Single-cell mechanoprofiling of aging and cancer using AFM probes coated with RGD ultrashort peptide nanofibers

Cell adhesion to the extracellular matrix (ECM) is fundamental in both cancer progression and aging. Many ECM proteins present on their surface several biofunctional adhesion cues, made of short peptide sequences that are recognized by adhesion molecules on the cell membrane. Incorporating these biofunctional peptides into hydrogels is a common strategy to promote cell adhesion to synthetic biomaterials, both for 2D and 3D cell cultures. Here, we combine atomic force microscopy (AFM) with ultrashort biofunctional peptide hydrogels to investigate the mechanobiology of single-cell adhesion in real time across three settings: young vs. senescent human dermal fibroblasts, human mesenchymal stromal cells (MSCs) from healthy donors and leukemia patients, and human neuroblastoma cells. When AFM probes were coated with self-assembling peptide hydrogels presenting cell adhesion motifs, AFM force spectroscopy profiles of the cells to biofunctional motif coated probes yield discriminating results. To our knowledge, this is the first application of ultrashort biofunctional peptide hydrogels coated on AFM tips enabling mechanobiological analysis. This work showcases a single-cell adhesion profiling tool using AFM in combination with ultrashort peptide hydrogels that simulate distinct features of the extra-cellular matrix. Our method not only provides a simple, controlled route to adhesion behavior studies and informed hydrogel sequence design decisions in label-free, physiological environment, but could also discriminate adhesion propensity of healthy, aging, and cancerous cell states, holding promise for bioengineering and regenerative medicine applications. By extending this concept, the AFM-peptide hydrogel approach offers a novel, label-free way to distinguish senescent cells and cancer-altered cells from their healthy counterparts, with potential in diagnostics and personalized therapeutics.

Global cancer incidence continues to rise, with an estimated 20 million new cases and 9.7 million deaths worldwide in 2022, making cancer one of the leading causes of morbidity and mortality across regions and income groups (Sung et al., 2021). Importantly, cancer is strongly age-associated: more than 60% of new cancer diagnoses and 70% of cancer deaths occur in individuals aged 65 years or older, reflecting the cumulative effects of genomic instability, chronic inflammation, and age-related tissue microenvironmental changes that promote malignant transformation (Campisi, 2013). As global life expectancy increases, the burden of age-related cancers is projected to grow substantially, underscoring the need for research that integrates cancer biology with the biology of aging.

Extracellular matrix (ECM) and cell adhesion molecules (CAMs) are proteins, integral to cellular functions, and their alterations are significant in both cancer and aging (Chastney et al., 2024). Proteins, including integrins, cadherins, fibronectin and laminin, mediate interactions between cells and the ECM, influencing processes such as migration, proliferation, and differentiation (Kanchanawong and Calderwood, 2023; Bonnans et al., 2014).

Particularly integrins are key receptors that mediate cell adhesion by interacting with specific adhesion motifs in the extracellular matrix (ECM), such as RGD, IKVAV, YIGSR, and GFOGER, that are biofunctional motifs of collagen, fibronectin, vitronectin, and laminin (Huettner et al., 2018). These interactions are critical in both normal physiological processes and pathological conditions, where they impact tissue integrity and disease progression (Balion et al., 2020).

In cancer, cell adhesion molecules like integrins play pivotal roles. In cancer cells, integrins can become downregulated, facilitating their detachment and movement through the ECM, which promotes metastasis (Liu et al., 2023; Hamidi and Ivaska, 2018; Winkler et al., 2020). On the other end, integrin overexpression can enhance cancer cell invasion of more distant tissue (Cooper and Giancotti, 2019). Additionally, ECM proteins such as fibronectin and laminin are involved in angiogenesis. The formation of new blood vessels can be beneficial but can also support unnecessary tumor growth. Cancer cells often secrete factors that alter the ECM’s composition, thereby promoting angiogenesis and supporting tumor expansion (Libby et al., 2024). The tumor microenvironment, shaped by ECM changes, provides structural and biochemical support for cancer progression. Altered ECM stiffness and composition can impact tumor cell behavior, influence the efficacy of therapies, and contribute to cancer progression (Mai et al., 2024). Immunosurveillance is also affected by integrins, where immune cells are critically impacted in cancer disease (Guenther et al., 2019).

Aging, similarly, is associated with significant changes in the composition of the ECM including adhesion molecules (Campisi et al., 2007; Ogrodnik and Gladyshev, 2023). With age, the ECM undergoes alterations including increased deposition of collagen and decreased levels of other proteins (Phillip et al., 2015). These changes can impair tissue structure and function, leading to age-related diseases. Integrins also show age-related functional changes, affecting cell adhesion and migration, which can impact wound healing and tissue repair (López-Otín et al., 2013). The increased stiffness of the ECM with aging contributes to chronic inflammation and fibrosis, conditions often seen in diseases such as arthritis and cardiovascular disorders. Additionally, the ability of cells to remodel the ECM effectively diminishes with age, leading to reduced tissue repair and regeneration capabilities (Campisi et al., 2007; Selman and Pardo, 2021).

It is straightforward that aging and cancer are intimately linked (Hanahan, 2022; López-Otín et al., 2023) at the cellular level through shared mechanisms involving integrin signaling and extracellular matrix (ECM) remodeling. As tissues age, senescent cells accumulate and secrete a pro-inflammatory molecules that disrupts ECM composition and stiffness. These changes enhance integrin engagement - particularly via RGD-binding integrins like αvβ3 and α5β1 - activating oncogenic pathways such as FAK and PI3K/AKT. The aged ECM, often stiffer and disorganized, fosters aberrant mechanotransduction that promotes stemness, survival, and invasion. Thus, dysregulated integrin-ECM interactions in aged tissues not only impair regeneration but also create a permissive niche for tumor initiation, progression, and therapeutic resistance (Fane and Weeraratna, 2020).

Mesenchymal stromal cells (MSCs) (Horwitz et al., 2007) serve as key regulators at the intersection of tissue regeneration, immune modulation, and disease progression. In cancer, MSCs promote tumor growth and metastasis by remodeling the tumor microenvironment, secreting cytokines and extracellular vesicles, and engaging RGD-binding integrins such as αvβ3 and α5β1. These integrin-mediated interactions with fibronectin- and vitronectin-rich ECMs facilitate MSC migration and differentiation into tumor-supportive stromal cells, activating pro-survival FAK signaling (Antoon et al., 2024). Understanding these adhesive and mechanical dynamics is essential for developing MSC-based therapies or targeting their roles in age-related diseases and cancer (Galderisi et al., 2022).

Ultrashort peptide (USP) hydrogels are innovative and smart biomaterials for biomedical applications, besides others (Hauser et al., 2011; Wu et al., 2012; Lakshmanan et al., 2013; Levin et al., 2020). These peptide-based hydrogels form self-assembled networks with high structural integrity under physiological conditions (Aisenbrey and Murphy, 2020). They offer versatility in tissue engineering and regenerative medicine due to their biocompatibility, ease of functionalization, and ability to support cell growth and tissue repair. They overcome many of the drawbacks of the standard 3D scaffold Matrigel, because they are xeno-free and their composition can be fully controlled by rational design and synthesis. This allows controlling stiffness, printability, and cell and tissue specificity of the matrix (Loo et al., 2014; Loo et al., 2019; Susapto et al., 2021; Alzanbaki et al., 2021; Pérez-Pedroza et al., 2021; Alshehri et al., 2021; Abdelrahman et al., 2022; Pérez-Pedroza et al., 2023; Abdelrahman et al., 2023; Alhattab et al., 2023a; Alhattab et al., 2023b; Xu et al., 2024).

Biofunctional peptides integrated into peptide hydrogel scaffolds offer a valuable approach for investigating cancer and cellular senescence. These peptides, designed to replicate essential biological signals, enable the manipulation of cellular behaviors and interactions within a controlled hydrogel environment (Qazi et al., 2020).

Atomic Force Microscopy (AFM) (Eibl and Bhushan, 2013) is a powerful tool for studying cell adhesion at the nanoscale (Franz et al., 2007; Jo et al., 2024; Le et al., 2019) and is increasingly combined with biofunctional peptides to provide detailed insights into cellular interactions with the ECM (Harjumäki et al., 2020). AFM can be used to image single molecules and cell surfaces (Moretti et al., 2022; Heath et al., 2021), measure adhesion forces (Moretti et al., 2023), and assess variations in mechanical properties (Moretti et al., 2021; Efremov et al., 2017; Garcia, 2020). Coating AFM probes with biofunctional molecules provides a simple, controlled method to study how these peptides influence cell adhesion, migration, and ECM interactions (Harjumäki et al., 2019; Na et al., 2022; Holuigue et al., 2023).

In this study, we establish a new method that integrates advanced Atomic Force Microscopy (AFM) characterization techniques with a range of synthetic biofunctional hydrogel formulations to examine the adhesion responses of various cell types, including fibroblasts (hDFs), mesenchymal stromal cells (MSCs), and neurons, in the contexts of cancer and aging, to different adhesion cues. Specifically, we used hDFs at low and high passage to simulate senescence; MSCs from a healthy donor and an acute myeloid leukemia patient; and neuroblastoma cells.

To the best of our knowledge, this is the first time biofunctional peptides capable of forming hydrogels have been coated onto an AFM probe to measure cell adhesion at the single-cell level over time. The ability to design and synthesize custom peptide sequences to target membrane-expressed proteins on living cells presents an exciting opportunity to characterize these complex systems at the single-cell level.

Overall, the combination of these two technologies provides detailed label-free information on adhesion behavior in living cells in response to different biofunctional peptides. This approach is useful for distinguishing divergent behavior across various cell types and subpopulations.

Considerations regarding the experimental setup of this work

In this study, we aim to understand two fundamental aspects of cell-material interactions. First, we seek to identify the most effective hydrogel adhesion cue for each cell type we examined, building on our efforts to optimize cellular adhesion to the synthetic parent peptides IIFK and IIZK. Second, we aim to determine whether this adhesion-based framework can distinguish different states of the same cell type, including transitions such as senescence or cancer.

Therefore, we coated the AFM spherical probe with the parent peptide IIFK (IIZK) and with the one to one mixture of the parent peptide and its biofunctional counterpart.

To explain the rationale behind this experimental design, let us understand how a cell anchors to a surface. Cells exhibit non-specific adhesion to positively charged surfaces due to the intrinsic anionic character of the plasma membrane, arising from phosphatidylserine, sialylated glycoproteins, and sulfated proteoglycans, which drives strong electrostatic attraction; when a cationic surface additionally presents an adhesion motif such as RGD, integrins engage this ligand while proteoglycans (e.g., syndecans) act as co-receptors that enhance clustering, force transmission, and matrix anchoring, thereby producing specific, cooperative adhesion (Morgan et al., 2007). Notably, there exist several integrin types specific for the same biofunctional motif.

The adhesion characteristics of a cell can be studied by depicting the interaction of the cell with a substrate, by pulling directly the cell on the biofunctional substrate or vice-versa, by providing the cell with an approaching biofunctional probe such as in this study. The advantages of using a movable probe acting on a cell are mostly to: i) arbitrarily decide the exposure time of the cell to the probe; ii) control the loading rate; these in turn can aid in understating the kinetic properties of the system under investigation in a dynamic way.

In fact, dynamic single molecule force spectroscopy experiments are critical to extract thermodynamic and kinetic parameters of ligand/receptor bond (Lostao et al., 2023; Evans and Ritchie, 1999). However, contrary to the typical experiment with single molecule F-D curves, the rationale behind this work is the understanding of the interaction of the cell with a hydrogel matrix that has a biofunctional motif. This tool is twofold: on the one side, it allows to depict the cell behavior towards designer biofunctional motifs, on the other side it allows to investigate which hydrogel sequence could give the better result in a 3D culture environment.

It is important to introduce the assumptions for this work: i) the spherical AFM probes have the same diameter (from the manufacturer); ii) the probe coating is homogeneous across the conditions (the same peptide concentration and coating protocol are used in all conditions); iii) the indentation area of the probe on the cell is assumed to be the same because the same cell types are compared only; iv) the hydrogel is adsorbed to the probe surface; v) the hydrogel is formed by a network of nanofibers held together by physically cross-linked bonds; vi) because of the large indentation area and the uniform coating of the probe with the hydrogels, several molecules interact with the cell surface; vii) the biofunctional motifs can be recognized by different receptors on the cell surface; viii) the charged peptides elicit an automatic adhesive response by the negatively charged cell surface; ix) the parent peptide is used as “negative control” with respect to the biofunctional hydrogel coated probe.

Because of the nature of the assumptions described earlier, no information on the kinetic behavior between ligand/receptor pair can be retrieved from this work.

In this framework, the parent peptide IIFK (IIZK) serves as the ideal negative control, rather than the use of blocking experiment, because it preserves the positive charge (and thus the non-specific electrostatic adhesion) while lacking the RGD motif, enabling direct comparison to IIFK-GGG-RGDS, which supports both non-specific and integrin-specific binding. Blocking integrins is therefore not meaningful, because electrostatic and proteoglycan-mediated adhesion persists regardless of integrin inhibition, preventing clean isolation of ligand-specific effects, whereas the parent peptide intrinsically controls for all non-specific interactions. Moreover, contrary to other F-D experiments, where the ligand on the probe is specifically recognized by only one receptor type on the cell surface, biofunctional motifs such as RGD are recognized by more than one type of integrin (Hynes, 2002). Therefore, the blocking experiments, would not provide additional information on a specific unfolding event. Finally, this study specifically investigates the interaction of the cell with an AFM coated with physically crosslinked hydrogels: the scope of this work is not to explore the kinetics or the energy landscape of a specific ligand/receptor pair, rather to explore the effect of biofunctional motifs hydrogels, on different cellular types over time.

In addition, for all the reasons mentioned above, we decided to evaluate the results by comparing the FA, the EA and FA/Dmax. In fact, the FA should provide information about non-specific interaction of the cell with the probe, thus mostly the negatively charged cell surface interacting with the positively charged hydrogel. Whereas, the EA and FA/Dmax will provide not only the non-specific interaction information, but also the one related to the integrin interaction with the biofunctional motifs.

Peptide synthesis

The peptides, IIFK (Ac-Ile-Ile-Phe-Lys-NH 2), IIFK-I (Ac-Ile-Ile-Phe-Lys- Gly-Gly-Gly-Ile-Lys-Val-Ala-Val-NH 2) IIFK-R (Ac-Ac-Ile-Ile-Phe-Lys-Gly-Gly-Gly-Arg-Gly-Asp-Ser-NH 2), IIFK-Y (Ac-Ile-Ile-Phe-Lys-Gly-Gly-Gly-Tyr-Ile-Gly-Ser-Arg-NH 2), IIZK (Ac-Ile-Ile-Cha-Lys-NH 2), IIZK-G (Ac-Ile-Ile-Cha-Lys- Gly-Gly-Gly-Gly-Phe-Hydroxypro-Gly-Glu-Arg-NH 2), IIZK-R (Ac-Ile-Ile-Cha-Lys- Gly-Gly-Gly-Arg-Gly-Asp-Ser-NH 2), IIZK-Y (Ac-Ile-Ile-Cha-Lys- Gly-Gly-Gly-Tyr-Ile-Gly-Ser-Arg-NH 2), were synthesized by 9-fluorenylmethoxycarbonyl (Fmoc)-based solid phase peptide synthesis. The Rink amide-methylbenzhydrylamine (MBHA) resin (GL Biochem) was pre-swelled with dichloromethane (DCM) (SRL) for 30 min, followed by deprotection of the Fmoc group with 20% (v/v) piperidine/dimethylformamide (Spectrum Chemical/Merck) before the first amino acid was introduced. For the coupling step, a N-methyl-2-pyrrolidone (NMP) solution (SRL), containing 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate (TBTU, GL Biochem), 1-hydroxy benzotriazole (HOBt, GL Biochem), N,N-diisopropylethylamine (DIPEA, Sigma-Aldrich), and Fmoc-protected amino acid, was added to the reaction vessel and agitated for 3 h. After the coupling of the last aminoacid, the resin was washed, and a capping process was conducted using a mixture of DIPEA (Sigma-Aldrich) and acetic anhydride (Sigma Aldrich) in dimethylformamide. The peptides were then cleaved from the resin using a mixture of trifluoroacetic acid (Thermo Fisher Scientific), triisopropylsilane (Sigma Aldrich), and water (95/2.5/2.5 v/v/v). The pure peptides were obtained through reverse phase-liquid chromatography purification. Purity and peptide mass were confirmed by mass spectra analysis.

Raman spectroscopy

Raman spectra of the parent peptides and peptide combinations were collected in a WiTec Apyron confocal Raman spectrometer. The samples were prepared by dissolving a total weight of 10 mg peptide powder or peptide mix powder (1w of parent peptide and 1w of biofunctional peptide) into 900 µL of MilliQ water and then adding 100 µL of 10X PBS (phosphate buffered saline) into a 4 mL glass vial. A small quantity of the formed hydrogel was then deposited onto a CaF 2 (Crystran Ltd.) Raman inactive glass slide. Several spectra were collected by using a 12 mW illumination power at the source with a 532 nm wavelength laser. The collected spectra were baseline subtracted, the counts normalized to the 1,450 cm -1 Raman shift value and averaged. Afterwards, the Amide I region from 1,500 to 1,800 cm -1 was investigated to depict the secondary structure of the peptides. Typically, the secondary structure of proteins contains α-helix, β-sheet, and disordered structures. Therefore, we fitted several Voigt functions to the experimental spectra to retrieve the composition of the peptide secondary structure. In particular, we introduced in the fitting procedure the peak at about 1,550 and 1,610 cm -1 related to the aromatic aminoacids, the water peak at about 1,640 cm -1, the peak at 1,650–1,658 cm -1 for the α-helix, at 1,660–1,676 cm -1 for the β-sheet/turn, and at 1,680–1,690 cm -1 for the disordered structure.

AFM microscopy

AFM topography

AFM topography was used to characterize the nanofibrous structure of the parent peptide and peptide combination. JPK Nanowizard III AFM (Bruker) was used to acquire the topography images. The parent peptides and peptide combinations were dissolved at a final concentration of 10 mg/mL in MilliQ water and left on the bench overnight. The day after 20 µL of a ten times diluted solution were spotted on a freshly cleaved mica sheet and dried after 2 min under an N 2 stream. The topography measurements were acquired by using the Olympus AFM probe AC-240TS-R3 with a nominal resonance frequency of 70 kHz in tapping mode in air. The topography images were acquired at 2 × 2 µm, and 1,024 × 1,024 pixel resolution per image. The images were processed with the Gwyddion software for the statistical analysis. Each image was flattened and leveled to zero. Then several orthogonal and longitudinal profiles were drawn in order to get the characteristics dimensions of the fibers: the diameter and the helical pitch.

AFM force spectroscopy

AFM force spectroscopy was performed in liquid environment with a JPK Nanowizard III AFM (Bruker). The cells were grown in 2D environment in TPP petri dishes specifically designed for JPK AFM. CP-qp-CPNT-BSGA probes were used for the adhesion measurements. The probe’s tip is a glass sphere of nominal 5 µm diameter. The choice for this probe diameter and not a larger one was dictated by two observations: i) larger diameters of the AFM probe would not allow the AFM tip to fully detach from the cell and ii) neuron cells detached easily from the petri dish surface when larger probes are used, given that a minimum of Force setpoint of 2 nN at least must be applied (see comments later).

AFM probe hydrogel adsorption

The peptide solutions were prepared by dissolving 10 mg/mL of peptide powder in MilliQ water; for the combined peptides, we used a 1:1 (w/w) ratio of parent peptide/biofunctional peptide. After dissolved, they were left undisturbed for 24 h to allow full fiber formation. This was verified beforehand by performing AFM topography measurements on mica surface as described in Section 5.4.1. A glass slide was cleaned with ethanol and one 5 µL drop of the peptide solution was combed by a pipette tip on the glass surface. Then, the AFM probe was washed in ethanol. It was then mounted on the AFM head, calibrated by thermal noise method and approached with a force setpoint of 5 nN at the edge of the peptide drop. After, the probe was kept in contact for 30 s with the hydrogel and then rapidly removed by zeta stepper motors with a step of 1,000 μm, while the glass slide with the hydrogel was removed.

In the instance that we aimed to verify that the hydrogel was absorbed onto the probe surface, we performed an AFM topography on the AFM probe itself. The same probe was unmounted from the AFM head and attached to a glass microscope slide by the use of double tape adhesive. Then, AC-240TS-R3 AFM probe was used as reported in Section 4.3.1, to perform a tapping mode topography on a small area of the probe itself. In this way, we could verify that nanofibers of hydrogel covered the surface of the probe (Supplementary Figure S1).

In the instance that we aimed to proceed with the measurements on the living cells, the probe was left undisturbed for 20 min after coating in the hydrogel layer and the glass slide with the hydrogel layer was removed. Afterwards, the petri dish with the cells were mounted on the petri dish holder that, in the meantime, was warmed up at 37 °C. The media was exchanged twice with Hank’s Balanced Medium (HBSS) 1X and the AFM probe was immersed in the buffer while mounted on the AFM head and allowed to equilibrate in the petri dish buffer for 10 min. Then, the cantilever spring constant was calibrated again with the thermal noise method, contact free. The quality factor, sensitivity and spring constant values were recorded. Viable cells were identified by using the Olympus optical inverted microscope on which the AFM head was mounted. The health of the cell was visually verified by observing the shape of its edges prior and after each measurement.

JPK AFM Nanowizard III was operated in contact mode for the force-distance measurements on single living cells in HBSS 1X buffer. A map of 4 × 4 force spectroscopy curves, with a total size of 3 × 3 µm was recorded on top of each cell, slightly aside from the nucleus. Three different recordings were acquired on the same area map to identify integrin formation after 1 s, 10 s, and 30 s of permanence of the AFM probe approached to the cell surface. Other parameters included approach force setpoint at 5 nN, displacement of the probe at 12 μm, velocity of probe displacement at 2 µm/s. The approach force setpoint for the neurons was kept at 2 nN, since the cells tend to detach at larger setpoint.

A minimum of 5 cells in 3 different replicates (N = 15 × 3 different time points × 16 points: N* = 720 acquired F-D curves per cell type and per hydrogel type) were acquired for the same cell type and hydrogel type, using the same hydrogel coated probe per individual replicate (see plot in Supplementary Figure S6). After the measurements, the probes were rinsed with MilliQ water and stored to proceed with AFM topography on the spherical probe itself as described above, to verify the persistence of the hydrogel fibers on the probe.

The F-D curves were analyzed by the proprietary JPK analysis software. Each F-D curve was baseline subtracted, the height corrected for the deflection of the cantilever, and the retract part of the curve was investigated by recording both the area under the curve (Energy of adhesion in Joule) and the maximum adhesion force (Force of adhesion in N).

The obtained data were then plotted according to the cell type and the time of probe permanence on the cells. Statistical significance between the populations was retrieved by applying a Mann-Whitney test of significance to the median values.

Cell culture

Cell cultures

Human dermal fibroblasts (Gibco, MA, United States) were cultured in a T75 flask at 37 °C, 5% CO 2. Complete media was prepared by adding 10% FBS (Gibco, MA, United States), and 1% Penicillin-Streptomycin (Gibco, MA, United States) to DMEM, high glucose (Gibco, MA, United States). The media was changed every 3–4 days, and cells were split after they reached a 90% confluence. Splitting was performed by washing the media from the flask and adding 5 mL of TrypLE Express (Gibco, MA, United States) for 5 min at 37 °C. The trypsinization solution was inactivated with 5 mL of complete medium and then centrifuged at 300 × g for 5 min. The supernatant was discarded by decantation. Cells were resuspended in 1 mL of complete media and split in a 1:5 ratio.

Human bone marrow mesenchymal stromal cells (hBM-MSCs) from a healthy donor and an AML patient were cultured as previously described (Alhattab et al., 2019). The culture medium consisted of α-modified minimum essential medium (α-MEM) supplemented with 10% mesenchymal stem cell qualified FBS, 2 mM L-glutamine, and 1% Penicillin/streptomycin (GIBCO, ThermoFisher, United States). Cells at passages 4–7 were used in cell culture experiments. hBM-MSCs from a healthy donor were a kind gift from Professor Abdalla Awidi at The University of Jordan/Cell Therapy Center in Jordan. Patient-derived AML-associated MSCs were collected from King Fahad Medical City, KSA, under ethical approvals 20–227, issued by the IRB Office of King Fahad Medical City for studies led by Prof. Malak Abedalthagafi and according to protocols approved by the IBEC office of King Abdullah University of Science and Technology.

The SH-SY5Y cell line (ATCC, CRL-2266), a thrice-subcloned derivative of the SK-N-SH neuroblastoma line, was utilized due to its widespread use as a model in neurodegenerative research. The SH-SY5Y cells’ capacity to differentiate into various functional neuron types makes them a valuable tool for investigating neurodegenerative processes.

Cell adhesion assay

Adhesion assays were prepared by placing dermal fibroblasts on top of a peptide hydrogel layer and inducing detachment using mechanical force. Adhesion was evaluated by the final amount of cells that tolerated the mechanical force and remained attached. For this, peptide solutions were prepared at 8 mg/mL in ultrapure, nuclease-free water (Ambion, MA, United States) in a glass vial. The peptide solutions were vortexed for 10 min, or until fully dissolved, and sterilized under UV light for at least 60 min. The peptide combinations included: IIFK, IIFK/IIFK-I, IIFK/IIFK-R, IIFK/IIFK-Y, IIZK, IIZK/IIZK-G, IIZK/IIZK-R, IIZK/IIZK-Y. The combinations are 1:1 (w/w).

Following sterilization, the peptide solutions were transferred to a 96-well plate with five replicates of each peptide solution at 35 μL per well. All plates were prepared in duplicate. Peptide hydrogels were established in aqueous solution using 2X PBS (Gibco, MA, United States), adjusting the final peptide concentration to 4 mg/mL. The plates were sealed with parafilm and incubated overnight at 37 °C to ensure complete gelation.

The next day, a cell solution was prepared at 0.6 × 10 5 cells/mL in complete media, and 50 μL of this solution was added to the peptide gels. One plate was incubated for 60 min at 37 °C, 5% CO 2, while the other plate was subjected to mechanical agitation vortexing at medium intensity for 5 s to represent the t = 0 min adhesion condition. After 60 min, the second plate was agitated in the same manner to represent the adhesion condition at t = 60 min.

Immediately after each agitation, the media was removed, and wells were gently washed with 1X PBS. Cells were fixed with 4% paraformaldehyde (Thermo Fisher Scientific, MA, United States) for 30 min at room temperature, followed by a wash with 1X PBS. DAPI (1:1,000 from a 10 mg/mL stock solution; Thermo Fisher Scientific, MA, United States) was added for 10 min in the dark for nuclear staining, then washed once with 1X PBS. Plates were stored at 4 °C in the dark.

Imaging was performed using an EVOS M7000 Cell Imaging System (Thermo Fisher Scientific, MA, United States) with the DAPI (357/447 nm) light cube. Stitched images covering 80% of each well were captured using a ×4 objective. Maximum intensity projection images of a Z-stack sweeping from top to bottom with a step of three focal planes were generated, with a final resolution of 4,072 × 3,042 px. Images were segmented into a 1,000 μm × 1,000 μm grid, with three squares randomly selected per well for manual cell nucleus counting. Statistical significance of the mean was performed by using a two-sample t-test on each pair of interest.

Cell culture for AFM force spectroscopy

Human dermal fibroblasts (hDF) were cultured on TPP Petri dishes for 1 day prior to the AFM measurements; cells at low and high passages (#5 and #17) were used. Both healthy and cancer derived MSCs were cultured on TPP Petri dishes for 3 days before the AFM measurements. Both HDFs and MSCs cells were seeded at a density of 4,000 cells/cm 2, to which 2 mL of culture media were added. The cells were kept in an incubator at 37 °C under 5% CO 2.

SH-SY5Y cells were seeded at a density of 8,000 per petri dish and incubated at 37 °C with 5% CO 2 for 10 min to allow cell adhesion. After this, 2 mL of phenol red-free DMEM (Gibco, United States) supplemented with 10% fetal bovine serum (FBS) (Gibco, United States) and 1% penicillin-streptomycin (P/S) (Gibco, United States) was added to each dish, and the cells were maintained in a humidified incubator at 37 °C with 5% CO 2. The cells were measured after 3 days in culture.

The results part is divided into two sections: Firstly, we report on the characterization of the peptides, and secondly, on the adhesion results of different types of cells to biofunctional peptide hydrogels, by means of AFM force spectroscopy measurements.

Peptide characterization

The rationally designed ultrashort peptides (USP) used for this study are composed of a short amphiphilic sequence containing a tail of mainly hydrophobic amino acids and a single positively charged amino acid head group. The amphiphilic peptide compounds are staggered in an anti-parallel self-assembly process to fibrils generating in aqueous solution nanofibrous scaffolds in form of stable three-dimensional hydrogels that support 3D cell growth. USP hydrogels can incorporate additional biofunctional motifs, selected from natural occurring sequences in the ECM. The two parent peptides that we investigated in this study were the peptides IIFK and IIZK. We have previously described their physico-chemical properties, and outlined how they form hydrogels, composed of entangled nanofibers that entrap water and saline solutions (Susapto et al., 2021). When adding biofunctional motifs to peptides IIFK and IIZK, we connected the peptides and the motif with a triple glycine linker. The added chosen motifs are known for supporting cell adhesion, namely, RGDS, IKVAV, YIGSR, GFOGER. We already reported on the promising results of several of these peptides when used for organoid growth and 3D bioprinting applications (Pérez-Pedroza et al., 2023; Xu et al., 2024).

For brevity, we indicate the biofunctional peptides as follows: IIFKGGGIKVAV (IIFK-I), IIFKGGGRGDS (IIFK-R), IIFKGGGYIGSR (IIFK-Y), IIZKGGGGFOGER (IIZK-G), IIZKGGGRGDS (IIZK-R), IIZKGGGYIGSR (IIZK-Y). We indicate the combinations of parent peptide with derived biofunctional peptide at a one to one concentration ratio (w/w), as follows: IIFK + IIFKGGGIKVAV (IIFK/IIFK-I), IIFK + IIFKGGGRGDS (IIFK/IIFK-R), IIFK + IIFKGGGYIGSR (IIFK/IIFK-Y), IIZK + IIZKGGGGFOGER (IIZK/IIZK-G), IIZK + IIZKGGGRGDS (IIZK/IIZK-R), IIZK + IIZKGGGYIGSR (IIZK/IIZK-Y).

The glycine linker gives flexibility to the peptide molecules and provides additional space for cells to access and interact with the biofunctional motif, while protecting the parent peptide sequence to self-assemble less disturbed into fibers. The more prominent the hydrophilic character of the added biofunctional motif, the less propensity there is for nanofiber formation. In fact, the peptides IIFK-R and IIZK-R alone do not self-assemble, and do not easily form hydrogels, if not at very high concentrations or with suitable ratios between parent peptide and biofunctional peptides (Xu et al., 2024). Consequently, we decided to combine the parent peptide with the derived biofunctional peptide, both to favor the hydrogel formation and to control the stiffness of the scaffold. Regulating the stiffness of the scaffold is crucial for the growth of different cell types.

The Raman spectroscopy performed on the hydrogels in 1X PBS with the two parent peptides and peptide combinations, shows the prevalent formation of a sheet type of secondary structure (Figure 1a; Table 1). Interestingly, we observed a preponderance of β-turns (Raman shift range 1,660–1,663 cm -1) in the mixture containing RGDS and GFOGER, while the combinations containing YIGSR and IKVAV show a preponderance of β-sheets (Raman shift range 1,664–1,670 cm -1). These Raman shifts are both characteristic of amyloid-like fibers (Susapto et al., 2021; Pérez-Pedroza et al., 2023; Moretti et al., 2023; Wan et al., 2023; Zhang et al., 2020; Moretti et al., 2018; Moretti et al., 2017; Moretti et al., 2013).

The AFM topography (height signal) reported in Figure 1b, shows how both parent peptides and the dual peptide combinations (1:1 = w/w), generate nanofibers, albeit with different characteristics. The IIFK nanofiber helix is left-handed while the IIZK nanofiber helix is right-handed, confirming our previous data (Susapto et al., 2021). In all peptide combinations where IIFK was investigated together with its biofunctional peptide, the resulting fibers are left-handed like the parent peptides, while in the case of IIZK in combination with IIZK-Y the handedness is flipped from right-handed to left-handed (Figure 1b). Concerning important fiber characteristics, such as the diameter and helical pitch, we observed that the interaction of the parent peptide with its biofunctional peptide gave raise to both different diameters and different helical pitches. For example, native IIFK fibers have a diameter of 5.84 ± 0.29 nm that changes to 6.31 ± 1.4 nm, 4.55 ± 0.42 nm, 7.06 ± 0.65 nm for the combination IIFK/IIFK-I, IIFK/IIFK-R and IIFK/IIFK-Y respectively. In addition, the IIFK/IIFK-Y presents a new helical pitch below 50 nm, while IIFK/IIFK-I presents the highest pitch variability (Table 2). In the case of IIZK and its combination, the diameter of the fibers is quite stable at around 8 nm while the pitch of IIZK combinations decreases with respect to IIZK. In particular, the fibers’ helical pitch goes from values above 100 nm in IIZK nanofibers, to pitch values between 50 and 70 nm. Interestingly, the combination of IIZK and its biofunctional derivate with YIGSR gives rise to an additional helical type with respect to the other two investigated combinations (Table 2).

AFM force spectroscopy with biofunctional peptides on single living cells

AFM can be applied on living cells in physiological environment. In addition to topography studies, AFM allows to acquire information on the adhesion properties of the material, in the so-called force spectroscopy mode. In this mode, force vs. distance (F-D) curves are collected typically in an array of x-y points. The characteristic data can then be represented as a Force vs. Z-piezo displacement graph, as exemplified in Figure 2a. In this graph we are able to identify an approach part (red line) of the curve, where the probe moves from far away until coming into contact with the cell, and a retract part of the curve (black line) where the probe moves from the point of contact with the cell back to its original position. Each curve can be repeated several times to collect an x-y array of curves on the cell surface. Typically, the retract part of the curve is investigated to gather information related to the adhesion of the probe to the cell. Therefore, the coating of the probe with specific molecules, aims to investigate specific interactions of the cell with these molecules. As can be visualized in Figure 2b (top), our strategy is to use an AFM spherical probe coated with the peptide hydrogels of interest, to measure the dynamic interaction of the cells with it. We verified that the coating was effective before and after measuring the cells. In particular, we observed that the sphere is coated with fibrous structures, identifiable as peptide nanofibers, as anticipated, both before and after measurement (Figure 2b, bottom, Supplementary Figure S1).

The retract part of the F-D curve can present different features: generally, three major outcomes are used to interpret the data: the Force of adhesion, the distance at which the maximum force of adhesion event occurs, and the Energy of adhesion. The Force of adhesion (FA) measures the maximum force required by the probe to detach from the surface, while the Energy of adhesion (EA) measures the work done by the probe to detach from the cell surface (Figure 2a). The FA, expressed in Newton, provides information on the detachment force of the strongest bond, when the probe is separated from the cell. The distance at which the maximum adhesion event occurs (Dmax) can shed light on the stiffness of the cell response to the specific ECM cue: this becomes especially evident when the parameter FA/Dmax, or “adhesion stiffness”, is evaluated (Na et al., 2022). The EA, expressed in Joule, gives information on the behavior of the adhesion profile in the Z dimension and includes three types of interactions that can be described as non-specific adhesion, sawtooth profiles and step (tether) profiles (Figure 2a). According to the literature, the sawtooth profile describes a specific interaction between the probe and a molecule on the cell surface, which is anchored to the cytoskeleton, resulting in a work at higher force; the step instead describes a specific interaction between the probe and a molecule that is not strongly anchored to the cell cytoskeleton and therefore, when pulled, the cell membrane extrudes in a long “membrane tether”, resulting in a work at low force (Sheetz, 2001; Evans and Calderwood, 2007; Davis and Sowinski, 2008; Müller et al., 2009). Intuitively, the curves that contain sawtooth profiles will have higher EA than the ones containing steps (Müller et al., 2009). However, the number of molecular interactions in each F-D profile plays a role in determining the EA that in addition, varies depending on the grade of clustering of the adhesion molecules. This is why the adoption of the FA/Dmax parameter can provide additional information on the occurring interaction. For more detailed explanation on the experimental setup, refer to the first paragraph of the Methods.

AFM F-D results

The first step of the experimental procedure was to coat the AFM spherical probe with a proper set of peptide hydrogels. Having decided that the negative control would be the parent peptide (see Methods), we needed to screen for the best biofunctional peptide sequence and hydrogel form. For example, from previous works we were aware that the peptide IIFK-R alone would not form a stable hydrogel (Xu et al., 2024) and it should be mixed with the parent peptide in order to allow the hydrogel formation.

Notwithstanding, we performed a probe coating pre-screening with the aim of deciding if the IIFK/IIFK-R would be more effective than the IIFK-R hydrogel alone. We used three types of coatings: the IIFK alone, the IIFK-R alone and a 1:1 (w/w) combination of the IIFK and IIFK-R for F-D measurements on HDFs cells alone (see methods section). The results show that the IIFK/IIFK-R is more effective in elicit the cell adhesive response, followed by the parent peptide and lastly by the IIFK-R peptide alone (Supplementary Figure S2). The cause of this effect is beyond the scope of this work, but this evidence allowed us to select the mixture of the parent peptide with the biofunctional peptide as a better candidate to perform AFM F-D measurement.

After, we performed a biofunctional peptide pre-screening by AFM F-D curves only on hDF cells to limit the use of the other, more delicate cells. The AFM probes were coated with the parent peptide and the combination of peptides solutions as described in the method section. Then, we performed F-D measurement on the hDF cells and extrapolated their adhesion energy to each single peptide. The median values of EA collected in the different experiments are plotted in Figure 2c. As can be appreciated in the graph, the peptide combinations that elicits the best response contains the RGDS sequence. From the Mann-Whitney statistical analysis we extrapolated that although at 1s and 30s IIFK/IIFK-R and IIZK/IIZK-R median values are not statistically different, they are different for the two populations at 10s permanence time.

To confirm the AFM results, we performed an adhesion test where the hDFs cells were cultured on top of parent peptides and biofunctional peptide combination hydrogels (Supplementary Figure S3). We derived that both at t = 0 min and t = 60 min the most adhesive hydrogel is the combination IIFK/IIFK-R (Figure 2d), therefore confirming the preliminary AFM F-D data.

Thus, we decided to perform the adhesion measurements by comparing the IIFK peptide and the combination of IIFK/IIFK-R on the following experiments.

For the adhesion studies dedicated to peptides IIFK and IIFK in combination with IIFKGGGRGDS (IIFK + IIFK-R, from now on named MIX), three cell categories were investigated: human dermal fibroblasts, human mesenchymal stromal cells, and human neuroblastoma cells. In the case of the hDFs we investigated cells at low passage (#3–4) and cells at high passage (#17–18) to simulate senescence (Ogrodnik and Gladyshev, 2023). The MSCs were collected from the bone marrow of a healthy donor and an AML patient. The neuroblastoma cells were derived from a cancer neuroblastoma cell line (See Methods section).

Figure 3, Supplementary Figure S4 show the optical images of the typical presentation of the cells cultured in 2D in a physiologically-relevant solution with the AFM probe (on the right side) facing the cells from the top, right before the measurement. This visual aid allowed us to identify the healthy cells by their margins adhesion to the surface and to stop the measurements when they did not look any longer viable (typically after 3 h). It also allowed to precisely positioning the AFM probe in an area aside to the nucleus. We need to perform the measurement in this area because the nucleus tends to slip away under the pressure of the probe, therefore altering the results. At the right side of each optical image we present a representative sample of the retract part of the F-D curves collected from one cell, for each type of cell and for each type of peptide. Importantly, we plot the curves of each cell type and peptide type acquired with 1, 10, and 30 s of probe permanence on the cell itself.

In Figure 3 we display the F-D curves of fibroblasts at early stage passage (YHDF in Figure 3a) and senescent (OHDF in Figure 3b), of MSCs from healthy donor (HMSC in Figure 3c) and from leukemia cancer patient (CMSC in Figure 3d), and of neuroblastoma cells (NB in Figure 3e). From this sample, we can already appreciate how the longer the exposure of the cell to the probe, the higher the EA, and mostly the EA value is higher with the MIX compared to IIFK, with some exceptions as we will explain later. In Supplementary Figure S5 we plot examples of the same F-D curves organized to compare IIFK and the MIX at the same exposure time. The adhesion to both peptides appears to be increased with time extension in all the cell types (Supplementary Figures S5a,c,d,e), except for the senescent fibroblasts OHDF (Supplementary Figure S5b). However, it is not possible here to discriminate between the effect of IIFK and the MIX in the same cell type.

Importantly, to evaluate if the hydrogel was detaching from the probe while performing the F-D measurements, we plotted the EA values for each experiment and each replicate, in temporal sequence (Supplementary Figure S6). This is an important information to understand not only the stability of the hydrogel on the probe, but also to avoid binding site saturation on the cell itself. We can observe that there is no proportional decrease of the EA over time; rather, it depends on the cell investigated. This supports the AFM topography scans acquired on the used AFM probes, showing the presence of the hydrogel fibers after measurement, as already reported in Figure 2, Supplementary Figure S1.

Typical datasets acquired from AFM force spectroscopy on diverse cell populations, present with skewed or multimodal distributions. To extrapolate the statistical significance between populations, we used the non-parametric Mann-Whitney test based on the median value. However, although this test is the preferred one in the case of skewed populations, the analysis of the median cannot take into account for large population skewness or for the presence of multi-modal behavior. Therefore, in the following paragraphs, we will not only analyze the median values of the populations, but also their distribution for EA, FA, Dmax, and FA/Dmax for all the experimental conditions.

The bar plots in Figures 4–6 represent the median values (with the related median absolute deviation as error bar) of the EA, FA, Dmax, and FA/Dmax values, with probes coated with IIFK or MIX hydrogels. For each of these plots, the two different cell populations for hDF, YHDFs vs. OHDFs, (Figure 4), and MSCs, HMSCs vs. CMSCs, (Figure 5), were investigated, while we had only one class of neurons, the human neuroblastoma cells (Figure 6). In Supplementary Figure S7 we plot the number of sawtooth profiles, step profiles, and equal combination of the two we observed in the retract part of the F-D curve, for each condition investigated. In Table 3, we summarize the change in EA expressed as EA MIX / EA IIFK and the change in adhesion stiffness FA/Dmax, expressed as (FA/Dmax MIX)/(FA/Dmax IIFK).

Fibroblasts population

A clear difference was observed between young and senescent fibroblasts (Figure 4). In YHDFs, adhesion energy increased with contact time and was significantly higher for the MIX hydrogel, particularly at short exposure times (1–10 s). This increase correlated with a higher prevalence of sawtooth rupture profiles, indicating sequential receptor–ligand interactions and suggesting active engagement with the biofunctional motif (Supplementary Figure S7). Adhesion force values also increased with exposure time and was already enhanced after 1 s contact with the MIX probe. Analysis of Dmax and FA/Dmax further indicated relatively high adhesion stiffness for both fibroblast populations; however, YHDFs displayed higher stiffness values earlier in time (around 10 s), suggesting a faster activation of adhesion domains.

In contrast, senescent fibroblasts (OHDF) exhibited a uniform response across conditions. Adhesion energy showed only modest changes between IIFK and MIX hydrogels, and the increase in adhesion force became apparent only after longer contact times (30 s). Moreover, the retract-curve profiles shifted toward a higher proportion of step and mixed events, indicating fewer sequential rupture events and a reduced specificity of adhesion. OHDF also exhibited a consistent Dmax, both with IIFK and the MIX, and the ratio FA/Dmax indicates a constant growth across timescales, contrary to YHDF, in both conditions. These findings are consistent with literature reports describing altered integrin signaling, reduced focal adhesion turnover, and decreased endocytosis in senescent cells. Such changes lead to a more stable but less dynamic adhesion phenotype, often associated with altered integrin expression and impaired cellular renewal (Shin et al., 2020; Shin et al., 2021; Pang et al., 2023).

Mesenchymal stem cells

A distinct behavior was observed when comparing healthy MSCs and cancer MSCs cells (Figure 5). In healthy MSCs (HMSC), the parent peptide IIFK generated higher adhesion forces, whereas adhesion energy increased with contact time in both hydrogel conditions. Analysis of retract profiles revealed a larger number of sawtooth events with the IIFK probe, explaining the higher adhesion energy observed under this condition (Supplementary Figure S7). However, analysis of Dmax and FA/Dmax indicated that adhesion stiffness was higher with the MIX hydrogel, particularly at 10 s contact time, suggesting that integrin engagement promotes stronger mechanical anchoring despite fewer rupture events.

In CMSC, adhesion energy remained similar between the two-hydrogel conditions and the distribution of retract profiles showed a balanced mixture of sawtooth and step signatures. Adhesion stiffness was also largely unaffected by the presence of the RGD motif. This behavior likely reflects the altered membrane organization and cytoskeletal architecture commonly observed in cancer-associated stromal cells (Chastney et al., 2024; Hamidi and Ivaska, 2018). Indeed, cancer-derived MSCs often display dysregulated integrin expression and signaling pathways that promote migration and tumor progression rather than stable matrix adhesion (Gou et al., 2024).

Neuroblastoma cells

Human neuroblastoma cells exhibited slightly higher adhesion energy and force when interacting with the IIFK hydrogel compared with the MIX condition (Figure 6). Consistently, sawtooth rupture profiles were more frequent in interactions with the parent peptide, particularly during short contact times (1–10 s) (Supplementary Figure S7). Adhesion stiffness also showed slightly higher values with the IIFK probe at intermediate contact times. These observations suggest that adhesion in neuroblastoma cells is primarily driven by electrostatic interactions with the positively charged peptide hydrogel rather than by specific recognition of the biofunctional motif.

Population level analysis

To further explore the heterogeneity within each cell population, the interquartile range (Q 3 -Q 1) and distribution of adhesion parameters were analyzed (Supplementary Figures S8–13; Supplementary Tables S1–S3). In several cases, the interquartile range increased under MIX conditions, indicating variability in the response of individual cells within the same population (Supplementary Figures S8–S10). For example, CMSCs showed increased EA variability with the MIX hydrogel across all contact times, while NB cells exhibited larger variability at longer exposure times. In fibroblasts, low-passage cells displayed broader FA distributions with the MIX hydrogel, suggesting heterogeneous engagement of adhesion receptors.

The adhesion stiffness distributions also revealed interesting trends. Higher FA/Dmax interquartile range values were observed at 10 s with the MIX hydrogel for YHDFs and HMSCs, whereas OHDFs and CMSCs showed higher stiffness values at the same time point with the parent peptide IIFK. This pattern suggests that healthy adherent cells respond more rapidly to the biofunctional motif, while senescent or cancer-related cells rely more strongly on non-specific electrostatic interactions.

Distribution analysis further showed that most populations were positively skewed and in some cases exhibited bimodal distributions, particularly at longer contact times (Supplementary Figures S11–S13). Bimodality was observed in fibroblast populations with the MIX hydrogel at 30 s, in CMSCs under both hydrogel conditions at 30 s, and in neuroblastoma cells for both hydrogels at longer contact times. Such distributions likely reflect the coexistence of subpopulations with different adhesion states or receptor activation levels.

Integrated interpretation

We would like to propose that the fundamental information about the systems investigated could be mostly retrieved from the analysis of EA and FA/Dmax. Comparison of the MIX and IIFK conditions using EA and FA/Dmax ratios (Table 3) highlights distinct adhesion phenotypes across cell types. Young fibroblasts exhibited a strong increase in adhesion energy in the MIX condition without a substantial change in adhesion stiffness, suggesting that the biofunctional motif primarily increases the number or persistence of adhesion interactions. In contrast, healthy MSCs showed a marked increase in adhesion stiffness with the MIX hydrogel, indicating fewer but mechanically reinforced adhesion events. Senescent fibroblasts and cancer-derived MSC-like cells behaved similarly: both showed limited sensitivity to the biofunctional motif and higher stiffness values under the IIFK condition, suggesting that adhesion in these cells is largely dominated by non-specific electrostatic interactions. Neuroblastoma cells also displayed only a modest response to the biofunctional hydrogel.

Overall, these findings indicate that biofunctional hydrogel coatings modulate adhesion mechanics in a cell-type-dependent manner. Healthy cells respond more specifically to the RGD motif, while senescent and cancer-associated cells exhibit reduced sensitivity and rely more strongly on non-specific adhesion mechanisms.

This study demonstrates that AFM force spectroscopy combined with biofunctional ultrashort peptide hydrogels can resolve cell-type- and state-dependent adhesion phenotypes at the single-cell level. The main finding is that the RGD-containing MIX probe does not simply increase adhesion in all cells, but instead reveals distinct modes of interaction. Young fibroblasts (YHDF) responded to the MIX mainly through increased adhesion energy (EA), whereas healthy mesenchymal stromal cells (HMSC) responded predominantly through increased adhesion stiffness (FA/Dmax). In contrast, senescent fibroblasts (OHDF), cancer-derived MSCs (CMSC), and, to a lesser extent, neuroblastoma cells (NB), showed weaker discrimination between the parent peptide and the biofunctional MIX, suggesting a larger contribution of non-specific electrostatic interactions and a reduced ability to mechanically reinforce the RGD cue.

These results support the concept that the chemistry of the AFM probe is itself biologically informative. Previous AFM force-spectroscopy studies using native extracellular matrix-coated probes showed that microenvironment-specific ligands can generate cell- and tissue-specific adhesion signatures (Holuigue et al., 2023). In the same way, the comparison between IIFK and MIX in the present work indicates that the peptide coating is not an inert adhesive layer, but a tunable biointerface capable of revealing how different cells interact with defined extracellular cues. This interpretation is also consistent with our previous studies showing that decorated ultrashort peptide matrices influence organoid growth and architecture in a motif-dependent manner, including lumen formation and improved epithelial organization in colon organoid models (Pérez-Pedroza et al., 2023; Xu et al., 2024).

The different responses of YHDF and HMSC to the MIX are particularly informative. In YHDF, the increase in EA together with the higher prevalence of sawtooth events suggests a larger number or persistence of sequential receptor-mediated detachment events. In HMSC, however, the strongest effect of the MIX was observed in FA/Dmax, indicating fewer but mechanically reinforced adhesion events. These two outcomes are not contradictory. Rather, they suggest that different healthy adherent cell types translate the same RGD cue into distinct mechanical behaviors: fibroblasts through increased adhesion work, and stromal cells through stronger effective anchoring. This interpretation is supported by studies showing that integrin activation state and ligand-specific biomaterial interactions strongly influence AFM-measured adhesion mechanics (Harjumäki et al., 2020).

The comparison between young and senescent fibroblasts further supports the biological relevance of the method. Senescent OHDF exhibited only modest differences between IIFK and MIX and tended to show delayed or weaker evidence of RGD-sensitive reinforcement. This is consistent with the known role of integrin-mediated adhesions in cellular senescence and with reports showing altered focal adhesion turnover, abnormal endocytosis, and dysregulated integrin signaling in senescent cells (Shin et al., 2020; Shin et al., 2021). In this context, senescent cells may remain adhesive, but their adhesion is less dynamic and less specific, which agrees with broader concepts linking aging to altered ECM mechanobiology and impaired tissue renewal (Phillip et al., 2015; Selman and Pardo, 2021; López-Otín et al., 2023).

A similar interpretation applies to CMSC. Compared with HMSC, cancer-derived MSCs showed weaker sensitivity to the RGD-containing MIX and more heterogeneous adhesion behavior. This may reflect the altered membrane organization, cytoskeletal architecture, and integrin signaling often observed in cancer-associated stromal cells and cancer stem-like states (Cooper and Giancotti, 2019; Antoon et al., 2024; Gou et al., 2024). The parallel between OHDF and CMSC is noteworthy: both populations appear less able to convert the biofunctional cue into a coherent reinforced adhesion response, and instead rely more strongly on the baseline adhesive properties of the parent peptide.

The neuroblastoma results were more modest, with slightly stronger adhesion often observed for IIFK than for MIX. However, the population-level analysis suggested heterogeneity, with subpopulations becoming more apparent at longer contact times. This is compatible with the known variability of integrin expression in neuroblastoma and with the possibility that the relatively low force setpoint used here mainly probed superficial membrane interactions rather than deeper integrin-cytoskeleton coupling (Tagalakis et al., 2021).

Overall, these findings indicate that biofunctional peptide hydrogels modulate adhesion mechanics in a cell-type-dependent manner. Healthy cells appear more capable of specifically engaging and mechanically reinforcing the RGD cue, whereas senescent and cancer-associated cells show reduced specificity and greater dependence on non-specific adhesion.

Taken together, this work illustrates how nanomechanical measurements combined with biofunctionalized materials can reveal the biophysical mechanisms underlying cell-matrix adhesion. Differences in adhesion force, energy, and adhesion stiffness reflect variations in integrin engagement and cytoskeletal anchoring across cell types and physiological states. Such mechanobiological signatures may provide valuable information for distinguishing healthy, senescent, and cancer-associated cells.

Specifically, the importance of integrin-targeted therapy in aging and cancer is recognized by the medical community. RGD sequences are not only used to block integrin activity but also to deliver drug cargoes to specific sites (Tagalakis et al., 2021; Rozo et al., 2016; Milla et al., 2011).

In line with this vision, AFM has already been suggested as a clinical tool for probing tumor cell mechanics and evaluating drug responses at the single-cell level (Li et al., 2015; Stylianou et al., 2018). Our study adds to that narrative by demonstrating the added value of a tailored biointerface: the self-assembling peptide coating enhances the relevance of the measurement to physiological cell-matrix interactions.

In this study, ultrashort self-assembling peptide hydrogels functionalized with an RGD motif were used to coat AFM probes and investigate cell–matrix adhesion at the single-cell level. This approach enabled the identification of distinct adhesion phenotypes across multiple biological contexts, including young versus senescent fibroblasts, healthy versus leukemia-associated mesenchymal stromal cells, and neuroblastoma cells. By combining measurements of adhesion energy (EA), maximum adhesion force (FA), and adhesion stiffness (FA/Dmax), we were able to dissect different aspects of cell–matrix interactions and highlight differences in how cells engage biofunctional extracellular cues.

Our results indicate that healthy cells respond more specifically to the biofunctional peptide motif. Young fibroblasts exhibited a pronounced increase in adhesion energy in the presence of the RGD-containing hydrogel, suggesting a higher number or persistence of receptor-mediated interactions. Healthy MSCs, instead, displayed a marked increase in adhesion stiffness, indicating fewer but mechanically reinforced adhesion events. In contrast, senescent fibroblasts and leukemia-derived MSCs showed weaker sensitivity to the RGD cue and tended to rely more strongly on the baseline adhesive properties of the parent peptide. This behavior suggests a reduced capacity to translate specific extracellular signals into reinforced cytoskeleton-anchored adhesion. Neuroblastoma cells displayed a more heterogeneous adhesion phenotype, consistent with the presence of subpopulations with different adhesive states.

Importantly, the population-level analysis revealed skewed and occasionally multimodal adhesion distributions, emphasizing the ability of this AFM-based platform to capture heterogeneity within cell populations. Such mechanophenotypic fingerprints may be particularly valuable for identifying functional subpopulations that are not easily distinguished by conventional biochemical markers.

Beyond the specific biological insights reported here, the AFM/peptide-hydrogel platform offers a versatile tool to investigate cell–matrix interactions in physiologically relevant conditions. By modifying the peptide coating, the system can be expanded to probe different integrin–ligand systems, such as laminin-derived motifs (e.g., IKVAV), collagen-binding sequences (e.g., GFOGER), or fibronectin-derived peptides. This modularity could enable selective adhesion profiling across diverse cell types and pathological conditions, including cancer, fibrosis, and aging-related diseases.

From a translational perspective, the method may also support future developments in drug screening and personalized medicine. Quantifying changes in adhesion forces in response to integrin inhibitors, focal adhesion kinase modulators, or cytoskeleton-targeting therapies could provide a sensitive, label-free readout of drug efficacy at the single-cell level. Furthermore, engineered peptide matrices with tunable biochemical and mechanical properties could be used to selectively capture or exclude specific cell populations, offering potential applications in regenerative medicine and cell therapy.

Future work should expand the current approach in several directions. First, combining AFM adhesion measurements with complementary molecular readouts, such as fluorescence imaging of integrins or cytoskeletal components, would further clarify the relationship between mechanical signatures and underlying adhesion machinery. Second, extending the platform to three-dimensional cell culture systems and organoid models may provide a more physiologically relevant context to study cell–matrix interactions. Finally, integrating this technology with microfluidic or high-throughput force spectroscopy systems could enable systematic screening of patient-derived cells, thereby facilitating mechanophenotypic diagnostics and therapeutic stratification.

Overall, the integration of AFM force spectroscopy with biofunctional ultrashort peptide hydrogels provides a powerful and adaptable strategy to probe integrin-mediated adhesion in living cells. By revealing differences in adhesion mechanics between healthy, senescent, and cancer-associated cells, this platform opens new opportunities for mechanobiology research and future translational applications in oncology, aging biology, and regenerative medicine.