Nanomedicine, Volume I: Basic Capabilities

© 1999 Robert A. Freitas Jr. All Rights Reserved.

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


8.2.5 Organography and Histonavigation

Measured by volumes, the typical ~0.06 m3 adult male body contains ~38% hydrated protein, ~15% fat, ~14% expandable volumes and ~13% fluids. More importantly, ~60% of the volume of the human body is comprised of well-defined and highly specialized organs (Table 8.9) if skin and blood are included, as is commonplace.

Earlier Sections have described numerous physical routes through which all the organs can be accessed by a traveling medical nanorobot. Simple maps describing each of these routes can be made available to the nanodevice within a modest onboard data storage budget.

In addition to map-reading and precision positional navigation to ~3 micron accuracy (Section 8.3.3), with the help of a navigational network a nanorobot might verify that it had arrived at its intended organ by detecting various asymmetries that exist throughout the human body. As trivial examples, the heart lies mostly on the left side of the chest while the liver lies mostly on the right.* The larger right lung has 3 lobes with 10 secondary bronchi segments, while the left lung has only 2 lobes and 9 secondary bronchi segments. The left ventricle of the heart has a muscular wall twice as thick as the right ventricle. The right lobe of the liver is much larger than the left lobe, and has a completely different shape. The right ear statistically is slightly more sensitive than the left ear,827 undoubtedly involving small-scale anatomical asymmetries. The left and right halves of the face have distinct physiognomies.

* Normally the major lobe of the liver is on the right side of the abdomen and the spleen is on the left side, but about one in 10,000 patients is born with situs inversus -- a reversal of the left-right positions of these organs.1127

Aside from such gross anatomical measurements, there are many other (and more convenient) ways for a nanorobot to determine its histological location in the body. For example, in the kidney, the tissue pressure (easily measurable by nanosensors) is lower than the pressure under the tight renal capsule, and tissue pressure is lower in the brain than in the surrounding cerebrospinal fluid.363 (See also Section 8.4.2.) The speed of sound varies markedly from tissue to tissue (Table 6.7). For instance, a grid of nanorobots equipped with acoustic radiators and ~nanosecond clocks (Section 10.1) could measure and distinguish the speed of sound in the brain (1550 m/sec), the kidney (1575 m/sec), and muscle (1600 m/sec) over a grid distance of ~100 microns (Section 8.3.3). In the human eye, at 20 MHz, the highest acoustic sound velocities are found in the sclera (1597 20.3 m/sec); the lens exhibits the highest attenuation in the eye.483

Each tissue has a unique spectral transmission signature at optical wavelengths. For example, liver looks distinct from bowel because of differences both in absorbance and in the way the tissue scatters light (see Section 4.9.4), making feasible automated discrimination among tissue types.738 Both scattering and absorption measurements will require at least ~100 micron path lengths (~5 cell widths) for reliable tissue discrimination, requiring a cooperative activity among several nanorobots using timed test pulses. A ~1 pW radiator power budget producing up to ~107 optical photons/sec omnidirectionally using a ~1 micron2 emitter transmits up to ~103 photons/sec to a ~1 micron2 receiver located ~100 microns away (Section 7.2.3), which should be adequate for tissue-type discrimination. Tissues might also be distinguished based on cell membrane electrical conductivity (Section 4.8.7).

Each organ has a unique macroscale chemical signature. For instance, the anterior lobe of the pituitary gland (the adenohypophysis) is particularly rich in endocrine hormones including somatotropin (growth hormone), thyroid-stimulating hormone (TSH), adrenocorticotropic hormone (ACTH), follicle-stimulating hormone (FSH) and luteinizing hormone (LH). By contrast, the posterior lobe of the pituitary gland (the neurohypophysis) is rich in neuropeptides such as oxytocin and vasopressin.750 Grey matter, white matter, and myelin in human brain tissues have distinctive lipid compositions.1014

In some cases, the local distribution of vitamin concentrations alone might be probative. Table 8.10 gives vitamin concentrations for various tissues,585 which may reflect chemical variations present in the intracellular fluids, the constituent cells, or both. Thiamine (B1) occurs both free and phosphorylated,749 with elevated concentrations in the heart, liver, and kidneys, and low amounts in the stomach and skin; larger stores of riboflavin (B2) exist in the kidney and liver than elsewhere; niacin (B3) is more prominent in the liver and skeletal muscle; inositol is concentrated most heavily in the brain, kidney, and spleen. Cobalamin (B12) is stored primarily in liver, kidney, lungs, and spleen.752 Vitamin A (retinol) is fat-soluble and is stored principally in ester form in the liver but also in the lungs and kidneys.749 Water-soluble vitamin C (ascorbic acid) is present in highest concentration in tissues of high metabolic activity, most notably the adrenal and pituitary glands and in the intestinal wall.749 Vitamin D (cholecalciferol) is most concentrated in the skin,752 while vitamin E is stored mainly in the heart, lungs, muscles, and fatty tissues,749 and vitamin K is absorbed through the colon and stored in the liver;752 all three are fat-soluble. With enough chemical markers to choose from, a nanorobot equipped with suitable chemosensors can make high-probability locational inferences.

Tissues also have unique extracellular matrices (ECM) that can be sampled and identified by traveling nanorobots (Section For example, 19 different forms of collagen have been identified in various tissues, including 5 fibrillar types, 1 network-forming type, 4 fibril-binding types, 2 short-chain types and 1 long-chain anchoring type.971,1497,1498 Types III and VIII are found only in cardiovascular tissue, Type VII is found only in the skin, and so forth.521 Glycosaminoglycans (a component of the glycocalyx and ECM, usually complexed with proteins to form mucoproteins) are also tissue-specific. For instance, chondroitin sulfates are found in cartilage, bone, and cornea. Keratin sulfate I is found in the cornea, keratin sulfate II in loose connective tissue, heparin in mast cells, heparan sulfate in skin fibroblasts and the aortic wall, and hyaluronic acid in synovial fluid, vitreous humor, and loose connective tissue.996 It appears that ECMs contain embedded structures and glycoproteins (e.g., osteonectin, tenascin) unique to each tissue type that should be recognizable to nanorobots even in the complete absence of cells. Intermediate filaments comprising the cytoskeleton inside the cell are also tissue-specific (Section

Finally, organs may be identified by examining the surface antigens of their constituent cells. This is the most powerful technique. For instance, T cell receptors may be organ-specific: those comprised of g 2 subchains are found in the spleen, g 3 in the skin, g 4 in the female reproductive organs and the tongue, and g 5 in the lining of the intestinal tract.882 Protein components of the matrix, such as laminin and fibronectin, bind to specific integrin molecules on cell surfaces; through these 20+ integrins, the ECM transduces signals that regulate intracellular tissue-specific gene activity.971 The plasma membrane class B scavenger receptor for high-density lipoproteins, known as SR-BI, is expressed primarily by liver, ovarian, and adrenal cells.1038 Cell type identification is a major topic that is addressed at length in Section

Must a bloodborne nanorobot exit the bloodstream and enter the tissues in order to determine its proximity to the target organ? Most endothelial cells that coat the blood vessels servicing an organ or tissue system are likely to display organ-specific antigens on their luminal surfaces (Section In these cases, organ homing can be accomplished without exiting the blood vessels, simply by sampling the antigenic signature of the vessel luminal walls. However, in some cases it may be necessary for a nanorobot to penetrate the endothelial layer and directly examine the surfaces of the underlying tissue cells or the underlying ECM to determine proximity to the target organ.

Once a nanorobot verifies that it has arrived at the intended organ, the next major challenge is intraorgan navigation. This may be accomplished by following standardized maps which may be customized to the individual patient by prior somatographic surveys (e.g., Section and supplemented with direct positional and chemonavigational techniques. A complete somatographic description of all the organs in the body is beyond the scope of this book, especially since relevant material is readily available in the well-known Handbook of Physiology series published by the American Physiological Society, and in anatomy,853,854,863,866 histology867,935,936 and cytology531,938,939 textbooks generally. However, it is instructive here to convey a sense of place and scale by briefly reviewing the internal structure of the largest gland in the human body, the liver (see position in Figure 8.16).

The liver is a soft, plastic organ whose surface is pushed in by the surrounding organs. Thus the smaller left lobe bears concave impressions of the esophagus and stomach, while the larger right lobe bears concave impressions of the duodenum, the transverse colon, and the kidney.

Histologically, the liver is comprised of ~250 billion hepatic cells averaging ~18 microns in size. The human liver contains ~6 different cell types, but hepatocytes constitute ~80% of the cell population of the liver.935 The hepatocyte has 13 spherical nuclei (depending upon cell volume and position in the structure), each containing a nucleolus. These cells are bathed in flowing blood and arranged into ~1 million roughly hexagonal functional units ~1 mm wide called hepatic lobules (Fig. 8.26A). There are no interlobular septa (membranes separating the lobules) in the human liver.936

Around each hepatic lobule, liver cells form single-cell thick trabecular or hepatic plates (Fig. 8.26B). Portal blood from the gastrointestinal tract enters the portal vein through the hepatic plate, carrying absorbed nutrients into the portal venules. Each portal venule opens directly into large numbers of venous sinusoids (open cavities) formed by liver cells. The hepatic sinusoids form rich vascular networks with extensive but irregular anastomoses, making a three-dimensional spongework. Blood filters through the sinusoids and exits via the central vein (Fig. 8.26C), from which it flows into the hepatic venous system and from there into the inferior vena cava. Hepatic arterioles diverging from the hepatic artery supply an additional ~25% of the blood flow through the lobule, enriching the venous blood with freshly oxygenated red cells.

Bile canaliculi 12 microns wide936 lie between adjacent liver cells, forming networks of polygonal meshes each surrounding an individual hepatic cell. The canaliculi empty into a bile ductule at the periphery of the lobule. The ductules, in turn, join other ductules, producing vessels of increasing diameter, leading eventually to the wide hepatic duct and the common bile duct beyond.

The lymphatic system of the liver begins in the spaces of Disse, which lie between the surfaces of the hepatic plates and the endothelial walls of the venous sinusoids (Fig. 8.27). This peri-sinusoidal space contains blood plasma but no red cells or platelets, and is lined with vast numbers of microvilli projecting from the sinusoidal surface of the hepatocytes bordering the space.936 The spaces of Disse empty into small lymphatic vessels that run alongside the portal venules.

The sinusoid endothelial cells are flat and thin, with only their nuclei protruding slightly into the sinusoid lumen. Their walls are perforated with large fenestrae (~1 micron openings) numbering ~0.1/micron2. The fenestrations are dynamic structures that undergo changes in size and number in response to hormones and cytoskeletal fiber inhibitors.884,885 There are also numerous smaller pores measuring ~0.1 microns in diameter, mostly organized into groups of 10-15 pores called sieve plates (Fig. 8.27). Sieve plates are ~0.7 microns wide and number ~0.3 plates/micron2.936 These openings admit particulate material into the Disse space, for disposal by Kupffer cells.

Kupffer cells are ~15-20 micron stellate macrophagic cells, mechanically attached to the sinusoid endothelial cells. They partially occlude the sinusoid lumen but have no functional attachments to the endothelial cells or the hepatocytes. Kupffer cells are motile and have the ability to phagocytize (engulf and digest) particles of dirt, inert nanodevices (engulf and transport; Section, worn-out blood cells, and bacteria. Their customary positioning, predominantly at the periportal end of the sinusoids, confirms that they monitor arriving blood, looking for particles to remove from the flow. There are ~25 billion Kupffer cells in the liver,884,886 with a lifespan of many months.

With ~106 hepatic lobules each containing ~250,000 liver cells, a connectivity map listing every hepatocyte requires a 20-bit lobular address plus an 18-bit cellular address (within the lobule). Thus a simple hepatic lobular map (~1 mm resolution) takes just 20 megabits but a complete hepatic cellular map (~18 micron resolution) requires 250 billion 38-bit addresses, for a total map size of ~9.5 terabits. Of course, such complete maps may rarely be needed, since most medical nanorobotic tasks involve operations to be performed on a relatively small number of cells, or on a serendipitous basis, or can be guided by chemonavigation (Section 8.4.3).


Last updated on 16 April 2004