© 2003 Robert A. Freitas Jr. All Rights Reserved.

Robert A. Freitas Jr., Nanomedicine, Volume IIA: Biocompatibility, Landes Bioscience, Georgetown, TX, 2003

15.2.2 Adhesive Interactions with Implant Surfaces

As the famous physiological chemist Leo Vroman once hyperbolized [950]: “Facing a hail of miscellaneous eggs, we cannot expect to come away clean. Unless they are hard-boiled ones, we are most likely to become coated rapidly with a relatively thin film of matter from the most numerous and most fragile eggs. Similarly, no interfaces may exist that, facing blood plasma, can escape being coated with the most abundant and fragile plasma proteins.”

Following the implantation of a biomaterial into a host tissue, the first event to occur at the tissue-material interface (which dictates biocompatibility) is the noncovalent adsorption of plasma proteins from blood onto the surface [517-520, 936, 954] if the material comes into direct contact with blood. (Osmotic minipumps delivering drugs subcutaneously would escape from this process.) Protein adsorption is much more rapid than the transport of host cells to foreign surfaces. Once proteins have adsorbed to the surface of the foreign material, host cells no longer see this underlying material, but only the protein-coated surface overlayer. This adsorbed protein overlayer – rather than the foreign material itself – then mediates the types of cells that may adhere to the surface, which ultimately can determine the type of tissue that forms in the vicinity [517]. Thus the type and state of adsorbed proteins, including their conformational changes, are critical determinants of biocompatibility [518-523], so pretreatment of surfaces can be a control mechanism (Section 15.2.2.1). Even by the late 1960s, Vroman and Adams [950-952] and Baier and Dutton [953] had found that within 10 seconds of exposure to blood or plasma, a uniform ~6 nm layer of fibrinogen formed on surfaces of Ge, Pt, Si, and Ta; after 60 sec, the layer was less uniform and averaged ~12.5 nm thick, but was still dominated by fibrinogen. Rudee and Price [793] determined that human serum albumin (HSA) (molecular dimensions 8 nm x 3.8 nm [1440], with a monomolecular radius of gyration in pH 5-7 solution of 3.2-3.4 nm [1441]) formed a continuous film on amorphous carbon surface in only 1.3 sec of exposure. Fibrinogen required 2.5 sec to form films. Protein adsorption on a range of hydrophobic and hydrophilic polymer surfaces is typically 0.3-12 mg/m² (~500-20,000 molecules/micron²) [1322].

Black [234] notes that many molecules synthesized naturally in the cell have a “tail” portion that inactivates them. Later on, enzymatic processes strip away this small segment, releasing the molecule into an active extracellular substrate pool. Contact with foreign surfaces, along with adhesive forces, may cause premature activation. Indeed, some molecules are specifically designed to become activated in this manner. One common example is fibrinogen, a molecule that is reduced slightly in molecular weight and ultimately converted to the active protein, fibrin, after foreign surface contact during blood clotting (Section 15.2.5).

Three-dimensional images of adsorbed human fibrinogen molecules have been obtained by atomic force microscopy (AFM) [562, 563], scanning force microscopy [564], and electron microscopy [565].* The mechanical kinetics of adsorption have also been examined by testing the adhesion strength of an individual fibrinogen molecule that is affixed to an AFM tip and is briefly touched to a silica surface [566].** A clearer molecular picture of the fibrinogen adsorption event on implant surfaces (and subsequent inflammatory response) is slowly emerging [567, 568].*** Similar tests need to be performed on, for example, diamond and sapphire surfaces.

* On hydrophobic mica, the adsorbed fibrinogen molecule has a bi- or trinodular slightly curved linear shape of mean length 65.9 nm and mean height 3.4 nm [563]. On hydrophobic silica, a trinodular 60-nm long form and a globular 40-nm diameter form are observed [564]. On quartz at low solution concentrations, fibrinogen molecules appear to form a 46-nm long multinodular rod with 6-7 nodes each 4 nm in diameter on hydrophilic surfaces, and a 40-nm long binodular or trinodular rod with node diameter 5-9 nm on hydrophobic surfaces. At high solution concentrations, the molecule forms end-to-end polymers on hydrophilic surfaces and spherical 18-24 nm diameter structures on hydrophobic surfaces [565].

** The minimum interaction time for strong binding was 50-200 millisec, producing mean adhesion strengths from 300 pN for a 5 millisec contact up to 1400 pN for a 2 sec contact. Consecutive ruptures indicating the detachment of multiple adhesion sites occurred 20-25 nm apart along the length of the molecule. This is comparable to the characteristic distance between D and E globules of a single fibrinogen molecule, suggesting that fibrinogen adsorbs mainly through its D and E globules [566].

*** Plasmin degradation of purified fibrinogen into defined domains reveals that the proinflammatory activity resides within the D fragment of the fibrinogen molecule, which contains neither the fibrin cross-linking sites nor RGD sequences [567]. After contact with blood or tissue fluid, the D domain tends to interact with biomaterial surfaces and is important in the tight binding of fibrinogen to implant surfaces [568]. The biomaterial surface then promotes conformational changes within the D domain, exposing P1 epitope (fibrinogen gamma 190-202, which interacts with phagocyte CD11b/CD18 (Mac-1) integrin) [568]. The engagement of Mac-1 integrin with P1 epitope then triggers subsequent phagocyte adherence and reactions [568], as demonstrated by experiments which show that phagocyte accumulation on experimental implants is almost completely abrogated by administration of recombinant neutrophil inhibitory factor (which blocks CD11b-fibrin(ogen) interaction) [567].

Once the precise molecular mechanisms of protein adhesion are fully understood, we may hypothesize that nanodevice surfaces could be designed for maximum proteophobicity and that this might be possible because numerous partially proteophobic molecular surfaces are already known, including:

(1) polyethylene glycol (PEG) [569] and other steric barrier coatings (Section 15.2.2.1);

(2) surface-immobilized fibrinolytic enzymes such as lumbrokinase [570-572] or, more generally, immobilized proteolytic enzymes [755] that can cleave and detach proteins that attach;

(3) hydrophilic cuprophane [573-577], chemically-modified cuprophane [578], polysulfone [576, 577], and polyacrylonitrile [579] hemodialysis membranes;

(4) albumin-passivated surfaces [2536-2538] that reduce fibrinogen adsorption [519], platelet adhesion and activation, and thrombogenicity [543, 580, 937] and accumulate very few adherent neutrophils or macrophages [581];

(5) tetraethylene glycol dimethyl ether glow-discharge plasma deposition surfaces, which can reduce fibrinogen adsorption to ~0.2 mg/m² (~350 molecules/micron²) on many different substrates [582];

(6) surface-immobilized heparin [1888, 2535], a natural anticoagulant (Section 15.2.5); and

(7) artificial glycocalyx-like engineered non-adhesive surfaces [753] (Section 15.2.2.1), low protein adsorbing films [754], graft polymerized acrylamide [5257], and other examples of biological adhesive surface engineering [2356, 2589].

More generally, nanomedical implants and instrumentalities may require surfaces of engineered bioadhesivity – for instance, diamond-like carbon coated surfaces with an additional overcoat of biological materials, perhaps including extracellular matrix proteins, laminin, fibronectin, albumin, and collagen IV, to either promote or inhibit cell growth and spreading [629]. Ratner [5293-5295] gives examples of biomaterials that inhibit nonspecific protein adsorption while simultaneously controlling the interactions of matricellular proteins at implant surfaces, thus reducing foreign body response while promoting healing. More systematic study is clearly needed. For instance, the Adhesion Dynamics model of Chang et al [2554] defines molecular characteristics of firm adhesion, rolling adhesion and non-adhesion, but only in the limited domain of leukocyte-vascular rolling interactions. In other studies, computer simulations suggest that organic molecules may be readily bonded to diamond or other nanorobotic surfaces to impart desired biocompatibility characteristics. Examples include a recent density functional theory (DFT) study of cycloadditions of dipolar molecules to the C(100)-(2x1) diamond surface [4738] and related experimental investigations [4748]), and the covalent immobilization of enzymes onto gamma-alumina surfaces [4772]. Conventional means can be employed to orient rod-shaped molecules on DLC surfaces (as in liquid crystal arrays [4742]) or to orient nanowires [4785] or carbon nanotubes [4793] on sapphire/alumina surfaces.

As of 2002, numerous companies [2281] including Advanced Surface Technologies (Billerica, MA), MetroLine Industries, Inc. (Corona, CA), Spire Corp. (Bedford, MA), SurModics Inc. (Eden Prairie, MN), Ultramet (Pacoima, CA), and Vitek Research Corp. (Derby, CT) were providing commercial design services to create customized biocompatible surfaces on implantable medical devices and medical materials.

Last updated on 30 April 2004