Biocompatibility of biomedical devices

Material/protein/cell interactions at the biointerface

Biomaterials and biomedical devices designed for blood-contacting applications have been practically implemented for many years [1,2,3]. Biocompatibility is among the most critical issues in biomaterial/device design, yet the mechanisms of biocompatibility are not understood at the molecular level. This focus review primarily addresses the “biocompatibility” of materials against various biological elements in blood flow systems. Biocompatibility is defined as being synonymous with bioinertness, anti-fouling, and stealth properties. When biomaterials comprising these devices encounter blood, the body’s innate defense mechanisms recognize them as foreign [1,2,3]. Foreign body reactions via blood coagulation on biomaterial surfaces proceed through a sequence of events (Fig. 1): (1) adsorption of water molecules from blood, leading to material hydration; (2) nonspecific adsorption of plasma proteins and conformational changes in adsorbed proteins, exposing cell adhesion sites; (3) platelet adhesion; (4) platelet activation; and (5) subsequent blood cell deposition, fibrin network formation and thrombus formation on blood-contacting materials. Water plays a crucial role in biointerfacial interactions, including protein adsorption and cell adhesion on biomaterials [4,5,6,7,8], yet water is often neglected at the interface between materials and biology [9]. Since the 1970s, some studies have investigated the relationships among hydrated water, hydrated materials, and biocompatibility from various perspectives, such as water content, hydrophilicity, water mobility, hydration energy, water density and orientation, and polymer chain mobility [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24] (Fig. 2). However, the relationship between the state of hydration and biocompatibility has not been fully elucidated.

Fig. 1
Fig. 1
Full size image

Initial events at the interface between biological fluids and materials. When biomaterials come into contact with blood/tissues, (1) water molecules immediately adsorb onto their surface. (2) This is followed by protein adsorption, protein restructuring, and denaturation of the adsorbed protein. (3) Cell adhesion. (4) Cell activation. (5) Fibrin formation and thrombus formation on blood-contacting materials

Fig. 2
Fig. 2
Full size image

Water States Linking Biocompatibility. Representative articles on the relationship between biocompatibility and water states in terms of the water content, water mobility, hydration energy, water density arrangement, and hydration surface molecular mobility. The concept of intermediate water (IW) can bridge these articles

Role of hydration states in biocompatibility

Materials properties—wettability, surface roughness, and elastic modulus—are critical for influencing protein adsorption and cell adhesion. However, these factors alone do not fully account for the biological responses observed. We hypothesize that the water states around materials are critically important for determining their biocompatibility. Compared with bulk water, water at biointerfaces exhibits fundamentally different behavior. There are three types of distinct water surrounding synthetic and biological molecules (Table 1): (i) strongly bound water, in which the water molecules are strongly bound to materials and with restricted mobility such that the water is unable to freeze—termed nonfreezing water; (ii) loosely bound water, in which the water molecules are close to the surface of biomolecules with less restricted mobility than nonfreezing water; these water molecules can freeze and melt but at temperatures well below 0 °C—termed intermediate water; and (iii) scarcely bound water, in which the water molecules are free from materials and freeze at 0 °C—termed free water. We can experimentally determine these three types of water using calorimetry, spectroscopy, and direct interfacial methods, such as differential scanning calorimetry (DSC), time-resolved in situ attenuated total reflection infrared spectroscopy (ATR-IR), Raman spectroscopy, sum-frequency generation (SFG) spectroscopy, terahertz spectroscopy, dielectric relaxation spectroscopy, soft X-ray emission spectroscopy, quasielastic neutron scattering, solid-state nuclear magnetic resonance (NMR), surface force measurements, and a wide variety of analytical techniques [25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50]. Among these, intermediate water was found in hydrated biopolymers and hydrated biocompatible synthetic polymers. Notably, intermediate water prevents the denaturation of proteins and facilitates cell adhesion on biointerfaces [51, 52]. However, the mechanism of how intermediate water is formed near materials is still elusive. Here, we discuss the fundamental concepts for determining the interactions of proteins and cells with hydrated materials along with selected examples corresponding to our recent studies [53, 54] on poly(2-methoxyethyl acrylate) (PMEA), PMEA derivatives, and self-assembled monolayers (SAMs).

Table 1 The three types of distinct water surrounding biological molecules and materials

Infrared spectroscopy of hydrated materials

Spectroscopic observations are promising techniques for investigating the detailed hydration structures of materials at the functional group level because of their sensitivity to the microscopic environment. Time-dependent attenuated total reflection infrared (ATR-IR) spectroscopic observations during the sorption process of water into a polymer film, as well as the subsequent data analysis based on machine learning, are described herein.

Time-dependent ATR-IR spectroscopy

Water has intense light absorption in the mid-infrared region. The absorption coefficients \(\alpha =4\pi k\left(\lambda \right)/\lambda\) of liquid water for O-H stretching at 3400 cm-1 and that for O-H bending at 1640 cm-1 are approximately 1.3 × 106 m-1 and 2.7 × 105 m-1, respectively, where \(k\left(\lambda \right)\) is an imaginary part of the refractive index as a function of wavelength \(\lambda .\)[55]. Therefore, transmission infrared (IR) spectroscopy for aqueous solutions requires controlling the optical path length to be approximately several micrometers, which can be difficult [56]. However, IR spectroscopy is a promising technique for investigating the molecular structures and interactions of compounds, including water molecules, because it reflects the microscopic environment around their chemical bonds [57,58,59]. Infrared spectroscopic observations of hydrated polymeric materials have also been well discussed [60,61,62,63]. We reported the hydration structures of PMEA and its analogous polymers by means of ATR-IR spectroscopy [64,65,66]. The equilibrium water content for PMEA is only 9.0 wt%, and three different types of hydrated water, namely, nonfreezing water, intermediate water and free water, were identified by differential scanning calorimetry (DSC) below 0 °C [67]. We captured the hydration structure of three different types of hydrated water in PMEA at ambient temperature by using spectroscopic techniques coupled with computational data analyses.

The ATR-IR spectra of dry and wet PMEA, measured using the flow-through cell shown in Fig. 3C, are shown in Fig. 3A[64]. A film sample of PMEA was deposited on a hemispherical internal reflection element (IRE) made of zinc selenide, which has a large refractive index of approximately 2.4 in the mid-infrared region. The thickness of the film sample was adjusted to exceed the penetration depth of the near-field light generated at the IRE/sample interface because of total internal reflection of the incident light. Therefore, the O-H stretching at approximately 3700–3000 cm-1 and the O-H bending at 1640 cm-1 arise from water sorbed into the polymer matrix, excluding information about bulk water in contact with the surface of the polymer film. The spectral waveform of the O-H stretching of water in PMEA looks somewhat different from that of bulk water. This suggests that the ATR-IR spectroscopic observations capture information about the hydration structure of water in the PMEA. Spectral intensities arising from PMEA below 1800 cm-1, excluding the band due to water at 1640 cm-1, decrease after contact with water. This is due to the expansion of the polymer film induced by water sorption.

Fig. 3
Fig. 3
Full size image

ATR-IR Analysis of Water States in PMEA. A ATR-IR spectra of dry and wet PMEA, B time-dependent ATR-IR spectra obtained during a sorption process of liquid water (adapted from Morita et al., Langmuir 2007;23:3750-61, © 2007 American Chemical Society, with permission), C a schematic illustration of the in situ ATR-IR flow-through cell, D three-component spectra of water in PMEA deconvoluted by MCR, A nonfreezing water, intermediate water and C free water

To clarify the hydration structures of PMEA in detail, time-dependent ATR-IR spectra during the sorption process of water into the PMEA film were observed (Fig. 3B) [64]. The spectral waveform of O-H stretching in the 3700–3000 cm-1 region changes gradually with increasing spectral intensity. This is evidence of a change in the water structure in the PMEA with increasing water content. In the case of water vapor sorption into PMEA, all of the sorbed water is nonfreezing water [67], and the spectral waveform of the nonfreezing water has a relatively high wavenumber contribution of approximately 3600 cm-1[64].

Machine learning using spectroscopic data

To clarify the overlapping spectral waveforms attributed to the three different types of water in PMEA, multivariate curve resolution (MCR) and two-dimensional correlation spectroscopy (2DOS) have been applied to the time-dependent ATR-IR spectra in the O-H stretching region [68, 69]. The three types of waveforms of water in the PMEA deconvoluted by the MCR are shown in Fig. 3D. Nonfreezing water has a higher wavenumber contribution of approximately 3600 cm-1, which corresponds to the same waveform of water vapor sorbed into the PMEA (Fig. 3D, A). The spectral waveform of free water is somewhat similar to that of bulk water, with a broad peak at approximately 3400–3200 cm-1 (Fig. 3D, C). Intermediate water has a characteristic spectral waveform in the O-H stretching region, with a maximum peak at approximately 3400 cm-1 and few contributions at approximately 3200 cm−1 (Fig. 3D, B). Note that the spectral contribution at approximately 3200 cm-1 is generally observed in the ice Ih, which has a hexagonal crystal structure because of hydrogen bonds among water molecules. These findings suggest that compared with free water or bulk water, intermediate water forms a small water cluster with a locally formed ice Ih structure in the liquid phase. The characteristic waveform for intermediate water, featuring a remarkable peak at approximately 3400 cm-1 and a smaller contribution at approximately 3200 cm-1, appears only on the surface of a self-assembled monolayer (SAM) modified on a surface-enhanced infrared absorption (SEIRA) substrate, which suppresses nonspecific protein adsorption and cell adhesion [70]. These results imply that a surface hydration structure similar to that of the intermediate water in the PMEA is responsible for the material function in the SAM. Crucially, DSC measurements of the SAM are challenging, and SEIRA spectroscopic observations enabled evaluation of the hydration structure on the SAM surface, which is similar to that of the intermediate water in the PMEA.

In the fingerprint region below 1800 cm-1, which was attributed mainly to the polymer chain, a characteristic peak splitting was identified at 1727 cm-1 and assigned to C = O stretching. The positive lower wavenumber contribution at 1712 cm-1 is due to hydrogen-bonded C = O, whereas the negative higher wavenumber contribution at 1727 cm-1 is due to C = O being free from hydrogen bonding. Our kinetic analysis [64] revealed that the increase in the relaxation time of hydrogen-bonded C = O is associated with that of the spectral waveform attributed to the nonfreezing water deconvoluted by MCR. Quantum chemical calculations also support the observed vibrational frequencies of the C = O・・・HO type of hydrogen bonding interaction [58]. Furthermore, 85.6% of the nonfreezing water in PMEA exhibited C = O・・・HOH・・・O = C-type hydrogen bonding interactions, i.e., one water molecule hydrated to two carbonyl groups. This type of water hardly crystallizes among water molecules even below –100 °C because of the strong interactions with the polymer side chains as a bridge structure.

A higher wavenumber shift of the band at 1155 cm-1 arising from the methoxy moiety was also found [64]. The relaxation time of the shift is in good agreement with the intensity variation of the intermediate water deconvoluted by MCR [68]. These findings suggest that the intermediate water hydrated to the methoxy moiety in the PMEA side chain terminal has a relatively smaller water cluster than the free water or bulk water does.

The hydration structure of PMEA described above was obtained by analyzing a large number of spectra acquired by time-dependent ATR-IR spectroscopy during the sorption process of water using several kinds of computational data analyses, including unsupervised machine learning and quantum chemical calculations. Recently, this interpretation has been supported by molecular dynamics (MD) simulations [71]. Furthermore, a phase separation structure in hydrated PMEA consisting of a water-rich phase and a water-poor or polymer segment-rich phase was found in a study using a model monomer for the PMEA of 2-methoxyethyl acetate (MEAc) [72]. Sum-frequency generation (SFG) [73] and atomic force microscopy (AFM) [74] studies revealed that the hydration structure of the water/PMEA interface is concentrated approximately four times greater than that of the polymer matrix because of inhomogeneous phase separation [72].

Surface force measurements of hydrated materials

The intermediate water concept is also valid for explaining the biocompatibility of self-assembled monolayers (SAMs), which are characterized by their high molecular packing density and relatively rigid, ordered structures. This section systematically reviews the direct experimental evidence that validates the water barrier model, beginning with foundational synthetic SAMs and extending to more complex biomimetic interfaces.

Water barriers near bioinert self-assembled monolayers (SAMs)

The most direct and compelling evidence for the existence and function of a water barrier on anti-fouling SAMs has been obtained through surface force measurements [75,76,77,78,79,80,81]. Seminal studies on canonical bioinert systems, such as SAMs terminated with oligo(ethylene glycol) (OEG) or zwitterionic sulfobetaine (SB) moieties, have utilized colloidal probe atomic force microscopy (AFM) to quantify the interaction forces between two such surfaces in physiological environments (Fig. 4). These experiments, conducted in phosphate-buffered saline (PBS), revealed a long-range repulsive force that is absent on bioadhesive surfaces such as methyl-terminated SAMs, which instead show attractive hydrophobic interactions. This water-induced repulsion is characterized by a range of approximately 4–6 nm. The nature and range of this force were found to correlate directly with the resistance of SAMs to protein adsorption and their overall blood compatibility, providing the first strong evidence of a water-mediated barrier [50].

Fig. 4
Fig. 4
Full size image

Chemical structures of the thiols used to fabricate the SAMs in this work (upper). SEM image of a colloidal probe for force measurements (lower left) and surface force–distance curves measured between the SAMs in PBS buffer (lower right). The range of the repulsive force was 4 to 6 nm, and half of the range (2 to 3 nm) was considered to be the thickness of the interfacial water inducing repulsion

Definitive validation of this model was achieved through visualization of the hydration structure with frequency-modulation AFM (FM-AFM). This 3D high-resolution imaging method was employed to examine the interface of a bioinert mixed-charge (MC) SAM in PBS (Fig. 5). The findings demonstrated a highly ordered layer of interfacial water with an approximate physical thickness of 2 nm. This observation constitutes the first direct visual evidence of a structured hydration layer that was hypothesized to generate a repulsive force [82].

Fig. 5
Fig. 5
Full size image

Chemical structure (upper left) and molecular packing configuration (upper right) of MC-SAMs. The molecular packing (in-plane imaging) (bottom left) and hydration structure (cross-sectional imaging) (bottom right) of interfacial water near a bioinert MC-SAM, as observed using frequency-modulated atomic force microscopy in PBS buffer. In the cross-sectional view, there was a region that induced a repulsive force to the AFM probe with a thickness of approximately 2 nm

Key insights were obtained when the physical thickness of the hydration layer was compared with its functional interaction range. The approximately 2 nm layer observed via FM-AFM represents the most tightly bound and structurally organized water molecules in the primary and secondary hydration shells of the SAM. However, the colloidal probe AFM detected a repulsive force at a much greater distance of 4–6 nm. This finding indicates that the influence of the SAM extends far beyond the physically bound water layer. The structured primary layer acts as a template, inducing some order in the surrounding water molecules and forming an extended hydrogen-bonding network. Therefore, the measured repulsive force reflects the energy needed to compress and disrupt this entire extended hydration region, not just the ~2 nm layer. Recognizing the difference between the physical thickness and the effective functional range of the water barrier is essential for understanding how these surfaces affect long-range biological interactions.

Water barriers in biomimetic systems

Chang et al. demonstrated this principle using zwitterionic peptide-based SAMs [82]. They fabricated SAMs composed of repeating dipeptide units of glutamic acid-lysine (EK) and aspartic acid-lysine (DK), which are the most abundant zwitterionic pairs found on the surfaces of proteins in the human body [83,84,85]. The biological performance of these surfaces significantly differed: the EK and DK SAMs exhibited excellent resistance to both protein adsorption and platelet adhesion, whereas the ER (R: arginine) and DR SAMs significantly fouled (Fig. 6). This dramatic difference in biological response was perfectly mirrored by the interfacial forces measured with AFM. The bioinert EK and DK SAMs generated a powerful, long-range repulsive force with an onset distance of 8–9 nm. In contrast, the bioadhesive ER and DR SAMs produced only weak, short-range repulsion with an onset of less than 4–5 nm.

Fig. 6
Fig. 6
Full size image

(upper) Amino acid sequences of the peptides used to fabricate zwitterionic peptide SAMs. (lower left) Amounts of fibrinogen adsorbed on the peptide SAMs. (lower right) Approaching surface force–distance curves obtained with the peptide SAMs. Identical SAMs were fabricated on both the probe and the substrate. All the curves were measured in phosphate-buffered saline (PBS)

Although all four peptide SAMs are hydrophilic, with static water contact angles ranging from 27° to 34°, simple macroscopic wettability does not predict their bioinertness. The key difference lies in the nature of the repulsive force. The decay length of the repulsion for the fouled ER and DR SAMs in PBS was measured to be approximately 0.7 nm, closely matching the theoretical Debye length of the buffer (0.745 nm). This suggests that the force is mainly standard electrostatic double-layer repulsion. In contrast, the decay length for the anti-fouling EK and DK SAMs was approximately 1.8 nm, much longer than the Debye length, indicating a non-DLVO, water-mediated interaction. This finding shows that the ability to establish a strong, long-range hydration barrier depends heavily on the molecular structure of the amino acid side chains—specifically, how water interacts with the primary amine of lysine versus the guanidinium group of arginine. Recent molecular dynamics simulations support this, confirming that zwitterionic polymers create a robust hydration energy barrier that resists protein adsorption and that this ability is precisely tuned by molecular parameters such as the charge separation distance and dipole moment [86].

A similar demonstration of water-mediated interaction control was reported by Kanayama et al. using gold nanoparticles functionalized with double-stranded DNA (dsDNA) (Fig. 7a) [87]. The study revealed that the colloidal dispersion stability of these nanoparticles under high-salt conditions, where electrostatic repulsion is screened, is determined entirely by the pairing status of the outermost terminal base pair (Fig. 7b). Nanoparticles coated with dsDNA terminated in complementary base pairs spontaneously aggregated. This aggregation is driven by attractive hydrophobic interactions between the DNA ends, a phenomenon known as blunt-end stacking. In contrast, nanoparticles functionalized with dsDNA terminated in mismatched, unpairing base pairs remained stably dispersed under identical high-salt conditions. This macroscopic difference in behavior, governed by a single molecular detail, suggests that fundamentally different interfacial forces are at play.

Fig. 7
Fig. 7
Full size image

Surface Forces of Hairpin DNA SAMs. a Hairpin structure-folded DNA strands were anchored on the gold surface through Au–S bond formation, resulting in the formation of homogenous dsDNA layers. b (upper) Base sequences and folded structures of hpDNA-SH. (lower) Photographs of the C- and M-DNA − GNP dispersions in sodium phosphate buffer (10 mM, pH 7.4) containing various concentrations of NaCl at 25 °C. c Approaching surface force–distance curves obtained with the DNA SAMs. Identical SAMs were fabricated on both the probe and the substrate. All the curves were measured in sodium phosphate buffer (10 mM, pH 7.4) containing various concentrations of NaCl

To elucidate the mechanism behind this terminal-specific behavior, surface force measurements were conducted between two dsDNA-coated surfaces using colloidal probe AFM (Fig. 7c). The results provide a direct link between the interfacial forces and the observed colloidal stability. For surfaces with complementary dsDNA, the interaction force transitioned from repulsive to strongly attractive as the salt concentration increased, with the attraction beginning at a distance of approximately 8 nm. These findings confirm that salt-facilitated blunt-end stacking is the driving force for aggregation [87, 88]. Conversely, for surfaces with mismatched dsDNA, a persistent, long-range repulsive force was measured, with an onset of approximately 8 nm. This repulsion was attributed to a hydration force, arising from the greater number of potential water-binding sites on the unpaired terminal nucleobases than on the fully paired ones. This robust hydration barrier prevents the nanoparticles from making contact, thus ensuring their colloidal stability and serving as a powerful example of how a single, discrete molecular change can fundamentally alter macroscopic interfacial forces by modulating the local hydration environment.

The consistent observation of non-DLVO repulsive forces across various anti-fouling systems from synthetic OEG and zwitterionic SAMs to biomimetic peptide and DNA monolayers indicates a common underlying physical cause. The “intermediate water concept” provides a solid theoretical foundation for explaining these experimental results. A key aspect of this model is that a high proportion of intermediate water at an interface is strongly linked to excellent blood compatibility and anti-fouling properties. In this framework, the long-range force measured by AFM can be viewed as the physical manifestation of this intermediate water layer. Although less ordered than the primary hydration shell, this layer has a unique hydrogen-bonding network different from that of bulk water, and compressing or displacing it requires a significant amount of energy [79, 89]. This idea establishes a crucial link between the microscopic state of interfacial water, measurable mesoscopic forces, and the broader biological effects associated with protein and cell resistance.

Concluding remarks and future perspective

Hydration water can be classified into three types: free water (scarcely bound water), intermediate water (loosely bound water), and nonfreezing water (tightly bound water). Among these, intermediate water was found in hydrated biopolymers and hydrated biocompatible synthetic polymers. Time-dependent ATR-IR spectroscopy combined with machine learning revealed the detailed hydration structures of PMEA at the functional group level. The results highlight the power of spectroscopic approaches to link molecular-level hydration to macroscopic material properties. We found that intermediate water is a key indicator of the biocompatibility of material surfaces under physiological conditions. The amount of intermediate water is influenced by the type of functional groups, local polymer configuration, and polymer chain mobility. Intermediate water also plays a key role in the bioinertness of SAMs and biomimetic surfaces. Long-range, water-mediated repulsive forces prevent protein adsorption, platelet adhesion, and overall biocompatibility. Molecular details, such as amino acid side chains or DNA terminal pairing, critically influence the strength and range of these hydration barriers. This water barrier model provides a unified framework for designing biocompatible interfaces.

This concept focuses on the presence and amount of intermediate water—a type of water common to both biopolymers functioning in aqueous environments and biocompatible synthetic polymers of nonbiological origin—and serves as a guiding principle for understanding and developing polymer properties and functions at the molecular level [90,91,92]. In the rational design of next-generation polymeric interfaces under complex biological conditions based on the intermediate water concept, we have demonstrated that subtle, systematic changes in molecular architecture can profoundly influence the bulk and surface properties and biological performance of polymers [7, 8, 20, 21, 52,53,54, 93,94,95,96,97,98,99,100,101]. For example, we successfully synthesized a series of several model polymers with identical elemental compositions but with the ester carbonyl group systematically repositioned along the side chain [53]. Our investigation revealed a clear relationship between the position of the carbonyl group, the resulting hydration states, the polymer dynamics under hydrated conditions, and the biocompatibility. A key finding is that compared with PMEA, poly[(2-acetoxyethyl)vinyl ether], in which the carbonyl group is located at the terminus of the side chain, is more resistant to platelet adhesion [53]. This superior performance is correlated with its unique hydration properties, including the largest amount of intermediate water, as quantified via DSC. The rational design of biocompatible polymers remains a significant challenge. Our work contributes a fundamental design principle, showing that precise control over the placement of a single functional group can be used to tune critical polymer properties that govern biological response. We believe these findings, which link a specific molecular architectural parameter to macroscopic physicochemical properties and ultimately to biological performance, will be of substantial interest in the fields of polymer chemistry, polymeric biomaterials science, and biomedical device engineering for novel biomedical interventions.

Traditionally, the development of biomaterials has relied largely on empirical screening strategies. In recent years, however, the introduction of the intermediate water concept has begun to redefine this paradigm. By providing a physiochemically grounded framework for characterizing biomaterial–water interactions, this approach enables more precise, quantitative, and rational screening of biomaterials, thereby improving both the efficiency and reliability of early-stage material selection. In fact, research and development are underway on molecular designs that achieve a balance between multiple functions, including not only the suppression of nonspecific adsorption and the prevention of fouling and adhesion but also the properties required in various fields where materials are used in aqueous environments—such as environmental, energy, electrical, electronic, mechanical, food, and agricultural applications. These functions include ionic conductivity, electrical properties, toughness, temperature responsiveness, biodegradability, selective adhesion to cancer cells, drug delivery systems, protein stabilization and antiaggregation properties, and pharmacological efficacy [102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141].