Tải bản đầy đủ - 0trang
2 Genentech's UK patent 2,119,904 on Tissue Plasminogen Activator (t-PA)
Patent Law f o r Biopharmaceuticals
data to support the author‘s conclusions. But for European patent law,
sufficiency and support are now coming to be considered as two sides of the
same coin, since both have to be assessed in relation to the patent claims,
which determine the scope of the legal protection. Both investigations
converge on the effectiveness of the patent description to teach the skilled
person how to put the invention to use. What the inventor has actually done
in the laboratory is, in a sense, less important than the ability to write a recipe
for others to follow.
For biotechnology, patent law first grappled with t h s problem in relation
to inventions in classical microbiology, especially those involving the use of
newly isolated or developed strains of micro-organism to produce some
useful product. The repeatability problem was solved by using culture
collections as official patent depositories of such strains as a substitute for the
well-nigh impossible task of providing an adequate written description both
of the new organism and how to obtain it. This became formalised in an
international convention (the Budapest Convention) whereby deposit in any
one such depository was recogtllsed by all member countries as complying
with their own national requirements. Thls convention now extends beyond
micro-organisms to any biologxal material that cannot be adequately
described by the written word. It follows that this principle must apply to
many of the materials referred to in gene cloning and monoclonal antibody
protocols whch are not available ‘off the shelf to the skilled person wishing
to repeat the prescribed procedure.
Erythropoietin in the European Patent Ofice
(Appeal case T412/93)
European patent No.148 605 was opposed by six Opponents. The case
has assumed considerable legal and t e c h c a l complexity, includmg
consideration of over 500 cited documents, and is not concluded at the time
of this writing. The following remarks are therefore limited to what can be
determined from published accounts.
Thls patent was issued with claims to certain specified DNA sequences
coding for expression of a polypeptide product having at least part of the
primary structural conformation of erythropoietin. The claim covers (i)
specific tabulated sequences, (ii) sequences which hybridize to them , and (iii)
those “whch but for the degeneracy of the genetic code would hybridize to
any of (i) or (ii)”. There are also claims to the polypeptide expression
product of an exogenous DNA sequence. The polypeptide claims are skilfully
worded product-by-process claims which do not cover the polypeptide as
formed in nature.
R Stephen Crespi
A considerable part of the t e c h c a l arguments of expert witnesses
revolved around whether the description was adequate to enable the
preparation of erythropoietin cDNA. This particular enquiry always involves
deciding who is the typical “skilled person” who has to repeat the described
method. The EPO Appeal Board decided that the notional skilled person is “a
Ph.D researcher with several years of experience and two laboratory
assistants having the necessary manual dexterity and lack of fatigue”. The
matter was heavily contested by the scientific experts for all sides. The end
result is that the cDNA claim has not survived but the broader DNA claim is
As to whether the claims to the recombinant polypeptide (r-Epo) were
permissible, the question was whether the product was new in the s&se of
being different from erythropoietin obtained from urine (u-Epo). As indicated
previously t l u s is crucial to the allowability of product claims. The scientific
experts were much exercised on this point also. The claim was finally
amended to specify that the product has greater molecular weight than uEpo.
Another decision of the Appeal Board on this issue is awaited on certain
procedural points following whch the matter will be concluded in the EPO. It
may of course resurface in national courts.
Recombinant Factor VII patent
In view of the inventiveness issues that appeared in the t-PA and epo
cases, it is instructive to note how these problems were anticipated by the
writer of the Factor W patents (US 4,784,950 and EP 0 200,421).
The patent points out that Factor W is a trace plasma protein and the
mRNA encoding it is rare. Purification without degradation was also
problematical. Consequently Factor W was poorly characterised and was
hfficult to obtain in quantities sufficient for sequence analysis.
It was nevertheless necessary to expand on tlus argument to persuade the
EPO examiner to withdraw her objection that, at the priority date of this
patent application (April 1985), it was obvious to clone this particular gene
and how it could be acheved. The applicant pointed out in detailed manner
the uncertainty of malung a cDNA library, the inadequacies of sequence
information for oligonucleotide screening, the uncertainty of antibody tools
then available, and the lack of success of other groups working on the
problem. The EPO examiners are open to t h l s lund of reasonable persuasion
and will concede in the face of such impressive arguments.
For reasons mentioned above, these patents do not have a product claim to
the naked recombinant protein. The claims are mostly directed to the DNA
constructs, recombinant plasmids, transfected cells and methods of
Patent Law for Biopharmaceuticals
production. However, the European Factor W patent has a commercially
very valuable claim to “a pharmaceutical preparation for the treatment of
bleedmg disorders containing a protein having an amino acid sequence as
shown in Figure 1b and free of contaminating human proteins”.
Recombinant Factor IX
In some countries, the Factor IX gene and the protein are covered in
separate patents. The recombinant Factor IX is covered in US 5,171,569
wherein it is claimed as a plasma-free preparation containing the
recombinantly-derived Factor IX protein defined, inter alia, as derived from a
single individual, free of pox viruses and other plasma constituents, and
having a specific activity (as defined in a certain way) of at least 90% of
average normal human plasma.
In Europe the gene for the Factor IX precursor polypeptide (convertible in
vivo into human Factor D( ) is the subject of EP 107,278 wherein it is claimed
as a specified 129 nucleotide sequence.
Hepatitis B patent (Biogen v Medeva)
Thls case is an even more strikmg example of different outcomes obtained
on equivalent patents in different jurisdictions. This patent covers
recombinant DNA molecules having sequences coding for polypeptides
displaymg HBV antigen specificity (claim 1) and HBV antigenicity (claim 2).
There are specific claims to the coding for HBV core antigen and HBV
surface antigen and product-by-process claims to the expression products of
these. Biogen’s European patent 182,442 survived opposition but its British
counterpart underwent the whole rigour of litigation in the UK courts.
Questions of adequacy of description and support for the broad claims figured
prominently throughout h s case ,the most remarkable feature of whch was
the divergmg opinions expressed on these issues by the various judges en
route to the final decision. Finally the UK House of Lords decided that its
claims were too broad. The complexities of h s case have been described in
detail elsewhere (8).
The patent literature is replete with inventions in the field of monoclonal
antibodies. Most of the early development of h s technology, as reflected in
patents, has been directed to the use of monoclonals for the separation and
R Stephen Crespi
purification of bio-molecules and for diagnostic applications. One example of
an early patent to be granted is US patent 4,361,549 whch describes an
OKT3 hybridoma which produces a complement fixing monoclonal antibody
to an antigen found on normal human T cells. The claims are restricted to
mouse monoclonals but the broadest is not restricted to any one T cell
antigen. Diagnostic and therapeutic applications are indicated. This product
is now on the market for use in controlling transplant rejection.
The use of mouse-derived antibodes in humans can induce a human antimouse antibody (HAMA) response which might be serious if repeated use is
necessary. The full therapeutic potential of monoclonal antibodes may
therefore depend on techniques for reducing or eliminating the anti-globulin
response to mouse and other non-human antibodies. The development of
chmeric and ‘humanised’ antibodies formed by recombinant DNA methods
as hybrid immunoglobulins now offers the promise of reducing the immune
response to such antibody products. For reasons touched on earlier,
(e.g. for the t-PA and Epo patent situations), protein engineering technology
gwes rise to the opportunity for new final product patents i.e. those
containing product-per-se claims for the new polypeptides as well as for
DNA sequences that result from these techques. Patents now being granted
in t h s field include some for new principles of immunoglobulin reshaping but
most are for specific products with partially or fully defined amino-acid
sequences. A selection of these is gven below.
The basic principle of antibody reshaping is covered by US patent
5,225,539 and its European equivalent EP 239,400. The claims granted on
t h ~ sinvention cover any ‘altered’ antibody in which the variable domain
framework regons and the complementarity-determining r w o n s (CDRs) are
derived from dlfferent immunoglobulins. The US claims specify this
difference as one of antigen binding specificity or affinity, species, class or
subclass. Inventions whch are the first to open a new field are often
described as “pioneering”, especially if they are followed by a stream of later
developments in the same field. The first patent usually issues with claims of
very broad scope because the patent examiner has not been able to cite
seriously damaging prior art relevant to either novelty or inventiveness.
Such patents are often called “master patents” because they dominate later
inventions. In these situations the requirement for adequate supporting
disclosure has often been met by providing a relatively small number of
‘proof of principle’ model examples. Thls is the case with the patent
numbered above in which one of the specific worked examples describes the
humanisation of one chain of a mouse antibody to a small chemical hapten.
The first example of an antibody humanised to a complex antigen in both
the light and heavy chains is the subject of European patent 328,404.
T h ~ santibody (known as Campath 1-H) binds to the CD52 antigen.
Patent Law for Biopharmaceuticals
The essential feature of t h ~ spatent is the choice of CDRs of specified
amino-acid sequence derived from a particular rat antibody. This patent
seems also to have been the first to include an amino-acid change in
the origmally fully human framework region of one of the immunoglobulin
chains. This favoured the paclung of the CDRs and improved the bindmg
power of the humanised antibody.
Framework changes designed to acheve the above effect are covered
in US patent 5,585,089. Thls is a more than usually complex document
and one of its European counterparts, EP451,216, has itself provoked a
response from others worlung in this field of research and in industry.
This patent covers a number of principles involved in selecting pairings of
CDRs from a donor Ig and frameworks from a human acceptor Ig showing a
hgh degree of homology to the framework of the donor Ig. These principles
include certain defined criteria for the replacement (if necessary) of aminoacids in the heavy or light chain human frameworks by the correspondmg
amino-acids in the donor Ig sequence.
Some other patents on products undergoing clinical evaluation are
US 5,585,097 which covers an aglycosyl anti-CD3 antibody claimed in
terms of fully sequence-defined light and heavy chains and aglycosylated
constant r w o n . Aglycosylation reduces the first-dose (cytolune) response
whch may be encountered with the parent antibody. As with many similar
US patents, there is also a claim in 5,585,097 to a method of treating a patient
with t h ~ santibody to prevent renal allograft rejection. Method claims of this
type are not allowed under European patent law. The reason for this is not
primarily an ethtcal one, as is sometimes thought, but because the procedural
act of treating humans is not an industrial process and therefore does not meet
the ‘susceptibility of industrial application’ requirement for patentability.
Another example of method claims is US patent 5,656,272 whch
describes the preparation of an anti-tumour necrosis factor-a chimeric
antibody for treating Crohn’s disease. This product has received US FDA
approval. The patent claims are all directed to the method of treatment.
A parallel patent application for the antibody itself may also exist.
The writing of patents for nucleic acids and proteins presents an
acute dilemma. One knows that the composition of these molecules may
be modified in various ways, leading to mutants, variants and derivative
forms which may either retain, enhance or reduce, or totally lose the original
R Stephen Crespi
biological activity. The patent draftsman attempts to guard against thud party
avoidance of claims tied too closely to the limited range of specific products
made by the inventors and presented as the patent examples. But in the
absence of data on the effect of compositional variation on activity there is
nothmg to guide h m . The patent examiners will normally stress the uncertain
effect of variation and will insist that the claims are limited to what has been
The patent draftsman will usually prepare the ground for a broad
interpretation of the claims by the use of a skilfully drawn "Definitions"
section. A comprehensive model of this tactic is to be found in the human
tissue plasminogen activator patents, especially US patent 4,766,075 whch is
the equivalent of UK patent 2,119,804 discussed above. These definitions of
t-PA embrace natural allelic variations and derivatives modified by single or
multiple amino-acid substitutions, deletions, additions or replacements in the
t-PA molecule so long as the essential biologcal function of t-PA is retained.
An infringement suit on the US t-PA patents provides an instructive
example of how these matters are treated by the courts.
Genentech v Wellcome Foundation and Genetics
Genentech sued these defendants for infringement of the following three
US patents relevant to t-PA (9) :4,752,603 (the '603 patent) is directed to Human plasminogen activator derived
from the Bowes melanoma cell line and it covers the original work carried out
by Leuven Research and Development. The claims are limited to material of
specific activity of 500,000 IU/mg against a specified reference standard. This
limitation was necessary because of an earlier publication by one of the
inventors describing material purified to an activity of 266,000 units.
4,766,075 (the '075 patent) covers "A DNA isolate consisting essentially of a
DNA sequence encoding human tissue plasminogen activator" and the
corresponding recombinant expression vectors.
4,853,330 covers the process of expressing the DNA of the '075 patent to
Wellcome's product, made in the UK and exported to the US, differs by
Wellcome product (met-t-PA) contained m&onine in place of valine at
only one amino-acid from human t-PA,"the product of the patent".
Patent Law f o r Biopharmaceuticals
The GI product (FElX) was a product having 81 amino acid deletions
from t-PA. It lacked the finger r w o n and most of the epidermal growth
region and had other differences in the knngle region of the native protein.
The District Court of Delaware (in March 1990) held that, on their literal
interpretation, the claims in both the '603 and '075 patents were limited to the
full-length amino-acid sequences of naturally occurring human t-PA and its
naturally occurring allelic variants. The court also decided that the specific
activity limitation in the '603 patent should also apply to the definition of t-PA
in the context of the '075 patent. l h s was particularly surprising because the
claims of the '075 patent were to the DNA coding and not to the protein,
either per se or at any level of purity.
It being admitted that neither met-t-PA nor FElX naturally occur in
humans and that their specific activities were lower than that specified (or
assumed) in the claims, these products were held not to infringe the patent on
the literal interpretation. Wellcome also argued successfully that importation
of met-t-PA into the US, which involved no use of the claimed DNA or
recombinant cell lines in the US, was non-infringing for this reason also.
But US law also has a "doctrine of equivalents'' which the Delaware court
described as "an equitable doctrine permitting a more expansive interpretation
of patent claims than the literal scope thereof'. In view of the material issues
of law involved, the court declined to rule on t l x s issue and reserved it for
further trial. Genentech then applied for a jury trial on the equivalents issue.
l h commenced 7 days later and resulted withn 15 days in verdicts of
infringement by equivalents for both products. One might well marvel at the
idea that a jury could be expected competently and fairly to assess t e c h c a l
and legal issues as complex as those involved in t h s case. However, h s
victory was relatively short-lived (as legal processes go) because it was
reversed by the Court of Appeals for the Federal Circuit (CAFC) in June
W l e the Appeal to CAFC was pending, Wellcome announced their
decision to discontinue development of a t-PA product. Genentech had also
decided not to cross-appeal on the issue of the literal interpretation of the
claims. Therefore the only issue for the CAFC was whether FElX infringed
under the doctrine of equivalents.
The court identified three key issues. The first was whether the specific
activity limitation in the '603 patent applied to the '075 and '330 patents, to
which the court gave a negative answer because these were in every way
independent patents. The second was the basis of measurement of the specific
activity figure. The court decided that this was to be measured by the bovine
fibrin plate assay. The thud and much the most important question for the
court was the meaning of "human tissue plasminogen activator".