A Real Engineering Challenge – Artificial Red Blood Cells

When you’re thinking about “what can we do next”, you can think big or you can think small – very, very small. Robert Freitas at the Institute for Molecular Manufacturing (IMM) has such an idea – artificial red blood cells (RBCs). These would be nano-machines which could augment the oxygen and carbon dioxide carrying capacity of natural RBCs.

Oxygen is fundamental to life – mitochondria inside cells use oxidation to drive power generation for a cell, without which they will quickly die. The hemoglobin in RBCs is essential in transporting oxygen from lungs to tissues and carrying the carbon-dioxide generated in the consumption of food back to the lungs.

Any loss of effectiveness in this task due to anemia, respiratory problems or a host of other conditions can be life-threatening. Some methods to address insufficient oxygenation, such as hemoglobin formulations (hemoglobin separated from RBCs) have lower carrying capacity than RBCs and very short lifetime in the vascular system. Alternatives with higher carrying capacity and significantly longer lifetime would be a major step forward.

Dr. Freitas’ proposal (this has not yet been built as far as I know) is to construct a spherical nano-machine, ~1μm in diameter, which can easily pass through capillaries. The machine acts as a pressure vessel, pumping O[SUB]2[/SUB] in and CO[SUB]2[/SUB] out when in the lungs and pumping O[SUB]2[/SUB]out and CO[SUB]2[/SUB] in when in other tissue. Pumps are driven by rotors which can separate O[SUB]2[/SUB] and CO[SUB]2[/SUB] molecules from blood plasma.

Rotors are arranged around the spherical chamber, some to pump in and some to pump out, along with sensors to detect O[SUB]2[/SUB] and CO[SUB]2[/SUB] concentrations and a glucose-powered engine to drive the rotors (glucose is readily available in the bloodstream). O[SUB]2[/SUB] and CO[SUB]2[/SUB] are stored in tanks around the surface of the sphere; the center of the sphere is reserved for a water ballast chamber and the computer which will monitor sensors, control rotors and perform other functions.

The ballast chamber is provided to control buoyancy. This feature is not thought to be required during normal operation of the artificial cells in the body but would be useful for extracting these cells after they have completed their therapeutic purpose. Blood can be circulated from the body to a centrifuge where, with suitably adjusted buoyancy control, the artificial cells can be encouraged to separate out.

The computer has interesting constraints. The author expects that adequate computing capacity could be contained inside a sphere 58nm in diameter, consuming ~10[SUP]-14[/SUP] W, which would be a small percentage of the output of the glucose engine. Would be interesting to hear what ARM users think about this.

So far, what is described is challenging but maybe not inconceivable. The real challenge comes in manufacturing at volumes required to make this useful. The author estimates that a full load required to replicate RBC carrying capacity for one patient is ~5.10[SUP]12[/SUP] devices. If you could make a million devices for a penny, a single transfusion would cost ~$50k. Scaling down the load doesn’t really help – dropping to a 10% load will reduce the value of the solution correspondingly. So we have to get to much better than a million devices per penny. Geometrically that doesn’t seem impossible, but I‘m sure this will take a high order of MEMS device physics (and chemistry).

Final devices will of course be separated (each wrapped in a diamond-like shell to avoid degradation of the device and toxicity to the blood). This will truly be silicon dust (or more exactly diamond dust). There’s a great description in the paper on how a therapeutic dose is prepared starting with this powder. A cold glucose solution is added to the powder, along with any necessary salts, proteins, etc. Sensors on the cells detect glucose which they start pumping into their tanks. They then fill the oxygen tanks and finally fill the ballast tanks, at which point they sink in the solution. Any powder left on top indicates defects in those cells which should be skimmed off. A command is sent to the cells to expel enough water to reach neutral buoyancy; from there they are ready to be injected into the patient through an IV drip.

Probably this isn’t something we’re going to see very soon, but it does define a great stretch goal for where we could get if we really try hard. You can get more detail from Dr. Freitas’ paper HERE.

More articles by Bernard…