Elan Pharmaceuticals, Inc., and Athena Neurosciences, Inc. v. Mayo Foundation for Medical Education and Research

Citations: 304 F.3d 1221; 64 U.S.P.Q. 2d (BNA) 1292; 2002 U.S. App. LEXIS 18007; 2002 WL 2003002Docket: 00-1467

Court: Court of Appeals for the Federal Circuit; August 30, 2002; Federal Appellate Court

SummaryOriginalCytator Case TreatmentBackward CitatorAnalysis

EnglishEspañolSimplified EnglishEspañol Fácil

Elan Pharmaceuticals, Inc. and Athena Neurosciences, Inc. appeal a summary judgment from the United States District Court for the Northern District of California favoring the Mayo Foundation for Medical Education and Research. The district court invalidated Elan's two patents—U.S. Patent No. 5,612,486 and U.S. Patent No. 5,850,003—claiming they were anticipated by U.S. Patent No. 5,455,169, known as the Mullan patent. The Federal Circuit, led by Circuit Judge Pauline Newman, reversed this ruling, determining that the legal criteria for anticipation were not satisfied based on the record. The case is remanded for further proceedings. 

Elan's patents involve genetically modified transgenic animals designed to be susceptible to Alzheimer's disease, incorporating a specific mutated gene known as the 'Swedish mutation.' This mutation is linked to a higher incidence of early-onset Alzheimer's in a Swedish family. The patents discuss the formation of beta-amyloid peptide (betaAP), a key component of Alzheimer's plaques, and the enzymatic processes involving amyloid precursor protein (APP) that lead to the production of betaAP.

Humans who do not develop Alzheimer's disease (AD) effectively break down amyloid precursor protein (APP) without producing significant beta-amyloid (betaAP). The district court's summary judgment of anticipation is based on the Mullan patent, which details the identification of a mutated gene linked to Alzheimer's from a Swedish family. Dr. Mullan discovered that the mutation alters the nucleotides at codons 670 and 671, replacing lysine and methionine with asparagine and leucine. The patent describes the creation of transgenic non-human animals that express this mutated gene, which can be utilized in AD research and treatment. Specifically, it outlines methods for screening agents that may treat AD by observing the expression and processing of APP in these transgenic models.

Mullan describes techniques for creating transgenic animals, including cloning and isolating the mutated human gene for integration into host organisms, particularly mice. The patent emphasizes that these transgenic animals will express variant polypeptides in a neuron-specific manner, ideally using the full human APP gene sequence. Various gene transfer methods are referenced, such as homologous recombination in embryonic stem cells and microinjection into embryos, along with citations of scientific literature supporting these approaches.

Viral vectors, such as adeno-associated virus, can be employed to introduce a mutated gene into stem cells or developed animals for screening purposes. Mullan notes that the mouse gene allele can be altered to reflect the Swedish mutation by methods such as site-directed mutagenesis or gene conversion, targeting the specific codon position that corresponds to the human APP gene. However, it is acknowledged that Mullan did not create a transgenic animal with this mutation or identify effective methods for doing so, with expert testimony highlighting the challenges and low success rates associated with creating transgenic animals.

The Elan patents detail the generation of transgenic mice with DNA containing the Swedish mutation in the APP gene, which leads to the production of human beta-amyloid peptide (betaAP) through mouse enzymes. The process of detecting betaAP in the mouse brain is complicated due to the small size of the betaAP molecule, necessitating detection via larger cleavage fragments. 

Claim 1 of the '486 patent specifies a transgenic rodent whose genome includes a transgene for a heterologous APP polypeptide with the Swedish mutation, specifically encoding asparagine and leucine at codon positions corresponding to human APP695, and producing detectable ATF-betaAPP in brain homogenates. The '003 patent includes similar claims but also specifies the presence of a promoter and a polyadenylation site in the transgene. Both patents assert that the transgene is expressed in a murine rodent and is nonhomologously integrated, although these aspects are not claimed as distinguishing features. The focus of the appeal is primarily on the '486 patent.

The Mullan patent is considered prior art due to its earlier filing date, and the Elan patents were granted only after Elan amended its claims to include a specific clause. Elan contends that its claims are limited to detectable levels of ATF-betaAPP in the rodent brain, a limitation not addressed by Mullan, which Elan argues means the claims cannot be anticipated. Elan asserts that ATF-betaAPP was not disclosed in humans until after Mullan's filing, making the limitation non-inherent in Mullan's disclosure. 

The district court determined that while Mullan does not explicitly mention ATF-betaAPP formation, it is inherently taught through Mullan’s description of transgenic mice with the Swedish mutation. Consequently, the court ruled that Mullan anticipates Elan's claims, declaring both patents invalid on summary judgment. 

Summary judgment requires no genuine issues of material fact when viewed favorably to the non-moving party. Elan disputes the anticipation criteria and claims that material facts were in contention, asserting it would win if these facts were resolved in its favor. To secure a patent, an invention must be novel, as defined by 35 U.S.C. 101, 102(a), meaning it cannot be known or used by others beforehand. Anticipation is a factual question, requiring that all elements and limitations of the claimed invention be explicitly or inherently disclosed in a single prior art reference. The reference must describe the invention clearly enough for a person skilled in the field to recognize it exists, and if anticipation relies on inherent limitations, it must be demonstrated that such information was known to be present in the prior art.

An inherent limitation is one that is necessarily present and cannot be established through probabilities or possibilities. In the case of Scaltech, Inc. v. Retec/Tetra, LLC, the district court determined that the claims made by Elan were anticipated by Mullan's prior art, which suggested that standard procedures would yield a small percentage of transgenic mice capable of producing detectable ATF-betaAPP. The court found the absence of specific mention of ATF-betaAPP in Mullan irrelevant, reasoning that the known low success rate of gene transfer implied some successful outcomes were statistically probable.

Elan contended that Mullan merely provided broad instructions for gene transfer, lacking assurance of success. Although Mullan outlined procedures for creating transgenic animals, he did not detail all elements of Elan's claims or clearly demonstrate the production of transgenic mice with detectable ATF-betaAPP. The court asserted that general instructions for gene transfer, which involve inherent uncertainties, do not meet legal standards for anticipation of a successful transgenic product. 

Elan emphasized that the ability to process human APPsw to create detectable ATF-betaAPP was a crucial element of its claims, distinguishing its invention from Mullan's. Elan noted that the patent examiner required this limitation to differentiate from Mullan, which did not address the formation of detectable ATF-betaAPP. Mayo, however, argued that this limitation was inherent in Mullan because successful transgenic procedures would produce ATF-betaAPP. The district court agreed but noted that Mullan did not provide evidence for this claim, and the existence of ATF-betaAPP in the transgenic mice was not established as common knowledge among skilled practitioners in the field. Inherency must be supported by prior art evidence, not merely the knowledge of the inventor, in order to avoid hindsight bias.

A prior patent or publication can only anticipate a later patent if it provides clear instructions for practicing the invention. If the earlier reference merely serves as a starting point for experimentation without guaranteeing success or offering comprehensive guidance, it does not count as anticipation. In this case, the court found that the legal criteria for anticipation were not satisfied, leading to the reversal of the summary judgment of invalidity and remanding the case for further proceedings.

Elan claims that its patent covers all transgenic mice with the Swedish mutation. If construed broadly, these claims could be deemed invalid under sections 103 or 112; however, the district court did not address these issues. The court clarified that a novel patented product cannot be considered anticipated if it did not previously exist. The dissenting opinion incorrectly suggested that the ruling allows for the patenting of existing inventions, while the court maintained that Elan's invention—a transgenic mouse—did not exist before.

Mullan had the idea of creating such a mouse but did not actually make one or specify effective methods from existing literature. The court emphasized that a mere proposal does not meet the anticipation standard. The dissent argued that this ruling could lead to the patenting of existing inventions by merely identifying an inherent characteristic. However, the court distinguished this case from prior examples, asserting that the Mullan mouse was not a known entity like the broccoli sprouts referenced in earlier litigation.

The dissent's view that any suggested but unmade product should be unpatentable could hinder the potential for patent protection on new transgenic products. The established law dictates that new products cannot be anticipated if they did not exist before, regardless of the general knowledge of their production methods. The court noted that the implications of this ruling raise significant policy questions, but did not alter established principles governing anticipation.

Anticipation in patent law refers to the prior knowledge of a subject matter. A mere suggestion of procedures that may produce a novel product does not constitute prior existence of that product. At the time of the Mullan application, it was agreed that no one, including Mullan, had created a mouse with the Swedish mutated gene or had definitive knowledge regarding the gene's acceptance by mouse DNA, expression of the human mutated protein, or cleavage of that protein by mouse enzymes. Expert testimony indicated significant uncertainty in the ability of mice to generate specific APP fragments for drug discovery. It was established that Mullan had not produced a mouse through any proposed method and that predicting the success of any method was not feasible. The dissent's assertion that a successful mouse could be considered prior art was rejected, as anticipation is not determined by an inventor's own disclosures but by prior art known to those skilled in the field. The scope of Elan's claims and whether they cover the Mayo mouse were not adjudicated in this appeal. Additionally, Elan has separate patents for detecting ATF-beta APP, which do not necessitate a detection method within the claims at issue. The decision was reversed and remanded.

The mutation positions at codons 670/671 and 596/597 are identical due to different starting points in the APP chain, as referenced in the '486 patent. Elan argues that producing a successful transgenic mouse is uncertain and difficult, noting that the Mayo mouse was the 2,576th screened. A dissenting opinion emphasizes that an inventor's disclosure cannot be used as proof of anticipation under 35 U.S.C. § 102, contradicting existing case law. The dissent warns that allowing patent rights based on inherent characteristics not previously identified in prior art undermines established patent law. The patents in question are U.S. Patent Nos. 5,612,486 and 5,850,003, collectively referred to as the Elan patents. The independent claim of the '486 patent describes a transgenic rodent with a transgene encoding a heterologous APP polypeptide with the Swedish mutation, specifically requiring detectable ATF-betaAPP in the brain homogenate. Similarly, the claim of the '003 patent outlines a transgenic rodent with a DNA segment encoding the same polypeptide. The claims are product-focused, not method-based, and arguments regarding methods of detection are deemed irrelevant. Elan concedes that the Mullan patent disclosed a transgenic mouse with the Swedish mutation, acknowledging the first two claim elements but disputing only the requirement for detectable ATF-betaAPP levels.

Elan asserts that its claimed rodents are novel due to the processing of amyloid precursor protein (APP) to ATF-betaAPP in detectable amounts within brain homogenates. Mayo acknowledges that Mullan does not explicitly disclose this aspect but argues it is inherently present in Mullan’s disclosure. The central question is whether Mullan’s rodent inherently produces detectable levels of ATF-betaAPP. The district court concluded that Mullan inherently anticipates Elan’s patent claims, stating that the claimed mice are a subset of those described by Mullan, which would naturally exhibit detectable ATF-betaAPP. The court indicated that if Elan's claims were otherwise, the patents would not be enabled. However, the majority opinion disagrees, noting a lack of extrinsic evidence supporting the inherency of ATF-betaAPP in Mullan's mice, and emphasizes that inherency cannot rely solely on the inventor's knowledge; it must be substantiated by prior art evidence. The majority's position is that there is no legal basis for requiring proof of inherency exclusively from prior art, acknowledging that extrinsic evidence can fill gaps in the prior art reference. The text critiques the majority's conclusion as legally incorrect and inconsistent with established case law, which states that discovering inherent characteristics of an existing product does not qualify for patentability.

Claimed inventions lose patentability if they lack novelty, regardless of any properties stated in prior art. For the Elan patents, the focus is whether they prove that the Mullan transgenic rodent inherently possesses the claimed characteristic, and they do. The '003 patent describes newly identified fragments that result from the cleavage of the APP protein, highlighting that these fragments, particularly ATF-βAPP, are produced through an alternative secretory pathway, potentially linked to diseases related to β-amyloid plaque. Transgenic mice expressing the Swedish mutation of the APP gene produce significantly higher levels of ATF-βAPP compared to wild-type. The Mullan patent also describes a transgenic mouse that not only carries the Swedish mutation but actively expresses and metabolizes it to produce β-amyloid peptide, thereby creating a model for studying APP and β-amyloid metabolism. This transgenic model is capable of producing detectable amounts of the ATF-βAPP metabolite, as required by the patent claims. Elan's appeal acknowledges the Mullan patent’s disclosure, which establishes a clear basis for the claims regarding the transgenic mouse's capabilities.

Elan's patents claim transgenic mice that are a small and unexpected subset of those potentially produced under the Mullan method. However, the claimed mice are not merely a small subset; Mullan's method does disclose the Swedish form of APP, which all successful mice produced will express. Therefore, any mouse that processes the Swedish form APP will also produce the ATF.APP characteristic claimed in Elan's patents. The principle of "inherency" does not require probabilities or possibilities to establish anticipation, and the district court correctly found that Elan's claims were inherently anticipated by Mullan. The dissent notes that while the majority suggests Mullan did not disclose a specific element, this issue was not raised on appeal. Mullan does disclose that the transgenic animals will express variant APP polypeptides, and there is no authority requiring a more detailed description. While Mullan's enablement could be questioned, Elan chose not to challenge it, likely due to the risk of invalidating their own claims.