Google Custom Search

IMM Report Number 28: Nanomedicine

In conjunction with Foresight Update 46

How Nanorobots Can Avoid Phagocytosis by White Cells, Part II

By Robert A. Freitas Jr.
Research Scientist, Zyvex Corp.

Robert A. Freitas  Jr.
Robert A. Freitas Jr.

Part I appeared in Foresight Update 45.

Ingestion or phagocytosis of medical nanorobots [1] by white cells will occur in a series of well-defined steps. Normally inactive white cells are activated when they encounter a foreign particle, producing a change in metabolic activity and cell shape. During contact and recognition of the foreign particle, the phagocyte plasma membrane develops a local invagination or dimple. The particle is drawn inside and the dimple closes, often pinching off to form a small vacuole or phagosome consisting of everted cell wall membrane, trapping the particle inside the cell. The phagosome then forms a phagolysosome by merging with a lysosome, whose contents (including degradative lysozymes) are released into the smaller vacuole, attacking the enclosed foreign or denatured proteins. Afterwards the phagolysosomal vacuole may be absorbed or released to the outside at the cell’s outer surface via exocytosis, producing a large membrane flow. In cultured macrophages an amount of membrane equal to the entire surface area of the cell is replaced in ~1800 sec, and macrophages may ingest up to ~25% of their volume per hour [1].

In Part I of this paper [2], we described the initial antiphagocytic strategy for medical nanorobots which is to avoid phagocyte contact, recognition, binding and activation. In Part II, we assume that this initial strategy has either failed or has not been used, in which case the medical nanorobot has been recognized by, and become transiently bound to, a phagocyte. The best nanorobot strategies at this point are first, to inhibit phagocytic engulfment, and second, to inhibit enclosure and scission of the phagosome if engulfment has begun. Let’s look at these two approaches in more detail.

Even if a medical nanorobot has been recognized and has attached to the phagocyte outer surface (typically across a ~20 nm gap bridged by ~12 nm strands), the device can still prevent complete engulfment from taking place. Macrophages challenged with a particular type of target usually bind many more targets than they ingest [3]. Fortunately, internalization is a relatively slow process and most particles that become bound to the phagocyte surface are not ingested [3]. On rare occasions, phagocytosed particles are actually expelled.

Phagocytosis is an uptake of large particles governed by the actin-based cytoskeleton. Complement-opsonized (CO) and antibody-opsonized (AO) particles are phagocytosed differently by macrophages [4] — CO particles sink into the cell, whereas AO particles are engulfed by lamellipodia which project from the cell surface. During the ingestion of CO particles, punctate structures rich in F-actin, vinculin, alpha-actinin, paxillin, and phosphotyrosine-containing proteins are distributed over the phagosome surface [4]. These foci can be detected underneath bound CO particles within 30 seconds of cell activation, and their formation requires active protein kinase C. Complement receptor-mediated internalization requires intact microtubules and is accompanied by the accumulation of vesicles beneath the forming phagosome [4]. By contrast, during the ingestion of AO particles (Fcgamma receptor mediated phagocytosis), all proteins are uniformly distributed on or near the phagosome surface. Ingestion of AO beads is blocked by tyrosine kinase inhibitors (e.g., released from or tethered to medical nanorobots), although the phagocytosis of CO particles is not [4].

Phagocytic particle ingestion can require actin assembly and pseudopod extension, two cellular events that may coincide spatially and temporally but apparently use distinct signal transduction events or pathways [5]. Medical nanorobots that have become bound to the extracellular phagocyte surface may attempt to inhibit either or both of these signal transduction pathways.

In the first case, during actin assembly, engagement of particle-bound immunoglobulin IgG ligands by receptors for the Fc portion of IgG results in receptor aggregation and recruitment of cytosolic tyrosine kinase, especially Syk [6]. The onset of uptake is accompanied by tyrosine phosphorylation of several proteins, which persists for up to 3 minutes, is concentrated at phagocytic cups and nascent phagosomes, and is correlated with the accumulation of actin filaments. Phosphorylation of tyrosine residues occurs within immunoreceptor tyrosine activation motif (ITAM) consensus sequences found in FcgammaR subunits, which allows further recruitment and activation of Syk [6]. Syk tyrosine kinase activity is required for FcgammaR-mediated actin assembly, which is controlled by several GTPases, including Rac1 and CDC42 [6]. Rac1 and CDC42 (two Rho proteins involved in the signal transduction through the Fc receptors) are required to coordinate actin filament organization and membrane extension to form phagocytic cups, to allow particle internalization during FcR-mediated phagocytosis, and are involved in the phosphotyrosine dephosphorylation required for particle internalization [7].

Actin assembly can be inhibited by Clostridium difficile toxin B, which is a general inhibitor of Rho GTP-binding proteins [7]. Inhibition of Rac1 or CDC42 severely inhibits particle internalization but not F-actin accumulation [7]. In laboratory tests with cells, inhibition (via knockout of gene expression in a mutant line) of CDC42 function results in pedestal-like structures with foreign particles at their tips on the phagocyte surface, whereas inhibition of Rac1 results in particles bound at the surface that are enclosed within thin unfused membrane protrusions, demonstrating that Rac1 and CDC42 have distinct functions and may act cooperatively in the assembly of the phagocytic cup [7]. Phagocytic cup closure and particle internalization has also been blocked when phosphotyrosine dephosphorylation is inhibited by treatment of the phagocytic cells with phenylarsine oxide, an inhibitor of protein phosphotyrosine phosphatases [7]. Ceramide also inhibits tyrosine phosphorylation in human neutrophils [8].

In the second case, during pseudopod extension, phosphatidylinositol 3-kinase (PI3K) is recruited to the plasma membrane, triggering exocytosis from an internal membrane source, as is required for pseudopod extension [6]. (Macrophage spreading on opsonized surface is accompanied by insertion into the plasma membrane of new membrane from intracellular sources [5].) One or more isoforms of PI3K are required for maximal pseudopod extension, though not for phagocytosis per se; PI3K is required for coordinating exocytic membrane insertion and pseudopod extension [5].

Pseudopod extension may be partially inhibited using wortmannin (WM) or LY294002, which are two inhibitors of PI3K [5]. Both of these specifically inhibit phagocytosis without inhibiting Fcgamma receptor-directed actin polymerization, by preventing maximal pseudopod extension. Decreasing the size of test beads, and hence the size of pseudopod extension required for particle engulfment, de-inhibited phagocytosis (in presence of these inhibitors) by up to 80% at the very smallest submicron particle sizes. For larger (nanorobot-sized) foreign particles, phagocytosis is blocked before phagosomal closure. Both compounds also inhibit macrophage spreading on opsonized surfaces (i.e., on substrate-bound IgG) [5].

Amphiphysin II associates with early phagosomes in macrophages and participates in receptor-mediated endocytosis by recruiting the GTPase dynamin to the nascent endosome. There is a signaling cascade in which PI3K is required to recruit amphiphysin II to the phagosome, after which the amphiphysin II in turn recruits dynamin to the phagosome [9]. A modified amphiphysin II molecule with its dynamin-binding site ablated away inhibits phagocytosis at the stage of membrane extension around the bound foreign particles [9]. Both phenylbutazone and chloramphenicol also have shown an inhibitory effect on the engulfment stage of phagocytosis of yeast particles by cultured human monocytes [10].

As might be expected, bacteria already employ a wide variety of strategies to avoid engulfment when physically contacted by host phagocytes [11]. Some of these strategies could in principle be mimicked by medical nanorobots. Most commonly, many important pathogenic bacteria bear substances on their surfaces that inhibit phagocytic adsorption or engulfment. Resistance to phagocytic ingestion is usually due to an antiphagocytic component of the bacterial cell surface, such as:

  1. Cell Wall Substances [11] — polysaccharide surface slime (alginate slime and biofilm polymers) produced by Pseudomonas aeruginosa; O antigen associated with LPS of E. coli (smooth strains); and K antigen (acidic polysaccharides) of E. coli or the analogous Vi (K) antigen (microcapsule) of Salmonella typhi.
  2. Fimbriae and M Protein — fimbriae in E. coli [12], and M protein and fimbriae of Group A streptococci [12]. For example, Streptococcus pyogenes has M protein, a fibrillar surface protein whose distal end bears a negative charge that interferes with phagocytosis. Enterococci also have antiphagocytic surface proteins [13], such as M protein.
  3. Capsules — polysaccharide capsules of S. pneumoniae (unless antibody is present), Treponema pallidum, Klebsiella pneumoniae, Bacteroides fragilis, and Clostridium perfringens, and the Enterococci inhibit engulfment [11-13]. Haemophilus influenzae expresses a mucoid polysaccharide capsule of thickness ~1 microbial diameter which prevents digestion by host phagocytes, although many of these bacteria remain susceptible to opsonization [12]. The protein capsule on cell surface of Yersinia pestis resists engulfment [11].

Macrophages can also bind and engulf a variety of particles in the absence of specific opsonins, a process known as nonspecific phagocytosis [14], nonopsonic phagocytosis [15], or opsonin-independent phagocytosis. Polystyrene microspheres are often used to demonstrate this. For instance, during patocytosis of hydrophobic >0.5-micron particles by phagocytes, actin-independent capping of hydrophobic polystyrene microspheres on the plasma membrane precedes actin-dependent uptake of the microspheres into the surface-connected compartments [16]. Microsphere transport from plasma membrane invaginations into spaces created by unfolding stacks of internal microvilli are inhibited by administering primaquine [16]. Studies of non-specific endocytosis and binding of liposomes by mouse peritoneal macrophages also found that particle internalization declined markedly after anchorage of the cells to polystyrene substrate [17]. Inhibitors are potentially available to medical nanorobots to halt these processes too. For example, staurosporine selectively inhibits nonspecific phagocytosis while having no effect on receptor-mediated phagocytosis [14].

What if a medical nanorobot has become partially or wholly engulfed by a phagocyte? Can the vacuole still be prevented from pinching off and separating into a free intracellular phagosome containing the nanorobot (i.e., enclosure and scission)? More research is needed, but the answer appears to be yes.

Cells normally internalize soluble ligands and small particles via endocytosis and large particles via actin-based phagocytosis. The dynamin family of GTPases mediates the membrane destabilization, constriction, fission (scission) and trafficking of endocytic vesicles from the plasma membrane, but dynamin-2 also has a role in phagocytosis by macrophages [18]. Experiments reveal that early phagosomes (vacuoles) are enriched in dynamin-2, and inactive mutant versions of this molecule, if expressed, inhibit particle internalization at the stage of membrane extension around the particle [18]. This arrest of phagocytosis resembles that seen with PI3K inhibitors, preventing the recruitment of dynamin to the site of particle binding. Dynamin is a microtubule-binding enzyme with a microtubule-activated GTPase activity; phosphorylation engages its activity. Dynamin can interact with the actin cytoskeleton to regulate actin reorganization and subsequently cell shape [19].

Observations suggest that dynamin mediates scission from the plasma membrane of both clathrin-coated pits and caveolae during distinct endocytic processes [20]. For example, dynamin-1 is a 100 kD GTPase involved in scission of endocytic vesicles from the plasma membrane. It is present in solution as tetramer, undergoes self-assembly (following its recruitment to coated pits) to form higher-order oligomers that resemble collars around the necks of nascent coated buds. GTPase hydrolysis by dynamin in these collars is thought to accompany the pinching off of endocytic vesicles — dynamin may use GTPase hydrolysis physically to drive vesiculation, or may act as a classical G protein switch, or both [21]. (Purified dynamin readily self-assembles into rings or spirals, suggesting that it probably wraps around the necks of budding vesicles and squeezes, pinching them off [22]; the large GTPase dynamin is a mechanoenzyme [23].) Different dynamin isoforms may be localized to distinct cellular compartments but provide similar scission functions during the biogenesis of nascent cytoplasmic vesicles [20]. Once again, inhibitory tools that might be employed by medical nanorobots are potentially available. For example, anti-dynamin antibodies have been used to specifically inhibit dynamin function in cultured mammalian epithelial cells, inhibiting cellular uptake of external particles in these cells [20], and to inhibit clathrin-mediated endocytosis in hepatocytes [24]. Ca++ inhibits both dynamin I GTPase [25] and dynamin II GTPase [26] and may also serve as vesiculation inhibitors for fully engulfed medical nanorobots. Alternatively, butanedione monoxime, a class II myosin inhibitor, has been shown to prevent the purse-string-like contraction that closes phagosomes while not inhibiting the initial pseudopod extension [27].

One important remaining practical question is how the various inhibitory methods that we have identified will be specifically expressed in each class of medical nanorobots, but this is an issue for another time.


The author thanks Stephen S. Flitman, M.D., and C. Christopher Hook, M.D., for helpful comments on an earlier version of this paper.

Copyright 2001 Robert A. Freitas Jr. All Rights Reserved


1. Robert A. Freitas Jr., Nanomedicine, Volume I: Basic Capabilities, Landes Bioscience, Georgetown, TX, 1999. See at:

2. Robert A. Freitas Jr., “How Nanorobots Can Avoid Phagocytosis by White Cells – Part I,” Foresight Update No. 45, 30 June 2001, pp. 10-12. See at:

3. H. Bos, W. de Souza, “Phagocytosis of yeast: a method for concurrent quantification of binding and internalization using differential interference contrast microscopy,” J. Immunol. Methods 238(21 April 2000):29-43.

4. L.A. Allen, A. Aderem, “Molecular definition of distinct cytoskeletal structures involved in complement- and Fc receptor-mediated phagocytosis in macrophages,” J. Exp. Med. 184(1 August 1996):627-637.

5. D. Cox et al., “A requirement for phosphatidylinositol 3-kinase in pseudopod extension,” J. Biol. Chem. 274(15 January 1999):1240-1247.

6. S. Greenberg, “Modular components of phagocytosis,” J. Leukoc. Biol. 66(November 1999):712-717.

7. P. Massol et al., “Fc receptor-mediated phagocytosis requires CDC42 and Rac1,” EMBO J. 17(2 November 1998):6219-6229.

8. S.J. Suchard et al., “Mitogen-activated protein kinase activation during IgG-dependent phagocytosis in human neutrophils: inhibition by ceramide,” J. Immunol. 158(15 May 1997):4961-4967.

9. E.S. Gold et al., “Amphiphysin IIm, a novel amphiphysin II isoform, is required for macrophage phagocytosis,” Immunity 12(March 2000):285-292.

10. A. Odegaard, J. Lamvik, “The effect of phenylbutazone and chloramphenicol on phagocytosis of radiolabelled Candida albicans by human monocytes cultured in vitro,” Acta Pathol. Microbiol. Scand. C 84(February 1976):37-44.

11. Kenneth Todar, “Evasion of Host Phagocytic Defenses,” University of Wisconsin-Madison, see at:

12. C.G. Gemmell, “Changes in expression of bacterial surface antigens induced by antibiotics and their influence on host defenses,” Pathol. Biol. (Paris) 35(December 1987):1377-1381.

13. Lynn E. Hancock, Michael S. Gilmore, “Pathogenicity of Enterococci,” in Vincent Fischetti et al., eds., Gram-Positive Pathogens, ASM Press, Washington, DC, 2000. See at:

14. T. Hishikawa et al., “Calcium transients during Fc receptor-mediated and nonspecific phagocytosis by murine peritoneal macrophages,” J. Cell Biol. 15(October 1991):59-66.

15. I. Ofek et al., “Nonopsonic phagocytosis of microorganisms,” Annu. Rev. Microbiol. 49(1995):239-276.

16. H.S. Kruth et al., “Characterization of patocytosis: endocytosis into macrophage surface-connected compartments,” Eur. J. Cell Biol. 78(February 1999):91-99.

17. V.S. Goldmacher, “Macrophages. Inhibition of endocytosis by anchorage to substrate,” Exp. Cell Res. 154(October 1984):632-635.

18. E.S. Gold et al., “Dynamin 2 is required for phagocytosis in macrophages,” J. Exp. Med. 190(20 December 1999):1849-1856.

19. M.A. McNiven et al., “Regulated interactions between dynamin and the actin-binding protein cortactin modulate cell shape,” J. Cell Biol. 151(2 October 2000):187-198.

20. J.R. Henley, H. Cao, M.A. McNiven, “Participation of dynamin in the biogenesis of cytoplasmic vesicles,” FASEB J. 13(December 1999):S243-S247.

21. W. Yang, R.A. Cerione, “Endocytosis: Is dynamin a blue collar or white collar worker?” Curr. Biol. 9(15 July 1999):R511-R514.

22. M.H. Stowell et al., “Nucleotide-dependent conformational changes in dynamin: evidence for a mechanochemical molecular spring,” Nat. Cell Biol. 1(May 1999):27-32; J.E. Hinshaw, “Dynamin and its role in membrane fission,” Annu. Rev. Cell Dev. Biol. 16(2000):483-519.

23. M.A. McNiven et al., “The dynamin family of mechanoenzymes: pinching in new places,” Trends Biochem. Sci. 25(March 2000):115-120.

24. J.R. Henley et al., “Dynamin-mediated internalization of caveolae,” J. Cell Biol. 141(6 April 1998):85-99.

25. J.P. Liu et al., “Calcium binds dynamin I and inhibits its GTPase activity,” J. Neurochem. 66(May 1996):2074-2081.

26. M.A. Cousin, P.J. Robinson, “Ca(2+) influx inhibits dynamin and arrests synaptic vesicle endocytosis at the active zone,” J. Neurosci. 20(1 February 2000):949-957.

27. J.A. Swanson et al., “A contractile activity that closes phagosomes in macrophages,” J. Cell Sci. 112(February 1999):307-316.

IMM would appreciate learning your thoughts on the above article.