Posted 17 January 2004 - 09:52 AM
Why Wait? Somatic Editing Now!
Ray Van De Walker ©2003
Abstract: Many futurists believe that cellular repair of organisms requires techniques beyond state of the art, i.e. “molecular nanotechnology.” Preliminary experiments have adapted computer-aided prototyping to the construction of tissues, in an emerging technique called “organ printing.” This article speculatively describes integration of optical scanning, in-silico editing, and direct assembly of bodies from cells, using photolithographically-produced micromachines. Speculatively, with validation and regulatory approval, whole functioning human beings might be reconstructed, with neural systems copied from recently deceased or cryobiologically preserved patients. An organizational initiative is recommended.
Ray Van De Walker is a biomedical software engineer. He designed and implemented real-time mechanism-control software used in four semi-automated surgical devices, three of which have entered FDA-approved clinical use. He currently works for Curlin Medical LLC, a maker of fine ambulatory infusion pumps. His opinions here are his own. He has a B.S. in computer science, with twenty years of experience developing high reliability embedded software for a variety of devices, including printers and scanners. He also has a B.A. in Philosophy, with concentrations in bioethics and political theory. Both degrees are from the University of California, Irvine. He is married with children, reads science fiction, sings and plays flute.
1. Introduction
Somatic editing using molecular nanotechnology will probably have several phases: 1. Scan a body. 2. Edit the data. 3. Assemble cells. 4. Assemble a body from cells.
Since healthy, long-lived mammalian cells can now be cultured, fermentation can replace 3 for now. Machinery for 1, 2 & 4 seems grossly feasible, because we can already construct micromachinery.
This plan substitutes natural molecular nanotechnology for artificial, and depends on the cells to reproduce and “heal” small problems with a less-precise assembly. It provides a starting point to develop increasingly advanced somatic assembly technology. When molecular nanotechnology becomes available, somatic assembly can only improve.
Even with an unlimited budget, the machinery casually described here will take at least five years to construct and validate. The risk of combining the mechanisms in these new ways is large commercially, but merely “challenging”, rather than “impossible”.
The major risk with this approach is learning how to construct real organisms. The distant techniques of molecular nanotechnology have obscured these problems, which are fundamental to somatic assembly.
First experiments can assemble small organisms with existing robotic micromanipulators. A mouse-sized set of machinery is reasonable as an early purpose-built somatic assembler. Once mice can be built, scaling to primates and people is a reasonable commercial risk.
Specialists can no doubt improve on the following technical approaches. Please do so, and thanks in advance.
2. The Process in Brief
1. The organism is scanned, including subcellular neural structures. Cells can be sampled in the scan, or biopsied. Nondestructive scans are not ethically required if a validated assembly process exists, and other therapeutic options are exhausted.
2. The data and cell samples are reduced to a standardized archival format, permitting secure storage, copying, and processing by the widest possible variety of procedures.
3. A medical team may edit the data file with automated tools to heal or improve the body. Optimizations may reduce the cost of assembly as well. A team is necessary because surgeons specialize. Expert systems and organ templates might reduce medical time and cost.
4. The anatomy is translated to an assembly format. The anatomy must have a functional vascular tree at every stage of construction.
5. In parallel with editing, the cells should be modified, selected and grown as needed for assembly, healing or improving the body at a cellular level.
6. The anatomy is assembled. The micromechanical assemblers described here have estimates between one and six days (see following discussion). The assembly time for a mouse should be a few hours with a 4cm2 micromechanical assembly surface, which would enable effective research programs for such problems as neuroassembly.
7. Conventional therapies then repair the anatomical and functional sequelae of assembly, excising nonfunctional tags of material, transfusing blood, closing the vascular ports, and synchronizing somatic systems.
8. There is likely to be an extended convalescence, perhaps with physical, sensory and cognitive therapies. Near-term somatic assembly technologies produce close approximations to an anatomy, and then depend on the cells to heal to confluence.
3. The Value of Somatic Editing
3.1 Positives
1. Cellular assembly and fusion to confluence have been grossly validated. Mammalian endothelial cells deposited as stacked rings in a gel matrix have grown together into a structure approximating a small blood vessel.[ ][ ][ ]
2. In principle, a somatic edit should be able to cure genetic disease, infectious disease and injuries, including fatal traumas, cancer and whole-body poisoning and freezing.
3. In principle, editing appears able to cure hypothesized causes of aging in an individual, including mitochondrial failures, somatic mutations, cellular senescence, telomeric shortening, oxidative damage, glycation, and waste accumulations in and out of cells.
4. An archival data format could mitigate accidental death by serving as a backup analogous to a computer backup. It might substitute for suspended animation.
Therefore, any group that opposes involuntary death should support somatic editing.
5. Editing could implement extreme changes to an organism’s structure or function, including cosmetic changes. Of course, extreme changes are extremely dangerous. Validation will be required.
6. Unlike other surgical or cellular therapies, editing can theoretically repair structural damage to soft tissues, such as scarring, circulatory defects, retinal and inner-ear damage.
7. Unlike conventional surgeries, the operation need not improvise, and can simulate and abort in silico without risk to the patient. Even if a reconstruction dies, the team can try again from the archive with a more conservative approach, or even wait years and try later with better technology.
8. In principle, editing provides a single cure for multiple conditions at a fixed or decreasing cost. It sets a ceiling on the cost of medical therapies. Society can expect rapid cost reductions, as health-care organizations apply it to more problems.
9. Editing is an extreme of tissue engineering, which has been providing transplantation products (skin for burn victims) for ten years[ ], and which is subject to continuing research[ ].
3.2 Negatives
1. The resulting organism is only a reproduction of the original. Both identity, and miscopying are concerns.
2. Available scanning technology is destructive. Clearly a destructive scan is a last resort, suitable only for fatal terminal disease. It makes frequent backups impractical.
3. Somatic editing can copy organisms, which then compete for exactly the same resources. Also, “habeas corpus” and kidnapping laws cannot prevent secret abuse of secret copies.
4. Cruel new chimeras become possible, e.g. an organism of one species’ shape, constructed of an unrelated species’ cells.
5. Most people have rare anatomical features in at least one organ system. Since all organ systems would be assembled, if the rare features cannot be replaced, or reproduced, the reconstruction could die or have poor health.
6. Though simple in principle, somatic editing will be complex in practice, and is supposed to produce a uniquely complex result. If it fails, a therapeutic response may be difficult.
7. The process might be very expensive. It takes a team of surgeons, working together, as well as rights to substantial intellectual property, autologous cell modification and culturing, and the use of expensive biomedical equipment over at least several days. However, many people have multimillion dollar health care coverage, and might choose a cure over paliation of a fatal disease.
4. Somatic Editing in Detail
4.1 Assembly
Assembly is technically the riskiest part of somatic editing. It is presented first to motivate interest in the feasibility of the other parts.
4.1.1 Somatic Assembly Estimates
Time is all that prevents surgeons from assembling tissues by hand. So let us estimate the time to assemble a person. If it fits in the time of a long surgical operation (a few days), with grossly feasible equipment, then it may be possible to build the assembler and assemble people.
The basic equation is
time = items-to-place/(manipulators*placement-rate)
Say that a (relatively) inexpensive machine assembles a slender human frame on the plane that has the smallest areal cross section. An assembly plane of 40x50cm will handle most human beings after editing. This is 2000 cm2.
Say that most of the body can be assembled from 20um cubical cells (mammalian cells range from 5 to 50um and are cubo-octahedral). In this case, the assembly planes will have to assemble 5x10^8 cells, times two meters of height. This is about 5x10^13 items (don’t despair). This test case is a block, 40x50x200cm, to see if such a machine is possible. Real people are smaller.
Let each micromanipulator (see following discussion) on the assembly plane be a third of a millimeter on a side. The ink-jet printers used for prototype “organ printing” are unsuitable because they cannot perform the 6-axis positioning needed for neuroassembly and osteoassembly.
Either piezoelectric or magnetostrictive actuators can raster the assembly surface.[ ] Let the raster be 1/3 by 2/3 mm. The raster allows a manipulator to bond hundreds of items per layer. The overlap lets adjacent manipulators place items for a failing manipulator.
Redundant machinery is more credible because the machine’s parts can be imperfect. With 50% overlap, the average assembly speed slows from 100% to 66%, because the two neighboring manipulators must do 1.5 times the work to compensate for a failed unit.
Assembly could start by attaching cells to a stretched, loosely-woven cloth of fibers similar to resorbable sutures. After that, a cradle could be assembled along with the body, a common practice in 3d printing.
Let each micromanipulator be able to place one hundred assemblies per second. This is fast, but there is margin. If the assemblies move 20um (toward the flesh) at a meter per second, they reach the assembly plane in 0.0005 seconds. If the cycle is “go down, bond, return, fetch next item,” then 90% of the cycle time can be used to grab the item from the transporter and bond it to the body. This is a reasonable cycle budget. If it is not, then perhaps it could be slower, and more micromanipulators can be brought into play.
The assembly surface might be controlled with computer logic grossly similar to that of a video screen. High-end graphic processors now perform billions of operations per second.
Let there be two assembly surfaces that start in the middle and work outwards. This doubles the assembly rate, greatly reducing the risk to the patient, but adds a slight risk from misregistration. No doubt the machine will have an alignment process. Two assembly surfaces can also clean and repair each other.
The assembly time is:
time = things-to-place/(surfaces*assembly-surface*assembler-density*items/sec)
time = 5x10^13/(2*2000*900*66) = 210,438 seconds = 2.4 days
This is a reasonable surgical time, so the machinery is grossly feasible.
4.1.2 Avoiding lesions from systematic defects
A test article should be produced and inspected before each assembly. This should find most problems, and allow the assembler to be diagnosed and repaired.
However, if one of the manipulators develops an undetected defect, the resulting tissue defects would stretch in a line or plane through the tissue, causing a lesion.
So, the positioning actuators and software should be designed so that, every few placements, the assembly surface jogs a pseudo-random distance and direction. This will distribute any defects, at the cost of some time, and more expensive nanopositioning actuators. Natural healing will fix such small defects.
4.1.3 Neuroassembly
No one knows how to assemble brains. Cooling studies have proven that long-term memory is structural, not functional. [ ] Memories are also not formed by the pattern of connections.[ ] Synaptic theories of memory are persuasive, but no one is sure that they are right.[ ] A machine that can build mice and their neuroanatomy from scratch should be able to authoritatively falsify the synaptic theory.
If memory is encoded in synaptic structure, there needs to be a way to encode them. Perhaps the machine could select from several hundred types of synapses and plug them into an astrocyte. This would assemble a close approximation of a brain. Similar problems appear for any other theory of memory, so one can still estimate feasibility. The author calculates a harmonic mean of less than 300 dendrites per astrocyte, each made from a harmonic mean of a 1/3mm-long fiber and having its own synapse. The astrocyte would also have a 2mm axon (another harmonic mean).
A human brain has about 200 million cells, so there’s less than 1.2x10^11 assembly items. These fibers could be placed with methods similar to automated wiring systems (e.g. wire-wrap[ ]).
An alternative, more difficult approach is to pick and place segments of fibers. This multiplies joints, and each joint is an opportunity to fail. However, it may be the only option to assemble the spinal cord. With 20um segments, this approach leads to 1.5x10^14 items to assemble in a human brain, which is impressively large, but still grossly feasible.
Estimates may vary, but neuroassembly seems grossly feasible.
4.1.4 Soft tissue assembly
Existing cellular assembly methods use ink-jet insertion of cells into a matrix of resorbable gel.[3] With these, the chief challenge appears to be the weakness of the matrix. At this time, these biocompatible gels cannot withstand normal blood pressures in typical blood vessels.
With direct assembly, strong resorbable fibers similar to real extracellular matrix can be oriented and embedded in the assembly. With this technique it is mechanically possible to construct a viable organism with a merely more expensive and difficult machine.
4.1.5 Osteoassembly
Units of bone fiber (calcium salts in ostin) can be positioned with direct assembly. This permits the constructed bone to have structures very similar to natural bone (harvesian channels, trebucular webbings and the like), but with all the advantages of computer-aided structural design. Related projects are in progress, though assembly methods are unclear. [ ]
4.1.6 Bonding
Several bonding systems are obvious: Resorbable staples and sutures, glue [ ], and ultrasonic welding. [ ] Note that welding is not the high-energy focused ultrasound process customarily used in surgery. [ ].
Ultrasonic welding is known to join dissimilar materials at (relatively) low temperatures with very localized welds. In cellular ultrasonic welding, the integument of a new cell would be rapidly bounced or rubbed against already emplaced cells. Ideally, the change of acoustic impedance when the surface touches the emplaced extracellular matrix would induce localized heating to weld the integuments. Cells bond to an extracellular matrix, so this matrix, which resembles a composite plastic, would be the material actually welded. The weldments will be temporary and digestible, like any cauterization.
A micromanipulator might perform ultrasonic welding almost unaided, without foreign substances or consumables, reducing mechanical, supply, sterility and biocompatibility complications.
If ultrasonic welding is impractical, electrical or RF resistance heating should be possible, with similar advantages, but perhaps more damage.
Resorbable staples could also aid automated assembly. They might be faster than welding, and much stronger than the extracellular matrix, to assemble functional bones. [ ] Staples also avoid heat or vibration damage, but they may cause other mechanical damage. Biocompatibility and sterility become design issues, but there are a wide variety of materials, and a micromanipulator should be able to emplace staples without grotesque design modifications.
Of course, thermoreversible resorbable gels have been established as cellular printing media. Composites with fibers might be suitable for real use.
4.1.7 Micromanipulator
The author envisions some combination of vacuum-pencil and claw on a photolithographed Stewart platform actuated by electrostatic linear motors. [ ][ ][ ] Force and position feedback could be integrated into the motors, or provided by separate capacitance cells. The joints could be labyrinthine springs. The mechanism would be sealed within a polymer sheet. The platform might have to be pulled from the lithographic surface by an assembly step during manufacturing.
Most likely, each type of assembly item would need its own assembly cycle. Knives should be in reach to trim items before bonding. For greater strength or to solve other problems, the manipulator should be able to place staples, in addition to other bonding techniques.
A micromanipulator with force feedback can trivially operate as a force microscope, if the data paths are in the design. This mode could be very useful for quality assurance.
Somatic editing could start development with a conventional robotic Huxley-Wall micromanipulator, [ ] although it would lack the speed for any but the smallest organisms.
4.1.8 Cellular Logistics
The assembler needs to move assembly parts such as cells and fibers from the bioreactors to an assembly surface. In the medical world, it would be helpful if the bioreactor and assembler did not need to be collocated. Culturing the cells is likely to be the most time-consuming part of the procedure. Also, surgeries and the manufacturing of surgical supplies are traditionally separated, because they require such different capitalizations, skills and tools.
So, there are several distinct logistic stages through which a cell will need to pass: Each stage is another R&D project, so minimizing them is good.
4.1.8.1 Carrier Surface
A “carrier surface” is a first step. Many mammalian cells need an attachment to prevent cell death by apoptosis.[ ] The first move would be from bioreactor solution to sheets of molded simulated extracellular matrix (ECM). [ ] Other parts might be molded in sheets. Elastomeric transfer printing reliably produces features as small as a few nanometers.[ ]
From the bioreactor facility to the surgical facility, the carrier sheets can be carried in cold vats of nutrient solutions, like other transplant tissues.
4.1.8.2 Physical queues
In the surgery, the carrier sheets would be inserted into the assembler. The assembler might cut strips of the carrier a single cell wide, and feed these into a sorting facility. Alignment and registration standards will be required, as well as each item’s assembly cycle.
The sorting facility would take pure cell types and other assembly items, and create a queue of items for each micromanipulator.
A predictable cycle time is needed so the assembly manipulators can raster together. The sorter therefore must queue items to arrive at fixed intervals. In a conveyor, a loop is needed between the sorting facility and the manipulator. If a real conveyor belt is used, then it has to be kept taut. A vacuum column similar to those of old computer tape drives would work.
A possible real design is to have a corrugated surface of “bucket-brigades” of micromanipulators on the back of the assembly surface to sort and queue items. This may be easier to configure and operate than a system of microconveyors.
4.1.8.3 Integrating the Assembler
The reference “design” above has at least two hundred thousand micromanipulators and conveyors. These might be formed by photolithography, but thousands of chips still need to be assembled to micronic precisions. Automated assembly will be required, and it might need to be present at the surgical site to set-up the assembler. Another assembly surface could clean, set-up, diagnose and repair an assembly surface.
4.2 Scanning
Scanning an organism is technology with a high social risk, even though its technical risk is lower than the assembler. Organismic scans are not required to develop transplantable organs, and there is no developed market for such a service. Also, the machine vision software to convert scans into schematics is a significant risk.
4.2.1 Optical Data Collection
The author envisions a destructive optical scan, using something like a large ultramicrotome to remove layers, and a digital confocal UV or confocal UV laser-microscopic [ ] histological scan, with a computer vision system identifying cells, connected fibers and organelles with fluorescent ligands (antibodies). Similar projects have occurred [ ], and the technology is almost off-the-shelf. [ ][ ]
Ultraviolet resolves features as small as 110nm, about 1/450 the diameter of an astrocyte (neuron).[ ] This can extract the gross interconnection map and timings of a neural system. After researching dye selection, lower-resolution data from the fluorescing frequencies should identify the cellular and chemical composition of all significant features, including (hopefully) the substrate implementing memory.
Some form of scanner and a neuroassembler can authoritatively falsify the synaptic memory hypothesis.
If optical scanning is inadequate, a higher-resolution scanning method (see below) may be used for the central nervous system.
If ultramicrotoming is not reproducible, a laser can vaporize controlled thicknesses of tissue. Holmium infrared lasers precisely debride cortical bone and other hard tissue. [ ] An optical focus can aim the laser at a shallow angle to the scan, to remove a shallow layer of tissue after it has been scanned.
4.2.2 Higher resolution scans
Arrays of scanning force microscopes would have even better resolution, but have less development as chemical sensors. Mass-produced micromechanical force microscopes have been developed as mass data storage devices. [ ] Related designs might operate in a wet environment and probe with antibodies to sense cell types.
If needed, scanning based on electron microscopy might use heavy metal stains to extract chemical information. It needs to mount the tissues, and cycle between staining and vacuum. Current manual processes are sufficiently delicate that specialists usually perform them.[ ], but with enough money, automation is possible.
4.2.3 Mapping stains to genetic expression
To reassemble an organism, cells with the correct genetic expression must be placed in similar locations. In research the stains used in scanning should be correlated with cell types and genetic expression. The basic plan would be to sample of a type of cells, stain part of them, and analyze another part with gene chips.
4.2.4 The Archival Format
Archiving ends the scanning process. The result is a data file and package of cells in a general-purpose, portable archival format, standardized across the industry.
A standardized “bridge data format” is crucial to an ethical editing technology. [ ] Essentially, it is a new form for a human being, and should be as reliable and standardized as a human body. Imagine the frustration and anger of a patient’s loved-ones, if his archive becomes unreadable. Or, if after scanning, a procedure is needed, and the loved one is in the wrong format. An open format also opens the parts of the process to competition, reducing costs.
To guard against interpretive errors, the original scan data should be preserved for reference. This is challenging. The scans produce color photos that can be compressed 200-fold by wavelet compression. [ ] At a 110nm (synaptic) resolution, just the brain (1300cc) requires 1.5x10^16 bytes. In movie-sized 5gb DVDs, that’s a solid cube 2.2 meters on an edge. Scanning the whole body (200x40x60cm) at a 20um (cellular) resolution only requires 180 DVDs.
Almost any schematic or CAD/CAM description can compress the scans by a thousand-fold or more. [ ]
4.3 Cell Sources
Each cell type can be saved during scanning, and later selectively cherished or modified in vitro, so that only healthy, youthful cells are retained. This is a substantial therapeutic step depending on known technologies. [ ][ ][ ]
4.3.1 Commercial cell lineages
Commercial cell lines could heal a patient of whole body DNA injuries like a massive radiation or mutagen exposure, or provide fast emergency repairs.
Commercial cell lines are likely to be safer and less expensive than groomed autologous cells, because the development expense is amortized over more users.
Therefore, insurance and other budget medical services might pay only to reconstruct people from banked groomed cell lines.
A number of therapies might restore the patient’s autologous genome. First, the commercial cells could be free of telomerase, and ALT. The groomed autologous cell line might then be a stem cell. Over time, it would replace the commercial cell lineages.
Another scheme is to pay for autologous cell grooming tissue by tissue over a period of years. A person’s next reconstruction could use the banked, groomed cells.
4.3.2 Advanced cellular technologies
In ten years, cellular modifications to end aging may be available. [ ][ ]
Later, each cell type might be generated from a groomed gene-scan of the organism. This organism would then restart with a perfectly homogeneous genome.
People might order genetic editing of their cells, and then have their immune system “tuned” to accept the changed expressions. Somatic editing could then “install” the revised genome.
4.4 In Silico Editing
A surgical team may simply play back a recording with an improved cell lineage. However, most people who opt for editing will have numerous subclinical defects, besides a few patent systemic defects. How can these be corrected economically?
4.4.1 Intellectual Property
In early animal research, simple playback may be used extensively. Certainly the machines will be validated to scan and play back a scan.
However, tissue engineers are developing robust models of human organs. Skin was successfully commercialized more than ten years ago, for grafting burn victims.
Patients may prefer engineered organs to those they grew. Tissue-engineered organs will have validated structures and functions. It is likely that they will perform at least as well as average natural organs.
Once standards appear for somatic file formats, validated organ designs are likely to become intellectual property, relatively inexpensive compared to the medical effort they replace.
Once the intellectual property is available, redesigning much of one’s body might be almost point and click assembly of parameterized parts. It will be substantially less complex than current transplant surgery because the surgery does not occur in real time (the surgeon will be working with a digital model of the patient), and the therapy has no problems with histocompatibility.
4.4.2 Vascular editing and correction
The circulatory system will probably have to be rearranged so that the capillaries (the smallest vessels) can be constructed in layers parallel to an assembly surface. This is the obvious way to assure that when a layer of tissue is completed, it is perfused as well.
Many organs might need vascular trees that do not grow in the direction of assembly. One technique to keep these perfused might be to build a temporary vascular tree, with emplaced sutures to close it.
All parts of the body need a controlled fluidic resistances. The resistances must be managed to minimize the instantaneous work of the heart and arteries, or the parts with higher resistance will be starved of blood, or the body will be created with circulatory disease. Further, every part of every tissue must be within two cell layers of a blood vessel.
It might be medically helpful to install extra valves, and venal and arterial shunts constructed of autologous tissues.
A software system to help design these conditions is credibly solble by a biomedical engineering team. Such editing software is best operated by a domain expert like a vascular surgeon.
4.4.3 Neural editing
A complete map of a patient’s neural network can be produced after the scan is processed. This will certainly enable physical rearrangement, and even provide the basic data needed for simulation or redesign.
The neural system extends through the whole body. A significant worry for those treated with “cryonic neurosuspension” is that it fails to preserve information about connections to the body. Perhaps a neural switching system, similar to a telephone switch, could solve this problem.
A neurosurgeon would supervise the translation of the scans for assembly. It might be very helpful to simulate the system to see if seizures are possible in the translated system. Many people have subclinical neurological defects. These might be detected and corrected, as well.
If the brain’s internal connections can be traced, analyzed, and reconnected it should, in principle be possible to replace the suborgans uninvolved with biographical memory and consciousness with generic designs, without affecting a person’s perception of their identity.
Such functional brain grafts are likely to be first developed to correct neural deformities and trauma. However, if they come to work better than most people’s natural organs, they may come to be preferred.
Also, if a properly weighted neuron can simulate a logic gate, then some digital facilities could be built into the brain A tiny computer might implement a “body control panel,” calculator, food value counter and clock. This is a very near-term possibility if the “user interface” is easy to integrate with the human sensorium, (e.g. morse code).
4.4.4 Reproductive systems
Germ cells of course should be groomed autologous cells. Oocytes in particular require developmental culturing techniques beyond the current state of the art, possibly because no speculative market has been obvious without tissue engineering of organs. Other than those, reproductive organs appear to be merely more organs, of the same sort as any others.
4.4.5 Orthopedic corrections
Many older persons have diseased joints and spines. In addition, most people have subclinical deformities from muscoskeletal damage, either breaks or bruises.
Many patients may retain an orthopedic surgeon to perform skeletal editing to correct these deformities. Suitable intellectual property is likely to be available. Tissue-engineered joints have already been produced. [ ]
An intriguing possibility is that normal skeletal materials and anatomy might be genuinely optimized using computer-based structural analysis.
4.4.6 Other corrections
Almost all people have imperfect eyes, ears or teeth. These tissues are quite odd, and may present special requirements, possibly including surgeries, implants or prostheses. Scanning is required because the unique neural structures of a patient’s senses will be very helpful or essential. However, a patient’s neural structures can probably be integrated in silico with intellectual property for an optimized physical structure.
The tissues of the eye, especially the cornea, are items of active tissue engineering research.[ ] Naturalistic enervated synthetic corneas have already been constructed, and are likely to be available as intellectual property.[ ]
The middle-ear’s bones may need to be supported with temporary scaffolds, and include the body’s hardest bone. If construction fails, prostheses are available. [ ]
Teeth also have specialized tissues. Two animal models have already produced tooth buds and regenerated dentin. [ ] Dental prostheses and implants are widely available.
4.4.7 Cosmetic corrections
Editing for cosmetic appearance is foolish, unless a patient needs editing for other conditions. However, any conceivable cosmetic correction seems possible with somatic editing, including fully functional sex and race changes.
An unobtrusive, practical set of cosmetic changes might optimize the amount of brown fat (to stay lean), thicken the skin, fix blemishes, make a patient’s face and body symmetric, and perhaps adjust the patient’s muscoskeletal and fat distributions to fit standard clothes.
4.5 Growing cells and Scheduling
The best sequence of operations may be to biopsy the cells and culture them before the scan. If the cells do not culture correctly, or cause delays, the scan can be deferred. Also the delay between scan and assembly would be shortened, lessening the period during which the patient lacks a body.
A body needs not just cells, but also other parts: soft, resorbable fiber, dendrite, axon, synapse and bone fiber. Culturing these may cause delays.
5. Required Research and Development
Tissue-engineering companies are aggressively pursuing development of synthetic transplantable organs. Typically, they plan to cure a disease with a single-organ cause, such as missing teeth, heart disease, corneal blindness, kidney or liver failure or juvenile diabetes.
This profit-oriented approach means that their research and development lacks an organismic focus. While somatic editing may be a logical application of tissue engineering, undirected development of editing might take longer than any person now alive would want to wait.
Note that molecular nanotechnology doesn’t solve most of the anatomical and medical problems of somatic editing. Any technology requires processing bits as well as atoms, and MNT does not develop safety-validated medical algorithms and devices. These need to start development now, if possible, and assemblers based on micromanipulators enable this.
A crucial step is to develop scanning systems. These have high social risks because there is no existing market. Very small (4cm2) systems would be genuinely usable for full-organism research, and only slightly more expensive than existing microscopes.
Another step is to produce small (4cm2) prototype assemblers able to perform neurologic and ostic assembly. Current organ printers use small fluid jets to place cells in gel. [1] These cannot orient ostic and neural structures in three dimensions, or even assemble working blood vessels. Initial research might start by extending existing robotic Huxley-Wall micromanipulators. [19]
Another step is to develop methods of culturing the needed cells and items, and mounting them for assembly. Another set of problems is to develop validated designs and models of organs and organ systems.
Since the organs and cells are part of their business, tissue-engineering firms may help with these.
Once these devices and services are available, researchers can begin to attempt assembly of small organisms and embryos.
The holy grail of this stage would be assembly of a fully-differentiated living mouse embryo. An embryo is likely to be easier than an adult mouse because the neural and gastrointestinal structures are less critical.
Once fully-differentiated mouse embryos can be assembled, the next step is to attempt assembly of adult mice. This stage will address system problems related to the larger size and less-resilient organism of adults.
Ultimately, the desire is to scan, and reproduce trained mice, demonstrating that neural memory has been conserved by the reassembly.
With trained mice assembled, it’s likely that scaling the techniques up to primates and humans will be a fairly ordinary biomedical engineering task.
6. Regulatory Considerations
The procedure is so radical that despite the expense, a conservative validation must complete a successful primate study, probably with M. Rhesus. Clear success in a primate would be 80 (+/-5) percent survival for two years, with no continuing therapeutic support, and significant preservation of training and social relationships. Note that the second and third make of machine would probably face similar requirements, even though supported by a smaller market-share.
A likely first medical use is reconstruction of a preconsenting recently-deceased patient with cancer or circulatory disease. No loss of life can occur, and the indicated set of patients is large. As validation proceeds, a successful therapy may be extended to riskier pathologies, perhaps including fatal trauma or whole-body freezing.
7. An organizational initiative
The author was bemused by the Methuselah Mouse Prize. Analogous prizes could guide research toward somatic editing. Somatic editing is in the interest of any group that opposes “involuntary death.” It is the only currently-feasible technique that offers backups to mitigate accidental death. Further, it can apply advanced cellular technologies to individuals.
An interested nonprofit group is uniquely qualified to coordinate research goals.
7.1 Develop the Equipment
The equipment does fill a needed market niche, even though it currently does not exist. No tissue-engineering project by itself can afford to develop scanners and cellular assemblers, but no competitive tissue-engineering firm could afford to lack them once they were developed.
Further, the developer could be practically assured of continuing sales of consumables. Profitable alliances between equipment manufacturers and cytologic companies seem likely.
One crucial step might be a series of grants or prizes to develop reliable somatic scanning equipment. This would concentrate resources to make such specialized equipment available for purchase by researchers. This is more critical than somatic assembly because there is no present market.
Another set of grants or prizes should further somatic assembly equipment, starting with a Huxley-Wall robot, and migrating to faster assembly systems.
Another important step would be to host, or support the development of public-domain archival formats to store organismic descriptions. It should be chaired by a very hard-nosed system engineer who can prevent the sillier parts of a design-by-committee.
Software development probably does not need to be supported. Once equipment and data interchange formats are available, software will arise naturally as researchers work with the data. Researchers will naturally migrate effective software to for-profit companies.
7.2 Get the research done
Another important step might be prizes for construction of working mouse organs. These are credible R&D steps for existing tissue-engineering firms. Winning such a prize would be a public relations coup for a tissue-engineering firm, while simultaneously advancing the overall paradigm of somatic editing. Since the intellectual property for mouse organs has relatively little commercial vale, it would form a loss-leading product line, most likely sold near cost to researchers.
Of course, these sales grow the future market, eventually providing a market (somatic editing) for the firms’ intellectual property in human organs.
The prizes should be progressive, and cumulative, concentrating funds on the development of the last few organs. The last few organs are likely to end up taking most work to reproduce.
The last prize could be for assembling a trained adult mouse
8. Conclusion
Near-term somatic assembly technology has been enabled by the combination of practical photolithographed micromachinery and cellular technologies. Development might not occur, or might take an unreasonably long time without guidance. The guidance could be as simple as a bread-crumb trail of prizes, offered by a non-profit institute.
