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

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

15.4.4 Biocompatibility of Nanorobot Fragments in vivo

The partial or complete disintegration of diamondoid medical nanorobots in vivo should be an exceedingly rare event (Chapter 17). Nevertheless, if and when this occurs, nanorobots that have lost physical integrity should be recognized as foreign matter and be engulfed by free macrophages or resident phagocytes such as Kupffer cells. The uncoated, rough exposed surfaces of these diamondoid devices should invite prompt opsonization (blood protein tagging) and removal from blood or tissue via geometrical (Section 15.4.2) or phagocytic (Section 15.4.3) processes.

Large nanorobot fragments consisting of sharp indigestible shards might destroy the phagocytosing cell, causing it to rupture and discharge its cytoplasmic contents. This could lead in turn to acute local inflammation (Section 15.2.4) and the probable release of chemotactic agents attracting mesenchymal cells to the site, which would then differentiate into fibroblasts, resulting in entombment of the shards in the adjacent tissue by a permanent fibrous spherical granulomatous capsule (Section 15.4.3.5). Studies show that jagged mechanically-produced PMMA wear particles elicit increased inflammatory responses compared to smooth round latex particles [646]. Attachment to phagocytic cells may be enhanced by the rough surface of mechanically-produced particles [3653], the increased surface area of rough particles [646], or by other factors [646]. Less jagged nanorobot fragments might ultimately be sequestered in the lymph nodes or other lymphatic organs, or may also suffer granulomatous entombment in place. Nanorobot fragments resulting from dental grinding (Section 9.5.1, Chapter 28) may be swallowed and eliminated from the body via the alimentary canal. Alternatively, these fragments may become embedded in the oral mucosa, subsequently either being encased in a permanent granuloma or being extruded from the mucosal epithelium (e.g., marsupialization) back into the oral fluids, then swallowed and excreted (unless reabsorbed in the gut; Section 15.4.3.3.2).

A medical nanorobot, like any machine, is most likely to rupture at its weakest links. Noncovalent (e.g., van der Waals) bonds will break before covalent bonds, so nanorobots that are subjected to catastrophic physical forces initially are most likely to fragment into relatively less reactive nanoparts and nanoassemblies, rather than into more highly reactive semirandom molecular fragments of such components. A few representative diamondoid nanoparts have already been designed (Section 2.4.1) by K.E. Drexler, R.C. Merkle, and R.A. Freitas Jr. These nanoparts are essentially very large molecules incorporating multiatom structures held together by a combination of internal covalent and noncovalent bonds (including steric hindrance), having, for example, the following basic chemical formulas:

C₂₀ H₂₄ Si₂ and C₂₀ H₂₄ Ge₂ for the dimer placement mechanosynthesis tool tip [36, 5683];

C₁₄₃₃ N₅₃₆ H₄₀₃ O₁₃₄ Si₄₄ S₃₄ F₁₂ for the fine motion controller [3654];

C₁₈₂₆ H₁₈₀₆ Si₁₆₄₅ O₃₆₇ N₂₂₄ S₂₂₀ P₇₇ for the neon gas pump [3655];

C₂₄₆₁ Si₂₇₉₂ H₈₆₄ N₆₂₈ P₄₅₂ O₃₆₇ S₃₅₆ for the differential gear [3656];

C₁₄₇₂ H₁₀₀₀ for the hydrocarbon bearing [3657]; and

C₄₇₀₈ H₂₀₂₀ for the hydrocarbon universal joint [3658].

Such nanocomponents clearly are not pure diamond structures – the mean atomic weight per atom for the above designs is 7.0-18.8 daltons/atom of structure, with an average of ~12 daltons/atom. Sapphire (aluminum oxide) based nanorobot components are also likely. In 2002 few such nanocomponents had been precisely defined at the molecular structural level, so their biochemical and biological reactivity is largely unknown. Will they be opsonizable and phagocytosable? Large free nanoparts with irregular exterior contours and surface charge distributions are probably immunogenic (Sections 15.2.3.3 and 15.3.7) and thus will likely become visible to the immune system and to the RES. It is already known that nanometer-size pieces of amorphous carbon and pure diamond can undergo photooxidation in the presence of light, oxygen (air) and water [3659], and sapphire can suffer slow chemical dissolution under certain conditions (Section 15.3.5.6). Designed to be built and operated in vacuo [10] and often containing relatively energetic strained-shell structures (Section 2.4.1), individual nanoparts could be chemically reactive in vivo with water, oxygen, ions and free radicals, proteins, or other biological substances that are abundantly available. Detached protein-based presentation semaphores might also be immunogenic (Section 15.2.3.3). Stably functionalized adamantanes, the smallest possible chunks of diamond that can exist, appear reasonably biocompatible and are generally excreted unchanged in the urine (Section 15.3.1.4(8)).

Will free nanoparts be cytotoxic, pyrogenic, or systemically toxic? A large solid pure hydrocarbon-like molecule with a lengthy stretch of exposed hydrogens might be nontoxic, as is the case with large chunks of long-chain linear hydrocarbons such as paraffin wax [3660]. However, some short-chain linear paraffin hydrocarbons (C_nH_2n+2) such as propane are considered “poisonous” [3661] with an official NFPA Health Hazard Rating of 1 or “slightly toxic” (scale 0-4) [3662]. Hydrogen-terminated diamond and diamond-like coatings appear relatively nontoxic and inert (Section 15.3.1), as do sapphire-like particles (Section 15.3.5.5). Nanoparts with exposed sulfur or nitrogen atoms could be more reactive in vivo. In vitro studies of ultrafine particles with living cells show an increased ability to produce free radicals which then cause cellular damage [6199-6201], manifested as genotoxicity [6200] or altered rates of apoptosis [6200-6203]. As noted by Howard [6188], the upper size limit for the lung toxicity of ultrafine particles is not fully known but is believed to lie between 65-200 nm [6190]. Endocytic vesicles in alveolar membranes may be 40-100 nm in size and are thought to be involved in protein macromolecule and occasionally virus transport into cells [6189]. Medical nanorobots or passive inert nanoparticles larger than 100-200 nm should present relatively low endocytic transport risk, although smaller particles, free nanoparts, or devices might pose some incremental risk.

It is possible that stray nanoparts may prove relatively more inflammatory than whole nanorobots, all else equal. One experiment [769] found that 14-nm carbon black particles (about the size of individual nanoparts; Section 2.4.1) produced a much greater alveolar neutrophil inflammation reaction than was elicited by larger 260-nm carbon black particles (closer to the size of whole medical nanorobots). This is especially important at the very low doses of free nanorobot parts anticipated in vivo where particle area dominates the inflammation response, as distinct from the relatively unlikely higher overload-inducing doses where total particle mass or volume of the instilled particles dominates without any influence of total surface area [769]. Several in vivo studies have found elevated inflammation in animal lungs exposed to ultrafine (<100 nm) particle aerosols [6190-6196]. For example, Donaldson [6194] notes that “ultrafine particles made of low-solubility, low-toxicity materials are more inflammogenic in the rat lung than fine respirable particles made from the same material. The property that drives the greater inflammogenicity of ultrafines is unknown but very likely relates to particle surface area and involves oxidative stress. Ultrafine particles can also impair the ability of macrophages to phagocytose and clear other particles, and this may be pro-inflammogenic.” Seaton et al [6197] have proposed that the chronic inhalation of nanoparticles can provoke alveolar inflammation that can damage the lining of the blood vessels, leading to arterial disease, though there is some evidence [6198] that nanoparticle-induced lung inflammation and peripheral vascular thrombogenesis can be partially decoupled. Oberdorster [6216] reported that exposing rats to air containing 20-nanometer-diameter PTFE (Teflon) nanoparticles for 15 minutes leads to death for most of the animals within 4 hours, whereas animals exposed to air with much larger 130-nm particles suffered no ill effects. Histology studies showed that macrophage cells that normally clear out foreign material had trouble ridding tissue of the smaller particles.

What about the biodistribution of stray nanoparts? Kreyling et al [6177] examined the distribution of nanopart-sized 15-nm and 80-nm particles of chemically inert radiolabeled iridium particles in rats. Inhaled particles (including particles deposited in the alveolar region) were cleared predominantly via airways into the gastrointestinal tract and feces, with only <1% of nanoparticles translocated into secondary organs such as liver, spleen, heart, and brain after systemic uptake from the lungs and the translocated fraction of 15-nm particles about ten times larger than for the 80-nm particles. Iridium nanoparticles injected intravenously were “rapidly and quantitatively accumulated in the liver and spleen and retained there,” and nanoparticles inserted gastrointestinally by gavage were not absorbed through gut walls. The study concluded that “only a rather small fraction of [the inert nanoparticles] has access from peripheral lungs to systemic circulation and extrapulmonary organs.” A similar study of 20-29 nm carbon particles by the same researchers [6178] found significant translocation only from lung to liver after 1 day post-exposure, but it was unclear whether translocation had occurred via the circulation or the GI tract. Oberdorster [6216] reportedly also found that rat-inhaled carbon-13 and manganese nanoparticles reached the olfactory bulbs and then migrated throughout the brain [6212].

R. Bradbury notes that the three most common unguided active chemical reactions that occur in biological tissues are oxidation/free-radical damage, nitrosylation (e.g., NO attacks tyrosine and perhaps other amino acids), and glycosylation. The potential for stray nanorobot parts or their randomly-structured fragments to catalyze or enhance the rates of these reactions should be studied, and the resistance of undefended nanorobot surfaces to attacks by these reaction molecules should also be investigated.

In the case of fragmented biorobots (Section 1.3.2.1, Chapter 19), the biocompatibility of cell parts could be of special interest [3663-3666]. For example, Glaumann and Trump [3664] injected loose mitochondria and microsome organelles intravenously into rats, and found that half of the injected dose was recovered in the liver, with smaller amounts found in the lungs, kidneys, spleen, and heart. Serum clearance half-life was 5-15 minutes for microsomes and 30-60 minutes for mitochondria [3663, 3664]. Glaumann and others have also examined the uptake of injected liver cell plasma membranes [5060], erythrocyte ghost cell membranes [2847], lysosomes [2848], microsomes [2851], mitochondria [3666, 5062], ribosomes [2851], and other subcellular organelles [5061] by Kupffer cells; of mitochondria [3665] by human glial cells; of sperm tails by oocytes [5066]; of collagen [5078-5083], fibronectin [5083] and melanosomes [5084] by fibroblasts; and of amyloid fibrils [5063], DNA/RNA (Section 15.3.6.1), ECM [5064], myelin debris [5065], and red cell ghost membranes [5049, 5380-5384] by various phagocytic cells. Actin (released from dying or lysed cells) can circulate at µM concentrations in peripheral blood and may modulate plasmin-dependent biological responses [5067]. Free alpha-actin is present in the blood of patients with angina pectoris and acute myocardial infarction or ischemia up to 0.112 mg/ml [5068]. Bloodborne actin is scavenged or sequestered by Vitamin D binding protein [5069-5071] and gelsolin [5071, 5072], with a serum clearance half-life for free actin of 30 minutes at the liver [5071]. Granulocytic fragments have been observed in blood during sepsis [5387]. Finally, an inhaled vaccine consisting mostly of free bacterial ribosomes has been tested as a treatment for respiratory infections [5073-5076], producing significantly increased serum concentrations of immunoglobulins [5077].

Nanosecretagoguery – e.g., triggered enzyme release by nanorobots or free nanoparts – is another ever-present concern. Under some circumstances phagocytes can release enzymes directly into the extracellular fluid in response to particles with certain physicochemical characteristics. For example, incubation of particulate activators of the alternative complement pathway such as zymosan or glucan (polyglucose) particles with monocyte monolayers in vitro causes the monocytes to release 9-18% of their internal stores of lysosomal enzyme, N-acetylglucosaminidase, directly into the culture medium [3667-3669]. Similar releases are observed in human monocytes exposed to latex beads at Nct ~ 5-10% [3667, 3668] and asbestos [3668], and occasionally in eosinophils and neutrophils [3670]. Interestingly, monocyte enzyme release due to inert latex beads (the closest analog to medical nanorobots) is almost completely inhibited by ~8 µg/ml of the fungal metabolite cytochalasin B [3667]. Nanorobot fragments that inadvertently mimicked the relevant molecular structures of glucan or other stimulatory particles might trigger similar unwelcome accidental releases in vivo – several adamantane derivatives are secretagogues for insulin release by mouse islets in vitro [5570], and proteins (analogous to secretagogic nanoparts) capable of serving more than one function (e.g., both ion channel and enzyme [3671]) are known.

Last updated on 30 April 2004