The Amazing Blood Machine

NATICK, MA--Like ancient alchemists who tried to turn lead into gold, engineers at ZymeQuest hope to turn Type A, B, and AB blood into the universal Type O that can be given to anyone. Unlike the alchemists, ZymeQuest appears to have succeeded.

The medical implications of this innovation are staggering. Each year, health-care professionals perform four million blood transfusions. Nearly 500 times per hour someone needs a blood transfusion. Hospitals spend an estimated $2 to $3 billion a year to test and retest this blood. And these same hospitals throw out about 6% of the blood they receive because it spoils before the hospitals get patients whose blood type matches the donor blood they have on hand.

ZymeQuest's Universal Blood Machine not only increases the safety of blood transfusions, but the process it enables costs less to carry out than repeated blood-type testing, saving hundreds of millions of dollars annually for hospitals and other health-care centers.

The engineering implications are also far-reaching in the lessons they hold for designers of other medical and non-medical devices. That's because many of the design challenges themselves were universal, among them the need to:

Achieve high tolerances.

• Meet tough performance requirements.

• Keep costs down so the company can make a profit in a small market.

Software--including Pro/ENGINEER from Parametric Technology Corp., Mastercam from CNC Software, and PartAdvisor from Moldflow--enabled engineers to design the machine's complex components.

Use of off-the-shelf plastics from a variety of vendors enabled engineers to make critical components--cassette, seal assembly, and bags--disposable, says Glen Jorgensen, director of ZymeQuest's systems development laboratory and project leader. Since there are only a few hundred blood centers in the U.S., Europe, and Japan who would be customers, making those components disposable so users will have to re-stock them regularly is key to profitability of the system.

In the beginning. The system had its origins some 14 years or more ago as the dream of Jack Goldstein of the Kimball Research Institute at the New York Blood Center. Goldstein's research into a way to alter the chemistry of the blood's red cell surface to convert all types of blood to type O seemed doable (see sidebar). What was needed was a working system to turn the dream into reality.

ZymeQuest, with headquarters in North Andover, MA, also had enough faith in the conversion process to license it from Goldstein in 1991 and arrange to continue the research using the New York facility as a scientific base.

Enter Jorgensen, who at the time was an independent consultant, and his business partner Don Barry. Wanting to get the conversion process developed as quickly as possible, ZymeQuest contracted with Jorgensen to build a prototype of the system. Working out of Jorgensen's basement, Jorgensen and Barry, ZymeQuest's manager of electronic systems, had a working model of the Universal Blood Machine ready for Phase I testing in eight months. Pleased with the results, ZymeQuest took the design team under its wings and set up a design and development laboratory in Natick, MA, to develop the next-generation conversion system.

Putting plastics to work. Key system components include a disposable cassette and a rotating seal assembly. The importance of this aspect of the design cannot be understated. Each component in the disposable system had its challenges when it came to the selection of the right material for the job. First, engineers decided to use FDA-approved Class VI implantable plastics across the board. Other criteria included: moldability, structural integrity, and ability to withstand gamma sterilization.

Design of the critical cassette and rotating seal assembly proved far from easy, Jorgensen notes. Tolerance and performance requirements of the components were especially demanding.

"The business depended on us to develop devices that not only met the performance challenge and were reliable--but they had to be inexpensive," Jorgensen recalls. "Since the components had to be made of plastic, this made dimensional stability and strength essential, particularly in the highly stressed, thin sections that support the seal faces of the rotating seal and the valve seats in the cassette."

Maximizing molding. Initially, engineers machined the plastic components to quickly put a working system together. However, the parts were far from ideal. Jeremy Fennelly, plastics engineer and consultant, joined the team to design the final configuration and optimize the injection-molding tool design and the materials used, starting with the design of the cassette.

Simulation through the use of molding software helped make this practical. The software employed, Part Adviser^TM from Moldflow Corp. (Lexington, MA), gave Fennelly the power to predict how a part would fill, offered a picture of the optimized gate location in the mold, and optimized injection-molding conditions. The material selected for the cassette: Makrolon 2458 polycarbonate (PC) from Bayer Polymers Div. (Pittsburgh).

ZymeQuest designers used the Moldflow software to model various gate configurations. With each iteration, they could look at flow-front temperatures and pressures, weld lines, air entrapment, fill time, and pressure drop.

"This analysis gave us the confidence to move forward with an expensive production tool," says Jorgensen. "We gave the software's recommendations to moldmaker GW Plastics Inc. (Bethel, VT) to set up the molding conditions. The initial parts came out of the mold usable. It's the first time we have seen this type of success with such a complex part. The savings in time-to-market proved significant."

Success of this cassette design prompted the engineers to use Part Adviser in the design of the even more complex rotating seal assembly. Plastic components (one fixed part and a component that moves at 3,000 rpm) form the heart of the centrifuge equipment. Flatness and tolerance had to be exact to ensure that liquids wouldn't leak. Molded-in stresses had to be minimized to ensure the parts would have a robust snap-fit.

For the top and lower halves of the rotating seal assembly that come in contact with blood, Fennelly again called on Makrolon PC. The material, he found, gave the components the needed dimensional stability and tensile strength, yet kept the two flat-face seals in place during the 3,000 rpm speeds achieved during centrifuge operation.

The other three components that shield the seal, since they don't come in contact with blood, were designed using a less expensive material. This time Fennelly turned to K-Resin KRO3, a styrene butadiene copolymer supplied by Phillips Chemical (Bartlesville, OK).

The bags used to hold the blood and chemical reagents presented a different problem when it came to the selection of a material. The material had to withstand autoclaving and heat sealing, resist extreme pH chemicals, and be plasticizer-free to avoid introducing any contaminants to the reagents. A three-layer polyester/polypropylene film met the requirements. The material: Cryovac M-312 film from Cryovac Sealed Air Corp. (Duncan, SC). "It's really a superior film," Glen enthuses.

Tying it all together. What made the design process a success, says Jorgensen, was the integration of state-of-the-art software into a "virtual" corporation. "It's the same logic that allows software tools to do the technical part of the job," he adds. "We created a seamless communications network throughout our three facilities, using modern electronic equipment to move data (high-speed modems, the Internet) around internally and externally (among the many suppliers) at a minimal cost. It's nearly a paperless process."

Since the ZymeQuest endeavor was small and brand new, the team was not tied to one system as is often the case in a large corporate setting. That fact enabled the electronic network to be shared with everyone. "We used off-the-shelf software to do the job," Barry explains, "but integrated it so that information flows effortlessly." This includes the firm's extensive design history file, all design changes, inventory management, vendor communications, and even training records. In fact, ZymeQuest often selects vendors because they have the electronic tools to make the process seamless.

From design to market. Currently, ZymeQuest has Phase II clinical trials of the system underway at two leading hospitals to determine whether converted B Cells have the same medical utility as unconverted O cells. So far, studies have shown that the converted cells behave like normal O cells--and they don't trigger an immune reaction.

The first FDA-approved system should hit the marketplace early in the first decade of the new millennium.

How the ZymeQuest blood conversion system operates

A unit of packed red blood cells (A, B, or AB) is sterile-connected to the ZymeQuest disposable set using a Terumo SCD^TM sterile connection device produced by Terumo Medical Inc. (Somerset, NJ). Sterile reagents supplied by ZymeQuest also are plugged into the disposable set.

Technicians activate the appropriate pair of valves, located in the disposable cassette, to allow the cells to drain into the centrifugal processing chamber

A typical process cycle includes the addition of a liquid reagent, mixing, centrifugal sedimentation of the cellular material, and the discharge of the non-cellular liquid to waste. This last step involves pumping a specially formulated hydraulic fluid into the lower of two cavities located within the centrifugal processing chamber and which are separated by a flexible membrane.

The cells are contained in a bag centrally positioned in the upper cavity. The expanding membrane discharges the sedimented fluid from the bag in the second cavity. This fluid stream is directed by the valving set within the cassette to a waste container. The washed cell mass remains in the processing bag ready for the next wash cycle.

Several washing cycles and reagents are required to convert a unit of red blood cells. A specified amount of each reagent is metered into the cells as a function of the starting mass. This sequential washing continuously reduces the pH in preparation for the enzymatic conversion, restoration of the pH back to physiologic levels, and the complete removal of the washing reagent to prepare the cells for long-term storage and subsequent transfusion.

In the middle of these cycles, technicians add a genetically engineered enzyme to the cells. The incubation temperature of the cells is directly measured with an infrared temperature sensor. Temperature adjustments are made by controlling the temperature and the circulation rate of the hydraulic fluid flowing through the lower cavity of the centrifugal chamber, which remains in thermal contact with the cells. After the last cycle, technicians move the cells to a collection bag. They meter a measured amount of storage solution to preserve the cells throughout the cells' shelf life.

Software streamlines design

To design the complex shapes envisioned for the Universal Blood Machine, ZymeQuest engineers used several software packages.

For 3D solid modeling, the design team used Parametric Technology Corp.'s (Waltham, MA) Pro/ENGINEER. To test the concept, the team e-mailed the files to one of its machining partners. The machine shop used CNC Software's (Tolland, CT) Mastercam^TM program to directly convert the digital model into machining instructions for one of its high-end milling centers. Directly, prototype parts were in the design team's hands. Following a testing period, engineers began design refinements to meet the performance and cost criteria.

Some components, like the long, thin, multi-port connector that snaps into the cassette, required sophisticated stress analysis to insure that the deflection of the plastic fell within the tolerance absorbed by the soft O-rings required to form a leakproof seal. Parametric's integrated Pro/MECHANICA^TM stress analysis package accomplished this task. Engineers incorporated modifications on every component into the next generation of detailed design efforts.

To find the best gate locations for injection molding, engineers used Part Adviser from Moldflow Corp. (Lexington, MA). "We had a seamless connection between the Pro/ENGINEER solid model and Part Adviser," says Plastics Engineer Jeremy Fennelly. "This link enabled us to transfer data back and forth to the molder on various models and analyses."

Not only did the Moldflow software identify some ultra-thin sections engineers did not see with the solid model, but it enabled them to test three or four gate locations using several resins to come up with the optimum gate locations and materials for each of the five parts, Fennelly notes. "The entire process took just five weeks. And the parts fit perfectly," he says.

Banking on blood conversion

Blood donors ready to give the gift of life must sit down with a pen and take a test. The questionnaire--along with biochemical tests of the blood--is part of a rigorous screen process that has made the U.S. blood supply one of the safest in the world.

Once that blood reaches the hospital, however, even the most careful initial screen efforts can't protect a patient from getting a fatal transfusion of the wrong blood type. Also, as much as 6% of the donor blood goes to waste because hospitals can't use those units within their 42-day shelf life. The solution: use type O blood!

How do hospitals get enough of the life-saving type O blood? That's the question Jack Goldstein of the Kimball Research Institute at the New York Blood Center asked. Conversion of all blood to type O would address supply and shipping imbalances, he envisioned.

With this in mind, Goldstein and other researchers set out to use an enzyme to alter the chemistry of the blood's red cell surface. Chains of sugars, which cover the cell surfaces of the four human blood types--A, B, AB, and O--all have the same basic sequence, with fucose at the end and galactose next in line.

A and B cells cannot be transfused into people with O blood. The extra sugar branch stimulates the immune system's antibodies to attack the foreign cells. Clipping off that added sugar branch from A and B cells transforms them into type O, averting the immune response.

The large number of sugar chains and their different orientations on cell surfaces made finding the right conditions for conversion a real challenge for Goldstein. Type B cells have over half a million sugar chains on their surface, while type A cells have twice that amount. Some are perpendicular to the surface; others lie parallel.

"This is what fascinated me," Goldstein reports, "to get enzymes and conditions where the enzymes would work." For instance, the enzyme for B conversion works best at a high acidity, but blood cells do their job of carrying oxygen in neutral conditions. Goldstein had to strike a balance that allowed the enzyme to clip off the extra galactose, without destroying the red cell's function. Eventually, he determined that the reaction could take place at 26C, rather than the higher temperature the blood cells are accustomed to, and at a slightly acidic pH of 5.5 or 5.6.

"When I started this work, no one had really treated red cells at such a low pH," Goldstein continues. "It was thought that they would just become nonviable." That turned out not to be the case. In Goldstein's process, not every sugar chain on every cell gets changed, he notes, but as long as enough are clipped, the body accepts the cells. It was at this juncture that ZymeQuest entered the scene to make Goldstein's system a reality.

Partners in design

Mechanical

Lab Engineering and Manufacturing, Inc.
Precision machining and assembly
Billerica, MA

Resh, Inc.
Precision Sheet metal
Franklin, MA

Ams Machine, Inc.
Precision Machining and assembly
Middleboro, MA

Electrical

Design Resources, Inc.
PCB layout and fab
Westboro, MA

Bodine Electric Co.
Motors
Chicago, IL

Carroll Touch, Inc.
Touch screen panels
Round Rock, TX

Technology 80, Inc.
Imbedded Logic Motor Controllers
Minneapolis, MN

Plastics

GW Plastics
Injection Molding, Assembly
Bethel, VT

AIM, Inc.
Injection Molding
Clinton, MI

Alps South Corp.
Silicone molding
St. Petersburg, FL

Apple Pattern, Inc.
Structural molded parts
Sterling, MA

CryoVac Sealed Air Corp
Specialty plastic film
Duncan, SC

Industrial Design and Artwork

Carroll Design Inc.
Lowell, MA
Jack Carroll and Lars Fischer

What this means to you

• Enzyme conversion system will transform A, B, and AB type blood into type O blood, the universal blood type that can be used by anyone needing a blood transfusion.

• Design and molding software played key roles in the system's design.

• Use of various plastic components made the system workable and affordable.

• An electronics network built around off-the-shelf devices enabled designs and communications to flow seamlessly between all parties involved in the design process.

Timeline for design

• Dr. Goldstein and staff conduct Phase I clinical trials of B-to-O conversion using existing laboratory cell washers

• Design & Fabricate 3 add-on controllers, called, "Zeke 1," for Phase II trials

• Design"next generation" prototype

• Strategic change: Redesign and rename to "Zeke 2"

• Design modifications: performance enhancements "Zeke 2.1"

• Design and prototype a multi-unit scale-up system for high thru-put; rename to "Zeke 3"