In conjunction with Foresight Update 45
How Nanorobots Can Avoid Phagocytosis by White Cells, Part I
By Robert A. Freitas Jr.
Research Scientist, Zyvex Corp.
|Robert A. Freitas Jr.|
Any invading microbe that readily attracts white cells (phagocytes) capable of eating it, and then allows itself easily to be ingested and killed, is generally unsuccessful as a parasite. That’s why most successful bacteria interfere to some extent with the activities of phagocytes or find some way to avoid their attention . Bacterial pathogens have devised numerous diverse strategies to avoid phagocytic engulfment and killing, mostly aimed at blocking one or more of the steps in phagocytosis, thereby halting the process .
Similarly, natural phagocytic cells presented with any significant concentration of medical nanorobots  also may attempt to internalize these nanorobots. How often will such an opportunity arise? There may be an average of one ~730 micron3 granulocyte (e.g., neutrophil) in every ~3 x 105 micron3 of human blood, one ~1525 micron3 monocyte in every ~2 x 106 micron3 of blood, and one >1525 micron3 macrophage in every ~2 x 105 micron3 of human tissues. By random thermal motions in a quiet fluid, a 2-micron nanorobot would trace out a volume containing one neutrophil in ~70 sec at 310 K (, Eqn. 3.1), or would diffuse the ~40 micron mean free distance (, Eqn. 9.72) between nanorobot and the nearest macrophage in quiet watery tissue in ~4000 sec (, Eqn. 3.1). In a small (1 mm diameter) artery with blood flowing at 100 mm/sec, each 2-micron nanorobot, in a total bloodstream population of 1012 such nanorobots, would collide with a white cell once every ~3 seconds near the periphery of the vessel but only once every ~300 seconds near the center of the vessel (, Section 22.214.171.124), a rheological disparity that will be amplified by phagocyte margination (, Section 126.96.36.199). Studies of macrophage particle-ingestion kinetics show that the number of particles ingested by each phagocytic cell may rise tenfold as the local particle concentration rises from 5 particles per cell to 150 particles per cell .
From these crude estimates, it becomes apparent that virtually every medical nanorobot placed inside the human body will encounter phagocytic cells many times during its mission. Thus all nanorobots which are of a size capable of ingestion by phagocytic cells must incorporate physical mechanisms and operational protocols for avoiding and escaping from phagocytes. Ingestion may require from many tens of seconds to half an hour to go to completion, depending upon the size of the internalized particle, so medical nanorobots should have plenty of time to detect and to actively prevent this process. The initial strategy for medical nanorobots is first to avoid phagocytic contact or recognition, and if this fails, then to avoid nanorobot binding to the phagocyte surface, and phagocytic activation.
One simple avoidance method employed by a few pathogens that may occasionally be practical for medical nanorobots is to confine activities to regions of the human body that are inaccessible to phagocytes. For example, certain internal tissues such as the lumens of glands, the urinary bladder and kidney tubules, and various surface tissues such as the skin are not regularly patrolled by phagocytes . The heart and muscle tissues also are relatively macrophage-poor. If reliable methods can be found for the remote (noncontact) detection of nearby phagocytes, akin to the detectability of bacterial metabolic chemical plumes (, Section 8.4.3), then most motile nanorobots should be able to outrun any “pursuing” phagocytes.
If remote phagocyte detection methods cannot be made reliably available, and for nonmotile nanorobots, other contact avoidance techniques must be employed. One potentially useful approach is to make use of the natural mediators of cellular chemotaxis (movement along a spatial gradient or directed cell locomotion) and chemokinesis (general random movement or nondirected cell locomotion) . Specific chemicals are known to be chemorepellents, chemotaxis antagonists, chemotactic factor enzymes or antibodies, or negative chemokinesis agents for various cell types.
For example, monocyte migratory inhibition factor inhibits macrophage migration, with a maximum inhibitory effect at 1 ng/ml for both unchallenged and particle-challenged macrophages . Excess zinc immobilizes macrophages , and mononuclear cells cultured from hyperimmunoglobulin-E (HIE) patients produced a ~61 kD protein factor that nontoxically inhibited normal neutrophil and monocyte chemotaxis  while serum from those patients contained a 30-40 kD inhibitor of granulocyte and monocyte chemotaxis . Phospholipase A2 inhibitors and a ubiquitin-like peptide  inhibit neutrophil chemotaxis, leukocyte-specific protein 1 (LSP1) is a negative regulator of neutrophil chemotaxis , and polyamines such as putrescine at 1 mM and spermidine at 0.1-0.5 mM inhibit chemotaxis (but not phagocytosis or engulfment) by neutrophils in vitro . Granulocyte locomotion is also inhibited by diclofenac sodium, a nonsteroidal anti-inflammatory agent, at concentrations below 10 micrograms/ml , and eicosapentaenoic acid somewhat rigidifies the plasma membrane of human neutrophils, leading to reduced chemotaxis . In other experiments, chemotaxis by human neutrophils toward several common chemoattractants was inhibited by 80%-95%, maximally at a concentration of ~50 microM of the protein kinase inhibitor 1-(5-isoquinolinesulfonyl) piperazine, without affecting the random migration of these white cells .
Much phagocyte chemorepellent research occurs in the context of elucidating bacterial avoidance strategies strategies that might be mimicked by medical nanorobots. Some bacteria or their products inhibit phagocyte chemotaxis. For example, Streptococcal streptolysin O (which also kills phagocytes) is a chemotactic repellent , even in very low concentrations. Staphylococcus aureus produces toxins that inhibit the movement of phagocytes; granulocytes are almost immobilized when administered 12 micrograms/ml of purified S. aureus lipase . Pertussis toxin, produced by the bacterium Bordetella pertussis, inhibits chemotaxis of neutrophils and other phagocytes; a PMN-inhibitory factor (PIF) extracted from B. pertussis cells showed little cytotoxicity and inhibited chemotaxis of neutrophils . Fractions of Mycobacterium tuberculosis inhibit leukocyte migration , the Clostridium perfringens phi toxin inhibits neutrophil chemotaxis , and other “specific antigen” can suppress basophil chemotaxis. Phagocyte chemotaxis is generally reduced by antibiotics such as cefotazime, rifampin, and teicoplanin . Rifampin and tetracyclines inhibit granulocyte chemotactic activity. Leukocyte, lymphocyte and monocyte chemotaxis is inhibited by methylprednisolone and azathioprine, whereas only lymphocytes are chemotactically inhibited by cyclosporine. More research is required to select, or more likely to design, the ideal chemorepellent agent that might be secreted (perhaps at nM concentrations, ~1 molecule/micron3, or less) by, or surface-tethered to, medical nanorobots seeking to avoid contact with phagocytes. Note that bioactive substances released locally by nanorobots can later be retrieved by similar means, thus avoiding nonlocal accumulations of these substances during nanomedical treatment.
Chemorepulsion is adequate for a few devices on simple missions of limited duration, but large numbers of medical nanorobots on longer more complex missions will inevitably come into physical contact with a phagocyte. The least disruption to normal immune processes is achieved if the nanorobot surface can deny recognition to the inquiring phagocyte at the moment of physical contact. Surface-bound moieties are generally preferable to free-released molecules when large numbers of in vivo nanorobots are involved. For example, each nanorobotic member of an internal communication network (, Section 7.3.2), stationed perhaps ~100 microns apart throughout the tissues, must continuously avoid being ingested by passing phagocytes. An approach that relies primarily on antiphagocytic chemical releases risks extinguishing all phagocytic activity throughout the body, severely compromising the natural immune system.
By 2001, “long-circulating” phagocytosis-resistant particles  and stealth drug carriers  have become the objects of active and extensive research. It is well-known that nanoparticle adsorption and internalization by phagocytes are inhibited by the presence of a coating of polysaccharide (e.g., heparin or dextran) chains in a brush-like configuration, or by very hydrophilic coatings. Low phagocytic uptake is achieved using a surface concentration of 2%-5% by weight of PEG, giving efficient steric stabilization (e.g., a distance of ~1.5 nm between two adjacent terminally-attached PEG chains in the covering brush) and avoiding uptake by neutrophils . Experiments by Davis and Illum  suggest that polystyrene particles sterically stabilized with adsorbed poloxamer polymer could achieve an extrapolated zero phagocytic uptake using a ~10 nm thick coating on 60 nm diameter particles or a ~23 nm thick coating for 5.25 micron diameter particles, thus eliminating nonspecific phagocytosis. Another study found that pegylated sheep red blood cells (RBCs) were ineffectively phagocytosed by human monocytes, unlike untreated sheep RBCs. Electrical characteristics also are important. Phagocytosis of polystyrene beads (as measured by cellular oxygen consumption) appears strongly dependent on surface potential and thus upon fixed surface charge, and surface charge heterogeneity across domains as small as 1-4 microns can greatly affect phagocytic ability.
Rather than coatings which phagocytes cannot recognize at all, medical nanorobots alternatively could carry surfaces that phagocytes will recognize as “friendly.” For example, coatings that mimic natural immune-privileged cells could be used. Nanorobot exteriors could be covalently bound with essential erythrocyte coat components — a simulated RBC surface could be useful in the bloodstream, but might provoke a response in the tissues. Similarly, fibroblast-like surface might be useful in the tissues, but is not normally seen in the bloodstream and phagocytes might respond to its presence there. Simulated neutrophil or monocyte surfaces would be better, since these cells normally migrate from blood to tissues, hence the immune system expects to see these surfaces virtually everywhere; lymphocytes are likewise normally present in both blood and tissues but are also adept at passing through the endothelial lining, the lymphatic processes, and the lymph nodes without being detained or trapped, eventually returning to the arterial circulation. The ideal solution may be for the medical nanorobot to display a specific set of self-markers at its surface, perhaps including moieties such as CD47. CD47 is a surface protein present on almost every cell type that provides an explicit phagocytic inhibitor signal to NK cells and to macrophages .
Microbial pathogens employ similar strategies to create antiphagocytic surfaces that avoid provoking an overwhelming inflammatory response, thus preventing the host from focusing the phagocytic defenses . Enveloped viruses and some bacterial pathogens can cover their external cell surface with components that are seen as “self” by the host’s phagocytes and immune system, a strategy that hides the true antigenic surface. Phagocytes then cannot recognize the bacteria upon contact and the possibility of opsonization by antibodies to enhance phagocytosis is minimized . For example, Group A streptococci can synthesize a capsule composed of hyaluronic acid, the “ground substance” (tissue cement) found in human connective tissue. The streptococcal hyaluronic acid capsule is nonantigenic and thus very effective in preventing attachment of the organism to the macrophage . Additionally, the cytoplasmic membrane of Streptococcus pyogenes contains antigens similar to those found on human cardiac, skeletal and smooth muscle cells, on heart valve fibroblasts, and in neuronal tissues, resulting in molecular mimicry and an immune tolerance response by the host . Other examples include pathogenic Staphylococcus aureus that produces cell-bound coagulase which clots fibrin on the bacterial surface , the syphilitic agent Treponema pallidum that binds human fibronectin to its surface , and a variety of bacteria that cause meningitis that avoid phagocytosis either by preventing deposition of complement by sialic acid on the surface or by modification of lipopolysaccharide (LPS). Haemophilus influenza expresses a mucoid polysaccharide capsule that prevents digestion by host phagocytes; a few strains resist opsonization and have become serum resistant by modification of their LPS O-antigen side chains, rendering them “invisible” to host immune defenses.
What if the nanorobot has been recognized as foreign by a white cell? As the next line of defense, medical nanorobots can directly inhibit phagocytic binding and activation. In the case of receptor-mediated binding, dansylcadaverine, amantadine, and rimantadine induce inhibition of endocytosis of complement-coated zymosan particles by human granulocytes.
These drugs block receptor-mediated endocytosis, possibly by their actions on phospholipid metabolism , although dansylcadaverine is not an endocytosis inhibitor in cells lacking transglutaminase activity. Cell-bound or soluble protein A produced by Staphylococcus aureus  attaches to the Fc region of IgG and blocks the cytophilic (cell-binding) domain of the antibody; thus the ability of IgG to act as an opsonic factor is inhibited, and opsonin-mediated ingestion of the bacteria is blocked. In the case of nonreceptor phagocytic binding, medical nanorobots could emit or expose on their surfaces chemical surfactants which would repel the lipid bilayer wall, e.g., by reducing the nanorobot’s coefficient of adhesion to very low or even negative values (, Section 9.2.3).
Phagocyte activation can also be directly inhibited. Several pathways of phagocytic signal transduction have been identified , including the activation of tyrosine kinases or serine/threonine kinase C, leading to phosphorylation of the receptors and other proteins which are recruited at the sites of phagocytosis. Monomeric GTPases of the Rho and ARF families which are engaged downstream of activated receptors, in cooperation with phosphatidylinositol 4-phosphate 5-kinase and phosphatidylinositol 3-kinase lipid modifying enzymes, can modulate locally the assembly of the submembranous actin filament system that leads to particle internalization. It may be possible for nanorobots to affirmatively influence, modulate, or even extinguish the phagocytic activation signal by physical, chemical, or other means, perhaps using GTPase or kinase inhibitors  such as genistein (50 microM), herbimycin (17 microM), staurosporine and trifluoperazine; in many cases there are two or more pathways that must be simultaneously inhibited, although in a few cases these pathways may share a common inhibitor. CNI-1493 is a potent and well-known macrophage deactivator or “pacifier” .
The author thanks C. Christopher Hook, M.D., and Stephen S. Flitman, M.D., for helpful comments on an earlier version of this paper.
Copyright 2001 Robert A. Freitas Jr. All Rights Reserved
1. Kenneth Todar, “Evasion of Host Phagocytic Defenses,” University of Wisconsin-Madison, see at: http://www.bact.wisc.edu/microtextbook/disease/evadephago.html.
2. Robert A. Freitas Jr., Nanomedicine, Volume I: Basic Capabilities, Landes Bioscience, Georgetown, TX, 1999. See at: http://www.nanomedicine.com.
3. Isabelle Catelas et al., “Flow cytometric analysis of macrophage response to ceramic and polyethylene particles: Effects of size, concentration, and composition,” J. Biomed. Mater. Res. 41(15 September 1998):600-607; “Cytotoxicity and macrophage cytokine release induced by ceramic and polyethylene particles in vitro,” J. Bone Joint Surg. Br. 81(May 1999):516-521.
4. A. Perianin et al., “Diclofenac sodium, a negative chemokinetic factor for neutrophil locomotion,” Biochem. Pharmacol. 34(1 October 1985):3433-3438.
5. M. Lind et al., “Chemotaxis and activation of particle-challenged human monocytes in response to monocyte migration inhibitory factor and C-C chemokines,” J. Biomed. Mater. Res. 48(1999):246-250.
6. Erle E. Peacock, Jr., Wound Repair, W.B. Saunders Company, Philadelphia, 1984.
7. H. Donabedian, J.I. Gallin, “Mononuclear cells from patients with the hyperimmunoglobulin E-recurrent infection syndrome produce an inhibitor of leukocyte chemotaxis,” J. Clin. Invest. 69(May 1982):1155-1163.
8. S. Chikazawa et al., “Hyperimmunoglobulin-E-associated recurrent infection syndrome accompanied by chemotactic inhibition of polymorphonuclear leukocytes and monocytes,” Pediatr. Res. 18(April 1984):365-369.
9. G. Cohen, M. Rudnicki, W.H. Horl, “A peptide isolated from a patient on continuous ambulatory peritoneal dialysis has homology to ubiquitin and inhibits the chemotactic response of polymorphonuclear leukocytes,” Miner. Electrolyte Metab. 23(1997):210-213.
10. J. Jongstra-Bilen et al., “LSP1 modulates leukocyte populations in resting and inflamed peritoneum,” Blood 96(1 September 2000):1827-1835.
11. J.D. Walters et al., “Polyamines found in gingival fluid inhibit chemotaxis by human polymorphonuclear leukocytes in vitro,” J. Periodontol. 66(April 1995):274-278.
12. S. Spika et al., “The mechanism of inhibitory effect of eicosapentaenoic acid on phagocytic activity and chemotaxis of human neutrophil granulocytes,” Clin. Immunol. Immunopathol. 79(June 1996):224-228.
13. L. Harvath et al., “Inhibition of human neutrophil chemotaxis by the protein kinase inhibitor, 1-(5-isoquinolinesulfonyl) piperazine,” J. Immunol. 139(1 November 1987):3055-3061.
14. J. Rallof et al., “Interference of Staphylococcus aureus lipase with human granulocyte function,” Eur. J. Clin. Microbiol. Infect. Dis. 7(August 1988):505-510.
15. S. Utsumi et al., “Polymorphonuclear leukocyte-inhibitory factor of Bordetella pertussis. I. Extraction and partial purification of phagocytosis- and chemotaxis-inhibitory activities,” Biken. J. 21(December 1978):121-135.
16. B. Van Vlem et al., “Immunomodulating effects of antibiotics: literature review,” Infection 24(July-August 1996):275-291.
17. V.P. Torchilin, “Polymer-coated long-circulating microparticulate pharmaceuticals,” J. Microencapsul. 15(January-February 1998):1-19.
18. S.S. Davis, L. Illum, “Polymeric microspheres as drug carriers,” Biomaterials 9(1988):111-115.
19. R. Gref et al., “Stealth corona-core nanoparticles surface modified by polyethylene glycol (PEG): influences of the corona (PEG chain length and surface density) and of the core composition on phagocytic uptake and plasma protein adsorption,” Colloids Surf. B Biointerfaces 18(1 October 2000):301-313.
20. Per-Arne Oldenborg et al., “Role of CD47 as a Marker of Self on Red Blood Cells,” Science 288(16 June 2000):2051-2054.
21. E. Whitnack, A.L. Bisno, E.H. Beachey, “Hyaluronate capsule prevents attachment of group A streptococci to mouse peritoneal macrophages,” Infect. Immun. 31(March 1981):985-991.
22. K. Krisher, M.W. Cunningham, “Myosin: a link between streptococci and heart,” Science 227(25 January 1985):413-415.
23. M. Garcia Gil, M. Sanchez Crespo, “Dansylcadaverine and rimantadine inhibition of phagocytosis, PAF-acether release, and phosphatidylcholine synthesis in human polymorphonuclear leukocytes,” Immunopharmacology 6(December 1983):317-325.
24. B. Jansson, M. Uhlen, P.A. Nygren, “All individual domains of staphylococcal protein A show Fab binding,” FEMS Immunol. Med. Microbiol. 20(January 1998):69-78.
25. K. Kwiatkowska, A. Sobota, “Signaling pathways in phagocytosis,” Bioessays 21(May 1999):422-431.
26. O. Dorseuil, M.T. Quinn, G.M. Bokoch, “Dissociation of Rac translocation from p47phox/p67phox movements in human neutrophils by tyrosine kinase inhibitors,” J. Leukoc. Biol. 58(July 1995):108-113.
27. C. Granert et al.., “Suppression of macrophage activation with CNI-1493 increases survival in infant rats with systemic Haemophilus influenzae infection,” Infect. Immun. 68(September 2000):5329-5334.
Note: An article by Robert A Freitas Jr. on the potential
applications of advanced nanotechnology to dental care appeared
in the November 2000 issue of the Journal of the American Dental
Association (JADA). That article is now available from the JADA
website (http://www.ada.org/prof/pubs/jada/index.asp). To
access the article, click on the link for Archives, and choose the
options for the November 2000 issue. In the listing of the issues
contents, choose the Nanodentistry article. The article is available
as either a HTML web page or an Acrobat PDF file.