells were seeded at 2.0× 104 cells/cm2 and culturedfor 1 day. Scale bar: 500 μm.ChemBioChem, 2026 11 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licenseadhered cells. This hypothesis will be further examined in con-junction with flow cytometry data from capture experimentsusing cell separation columns.3.5 | Selective Cell Separation Using CD44BP-Functionalized Polymer Columns and PostsortingEvaluation of hMSC FunctionalityBased on these findings, the PC7P2M1–CD44BP-modified surfacewas identified as the optimal configuration for capturing hMSCsexpressing CD44 antigens. Therefore, a cell separation columnwas fabricated by packing silica microparticles modified withPC7P2M1–CD44BP on their surfaces, and the selective separationof hMSCs was investigated. A model cell suspension containing amixture of hMSCs and HEK293 cells (lacking CD44 antigens)was used to assess separation performance in detail. The eluatewas collected in 1mL fractions, and both cell types and theirnumbers were quantitatively evaluated.Silica microparticles with diameters ranging from 250 to 350 μmwere selected as the optimal packing material. Preliminaryexperiments (data not shown) demonstrated that the use ofporous silica microparticles allowed protein infiltration, leadingto reduced biocompatibility. The 250–350 μm particle size rangewas found to provide an appropriate balance between contactefficiency and flow rate. Specifically, columns prepared with75–180 μm particles required �2 h for cell separation due toextremely slow elution, while those prepared with 600–850 μmparticles exhibited poor separation performance because ofexcessively rapid elution (within 10 min), which limited cell–surface interaction (data not shown).In a preliminary examination using a cell separation columnpacked with nonporous silica beads of 250–350 μm in diameter,it was observed that HEK293 cells began to elute �6min afterthe start of the experiment, with about 80% of the loadedHEK293 cells eluting within 15min. In contrast, hMSCs beganto elute �8min after initiation, and even 25min after collection,a considerable proportion of hMSCs appeared to remain within thecolumn. Upon subsequent addition of 5mM EDTA/PBS solution,nearly all the remaining hMSCs were recovered. The fractioneluted with the EDTA solution also contained a small proportionof HEK293 cells. These preliminary findings can be interpreted asfollows. The delayed elution of hMSCs is attributed to their specificinteraction with the PC7P2M1–CD44BP-modified silica microparti-cle surface. The substantial recovery of hMSCs upon treatmentwith 5mM EDTA/PBS likely reflects the strong multivalent inter-actions between CD44 antigens on hMSCs and CD44BP on thepolymer-coated surface. Furthermore, the presence of HEK293cells in the EDTA-eluted fraction may be explained by the partialblockage of flow paths caused by the hMSCs bound at multiplecontact points, as illustrated in Figure S6, which may have physi-cally trapped the HEK293 cells within the column.FIGURE 4 | Selective adhesion of hMSCs and HEK293 Cells on copolymer surfaces functionalized with CD44BP of different linker lengths.(A) Quantitative analysis of HEK293 (dark gray bars) and hMSC (black bars) adhesion on surfaces functionalized with CD44BP peptides of differentlinker lengths (L0, L3, L14) (n= 3). Cells were seeded at 2.0 × 104 cells/cm2 and cultured for 1 day. For each copolymer composition, the sample labelsare as follows: (a) nonpeptide-modified surface, (b) CD44BP(L0), (c) CD44BP(L3), and (d) CD44BP(L14). Data are presented as mean ± SD (n= 3).Statistical significance was evaluated using one-way ANOVA followed by Tukey’s HSD test (p< 0.001). Bars labeled with different letters representstatistically distinct groups. (B,C) (B) Representative microscopy images of HEK293 cells and (C) hMSCs adhered to surfaces coated with (a–c) PC8P1M1,(d–f ) PC7P2M1, and (g–i) PC6P3M1, functionalized with CD44BP peptides of different linker lengths: (a,d,g) L0, (b,e,h) L3, and (c,f,i) L14. Scale bar: 500 μm.12 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseTherefore, two columns were prepared for the sequential separa-tion experiments. The first column was operated under the sameconditions used in the preliminary examination. Twenty minutesafter the start of the experiment, all remaining cells retainedwithin the column were recovered by adding 5mM EDTA/PBS solution. The recovered cells were then washed to removethe residual EDTA and subsequently reintroduced into the col-umn for a second separation (corresponding to the graph x-axisrange of 31–50.5 min). Finally, 5 mM ethylenediaminetetraaceticacid/phosphate-buffered saline (EDTA/PBS) was added to elutethe remaining bound cells. As shown in Figure 5A, �94% of theHEK293 cells were eluted within 15min from the start of theexperiment in the first column, whereas 67% of the loadedhMSCs were contained in the fractions collected between 15and 25min and in those eluted with 5mM EDTA/PBS solution.When the EDTA-recovered fraction, comprising 51% of the ini-tially added hMSCs and 6% of HEK293 cells, was reapplied to thecolumn, all HEK293 cells eluted within 35 min, and only hMSCswere recovered thereafter.Although this procedure was not further optimized in the pres-ent study, it can be inferred that if high-purity hMSCs arerequired, a large quantity of highly pure hMSCs can beobtained by collecting all eluates recovered 15 min after thefirst separation and performing an additional round of columnseparation.The eluate was subsequently divided into groups according tocollection time intervals (5–15 min, 16–25min, 26–50.5 min,and the fraction recovered using 5mM EDTA/PBS solution), andeach group was analyzed by flow cytometry (Figure 5B–G).As a reference, flow cytometric analyses of control hMSCs andHEK293 cells were performed (Figure 5B,C). In the flow cytom-etry plots, the vertical axis represents the fluorescence intensitycorresponding to CD105 antigen expression, whereas theFIGURE 5 | Cell separation performance, CD44BP-mediated binding specificity, and postsorting differentiation capacity of hMSCs using a PC7P2M1–CD44BP(L3)-based sorting column. (A) Fractionation profile obtained from amixed suspension of hMSCs and HEK293 cells using a column packed withPC7P2M1-CD44BP(L3)-modified silica beads. The graph shows the number of cells collected in each time fraction based on three independent experi-ments (n= 3). Because the fractions represent continuous, time-dependent outputs from the column, no statistical analysis was applied to this dataset.(B–G) Flow cytometry analysis of cells before and after column separation. (B,C) Flow cytometry profiles of untreated control (B) hMSCs and(C) HEK293 cells. (D) Fraction collected between 5–15min. (E) Fraction collected between 16–25min. (F) Fraction collected between 26–50.5 min.(G) Cells collected after 50.5 min. (H) Fractionation profile obtained when hMSCs were preincubated with excess free CD44BP prior to loading ontothe PC7P2M1-CD44BP(L3)-modified column. Preincubation with the free peptide competitively inhibited CD44BP-mediated binding, resulting in aloss of selective retention within the column. Data represent three independent experiments (n= 3). Because the fractions are continuous andtime-dependent, no statistical analysis was performed. (I) Evaluation of the multilineage differentiation capacity of hMSCs before and after sorting:(a) Fluorescence micrographs of ALP-stained osteoblasts. (b) Images of Oil Red O–stained adipocytes.ChemBioChem, 2026 13 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehorizontal axis represents the fluorescence intensity correspond-ing to CD44 antigen expression.HEK293 cells accounted for 98.7% of the cell population in thefractions collected between 5 and 15min (Figure 5D). In the frac-tion collected between 16 and 25min, 61.6% of the cells werehMSCs coexpressing both CD44 and CD105 antigens, 38.1% wereHEK293 cells, and only 0.1% expressed either CD44 or CD105alone (Figure 5E). In the fraction collected between 26–50.5 min, 98.3% of the cells were hMSCs (Figure 5F), and inthe EDTA-eluted fraction, the hMSC purity reached 99.6%.These results are consistent with those of the quantitative analy-sis presented in the previous section. Notably, as the elution pro-gressed to later fractions, the double-positive cell populationexpressing both CD44 and CD105 exhibited a shift toward higherfluorescence intensities (upward and rightward on the scatterplots). This trend suggests that the hMSCs eluted in the laterfractions displayed higher expression levels of CD44 and CD105antigens. Consequently, these results indicated that the cells col-lected following repeated column separation possessed enhancedexpression of stemness markers, implying a greater degree ofmaintenance of their undifferentiated state.In the previous section (specific cell culture on the substrate modi-fied with CD44BP-anchored ternary polymer), we discussed whyonly approximately half of the seeded hMSCs adhered to thePC7P2M1–CD44BP-modified surface. Based on the flow cytometryresults, it can be inferred that cells with low CD44 antigen expres-sion failed to adhere to the modified surface. In other words, thisseparation system not only enables the isolation of specific celltypes but also allows the selective capture of hMSCs with highCD44 expression, indicative of a highly undifferentiated state.Since maintaining an undifferentiated state is crucial for efficientlineage-specific differentiation, a system capable of isolating undif-ferentiated hMSCs provides a significant functional advantage overconventional separation techniques.Considering the above findings, it was demonstrated that byintroducing a cell suspension into this separation column, col-lecting the EDTA-eluted fraction after 15 min, removing theEDTA by centrifugation (�10 min), and reapplying the cellsfor a second separation, high-purity hMSCs could be isolatedwithin 40–50 min. Therefore, this column device represents anefficient, rapid, and low-damage method for isolating hMSCs,offering a substantial improvement over currently available cellseparation approaches.To confirm that the separation of hMSCs by the PC7P2M1-CD44BP-modified column was specifically mediated by the interactionbetween the CD44 antigen on the cell surface and CD44BP immo-bilized on the silica particle surface, a competitive inhibition exper-iment was performed by preincubating the cells with free CD44BPbefore column loading (Figure 5H). HEK293 cells (gray bars) andhMSCs (black bars) were eluted almost simultaneously, indicatingthat free CD44BP bound to CD44 antigens on the surface ofhMSCs, thereby preventing their binding to CD44BP immobilizedon the column. This finding demonstrates that the cell separationmechanism of the PC7P2M1-CD44BP-modified column was pri-marily governed by specific CD44–CD44BP interactions.The proliferative capacity and maintenance of multipotency of thehMSCs isolated using the constructed separation column (cells col-lected after 40min) were evaluated. Proliferative capacity wasassessed by determining doubling time, and multipotency wasexamined by inducing differentiation into osteoblasts andadipocytes.The doubling time of hMSCs before separation was 40± 2.0 h,whereas that of cells after separation was 42± 1.5 h. These valuesare consistent with previously reported doubling times forhMSCs [25–29], indicating that the separation process usingthe column does not affect proliferative capacity.To evaluate the differentiation potential of hMSCs before and afterseparation, the expression of alkaline phosphatase (ALP), a markerof osteogenic differentiation, was analyzed by fluorescent immu-nostaining, while lipid droplet formation during adipogenic differ-entiation was assessed by Oil Red O staining. ALP expressionpatterns, including the stained area and fluorescence intensity,showed no difference between the pre- and postseparation cells(Figure 5I,a). Similarly, the formation of lipid droplets by adipo-cytes showed no observable differences before and after separation(Figure 5I,b). These findings demonstrated that the osteogenic andadipogenic differentiation abilities of hMSCs were retained follow-ing column-based separation, confirming that the separation pro-cedure did not adversely affect cellular functions.4 | ConclusionA surface functionalized with PCmPnM1–CD44BP was successfullyfabricated, enabling the suppression of nonspecific protein adsorp-tion and nonspecific cell adhesion, while permitting the selectivecapture of hMSCs, with the PC7P2M1–CD44BP formulation exhib-iting the most favorable performance. Based on these optimi-zed surface properties, a cell separation column packed withPC7P2M1-CD44BP-modified silica microparticles was constructed,and efficient purification of hMSCs from mixed cell suspensionswas demonstrated. Flow cytometry analysis further revealed thathMSCs eluted in the later fractions exhibited high expression ofCD44 and CD105, indicating that the separation efficiency was cor-related with CD44 expression levels. Notably, the entire separationprocess was completed within 35min, and postseparation, hMSCsretained both their proliferative activity and multilineage differen-tiation capacity. Collectively, these findings demonstrate that thispeptide-functionalized antifouling polymer system enables therapid, selective, and minimally invasive isolation of high-qualityhMSCs, offering a promising platform for stem cell–based regener-ative medicine.Author ContributionsTadashi Nakaji-Hirabayashi: conceptualization (lead), data curation(equal), funding acquisition (lead), investigation (equal), methodology(lead), project administration (lead), resources (lead), supervision (lead),visualization (equal), writing – original draft (lead), writing – review andediting (lead).Moe Kato: conceptualization (equal), data curation (lead),formal analysis (lead), investigation (lead), methodology (lead), visualiza-tion (lead), writing – original draft (lead). Kazuaki Matsumura: resour-ces (equal), validation (equal), writing – review and editing (supporting).Chiaki Yoshikawa: funding acquisition (equal), methodology (equal),resources (equal), supervision (equal), writing – review and editing(supporting). Yuki Usui: funding acquisition (equal), methodology(supporting), resources (supporting). Takahiro Kishioka: fundingacquisition (equal), methodology (supporting), resources (supporting).Taito Nishino: funding acquisition (equal), methodology (supporting),resources (supporting).14 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons LicenseAcknowledgmentsThe authors gratefully acknowledge the support of a Grant-in-Aid forScientific Research from the Japan Society for the Promotion ofScience (JSPS) (Grant Numbers 15H05353, 17KK0130, 18K19907, and22H3951) and the NIMS Joint Hub Program. We also extend our sincereappreciation to the Nissan Chemical Corporation for the continuous sup-port of this research. Part of this work was conducted using NIMS molec-ular and material synthesis platforms. This study was supported by theJST SPRING (Grant Number JPMJSP2145).FundingThis Study was supported by the Japan Society for the Promotion of Science(JSPS), Grant-in-Aid for Scientific Research (Grant 15H05353, 17KK0130,18K19907, and 22H3951) and JST SPRING (Grant JPMJSP2145).Conflicts of InterestUsui, Kishioka, and Nishino are employes of Nissan Chemical Corporation.The authors declare no conflicts of interest. Collaboration with NissanChemical Corporation did not influence the objectivity of the study or inter-pretation of the results.Data Availability StatementAll data supporting the findings of this study are available in the articleand Supplementary Information files.References1. C. M. Madl, S. C. Heilshorn, and H. M. Blau, “Bioengineering Strategiesto Accelerate Stem Cell Therapeutics,” Nature 557, no. 7705 (2018): 335–342, https://doi.org/10.1038/s41586-018-0089-z.2. T. G. Fernandes, C. A. V. Rodrigues, M. M. Diogo, and J. M. S. Cabral,“Stem Cell Bioprocessing for Regenerative Medicine,” Journal ofChemical Technology and Biotechnology 89, no. 1 (2014): 34–47,https://doi.org/10.1002/jctb.4189.3. N. K. Satija, V. K. Singh, Y. K. Verma, et al., “Mesenchymal Stem Cell-Based Therapy: A New Paradigm in Regenerative Medicine,” Journal ofCellular and Molecular Medicine 13, no. 11-12 (2009): 4385–4402, https://doi.org/10.1111/j.1582-4934.2009.00857.x.4. A. Nicodemou and L. Danisovic, “Mesenchymal Stromal/Stem CellSeparation Methods: Concise Review,” Cell and Tissue Banking 18, no.4 (2017): 443–460, https://doi.org/10.1007/s10561-017-9658-x.5. T. Squillaro, G. Peluso, and U. Galderisi, “Clinical Trials withMesenchymal Stem Cells: An Update,” Cell Transplantation 25, no. 5(2016): 829–848, https://doi.org/10.3727/096368915X689622.6. M. J. Tomlinson, S. Tomlinson, X. B. Yang, and J. Kirkham, “CellSeparation: Terminology and Practical Considerations,” Journal ofTissue Engineering 4 (2012): 2041731412472690, https://doi.org/10.1177/2041731412472690.7. A. Bosio, V. Huppert, S. Donath, P. Hennemann, M. Malchow, andU. A. Heinlein, “Isolation and Enrichment of Stem Cells,” Advances inBiochemical Engineering/Biotechnology 114 (2009): 23–72, https://doi.org/10.1007/10_2008_38.8. H. Pertoft, “Fractionation of Cells and Subcellular Particles withPercoll,” Journal of Biochemical and Biophysical Methods 44, no. 1-2(2000): 1–30, https://doi.org/10.1016/S0165-022X(00)00066-X.9. M. Almeida, A. C. García-Montero, and A. Orfao, “Cell Purification: ANew Challenge for Biobanks,” Pathobiology 81, no. 5-6 (2014): 261–275,https://doi.org/10.1159/000358306.10. M. Nishida, T. Nakaji-Hirabayashi, Kitano Hiromi, et al.,“Optimization of the Composition of Zwitterionic PC7P2M1s for theEasy Construction of Bio-Inactive Surfaces,” Journal of BiomedicalMaterials Research. Part A 104a (2016): 2029–2036, https://doi.org/10.1002/jbm.a.35737.11. Y. Zhong, M. Yan, S. Wu, A. Su, and J. Pu, “Self-Renewing ZwitterionicSilicone Hybrid Coatings with Multi-Modal Bacterial Resistance forSuperior Anti-Biofouling Performance,” Chemical Engineering Journal524 (2025): 169828, https://doi.org/10.1016/j.cej.2025.169828.12. T. Nakaji-Hirabayashi, K. Fujimoto, C. Yoshikawa, and H. Kitano,“Functional Surfaces for Efficient Differentiation of Neural Stem/Progenitor Cells into Dopaminergic Neurons,” Journal of BiomedicalMaterials Research. Part A 107, no. 4 (2019): 860–871, https://doi.org/10.1002/jbm.a.36602.13. C. Yoshikawa, T. Nakaji-Hirabayashi, N. Nishijima, et al., “Ultra-LowFouling Photocrosslinked Coatings for the Selective Capture of CellsExpressing CD44,” Materials Science & Engineering. C, Materials forBiological Applications 120 (2021): 111630, https://doi.org/10.1016/j.msec.2020.111630.14. M. F. Pittenger, A. M. Mackay, S. C. Beck, et al., “MultilineagePotential of Adult Human Mesenchymal Stem Cells,” Science 284, no.5411 (1999): 143–147, https://doi.org/10.1126/science.284.5411.143.15. M. Yan, C. Zhang, G. Sun, et al., “Self-Adjusting Lubricant-InfusedPorous Hydrophobic Sticky Surfaces: Programmable Time Delay Switchfor Smart Control of the Drop’s Slide,” ACS Applied Materials &Interfaces 11, no. 46 (2019): 43861–43688, https://doi.org/10.1021/acsami.9b14064.16. P. K. Smith, R. I. Krohn, G. T. Hermanson, et al., “Measurement ofProtein Using Bicinchoninic Acid,” Analytical Biochemistry 150, no. 1(1985): 76–85, https://doi.org/10.1016/0003-2697(85)90442-7.17. R. Townend, T. F. Kumosinski, S. N. Timasheff, G. D. Fasman, andB. Davidson, “The Circular Dichroism of the Beta Structure of Poly-L-Lysine,” Biochemical and Biophysical Research Communications 23, no.2 (1966): 163–169, https://doi.org/10.1016/0006-291X(66)90522-5.18. B. Alberts, A. Johnson, and J. Wewis, et al., Molecular Biology of theCell, 4th ed. (Garland Science, 2002), 602–604.19. Y. Luo, H. Shen, Y. Fang, et al., “Enhanced Proliferation andOsteogenic Differentiation of Mesenchymal Stem Cells on GrapheneOxide-Incorporated Electrospun Poly(Lactic-Co-Glycolic Acid)Nanofibrous Mats,” ACS Applied Materials & Interfaces 7, no. 11(2015): 6331–6339, https://doi.org/10.1021/acsami.5b00862.20. M. Köllmer, V. Keskar, T. G. Hauk, J. M. Collins, B. Russell, andR. A. Gemeinhart, “Stem Cell-Derived Extracellular Matrix EnablesSurvival and Multilineage Differentiation within SuperporousHydrogels,” Biomacromolecules 13, no. 4 (2012): 963–973, https://doi.org/10.1021/bm300332w.21. S. Fickert, U. Schröter-Bobsin, A. F. Gross, et al., “HumanMesenchymal Stem Cell Proliferation and Osteogenic Differentiationduring Long-Term Ex Vivo Cultivation Is Not Age Dependent,”Journal of Bone and Mineral Metabolism 29, no. 2 (2011): 224–235,https://doi.org/10.1007/s00774-010-0215-y.22. Y. Y. Lipsitz, N. E. Timmins, and P. W. Zandstra, “Quality CellTherapy Manufacturing by Design,” Nature Biotechnology 34, no. 4(2016): 393–400, https://doi.org/10.1038/nbt.3525.23. B. Lawrenz, H. Schiller, E. Willbold, M. Ruediger, A. Muhs, andS. Esser, “Highly Sensitive Biosafety Model for Stem-Cell-DerivedGrafts,” Cytotherapy 6, no. 3 (2004): 212–222, https://doi.org/10.1080/14653240410006031.24. J. A. Heslop, T. G. Hammond, I. Santeramo, et al., “Concise Review:Workshop Review: Understanding and Assessing the Risks of Stem Cell-Based Therapies,” Stem Cells Translational Medicine 4, no. 4 (2015): 389–400, https://doi.org/10.5966/sctm.2014-0110.25. T. Fujikawa, S. H. Oh, L. Pi, H. M. Hatch, T. Shupe, and B. E. Petersen,“Teratoma Formation Leads to Failure of Treatment for Type I DiabetesUsing Embryonic Stem Cell-Derived Insulin-Producing Cells,” AmericanChemBioChem, 2026 15 of 16 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttps://doi.org/10.1038/s41586-018-0089-zhttps://doi.org/10.1002/jctb.4189https://doi.org/10.1111/j.1582-4934.2009.00857.xhttps://doi.org/10.1111/j.1582-4934.2009.00857.xhttps://doi.org/10.1007/s10561-017-9658-xhttps://doi.org/10.3727/096368915X689622https://doi.org/10.1177/2041731412472690https://doi.org/10.1177/2041731412472690https://doi.org/10.1007/10_2008_38https://doi.org/10.1007/10_2008_38https://doi.org/10.1016/S0165-022X(00)00066-Xhttps://doi.org/10.1159/000358306https://doi.org/10.1002/jbm.a.35737https://doi.org/10.1002/jbm.a.35737https://doi.org/10.1016/j.cej.2025.169828https://doi.org/10.1002/jbm.a.36602https://doi.org/10.1002/jbm.a.36602https://doi.org/10.1016/j.msec.2020.111630https://doi.org/10.1016/j.msec.2020.111630https://doi.org/10.1126/science.284.5411.143https://doi.org/10.1021/acsami.9b14064https://doi.org/10.1021/acsami.9b14064https://doi.org/10.1016/0003-2697(85)90442-7https://doi.org/10.1016/0006-291X(66)90522-5https://doi.org/10.1021/acsami.5b00862https://doi.org/10.1021/bm300332whttps://doi.org/10.1021/bm300332whttps://doi.org/10.1007/s00774-010-0215-yhttps://doi.org/10.1038/nbt.3525https://doi.org/10.1080/14653240410006031https://doi.org/10.1080/14653240410006031https://doi.org/10.5966/sctm.2014-0110Journal of Pathology 166, no. 6 (2005): 1781–1791, https://doi.org/10.1016/S0002-9440(10)62488-1.26. M. A. Ghoneim, A. F. Refaie, B. L. Elbassiouny, M. M. Gabr, andM. M. Zakaria, “From Mesenchymal Stromal/Stem Cells to Insulin-Producing Cells: Progress and Challenges,” Stem Cell Reviews andReports 16, no. 6 (2020): 1156–1172, https://doi.org/10.1007/s12015-020-10036-3.27. C. M. Digirolamo, D. Stokes, D. Colter, D. G. Phinney, R. Class, andD. J. Prockop, “Propagation and Senescence of Human Marrow StromalCells in Culture: A Simple Colony-Forming Assay Identifies Samples withthe Greatest Potential to Propagate and Differentiate,” British Journal ofHaematology 107, no. 2 (1999): 275–281, https://doi.org/10.1046/j.1365-2141.1999.01715.x.28. W. L. Grayson, F. Zhao, B. Bunnell, and T. Ma, “Hypoxia EnhancesProliferation and Tissue Formation of HumanMesenchymal Stem Cells,”Biochemical and Biophysical Research Communications 358, no. 3 (2007):948–953, https://doi.org/10.1016/j.bbrc.2007.05.054.29. P. V. Pham, N. B. Vu, V. M. Pham, et al., “Good ManufacturingPractice-Compliant Isolation and Culture of Human Umbilical CordBlood-Derived Mesenchymal Stem Cells,” Journal of TranslationalMedicine 12, no. 56 (2014): 56, https://doi.org/10.1186/1479-5876-12-56.Supporting InformationAdditional supporting information can be found online in the SupportingInformation Section. Supporting Scheme S1: Schematic illustration ofthe surface coating process of a glass substrate or silica beads with theCD44BP–functionalized ternary copolymer. Supporting Scheme S2:Schematic illustration of silica beads coated with PCmPnM1–CD44BP.Supporting Fig. S1: (A) The cell separation column system packed withPC7P2M1–CD44BP–modified silica beads inside a syringe. (B) The proce-dure for the cell separation experiments. Supporting Fig. S2: LC–MS/MS spectra of (A) CD44BP(L3)–Aha and (B) CD44BP(L3). SupportingFig. S3: Time-dependent water contact angle measurements recordedevery 5 s over a 600 s period. Data represent the mean values obtainedfrom five independent measurements. Supporting Fig. S4: IR spectra ofsurfaces modified with PCmPnM1 (black line) and PCmPnM1–CD44BP(L3)(gray line): (A) PC8P1M1, (B) PC6P3M1, and (C) PC5P4M1. SupportingFig. S5: (A,B) SEM images of various polymer-modified substrates.(A) Images acquired at 100×magnification (scale bar: 500 μm). The brightregion at the top corresponds to the cross-sectional edge of the glass sub-strate, indicating that the focus is correctly adjusted due to the absence ofsurface structures. (B) Images acquired at 5000×magnification (scale bar:10 μm). The examined surfaces were as follows: (a) glass, (b) PC8P1M1,(c) PC8P1M1–CD44BP(L3), (d) PC7P2M1, (e) PC7P2M1–CD44BP(L3),(f ) PC6P3M1, (g) PC6P3M1–CD44BP(L3). Supporting Fig. S6: Microscopicimage of cells within the copolymer–CD44BP–modified column. hMSCsinteract with CD44BP on the silica bead surface, leading to partial blockageof the flow path by adhering cells.16 of 16 ChemBioChem, 2026 14397633, 2026, 3, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cbic.202500822 by National Institute For, Wiley Online Library on [25/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttps://doi.org/10.1016/S0002-9440(10)62488-1https://doi.org/10.1016/S0002-9440(10)62488-1https://doi.org/10.1007/s12015-020-10036-3https://doi.org/10.1007/s12015-020-10036-3https://doi.org/10.1046/j.1365-2141.1999.01715.xhttps://doi.org/10.1046/j.1365-2141.1999.01715.xhttps://doi.org/10.1016/j.bbrc.2007.05.054https://doi.org/10.1186/1479-5876-12-56 CD44-Binding Peptide-Functionalized Antibiofouling Polymer Surface for High-Performance Separation of Human Mesenchymal Stromal Cells 1. Introduction 2. Materials and Methods 2.1. Synthesis of Ternary Copolymer: Poly(carboxymethylbetaine-co-propargyl methacrylate-co-3-methacryloyloxypropyl trimethoxysilane) 2.2. Synthesis of CD44-Binding Peptide 2.3. Construction of Surface Modified With Ternary Copolymer Immobilized With CD44BP 2.4. Characterization of PCmPnM1 2.5. Characterization of CD44BP 2.6. Wettability of Ternary Polymer- and CD44BP-Carried Ternary Polymer-Modified Surfaces 2.7. Characterization of Ternary Polymer- and CD44BP-Carried Ternary Polymer-Modified Surface 2.8. Antibiofouling Property of Ternary Polymer- and CD44BP-Carried Ternary Polymer-Modified Surfaces 2.9. Selectivity of Cell Capture on CD44BP-Carried Ternary Polymer-Modified Surfaces 2.10. Construction of Cell-Separating Column 2.11. Fluorescent Labeling of Cells for Cell Separation Analysis 2.12. Specific Cell Separation 2.13. Analysis of Separated Cells Using Flow Cytometry 2.14. Evaluation of Multipotency of Separated Cells 2.15. Statistical Analysis 3. Results and Discussion 3.1. Characterization of Ternary Polymer 3.2. Characterization of the CD44 Binding Peptide 3.3. Surface Characterization of the CD44BP-Carried Ternary Polymer Surface 3.4. Suppression of Nonspecific Adsorption and Selective Cell Adhesion on CD44BP-Functionalized Ternary Polymer Surfaces 3.5. Selective Cell Separation Using CD44BP-Functionalized Polymer Columns and Postsorting Evaluation of hMSC Functionality 4. Conclusion