Portal de Periódicos da CAPES

Remembrances of factor VIII. Part 1: The race to the gene

2004; Elsevier BV; Volume: 2; Issue: 3 Linguagem: Inglês

10.1111/j.1538-7933.2004.00614.x

1538-7933

Jane Gitschier,

Platelet Disorders and Treatments

I have often been credited with the cloning of factor (F)VIII, but the credit does not belong to me. If any one person should be attributed with this discovery, it is my friend and colleague Bill Wood, who took the first peek at this massive gene. But even that statement misses the mark. This project succeeded because of teamwork. The research I will describe took place in the biotech crucible, Genentech, 20 years ago. Genentech in those days could still be considered a start‐up company. I was the 400‐somethingth employee. We were all were housed in one mud‐colored, windowless warehouse at the bayside of an industrial park in South San Francisco. The company seemed to operate like a boy's locker room, and the place reeked of testosterone. No prank was too outrageous, no poker bet was too high, and no woman was part of the inner circle. The head of organic chemistry came out of the building one day to find his car had been painted pink. When a coworker of mine was asked where Dave Goeddel, the most famous, productive and senior scientist at the time, could be found, she retorted, ‘Just go upstairs and look for the tallest 14‐year‐old’. You get the idea. But the place was magical. I found myself in an unbelievably lucky and exciting situation and I loved every minute of it. Bill Wood and I joined Genentech within a few months of each other in the winter of 1981/1982. We were members of Dick Lawn's lab, I as a postdoc and Bill as a scientist. Bill immediately became my scientific mentor, as I had no background in nucleic acid biochemistry. We spent many hours working together, he often giving me the seed of an idea on how to proceed and I offering up results. Bill and I both loved looking at data, a trait I enjoy to this day, and we developed a special and productive partnership. Bill had come to Dick's lab specifically to work on FVIII. FVIII was considered not just the pharmaceutical prize of its time, but the cloning challenge of the moment. The protein was known to be enormous and labile. Combined with its exceptionally low abundance in plasma, it seemed an unlikely candidate for cloning success. However, if potential monetary gain and intellectual challenge were not sufficient motivators, for all of us in Dick's lab the plight of hemophiliacs was the crucial driving force. Readers of this journal are only too familiar with problems associated with past remedies for hemophilia—whole plasma, then frozen concentrate, or even porcine FVIII. The morbidity, the fear and the ineffectiveness in patients with severe hemophilia who also produced ‘inhibitors’ (antibodies against FVIII) made treatment a constant battle. Worse yet, in the spring of 1982, a new disease, a lymphadenopathy later called AIDS, started to afflict hemophiliacs because of the contaminated blood supply. We knew we needed to come up with a recombinant, virus‐free version of this protein. We were on a mission. But how to find the gene for this elusive protein? At the time, there was almost nothing to go on. One, we needed a source of FVIII protein from which some amino acid sequence could be determined. From this sequence, working backwards, we could devise oligonucleotides composed of codons for the amino acid and hybridize them, by Watson–Crick base‐pairing, to a FVIII‐encoding nucleic acid sequence. Two, we needed a source for that nucleic acid sequence. The first problem was solved by a successful collaboration. Last year in these pages, Ted Tuddenham described his work on purification of FVIII, starting in the 1970s [1Tuddenham E.G.D. In search of the eighth factor: a personal reminiscence.J Thromb Haemost. 2003; 1: 403-9Abstract Full Text Full Text PDF PubMed Scopus (7) Google Scholar]. By the early 1980s it was clear that FVIII was a protein distinct from von Willebrand factor and alone provided the crucial activity for treatment of patients with classic hemophilia. Genentech partnered with Speywood Laboratories, who had funded Ted's purification efforts. Ted provided Genentech with antibodies against human FVIII, both from patients who produced ‘inhibitors’ as well as monoclonal antibodies made by Frances Rotblat in his laboratory at the Royal Free Hospital in London. He also supplied preparations of human FVIII purified from the blood of thousands of donors. The second problem, finding a source for the FVIII nucleic acid, was more vexing. Until then, genes were typically cloned as ‘cDNAs’ (complementary DNAs), sequences synthetically produced in the laboratory to recapitulate coding mRNAs. FVIII circulated at such low levels that we had to assume its mRNA was also rare. It appeared to be a very large protein, indicating that its corresponding mRNA would also be large and therefore vulnerable to degradation. Adding to the challenge was the question of its site of synthesis. The liver seemed a very good guess, as most coagulation proteins were known to be produced there. But could we count on it? Dick proffered a different approach. He reasoned that a ‘library’, or collection, of cDNA from liver may or may not have a FVIII cDNA in it, or for that matter, the particular bit of cDNA for which we were probing. On the other hand, he suggested, a library made from human genomic DNA (the whole human genome chopped up and cloned into a phage vector) should have ‘all’ the human sequences in it. This is not completely true, of course; on statistical grounds alone, sequences of the human genome will be missing, as will ‘unclonable’ sequences. Months before we entered the collaboration with Ted and his group, Bill and I and other members of Dick's lab had begun developing methods to clone the gene directly from genomic DNA. We used factor (F)IX, another gene located on the human X‐chromosome, as our guinea pig. Until then, the most commonly used genomic DNA library, one in fact made by Dick when he had been a postdoc, came from male DNA. Consequently, DNA sequences from the X‐chromosome, compared with those from the other chromosomes, were underrepresented by one‐half. Bill found a cell line from a rare individual with four X‐chromosomes and made a new genomic library to enrich for sequences on the X‐chromosome. He also made a series of blots of DNA from a male (having one X‐chromosome) and the 4X individual for testing. If our oligonucleotides were to hybridize at four times the intensity to the 4X DNA compared with the 1X DNA, it would probably indicate an X‐linked gene—hopefully the FVIII gene. Our team also experimented with various kinds of oligonucleotides. An effective cDNA cloning strategy, using pools of relatively short oligonucleotides to cover all possible codon choices, had to be re‐evaluated in the face of the massive complexity of the human genome, with 3 billion base pairs of DNA. If, for example, we had used 16 14‐mers to cover all possible codon choices corresponding to a stretch of five amino acids, we might expect hundreds of matches just by chance. Exacerbating the problem was that oligonucleotides rich in cytosines and guanines (C and G) hybridize much more tightly to their cognate sequences than do those rich in adenines and thymines (A and T). This meant that if an experiment were conducted at a temperature high enough to ensure specificity for binding of a CG‐rich probe to its target sequence, the AT‐rich probe in the pool of sequences did not stand a chance. Bill's attitude was ‘this will never do’, and with me in tow, he marched off to the office of Larry Lasky, a new Genetech recruit and a friend of Dick's. How Bill knew Larry would have a good idea on how to solve this problem I will never know, but within seconds Larry uttered a single word: ‘TEACL’. Larry was talking about tetraethylammonium chloride, a compound that would make all A–T base pairs melt at the same temperature as G–C base pairs. For technical reasons we settled on a related salt, tetramethylammonium chloride, and did a few pilot experiments to be sure we got all oligonucleotides to melt at the same temperature [2Wood W.I. Gitschier J. Lasky L.A. Lawn R.M. Base composition‐independent hybridization in tetramethylammonium chloride: a method for oligonucleotide screening of highly complex gene libraries.Proc Natl Acad Sci USA. 1984; 82: 1585-8Crossref Scopus (588) Google Scholar]. Ever since then I have regarded Larry as having a mind like a steel trap. Another strategy involved use of a single ‘long probe’ synthesized to match one set of codon choices for a longer stretch of amino acids. This approach had been used successfully by an adjacent lab and we planned to give it a try as well. Sometime in early 1983, protein arrived at Genentech and the work toward cloning began in earnest. Gordon Vehar was the head of the protein chemistry effort and was an experienced ‘clotter’, having worked on purification of bovine FVIII in Earl Davie's lab. Gordon was joined by postdoc Bruce Keyt and later Dan Eaton, and together they analyzed the purified material. They measured FVIII activity, correlated it with proteolytic cleavage patterns in response to thrombin digestion, and assessed fragmented products by tryptic peptide mapping. What emerged was a complicated pattern of peptides that together were associated with activity. Some of the fragments, namely 50‐ and 43‐kDa fragments, were clearly related to a 90‐kDa fragment, which in turn was a derivative of the largest fragment of 210 kDa. Another species, the 73‐kDa fragment, was a degradation product of an 80‐kDa fragment, and these were distinct in tryptic‐digestion pattern from all the other species. Somehow all of these peptides had to be related to the single‐chain, 300‐kDa species previously described by Ted's group [3Rotblat F. Goodall A.H. O'Brien D.P. Rawlings E. Middleton S. Tuddenham E.G. Monoclonal antibodies to human procoagulant factor VIII.J Lab Clin Med. 1983; 101: 736-46PubMed Google Scholar]. In March of 1983, somewhere between page 29 and page 30 in my lab notebook no. 6, I jumped ship from pilot studies on FIX to the full‐court press on FVIII. The first entry in my black‐and‐white speckled record of FVIII project team meetings is dated 3/3/83. Within just 4 weeks of the first entry, while I was mucking about with some of the FVIII antibodies, the protein chemistry team had determined the amino‐terminal sequence of several tryptic peptides from the 80‐kDa fragment. Based on one peptide sequence, AWAYFSDVDLEK, Dick ordered synthesis of a single ‘long’ oligonucleotide of 36 nucleotides, and Bill tried it on his 1X/4X panels. Encouraged by the 4‐fold intensity difference, Bill screened the library he had prepared, and by mid May he had isolated the first of what proved to be 26 exons encoding human FVIII. He presciently named this exon ‘A’, not knowing where A fitted into the overall sequence of the gene. All we knew was that part of the 80‐kDa fragment had been cloned. Remarkably, only 10 of the 12 amino acids from this peptide were encoded by exon A. Suddenly what seemed almost impossible was a reality, and more people leapt into the fray. Dan Capon, one of Dave Goeddel's Wunderkindern, took over the cDNA cloning aspect of the project. Dan was one of the most intense, gifted and fiercely competitive researchers you might come across. With his thick black hair and piercing gray eyes, his forward‐bent body seemed to suggest a constant state of attack. Dan and I had been graduate students together at MIT, and I knew he was unstoppable. He and Karen Fisher (then Wion) screened 80 human cell lines to see if they could come up with a decent FVIII producer. They found it in an unlikely source, a T‐cell hybridoma line called AL‐7. Simultaneously, Bill invented a procedure whereby large genomic stretches of FVIII could be made to produce a spliced messenger RNA. Bill never published this idea, but with it he generated sufficient cDNA sequence emanating from exon A to help Dan pull out longer FVIII cDNA clones. Years later an identical technique dubbed ‘exon‐trapping’ was described and developed into a cloning kit, but Bill really got there first. Screening with a set of oligonucleotide probes corresponding to a different peptide sequence subsequently turned up the second exon, ‘B’. When this genomic clone was subdivided and sent off for DNA sequencing, we discovered we had hit a genomic jackpot. The extraordinarily long open reading frame continued for 3100 nucleotides, which ultimately proved to be a third of the entire mRNA. This gave the cDNA cloning efforts a quick shot in the 5′ direction. As there was so much work to be done on so many fronts, I became responsible for the genomic cloning effort. This project turned out to be a tour de force for which I was well suited; it drew on my strengths in logic and maths as well as my organizational abilities. Luckily, I was teamed up with a spunky young technician named Terri Goralka, and over the course of the next 9 months, Terri and I were joined at the hip. We screened genomic libraries, made restriction maps with a set of 10 enzymes, subcloned all BamHI and EcoRI fragments for more refined mapping, identified single‐copy probes for genomic ‘walking’ in both directions, and identified and sequenced exons. We made grids of all the clones so they all could be rapidly tested whenever a new peptide sequence became available. Added to this interplay were the data that poured in from the other members of Dick's lab, including Karen, Dale Hansen, Philip Hollingshead, and Dick himself, a soft‐spoken guy who cultivated our pleasant and highly cooperative lab atmosphere. I developed a record‐keeping strategy and each of us adopted a unique tape color for our Eppendorf tubes. I continue to prize my aliquot of Bill's first FVIII genomic clone, now dried up on the side of a bright‐green labeled tube. One little hiccup in this well‐oiled system was the DNA sequencing lab, run by Peter Seeberg, a lanky German of growth hormone fame, but of later notoriety for claiming to have misappropriated those sequences from the University of California after his postdoc there. Peter hated working with women, a well‐known fact that he readily acknowledged. Terri, Karen and I felt we had to walk on glass whenever we ventured into the sequencing lab, trying to evade Peter. Often, I just sequenced fragments myself to avoid going down there. On reflection, it is hard to appreciate how painstaking this work was. In those days, there were no DNA sequencing machines, no robotics for minipreping DNA and no polymerase chain reaction! We actually had to READ sequencing gels, enter the sequence manually and proofread it. A typical morning consisted of 96 minipreps, and an afternoon of reading sequencing gels and entering the data into the computer by hand. Between experiments, Terri goaded me into tagging along with her on windy 3‐mile runs through the industrial park and to Jazzercize classes at the local athletic club. I was temporarily not just a DNA jock, but an actual JOCK with my own smelly towel in the Genentech ladies' room to prove it. The iterative cloning and mapping process we developed led to a complete map of the human FVIII gene. At 186 bp, it was the largest gene identified at the time. The map that began on 23 May 1983 on a single piece of blue‐grid paper grew to 110 cm and still hangs in my office, a reminder of one of the happiest and most productive periods of my life. By December of 1983, based largely on Dan's cDNA work, we knew the entire sequence of the protein and how the mysterious fragments of 43, 50, 73, 80, 90 and 210 kDa were related. We saw how exon B encoded the entire B domain, a region that proved completely dispensable for activity. We scratched our heads when we realized FVIII is a distant relative of ceruloplasm. Genentech was not the only company hell‐bent on cloning the FVIII gene and putting the factor in a bottle. Our most vigorous competitor was Genetics Institute (GI), a smaller Boston‐based biotech company populated by former MIT and Harvard postdocs. GI teamed up with Mayo Clinic's David Fass, who had purified porcine FVIII and had agreed to supply GI with reagents as Ted had for us. GI's plan was to isolate the porcine gene and pull out the human gene by homology. Competing biotech companies do a special dance with one another. Spirits, and stock prices, can rise and fall with innuendo, hearsay, and on occasion, presentations of actual data. In the summer of 1983, for example, Bill and I attended a Gordon Conference on Hemostasis in New Hampshire. David Fass was a friendly guy and could often be found either bounding off with his fishing pole or heading back with a catch. We were all ears when he reported a porcine peptide sequence with what we recognized as a signature motif of three tyrosines. Fass's finding was identical to the result from an experiment Hermann Oppenheimer at Genentech had performed months earlier on iodinated 90‐kDa fragment, indicating tyrosine residues at positions 5, 6 and 16 both in human and bovine FVIII. We slunk about, scrambling to find a pay phone somewhere on the campus (this was in the days before cell phones) to call the lab. We knew we were way ahead. Little wonder, then, that we were stunned several months later when, on 2 December 1983, we awoke to newspaper reports of GI cloning the FVIII gene. Careful reading of one of these articles convinced us that GI had only a partial human clone, and we breathed a cautious sigh of relief. Bob Kamen, the president of GI, was later quoted by the Wall Street Journal as saying, ‘We made the announcement because we had a large number of people who had pulled off a minor miracle’. We knew just how they felt. We later realized that GI had also filed a patent application based on partial sequence. When GI made its announcement, we were on the verge of the complete sequence. But as excited as we were to solve the FVIII puzzle, we knew we still had a long road ahead. The gold standard was ‘expressing’ the cloned gene and putting the ‘recombinant’ FVIII in a bottle. Again, this was no easy feat, since FVIII, at 2351 amino acids, was far larger than anything that had been produced by recombinant DNA up to that point. FVIII cDNA had to be stitched together in one piece and then put into an expression vector that would promote production of FVIII protein in a mammalian cell. Dan and Bill teamed up with Chris Simonsen and his group, who specialized in expression, and held their breath while the protein chemists monitored the baby hamster kidney cell line's progress. On 10 April 1984, Gordon came by Dick's lab and left a note on Bill's desk: ‘See me’. They had activity! For scientists, the next move would naturally be to publish, but in biotechland, the next move is to secure a patent position. A seminar given by Dick and Gordon was hastily organized at the University of California, San Francisco. The seminar was timed to occur after the patent was submitted on 20 April 1984. By the time Dick and Gordon spoke late that afternoon, Bill had hand‐delivered the application to the US Patent Office in Washington. This seminar served as the public announcement, which was then reported on the front page of the New York Times and that evening on the major network news programs. Everyone was in a tizzy to get the data submitted to Nature, and there were the inevitable discussions about authorship order. I remember one that had to be adjudicated by Dave Goeddel in his office. It was resolved that I would be the first author on the genomic cloning paper and Bill first on the cDNA cloning paper. This understandably rankled Dan, who had worked so hard on the project. Worse than that though, the Genentech legal department, which had to sign off on manuscript submissions, was loathe to let us publish such valuable findings. We started losing our precious lead time to press. Miffed that publication took so long (it eventually occurred in November of 1984 [4Gitschier J. Wood W.I. Goralka T.M. Wion K.L. Chen E.Y. Eaton D.H. Vehar G.A. Capon D.J. Lawn R.M. Characterization of the human factor VIII gene.Nature. 1984; 312: 326-30Crossref PubMed Scopus (723) Google Scholar, 5Wood W.I. Capon D.J. Simonsen C.C. Eaton D.L. Gitschier J. Keyt B. Seeburg P.H. Smith D.H. Hollinshead P. Wion K.L. Delwort H. Teddenham E.G.D. Vehar G.A. Lawn R.M. Expression of active human factor VIII from recombinant DNA clones.Nature. 1984; 312: 330-7Crossref PubMed Scopus (520) Google Scholar, 6Vehar G.A. Keyt B. Eaton D. Rodriguez H. O'Brien D.P. Rotblat F. Oppermann H. Keck R. Wood W.I. Harkins R.N. Tuddenham E.G.D. Lawn R.M. Capon D.J. Structure of human factor VIII.Nature. 1984; 312: 337-42Crossref PubMed Scopus (655) Google Scholar]), we were nonetheless pleased that the well‐deserving GI team had the opportunity to publish their findings in the same issue [7Toole J.J. Knopf J.L. Wosney J.M. Sultzman L.A. Buecker J.L. Pittman D.D. Kaufman R.J. Brown E. Shoemaker C. Orr E.C. Amphlett G.W. Foster W.B. Coe M.L. Knutson G.J. Fass D.N. Hewick R.M. Molecular cloning of a cDNA encoding human antihaemophilic factor.Nature. 1984; 312: 342-7Crossref PubMed Scopus (657) Google Scholar]. In fact, the GI team had used some of the same strategies we had, including screening with a long probe, cloning directly from genomic DNA, and testing their sequence for X‐linkage with the 4X cell line. Even though we did not know our competitors on the Atlantic, we knew what they had been through and we felt a kindred spirit. Serendipity led Nature to publish the paper entitled ‘Characterization of the human factor VIII gene’, on which I was lead author, first in its series of four articles on FVIII. We had assumed that honor would go to the paper by Wood et al. on the cDNA cloning and production of active human FVIII. It was a lucky break for me, but paled in comparison with the joy of discovery, the camaraderie, the intellectual excitement, and the good feeling of having accomplished something worthwhile. Both Genentech and GI realized their goals for treating hemophiliacs. GI's product made it first to the market, with Food and Drug Administration (FDA) approval in late 1992 under the name Recombinate. Miles/Bayer licensed Genentech's technology and currently sells it as Kogenate, approved by the FDA in 1993. Since GI filed its patent application first, a patent interference case, the longest in US history, ensued. The case was ultimately decided in favor of Genentech, who had the stronger claim based on full‐length sequence and expression of active protein, and Genentech agreed to cross‐license with GI. GI is now a subdivision of Wyeth Laboratories and produces a B‐domain‐less version of FVIII under the name ‘ReFacto’, approved in 2000. The companies have had trouble meeting the demands for recombinant FVIII in recent years.