Immunogenicity of Biopharmaceuticals: An Example from Erythropoietin

Published on: 
, ,
BioPharm International, BioPharm International-08-01-2005, Volume 18, Issue 8

Subcutaneous administration is likely to be an important factor in generating an immunogenic response.

Biotechnology-derived therapeutic proteins are playing an ever-increasing role in the pharmaceutical market. These proteins are mostly used in the treatment of life-threatening and severe diseases. The aim is to correct an acquired or inherited deficiency of a native protein or to alter a disease process. Interestingly, most therapeutic proteins have been shown to be immunogenic, despite the fact that the amino acid sequence is identical or nearly identical to endogenous proteins. The formation of antibodies appears, in many cases, not to have a clinical effect.

Marc H. V. van Regenmortel, Ph.D

There are cases, however, in which antibodies were found to be neutralizing, which can negate the effect of the therapy or when cross-reacting with the endogenous protein, deplete the effect of this protein. The clinical effects of such cases are significant, and can result in a more severe disease and a need for acute intervention. Several potential mechanisms can cause immunogenicity of therapeutic proteins such as aggregates, molecular mimicry, adjuvants, neo-antigens, and modification of the protein. Epoetin is an example of one display of this mechanism.


Over the past 20 years our understanding of immunology has grown exponentially, and there is no end in sight. Because of the complexity of this field, any discussion related to the formation of antibodies requires simplification since there are multiple potential routes for the production of antibodies against a therapeutic protein. In general, we focus on the most common pathways. For a more detailed discussion, the book by Abbas and Lichtman serves as a good reference.1

The number of biotechnology-derived therapeutic proteins now available includes over 75 products representing 65 different types of molecules.2 In spite of the fact that biopharmaceuticals currently have only a small share of the pharmaceutical market, their medical impact is enormous. Examples of deficiency correction are recombinant human-insulin therapy for diabetes mellitus, therapy with factor VIII in patients with hemophilia A, and recombinant human erythropoietin (epoetin) therapy for the treatment of anemia. An example of a biopharmaceutical that alters disease processes is the use of recombinant interferons in patients with multiple sclerosis, hepatitis, and cancer.

The majority of therapeutic proteins have an identical or nearly identical amino acid sequence of endogenous proteins, and may not be immunogenic if they are recognized as self through mechanisms of immune tolerance (Figure 1). Yet, nearly all of the biopharmaceuticals currently approved for clinical use are known to be immunogenic.3 Subtle differences in protein structure, a lack of tolerance in the patient, or the presence of an adjuvant may result in the generation of an antibody response. In many cases of antibody formation to biopharmaceuticals, there is no clinical effect.3-5

Figure 1. Development of Central and Peripheral Tolerance to Self Antigens

In some cases antibodies may be neutralizing, which can negate the effect of the therapy.6-16 Cross-reaction with the endogenous protein can deplete the effect of this protein.17-19 This effect has been observed in the management of anemia in chronic kidney disease patients, where cross-reactive neutralizing antibodies were found to inhibit the activity of both the administered epoetin and the patient's own residual erythropoietin.18,19 This created a more severe anemia, called pure red-cell aplasia (PRCA), characterized by an almost complete absence of red-blood-cell precursors in bone marrow.18,19 In the majority of reported cases of PRCA, the presence of anti-erythropoietin antibodies has been confirmed, resulting in the condition named antibody-mediated PRCA (Ab+PRCA).


Since epoetin alfa was introduced in 1989 as medicine to treat anemia, many hundreds of thousands of patients have received doses. In the first decade of its use, three patients with chronic kidney disease (CKD) were reported to have developed Ab+PRCA.20-22 From 1998 onwards however, the incidence of Ab+PRCA increased,19,23 reaching a peak in 2002.24 Over 200 cases of Ab+PRCA have now been reported in patients with CKD treated with subcutaneous epoetin.24 The incidence of Ab+PRCA ranges from rare (<1 × 10-3 patient-years) to very rare (<1 × 10-4 patient-years), but it nevertheless represents a serious clinical problem that was poorly understood.25

Statistics associate the problem with subcutaneous administration of epoetin and not intravenous administration. AB+PRCA occurs in patients treated for anemia associated with CKD but not in cancer patients.26,27 Although the majority of cases are associated with exposure to epoetin alfa, there are also a number of cases where patients have received only epoetin beta, which differs from epoetin alfa in glycosylation, isoform composition, and formulation.26-28

In two studies conducted to determine the nature of anti-erythropoietin antibodies present in blood samples of PRCA patients also receiving epoetin therapy, all the patients had antibodies that prevented the proliferation of erythroid progenitor cells from normal bone marrow. Moreover this inhibition could not be overcome by additional epoetin.19,27 In the first study, antibodies from 12 of the 13 patients would bind only to conformational epitopes in the protein moiety of 125I-labeled erythropoietin and were of low-to-moderate affinity, while antibodies from the remaining patient (who had received only epoetin beta) would bind to both conformational and linear epitopes and were of high affinity.19 The presence or absence of the carbohydrate moieties had little effect on erythropoietin binding to the antibodies. The second study, which included an additional seven patients, reported that all patients received subcutaneous epoetin to correct anemia of chronic renal failure.27

Use of surface plasmon resonance (SPR) to detect binding of anti-erythropoietin antibodies to immobilized erythropoietin confirmed that these antibodies are of low-to-moderate affinity. The same technique has shown that human anti-erythropoietin antibodies detected in the PRCA subjects consist of immunoglobulin IgG isotypes, with no indication of the presence of IgE, IgA, or IgM isotypes. Although T-cell-mediated immune responses begin as IgM, the IgM response is short-lived, and it is likely to have switched to IgG by the time patients are screened for anti-erythropoietin antibodies. Epitope determination using a series of glycosylated erythropoietin muteins indicated that the antibody response was oligoclonal, and that the epitopes recognized varied both qualitatively and quantitatively between individuals.29 Discussion of possible mechanisms of erythropoietin immunogenicity induction follows a brief overview of a typical immune response.


Mature antigen-responsive B-cells develop from bone marrow precursors and populate peripheral lymphoid tissues, which are the sites of interaction with antigens. B-cells have a specific membrane IgM- or IgD-type receptor. When these receptors recognize an antigen, multiple B-cell receptors cross-link to multiple copies of an antigenic protein on the pathogen surface (Figure 2). The B-cell engulfs the pathogen, digests it, and presents viral peptides (8 to 15 amino acids long) in a class II major histocompatibility complex (MHC) on its surface. The presence of an adjuvant, often present with a virus or bacteria, will further activate the B-cell. At this stage, the B-cell proliferates, secretes IgM antibodies, and expresses the co-stimulating factor B7 on its surface. The B-cell is now activated, but this IgM immune response will be short-lived without T-cell help.

Figure 2. The Activation of B-cells in the IgM Immune Response

T-cells recognize short peptides on the surface of B-cells or other antigen-presenting cells (APCs), such as dendritic cells, in association with class II MHC (Figure 3).1 If an activated B-cell encounters a T-cell that recognizes a class II MHC-linked peptide it is presenting, then this recognition and binding to the T-cell receptor, combined with the interaction of CD28 molecules on the T-cell with B7 molecules on the B-cell, activates the T-cell.

Figure 3. The Activation of T-cells in the IgG Immune Response

The T-cell also produces other factors (e.g., CD40-ligand and cytokines) that stimulate growth and differentiation of B-cells. This classic T-cell mediated immune response demonstrates the full potential of the B-cell response. This is the typical response of the body to viruses and microorganisms, and is a likely mechanism for B-cell activation against proteins when macro-aggregates are present. At this stage, a number of changes occur in the B-cell: 1

1. Proliferation. The B-cell begins to multiply to generate a larger population of B-cells.

2. Isotype Switching. The B-cell begins to switch the heavy-chain class of its antibody from IgM or IgD to IgG, IgA, or IgE. From a single B-cell, a population of different antibody classes specific to the same epitope can be generated.

3. Affinity Maturation. Through successive genetic modifications, the B-cells will modify the antigen-binding site of the antibody. A natural selection process occurs, where B-cells with a higher affinity receptor continue to proliferate as B-cells with weaker affinity receptors are eliminated.

4. Epitope Spreading. As a result of small modifications of the antigen-binding site, the exact epitope on the foreign protein to which the antibody binds can also change.

5. Differentiation. As the B-cells proliferate, some will differentiate into antibody-secreting cells (called plasma cells) and others will become long-lived memory B-cells, which can initiate a rapid secondary response if the recipient is exposed to the same immunogen later in life.

Soluble Proteins

When only soluble protein, such as erythropoietin, is present (no bacterial, viral, or aggregate structure), a similar type of immune response occurs. This case is shown in Figure 4, and begins generally in the subcutaneous space where APCs (frequently dendritic cells) take up a protein, digest it, and present protein fragments via the class II MHC complex. If there are T-cells present that can recognize these protein fragments (peptides) presented by MHC, they can be stimulated if the APCs are activated (by an adjuvant) to provide a second signal.

Figure 4. Induction of Anti-EPO Antibodies in CKD Patients


This activation of T-cells can occur in the subcutaneous space, or when dendritic cells loaded with erythropoietin migrate to a peripheral lymph node where the T-cell receptor is engaged. B-cells, capable of recognizing epoetin, can become activated by the T-cells and begin to proliferate and mature. This mechanism is considered to be the most probable in breaking tolerance to a self-antigen administered subcutaneously.

Multiple factors are required to trigger a T-cell-mediated immune response to a recombinant therapeutic protein. In the case of epoetin, these include the presence of a sufficient number of erythropoietin-recognizing B- and T-cells in the patient, as well as epoetin and an adjuvant (Figure 4). When the therapeutic protein is homologous to a self-antigen, as is the case with epoetin, the level of both B- and T-cells that would respond to epoetin is very low because of self-tolerance, making this response a rare event. However, repetitive dosing in the presence of an adjuvant could eventually break that tolerance and lead to the production of anti-erythropoietin antibodies and PRCA. These basic features of immune responses were taken into account when designing the technical investigation into the cause of epoetin-associated PRCA.


The investigation into the particular source of increased immunogenicity related to Eprex brand of epoetin was triggered by Casadevall's report.19 Her discussion of the possible immunological factors that could affect the production of anti-erythropoietin antibodies in CKD patients treated with epoetin served as a basis of the technical investigation into specific Eprex-related factors. Although this discussion focused on a specific product, it is applicable to all biopharmaceutical proteins.

Table 1. Mechanisms of Tolerance in B-cells and T-cells

Because epoetin is a homologue of the endogenous (self) protein erythropoietin, antibody production to erythropoietin is most likely caused by a breakdown of immune tolerance. The mechanisms of tolerance and breakdown of tolerance are outlined in Tables 1 and 2.

Table 2. Factors That May Influence the Immunogenicity and Tolerogenicity of Protein Antigens

The breakdown of tolerance results in the formation of anti-erythropoietin IgG antibodies, and current data support the concept that these antibodies recognize the native three-dimensional structure of erythropoietin common to all epoetin products.19 The oligoclonality of the antibodies, the presence of IgG isotypes, and the time frame of onset of the antibody response (which may be many months after the first exposure) all suggest that T-cell-dependent activation of B-cells is responsible for the formation of antibodies to erythropoietin.19,23 Although T-cell-mediated immune responses begin as IgM, this response is short-lived and is likely to have switched to IgG by the time patients are screened for anti-erythropoietin antibodies, offering a probable explanation as to why no IgM isotypes have been detected in patients with PRCA.

The central question then emerges of whether (and how) epoetin induces a breakdown of the immunological tolerance that normally prevents the body from responding to self-antigens, or alternatively, whether its structure has been modified so that it has become a neo-antigen (i.e., one that the body has not seen before). Table 3 provides a summary list of known and suspected causes of increased immunogenicity to biopharmaceutical products. Although this discussion focuses on epoetin alfa, it may also be relevant to immune reactions of other biopharmaceuticals. We had to eliminate many of the possible causes listed in the table, which is why this was such a complex investigation.

Table 3. Possible Causes of Immunogenicity of Injectable Biopharmaceutical Self-proteins: A Guide for Investigation


Tolerance to erythropoietin could be broken if the epoetin formed large multivalent aggregates with a highly organized, repetitive display of antigenic determinants (spacing in the 5 to 10 nm range), which resemble virus-like particles.30,31 Epoetin-alfa molecules can aggregate, when exposed to elevated temperatures. If such aggregates are large enough, they could hypothetically act as polyvalent neo-antigens, and thus activate B-cells and induce short-lived IgM production. In the case of Eprex, no changes in aggregate content were found between recent bulk and historical lots. Levels of aggregate found at the end of shelf life for final product were well below the 2% level allowed by the European Monograph for bulk drug substance.51 Heat induced aggregates in concentrations higher than 50% did not induce an immune response to epoetin alfa in mice. It is difficult to correlate animal immune responses to human responses, so any animal data have limited predictive value. However, the data indicate that aggregation of epoetin did not generate any new epitopes recognized by the mice. For other proteins, such as interferon-α2a and human growth hormone, the formation of aggregates was suggested to contribute to immunogenicity associated with these proteins.32,33 Large structures are the focus of any investigation on the role of aggregates, and there were no large structures detected in the Eprex investigation.

One factor that seems to be associated in the last decade with increased incidence of Ab+PRCA was the replacement of human serum albumin with polysorbate 80 as a stabilizer in the Eprex formulation of epoetin alfa sold outside the US. Polysorbate 80 can form micelles above a certain critical concentration which could act as a surface for epoetin molecules to gather and form a large multivalent structure. Subcutaneous administration might increase immunogenicity because the virus-like micelles would remain more concentrated after subcutaneous administration than after intravenous injection.34,35 Subsequent investigation has indicated that erythropoietin–polysorbate-80 micelle structures do not exist, and that polysorbate-80 micelles are too unstable to act as a multivalent structure.36,37 Neverthe-less, when investigating immune reactions to proteins, the potential of multivalent structures should be carefully evaluated.

Molecular Mimicry

The incidence of Ab+PRCA in CKD patients treated with epoetin is very low, and is in the same range as the incidence of some autoimmune diseases, such as Guillain–Barré syndrome38 and Goodpasture's syndrome.39 One mechanism that could explain the infrequent induction of erythropoietin antibodies is molecular mimicry between epoetin and an antigen on an infecting microorganism. Molecular mimicry can involve carbohydrate modification of proteins, and sugar profiling has demonstrated that epoetin differs in carbohydrate structure from human serum erythropoietin.40 However, carbohydrates usually do not elicit a robust immune response, and anti-carbohydrate antibodies were not found in patients with Ab+PRCA.19,27 Although the theoretical possibility of B-cell epitope mimicry cannot be discounted, there is no known link between Ab+PRCA and any type of infection that could produce a mimicking foreign-carbohydrate epitope.


There are a number of potential adjuvants that could exist in biopharmaceutical products. Host-cell proteins from the production cell line can co-purify with a recombinant protein and could, if present in sufficient quantity, induce an immune reaction.13,41,42 This is especially true for proteins that are produced in microbial expression systems.13,42 Because epoetin alfa is produced in Chinese Hamster Ovary (CHO) mammalian cells, and the CHO host-cell protein level is in the part-per-billion range, host-cell proteins represent a relatively low level of risk.

Silicone oil, used in the siliconization of glass pre-filled syringes, could act as an adjuvant.43-46 There is evidence in the literature that silicone gels can serve as weak adjuvants; however, silicone oils are weaker adjuvants than silicone gels and low-molecular-weight silicone oils have not been demonstrated to stimulate an immune response in animals.47 In addition, the siliconization process for Eprex has been in place since 1994, which was four years prior to the observed increase in immunogenicity. Subsequent investigation was not supportive of an adjuvant effect of silicone oil in Eprex preparations.

Chemical adjuvants could be present in some of the materials used in the formulation of the product, or they could occur as compounds that leach from the container closure system. Sharma and colleagues showed that the addition of polysorbate 80 to the formulation resulted in leached compounds in the product, and the adjuvant activity of these compounds was demonstrated through immunogenicity studies in mice.37 In addition, clinical investigations indicated that Eprex associated PRCA was significantly correlated with exposure to pre-filled syringes with uncoated rubber stoppers.48


Neo-antigens are foreign epitopes that are not tolerogenic for the immune system. For some indications ofbiopharmaceutical protein usage, the ability to produce an endogenous form of the protein never existed in the patient. In such cases, (complete factor VIII deficiency, growth hormone deficiency) the biopharmaceutical protein represents a neo-antigen, and formation of antibodies is common and expected.11,49 The case of epoetin is unique; normal B- and T-cell tolerance to epoetin is expected because all CKD patients produced normal levels of erythropoietin before suffering from CKD. However, because CKD patients produce much less erythropoietin, it is possible that a higher number of B- or T-cells that respond to erythropoietin may survive the normal tolerogenic mechanisms.

Modification of the Protein

Neo-epitopes could be present if the epoetin molecule has undergone significant misfolding, disulfide scrambling, or reactions with other proteins. The normal release- and characterization-testing of Eprex has not indicated the presence of abnormal structures. The loss of carbohydrate could expose areas of the molecule that are not normally seen by the human immune system; however, all patient antibodies bind to the fully glycosylated molecule, making this an unlikely explanation.19 It is also possible that abnormal carbohydrate structures could elicit antibody formation; but this is unlikely because patient antibodies bind to the deglycosylated molecule.19

Modification of the protein during production or storage could also produce neo-epitopes. Typical modifications that have been reported for proteins are post-translational changes such as oxidation or deamidation. Highly sensitive quantitative methods to determine if changes in the protein occurred coincidently with the increase in PRCA were recommended and subsequently investigated. No changes in the epoetin molecule were detected.


There are many interacting components in the immunological mechanisms of tolerance and antibody formation, and it is likely that the cause of anti-erythropoietin antibodies is multifactorial. In addition to the potential mechanisms of immunogenicity of epoetin described earlier, a number of factors may also play a role.

Route of Administration

Subcutaneous administration is likely to be an important factor in generating an immunogenic response. Administration of epoetin via the subcutaneous route has increased — particularly for predialysis and peritoneal dialysis patients where intravenous access is unavailable. Subcutaneous administration localizes the drug to a small area with a short path to drain into the lymph nodes where B- and T-cells are present.

Localization can also allow dendritic cells to infiltrate the site where they could present MHC-restricted peptides to T-cells. If an adjuvant is present, the dendritic cells can be activated by the adjuvant and subsequently activate T-cells. Many investigators have found a lower bioavailability of epoetin after subcutaneous injection when compared with intravenous injection. This indicates incomplete absorption from subcutaneous tissue. Patients who self-administer epoetin subcutaneously may accidentally do so intradermally, increasing the possibility of an immunogenic reaction. Furthermore, the linked T-cell activity required to induce high-titer anti-erythropoietin IgG antibodies in patients with Ab+PRCA could be generated by co-stimulatory activities that might include adjuvants, foreign protein products (either in the preparation or introduced at the injection site), microbial infection, or simply the trauma of injection. When intravenous injection is used, there is an immediate dilution and dispersion of all of these potential factors.

There are no cases in the Johnson & Johnson database of Ab+PRCA associated with a patient who has only received Eprex intravenously.48 Based upon this, it is a rule that epoetins should always be administered via the intravenous route in CKD patients, if intravenous access is available.24,50

Patient Characteristics

Genetic factors are likely to be important predisposing factors for Ab+PRCA just like for other autoimmune diseases. Studies of erythropoietin gene polymorphisms and MHC subtypes have shown no association with Ab+PRCA so far,52 but these studies were very small and do not completely rule out genetic influences. We feel it is unlikely that such genetic factors will be determined.

To date, Ab+PRCA has primarily occurred in CKD patients.19,23 There have been no cases of Ab+PRCA in cancer patients, which is why subcutaneous administration is still indicated for this population. This suggests that normal tolerance mechanisms have been weakened in some CKD patients due to their decreased erythropoietin production. Non-CKD cancer patients likely produce normal to high levels of erythropoietin to counter the reduction of hematopoesis in their bone marrow. Alternatively, the apparent protection of cancer patients may be due to immunosuppressive therapy.


The immunological mechanisms of tolerance and antibody formation are complex, with many interacting components. It is likely that the cause of anti-erythropoietin antibody formation is multifactorial, involving both B-cell and T-cell responses to the epoetin. While genetic factors are likely to be important predisposing factors for Ab+PRCA the predisposing factors in the population are not likely to have changed in 1998, when the incidence of Ab+PRCA suddenly increased. Therefore, the most probable explanation for the increase in the incidence of Ab+PRCA with epoetin treatment in CKD patients points to a 1998 change in formulation that led to the presence of leachates that act as adjuvants. Additional technical analyses have been done that support this hypothesis.37 In addition, animal studies and epidemiological data also support the hypothesis.37,48

The product factors that have changed include the major route of administration (from intravenous to subcutaneous) and the substitution of human serum albumin with polysorbate 80. Neither subcutaneous administration nor the use of polysorbate 80 represent a sufficient explanation for the increase in PRCA. The increase did not occur with subcutaneous administration of other epoetins, including Eprex dosage forms where uncoated stoppers were not used.48

The cause of the increase in the immunogenicity of Eprex has been determined to be the interaction of the polysorbate 80 with the uncoated rubber stopper in syringes, which led to the leaching of organic compounds (leachates) that have been demonstrated to act as adjuvants in mice.37 The replacement of the uncoated rubber stopper with a FluroTec-coated syringe stopper has eliminated the leachates and together with other risk mitigation efforts have reduced the incidence of Ab+PRCA.37,48


When investigating the immunogenicity of recombinant therapeutic proteins, it is important to remember that every stage of the production process must be carefully monitored, because even factors such as formulation and packaging are important for biopharmaceutical proteins. The B-cells, T-cells, and dendritic APCs are patient-derived and as such will be related to the patients genetics and type of disease, and can be part of the clinical investigation.

In contrast, the recombinant protein and possible adjuvants are likely to be product-derived and need to be explored through a technical product-related investigation. If an adjuvant is involved, it could be formulation- or patient-related, or opportunistic, such as from an infection. Awareness of factors that influence the immunogenicity of recombinant therapeutic proteins is likely to become more important as the number of biopharmaceuticals on the market increases and new versions of these products begin to emerge. When dealing with immunogenicity, the structure of the protein, contaminants from the manufacturing process, the formulation, and the container and closure can all affect the immunogenicity of the product.


For a complete listing, refer to BioPharm International's website at

Marc H.V. van Regenmortel, Ph.D. is emeritus research director at Ecole Supérieure de Biotechnologie de Strasbourg,Boulevard Sébastien Brandt,F-67412 Illkirch,France,33.390.24.48.12,fax 33.390.24.48.11,

Katia Boven,M.D. is a medical researcher at Johnson & Johnson,Pharmaceutical Research and Development, 920 Route 202, Titusville, NJ 08869,908.707.3436, fax 908.526.2650,

Fred Bader, Ph.D. is vice president of Johnson & Johnson Global Biologics Supply Chain LLC, 200 Great Valley Parkway, Malvern, PA 19355,610.889.4619, fax 610.651.6482,


1. Abbas AK, Lichtman AH. Cellular and Molecular Immunology, 5th edition. Saunders: Philadelphia; 2003.

2. Evens RP, Sindelar RD. Biotechnology products in the pipeline. In: Crommelin DJA, Sindelar RD (Ed.). Pharmaceutical Biotechnology, 2nd edition. London: Francis & Taylor; 2002.

3. Porter S. Human immune response to recombinant human proteins. J. Pharm. Sci. 2001; 90:1-11.

4. Schernthaner G. Immunogenicity and allergenic potential of animal and human insulins. Diabetes Care. 1993; 16:155-165.

5. Laricchia-Robbio L, Moscato S, Genua A, et al. Naturally occurring and therapy-induced antibodies to human granulocyte colony-stimulating factor (G-CSF) in human serum. J. Cell Physiol. 1997; 173:219-226.

6. Oberg K, Alm G, Magnusson A, et al. Treatment of malignant carcinoid tumors with recombinant interferon alfa-2b: development of neutralizing interferon antibodies and possible loss of antitumor activity. J. Natl. Cancer Inst. 1989; 81:531-535.

7. Olsen E, Duvic M, Frankel a, et al. Pivotal phase III trial of two dose levels of denileukin diftitox for the treatment of cutaneous T-cell lymphoma. J. Clin. Oncol. 2001; 19:376-388.

8. Rosenschein U, Lenz R, Radnay J, et al. Streptokinase immunogenicity in thrombolytic therapy for acute myocardial infarction. Isr. J. Med. Sci. 1991; 27:541-545.

9. Vanderschueren SM, Stassen JM, Collen D. On the immunogenicity of recombinant staphylokinase in patients and in animal models. Thromb. Haemost. 1994; 72:297-301.

10. Grauer A, Ziegler R, Raue F. Clinical significance of antibodies against calcitonin. Exp. Clin. Endocrinol. Diabetes. 1995; 103:345-351.

11. Jacquemin MG, Saint-Remy JM. Factor VIII immunogenicity. Haemophilia 1998; 4:552-557.

12. Giannelli G, Antonelli G, Fera G, et al. Biological and clinical significance of neutralizing and binding antibodies to interferon-alpha (IFN-alpha) during therapy for chronic hepatitis C. Clin. Exp. Immunol. 1994; 97:4-9.

13. Prümmer O. Treatment-induced antibodies to interleukin-2. Biotherapy 1997; 10:15-24.

14. Blumenfeld Z, Frisch L, Conn PM. Gonadotropin-releasing hormone (GnRH) antibodies formation in hypogonadotropic azoospermic men treated with pulsatile GnRH–diagnosis and possible alternative treatment. Fertil. Steril. 1988; 50:622-629.

15. Claustrat B, David L, Faure A, Francois R. Development of anti-human chorionic gonadotropin antibodies in patients with hypogonadotropic hypogonadism. A study of four patients. J. Clin. Endocrinol. Metab. 1983; 57:1041-1047.

16. Ragnhammar P, Wadhwa M. Neutralizing antibodies to granulocyte-macrophage colony stimulating factor (GM-CSF) in carcinoma patients following GM-CSF combination therapy. Med. Oncol. 1996; 13:161-166.

17. Zipkin, I. Amgen lays MGDF to rest. BioCentury 1998 Sept.

18. Casadevall N, Dupuy E, Molho-Sabatier P, Tobelem G, Varet B, Mayeux P. Autoantibodies against erythropoietin in a patient with pure red-cell aplasia. New Engl. J. Med. 1996; 334:630-633.

19. Casadevall N, Nataf J, Viron B, Kolta A, Kiladjian JJ, Martin-Dupont P, et al. Pure red-cell aplasia and antierythropoietin antibodies in patients treated with recombinant erythropoietin. New Engl. J. Med. 2002; 346:469-475.

20. Bergrem H, Danielson BG, Eckardt KU, Kurtz A, Stridsberg M. A case of antierythropoietin antibodies following recombinant human erythropoietin treatment. In: Bauer C, Koch KM, Scigalla P, Wieczorek L (Ed.). Erythropoietin: molecular physiology and clinical application. New York: Marcel Dekker; 1993.

21. Prabhakar SS, Muhlfelder T. Antibodies to recombinant human erythropoietin causing pure red cell aplasia. Clin. Nephrol. 1997; 47:331-335.

22. Peces R, de la Torre M, Alcazar R, Urra JM. Antibodies against recombinant human erythropoietin in a patient with erythropoietin-resistant anemia. New Engl. J. Med. 1996; 335:523-524.

23. Eckardt KU, Casadevall N. Pure red-cell aplasia due to anti-erythropoietin antibodies. Nephrol. Dial. Transplant. 2003; 18:865-869.

24. Johnson & Johnson. Press release. Summary of PRCA case reports: update on our company's effort to reduce case reports of pure red cell aplasia. October 10, 2003. Available at: Accessed November 2004.

25. Bennett C, Luminari S, Nissenson A, et al: Pure red-cell aplasia and recombinant epoetin therapy - a follow-up report from the research on adverse drug events and reports project. New Engl. J. Med. 2004; 351:1403–1408.

26. Krüger A, Schröer W, Röhrs F, Vescio G. PRCA in a patient treated with epoetin beta. Nephrol. Dial. Transplant. 2003; 18:1033-1034.

27. Casadevall N. Antibodies against rHuEPO: native and recombinant. Nephrol. Dial. Transplant. 2002; 17 Suppl 5:42-47.

28. Storring PL, Tiplady RJ, Gaines Das RE, Stenning BE, Lamikanra A, Rafferty B, et al. Epoetin alfa and beta differ in their erythropoietin isoform compositions and biological properties. Br. J. Haematol. 1998; 100:79-89.

29. Heavner G. Personal communication. 2003 May.

30. Bachmann MF, Zinkernagel RM. The influence of virus structure on antibody responses and virus serotype formation. Immunol. Today. 1996; 17:553-558.

31. Chackerian B, Lenz P, Lowy DR, Schiller JT. Determinants of autoantibody induction by conjugated papillomavirus virus-like particles. J. Immunol. 2002; 169:6120-6126.

32. Hochuli E. Interferon immunogenicity: technical evaluation of interferon-alpha 2a. J. Interferon Cytokine Res. 1997; 17:S15-S21.

33. Moore WV, Leppert P. Role of aggregated human growth hormone (hGH) in development of antibodies to hGH. J. Clin. Endocrinol. Metab. 1980; 51:691-697.

34. Hermeling S, Schellekens H, Crommelin DJ, Jiskoot W. Micelle-associated protein in epoetin formulations: a risk factor for immunogenicity? Pharm. Res. 2003; 20:1903-1907.

35. Schellekens H. Immunogenicity of biopharmaceuticals: why proteins should be treated with respect. Eur. J. Hosp. Pharm. 2003; 8:76-81.

36. Kerwin B, Deechongkit S, Park S, Kim J, Burnett H. Effects of polysorbates 20 and 80 on the structure and stability of darbepoetin alfa and epoetin alfa. European Renal Association-European Dialysis and Transplant Association (ERA-EDTA). 17 May 2004: 321.

37. Sharma B, Lisi P, Ryan M, Bader F. Technical investigations into the potential cause of the increased incidence of epoetin-associated pure red cell aplasia. Eur. J. Hosp. Pharm. 2004; 10:86-91.

38. Ho T, Griffin J. Guillain-Barre syndrome. Curr. Opin. Neurol. 1999; 12:389-394.

39. Phelps RG, Rees AJ. The HLA complex in Goodpasture's disease: a model for analyzing susceptibility to autoimmunity. Kidney Int. 1999; 56:1638-1653.

40. Skibeli V, Nissen-Lie G, Torjesen P. Sugar profiling proves that human serum erythropoietin differs from recombinant human erythropoietin. Blood 2001; 98:3626-3634.

41. Wadhwa M, Skog AL, Bird C, et al. Immunogenicity of granulocyte-macrophage colony-stimulating factor (GM-CSF) products in patients undergoing combination therapy with GM-CSF. Clin. Cancer Res. 1999; 5:1353-1361.

42. Antonelli, G, Simeoni E, Currenti M, et al. Interferon antibodies in patients with infectious diseases. Anti-interferon antibodies. Biotherapy 1997; 10:7-14.

43. van Oss CJ, Singer JM, Gillman CF. The influence of particulate carriers and of mineral oil in adjuvants on the antibody response in rabbits to human gamma globulin. Immunol. Commun. 1976; 5:181-188.

44. Naim JO, Lanzafame RJ. The adjuvant effect of silicone-gel on antibody formation in rats. Immunol. Invest. 1993; 22:151-161.

45. Naim JO, Lanzafame RJ, van Oss CJ. The effect of silicone-gel on the immune response. J. Biomater. Sci. Polymer Ed. 1995; 7:123-132.

46. Naim JO, Ippolito KML, Lanzafame RJ, van Oss CJ. Induction of type II collagen arthritis in the DA rat using silicone gels and oils as adjuvant. J. Autoimmun. 1995; 8:751-761.

47. Naim JO, Ippolito KML, Lanzafame RJ. The effect of molecular weight and gel preparation on humoral adjuvancy of silicone oils and silicone gels. Immunol. Invest. 1995; 24:537-547.

48. Boven K, Stryker S, Knight J et al. The increased incidence of pure red cell aplasia with an Eprex formulation in uncoated rubber stopper syringes. Kidney Int. 2005; in press.

49. Frasier SD. Human pituitary growth hormone (hGH) therapy in growth hormone deficiency. Endocr. Rev. 1983; 4:155-170.

50. Johnson & Johnson. Press release. Intravenous administration required when using EPREX® /ERYPO® (epoetin alfa in chronic renal failure patients. Subcutaneous administration still available for oncology and other patients. December 2, 2002. Available at: Accessed November 2004.

51. Erythropoitin concentrated solution. Pharmeuropa 14.1. 2002 January: 94-9.

52. Casadevall N. Personal communication. 2003 August.