The following article was reprinted with permission from J A Kemp.

The European Patent Office (EPO) continues to grant many patents relating to antibodies, and in doing so applies the same patentability criteria as to other inventions.  However, some commentators have suggested that antibodies are regarded as a special case by the EPO when evaluating inventive step / obviousness.  It is implied that the EPO sets a higher patentability threshold for inventions in this field as compared to, for example, small molecule therapeutics.

This perception likely reflects no more than the crowded nature of the antibody field, which has matured rapidly in recent years.  As a result, EPO Examiners are able to make a large (and increasing) number of assumptions regarding the background knowledge of the skilled person.  For example, in the absence of evidence to the contrary, the EPO considers it to be routine to raise an antibody to a known target.  The EPO also considers it to be routine to optimise certain characteristics of an antibody, such as affinity or immunogenicity.

Although it is not unique to the field, the approach of the EPO can present significant challenges to applicants seeking to pursue claims to antibodies, particularly antibodies to known targets.  We have set out below our suggestions for how grant of antibody claims at the EPO might nonetheless be achieved in different factual scenarios.  We have also set out the types of supporting evidence that may be required in each case.  Figure 1 provided at the end of this briefing summarises the suggestions in the form of a decision tree, which may help to guide you to the most appropriate strategy at different stages in an antibody development project.

Scenarios likely to Result in Broad claims

Broad antibody claims typically arise from two basic scenarios.  In the first scenario, the target is a newly-identified molecule, or is a molecule to which it has previously proven difficult to raise an antibody.  In such cases, the antibody may be defined solely in terms of the target bound.  For example:

"An antibody which specifically binds to "

In the second scenario, the target is found to have a previously unappreciated role in a disease.  That is, the target is found to be associated with a new medical use.  In such cases, the antibody may be defined by reference to the target bound and the new medical use.  For example:

"An antibody which specifically binds to for use in a method of treating ".

These two basic scenarios are discussed in more detail below.

Newly-Identified or Difficult Target

The extensive mapping of genomes and proteomes means that truly "new" targets are increasingly rare.  However, an applicant who identifies, for example, a new polypeptide (potentially including new mutant forms of a known polypeptide) can expect to obtain broad claims to any antibody that specifically binds to the new polypeptide.  It is not typically necessary to provide evidence that an antibody has actually been produced if the target is susceptible to routine methods of antibody production.

An applicant who develops a method allowing an antibody to be raised for a known target which was not susceptible to routine methods can also expect to obtain broad claims.  The EPO effectively views such a target as a special category of new target -- it is "newly available" for antibody production.  Inventions of this type are though increasingly rare.  Methods of antibody production have developed such that few targets are not now susceptible to them, and overcoming technical difficulties is frequently seen as routine.  If the EPO can be persuaded otherwise for a particular case, the patent application will need to include clear evidence that an antibody to the target has indeed been raised, as well as sufficient information to allow this to be repeated.

For antibodies to such newly-identified targets or difficult targets, medical use claims may be allowable even without direct evidence of a clinically-relevant functional effect.

New Medical Use

Where a known target molecule is found to have a previously unappreciated role in a disease, it may be possible to obtain claims directed to any antibody that specifically binds to the target for use in a particular method of treatment.  In a typical scenario of this type, the target has a known association with one disease, but is subsequently found to have a role in a different (new) disease that was not previously appreciated.  Under such circumstances the claims will be limited to an antibody for use in a method of treating the new disease, but will cover any anti-target antibody for the said use.

It will be necessary to establish that the new medical use is not obvious from any previous disclosure regarding the target or existing antibodies that bind to it.  It will also be necessary to establish that it is at least plausible that an antibody binding to the target will have a therapeutic effect.  At present this is a relatively low threshold at the EPO and in vitro functional data will likely be sufficient, particularly if it derives from a disease model or other assay relevant to the indication.  In vivo data from animal models is helpful but not required.  It is not normally expected that a patent application will include in vivo data from human clinical trials.

Similar considerations may apply if it was previously thought that the only antibodies that could be produced to a target were pharmaceutically irrelevant (for example due to low affinity), but it is subsequently found that they have a therapeutic effect¹.  In such cases a claim to a pharmaceutical composition comprising any antibody specific for the target may be possible.

Scenarios which may Result in Narrower Claims

It is unlikely that broad antibody claims will be available if both the target and associated medical uses are known.  However, grant of narrower claims may still be possible.  Such claims will likely need to incorporate limitations from either or both of the following categories.

New or Improved Characteristic of an Antibody

Limitations to specific functional characteristics of the antibody, such as a required level of affinity or a down-stream functional effect, likely also with at least some structural (sequence) characteristics.

New Development of a Known Medical Use

Limitations to specific aspects of the medical use, such as a particular patient group or a dosing/administration regime.

Both categories are discussed in more detail in the following section.  The categories are not mutually exclusive, and claims including both should be pursued wherever possible as part of a comprehensive filing strategy (see additional comments on filing strategy in the separate section below).  However, in our experience, at any given stage in development a particular antibody project is likely to favour limitations from one category over the other.

New or Improved Characteristic of an Antibody

Examples of the types of functional characteristic that applicants may seek to rely upon are described in more detail below.  These examples should not be viewed as exhaustive or mutually exclusive.  In practice a combination of these and other characteristics may well provide the best route to allowable claims.

Irrespective of the characteristic(s) relied upon, pursuit of this approach will likely require that applicants have provided a significant amount of structural information (primary amino acid sequences) in the patent application.  Whether a particular antibody is claimed directly or is recited as a reference antibody, the EPO often insists that structural detail be provided at a level at least sufficient to define the target binding region of an antibody.  It may alternatively be possible to meet these requirements by reference to a biological deposit of, for example, a hybridoma, provided the deposit is made in accordance with the Budapest Treaty.

Historically, the EPO has accepted structural definitions of antibodies which refer only to one or two CDR sequences.  However, EPO Examiners are likely now to insist upon all six CDRs as a minimum requirement, unless data is available to show that any characteristic relied upon is achieved with fewer CDRs.  There is though an increasing appreciation of the relevance of framework regions to target binding, and so there is a trend towards requiring that the heavy and light chain variable region sequences be recited in full.  In some cases it may even be necessary to specify the isotype of the constant region if this is relevant to the characteristic(s) relied upon.

By contrast to its approach for small molecule therapeutics, the EPO does not in general consider that a unique structure can confer inventive step on an antibody to a known target.  Thus, whilst an antibody comprising a unique primary sequence will be considered to be novel, structural non-obviousness arguments are unlikely to be accepted even when both variable region sequences are recited in full.

This means that applicants will generally be unable to rely upon, for example, the determination of unique CDR and framework sequences to provide an inventive step.

Instead, the EPO has developed a requirement that a new antibody to a known target must demonstrate "an unexpected effect" relative to pre-existing antibodies to the same target² in order for inventive step to be acknowledged.  It will therefore generally be necessary to demonstrate that the antibody as claimed possesses a functional characteristic (or combination of characteristics) which could not reasonably have been predicted from the prior art.

Ideally, at least some direct experimental evidence of the characteristic(s) relied upon will be provided in the patent application, for at least one exemplary antibody.

Under some circumstances it may be possible to rely upon supplementary data filed later, but there must be a connection to the characteristic(s) shown in the application.  The characteristic(s) relied upon must be at least plausible at the filing date³.

This test has been applied very strictly by the EPO on at least one occasion.  In a specific case⁴ the Board of Appeal held that a particular effect relied upon by an applicant to support inventive step could not be derived from the patent application itself, even though an arguably related effect was explicitly demonstrated in the Examples.  Accordingly the effect relied upon was considered to have been demonstrated only after the filing date of the application, and was not taken into account when assessing inventive step.

If the EPO is persuaded that "an unexpected effect" is present and plausible, the breadth of claim that is allowed will depend upon the circumstances of each case -- in particular the state of the prior art and the data available.

At the narrowest end of the spectrum, it may be necessary to limit the claims to the specific sequences of an antibody which has been shown to possess the unexpected effect.  For example:

"An antibody which binds specifically to and which comprises a VH region comprising SEQ ID NO: x and a VL region comprising SEQ ID NO: y."

At the opposite end of the spectrum, and more rarely, it may be possible to claim a general class of antibodies which achieves the said effect, defined solely in functional terms. For example:

"An antibody which binds specifically to and which has ."

It will be necessary for the application to include sufficient information such that the skilled person can generate further examples of an antibody which meets the functional definition.  The identification in the patent application of a specific, structurally-defined antibody which can be used as a reference / control in a screening method is usually acceptable.  However, it is not unusual for an Examiner to require that the claim also include some structural information about the claimed antibody or the reference antibody.  As such, allowable claims may be a hybrid of the two types shown above.  For example:

"An antibody which binds specifically to and which has , wherein the antibody competes for binding to with an antibody which comprises a VH region comprising SEQ ID NO: x and a VL region comprising SEQ ID NO: y."

Examples of Possible New Characteristics

As mentioned above, the following examples should not be viewed as exhaustive or mutually exclusive.  In practice a combination of these and other characteristics may well provide the best route to allowable claims.  The goal is to establish that there is a technical effect or characteristic (or multiple such effects or characteristics) which can be shown to have been unexpected based on the relevant prior art.

Affinity

The EPO considers the generation of an antibody with nanomolar affinity to be achievable by routine methods for the vast majority of targets.  Whilst there may be exceptions for particularly difficult targets, high affinity alone is now unlikely to be sufficient to establish an inventive step.  However, high affinity in combination with another advantageous characteristic is nonetheless likely to be helpful.

Specificity

The EPO tends to view high specificity for target to be an inherent characteristic of all antibodies and so an increase in specificity alone is unlikely to be sufficient to establish an inventive step.  However, an increase in specificity which gives rise to an unexpected technical advantage, or which was previously thought to be difficult to achieve, may be enough.  For example if a new antibody is able to distinguish between closely-related targets to a greater extent than was previously possible.

Epitope

An antibody which can bind to an epitope in a target that was previously unrecognised and/or previously considered to be inaccessible is likely to be viewed as inventive.  Under certain circumstances this may even be seen as a sub-category of the "Newly-identified or difficult target" scenario described above.  If binding to the particular epitope also provides a functional advantage (e.g. enhanced specificity) this is likely to be helpful.

Manufacturability

Routine optimisation for improved solubility etc is unlikely to be viewed as inventive.  However, there may be exceptions if the nature or position of any substitutions, and the subsequent effect, can be shown to have been unexpected.  For example, if a surprising increase in affinity or specificity occurs.

Immunogenicity

Humanisation of antibodies to reduce anti-isotypic or idiotypic responses is now largely viewed as routine, as is the production of fully human antibodies.  However, as with manufacturability there may be exceptions if the nature or position of any substitutions, and the subsequent effect, can be shown to have been unexpected.

The Board of Appeal has held in one case⁵ that particular point mutations to introduce human sequences in the framework regions of a specific mouse antibody were sufficient to acknowledge inventive step for that antibody.  The evidence in that case persuaded the Board that the skilled person could not reasonably have expected that those specific mutations would reduce immunogenicity without significantly reducing affinity.

Down-Stream Functional Effects

Any improvement in down-stream function is likely to be helpful when seeking to establish an inventive step.  However, adaptations of an antibody to enhance down-stream functions such as ADCC are unlikely to be viewed as inventive where they rely upon known phenomena such as altered glycosylation of Fc regions.

New Development of a Known Medical Use

It seems it is becoming ever more challenging to persuade the EPO that a new or improved characteristic of an antibody is enough alone to establish inventive step.  However, in our experience antibody inventions are treated no differently to other therapeutics when assessing further developments of a known medical use, in which a new and unexpected technical effect can be demonstrated.

The scope of claim that may be available via this route will depend upon the data produced, since it will be necessary to show that the effect relied upon is achieved across substantially the whole scope claimed.  This may require limitation to an antibody defined by a complete structural definition (including constant regions), if the EPO do not accept that the effect relied upon applies more generally, for example to all antibodies possessing the same six CDRs.

Ideally the data showing the effect should be present in the patent application.  However, data generated post-filing may be used to supplement arguments later in prosecution, provided the effect relied upon is plausible at the filing date⁶.

Bearing this in mind, we consider that inventions of this type are generally in one of two categories:

Combinations

(Bispecific Molecules/Combination or Co-Administration with other Therapeutics/Conjugates)

Any form of therapy which combines an antibody with another functionality may give rise to allowable claims.  Examples include a combination with another binding functionality in the same molecule (a bispecific molecule), a combination with another effector functionality in the same molecule (a drug/toxin conjugate), or the co-administration of the antibody with another therapeutic (combination or co-administration), or any combination of these types.  In each case, as with any combination of known therapeutics, it will be necessary to establish that the combination results in a technical effect (typically an advantage) which was not obvious from the prior art (e.g. synergistic increase in efficacy).  It will likely be necessary to include experimental evidence of the technical effect in the patent application.

Treatment Methodologies

(Patient Group/Dose or Formulation/Administration Route or Regime)

The Enlarged Board of Appeal of the EPO has established⁷ that claims to second/further medical uses of a substance may be based not only on the treatment of a different disease, but also on the treatment of the same disease by a method which differs for example in the dosage, formulation, administration regime, group of subjects or route of administration.  In each case, as with any known therapeutic, it will be necessary to establish that the difference gives rise to a technical effect (typically an advantage) which was not obvious from the prior art (e.g. surprisingly enhanced efficacy for topical versus parenteral administration).  It will likely be necessary to include experimental evidence of the technical effect in the patent application.

New Types of Antibody

An invention which relates to a new type of antibody may give rise to broad claims which are not limited to a specific antibody or target.  Examples of possible inventions in this category could include a novel and inventive arrangement of the binding domains (a new antibody "format"), a novel and inventive effector region (such as a modified Fc domain), or an entirely different class of antibody molecule isolated from a newly-discovered species.

For inventions of this type, the assumptions that the EPO would otherwise make regarding production and optimisation of known antibody types should largely be negated.  The EPO should treat such inventions in the same way as any other technology, since normal considerations of novelty, inventive step and sufficiency of disclosure should apply.  As such the assessment of each case will depend upon its merits.

Comprehensive Filing Strategy

A typical antibody development project will pass through different stages, and at each stage it may correspond more closely to one or the other of the scenarios outlined above.  This briefing is intended to provide some suggestions as to how patentable claims may be achieved in each scenario.  These suggestions are also summarised in Figure 1.

However, the scenarios that are described are not mutually exclusive.  As with any therapeutic invention, a comprehensive filing strategy should seek to explore all of the available options for patent protection throughout the lifespan of the project and beyond.  For example, an initial broad filing to a new medical indication should be followed with multiple narrower filings to antibodies with improved characteristics and/or developments of the treatment methodology.

Figure 1

¹ Board of Appeal decision T601/05: claims to a pharmaceutical composition comprising human anti-TNFα antibodies were found to be inventive. At the time only low affinity human antibodies to the target could be generated, and it was thought that these were not suitable for pharmaceutical use.
² Board of Appeal decision T735/00
³ Board of Appeal decision T1329/04
⁴ Board of Appeal decision T2637/11
⁵ Board of Appeal decision T0067/11
⁶ Board of Appeal decision T1329/04
⁷ Enlarged Board of Appeal decision G2/08

This article was reprinted with permission from J A Kemp.