Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc. (Fed. Cir. 2013)

By Andrew Williams --

A claim term that can have different meanings or values depending on the method used to measure it renders the claim indefinite because it is impossible for a potential infringer to discern the boundaries of the claim.  This is because it is not possible to determine whether potential infringing activity will fall within the scope of the claims.  This principle of patent law was reaffirmed by the Federal Circuit last week in Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc.  The Court held that because the claim term "molecular weight" can be calculated at least three different ways, resulting in three different values, the asserted claims that included this value were invalid as indefinite.  However, the Court also held that at least some of the claims that included this term, but which explained in the claim how to calculate "molecular weight," were correctly found to be definite.  In addition, the Court determined that there was insufficient evidence to overturn the lower court's determinations that the asserted claims were enabled, not obvious, and infringed.  The impact, therefore, was that Teva's injunction of the various ANDA filers was maintained, but the issue of the length of the injunction was remanded to the lower court.

As with most ANDA cases, it is important to have a grasp of the technology to understand the issues involved in this case.  The drug at issue in this case was Copaxone®, which is marketed by Teva for treating multiple sclerosis.  The active ingredient is copolymer-1, or glatiramer acetate, which is a polypeptide product that consists of four different amino acids (alanine, glutamic acid, lysine, and tyrosine).  Copolymer-1, however, is a mixture of individual polymer molecules with different constituent ratios, and different molecular weights.  Therefore, it is often expressed as "average molecular weight."  However, statistically speaking, there are at least three ways to calculate average molecular weight:  peak average molecular weight (M_p), number average molecular weight (M_n), and weight average molecular weight (M_w).  One way to measure "molecular weight" of these molecules is to use Size Exclusion Chromatography (SEC).  A simplistic way to think about SEC is as a technique that utilizes a column "sieve" that allows different size (or weight) molecules to leave (elute) the column at different times.  If the column is calibrated, you can determine the size or weight of the molecules in question depending on when they elute from the column.  The peak average molecular weight (M_p) is the molecular weight of the most abundant molecule in the sample, which occurs as the highest peak in the elution profile.  The number average molecular weight (M_n) is the combined total mass of all of the molecules in the sample divided by the total number of molecules.  Judge Moore, writing for the three-judge panel, chose to not define weight average molecular weight (M_w), instead referring to it as "still another average molecular weight measure that is calculated differently from M_p and M_n."  For readers that are curious, weight average molecular weight (M_w) is similar to number average molecular weight, except instead of dividing the sample by total number of molecules, you take into account the weight fraction of each molecule.  Therefore, the larger the molecule, the more "weight" it will have in the value.

There is a second way to express molecular weight, which was used in the patents-in-suit in the so-called "Group II" claims.  Claim 1 of U.S. Patent No. 6,054,430 is representative of these Group II claims:

    1. Copolymer-1 having over 75% of its molar fraction within the molecular weight range from about 2 kDa to about 20 kDa, prepared by a process comprising the steps of:
    reacting protected copolymer-1 with hydrobromic acid to form trifluoroacetyl copolymer-1 having over 75% of its molar fraction within the molecular weight range from about 2 kDa to about 20 kDa, wherein said reaction takes place for a time and at a temperature predetermined by test reaction, and
    treating said trifluoroacetyl copolymer-1 having over 75% of its molar fraction within the molecular weight range from about 2 kDa to about 20 kDa with aqueous piperidine solution to form copolymer-1 having over 75% of its molar fraction within the molecular weight range from about 2kDa to about 20kDa.

In this case, the molecular weight is not expressed as a statistical measure, but rather the percentage of copolymer-1 molecules that fall within a defined range.  As a result, the measurement is not dependent on the type of calculation performed because it requires a determination of the actual values.

The procedural posture of this case is typical for an ANDA case (notwithstanding the fact that there appear to be generic companies on both sides of the "V").  Teva holds the NDA for Copaxone®, for which eight patents are listed in the Orange Book.  Several companies separately submitted Abbreviated New Drug Applications (ANDAs) to market generic versions of this drug.  Teva sued these companies for patent infringement, and at least seven of the eight Orange Book listed patents were asserted.  All seven of these patents shared the same specification, because they descended from the same parent, U.S. Patent No. 5,800,808 ("the '808 patent"), which was also asserted in this case.  This is significant because the '808 patent was filed on May 22, 1995, and therefore has a later expiration date than all of the other patents (seventeen years from date of issuance, or Sept. 1, 2015, instead of May 24, 2014, the expiration date of the remaining patents).  Thus, when the Court determined the claims of the '808 patent to be indefinite, it remanded the case to the lower court to determine the impact this would have on the length of the injunction.  As for the lower court's findings, it defined the term "molecular weight" in the Group I claims as peak average molecular weight (M_p), and rejected the argument that the term was insolubly ambiguous.  It also held that all of the asserted claims were infringed, and that they were not invalid for obviousness or lack of enablement.  The ANDA filers appealed.

Indefiniteness

The Appellants challenged the lower court's determination that "molecular weight" could be defined, because one of skill in the art would not understand which measurement (M_p, M_w, M_n, or something else) should be used.  The specification of the patents-in-suit was silent as to which measurement was appropriate, other than to describe the use of SEC (which can be used to obtain all three values).  Statements made in the prosecution history compounded the problem.  In response to an indefiniteness rejection in one of the patents, the patentees said that "[o]ne of ordinary skill in the art, upon reviewing the specification, would understand that 'average molecular weight' refers to the molecular weight at the peak of the molecular weight distribution curve," or M_p.  However, in response to an identical rejection in an earlier related application, the patentees said that ""[o]ne of ordinary skill in the art could understand that kilodalton units implies [sic] a weight average molecular weight," or M_w.

Teva argued that one skilled in the art would have understood that average molecular weight could only be M_p.  Teva asserted that use of SEC without more suggests M_p because determining M_n and M_w required additional calculations.  The Court appeared skeptical of this argument, because SEC can be used to calculate all three values.  Teva also downplayed the prosecution history by explaining that a skilled artisan would recognize that the statement in the earlier application was false -- because "kilodalton units" does not imply weight average.  Teva appears to have alleged that the second statement rectified the earlier misstatement.  However, there is no indication that the patentees in the later application retracted the earlier statement.  Therefore, the Court was left with two contradictory statements that could not be rectified.  As such, the claims in Group I were held to be indefinite.

However, the claims in Group II survived the invalidity challenge.  The Appellants argued that the claim term "molecular weight" was the same in all of the asserted claims, and therefore they should all rise and fall together.  However, the Federal Circuit pointed out that the use of the term "molecular weight" in these claims was not in reference to an average, but rather to a range, as shown in the representative claim above.  Therefore, because the claim had exact values instead of statistical measures, the scope was readily ascertainable.

Enablement

Appellants also alleged that the claims were not enabled because the specification did not specify how the SEC column should be calibrated, and therefore either self-standards or universal calibration could be used.  They argued that copolymer-1 is complex and exhibits an unpredictable behavior.  In fact, they pointed to Teva's difficulties in identifying an appropriate calibration method.

The Federal Circuit found no clear error with the fact finding of the lower court.  The lower court weighed the testimony of the experts, and found Teva's expert to be more convincing.  Moreover, there was no evidence that Teva's own alleged difficulties were based on faulty measurements as opposed to regulatory or scale-up issues.  As a result, the lower court's determination was affirmed.

Obviousness

Appellants also challenged the lower court's holding that the patents at issue would not have been obvious in view of the prior art.  The closest prior art reference disclosed compounds with a molecular weight higher than 10 kDa.  Appellants alleged that there was no evidence that copolymer-1 with a molecular weight less than 10 kDa exhibits an improved toxicity profile over the prior art.  However, the Federal Circuit found no error with the District Court's determination that there was a preference in the prior art for higher molecule weight compounds, and therefore it taught away from the claimed invention.  Also, the Court noted that Copaxone® was coextensive with the asserted claims.  Therefore, a nexus was presumed for purposes of the undisputed commercial success of the drug product, and Appellants had not rebutted this presumption.  As a result, because the Appellants failed to establish the obviousness of the claimed invention, the Federal Circuit affirmed.

Infringement

Finally, Appellants challenged the finding of infringement.  First, they challenged the Court's construction of the term "approximately" in relation to the molar ratio of the amino acid subunits.  However, "approximately" was not a claim term, but instead it was found in a claim term definition.  The term at issue was "copolymer-1" and the District Court construed it to mean "a mixture of polypeptides composed of alanine, glutamic acid, lysine, and tyrosine in a molar ratio of approximately 6:2:5:1."  As such, the application of the term "approximately" was part of the infringement analysis, not of claim construction, and therefore the standard of review was clearly error.

In making the infringement analysis, the lower court determined that a ratio of 6:2:5:1 was equivalent to a percentage ratio of 42.9%, 14.3%, 35.7%, and 7.1%.  In comparison, Mylan's ratio was 42.7%, 14.4%, 33.6%, 9.2% (the ratio for Sandoz was comparable, so only Mylan's product will be discussed here).  The court calculated that this aggregate difference was only 4.5%.  As compared with the prior art, which tolerated aggregate differences of 12%, it found the accused products to fall within the scope of the claims.  Appellants asserted that the relative proportions were important, not the aggregate.  Therefore, because Mylan's ratios were 4.6:1.6:3.7:1.0, the ratio of lysine to tyrosine was 3.7:1.0, and therefore deviated more than 16% from the ratio of 5.0:1.0.  The Federal Circuit agreed with Teva, however, that Appellants were performing a "mathematical sleight of hand."  In essence, they were normalizing the ratio to the least abundant amino acid, tyrosine, which will skew the results.  Ultimately, the Federal Circuit found that the lower court did not clearly err by looking at the aggregate percentage difference and finding it less than 5% from the ideal.

Finally, Appellants asserted that the Teva had disclaimed copolymer-1 compositions having M_w greater than 10kDa.  The basis for this assertion was that the patentees had compared the pending claims to a reference describing copolymer-1 with a minimum molecular weight of 10 kDa, stating that the then-pending claims cover copolymer-1 having a molecular weight of about 5 to 9 kDa.  Somewhat ironically, the Federal Circuit found that there was no a clear disavowal of the claim scope, because it was unclear that the patentees were referring to M_w.  Therefore, for almost the same reason that some of the claims were found to be indefinite, the Court found that there was no disavowal, and therefore infringement determination was affirmed.

Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc. (Fed. Cir. 2013)
Panel:  Chief Judge Rader, Circuit Judge Moore, and District Judge Benson
Opinion by Circuit Judge Moore

Comments

[Manisha Chhabra]

Patent Expiry of U.S. Patent No. 5,800,808 is mentioned as "(twenty years from date of issuance, or Sept. 1, 2018".

According to my understanding it should be "(seventeen years from date of issuance, or Sept. 1, 2015".

[Andrew Williams]

Dear Manisha,

Thank you for catching that mistake. Of course, you are correct, it should be seventeen years from the date of issuance (and we will try to correct the original post). However, the original point is still the same – the ‘808 patent has a later expiration date than the other patents at issue. Oddly, the Federal Circuit glossed over this fact, only looking to the Orange Book for patent expiration information. In fact, during oral argument, the Court seemed to be confused by the fact that there could be non-Orange Book listed patents in suit. The result should be a shortening of the injunction because of the invalidation of the ‘808 patent.

Thank you again for catching that mistake.

Andrew

[Igor Faynshteyn, Esq.]

Although this is somewhat off topic, I wanted to ask: do federal district courts sometimes use neutral expert witnesses, to clarify scientific, or otherwise technical issues for the jury?
In a fairly recent debate about patents, between Judge R. Posner and Professor R. Epstein, Judge Posner said that he likes using neutral witnesses when he volunteers as a judge in patent cases, at the district court level, because they are less "coached" by their lawyers and are therefore more honest and straight-forward in volunteering their technical expertise.
Is this practice picking up some steam, or do courts remain generally conservative in this regard and maintain a robust adversarial legal structure?

Igor Faynshteyn, Esq.