[2014 EWHC] 3857 (Pat) is a decision by Mr Justice Birss in the Patents Court, England and Wales. This case is about a stable formulation containing the antibody Herceptin (or ‘trastuzumab’) which is used to treat breast cancer. The formulation contains Herceptin and three specified ‘excipients’. The decision concerns Genentech’s two patents EP 1516628 and EP 2275119, which are divisionals from the same parent case. Herceptin itself was protected by a different patent, EP 0590058 for which the SPC expired on 29 July 2014. Hospira now wish to sell generic Herceptin and in order to do so have to obtain a generic authorisation based on the existing marketing authorisation for Herceptin. That would require them to use the same formulation as claimed in the present two patents, and so in this action Hospira sought revocation of Genentech’s patents to ‘clear the way’ for Hospira to sell generic Herceptin.
The decision has the following interesting points:
- the importance of common general knowledge in the field of ‘formulations’,
- the dangers of deleting features from claims post-grant and the complexity of generalising from specific examples,
- the ‘esoteric’ nature of product by process claims and how they can have different scopes for validity and infringement,
- an illustration of how to determine whether ‘new information’ is present in an amendment (which is different from the EPO’s approach to added matter),
- the relevance of the EPO’s would/could test to inventive step analysis in the present case, and
- a discussion of why the judgment has not fallen foul of the ‘Technograph’ test, in which a ‘step by step’ analysis can be based on hindsight.
The technology and the issues to be decided
The invention concerns a stable lyophilised (i.e. freeze dried) formulation for Herceptin which can be stored for long periods of time and which is reconstituted with water before administration to a patient. The formulation comprises four specified components: Herceptin, trehalose (a lyoprotectant), histidine (a buffer) and polysorbate 20 (a surfactant). Herceptin was known at the priority date and therefore the invention only lies in deriving this specific four-component formulation. The claims place limitations on the ratios and concentrations of the components, but these limitations were not deemed to make an inventive contribution.
Genentech requested significant amendment to the claims of both patents, and much of the decision deals with issues raised by the amendments. The main issues that were decided were:
- whether the amended claims extend the scope of protection,
- whether the amendments should be allowed given the introduction of a product by process feature,
- whether the amended claims add matter, and
- whether the claims were obvious.
Claims 1 and 2 from both patents are below showing the amendments that Genentech requested (the underlining shows text that has been introduced into the claims). Similar amendments were made to generate new claims 3 and 4 for EP 2275119 which are not shown here. The amendments include deletion of the feature relating to the 200-600:1 ratio of lyoprotectant to antibody and insertion of an ‘obtainable by’ product by process feature.
1. A formulation comprising a lyophilized mixture of a lyoprotectant, a buffer, a surfactant and an antibody,
wherein the molar ratio of lyoprotectant : antibody is 200 600 moles lyoprotectant : 1 mole antibody,wherein the lyoprotectant is trehalose or sucrosewherein the buffer is histidine, wherein the surfactant is polysorbate 20 and wherein the antibody is a monoclonal antibodyhuMAb4D5-8, obtainable by lyophilizing a solution containing 25 mg/ml huMAb4D5-8, 5mM histidine pH 6.0, 60 mM trehalose and 0.01% polysorbate 20.
2. A formulation comprising a lyophilized mixture of a lyoprotectant. a buffer, a surfactant and an antibody, wherein the lyoprotectant is trehalose, wherein the buffer is histidine, wherein the surfactant is polysorbate 20 and wherein the antibody is huMAb4D5-8, such that an amount of said lyophilized mixture containing 450 mg of said antibody can be reconstituted with 20 ml of BWFI (0.9 or 1.1% benzyl alcohol) to yield a concentrated protein solution containing 22 mg/ml of said antibody, 52 mM trehalose, 4 mM histidine, pH 6.0, 0.009% polysorbate 20.
1. Use of a lyophilized formulation comprising a monoclonal antibody, a buffer, a surfactant and a lyoprotectant
, wherein the molar ratio of lyoprotectant : antibody is 200 600 moles lyoprotectant : 1 mole antibody, wherein the lyoprotectant is trehalose, wherein the buffer is histidine, wherein the surfactant is polysorbate 20, and wherein the monoclonal antibody is huMAb4D5-8 an anti HER2 antibody, obtainable by lyophilizing a solution containing 25 mg/ml huMAb4D5-8, 5mM histidine pH 6.0, 60 mM trehalose and 0.01% polysorbate 20, in the preparation of a medicament for the treatment of abreast cancer characterised by overexpression of the HER2 receptor.
2. Use of a formulation comprising a lyophilized mixture of a lyoprotectant. a buffer, a surfactant and a monoclonal antibody, wherein the lyoprotectant is trehalose, wherein the buffer is histidine, wherein the surfactant is polysorbate 20 and wherein the monoclonal antibody is huMAb4D5-8, such that an amount of said lyophilized mixture containing 450 mg of said antibody can be reconstituted with 20 ml of BWFI (0.9 or 1.1% benzyl alcohol) to yield a concentrated protein solution containing 22 mg/ml of said antibody, 52 mM trehalose, 4 mM histidine, pH 6.0, 0.009% polysorbate 20, in the preparation of a medicament for the treatment of breast cancer characterised by overexpression of the HER2 receptor.
Common General Knowledge
The expert evidence was used to establish the common general knowledge in the field of formulations, including determining how the skilled formulator would have acted. At the priority date of March 1996 a phase II clinical study of Herceptin had been published (Baselga et al) and it was known that a phase III clinical trial was happening. However Hospira could not establish that these would be known to all in the art and therefore they were not deemed to be part of the common general knowledge.
Birss J established that in 1996 the task of providing a formulation for a large molecule such as the Herceptin antibody was a challenge and essentially empirical, but this had been successfully undertaken for a number of molecules, such as insulin, human growth hormone and erythropoietin. He felt the skilled formulator would not underestimate the potential risks and would not give up even if the early results were negative. The skilled formulator would deem the exercise well worth carrying out and worth pursuing seriously.
Birss J also noted the ‘very different point’ that the skilled formulator could not say in advance which putative formulation would work. This seems relevant to ‘expectation of success’ but did not factor into the inventive step analysis in the decision.
Birss J found that the general approach to identifying a formulation was established in 1996, that of testing candidate excipients in various conditions. Lyophilisation was also well known, as was the use of a buffer, surfactant and lyoprotectant in a lyophilised formulation. Histidine was known as a buffer and polysorbate 20 as a surfactant. However while the common general knowledge included the ‘idea of using trehalose as a possible lyoprotectant for proteins’ its toxicity when given as part of a therapeutic composition was not known.
Extension of scope by post-grant amendment
Birss J had to decide whether deletion of the feature of the molar ratio of lyoprotectant to antibody being 200-600 moles to 1 mole and insertion of the product by process feature led broadening of claim scope. Genentech argued that the product by process feature led to the ratio being 360:1 of lyoprotectant to Herceptin. However Hospira pointed out that the claims use ‘comprising’ language and therefore other proteins and lyoprotectants could be present in the composition, and that could take the ratio outside 200-600:1. Birss J agreed and found that the amended claims extended scope beyond the granted claims, but also noted that an amendment to replace ‘comprising’ with ‘consisting of’ would overcome the issue.
Product by process claims
In the decision Birss J introduces his discussion of product by process claims by saying they are ‘tricky’. He first discusses product by process claims which are written in the form of ‘obtainable by’ or ‘obtained by’ a defined process, which he describes as ‘overt’ product by process claims. Kirin Amgen  UKHL 46, reported by IPKat here. established that the EPO position had to be followed when determining the validity of such claims. That meant that the ‘obtained by’ form of claim could be anticipated by any product that had the same features as the product obtained by the process, even if it was not made by the process, i.e. for validity the scope of the claim was being interpreted in the same way as for the ‘obtainable by’ form of claim. However Birss J noted that for infringement purposes the ‘obtained by’ feature had to be read as only covering products that had been made by the defined process. Birss J admitted this was a ‘little paradoxical’.
He also noted that apart from ‘overt’ product by process claims (that used the language ‘obtainable by’ or ‘obtained by’) claims could have other product by process features such as ‘lyophilised’ in the present claims where again the same issues arose of the scope of the claims for validity and infringement.
Birss J discussed the circumstances in which the EPO will allow product by process claims. They are only allowed when there is no other way of defining the product (i.e. the product cannot be adequately defined using structural features). This is essentially a trade off between clarity and fairness to the patentee. Turning to the present claims he noted that it was difficult to determine which characteristics need to be present in a product for it to be ‘obtainable by’ the defined lyophilisation process, for example it is not clear what the water content of the product could be. The reader of the claims would be faced with the ‘impossible task of having to create for themselves a list of the relevant attributes’ of the product which is covered. In these circumstances, and in line with the EPO’s approach, Birss J refused to allow the amendment.
The added matter issue arises because the amendments represent a generalisation from Example 1 of the patent. Hospira argued that this represented an impermissible intermediate generalisation, and cited Palmez  RPC 47 . Genetech submitted that the amendment does not disclose ‘new information’ and cited AC Edwards v Acme  RPC 131. Birss J noted that it was not always easy to know which cases fall on the AC Edwards side of the line and which on the Palmaz side, admitting he made that mistake in AP Racing v Alcon  EWPCC 3 (the IPKat’s report of the appeal can be found here). He also noted that the ‘task of applying the law is made more difficult by the fact that the EPO does not approach added matter this way at all’. Turning to the amended claims he noted that they have the effect of only limiting the molar ratio of the components. However in the context of the original disclosure the molar ratios would be tied to particular protein concentrations, and therefore in the absence of a limitation to protein concentration the amended claims present ‘new information’ and therefore add matter.
Carter was the primary document for determining obviousness. It disclosed that Herceptin was in phase II clinical trials for breast cancer. Birss J felt that this would provide the skilled person with motivation to produce a formulation. There was motivation to investigate lyophilisation given that it was known liquid formulations of Herceptin degraded. Hospira’s case was that the required screening of excipients was wholly conventional and the three excipients present in the formulation were obvious candidates to include in the screen. The particular concentrations and/or molar ratios specified in the claims ‘are not suggested to be the product of inventive step’.
Birss J found that the skilled formulator would identify histidine as an appropriate buffer and polysorbate 20 as an appropriate surfactant from common general knowledge. Whether the skilled formulator would choose trehalose as a lyoprotectant was more complex to determine. Genentech argued that its toxicity was unknown and so regulatory approval would be needed for its use. Further it would not be present in a ‘first’ screen, and would only be tested after a ‘second’ screen. If the skilled formulator achieved success before that then trehalose might never be tested. Genentech also argued that the focus on trehalose was due to hindsight.
However Birss J disagreed, viewing trehalose as a ‘very promising lyoprotectant’ and finding that concerns about toxicity would not put the skilled formulator off from testing it. In Birss J’s opinion trehalose would be in a list of eight lyoprotectants to test. That makes it an obvious ingredient to test.
Would/Could, arbitrary selection and the Technograph test
Before concluding on obviousness Birss J went on to deal with other points that arose as part of looking at obviousness.
Genentech argued that the EPO’s approach of asking whether the skilled person would have arrived at the invention, not whether they could have, should be used. However Birss J felt that the law of obviousness could not be accurately summarised in this way, and that the issue was multifactorial and based on the circumstances. The well known 9 ½ inch plate is not something a skilled person would make. In this case it was more appropriate to look at whether the invention could be arrived at by using routine methods applied to excipients that were part of common general knowledge.
Birss J also considered the issue of whether the claims represented a ‘selection’ that could be inventive. He found that it was not an ‘arbitrary’ selection as the invention does give benefits, but that does not necessarily mean it involves inventive step. At the end of their work the skilled team would not be surprised that they had made a stable formulation, and nor would they be surprised by the identity of the ingredients.
Birss J also looked at whether the deriving the invention represented a ‘research programme’ which he interpreted as being an undue burden. However he felt that for this to be an issue it would require ‘a degree of persistence through unpromising results’, but that had not happened here.
Finally the Technograph test was looked at (from Technograph Printed Circuits Ltd v Mills and Rockley (Electronics)Ltd  RPC 346). The test found it wrong to find an invention obvious if it was arrived at after a series of steps which involved the cumulative application of hindsight, for example where knowledge of the target (based on hindsight) motivates the skilled person to take each step. In the present case Birss J found this was not the situation here, in particular because the tests are run in parallel rather than in series.
In view of these findings the claims were obvious from Carter.
The inventive step issues in the present case are similar to another recent case decided by Birss J, Teva UK Limited & Teva Pharmaceuticals Limited v Leo Pharma A/S & Leo Laboratories Limited  EWHC 3096 (Pat), reported by IPKat here. In that case also common general knowledge was central to deciding that the formulation of the invention was obvious. Given that common general knowledge, rather than prior art documents, seems to be most relevant in the field of formulations, it is not surprising that such cases are granted by Patent Offices where examination is document-based.
In his discussion of ‘arbitrary selection’ Birss J notes that the skilled team would not be ‘surprised’ by the formulation they had identified. One wonders if the result would have been different if Genentech could have identified at least one surprising feature of the formulation. In addition one wonders whether a ‘reasonable expectation of success’ argument could have been used in support of inventive step, given Birss J’s acceptance that one could not determine beforehand whether any particular formulation would work. Such an argument would probably not have succeeded but may have led to Birss J providing guidance as to when ‘expectation of success’ is relevant to inventive step.