IP Metalworks!

Patent language was once stodgy stuff, a special domain of meaning proscribed by experts.  "No so fast", the new courts have seemed to rule, the duty of public notice is met only when the claim language is transparent to an ordinary person skilled in the art, ie. a non-attorney.

Included here is a sample 'biotech" claim revised and improved by the author.   The amended claim recrafts language of a published patent (US 5401629 - invalidated 99-1381 CAFC).

Back to Index

  Clarity and Jargon in Patents:

Thirty years ago, the predecessor court to our new Court of Appeals for the Federal Circuit (CAFC) was empanelled by judges selected from the patent bar.  Patent language was a private professional language, the language of a guild, clear only to those who had been initiated in its syllogisms and had studied the volumes of case law from which that formalized meaning derived.  Now, CAFC is staffed by political appointees with an apparent determination to root out meaning that is not spelled out in the claim itself, or supported in the specification.  The new battle cry is "Plain English".  The court's point is that the public has a right to know exactly what is being patented.  The court seems to be looking not only for lack of ambiguity, but also for plain language, and will interpret patent claims as would be readily understood by one skilled in the art, with little regard for the structured but esoteric language of Patent Attorneys.  

As has been made abundantly clear in recent years, a patent claim must first draw a "bright line" around the claimed subject material, the res, and only then will the court affirm the patentee's right to exclude others from the practice of the invention.  If the line is not sufficiently bright, or tight, then the patent may well be invalid.  Judges are not scientists, but their decision is final. A prospective litigant should be forewarned that their decisions are sometimes not predictable, and for that reason, the art of patent drafting has become a balancing of risk.

Each patent claim must consist of a single sentence.  Given the extreme degree of parsimony and clarity required to condense a complex biotechnological or chemical invention into a single sentence, it is not surprising that we sometimes fall short.  Yet the implications are profound because the validity of the patent is at stake.   When sued, a potential infringer's first line of defense will unfailingly be an attack on the patent's validity, either by fault in novelty, non-obviousness, or by failure of the written description.  Lack of clarity has become a common thread in many cases that are subsequently litigated.  It has become increasingly clear that the court is simply annoyed by claims that it cannot understand, and this puts the plaintiff at a great disadvantage in enforcing an otherwise valid patent.

The problem is perhaps more obvious if we consider a real-life example.  The claim shown below was drawn from a patent that was allowed by the US PTO, and had a presumption of validity under law (35 USC 282), but was invalidated on appeal ten years later in a scathing rebuke by CAFC.  Surprisingly, CAFC's reason for this claim's invalidity was its "obviousness" (35 USC 103a, see dissent in published decision), in the sense that it lacked a truly inventive step, and we shall see how the claim's poor construction may have contributed to the court's perception.

On the right below is the claim edited for clarity; on the left is the claim in its original splendor.  This was not a trivial claim.  Its commercial value was in the tens or perhaps hundreds of millions of dollars.  The claim describes an early cell-based assay for high-throughput screening (HTS), a way of rapidly identifying and selecting a few pharmacologically active compounds out of the billions of inactive compounds known to man.

We claim, (original)

1.  A method for identifying compounds that modulate cell surface-protein mediated activity by detecting intracellular transduction of a signal generated upon interaction of the compound with the cell surface protein, comprising:

comparing the amount of transcription of a reporter gene or the amount of reporter gene product expressed in a first recombinant cell in the presence of the compound with the amount of transcription or product in the absence of the compound, or with the amount of transcription or product in a second recombinant cell; and

selecting compounds that change the amount of transcription of a reporter gene or the amount of reporter gene product expressed in the first recombinant cell in the presence of the compound compared to the amount of transcription or product in the absence of the compound, or compared to the amount of transcription or product in the second recombinant cell, wherein:

the cell surface protein is a surface receptor or ion channel;

the first recombinant cell contains a reporter gene construct and expresses the cell surface protein;

the second recombinant cell is identical to the first recombinant cell, except that it does not express the cell surface protein; and

the reporter gene construct contains:

(a) a transcriptional control element that is responsive to the intracellular signal that is generated by interaction of an agonist with the cell surface protein; and

(b) a reporter gene that encodes a detectable transcriptional or translational product and that is in operative association with the transcriptional control element.


We claim, (amended)

1.A method for selecting pharmaceutical compounds comprising [consisting essentially of*]:

1)  genetically modifying a suitable host cell to express a functional mammalian protein selected from cell surface receptor or ion channel, wherein said host cell is suitable for modification because it is shown to contain no functional endogenous form of said protein;

2)  genetically modifying said host cell with a reporter gene construct, further providing that the transcriptional activity of said reporter gene construct is modulated by a function of said mammalian protein, and that an assay for said transcriptional activity is known;

3)  then treating the genetically modified host cell with at least one exogenous compound;

4)  then assaying said host cell or the cellular milieu for said transcriptional activity;  and,

5)  selecting such compound or compounds that modulate said transcriptional activity.

*** *** ***

[Note: claim reconstructed here for clarity.  Assume "modulate". "assay," "interact", "suitable", "functional", "milieu", and "operatively linked" are defined in specification, and that positive and negative modulation, various classes of signal compounds, various reporter genes, various types of host cells, etc., are the literal subject of multiple dependent claims, with no reliance on claim differentiation.   In the specification, exogenous compounds and mixtures must be defined as pharmacologues, as distinguished from endogenous signaling molecules, in order to overcome the prior art on transcriptional modulation via cell receptors.]

* "Consisting essentially  of" is a more limiting transitional phrase that serves as a "fall back" language if the preferred, broader, "comprising" phrase is rejected on grounds of lack of predictability in the biotechnical arts.

What's wrong with this original claim?  The Federal Court of Appeals overturned a lower court jury decision and found it to be invalid (and hence not infringed).  Why?  At the very least, the language is complex.  Comparing the original claim (which was subsequently judged to be invalid) with the redrafted one (right panel) note that the original claim  begins by describing two cells and several possible combinations, but the amended claim on the left describes only one cell and one assay.  Upon further examination, it becomes clear that the original claim has introduced several experimental controls, basic elements of the scientific method, an unnecessary and confusing complication, as part of a patent claim.

Immediately, two possible problems arise.  First, what if a potential infringer simply did not use a second cell or a negative control?  Answer: the claim is not infringed.  In a simple but very effective method of screening that does not infringe, unknown compounds would be compared with known active compounds in a one cell model, and those compounds that were apparently more potent transcriptional modulators would be selected for further study by another method.

The second problem is more psychological. Claim language must be definite and distinct.  Basic patent law states that a claim must answer the question, "What is the invention?"   Admittedly, in more recent practice imported from Europe, claims with "not" phrases sometimes also describe what is not the invention.  But it is never good practice to claim something that is not patentable, and here the scientific method is a key element in what is claimed.  Any fool scientist knows how to construct a negative control, and it would be obvious to do so when validating the method.  Actually, there are two or three such controls that might be run, and again, there is nothing inventive about doing this.  So is the claim invalid for that reason?  Probably not, but the claim is tainted with the stain of familiarity by its recitation of what is obvious to a skilled artisan.  A judge, in reviewing this claim, must be able to distinguish what is inventive from what is not, and an explanation of the basic scientific method in a claim is both distracting and potentially dilutes the clarity of the inventive step.

Other problems are even more disadvantageous.  The claim recites agonists but not antagonists, and recites "compounds", a very broad word that came close to reading on prior art, where endogenous compounds such as native insulin had already been studied in similar cell constructs, albeit without motivation to screen exogenous pharmacologues.  The overly broad use of "compounds" in the claim strongly supported the "obviousness" defense and was a major factor in arguments over validity.  In revisiting the claim language, the word "compounds" is to be modified by inserting a definition into the specification, and by limiting the claim to "exogenous compounds" only.  By the time you get to court, it's too late to fix the claim language.

The preamble, "detecting intracellular transduction of a signal generated upon interaction", is technobabble.  How is a signal transduced?  One skilled in the art would recognize "transcription", "translation", and "expression", but here the inventors seem to borrow from digital electronics, using functional language without a structural basis.  In electronics jargon, a transducer is a sensor, and generates a signal.   Mis-adoptation of this jargon in a legal claim is risky because the Judge will insist on finding a full explanation of the process and structure in the specification.  Where is the transducer?  Under strict means-plus-function rules, if only one structural means is described in the specification, then only that specific means can be infringed (eg. TATA-type promotor; other upstream and downstream transcription-enhancing sequences would not necessarily infringe).   The preamble becomes a limitation on the claim.  Consider also the Judges frustration in turning to a dictionary, where transducer is defined as an electrical means to convert one form of energy into another, or in biotechnology, an experiment in which DNA is introduced into a host genome by means of a phage.  Jargon may seem acceptable to scientists familiar with the language, but it should be avoided if the goal is make a clear case in court without exhausting a Judge's intelligence or patience.

The scope of these claims may also be limited by the nature of the genetic modification.  It is now known that fully functional receptor proteins or ion channels are not necessary to assay ligand docking.  Protein fragments, partial protein complexes, and chimeric receptor hybrids were not considered.  This omission is perhaps understandable, because some added inventive effort was required to develop those systems.  By claiming "functional mammalian protein" or "native mammalian protein" here, the inventors could have filed subsequently on modified protein constructs without fear of terminal disclaimer.  The use of the word "native", by claim differentiation, would have buttressed language in the specification that allowed for, but did not fully anticipate or enable, the use of modified proteins, such as were engineered by the Cadus team.  These considerations are necessary if the parent filing is not to become 102 prior art that blocks subsequent filings.  The strategy is entirely justified because the initial specification in fact did not enable such workings, the inventors had worked only with a single system in E coli.  It is this author's view that the Cadus yeast constructs were inventive, and in fact achieved by intensive experimentation well outside the bounds established by the Wand's factors.  Not surprisingly, overreaching claims will also contribute to Judicial discomfort, and may in this case have catalyzed invalidation of the claims.

As it turned out, educating the judge, jury, and then the appellate panel, as to the molecular biology of the invention in its historical context was a major effort for both the defense and plaintiff.  At the time the patent was filed, a cell with an insulin receptor and a reporter gene construct had been used to study the mechanism of insulin-dependent regulation of transcription.  But the patent claim does not claim the cell construct as a composition of matter, it claims the use of such cell constructs in identifying new drugs.  Novelty was not contested, and a jury found the patent claim valid on the issue of non-obviousness.  CAFC, in overturning the Jury's finding on non-obviousness, may have errored by hindsight, seeing a motivation for substituting a pharmacologue for an endocrine where no explicit motivation was present.  Was there an inventive step between studying the mechanism of insulin action by constructing a reporter gene model for transcriptional control, and then conceiving the general use of  reporter genes to screen for drugs as calcium channel blockers or estrogen receptor antagonists?   The courts have long struggled with the question as to how "big" an inventive step has to be to merit a patent, and how to avoid hindsight in assessing the conceptual leap at the instant it was made.  But was it also the failure of the claim to clearly point out what was inventive that sunk this patent on appeal? 

I've reviewed the litigation in much more detail in the first installment of this journal [link to "A Cautionary Tale"].  Here I try to suggest how an erroneous decision on obviousness might have been avoided.  A claim must clearly describe the invention, and indeed, it must encapsulate the inventive step, so that a Judge can clearly see it.  If the Judge can't see it, the merits of the case are not going to decide the issue.

You be the judge as to whether the original claim language is successful.  I welcome your comments as to whether the amended claim recited above fixes the problem.  In this case, a $20M Judgement and the death of a company was decided with the choice of every single word in the claim.  Patent Attorneys (and Patent Agents) must  successfully draft each claim so that it withstands de novo claim construction at two levels of judical review.

There was another issue at trial which is also illustrative of the size of the task that education of Judge and Jury must be.  The above claim does not distinguish the classes of host cells suitable for genetic modification by transfection (or transformation).  The specification was also silent, and a single example described only mammalian tissue culture cells.  The alleged infringer was actually using a yeast cell model, and the teachings of the patent were shown to be not operative in yeast, but this was not the deciding issue at trial, and the argument bounced off the appellate Judges, again demonstrating the level of damage that comes from not successfully educating a Jury or Judge.

The controversy discussed here highlights use of plain language. Sharp critics may rush to point out that the amended claim language still includes some peculiar language such as "said element", etc., but this quaint patentese remains embedded in the traditions of law surrounding patents, and once the reader gets the hang of it, patent claims are more rigorous and explicit by its usage.  Antecedent claiming is a statutory requirement [112 2], hence the substitution of "said" by a plain "the" could have significant legal consequences.  The court has not found a problem with this embellishment of everyday English.  The substitution of "further providing" (above) in place of the more traditional "wherein" is unconventional and should be used with caution.  The message here is not that the court requires everyday, common language, rather that the court requires explicit, transparent language.  In Chemical patents, Markush claims may run on for pages, a grammarian's nightmare, but nonetheless have a literal meaning that can be unentangled phrase by phrase, substituent by substituent, with patience. 

Cases have been decided on the edition of the dictionary to consult in determining the plain meaning of a word, but this issue is readily avoided by simply including definitions for all the technical words used in the claims.  The specification is the place for those definitions.  In some instances, it may be useful to refer the reader to another citation where suitable definitions are given, but too many broad or, worse,  inconsistent definitions "incorporated by reference" can result in ambiguity that the judge must untangle (most often at the Inventor's expense). 

Inventors are invited by law to be their own lexographers, and its just common sense to write the specification to educate the judge, jury or panel exactly what the inventor means by "cell", "reporter", "modulate", "control", "transduce" or "operatively linked", etc, without rewriting the basic textbook on genetics and molecular biology.  Good judgement in limiting the definitions is approximated by the common sense given to a person skilled in the art.  Clearly, we must all take that view.

In the next edition, we'll try to improve on the claim language of a valid patent. 

To Table of Contents


Please address any comments to KLambert@4iPT.com.

  Copyright K K Lambert 2001-2005.  All rights reserved.   For permissions, contact the author.