CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`
`
`
`Patent
`Attorney's Docket No. 22338-10230
`
`IN THE UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`Group Art Unit:
`
`3991
`
`Examiner:
`
`
`
`B.M. Celsa
`
`(’542)
`(’859)
`
`
`
`
`
`90/007,542
`Control Nos.:
`90/007,859
`
`
`
`Confirmation Nos.: 7585 (’542)
`
`
`
`6447 (’859)
`Filed:
`
`13 May 2005
`
`
`
`23 December 2005
`Patent Owner:
`Genentech, Inc. and
`
`
`
`City of Hope
`Merged Reexaminations of U.S. Patent No. 6,331,415 (Cabilly et al.)
`For:
`
`
`SECOND DECLARATION OF DR. STEVEN L. McKNIGHT
`UNDER 37 C.F.R. § 1.132
`
`
`
`I, Dr. Steven L. McKnight, do hereby declare and state:
`
`1.
`
`2.
`
`3.
`
`I am a citizen of the United States, and reside in Dallas, Texas. I previously provided a
`declaration in this reexamination proceeding. The circumstances of my involvement in this
`case remain the same as I had described them in my earlier declaration.
`
`I analyzed the Final Action and the references discussed in the Final Action.1 My opinions
`about the references and the Final Action are provided in this declaration.
`
`The opinions I provide in this declaration are what I believe would have been the views of
`a person of ordinary skill in the art as of early April 1983. I believe I can accurately
`describe that person’s perspective. In early April 1983, I was actively experimenting in the
`area of recombinant DNA technology, including cloning and expressing recombinant
`eukaryotic genes. By that time, I had worked with both bacterial and mammalian
`expression systems, and had extensive experience using the Xenopus oocyte microinjection
`technique.
`
`
`1 U.S. Patent No. 4,399,216 (the Axel patent); Ochi et al., Nature 302:340-42 (1983) (Ochi); Oi et al.,
`Proc. Nat’l Acad. Sci. (USA) 80:825-29 (1983) (Oi); Rice & Baltimore, Proc. Nat’l Acad. Sci. (USA)
`79:7862-65 (1982) (Rice); Deacon & Ebringer, Biochemical Society Transactions 4:818-20 (1976)
`(Deacon); Valle et al., Nature 291:338-340 (1981) (Valle 1981); U.S. Patent No. 5,840,545 (Moore) ;
`EP004722 (Kaplan); U.S. Patent No. 4,511,502 (Builder); Accolla et al., Proc. Nat’l Acad. Sci. (USA)
`77(1):563-566 (1980) (Accolla); PCT Patent Publication No. WO 82/03088 (Dallas); and claims 1-7
`of U.S. Patent No. 4,816,567 (the ’567 patent).
`Mylan v. Genentech
`IPR2016-00710
`Genentech Exhibit 2041
`
`PAGE 1
`
`
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`The Patented Invention and the State of the Art in April 1983
`
`4.
`
`The ’415 patent requires the production of an immunoglobulin molecule or
`immunologically functional fragment by expression of DNA sequences encoding both
`heavy and light immunoglobulin chain polypeptides in a single transformed host cell. This
`means that all of the following things must happen:
`
` (i) host cells must have been successfully transformed with DNA sequences encoding
`the heavy and the light chain polypeptide sequences;
`
`(ii) the transformed host cell must independently express both sequences (e.g., each DNA
`sequence must be accurately transcribed into an mRNA, and each mRNA must be
`translated into an appropriate amino acid sequence corresponding to each chain); and
`
` (iii) the polypeptides must be assembled into an immunoglobulin tetramer or antigen
`binding fragment either inside or outside of the cell.
`
`5. None of the references cited by the Office (or any other publications of which I was aware
`in April 1983) describe or suggest performing an experiment comparable in complexity to
`what is required by the ’415 patent. In fact, I was not aware of a single paper published by
`April 1983 that even suggested the concept of producing more than one eukaryotic
`polypeptide at a time in a single recombinantly transformed host cell.
`
`6. Where experimental results are reported in these references, the results show a significant
`amount of unpredictability. Experimental results would have been important to a person of
`ordinary skill in the art in April 1983 because many of the biological mechanisms that
`controlled expression of foreign DNA and assembly of proteins were not well understood
`at that time. As Dr. Harris observed in his article, “it is clear that not all the rules
`governing the expression of cloned genes have been elaborated and those rules that do exist
`are still largely empirical.”2
`
`7.
`
`In my opinion, the publications and patents cited in the Final Action would not have led me
`(or any other person of ordinary skill in the art) in April 1983 to believe that what was
`required by the ’415 patent could be predictably achieved. Each of the cited references
`discloses something far less complicated than what the ’415 patent requires, and those that
`report results show significant unpredictability in achieving success in these simpler
`experiments. In addition, none of the references provide any answers to the questions that
`these references would have raised in the mind of a person of ordinary skill in the art in
`April 1983 about making an immunoglobulin molecule or fragment by producing the heavy
`and light chain polypeptides together in one transformed host cell.
`
`8. Considering these scientific observations in aggregate, I believe these references would
`have told a person of ordinary skill in the art in April 1983 to not attempt to produce an
`immunoglobulin molecule by expressing two different DNA sequences encoding the heavy
`and light chains in one transformed host cell. Instead, I believe the references suggested
`
`2 Harris, Genetic Engineering 4: 127-85, at p. 129 (1983).
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 2
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`taking the opposite approach, namely, to produce each chain in a separate cell culture and
`then (if that succeeds) attempt to assemble the immunoglobulin using these individually
`produced chains. Trying to produce the immunoglobulin this way would reduce some of
`the uncertainty by breaking the process down into more manageable steps. Only this
`approach would have been consistent with the prevailing mindset in April 1983 of
`producing only one eukaryotic polypeptide at a time in a transformed host cell.
`
`9.
`
`This mindset is shown in Dr. Harris’ paper, which listed all of the published reports of
`production of eukaryotic proteins using recombinant DNA techniques in bacterial host cells
`as of March 1983. Every example, without exception, reports production of only one
`polypeptide at a time in a transformed host cell.
`
`10. This mindset is also shown by the approach people had taken to produce insulin. Insulin
`was the only multimeric protein that had been produced using recombinant DNA
`techniques before April 1983. Insulin is a relatively simple multimeric3 protein made up of
`two polypeptide chains linked by two inter-chain disulfide bonds (and containing one intra-
`chain disulfide bond). Each of the insulin polypeptides is small (i.e., 21 and 30 amino acid
`residues). An immunoglobulin molecule is a much larger and more complicated protein
`than insulin. It is made up of two heterodimers and has many inter- and intra-chain
`disulfide bonds. Each immunoglobulin chain is also substantially larger than either chain
`of insulin (i.e., light chains have between 210 to 220 residues while heavy chains have
`between 455 to 550 residues).
`
`11. Goeddel et al.,4 for example, reported production of insulin by expressing each insulin
`chain in a separate host cell culture. Then after each chain had been expressed and
`recovered, the two chains were combined in a test tube to form the insulin structure. The
`other approach that had been proposed by April 1983 was to produce a single chain insulin
`precursor polypeptide, isolate that polypeptide from the cell culture, cleave it in a test tube
`to produce the two insulin chains, and then form the insulin multimer in the test tube.5
`
`12. The Moore patent also clearly reflects this one polypeptide-one host cell mindset. This
`patent describes a way of producing a “multimeric” antigen-binding molecule made up of
`short polypeptides corresponding to variable domain sequences found in heavy and light
`immunoglobulin chains.6 What Moore says to do is produce each of the heavy and light
`chain polypeptides in separate host cell cultures, and then combine them in a test tube to
`form the rFv.7
`
`3 A multimeric protein is a protein complex made up of more than one polypeptide subunit. The
`polypeptides form a stable complex through disulfide bonds and/or non-covalent interactions.
`4 Proc. Nat’l Acad. Sci. (USA) 76:106-110 (1979).
`5 See, e.g., Harris, supra note 2, at p. 138; Wetzel et al., Gene 16:63-71 (1981); Brousseau et al., Gene
`17:279-289 (1982).
`6 See, e.g., Moore at col. 2, lns. 22-35. The Moore patent does not include any experimental results
`showing that a functional rFv molecule was actually made. See also May 18, 2007 Declaration of
`Steven McKnight at ¶¶ 48-54.
`7 See, e.g., May 18, 2007 Declaration of Steven McKnight at ¶¶ 8-31.
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 3
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`13. The same approach of producing each chain of the immunoglobulin molecule in a separate
`host cell is what the Kaplan publication also says to do. For example, at page 10, the
`Kaplan publication says to produce each of the heavy and light chains in separate host cells,
`isolate each chain, and then attempt to produce the immunoglobulin molecule by
`combining the chains under mildly oxidizing conditions in a test tube (which they do not
`identify).8
`
`14. The Cabilly ’567 patent also follows this same mindset. The ’567 patent claims require
`production of only one chimeric heavy or light immunoglobulin chain at a time in a host
`cell.
`
`15. The Cabilly specification also identifies certain benefits of producing the different chains in
`separate host cell cultures. For example, it indicates that an “additional area of flexibility
`which arises from the use of recombinant techniques results from the power to produce
`heavy and light chains or fragments thereof in separate cultures . . . and to prevent
`reconstitution of the antibody or immunoglobulin aggregation until the suitable
`components are assembled.”9 It also explains that different types of immunoglobulin
`molecules can be made using separately produced heavy and light chains.10
`
`16. All of these references clearly call for production of only one desired polypeptide at a time
`in a recombinant host cell, even if the ultimate objective might have been to produce a
`multimeric protein. This is the opposite of what the ’415 patent requires (i.e., production
`of two different immunoglobulin polypeptides in one host cell).
`
`Co-Transformation of Host Cells is Not Equivalent to Co-Expression of Two DNA Sequences
`
`17. The Axel patent describes a technique where the goal was transformation and expression of
`foreign DNA sequences in eukaryotic host cells. The experimental results reported in the
`patent show that eukaryotic host cells could be co-transformed with two different DNA
`sequences, but that these co-transformed host cells did not properly transcribe both DNA
`sequences and did not produce the desired protein.
`
`18. The focus of the Axel patent is its technique for transforming a eukaryotic host cell with a
`gene encoding a selectable marker. The patent also shows that cells could be “co-
`transformed” with a second DNA sequence along with the marker gene. The second DNA
`
`8 Kaplan refers to one dsDNA per vector, per host followed by “separately purified” light and heavy
`chains (page 10). Even with the minimal detail in Kaplan, it is clear that the chains should be
`separately purified and then assembled. Kaplan refers to “assembling of the light and heavy chains”
`(page 3), and, at page 10, to “combining” the purified light and heavy chains under “mildly oxidizing
`conditions,” so it is clear to me that separate Ig chain production is all that is described and intended.
`9
`’567 patent at col. 14, ln. 65 to col. 15, ln. 4.
`10 See, e.g., id. at col. 15, lns. 44-57 (hybrid antibodies) (“Pairs of heavy and light chains . . . are
`prepared in four separate cultures, thus preventing premature assembly of the tetramer”); col. 16, lns.
`33-54 (univalent antibodies) (“[T]he desired Fc region [is] expressed . . . . This portion is then bound
`using the technique of D.2 to separately produced heavy chain . . . and separately produced light chain
`[is] added.”).
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 4
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`sequence was shown to “go along for the ride” and become stably integrated into the
`chromosomal DNA of the transformed cell, meaning that both DNA sequences would be
`passed on to the progeny of the “co-transformed” cell.11
`
`19. The Axel patent outlines a strategy of using the co-transformation technique to produce a
`“desired proteinaceous material.” The process described in the Axel patent uses two DNA
`sequences, each encoding one polypeptide.12 One DNA sequence is the gene that encodes
`the selectable marker (DNA II). The other encodes the protein of interest to be produced
`and recovered from the cell (DNA I). This DNA I + DNA II process is designed to
`produce only one polypeptide that is recovered from the cell -- the marker gene protein is
`not recovered from the transformed cells under the Axel patent process.
`
`20. The Axel patent reports that co-transformed cells successfully expressed the selectable
`marker gene (i.e., DNA II) and produced “functional” marker gene protein. As a result of
`expressing the functional marker gene protein, the transformed cells exhibited a changed
`phenotype that made them resistant to a chemical that was toxic to untransformed cells.13
`
`21. The Axel patent does not show production of any “functional” protein encoded by DNA I,
`much less a functional multimeric protein.14 Instead, it reports experimental results
`showing that the two attempts to express a “DNA I” sequence (i.e., a gene encoding a
`desired polypeptide) in a co-transformed cell both failed.15 In both experiments, the Axel
`patent reports that host cells were successfully “co-transformed” with the “DNA I”
`
`
`11 See Axel patent at col. 4, lns. 15-21.
`12 Axel explains that a proteinaceous material is a biopolymer formed from amino acids. See Axel
`patent at col. 4, lns. 28-29. I read this as meaning a single polypeptide (i.e., a sequence of amino acid
`residues linked by peptide bonds) rather than a multimeric protein complex made up of different
`polypeptides associated through non-covalent interactions or disulfide bonds.
`13 See, e.g., Axel patent at col. 2, lns. 16-27 (discussing Wigler et al., Cell 11:223-232 (1977)). This
`passage shows that this section discusses production of a functional “selectable phenotype” protein
`and not production of “functional” proteins encoded by DNA I sequences. I note this because the
`Office is relying on this incorrect assumption about which “functional” protein was made to justify its
`conclusion that Axel describes procedures for producing “functional antibodies.” See Final Action at
`pp. 30-31.
`14 The Axel patent lists interferon as one of the types of proteins that could be made by its procedures.
`In February 1980, when the Axel patent was filed, the only interferon proteins known were
`monomeric proteins – meaning they only had one polypeptide chain. The Final Action (at page 30)
`mistakenly states that interferon is a multimeric protein.
`15 The first experiment used the rabbit ß-globin gene as DNA I. See Axel patent, First, Second, and
`Third Series of Experiments, at col. 9, ln. 59 to col. 25, ln. 68. The second experiment used human ß-
`globin gene as DNA I. See id. at Fifth Series of Experiments, at col. 30, ln. 60 to col. 42, ln. 10. The
`Axel patent also reports results of co-transformation experiments using a marker gene and a second
`model DNA sequence (i.e., the pBR322 or ΦX174 bacteriophage sequence). See id. at col. 16, lns.
`52-54 (“The stable transfer of Φ DNA sequences to mammalian cells serves as a model system for the
`introduction of defined genes for which no selective criteria exist.”). Nothing is reported in the patent
`about expression of these model sequences.
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 5
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`sequence and the marker gene,16 but that the polypeptide encoded by the DNA I sequence
`was not produced in these co-transformed host cells.17
`
`22. The Axel patent also reports that in one of these experiments, abnormal transcription of the
`DNA I sequence was observed (i.e., the rabbit (cid:533)-globin DNA I sequence was not properly
`transcribed in the co-transformed cell).18 The Axel inventors apparently did not investigate
`whether the other DNA I sequence tested (the human (cid:533)-globin gene) was properly
`transcribed or translated, because nothing is reported about this in the patent.
`
`23. The results reported in the Axel patent clearly show that “co-transformation” of cells was
`not being equated with successful “co-expression” of two foreign DNA sequences. Instead,
`these results show that successfully co-transformed cells did not properly transcribe DNA I
`to produce correct mRNA, and did not produce any DNA I polypeptide.
`
`24.
`
`In April 1983, I would not have read the Axel patent as describing – or even suggesting –
`production of three different proteins (i.e., two different proteins of interest plus a marker
`protein) in a single transformed host cell. That does not match the DNA I + DNA II
`process described in the patent, and ignores the experimental results showing unsuccessful
`expression of each DNA I sequence actually tested.
`
`25. The Axel patent certainly does not explain how to make three different proteins in one host
`cell. For example, it does not contain any kind of plan for transforming a host cell with
`three different DNA sequences. It also does not contain any suggestions for improving the
`odds of successful transformation or expression of the three genes (as it does for
`experiments using just two genes).19
`
`26. A person of ordinary skill in the art would not have read the Axel patent as specifically
`suggesting production of two different antibody polypeptides in a single transformed host
`cell. This is because if that person followed the DNA I + DNA II scheme outlined in the
`patent, and used the procedure described for obtaining the DNA I sequence (i.e., restriction
`endonuclease digestion of chromosomal DNA), he or she would have obtained a DNA I
`sequence that encoded only one antibody polypeptide. The restriction endonuclease
`
`
`16 See id. at col. 13, ln. 56 to col. 14, ln. 23; see also id. at col. 15, lns. 12-22; col. 17, lns. 43-46; col. 22,
`lns. 39-41.
`17 See, e.g., id. at col. 21, lns. 37-57 (“Attempts to detect this protein [the rabbit ß-globin encoded by
`DNA I] in cell lysates using a purified anti-rabbit ß-globin antibody have thus far been
`unsuccessful.”).
`18 See id. at col. 20, lns. 13-15 (“Taken together, these results indicate that although the intervening
`sequences expressed in transformed mouse fibroblast are removed from the RNA transcripts precisely,
`the 5' termini of the cytoplasmic transcripts observed do not contain about 48±5 nucleotides present
`in mature 9S RNA of rabbit erythroblasts.”); see also id. at col. 19, ln. 7 to col. 20, ln. 67 (reporting
`the nature of aberrant (i.e., incorrect) transcription of the rabbit ß-globin DNA I sequence).
`19 See id. at col. 5, lns. 29-50; col.6, lns. 47-53; col. 7, lns. 3-26. These techniques of using excess ratios
`of copies of the DNA I sequence to the DNA II sequence or gene amplification of linked DNA I all
`result in cells that will have identical copies of the same DNA I sequence in the cell.
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 6
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`technique digests chromosomal DNA to recover the desired sequence. Because the
`antibody genes are located on different chromosomes, digestion will produce a DNA
`sequence that has only one antibody gene, not both.
`
`27. A person of ordinary skill also would not have read the Axel patent as saying that the
`heavy and light chains should be produced in one transformed host cell if that person
`wanted to use its procedures to try to make an antibody molecule. Instead, that person
`would have understood that the way to approach producing an antibody (or any other
`multimeric protein) using the Axel methodology would have been to produce only one
`antibody polypeptide at a time using a host cell co-transformed with a marker gene and a
`DNA I sequence encoding the desired antibody polypeptide (i.e., produce one co-
`transformed host cell with a DNA I encoding the antibody heavy chain, and a different co-
`transformed host cell with a DNA I encoding the antibody light chain). If each chain were
`successfully produced and isolated from the separate host cell cultures, then the next step
`would have been to try to combine the chains in a test tube to form the immunoglobulin
`tetramer or binding fragment. This is the only approach that is consistent with the DNA I +
`DNA II scheme outlined in the Axel patent, and with the recommendations in the Moore
`and Kaplan references.
`
`28.
`
`I note that the Office has pointed out that the Axel patent identifies potential benefits of
`producing eukaryotic proteins in eukaryotic cells. These benefits include that the proteins
`may be glycosylated or subjected to other types of post-translational chemical modification.
`These properties of eukaryotic cells were known before April 1983. I do not believe a
`person of ordinary skill in the art would have read these comments as suggesting that
`different antibody polypeptides should be produced in a single transformed host cell.
`
`29. The Office also indicates that the abstract of the Axel patent makes references to using
`multiple copies of genes. These comments are not suggesting that different proteins of
`interest should be produced in one host cell. The more detailed explanation in the body of
`the patent20 explains that by including multiple copies of the DNA I sequence relative to
`the DNA II sequence, or by using gene amplification of DNA I linked to DNA II, one can
`generate a host cell with multiple copies of the DNA I sequence. If the Axel patent had
`intended “multiple genes” to mean co-transformation and co-expression of three (or more)
`different DNA sequences, the patent would have clearly said so and would have explained
`how to accomplish that goal. For example, the patent could have said to prepare a “DNA
`III” sequence and use it to “co-transform” the host cell with the DNA I and DNA II
`sequences. Since it does not, I do not believe the Axel patent is saying to do this.
`
`The Ochi, Oi, and Rice Publications Illustrate the Unpredictability in the Field as of April 1983
`
`30. The uncertainty reported in the Axel patent in achieving successful expression of even one
`mammalian protein of interest is also seen in the Rice, Ochi, and Oi papers. Each of these
`papers describes efforts to express one immunoglobulin light chain gene in a lymphoid cell.
`
`
`20 See, e.g., id. at col. 6, ln. 44 to col. 7, ln. 26.
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 7
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`31. The Ochi paper provides a good illustration of the unpredictability reported in these
`transformation experiments. The Ochi researchers selected a mutant of an otherwise
`functional hybridoma cell line.21 The mutant had lost the ability to express its endogenous
`light chain gene, but continued to produce its heavy chain.22 The Ochi researchers then
`isolated the light chain gene from the functional parent hybridoma line, and transformed
`the mutant cell line with that gene. In other words, they put the same light chain gene back
`into the mutant cell line that had lost the ability to express that gene.
`
`32. Despite the simple design of this experiment, Ochi reports that 8 of the 14 successfully
`transformed B-cell lines failed to regain any antibody production.23 Of the remaining 6
`successfully transformed cell lines, only one produced antibodies at levels comparable to
`the parent line. In other words, the Ochi researchers report that they failed to achieve
`success 13 out of 14 times in the simplest type of lymphoid cell experiment they could
`design.
`
`33. The Ochi researchers report the same types of problems that the Axel patent reports: they
`were able to successfully transform host cells with the gene encoding their protein of
`interest – an immunoglobulin light chain – but most of the transformed cell lines did not
`successfully express that gene.24 Rice and Oi also report unsuccessful expression in
`successfully transformed lymphoid cells.25
`
`34. Each of these papers shows that successful transformation and expression of even one
`foreign immunoglobulin gene in a lymphoid host cell could not be reasonably expected in
`April 1983. I do not believe these references can be read as suggesting that something
`even more challenging – expressing two different foreign immunoglobulin genes in one
`transformed cell – would have been something that could be predictably achieved at that
`time.
`
`
`21 See Ochi at p. 340, col. 2 (“As recipient cells, we used the mutant cell line igk-14 which was derived
`from Sp603 and does not produce the (cid:539)TNP chain.”).
`22 See id. (“As shown in Fig. 3, the (cid:539)TNP gene is apparently deleted from the igk-14 cell line. Because
`the igk-14 cells still produce the TNP-specific p heavy chain, it would be expected that the expression
`of the (cid:539)TNP gene in these cells would restore the production of TNP-specific IgM.”).
`23 See id. at p. 341 (Table 1).
`24 Ochi reports that each of the 14 cell lines was successfully “transformed” with the foreign light chain
`gene. See id. p. 341, col. 1, middle paragraph. It then reports that 10 of these transformed cell lines
`produced virtually no protein, and all but one of the remaining “successfully transformed” lines
`produced protein at significantly lower levels than the parental line. See id. at p. 341 (Table 1); pp.
`341-342. The Ochi paper also points out that many of the successfully transformed B-cell lines did
`not successfully express the introduced light chain gene.
`25 The Rice paper reports that several of the transformed cell lines showed aberrant transcription of the
`introduced light chain gene. See Rice at p. 7864, col. 1. The Oi paper also reports problems with
`transcription. See Oi at p. 827 (“Twelve independently transformed Y3 and seven BW5147 cell lines
`did not produce detectable amounts of the S107 light chain, as judged by immunoprecipitation and
`gel analysis. XGRPT analyses verified that these cells were, indeed, transformants.”).
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 8
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`35. Dr. Baltimore’s declaration in the reexamination proceeding recognizes this uncertainty.
`He states in his declaration that “if the two chains were expressed” in the same suitably
`transformed mammalian cell, he believes that the cell would properly fold two chains and
`assemble them into a functional antibody. He avoids saying two important things in his
`declaration. First, he does not say that producing a transformed mammalian cell that
`successfully transcribed foreign heavy and foreign light chain genes and also properly
`produced the foreign heavy and foreign light chain polypeptides would have been
`predictable. Second, he does not say that a person of ordinary skill would have been
`motivated by his paper or anything else in the literature at the time of the invention to try to
`produce an immunoglobulin by independently expressing foreign heavy and foreign light
`chain genes in a single transformed cell. I believe Dr. Baltimore did not say anything about
`the prospects or idea of making his hypothetical co-transformed cell because even he
`would have not considered this predictable based on what was known in April 1983.
`
`36.
`
`I also find it important that the Rice, Ochi, and Oi papers describe experiments done by the
`preeminent researchers in the field of immunoglobulin gene expression in April 1983.
`Each was also published in a leading scientific journal. Despite this, none of the papers
`even suggests the idea of transforming a cell line with more than one immunoglobulin gene.
`If these authors believed their work was showing people how to make a recombinant
`antibody by expressing two different foreign immunoglobulin genes in one host cell, they
`would have said something about this somewhere in their papers. Of the three, none
`accomplished it, tried it, or even suggested it.
`
`37. The Rice, Ochi, and Oi papers, considered with the Axel patent, reinforce my belief that if
`someone wanted to try to produce an immunoglobulin molecule using recombinant DNA
`techniques in April 1983, that person would have tried to produce each of the two chains in
`separate host cells, rather than trying to produce both chains in one host cell. That is the
`same message delivered by the Kaplan and Moore references, and is consistent with the
`one polypeptide-one host cell mindset that existed in April 1983.
`
`Dallas Focuses On an Unrelated Goal and Would Not Have Influenced the Beliefs or
`Expectations of a Person of Ordinary Skill in the Art
`
`38. All of the references I have discussed above tell me that if a person of ordinary skill in the
`art wanted to try to produce an immunoglobulin molecule using recombinant DNA
`techniques in April 1983, that person would not have tried to do this by producing both
`immunoglobulin chains in one transformed host cell. The Axel, Rice, Ochi, and Oi
`references make this clear in the context of eukaryotic host cells, and the Kaplan and
`Moore references explicitly say to produce only one immunoglobulin chain at a time in a
`single prokaryotic host cell. The experimental results reported in these references would
`have given a person of ordinary skill no basis to predict or expect that it would be possible
`to achieve what is required by the ’415 patent.
`
`39. The Dallas publication would not have changed the clear message I see in these references.
`This is because the Dallas publication is describing a simple experiment where an E. coli
`cell was transformed with two different E. coli genes, and the expression product of these E.
`coli genes was not isolated or recovered. The Dallas publication simply would not have
`
`SECOND McKNIGHT § 1.132 DECLARATION
`
`
`
`PAGE 9
`
`

`
`CONTROL NOS. 90/007,542, 90/007,859
`
`
`
`ATTORNEY DOCKET NOS. 22338-10230, -10231
`
`provided any relevant guidance or insights into how to successfully produce and recover
`multiple eukaryotic proteins from a single host cell using recombinant DNA techniques in
`April 1983.
`
`40. One reason for my conclusion is the very different goal of the Dallas experiments –
`production of a whole cell bacterial vaccine. In these experiments, there was no need to
`isolate, purify, or even evaluate the proteins encoded by the E. coli genes. There also was
`no need to attempt to assemble the isolated bacterial proteins into a more complicated
`multimeric structure. Instead, the Dallas researchers just had to get an E. coli cell to
`express one or two of the E. coli genes for the experiments to be considered a success.26
`The desired product of the Dallas researchers was the transformed E. coli cell, not an
`isolated functional protein.
`
`41. Another reason for my conclusion is that the Dallas experiments involved transforming and
`expressing E. coli genes in an E. coli cell. These experiments would not