The Cabilly patents

Abstract

This article discusses the status quo of the Cabilly patents, their scope of protection and the role these patents play for the therapeutic antibody industry in Europe and the US.

Introduction

Much has been written about the history of the Cabilly patents,¹ which cover key steps of therapeutic antibody production. After a long-winded prosecution involving interference proceedings and re-examination, United States (US) Patent 6,331,415, also known as “Cabilly II,” was eventually upheld by a US Patent and Trademark Office (“USPTO”) appeal decision in 2009 despite having been declared dead several times before. Because the patent was granted in 2001, it will expire on December 18, 2018, although it has a priority claim dating back to 1983. Table 1 gives an overview of the Cabilly family of patents.

Table 1.

Overview of the Cabilly family of patents

Alias name	Inventor	Assignee	Patent number	Priority date	Date of Grant	Status	Expiry date
Cabilly I	Shmuel Cabilly et al.	Genentech	US4816567	April 8, 1983	March 28, 1989	expired	March 28, 2006
Cabilly II	Shmuel Cabilly et al.	Genentech	US6331415	April 8, 1983	Dec. 18, 2001	in force	Dec. 18, 2018
Cabilly III	Shmuel Cabilly et al.	Genentech	US7923221	April 8, 1983	April 12, 2011	in force	Dec. 18, 2018

Because Cabilly II protects a crucial step in the state-of-the-art production of therapeutic antibodies, major antibody drugmakers such as Abbott, Johnson and Johnson, ImClone and MedImmune have acquired licenses to it. According to Genentech's Form 10K filed with the US Securities and Exchange Commission (SEC) in 2008, Genentech's 2007 royalties for Cabilly II were $256 million. At the same time, Cabilly II has been the subject of a number of lawsuits between Genentech and its competitors.

On April 12, 2011, the USPTO issued Cabilly III, the youngest member of the Cabilly patent family. Cabilly III is a continuation of Cabilly II, which is in turn a continuation of Cabilly I.

Although Cabilly II and III largely overlap, considerable differences exist between them. It seems that the scope of Cabilly III covers gaps not encompassed by Cabilly II. Further, the Cabilly patents have expired in all countries except the United States. Although the covered methods are thus free to be used outside of the United States, antibody manufacturers must nevertheless assess the legal situation before importing their antibodies into the United States.

This article discusses the status quo of the Cabilly patents, their scope and relevance for antibody companies both in the United States and the rest of the world, especially in Europe.

Patent Lifetime

A patent's lifetime defines the period during which it is in force. A proper determination of a patent's lifetime is thus of paramount importance for the patent's assignee and its competitors alike. European patents expire 20 y after filing [Art. 63 (1) EPC], which may add up to an effective period of 21 y in case a priority has been claimed. Accordingly, Cabilly's European counterpart EP 0125023 expired in 2004, i.e., 21 y after its 1983 priority date.

Similarly, a US patent based on a US non-provisional application filed on or after June 8, 1995, expires 20 y after filing [35 USC §154(a)(2)], and up to 21 y after filing via certain priority claims. In contrast, a US patent based on an application filed before June 8, 1995 expires either 20 years after the earliest claimed non-provisional US priority date or 17 years after issuance, whichever is later.

Genentech has successfully made use of this transitional rule in the United States. Although its earliest priority application dates back to 1983, Cabilly II was issued 18 y after the priority date, while Cabilly III was issued even 28 y after the priority date. In both cases, the US transitional rule applies because the applications issuing into Cabilly II and III were filed before June 8, 1995. Hence, the “effective lifetime” of Cabilly II (i.e., the time between priority date and expiry date) adds up to an extraordinary 35 y. The same applies to Cabilly III, which was issued later but will nevertheless expire when Cabilly II does, due to a so-called terminal disclaimer required by the USPTO before the patent issued.

This leads to an imbalanced situation because in Europe, the methods covered by Cabilly's European counterpart patent have been in the public domain since 2004, i.e., they can be used by European companies without the assignee's consent, while these same methods will remain protected in the United States until 2018.

The Scope of Patents that Protect Methods of Production

The difference in patent lifetime between the United States and Europe creates a number or specific problems, particularly for the European antibody industry. Both legal systems grant claims which are related to methods or processes for the production of matter. In most cases, the products made by a protected method are protected as well [e.g., 35 USC §271(g), Art. 64 (2) EPC] even if no explicit product claims exist in the respective patent.

The methods claimed in the Cabilly patents are veritable methods for the production of matter. Therefore, the patents not only encompass the use of the protected methods per se, but also, under certain circumstances, the use, offer to sell, sale or importation of compounds produced thereby. However, it is the assignee's burden to prove that the antibody in question has in fact been produced with methods protected by the Cabilly patents.

Making antibodies in Europe using the method claimed in the Cabilly patents is legal, because the corresponding EP patent is expired, and the US patents cannot be enforced in Europe due to the territorial effect of patents. Antibody manufacturers may therefore use these methods in Europe, provided they do not violate other third party patents. However, importing the resulting antibodies into the United States may constitute patent infringement under 35 USC §271(g).

There are, however, certain limits to this protection. While importation is barred where an imported antibody is produced by a manufacturing process claimed in the United States, no such bar exists when the product has been “materially changed” by subsequent processes.

That is, in the United States, 35 USC §271(g)(1) provides an exemption from infringement regarding imported products which have either been “materially changed by subsequent processes”; or become “a trivial and nonessential component of another product.” Unfortunately, it is not clear how significant the “material change” must be to qualify for the exemption under 35 USC §271(g)(1). Antibody companies should have this issue analyzed on a case-to-case basis by qualified US counsel. The above exception is, furthermore, restricted by 19 USC §1337, under which a patentee my bring action to the US International Trade commission (ITS). Nineteen USC §1337 defines, among others, unfair acts in the importation of products, and has no exception for “trivial” products or products which have been “materially changed,” in the meaning of 35 USC §271(g)(1). However, according to 19 USC §1337 (a)(2) the patentee must demonstrate that “an industry in the United States, relating to the articles protected by the patent exists” or is being established.

Another exemption from infringement is provided by 35 USC §271(e)(1), according to which the importation of a patented invention “solely for uses reasonably related to the development and submission of information” to the US Food and Drug Administration (FDA) does not qualify as an act of infringement.

Under this provision, shipping an antibody produced by a Cabilly method into the United States may be exempt from patent infringement, if such activity is reasonably expected to lead, ultimately, to an FDA submission. The question of whether or not a particular importation is exempt under 35 USC §271(e)(1) must be assessed on a case-to-case basis. Any company planning to make an antibody using a Cabilly method and import it into the United States for testing purposes should therefore first seek US legal counsel. In any case, however, the exemption under §271(e)(1), if applicable, expires once FDA marketing approval has been obtained.

Cabilly's Scope of Protection

The scopes of Cabilly II and III overlap to a great extent. The two independent claims (i.e., claim 1 of Cabilly II and claim 25 of Cabilly III) characteristic of these patents and having the broadest scope of protection are shown in Table 2.

Table 2.

The two characteristic claims of Cabilly II and III

Cabilly II

Cabilly III

(1) A process for producing an immunoglobulin molecule (Ig) or an immunologically functional immunoglobulin fragment comprising at least the variable domains (Fv) of the immunoglobulin heavy and light chains, in a single host cell, comprising the steps of: (i) transforming said single host cell with a first DNA sequence encoding at least the variable domain of the immunoglobulin heavy chain (HC) and a second DNA sequence encoding at least the variable domain of the immunoglobulin light chain (LC), and (ii) independently expressing said first DNA sequence and said second DNA sequence so that said immunoglobulin heavy and light chains are produced as separate molecules in said transformed single host cell.

(25) A method for making an antibody (mAb) heavy chain (HC) or fragment thereof and an antibody light chain (LC) or fragment thereof each having specificity for a desired antigen, wherein the heavy chain or fragment thereof comprises a variable region (Fv) sequence and a human constant region (Fc) sequence, the method comprising culturing a recombinant host cell comprising DNA encoding the heavy chain or fragment thereof and the light chain or fragment thereof and recovering the heavy chain or fragment thereof and light chain or fragment thereof from the host cell culture.

Abbreviations in italics and emphasis added by author.

Despite the large overlap between the two claims, there are considerable differences. The claims of Cabilly III have been modified significantly relative to Cabilly II, presumably to cover subject matter not encompassed by Cabilly II. Take, for example, the actual transformation step of the host cell. While Cabilly II requires that a single cell is simultaneously transformed with two DNAs encoding the heavy chain and the light chain, respectively, Cabilly III has no such restriction. On the other hand, Cabilly III requires that at least the heavy chain has a human constant region sequence. Table 3 shows the major differences between these two claims of Cabilly II and III.

Table 3.

Major differences between the two characteristic claims of Cabilly II and III, and potential relevance thereof

	Cabilly II	Cabilly III	Remarks
Required compound features
Type of molecule	Immunoglobulin molecule or functional fragment thereof	Antibody or fragment thereof having specificity for a desired antigen	Term “having specificity for a desired antigen” has probably been introduced to Cabilly III to anticipate a claim construction order with respect to the term immunoglobulin, as requested by a plaintiff in an earlier trial.
Variable domains?	Fv of HC	HC or fragment with Fv sequence	Cabilly III adds the term “sequence” to more explicitly include also subsequences of HC and LC.
	Fv of LC	LC or fragment with Fv sequence	Neither Cabilly II nor II protect antibodies which have only HC domains.
Constant domain?	Not required	Human Fc sequence of at least HC	Cabilly III does not protect antibodies devoid of a Fc, or antibodies which have a non-human HC Fc.
Required method features
Type of transformation	A single host cell transformed with DNA for Fv of HC and DNA for Fv of LC	Not specified	Scope of Cabilly II is restricted to methods in which one cell is transformed with two DNAs. Cabilly III has no such restriction.
Type of expression	Both DNAs are independently expressed in said host cell	Not specified (claim 38 says “coexpression”)	Neither “independent expression” nor “coexpression” are defined in the respective patents, although the former implies that both subunits can be expressed independently from one another (e.g., the DNAs are under control of different promotors), while the latter implies a synchronous and possible also quantitative expression of both subunits.
Type of antibody production	HCs and LCs are produced as separate molecules in said host cell	Culturing a recombinant host cell comprising DNA encoding for both chains or fragments and recovering the HC or fragment and LC or fragment

As already mentioned, the two claims shown in Table 2 represent the broadest scope of protection conferred by Cabilly II and III with respect to antibody production methods. The remaining method claims recite at least the features mentioned in the said two claims, plus further features.

As a common rule in patent law, no literal infringement of a patent occurs when a third party method contains fewer features than those recited in the patent's broadest claim. (Aspects of contributory infringement, induced infringement and infringement under the doctrine of equivalents are not dealt with here).

This means, for example, that the literal scope of Cabilly II does not include particular heterokaryon or conjugation methods, or transgenic mammal approaches, which do not require that a single cell be simultaneously transformed with two DNAs encoding the heavy chain, and the light chain, respectively. Likewise, antibodies devoid of a constant region or antibodies which have a non-human heavy chain constant region, are outside the literal scope of Cabilly II.

Tables 4 and 5 show examples of state-of-the-art antibody formats and antibody production methods, and their dependence on Cabilly II and Cabilly III is discussed.

Table 4.

State of the art antibody formats and their dependence on Cabilly II and III

Antibody format		Example	Example configuration	Key IP right	Protected by Cabilly II ?	Protected by Cabilly III ?
Murine Immunoglobulin	IgG	Tositumomab (GSK)		Many	Yes	No, because no human Fc
Human, humanized or chimeric Immunoglobulin	IgG	Many		Many	Yes	Yes, because human Fc
F(ab)2		Many		Many	Yes	Yes
Fab		Many		Many	Yes	Yes
single-chain antibody	scFv	Many		Many	Yes	No, because no Fc
Diabodies	scFv(n)	Tandabs (Affirmed)		US2005089519	Yes	No, because no Fc
Nanobodies (single domain antibodies)	VHH	Ablynx		US2003088074	No, because only HC	No, because only HC
Camelid	VHH-(CH2–3)2	hcIgG		EP0698097	No, because only HC	No, because only HC
Shark	VHH-(CH1–5)2	IgNAR (AdAlta)		US2009148438	No, because only HC	No, because only HC
Primate IgG		Arana (now Cephalon/Teva)		EP 1945668	Yes	No, because New World primate framework regions (= no human Fc)
Receptor-Ig fusion protein	(Receptor-CH2-CH3)2	Etanercept (Pfizer)		US5447851	No, because no Fv	No, because no Fv
Immunotoxin	Different formats	L19-TNFa (Philogen)		EP1257297	Depends on the nature of the Immunoglobulin	Depends on the nature of the Immunoglobulin; L19-TNFa has an scFv, thus does not fall under Cabilly III because of the absence of an Fc

Table 5.

State of the art antibody production methods and their dependence on Cabilly II and III

Production method	Company		Patent	Protected by Cabilly II?	Protected by Cabilly III?
Heterologous Dimeric Protein Production in Heterokaryons	Neugenesis		US5643745	No, because two cell lines are independently transformed with HC DNA and LC DNA, and are then fused	Probably yes
Milk secretion in transgenic mammals	GTC Biotherapeutics		US5827690	No, because transgenic animal comprises a full repertoire of heavy and LC genes, i.e. no transformation of a host cell, and expression takes place in hybridoma cells	No, because no recombinant host cells (cells which have been transformed with vectors used)
Hybridoma cells obtained from transgenic rodents or mice	Regeneron VelocImmune		US7105348
Expression of monoclonal antibodies in ciliate host cells with different mating types	Cilian AG		WO2011107520	No, because two cells which have been trasformed with HC and Lc, respectively, can be conjugated to express full antibody	Probably yes

It is important to mention that neither patent covers the production of antibody mimetics, i.e., non-immunoglobulin proteins having antibody-like target-binding activity. Therapeutic proteins falling under this category include, for example, DARPins (designed ankyrin repeat proteins having target binding properties), and Affilins (target binding molecules derived from human ubiquitin).^2,3 In these molecules, the heavy/light chain language and the constant/variable domain language used in the Cabilly patents does not apply at all.

Lawsuits

The history of the Cabilly patents is also a history of epic patent lawsuits. Points of dispute have included, among others, the right to the patent (interference proceedings between Celltech and Genentech), issues of unfair competition and the amicable relationship between a licensee and a licensor (so-called “licensee estoppel,” MedImmune vs. Genentech),⁴ double patenting and lack of novelty (inter-partes re-examination filed by a strawman), invalidity, unenforceability and non-infringement by a third party antibody (Centocor Inc., vs. Genentech,⁵ Glaxo Group Ltd., vs. Genentech, Inc.⁶), and tortious interference and civil conspiracy (two cases, Human Genome Sciences vs. Genentech).⁷ The last two cases were filed in the Delaware District Court.

Not surprisingly, Human Genome Sciences (HGS) filed a new case at the Delaware District Court on the day Cabilly III was issued. HGS is asking for a declaratory judgment of invalidity, unenforceability and non-infringement, and seeks freedom to operate for belimumab (Benlysta^®), an antibody which targets the B-lymphocyte stimulator.⁸ On the same day, Genentech sued HGS and GlaxoSmithKline (GSK) before the Central District Court of California for infringing Cabilly III by selling belimumab and ofatumumab (Arzerra^®), which is GSK's anti-CD20 antibody.⁹ In the meantime, the Delaware court has granted Genentech's motion to transfer the three pending cases filed by HGS to the Central District Court of California.

The Cabilly patents have thus constituted and still constitute, a bottleneck with respect to antibody expression technologies. US companies using these technologies must either go through this bottleneck or face legal action, either as a claimant (e.g., in an invalidity action or a non-infringement action) or as a defendant (e.g., in an infringement action).

Conclusion

The Cabilly family of patents has been looked upon in the literature as controversial, particularly because the patents protect key steps of therapeutic antibody production, and their lifetimes are extraordinarily long.

The issuance of Cabilly III has further raised the potential for the Cabilly family of patents to cause conflict. Despite the large overlap between Cabilly II and Cabilly III, the scope of Cabilly III covers blind spots which are not encompassed by Cabilly II. This applies, for example, to the actual transformation process of the host cell. While Cabilly II requires that a single cell is simultaneously transformed with two DNAs encoding the heavy chain and the light chain, respectively, Cabilly III has no such restrictions.

While both Cabilly patents have a direct blocking effect on antibody production in the United States, their effect on antibody production in Europe is only indirect. That is, the importation into the United States of antibodies produced in Europe or elsewhere outside the United States by a method according to either Cabilly II or Cabilly III will likely be considered infringing under certain circumstances.

Companies outside of the United States that use antibody production methods and are contemplating importation into the United States sshould first determine whether their methods or compounds fall under the scope of the Cabilly family of patents. If so, such companies must prepare accordingly.

If importation is likely to occur before the expiry date, companies should first seek counsel by a qualified US patent attorney to determine whether their planned activities would be exempt from infringement under 35 USC §271(g)(1) or 35 USC §271(e)(1).

If such activity would not be exempt, companies should determine if the patents are valid and enforceable and, if so, prepare a license request, or seek ways to design around. Further, it is highly advisable to establish alternative routes of production which avoid the scope of the Cabilly family of patents in case a license is denied or commercially unreasonable, and because the burden of proving infringement is on the assignee.

Disclosure of Potential Conflicts of Interest

The information and opinions provided herein reflect the personal views and considerations of the author. They do not represent legal counsel and should not be attributed to Michalski-Hüttermann and Partner Patent Attorneys or to any of its clients. Patent numbers and patent lifetimes have been verified with the utmost care, but no liability is assumed for their correctness.

Note Added in Proof

Genentech is not the only company that has made use of said transitional rule. On November 23, 2011, US company Amgen has announced that a new patent had been granted which protects their anti-TNFα fusion protein etanercept (Enbrel^®).

The patent, US 8063182, which is based on a divisional application filed May 19, 1995, claims the priority of an earlier application filed September 10, 1990. The patent was granted November 23, 2011 and provides thus another 17 years of protection for Enbrel. In contrast to the usual 20 year patent lifetime, Amgen has thus successfully managed to extend the patent term in the US to 38 years.

Because US patent applications filed before November 29, 2000 are not published and remain secret until granted, the new patent came quite surprisingly to the Biologic industry. The basic US patent protecting Enbrel (US5610279), which claims the same priority, was set to expire October 23, 2012. Companies who had biosimilar versions of Enbrel in their pipeline in view of the assumedly approaching patent expiry will have to rethink their market strategy at least for the United States.