Back to GetFilings.com




QuickLinks -- Click here to rapidly navigate through this document

UNITED STATES
SECURITIES AND EXCHANGE COMMISSION
WASHINGTON, D.C. 20549

FORM 10-K

(Mark one)  

ý
ANNUAL REPORT PURSUANT TO SECTION 13 OR 15(d) OF THE SECURITIES EXCHANGE ACT OF 1934

For the fiscal year ended December 31, 2002

OR

o
TRANSITION REPORT PURSUANT TO SECTION 13 OR 15(d) OF THE SECURITIES EXCHANGE ACT OF 1934

For the transition period from                            to                             

Commission file number: 000-24207

ABGENIX, INC.
(Exact name of registrant as specified in its charter)

Delaware
(State or other jurisdiction of
incorporation or organization)		   94-3248826
(IRS employer Identification number)

6701 Kaiser Drive, Fremont, CA
(Address of principal executive office)
 
94555
(Zip Code)

(510) 608-6500
(Registrant's telephone number, including area code)

Securities registered pursuant to Section 12(b) if the Act: None
Securities registered pursuant to Section 12(g) of the act: Common Stock, $0.0001 par value;

Preferred Stock Purchase Rights
(Title of Class)

        Indicate by check mark whether the registrant (1) has filed all reports required to be filed by Section 13 or 15(d) of the Securities Exchange Act of 1934 during the preceding in 12 months (or for such shorter period that the registrant was required to file such reports), and (2) has been subject to such filing requirements for the past 90 days. Yes ý    No o

        Indicate by check mark if disclosure of delinquent filers pursuant to Item 405 of Regulation S-K is not contained herein, and will not be contained, to the best of registrant's knowledge, in definitive proxy or information statements incorporated by reference in Part III of this Form 10-K or any amendment to this Form 10-K. ý

        Indicate by check mark whether the registrant is an accelerated filer (as defined in Rule 12b-2 of the Exchange Act. Yes ý    No o

        The aggregate market value of the voting stock held by non-affiliates of the Registrant as of June 28, 2002 was $759,314,800. The number of shares of Common Stock, $0.0001 par value, outstanding on February 28, 2003, was 87,747,845.

        Documents incorporated by reference: Portions of the Proxy Statement for Registrant's Annual Meeting of Shareholders to be held June 27, 2003 (the Proxy Statement), are incorporated herein by reference into Part III.


TABLE OF CONTENTS

 
    
   Pages
Item 1.   Business   3
Item 2.   Properties   48
Item 3.   Legal Proceedings   48
Item 4.   Submission of Matters to a Vote of Security Holders   48
Item 5.   Market for the Registrant's Common Equity and Related Stockholder Matters   48
Item 6.   Selected Financial Data   49
Item 7.   Management's Discussion and Analysis of Financial Condition and Results of Operations   50
Item 7A.   Quantitative and Qualitative Disclosures about Market Risk   65
Item 8.   Financial Statements and Supplementary Data   67
Item 9.   Change in and Disagreements with Accountants on Accounting and Financial Disclosure   91
Item 10.   Directors and Executive Officers of the Registrant   91
Item 11.   Executive Compensation   91
Item 12.   Security Ownership of Certain Beneficial Owners and Management   91
Item 13.   Certain Relationships and Related Transactions   91
Item 14.   Controls and Procedures   91
Item 15.   Exhibits, Financial Statement Schedules, and Reports on Form 8-K   92
    Signatures   99
    Certifications   101

2


PART I

Item 1. Business.

        The following description of our business should be read in conjunction with the information included elsewhere in this Annual Report on Form 10-K. The description contains certain forward-looking statements that involve risks and uncertainties. When used in this Annual Report on Form 10-K, the words "intend," "anticipate," "believe," "estimate," "plan" and "expect" and similar expressions as they relate to us are included to identify forward-looking statements. Our actual results could differ materially from the results discussed in the forward-looking statements as a result of certain of the risk factors set forth below and in the documents incorporated herein by reference, and those factors described under "Additional Factors that Might Affect Future Results". In this Annual Report on Form 10-K, references to "Abgenix," "we," "us" and "our" are to Abgenix, Inc. and its subsidiaries.

Abgenix

        We are a biopharmaceutical company that is focused on the discovery, development and manufacture of human therapeutic antibodies for the treatment of a variety of disease conditions, including cancer, inflammation, metabolic disease, transplant-related diseases, cardiovascular disease and infectious diseases.

        We have proprietary technologies that facilitate rapid generation of highly specific, antibody therapeutic product candidates that contain fully human protein sequences and that bind to disease targets appropriate for antibody therapy. In this Annual Report on Form 10-K we refer to these candidates as fully human antibody therapeutic product candidates. We developed our XenoMouse® technology, a technology using genetically modified mice to generate fully human antibodies. We also own a technology that enables the rapid identification of antibodies with desired function and characteristics, referred to as SLAM™ technology. In our XenoMax™ technology, we use SLAM technology to select and isolate antibodies with particular function and characteristics from antibody-producing cells generated by XenoMouse animals. We believe XenoMax technology enhances our capabilities in product development and flexibility in manufacturing. We intend to use our technologies to build a large and diversified product portfolio that we expect to develop and commercialize through joint development and licensing arrangements with pharmaceutical companies and others, and through internal product development programs. We have entered into a variety of contractual arrangements with multiple pharmaceutical, biotechnology and genomics companies involving our XenoMouse and XenoMax technologies. Two of our customers, Pfizer, Inc. and Amgen, Inc., have initiated clinical trials with fully human antibodies generated from XenoMouse animals. In addition, under a joint development and commercialization agreement we are co-developing ABX-EGF, our leading proprietary antibody product candidate, with Amgen and Immunex Corporation, a wholly-owned subsidiary of Amgen.

Overview of Product Development

Preclinical Research

        Our product development activities begin with preclinical research and development. Our preclinical research and development efforts have been focused on:

3

        We identify antigen targets largely through licensing or collaboration agreements with other companies that have ownership interests or intellectual property rights in antigen targets that are of interest to us, or have particular methods of identifying potential antigen targets. We also conduct our own preclinical antigen target validation research. We generate and screen antibodies through use of our XenoMouse and XenoMax technologies. After we have identified antibodies of interest, we conduct in vitro experiments and in vivo experiments using animal models to provide further data about the potential therapeutic value of the antibodies for treatment of a variety of diseases or indications. Our preclinical activities also include improvement of production methods and support of collaborations.

Proprietary Product Development and Customer Product Development

        Our leading proprietary antibody therapeutic product candidate is ABX-EGF. Generated using XenoMouse technology, ABX-EGF is a fully human antibody therapeutic product candidate for the treatment of a variety of cancers. We are co-developing this candidate with Amgen and Immunex under a joint development and commercialization agreement. The status of clinical trials for ABX-EGF is as follows:

        "Phase 1" indicates safety and proof of concept testing in a limited patient population and toxicology testing in animal models. "Phase 2" indicates safety, dosing and efficacy testing in a limited patient population. "Phase 3" indicates safety and efficacy testing with a larger patient population.

        We agreed to jointly develop and commercialize ABX-EGF. We intend to enter into additional joint development agreements for other product candidates earlier in the product development lifecycle than when we entered into previous joint development agreements. We will expend significant capital to conduct clinical trials or share in the costs of conducting clinical trials, for our proprietary product candidates. We expect that this will substantially increase our operating expenses over the next few years in comparison to prior periods.

        In addition to ABX-EGF, we have three other proprietary antibody therapeutic products that are in clinical trials. These products are further described under the caption Proprietary Product Development Programs.

4


        In addition to our proprietary antibody therapeutic product candidates in clinical trials, there are two customer-developed antibodies generated with XenoMouse technology in clinical trials as follows:

Overview of Production Services

        Our antibody production services, also referred to as production services, include closely integrated process sciences and manufacturing capabilities for the manufacture of therapeutic product candidates. Within our pilot plant, our process sciences services include cell line development, optimization and production scale up. The resulting process can be transitioned to our manufacturing facility, portions of which are now operational. This facility is designed to manufacture product candidates for clinical trials and to support the early commercial launch of a limited number of products in compliance with applicable FDA good manufacturing practices. With both process sciences and manufacturing capabilities, we offer integrated antibody production services.

Overview of Technology Licensing

        We license our XenoMouse technology to pharmaceutical and biotechnology companies interested in developing antibody-based products. In some cases, we provide our mice to the customer who then carries out immunizations with its specific antigens. In other cases, we immunize the mice with the customer's antigens for additional compensation. Using our XenoMax technology, we can provide our customers a larger pool of high-quality antibodies from which to choose optimal therapeutic product candidates. The customer generally has an option for a period of time to acquire a product license for any antibody identified using our XenoMouse or XenoMax technology that the customer wishes to develop and commercialize.

Background

The Normal Antibody Response

        The human immune system protects the body against a variety of infections and other illnesses. Specialized cells, which include B-cells and T-cells, work in concert with the other components of the immune system to recognize, neutralize and eliminate from the body numerous foreign substances, infectious organisms and malignant cells. In particular, B-cells generally produce protein molecules, known as antibodies, which are capable of recognizing substances potentially harmful to the human body. Such substances are called antigens. Upon being bound by an antibody, antigens can be neutralized or blocked from interacting with and causing damage to the body. In order to effectively neutralize or eliminate an antigen without harming normal cells, the immune system must be able to generate antibodies that bind tightly (i.e., with high affinity) to one specific antigen (i.e., with specificity).

        All antibodies have a common core structure composed of four subunits, two identical light (L) chains and two identical heavy (H) chains, named according to their relative size. The heavy and light chains are assembled within the B-cell to form an antibody molecule that consists of a constant region and a variable region. As shown in the diagram below, one can represent an antibody molecule schematically in the form of a "Y" structure.

5

GRAPHIC

        The base of the "Y," together with the part of each arm immediately next to the base, is called the constant region because its structure tends to be very similar across all antibodies. In contrast, the variable regions are at the end of the two arms and are unique to each antibody with respect to their three-dimensional structures and protein sequences. Because variable regions define the specific binding sites for a variety of antigens, there is a need for significant structural diversity in this portion of the antibody molecule. The immune system achieves such diversity primarily through a unique mode of assembly involving a complex series of recombination steps for various gene segments of the variable region, including the V, D and J segments (see the diagram below).

GRAPHIC

6

        The human body is repeatedly exposed to a variety of different antigens. Accordingly, the immune system must be able to generate a diverse repertoire of antibodies that are capable of recognizing these multiple antigen structures with a high degree of specificity. The immune system has evolved a two-step mechanism in order to accomplish this objective. The first step, immune surveillance, is achieved through the generation of diverse circulating B-cells, each of which assembles different antibody gene segments in a semi-random fashion to produce and display on its surface a specific antibody. As a result, the body generates a large number of distinct, albeit lower affinity, circulating antibodies so as to recognize essentially any foreign antigen that enters the body. While capable of recognizing the antigens as foreign, these lower affinity antibodies are generally incapable of effectively neutralizing them.

        This limitation of the immune surveillance process is generally overcome by the normal immune system in a second step called "affinity maturation." Triggered by the initial binding to a specific antigen, the immune system then primes the small fraction of B-cells that recognize this antigen to progressively generate antibodies with higher and higher affinity through a process of repeated mutation and selection. As a result, the reactive antibodies develop increasingly higher specificity and affinity with the latter being potentially a hundred to a thousand times higher than those generated in the immune surveillance process. These more specific, higher affinity antibodies have a greater likelihood of effectively neutralizing or eliminating the antigen while minimizing the potential of damaging healthy cells.

Antibodies as Products

        Recent advances in the technologies for creating and producing antibody products, coupled with a better understanding of how antibodies and the immune system function in key disease states, have led to renewed interest in the commercial development of antibodies as therapeutic products. According to a survey by the Pharmaceutical Research and Manufacturers of America, antibodies accounted for over 20% of all biopharmaceutical products in clinical development in February 2000. We are currently aware of twelve antibody therapeutic products approved for marketing in the United States. These products are Orthoclone, ReoPro, Rituxan, Zenapax, Herceptin, Synagis, Remicade, Simulect, Mylotarg, Campath, Zevalin and Humira. These products are currently being marketed for a wide range of medical disorders such as transplant rejection, cardiovascular disease, cancer and infectious diseases.

        We believe that, as products, antibodies have several potential clinical and commercial advantages over traditional therapies. These advantages include the following:

Limitations of Current Approaches to Development of Antibody Therapeutic Products

        Despite the early recognition of antibodies as promising therapeutic agents, a number of commercial and technical limitations have thus far hampered most approaches to developing antibodies as products. Researchers aimed their initial efforts at the development of hybridoma cells from mice. Such hybridoma cells are immortalized mouse antibody-secreting B-cells. Researchers derive these hybridoma cells from normal mouse B-cells that have been fused with a perpetually-growing cell so that they are capable of reproducing over an indefinite period of time. They are then cloned to produce a

7


homogeneous population of identical cells that produce antibodies called monoclonal antibodies that are identical in their structure and functional characteristics.

        While mouse monoclonal antibodies can be generated to bind to a number of antigens, they contain mouse protein sequences and tend to be recognized as foreign by the human immune system. As a result, the human body quickly eliminates them and they have to be administered frequently. When patients are repeatedly treated with mouse antibodies, they will begin to produce antibodies that effectively neutralize the mouse antibody, a reaction referred to as a Human Anti-Mouse Antibody, or HAMA, response. In many cases, the HAMA response prevents the mouse antibodies from having the desired therapeutic effect and may cause the patient to have an allergic reaction. The potential use of mouse antibodies is thus best suited to situations where the patient's immune system is compromised or where only short-term therapy is required. In such settings, the patient is often incapable of producing antibodies that neutralize the mouse antibodies or has insufficient time to do so.

        Recognizing the limitations of mouse monoclonal antibodies, researchers have developed a number of approaches to make them appear more human-like to a patient's immune system. For example, improved forms of mouse antibodies, referred to as "chimeric" and "humanized" antibodies, are genetically engineered and assembled from portions of mouse and human antibody gene fragments. While these chimeric and humanized antibodies are more human-like, they still retain a varying amount of the mouse antibody protein sequence, and accordingly may continue to trigger the HAMA response.

        Additionally, the humanization process can be expensive and time consuming, requiring at least two months and sometimes over a year of secondary manipulation after the initial generation of the mouse antibody. Once the humanization process is complete, the remodeled antibody gene must then be expressed in a recombinant cell line appropriate for antibody manufacturing, adding additional time before the production of preclinical and clinical material can be initiated. In addition, the combination of mouse and human antibody gene fragments can result in a final antibody product that is sufficiently different in structure from the original mouse antibody that a decrease in specificity or a loss of affinity results.

LOGO

8

Human Antibodies

        The HAMA response can potentially be avoided through the generation of antibody therapeutic products with fully human protein sequences. Such fully human antibodies may increase the market acceptance and expand the use of antibody therapeutics. Researchers have developed several antibody technologies to produce antibodies with 100% human protein sequences (see the diagram above). One approach to generating human antibodies, called "phage display" technology, involves the cloning of human antibody genes into bacteriophages, viruses that infect bacteria, in order to display antibody fragments on the surfaces of bacteriophage particles. This approach attempts to mimic in vitro the immune surveillance and affinity maturation processes that occur in the body. Because phage display technology cannot take advantage of the naturally occurring in vivo affinity maturation process, the antibody fragments initially isolated by this approach are typically of moderate affinity. In addition, further genetic engineering is required to convert the antibody fragments into fully assembled antibodies and significant manipulation, taking from several months to a year, may be required to increase their affinities to a level appropriate for human therapy. Before pre-clinical or clinical material can be produced, the gene encoding the antibody derived from phage display technology must, as with a humanized antibody, be introduced into a recombinant cell line.

        Two additional approaches involving the isolation of human immune cells have been developed to generate human antibodies. One such approach is the utilization of immunodeficient mice that lack both B- and T-cells. Researchers transplant human B-cells and other immune tissue into these mice which are then subsequently immunized with target antigens to stimulate the production of human antibodies. However, this process is generally limited to generating antibodies only to nonhuman antigens or antigens to which the human B-cell donor had previously responded. Accordingly, this approach may not be suitable for targeting many key diseases such as cancer, and inflammatory and autoimmune disorders for which appropriate therapy might require antibodies to human antigens. The other approach involves collecting human B cells that have been producing desired antibodies from patients exposed to a specific virus or pathogen. As with the previous approach, this process may not be suitable for targeting diseases where antibodies to human antigens are required, and therefore is generally limited to infectious disease targets which will be recognized as foreign by the human immune system.

The Abgenix Solution—XenoMouse and XenoMax Technologies

        Our approach to generating human antibodies with fully human protein sequences is to use genetically engineered strains of mice in which mouse antibody gene expression is suppressed and functionally replaced with human antibody gene expression, while leaving intact the rest of the mouse immune system. Rather than engineering each antibody product candidate, these transgenic mice capitalize on the natural power of the mouse immune system in surveillance and affinity maturation to produce a broad repertoire of high affinity antibodies. By introducing human antibody genes into the mouse genome, transgenic mice with such traits can be bred indefinitely. Importantly, these transgenic mice are capable of generating human antibodies to human antigens because the only human products expressed in the mice (and therefore recognized as "self") are the antibodies themselves. The mouse thus recognizes any other human tissue or protein as a foreign antigen and the mouse will mount an immune response. Abnormal production of certain human proteins, such as cytokines and growth factors or their receptors, has been implicated in various human diseases. Neutralization or elimination of these abnormally produced or regulated human proteins with the use of human antibodies could ameliorate or suppress the target disease. Therefore, the ability of these transgenic mice to generate human antibodies against human antigens could offer an advantage to drug developers compared with some of the other approaches described previously. A challenge with this approach, however, has been to introduce enough of the human antibody genes in appropriate configuration into the mouse genome

9


to ensure that these mice are capable of recognizing the broad diversity of antigens relevant for human therapies.

        To make our transgenic mice a robust tool capable of consistently generating high affinity antibodies that can recognize a broad range of antigens, we equipped the XenoMouse with approximately 80% of the human heavy chain antibody genes and a significant amount of the human light chain genes. We believe that the complex assembly of these genes together with their semi-random pairing allows XenoMouse animals to recognize a diverse repertoire of antigen structures. XenoMouse technology further capitalizes on the natural in vivo affinity maturation process to generate high affinity, fully human antibodies. In addition, we have developed multiple strains of XenoMouse animals, each of which is capable of producing a different class of antibody to perform different therapeutic functions. We believe that our various XenoMouse strains will provide maximum flexibility for drug developers in generating antibodies of the specific type best suited for a given disease indication.

        We obtain the antibodies generated by XenoMouse animals by extracting the antibody-producing B cells. We can transform these B-cells into hybridomas to generate the quantities of antibodies needed for standard methods of assaying and selecting antibodies for further development. Hybridoma technology captures only about 1% of the antibodies originally generated by the mouse. Alternatively, we can submit the B-cells to our proprietary Selected Lymphocyte Antibody Method (SLAM) technology, which we acquired through our November 2000 acquisition of Abgenix Biopharma Inc. SLAM technology cultures the B-cells directly and rapidly assays them over a period of several days using a microplate-based, high throughput system. Using SLAM, we can typically increase the number of different antigen-reactive monoclonal antibodies identified in a single experiment by 100 to 1000-fold compared to hybridoma technology.

        We use the term XenoMax technology to refer to the use of XenoMouse technology together with SLAM technology. Our XenoMax technology enhances the speed and capability of generating fully human, high affinity antibodies. XenoMax technology allows researchers to rapidly scan the majority of the immune repertoire of an immunized XenoMouse animal, and to identify B-cells that produce antibodies with the desired functional properties and optimal affinities. Using rapid microplate-based assays to measure and rank antibodies according to design goals (e.g., potency, affinity, specificity), XenoMax technology can identify individual B-cells producing extremely high-quality antibodies. It can also recover the antibody encoding genes. Within three to five weeks after immunizing XenoMouse animals, XenoMax technology can produce a ranked set of recombinant antibody candidates resulting from the harvested B-cells. We believe XenoMax technology can speed product development timelines by allowing researchers to move directly into pre-clinical assessment of panels of suitable recombinant candidate antibody products, each ready for manufacturing scale-up. XenoMax technology samples up to 2 million B-cells per immunized XenoMouse animal, dramatically increasing the number of antibodies from which to choose optimal therapeutic product candidates. In contrast to phage display technology, antibodies derived from XenoMax technology retain their native pairing of heavy and light chains, and do not require in vitro affinity and /or potency maturation.

        Other approaches to generating fully human antibodies from mice that we understand are being pursued by competitors include: (i) transgenic mice containing heavy human chain and human light chain genes on a "minilocus" (which are mice that possess a relatively small number of representative human heavy and light chain genes in their genome), (ii) "transchromosomic" mice that contain large numbers of human heavy chain and light chain genes on one or more separate, or extra, chromosomes, and (iii) "UltiMab™" mice that are generated as a result of breeding "minilocus" containing mice with "transchromosomic" mice. "Transchromosomic" mice were developed by Kirin Brewing Co., Ltd. It is our understanding that "UltiMab" mice were developed by a collaboration between Medarex, Inc. and Kirin Brewing Co. and are currently used by Medarex, Kirin, GenPharm International, Inc. and

10


GenMab A/S. Also, Xenerex Biosciences, a subsidiary of Avanir Pharmaceuticals, uses a technology in which human B cells and T cells are implanted in mice with compromised immune systems.

        In addition to the generation of human antibodies from mice, we understand that competitors such as Cambridge Antibody Technology Group plc, MorphoSys AG and Dyax Corporation utilize phage display technology for the generation of human antibodies from phage display libraries derived from human samples. BioSite Incorporated, through a collaboration with Medarex, generates human antibody phage display libraries from immunized "UltiMab" mice. It is our understanding that these libraries are not used for deriving therapeutic antibody products.

Our Technology Advantages

        We believe that our technologies offer the following advantages:

        Producing antibodies with fully human protein sequences.    Our XenoMouse technology, unlike chimeric and humanization technologies, allows the generation of antibodies with 100% human protein sequences. We do not expect antibodies created using XenoMouse technology to cause a HAMA response even when administered repeatedly to patients without compromised immune systems. For this reason, we expect antibodies produced using XenoMouse technology to offer a better safety profile and to be eliminated less quickly from the human body, reducing the frequency of dosing.

        Generating a diverse antibody response to essentially any disease target appropriate for antibody therapy.    Because we have introduced a substantial majority of human antibody genes into XenoMouse animals, the technology has the potential to generate high affinity antibodies that recognize more antigen structures than some other transgenic technologies. In addition, through immune surveillance, we expect XenoMouse technology to be capable of generating antibodies to almost any medically relevant antigen, human or otherwise. For a given antigen target, having multiple antibodies to choose from could be important in selecting the optimal antibody product.

        Generating high affinity antibodies that do not require further engineering.    XenoMouse technology uses the natural in vivo affinity maturation process to generate antibody product candidates, usually in two to four months. These antibody product candidates may have affinities as much as a hundred to a thousand times higher than those seen in phage display. In contrast to antibodies generated using humanization and phage display technology, we and our customers can produce XenoMouse antibodies without the need for any subsequent engineering, a process that at times has proven to be challenging and time consuming. By avoiding the need to further engineer antibodies, we reduce the risk that an antibody's structure and therefore functionality will be altered between the initial antibody selected and the final antibody placed into production.

        Enabling more efficient product development.    XenoMouse technology can potentially produce multiple product candidates more quickly than humanization and phage display technology and we and our customers can conduct pre-clinical testing on several antibodies in parallel to identify the optimal product candidate that will be tested in clinical trials.

        Providing flexibility in choosing manufacturing processes.    Once we have identified an antibody with the desired characteristics, we can produce preclinical material either directly from hybridomas or from recombinant cell lines. Humanized and phage display antibodies, having been engineered, cannot be produced in hybridomas. In addition to potential timesaving, production in hybridomas avoids the need to license certain third party intellectual property rights covering certain processes for production of antibodies in recombinant cell lines.

        Enhancing the speed and capability of generating fully human, high affinity antibodies.    Our XenoMax technology allows researchers to rapidly scan the majority of the immune repertoire of an immunized XenoMouse animal to identify B-cells that produce antibodies with the desired functional

11


properties and optimal affinities. We believe XenoMax technology can speed product development timelines by allowing researchers to move directly into preclinical assessment of panels of suitable recombinant candidate antibody products, each ready for manufacturing scale-up.

        Providing an integrated production platform.    Our integrated production platform has been designed to minimize the risks associated with process, scale and site changes. We believe that our platform, which integrates a comprehensive range of process sciences services, including cell line, cell culture, purification, formulation and assay development, and our manufacturing facility can enable us to rapidly advance product candidates from cell line generation to production. This integrated approach may reduce the variability and risk associated with technology transfer and improve production quality and efficiency.

Abgenix Strategy

        Our objective is to be a leader in the discovery, development and manufacture of antibody-based biopharmaceutical products. Key elements of our strategy to accomplish this objective include the following:

        Building a large and diversified product portfolio by applying our technology to antigens we source.    Utilizing our XenoMouse and XenoMax technologies, our strategy is to develop antibody therapeutic product candidates by using antibodies that we generate under antigen sourcing contracts. This strategy includes sourcing antigens by entering into contractual agreements with leading academic researchers and companies involved in the identification and development of novel antigens, such as those we have entered with several genomics and biopharmaceutical companies. Sourcing antigen targets is a cost effective means for us to gain access to targets that might not otherwise be available to us. Using this strategy, we believe we can create a package that includes antigen rights, human antibodies, and preclinical and clinical data for use by us in self-funded product development efforts or for marketing to potential contract parties to establish collaborations for joint development of proprietary product candidates. We are targeting serious medical conditions, including cancer, inflammation, metabolic diseases, transplant rejection, cardiovascular disease, growth factor modulation, neurological diseases and infectious diseases.

        Establishing collaborations for proprietary product candidates.    Another key strategy is to build our product portfolio and generate revenues by licensing proprietary product candidates, including co-development arrangements, such as the product contracts we have with Amgen and Immunex for ABX-EGF. These proprietary product collaborations involve antibodies made to antigen targets that we source. After generating antibody product candidates and performing limited pre-clinical and clinical development, we intend to license these product candidates, and enter joint development and commercialization agreements. For most of our products, we may enter into proprietary product contracts before entering the Phase 2 clinical development stage, which would allow the contract parties to complete development and to market the product. In some limited circumstances, we may develop the product through later stage clinical trials and license the product candidate to a contract party for marketing. By licensing and entering into co-development arrangements, we can pursue multiple product candidates in the development stage, enabling us to spread our risk of product development, make cost-effective use of available human and capital resources and generate licensing and milestone revenues in the short term.

        Leveraging XenoMouse and XenoMax technology through licensing and other contracts.    We will continue to diversify our product portfolio and generate revenues by entering into contracts with pharmaceutical and biotechnology companies interested in using our XenoMouse and XenoMax technologies to develop antibody-based products. We have established agreements with over thirty customers covering numerous antigen targets. To date, many of these parties have entered into new or expanded agreements with us that allow them to specify additional antigens for antibody development.

12


We expect to enter into additional XenoMouse and XenoMax agreements over time. These agreements typically allow our customer to generate fully human antibodies to one or more specific antigen targets provided by the customer. In some cases, we have provided our mice to the customer who then carries out immunizations with its specific antigen target. In other cases, we immunize the mice with the customer's antigen target for additional compensation. Customers generally have an option for a period of time to acquire product licenses for any antibody product they wish to develop and commercialize.

        Production Services.    With our new manufacturing facility and our existing pilot plant, we intend to offer integrated process sciences and manufacturing services to existing and new collaborators and customers. We will offer these production services to existing customers to further enable their own development efforts, including services related to antibody product candidates that we have developed with them pursuant to existing agreements. Further, we intend to utilize available capacity to manufacture proprietary products that we are co-developing with current collaborators or that will be the subject of future co-development agreements. Finally, we also intend to offer production services to current XenoMouse and XenoMax customers and to new customers for the manufacture of their product candidates for clinical trials and to support the potential early commercial launch of a limited number of their products. We have entered into one such production services agreement, with CuraGen Corporation under which we will provide process sciences and manufacturing services for an antibody product candidate that CuraGen selected from a pool of antibodies we generated pursuant to our existing antigen sourcing contract with CuraGen.

Proprietary Product Development Programs

        We are currently developing antibody therapeutics for a variety of indications. The table below sets forth the current development status of our proprietary product candidates:

Proprietary Product Candidate
   Indication
   Status
ABX-EGF   Various cancers   Phase 1
    Renal cell cancer   Phase 2
    Non-small cell lung cancer   Phase 2(1)
    Colorectal cancer   Phase 2(1)
    Colorectal cancer (with chemotherapy)   Phase 2(1)
    Prostate cancer   Phase 2
ABX-MA1   Metastatic melanoma   Phase 1
ABX-IL8   Chronic obstructive pulmonary disease   Phase 2a(2)
ABX-CBL   Graft versus host disease   Phase 2/3(3)
(1)
Clinical trial managed by Amgen and Immunex.

(2)
In June 2002, we decided to conclude the Phase 2a clinical trial for chronic obstructive pulmonary disease based on a review of the results of our Phase 2 clinical trials in rheumatoid arthritis and psoriasis.

(3)
In February 2003, we completed a preliminary analysis of the Phase 2/3 clinical trial of ABX-CBL and concluded that the study did not meet its primary endpoint. While the Phase 2/3 clinical trial will continue until data collection is complete, we and SangStat do not plan to further develop ABX-CBL and we are winding down this clinical trial.

ABX-EGF

        Tumor cells that overexpress the epidermal growth factor receptor, or EGFr, on their surface often depend on EGFr's activation for growth. EGFr is overexpressed in a variety of cancers including lung, breast, ovarian, bladder, prostate, colorectal, kidney and head and neck. The activation of EGFr is

13


triggered by the binding to EGFr by epidermal growth factor, or EGF, or Transforming Growth Factor alpha, or TGFa, both of which are expressed by the tumor or by neighboring cells. We believe that blocking the ability of EGF and TGFa to bind with EGFr may offer a treatment for certain cancers. ABX-EGF, a fully human monoclonal antibody generated using XenoMouse technology, binds to EGFr with high affinity and has been shown to inhibit tumor cell proliferation in vivo and cause eradication of EGF dependent human tumors established in mouse models. We are conducting pre-clinical studies and assessing which tumor types to pursue as possible targets for treatment with ABX-EGF. Published studies have shown that ABX-EGF can inhibit growth of EGF-dependent human tumors cells in mouse models. ABX-EGF has also demonstrated the ability to reverse cancer cell growth and cause eradication of established tumors in mice even when administered after significant tumor growth has occurred. Furthermore, in these models where tumors were eradicated, researchers did not observe any relapse of the tumor after discontinuation of the antibody treatment.

        Clinical Status.    In July 1999, we initiated a Phase 1 dose-escalating human clinical trial examining the safety, pharmacokinetics and biological activity of multiple doses of ABX-EGF as monotherapy in patients with a variety of advanced cancers. We first reported data on this ongoing study in November 2001 and presented updated information at the annual meeting of the American Society for Clinical Oncology in May 2002. Forty-six patients had been recruited to this study at that time. ABX-EGF appeared to be well tolerated at weekly doses ranging up to 3.5 mg/kg. We did not observe any allergic reactions, clinically significant infusion-related reactions or human anti-human antibody formation. At doses greater than or equal to 2.0 mg/kg, typical EGF receptor mediated skin rashes were seen in 100% of patients. Six patients who had received ABX-EGF (doses of 0.1 or 0.75, 2.5 or 3.5 mg/kg) achieved a partial response, minor response or disease stabilization.

        On the basis of preliminary results from the ongoing Phase 1 clinical study, we and Immunex initiated five Phase 2 studies in April, July and December 2001 and January 2002. The first Phase 2 study is evaluating the effect of ABX-EGF monotherapy in patients with renal cell cancer. An interim analysis of this study was reported at the annual meeting of the American Society of Clinical Oncology in May 2002. A total of 88 patients with metastatic renal cell cancer had been included and treated in this ABX-EGF monotherapy study at the time. ABX-EGF was given weekly in doses of 1.0, 1.5, 2.0, and 2.5 mg/kg to cohorts of approximately 20 patients each. ABX-EGF was administered for eight weeks or until patients demonstrated progressive disease. Eighty-nine percent of patients included in this study had received prior systemic therapy and the majority of patients had received more than one prior systemic regimen. ABX-EGF was generally well tolerated. No allergic reactions, clinically significant infusion-related reactions, or human anti-human antibody formation were observed. A dose-related typical EGFr mediated skin rash was observed with an incidence of 100% at a dose level of 2.5 mg/kg. Single agent biological activity was seen in this heavily pre-treated patient population with three partial responses, two minor response and 50 percent stable disease reported.

        We are conducting the second Phase 2 study in patients with non small cell lung cancer receiving either standard chemotherapy with carboplatin and paclitaxel alone or in combination with ABX-EGF. The third Phase 2 study is evaluating the effect of ABX-EGF monotherapy in patients with metastatic colorectal cancer who have previously failed chemotherapy. The fourth Phase 2 study is evaluating the effect of ABX-EGF monotherapy in patients with hormone resistant prostate cancer without metastasis. The fifth Phase 2 study is evaluating the effect of ABX-EGF in combination with standard chemotherapy, as first-line treatment in patients with metastatic colorectal cancer.

ABX-MA1

        Melanoma is the most serious cancer of the skin. Currently, it is the seventh most common cancer in the United States. The projected 2003 incidence rate in the U.S. is 54,200 and the projected mortality rate is 7,600. Melanoma can spread in the body through the blood and lymphatic system. Organ involvement by metastasis, most commonly to the lungs and liver, is the leading cause of death

14


from the disease. Melanomas that have not spread beyond the site at which they developed are curable by surgical excision. Melanoma that has spread to distant sites is infrequently curable with surgery, although long-term survival is occasionally achieved by resection of metastases. Radiation therapy may provide symptomatic relief for metastases to brain, bones and viscera. Although advanced melanoma is relatively resistant to standard chemotherapy, some biologic therapies, such as interferon alfa and interleukin-2 have been reported to produce a low percentage of objective responses.

        ABX-MA1 targets a protein called MUC18, a cell surface adhesion molecule that is highly expressed on metastatic melanoma cells but not on normal skin cells. MUC18 has been demonstrated to play a critical role in melanoma growth and metastasis by regulating the adhesion and interaction between melanoma cells and surrounding skin cells and new blood vessel cells. In preclinical studies, binding of the MUC18 antigen by ABX-MA1 inhibited primary melanoma tumor growth and the formation of tumor metastases. MUC18 is also expressed on sarcomas, including smooth muscle and blood vessel-derived sarcomas, prostate cancer and renal cell cancers.

        Clinical Status.    In December 2001, we filed an IND and in February 2002 we initiated a Phase 1 clinical trial of ABX-MA1 for the treatment of metastatic melanoma. Enrollment is ongoing.

ABX-IL8

        IL-8, an inflammatory cytokine produced at sites of inflammation, attracts and activates white blood cells that mediate the inflammation process. A number of pre-clinical studies suggest that excess IL-8 may contribute to the pathology and clinical symptoms associated with some inflammatory disorders. Clinical studies have demonstrated significantly increased levels of IL-8 in tissues or body fluids of patients with certain inflammatory diseases, including psoriasis, reperfusion injury and inflammatory bowel disease. Antibodies to IL-8 have been shown to block immune cell infiltration and the associated pathology in animal models of several of these diseases. Using our XenoMouse technology, we have generated ABX-IL8, a proprietary fully human monoclonal antibody that binds to IL-8 with high affinity. We are currently evaluating ABX-IL8 for possible use in the treatment of chronic obstructive pulmonary disease.

        Chronic obstructive pulmonary disease.    Chronic obstructive pulmonary disease (COPD) is a chronic and debilitating disease marked by inflammation and progressive destruction of lung tissue resulting in shortness of breath, persistent cough, recurrent infections and chronic debilitation. COPD is currently the fourth-leading cause of death in the world and has been estimated to affect over 15 million people in the United States, 60 percent of whom have a severe form of the disease. Studies have correlated elevated levels of IL-8 in the broncho-aveolar fluid and lung tissue of COPD patients with inflammatory cells such as neutrophils, which have been implicated in the chronic destruction of lung tissue in patients with COPD. Pre-clinical studies have shown that antibodies to IL-8 have blocked the migration of neutophils.

        Clinical Status.    In September 2001, we submitted an IND to initiate a Phase 2a double-blind, placebo-controlled study designed to evaluate the efficacy and safety of ABX-IL8 in COPD. We designed the study to include a total of 150 patients across approximately 20 clinical sites in the United States. Patients received a total of three doses of ABX-IL8 administered monthly over a two-month period. Efficacy analyses focused on change in airflow, shortness of breath and disease-related quality of life. After studying the results of our Phase 2 clinical trials in rheumatoid arthritis and psoriasis, we announced in January 2002 and May 2002 that these results did not support further clinical trials of ABX-IL8. Further, based on a review of those results, we decided in June 2002 to conclude the COPD study as quickly as possible, consistent with patient safety follow-up.

15


ABX-CBL

        The CBL antigen is selectively over-expressed on activated immune cells including T-cells, B-cells and certain macrophages. We obtained an exclusive license to ABX-CBL, a mouse antibody, in February 1997. In August 2000, we entered into a joint development and commercialization agreement with SangStat for ABX-CBL.

        The goal for the development program was to reduce unwanted immune responses that occur in graft versus host disease, or GVHD, a life-threatening complication that frequently occurs following an allogeneic bone marrow transplant, or BMT. BMTs are used in the treatment of patients with leukemia, certain other serious cancers and immune system disorders. An allogeneic BMT procedure involves transferring marrow, the graft, from a healthy person into an immunosuppressed patient, the host. Often a portion of the graft recognizes the host's own cells as foreign, becomes activated and attacks them, resulting in GVHD. It typically involves damage to multiple organ systems, including the skin, liver and intestines and is the primary cause of death in allogeneic BMT patients. Current treatments consist of corticosteroids and other drug treatments to suppress the grafted immune cells. Approximately 2,000 patients per year in the United States contract steroid-resistant GVHD.

        Clinical Status.    In December 1999, we initiated a Phase 2/3 clinical trial to evaluate the survival benefit from ABX-CBL in patients with GVHD. The results of this study indicate that ABX-CBL demonstrated a survival rate at 180 days in patients with acute steroid-resistant GVHD that was similar to the drug administered in the study's control arm. The study was designed to demonstrate superior survival with ABX-CBL, and, therefore, did not meet its primary endpoint. We and SangStat do not plan further development of ABX-CBL.

Summary of Contractual Arrangements

Overview

        As of February 28, 2003, we had entered into contracts covering numerous antigen targets with over thirty customers to use our XenoMouse and XenoMax technologies to generate and/or develop the resulting fully human antibodies. Also as of February 28, 2003, we had entered into a contract with one customer to provide process sciences and manufacturing services related to an antibody product candidate that we developed with the customer pursuant to an antigen sourcing contract. Pursuant to our XenoMouse contracts, we and our customers intend to generate antibodies for development as product candidates for the treatment of cancer, inflammation, autoimmune diseases, transplant rejection, cardiovascular disease, growth factor modulation, neurological diseases, metabolic diseases and infectious diseases. We also plan to enter into additional contracts to provide antibody process sciences and manufacturing services. We expect that substantially all of our revenues for the foreseeable future will result from payments under these and other contracts. We have also licensed technology from third parties for use in conjunction with our proprietary technologies. The terms of our current contractual arrangements vary, but can generally be categorized as follows:

16

17

Summary of Payment Terms of Contractual Arrangements

        We derive our contract revenues from our target sourcing contracts, our proprietary product development agreements and our technology out-licensing contracts. We also expect to generate contract revenue from production services agreements. Generally, contract revenues consist of license, option, milestone, service and royalty payments. To date, we have received license, option and milestone payments from various parties but have yet to receive any production service payments or royalty payments.

        Pursuant to our target sourcing contracts and our technology out-licensing contracts, in 2002, we have recognized individual license, option and milestone payments that represented between approximately 0.08% and 4.63% of our recognized contract revenues for 2002 (not including revenues we recognized from Celltech Ltd.).

        Under our proprietary co-development agreement with Amgen and Immunex, we recognized revenues of $3.2 million in 2002, which represented approximately 17% of our contract revenues for that year. Under our proprietary co-development agreement with SangStat, we recognized revenues of $2.0 million in 2002, which represented approximately 10% of our contract revenues for that year. We expect that the estimated payments we may receive from SangStat for 2003 will be less than the payments we received for 2002 as a result of discontinuing the development of ABX-CBL.

        We expect to receive future license, option and milestone payments from our customers and collaborators; however, the amount and timing of these payments, if any, is uncertain because they depend to a large extent on the success of the research and development efforts of these parties.

        To date we have entered into one production services contract. Pursuant to this agreement, which we entered into in January 2003, we will be paid for the production services we provide pursuant to a work plan.

18

        While most of our target sourcing, proprietary product development and technology out-licensing contracts entitle us, under certain circumstances, to royalty payments, we have not received any royalty payments to date and do not anticipate receiving any such payments for a least a few years. We have entered into a production services contract that entitles us, under certain circumstances, to royalty payments. We will not be entitled to royalty payments unless our customers or collaborators are successful in developing and commercializing products derived from our technology. The likelihood that we or our collaborators will be successful is dependent on the outcome of research and development efforts and regulatory decisions with respect to our product candidates, and is therefore uncertain and speculative.

Summary of Expense Terms of Contractual Arrangements

        We have incurred expenses, including license, option or milestone payments, under our in-license agreements and we may incur future expenses of this sort under our target sourcing contracts. We may also incur future expenses in the form of royalty fees under one or more of these agreements.

        Under our in-licensing agreements and target sourcing contracts, in 2002, we made individual license, option or milestone payments that represented between approximately 0.01% and 0.09% of our research and development expenses for 2002.

        While most of our technology in-licensing and proprietary product development contracts include provisions for the payment of royalties by us under certain circumstances, we have not made any royalty payments to date and believe we are at least a few years away from selling any products that would require us to make any royalty payments. Whether we will ever be obligated to make royalty payments to third parties is subject to the future success of our research and development efforts, as well as the favorable decisions of regulators and, accordingly, is inherently uncertain.

Circumstances that Trigger Milestone Payments under Contractual Arrangements

        Under our target sourcing contracts, milestone payments with respect to therapeutic products may become payable, to us or by us, in the following circumstances:

        Under these contracts, milestone payments with respect to diagnostic products may become payable, to us or by us, in the following circumstances:

19

        Our proprietary co-development agreement with Immunex provides for no milestone payments by either party. Under our proprietary co-development agreement with SangStat, SangStat is obligated to make milestone payments to us upon the completion of certain clinical trials. Because we have decided to discontinue development of ABX-CBL, we do not expect any further milestone payments.

        Under our technology out-licensing agreements, milestone payments may become payable to us in the following circumstances:

        Under our technology in-licensing agreements, milestone payments may become payable by us in the following circumstances:

Termination Provisions of Contractual Arrangements

        Our agreements generally do not have definite termination dates; rather, these agreements typically terminate upon the expiration of the underlying royalty obligations. Whether these royalty obligations will be triggered and, if so, when, is dependent on the successful development and commercialization of products from the subject technology.

        Our target sourcing contracts generally terminate upon the expiration of all royalty obligations, if any, due under the relevant contract.

20

        Under our out-licensing agreements, the licensee typically can terminate the agreement at any time and we generally can terminate upon a breach by the licensee. Absent early termination, our out-licensing agreements typically continue in effect until the expiration of the licensee's payment obligations.

        In some cases, we can terminate in-licensing agreements after a certain period of time. Other in-licensing agreements do not provide for early termination by us (except in the case of the other party's breach), but provide that the agreement terminates upon the expiration of all of our royalty payment obligations.

Xenotech and the XenoMouse Technology

        In 1989, Cell Genesys started our business and operations as a subsidiary. In June 1991, Cell Genesys entered into several agreements with Japan Tobacco, Inc. for the purpose of forming Xenotech. In connection with the formation of Xenotech, both Cell Genesys and Japan Tobacco contributed cash, and Cell Genesys contributed the exclusive right to certain of its technology for the research and development of genetically modified strains of mice that can produce fully human antibodies. Cell Genesys assigned its rights in Xenotech to us in connection with our formation as an independent company in 1996. Through 1998, we made capital contributions to Xenotech, and provided research and development to Xenotech related to the development of XenoMouse technology in exchange for cash payments.

        Under several agreements with Japan Tobacco that became effective December 31, 1999, we acquired Japan Tobacco's fifty percent interest in the Xenotech joint venture and became the sole owner of Xenotech and the XenoMouse technology. Under these agreements, Japan Tobacco acquired a license to use certain existing XenoMouse technology and future XenoMouse technology that we develop and a license to certain new technology related to the generation of mouse models of certain human diseases, in exchange for cash payments and future royalty obligations.

Gene Therapy Rights Agreement with Cell Genesys

        In connection with the formation of Abgenix by Cell Genesys, Abgenix entered into the Gene Therapy Rights Agreement, or GTRA, which provides Cell Genesys with certain rights to commercialize products based on antibodies generated with XenoMouse technology in the field of gene therapy. Under the GTRA, Cell Genesys has certain rights to direct us to make antibodies to two antigens per year and has an option for a license to commercialize antibodies binding to such antigens in the field of gene therapy. The GTRA obligates Cell Genesys to make certain payments to us for these rights, including reimbursement of license fees and royalties on future product sales. The GTRA also prohibits us from granting any third-party licenses for antibody products based on antigens where the primary field of use is gene therapy. In the case of third-party licenses granted by us where gene therapy is a secondary field, the GTRA obligates us to share with Cell Genesys a portion of the cash milestone payments and royalties resulting from any products in the field of gene therapy.

21


Intellectual Property

        We rely on patents and trade secrets to protect our intellectual property rights. We own seven issued patents in the United States, one granted patent in Europe, three granted patents in Japan and several granted patents in other foreign countries. In addition we have 39 pending patent applications in the United States and 156 pending patent applications abroad relating to XenoMouse technology. Our wholly-owned subsidiary, Xenotech, owns three issued U.S. patents, one Australian patent and several granted patents in other foreign countries and has two issued U.S. patents and three pending foreign patent applications related to methods of treatment of bone disease in cancer patients, and has one U.S. patent relating to genetic manipulation. Our wholly owned subsidiary Abgenix Biopharma, owns one issued U.S. patent and has one pending patent in Canada and Europe relating to the SLAM technology. Our wholly owned subsidiary IntraImmune Therapies, Inc. has three pending applications in the United States and nine pending applications in other foreign countries related to intrabody technology, which may give antibodies access to intracellular targets. In addition, we have nine issued U.S. patents, several granted patents in other foreign countries, seven pending patent applications in the United States and eighteen pending patent applications abroad that we jointly own with Japan Tobacco relating to antibody technology or genetic manipulation. While we rely on U.S. and foreign patent laws to protect our proprietary technology, any patents, if issued, may provide us with little protection, especially in foreign countries.

        We also attempt to protect our technologies by maintaining trade secrets and proprietary know-how. However, the agreements we enter into for these purposes may not be enforced or our counter parties may breach them. In addition, these agreements may not prevent third parties from discovering our trade secrets or know-how or independently developing the same or similar technologies.

        Scientists have conducted research for many years in the antibody and transgenic animal fields. This has resulted in a substantial number of issued patents and an even larger number of pending patent applications. Patent applications in the United States are, in most cases, maintained in secrecy until patents are issued. The publication of discoveries in the scientific or patent literature frequently occurs substantially later than the date on which the underlying discoveries were made. Our commercial success depends significantly on our ability to operate without infringing the patents and other proprietary rights of third parties. Our technologies may unintentionally infringe the patents or violate other proprietary rights of third parties. Such infringement or violation may prevent us and our contract parties from pursuing product development or commercialization. Such a result would materially harm our business, financial condition and results of operations.

        GlaxoSmithKline plc, or Glaxo, has a family of patents relating to certain methods for generating monoclonal antibodies that Glaxo is asserting against Genentech, Inc. in litigation that was commenced in 1999. On May 4, 2001, Genentech announced that a jury had determined that Genentech had not infringed Glaxo's patents and that all of the patent claims asserted against Genentech are invalid. We understand that Glaxo has filed a notice of appeal with the Court of Appeals for the Federal Circuit. If any of the claims of these patents are finally determined in the litigation to be valid, and if we were to use manufacturing processes covered by the patents to make our products, we may then need to obtain a license should one be available. Our failure to obtain a license at all or on commercially reasonable terms could impede commercialization of one or more of our products in any territories in which these claims were in force.

        Genentech, Johnson & Johnson, Glaxo and Transkaryotic Therapies, Inc. each owns or controls a U.S. patent that relates to recombinant cell lines or methods of generating recombinant cell lines for the production of antibodies. If we were to use a production system covered by any of these patents, we may then need to obtain a license should one be available. Under these circumstances, our failure

22


to obtain a license at all or on commercially reasonable terms could impede commercialization of one or more of our products in any territories in which these patent claims were in force.

        Genentech, owns a U.S. patent that issued in June 1998 relating to inhibiting the growth of tumor cells that involves an anti-EGF receptor antibody in combination with a cytotoxic factor. ImClone Systems, Inc. owns or is licensed under a U.S. patent that issued in April 2001, relating to inhibiting the growth of tumor cells that involves an anti-EGF receptor antibody in combination with an anti-neoplastic agent. However, we do not believe that either the Genentech or ImClone patent would be successfully asserted against any planned commercial sales of ABX-EGF. We believe that currently all of the Company's activities relating to anti-EGFr receptor monoclonal antibodies are within the exemption provided by the U.S. patent laws for uses reasonably related to obtaining FDA approval of a drug. We do not expect the scope of our product development plans to change in the future prior to filing an application for a biologic license with the FDA. If a court determines that the claims of either the Genentech patent or the ImClone patent cover our activities with ABX-EGF and are valid, such a decision may require us to obtain a license to Genentech's or ImClone's patent, as the case may be, to label and sell ABX-EGF for certain combination therapies. Our failure to obtain a license at all or on commercially reasonable terms could impede our commercialization of ABX-EGF in the United States.

        In 2000, the Japanese Patent Office granted a patent to Kirin Beer Kabushiki Kaisha, one of our competitors, relating to non-human transgenic mammals. Kirin has filed corresponding patent applications in Europe and Australia. Kirin may also have filed a corresponding patent application in the United States. Our licensee, Japan Tobacco, has filed opposition proceedings against the Kirin patent. We cannot predict the outcome of those opposition proceedings, which may take years to be resolved. In any event, based on our analysis of the Kirin patent, we believe that the patent will not adversely affect our business.

        Extensive litigation regarding patents and other intellectual property rights has been common in the biotechnology and biopharmaceutical industries. The defense and prosecution of intellectual property suits, United States Patent and Trademark Office interference proceedings and related legal and administrative proceedings in the United States and internationally involve complex legal and factual questions. As a result, such proceedings are costly and time-consuming to pursue and their outcome is uncertain. Litigation may be necessary to:

        If we become involved in any litigation, interference or other administrative proceedings, we will incur substantial expense and the efforts of our technical and management personnel will be significantly diverted. An adverse determination may subject us to loss of our proprietary position or to significant liabilities, or require us to seek licenses that may not be available from third parties. An adverse determination in a judicial or administrative proceeding or our failure to obtain necessary licenses could restrict or prevent us from manufacturing and selling our products, if any. Costs associated with such arrangements may be substantial and may include ongoing royalties. Furthermore, we may not be able to obtain the necessary licenses on satisfactory terms, if at all. These outcomes will materially harm our business, financial condition and results of operations.

Patent Cross-License with GenPharm

        In March 1997, we along with Cell Genesys, Xenotech and Japan Tobacco, signed a comprehensive patent cross-license with GenPharm. Under the cross-license, we have licensed on a non-exclusive basis certain patents, patent applications, third-party licenses and inventions pertaining to the development and use of certain transgenic rodents, including mice that produce fully human antibodies. We use our

23


XenoMouse technology to generate fully human antibody products and have not licensed the use of, and do not use, any transgenic rodents developed or used by GenPharm. All of our financial obligations in connection with the cross-license were recognized in 1997.

Government Regulation

        Our product candidates under development are subject to extensive and rigorous domestic government regulation. The FDA regulates, among other things, the development, testing, manufacture, safety, efficacy, record keeping, labeling, storage, approval, advertising, promotion, sale and distribution of biopharmaceutical products. If we market our products abroad, they also are subject to extensive regulation by foreign governments. Non-compliance with applicable requirements can result in fines, warning letters, recall or seizure of products, clinical study holds, total or partial suspension of production, refusal of the government to grant approvals, withdrawal of approval, and civil and criminal penalties.

        We believe our antibody therapeutic products will be classified by the FDA as "biologic products" as opposed to "drug products." The steps ordinarily required before a biological product may be marketed in the United States include:

        Preclinical testing includes laboratory evaluation of product chemistry, formulation and stability, as well as animal studies to assess the potential safety and efficacy of each product. Laboratories that comply with FDA regulations regarding good laboratory practices must conduct preclinical safety tests. We submit the results of the preclinical tests, together with manufacturing information, analytical data and clinical study plans, to the FDA as part of the IND and the FDA reviews those results before the commencement of clinical trials. Unless the FDA objects to an IND, the IND will become effective 30 days following its receipt by the FDA. If we submit an IND, our submission may not result in FDA authorization to commence clinical trials. Also, the lack of an objection by the FDA does not mean it will ultimately approve an application for marketing approval. Furthermore, we may encounter problems in clinical trials that cause us or the FDA to delay, suspend or terminate our trials.

        Clinical trials involve the administration of the investigational product to humans under the supervision of a qualified principal investigator. We must conduct clinical trials in accordance with Good Clinical Practice under protocols submitted to the FDA as part of the IND. In addition, each clinical trial must be approved and conducted under the auspices of an Institutional Review Board and with patient informed consent. The Institutional Review Board will consider, among other things, ethical factors, the safety of human subjects and the possibility of liability of the institution conducting the trial.

        We conduct clinical trials in three sequential phases that may overlap. Phase 1 clinical trials may be performed in healthy human subjects or, depending on the disease, in patients. The goal of a Phase 1 clinical trial is to establish initial data about safety and tolerance of the biologic agent in humans. In Phase 2 clinical trials, we seek evidence about the desired therapeutic efficacy of a biologic agent in limited studies of patients with the target disease. We make efforts to evaluate the effects of various dosages and to establish an optimal dosage level and dosage schedule. We also gather additional safety data from these studies. The Phase 3 clinical trial program consists of expanded, large-

24


scale, multi-center studies of persons who are susceptible to or have developed the disease. The goal of these studies is to obtain definitive statistical evidence of the efficacy and safety of the proposed product and dosage regimen.

        Historically, the results from preclinical testing and early clinical trials have often not predicted results obtained in later clinical trials. A number of new drugs and biologics have shown promising results in clinical trials, but subsequently failed to establish sufficient safety and efficacy data to obtain necessary regulatory approvals. Data obtained from preclinical and clinical activities are susceptible to varying interpretations, which could delay, limit or prevent regulatory approval. In addition, we may encounter delays or rejections by regulatory authorities as a result of many factors, including changes in regulatory policy during the period of product development.

        Completion of clinical trials may take several years or more. The length of time generally varies substantially according to the type, complexity, novelty and intended use of the product candidate. Many factors may delay our commencement and rate of completion of clinical trials including:

        We have limited experience in conducting and managing clinical trials. We rely in part on third parties, including our collaborators, to assist us in managing and monitoring clinical trials. Our reliance on third parties may result in delays in completing, or failing to complete, clinical trials if they fail to perform under our agreements with them.

        Only four of our product candidates, ABX-EGF, ABX-MA1, ABX-CBL and ABX-IL8, have been in clinical trials. With respect to ABX-EGF and ABX-MA1, we have not obtained enough data from these clinical trials to date to demonstrate safety and efficacy under applicable FDA guidelines. As a result, such data will not support an application for regulatory approval without further clinical trials. With respect to ABX-CBL and ABX-IL8, the results of the clinical trials we conducted did not support further clinical studies. Clinical trials that we conduct or that third parties conduct on our behalf for any product candidate may not demonstrate sufficient safety and efficacy to obtain the requisite regulatory approvals. Regulatory authorities may not permit us to undertake any additional clinical trials for our product candidates.

        Our product candidates may fail to demonstrate safety and efficacy in clinical trials. For example, in January 2002 and May 2002, respectively, we announced that clinical trials of our proprietary product candidate ABX-IL8 as treatment for rheumatoid arthritis and psoriasis did not support further clinical studies of that product candidate. Additionally, in February 2003, we completed a preliminary analysis of the Phase 2/3 clinical trial of ABX-CBL and concluded that the study did not meet its primary endpoint and did not support further clinical studies of that product candidate. These and other potential failures may delay development of other product candidates, and hinder our ability to conduct related preclinical testing and clinical trials. As a result of such failures, we may also be unable to obtain additional financing. The failure of clinical trials can also result in research and development charges, such as those we incurred in the second quarter of 2002 in connection with our decision to wind down our clinical trials for ABX-IL8. Any delays in, or termination of, our clinical trials would materially harm our business, financial condition and results of operations.

        We have ongoing research projects that may produce product candidates, and we have not submitted INDs or begun clinical trials for these projects. We may not successfully complete our pre-clinical or clinical development efforts. We may not file further INDs and we may not commence clinical trials as planned.

25


        We and our third-party manufacturers also are required to comply with the applicable FDA current good manufacturing practice regulations and other regulatory requirements. Good manufacturing practice regulations include requirements relating to quality control and quality assurance as well as the corresponding maintenance of records and documentation. Manufacturing facilities are subject to inspection by the FDA and the facilities must pass an inspection by the FDA before they can be used in commercial manufacturing of any product. Manufacturing facilities in California, including our facility, are also subject to the licensing requirements of and inspection by the California Department of Health Services. We or our third-party manufacturers may not be able to comply with the applicable good manufacturing practice requirements and other regulatory requirements. If we or our third-party manufacturers fail to comply, our business, financial condition and results of operations will be materially harmed.

        For clinical investigation and marketing outside the United States, we may be subject to the regulatory requirements of other countries, which vary from country to country. The regulatory approval process in other countries includes requirements similar to those associated with FDA approval set forth above.

Competition

        The biotechnology and pharmaceutical industries are highly competitive and subject to significant and rapid technological change. We are aware of several pharmaceutical and biotechnology companies that are actively engaged in research and development in areas related to antibody therapy. These companies have commenced clinical trials of antibody therapeutic product candidates or have successfully commercialized antibody therapeutic products. Many of these companies are addressing the same diseases and disease indications as we or our customers are. Also, we compete with companies that offer antibody generation services to companies that have antigens. These competitors have specific expertise or technology related to antibody development and introduce new or modified technologies from time to time. These companies include GenPharm; Kirin Brewing Co.; GenMab; Cambridge Antibody Technology Group; Protein Design Labs, Inc.; MorphoSys; Xenerex Biosciences; XLT Biopharmaceuticals Ltd.; and Alexion Pharmaceuticals, Inc.

        Some of our competitors have received regulatory approval of or are developing or testing product candidates that may compete directly with our product candidates. For example, ImClone, AstraZeneca, Glaxo and a collaboration of OSI Pharmaceuticals, Inc., Genentech, and Roche have potential antibody and small molecule product candidates in clinical development that may compete with ABX-EGF, which is also in clinical trials. Furthermore, we are aware that AstraZeneca has received licensing approval in Japan for its experimental cancer drug Iressa, which may compete with ABX-EGF.

        Many of these companies and institutions, either alone or together with their customers, have substantially greater financial resources and larger research and development staffs than we do. In addition, many of these competitors, either alone or togeth