Chapter 02: Building the Continuous-Flow Blood Separator


Chapter 02: Building the Continuous-Flow Blood Separator



Media is loading


In this chapter, Dr. Freireich describes treating infections by transfusing white blood cells. He goes over the technical difficulties of separating white cells from platelets. (He describes his lab at NCI, festooned with 50 feet of tubing.) Here he also specifies why a continuous flow of blood was needed: using the analogy of an artificial kidney, he explains that leukemia patients required a huge number of either platelets or white blood cells, so the aim was to process a donor's entire blood supply, while mobilizing the donor's body to replace the elements removed for the transfusion. The next phase of the blood separator story begins when an IBM engineer, Al Judson, appears and asks if there's something he might do to help cure leukemia. (His son was afflicted by the disease.) Dr. Freireich describes the materials and technical challenges of creating an instrument that would channel blood from a donor into a centrifuge, collect the proper layer of separated elements for the patient, and deliver plasma back to the donor.

Publication Date



The Historical Resources Center, The Research Medical Library, The University of Texas MD Anderson Cancer Center


Houston, Texas

Topics Covered

The Interview Subject's Story - The Researcher; Overview; Definitions, Explanations, Translations; On Research and Researchers; Understanding Cancer, the History of Science, Cancer Research; The History of Health Care, Patient Care; The Researcher; Discovery, Creativity and Innovation; Professional Practice; The Professional at Work; Collaborations; Devices, Drugs, Procedures; Portraits; Patients; Discovery and Success


History of Science, Technology, and Medicine | Oncology | Oral History


Emil J Freireich, MD:

That success occurred. We said, well, now they’re all dying of infection. So we did the same thing; we did a retrospective study—Dr. [Gerald] Bodey [Oral History Interview], who is now retired—still here—number of white cells compared to the occurrence of serious infection. He showed another citation classic that if you lower the white count, the longer it’s low the more likely you have a major infection. Obviously, we’ve got to have white cells.

Well, the problem was that unlike the platelets, which have a half-life and a circulation of about five days, the volume of distribution of platelets—you understand, the circulating blood has red cells and all the other cells suspended in plasma, but the concentration is not uniform throughout the body. It’s higher in the spleen and in peripheral, small capillaries, so what we have to measure is kind of an average, and on average the volume of distribution of the red cells defines the blood volume because the red cells don’t go outside the blood vessels unless you’re bleeding. I’ll come back to that too.

So when you inject the platelet, the volume of distributions are about twice that of the red cells because the concentration of platelets in organs like the spleen is higher than that of the red cells. They get through, and the platelets, which are small, tend to peripherally circulate and they drag, so their volume of distribution is about twice, so when you inject X amount of red cells, you get Y amount of increase. When you inject X amount of platelets, you get Y/2—you get half the increment.

So we radio labeled granulocytes from the normal donors, and the first thing we learned was that the volume of distribution was twenty times normal, and the reason for that is that the granulocytes operate like the fire department. That is, the normal situation for the differentiate granulocytes is in the marrow granulocyte reservoir. They don’t circulate unless they’re called upon. There has to be a fire alarm. The red cells all circulate continuously—the platelets—but the white cells don’t. So when you inject white cells, the more granulocytopenic the recipient, the larger the volume of distribution.

Problem number two is if you label the granulocytes, once you inject the granulocytes labeled into a recipient, the half life and the circulation is six hours, because once the call goes out for granulocytes from the fire house, once they get into circulation, it’s because they’re needed. The message came from either a site of inflammation—a break in the vessel—so they’re consumed peripherally, very rapidly—half life six hours.

So if you do blood counts on a normal person—you probably know all this. If I’m getting repetitive, you can stop me. In a normal person, when you do a blood count, he has 5000 to 6000 per mL circulating granulocytes. Why is that? It’s because you’ve got infection all over the place. In a germ-free animal, his white count is ten. They’re all in the fire house. When you inject them, they immediately go along the gum margins, around the rectum, all the sites of infection—the nose, the lungs—so they exit immediately. It’s not that their life span is short; it’s that their physiology is short. They want to go where the problem is.

So we tried our bags, and we got as many as we could and you shoot them in and nothing happens. The problem was obvious. It was dose. So my friend Alan Kliman and I—we had a cookout in his backyard, and we started to talk about how we could get more granulocytes. We came up with the idea I take credit for and he takes credit for. We got the idea of turning to patients who have chronic granulocytic leukemia and have white counts 300 times normal. They’re diseased, of course, but the in vitro studies of granulocytes from granulocytic leukemia patients indicate that they’re damaged but they’re only about half as good. So we said, let’s get white cells from leukemia patients and give them to our dying leukemia children to control infection.

Great idea. Can you imagine doing that in 2011 in the United States? It’s insane. You’re giving leukemia cells to dying children with leukemia. They’re going to die of leukemia. They’re going to get—but only in the clinical center, only with Frei and Zubrod would that ever have occurred, and we did it. We got patients with CML-benign phase that can live with these high counts for very long periods of time, and we asked them to volunteer as donors, and we bled them continuously as donors. Of course, they had so many that as soon as we took them out, they put more in there. We actually published a paper showing that doing leukapheresis on these CML patients for a period of a month did not affect their survival or didn’t make them sick in any way, so it was safe.

We didn’t do that in beginning. We had to prove that, and when we gave these to the children, we worked out the whole physiology of granulocyte transfusion. We found the volume and distribution twenty times normal. We found the half life in the circulation, which ended up being about twenty-four hours instead of six, because once you fill the marrow granulocyte reservoir, they don’t immediately go back to the marrow, they stay in the blood and do what they’re doing. We showed the relationship between the pre-count, and we had a dose response. The higher the level of granulocytes in the recipient, the more likely they were to get cured. When we gave 1011 granulocytes to these children, ninety percent of them had their bloodstream cured of infection and they were cured.

Tacey Ann Rosolowski, PhD:

They were cured?

Emil J Freireich, MD:

Temporarily. So we had the solution. The problem is there are lots of people with no white cells and there are very few patients with CML and most of them aren’t going to volunteer to do nothing for a year. So we had to figure out how to get 1010—ten billion granulocytes from a donor. Well, obviously it’s impossible because the number—that’s the number in your whole body. If we got every white cell out of your peripheral blood we’d have enough for one transfusion.

But being stupid and young, digging ditches, I tried. So I went to work, read all the literature and studied flow dynamics, rheology. We studied the physiology of blood in the small vessels. We found that like the platelets, the granulocytes tend to roll along the margins so that the axial flow is red cells. We tried pushing blood through capillaries where the ratio of axial flow to peripheral flow was maximum, which is a very small tube. Then we tried to make it long enough so we could get, at the end, a very high concentration of white cells. This Rube Goldberg thing that I was working on required enormous positive pressure to push it through these capillaries, and it required a very long path for the blood, so if you came in my lab you’d see this tubing.

Tacey Ann Rosolowski, PhD:

How long was the tubing that was required?

Emil J Freireich, MD:

Oh, in order to get a reasonable separation it was maybe fifty feet—fifty feet of tubing through my lab. You had to have a huge machine pushing it very, very slowly.

Tacey Ann Rosolowski, PhD:

So this was really the next evolution of the blood separator?

Emil J Freireich, MD:

Well, I’m trying to figure out a way to separate them continuously, but it was obvious that the only way it would work would be centrifusion, and the problem with centrifusion is that the specific gravity of the white cell is only slightly less than the red cell. The way you separate them is you have to get the red cells to rouleaux. The rouleaux in the presence of macromolecules. They stack up like coins. That way the particles are larger than the white cells, and when you put them in the centrifuge, the sedimenting particles push the plasma out with the granulocytes and they end up in the buffy coat.

So we did all these experiments in tubes. We got blood from the blood bank that was contaminated with syphilis and stuff and had to be discarded. We did all that. If you came into my lab, it was like the hospital ward—blood all over the place.

Tacey Ann Rosolowski, PhD:

Why were you trying to work for the continuous flow as opposed to the centrifuge? What was the advantage?

Emil J Freireich, MD:

Well, the idea was that if I’m going to get enough white cells for a transfusion, I have to get the blood cells from your entire blood volume. Now, the one thing we learned was that when we started leukapheresing with the bags, no matter how many white cells we removed, the white cells in the marrow granulocyte reservoir replaced them very quickly. We knew that there was a large reservoir of granulocytes, but not in the blood. So obviously we had to process the blood. The obvious image was an artificial kidney, which removes pollutants and leaves the good stuff. So we needed something that would process the blood, remove the white cells, leave all the other formed elements, and mobilize the cells from the bone marrow, so we knew what we had to do.

So I was up there doing this stuff in the lab, and everybody knew crazy Freireich and all his tubes. One day, Mr. Judson appeared in my office. Why did he appear? Well, his son had chronic monocytic leukemia. He worked for IBM; he was an engineer. His engineering experience was with jet engines. He was one of the project officers that made all the—how to get the air and the fluid and the fire. His doctor happened to be Jerome Block, who I think is now dead. He ended up in California doing oncology. Jerry Block was one of our trainees. He said, “Is there anything an engineer can do?” Block said, “This crazy Freireich on the trial floor is trying to do blood things. Go talk to him.” He came to me.

In our first publication, I describe what we actually did. He came in, and this is what I said, “This is what I want to accomplish.” He said, “Okay, tell me what you need.” So I sat down with pencil and paper and wrote the ten things that I needed—continuous flow, closed system, no hemolysis, rapid flow—it’s in this paper. So Judson listened very carefully. He was a very straightforward guy. He took this piece of paper—

Tacey Ann Rosolowski, PhD:

Yep. Here are the ten things. It’s a really great list. Do you want to read it?

Emil J Freireich, MD:

No, you can read it.

Tacey Ann Rosolowski, PhD:

The leukocytes could be separated—

Emil J Freireich, MD:

So Judson wrote them all down. I said, “There’s the job.” So he said, “Okay.” So he went away and I forgot all about it. I mean, crazy, simple-minded guy who’s never done anything biological, only works on jet engines. So I continued with my things and the tubes and the work. In two or three months—I don’t remember the exact time—two or three months later, who appears in my lab? Mr. Judson, and he’s got a briefcase full of mechanical things—metal screws, plastic tubes. He said, “Here it is.” No one could have been more shocked than me. I don’t think I have pictures in that paper, but I have pictures of the original equipment. It’s amazing.

So we put this together with screws and screwdrivers and bolts and glue and O-rings. We built the thing. You see, the secret to a continuous-flow instrument is how you connect the rotating and the stationary parts. That was the biggest challenge we faced. So we put these things together, and we got heart pumps from the heart machines. We got tubing. It was all done with make-up things that he could salvage from IBM’s—when projects failed they threw them in the storehouse and he salvaged them.

Tacey Ann Rosolowski, PhD:

Why were you surprised when he showed up?

Emil J Freireich, MD:

Well, because he didn’t know anything about it, and I had been working my head off, and I’m ten times smarter than he is. Why didn’t I figure it out? But he went to work—you know—he was an engineer.

Tacey Ann Rosolowski, PhD:

It sounds like he’s a trenches guy too.

Emil J Freireich, MD:

And he came with a machine. I thought, whoa! So we tested it, and although it didn’t work, you could see immediately that the concepts were there. So we got the centrifuge, we learned how to put—we had the bolt. The original centrifuge was upside down, so the blood flowed from top to bottom and the centrifugal force was horizontal, so how to connect the rotating and stationary parts was a very difficult problem. How do you do that?

Well, Judson looked in the literature, and he discovered that during the war the atomic energy commission people were separating isotopes centrifugally in ultracentrifuge, and they had worked out what’s called face seal. A face seal is a combination of stainless steel and very high-quality Teflon plastic which has been engineered to be less than one molecule of water thick, so they’re absolutely flat. In that circumstance, the flat faces can rotate continuously if the mechanics are right so they don’t jiggle and there will be no crossing between the channels, so the thing is a circle.

Well, he went to Oak Ridge and he saw the things, and he went back to IBM and they made face seals. We had a beginning on that problem, and then the problem was every instrument that was continuous flow could separate, precipitate, and supernate. That’s not a problem. That’s not a problem. The problem was how to connect from the center of the centrifuge—how to collect the buffy coat.

Well, that was a real challenge because we had to devise a collecting system that would allow you to view the separation—we’ve got some beautiful pictures of that—so that you could locate the collector over the buffy coat, because the plasma comes from the top, the precipitate from the bottom, but the buffy coat is in the middle, so if you don’t get the right ratio between the flow of platelets and red cells, you have to locate the buffy coat over a fixed collecting point.

So he had a plastic bottom. We did it visually and manually, and, by golly, we invented the first continuous-flow machine that could collect supernate, precipitate, and buffy coat.

Now, the collection efficiency was still very, very poor because of this problem of red cells, but I had done a lot of work on centrifuge and white cells, and we had studied all the macromolecules—fibrinogen and so on—and I came across a paper by a guy who is still working, Craig Thompson, who had invented a drug called hydroxyethyl starch. It’s a synthetic macromolecule that is susceptible to glycolysis by human enzymes. He developed this in order to use it as a plasma substitute, so that in battle, instead of having to have plasma, which deteriorates and gets infected, you can use a synthetic macromolecule to replace the blood volume temporarily until they get out of shock and you can get blood and so on. It was very important during the war and during major surgery—during cardiovascular surgery. They had to use plasma substitutes instead of blood.

So I wrote to him, and he sent me some hydroxyethyl starch. It comes in various size—variety. There’s a lot of stuff that has to go through—you don’t want to know all the details. It worked beautifully in the test tube. Then we tried it in vivo, and it worked beautifully in vivo. It was about ninety-eight percent degraded in the recipient so that, although you retained some of it in your organs, it was not carcinogenic, and it didn’t stay there forever because eventually your hydrolytic enzymes would break it down. If you used the right size hydroxyethyl starch, most of it was degraded or excreted by the kidneys, so we had a relatively safe product. We had to go through all that. It worked beautifully.

Chapter 02: Building the Continuous-Flow Blood Separator