Manufacturing Process Development for an Epidermal Growth Factor-Based Cancer Vaccine

ABSTRACT

The Center for Molecular Immunology (Havana, Cuba) has been working on a novel cancer immunotherapy targeting the epidermal growth factor (EGF). The vaccine is composed of a chemical conjugate of EGF and a carrier protein (rP64k), designed to trigger an anti-EGF antibody response. Advanced clinical trials demanded further manufacturing process development to address scale-up and validation limitations in the initial process. The process development strategy focused on replacing the membrane dialysis purification step with a method that could be scaled up more easily; incorporating disposable technology to further facilitate scale-up and cleaning validation; and improving process and product characterization. The consistency of the new process was evaluated, and the equivalence of the vaccine preparations was assessed. The results of studies of molecular characterization, immunogenic activity, and clinical data are presented here.

Cancer vaccines were originally designed to target tumor-specific antigens: non-self antigens expressed only in tumor cells due to malignant mutations. It is now known that only a minority of cancer cells truly express no self antigens. Several studies have also been carried out on the immune recognition of non-mutated self antigens on tumor cells. The results of these studies support the feasibility of cancer vaccines targeting fully self antigens. ^1,2

CENTRO DE INMUNOLOGÍA MOLECULAR

The epidermal growth factor receptor (EGF-R), a member of a family of membrane receptors with tyrosine kinase activity, is emerging as a new target candidate for anti-cancer therapy.³ Several agents targeting the EGF-R or its ligands are already in clinical testing, including small-molecule tyrosine kinase inhibitors,⁴ monoclonal antibodies,^5,6 and cancer vaccines.^7,8

Rationale for an EGF–Based Therapeutic Cancer Vaccine

A novel active immunotherapy approach, designed to provoke an anti-EGF antibody response, has been developed at the Center for Molecular Immunology (Havana, Cuba). The vaccine (under the commercial name CIMAvax-EGF) comprises the EGF coupled to an immunogenic carrier protein, administered together with an adjuvant.^10–19

It is known that tumor cell proliferation begins with binding between the EGF and its cell membrane receptor (EGF-R). The anti-EGF antibodies, induced by vaccination with CIMAvax EGF, bind to the EGF (self molecule) inhibiting its binding to the EGF-R, and thus preventing activation of cell proliferation mechanisms derived from this ligand–receptor interaction (Figure 1).

Figure 1. Scheme showing the mechanism of action of CIMAvax-EGF. Patients are immunized with a vaccine comprising an epidermal growth factor (EGF) linked to an immunogenic carrier protein and administered with an adjuvant. Vaccination provokes the formation of anti-EGF antibodies that bind to circulating EGF, blocking it from binding to its membrane receptor (EGF-R) and thus preventing the cell proliferation mechanisms initiated by EGF-EGF-R binding. In vaccinated patients, anti-EGF antibodies increase while EGF sera concentrations decrease over time.

To validate the approach of vaccination with EGF in lung cancer therapy, clinical trials were designed to select the proper vaccine formulation (immunogenic carrier protein and adjuvant), as well as the proper scheme and therapeutic dose that would have a meaningful clinical effect.^{12–14,17–19}

Clinical experiences with CIMAvax-EGF in advanced lung cancer patients demonstrated that vaccination provoked an increase in anti-EGF antibody titers and a decrease in EGF sera concentration. The increase in anti-EGF antibodies directly correlated with increased survival of vaccinated patients. Decreases in EGF sera concentration also correlated with increased survival of vaccinated patients.^{14,17, 19} In randomized controlled trials, it was demonstrated that overall, more vaccinated patients survived than non-vaccinated controls, an effect that is much greater in patients under 60 years old.¹⁹

The long path for translating this basic concept into a real product (CIMAvax-EGF) began in 1992. The first challenge was trying to prepare an immunogenic preparation with a self antigen, because any cancer vaccine based on self antigens must be designed to create an adequate presentation environment to provoke a clinically significant immune response.⁹ To achieve this, the self protein (EGF) was conjugated to a immunogenic carrier protein. Several carrier proteins were tested in the preclinical^10,11 and clinical settings. ^{12–14,17–19} Based on immunogenicity results, the recombinant protein rP64k from Neisseria meningitides was selected for continued product development.¹³

Two adjuvants, aluminum hydroxide and Montanide ISA-51 (Seppic, France), were tested. The best results, in terms of immunogenicity of the vaccine formulation, were obtained using Montanide ISA-51, so this adjuvant was selected for further product development.¹³

For development of this vaccine through proof-of-concept (POC) clinical trials, a fairly simple manufacturing process was used. This initial production process consisted of chemical conjugation using a linker reagent, followed by an impurity-removal step with a dialysis membrane. However, this process had practical disadvantages for scale-up and compliance with good manufacturing practice (GMP) requirements.

Process Development

As the product development cycle advanced to late clinical trials, the manufacturing process needed to be improved to comply with GMP requirements and undergo validation. The challenge was to develop a new process for advanced stages of development (ASD) while maintaining the performance equivalence with the vaccine preparation used for the POC studies.

The process development strategy focused on the following goals:

• Optimizing the conjugation reaction

• Replacing the membrane dialysis purification process with a step that could be scaled up more easily

• Incorporating disposable tech-nology to further facilitate scale-up and cleaning validation

• Improving process and product characterization

• Evaluating the process con-sistency of the new process

• Evaluating the equivalence of the vaccine preparations.

Optimizing the Conjugation Reaction

The chemical conjugation method developed for the POC vaccine preparation allowed unspecific binding of the immunogenic carrier protein (rP64K) and the autologous protein (rEGF). This procedure required a high molar quantity of rEGF to avoid reaching the reaction limit. To ensure that the conjugation reaction yielded reproducible conjugation products following scale-up, conjugation reaction kinetics were studied to understand reaction times, the effect of reactant feeding strategies, and mixing requirements. As a result of the optimization studies, mixing and reaction times were defined that allowed reproducible conjugation results following scale-up of the process more than ten- fold. The working ranges for the main process variables were set for commercial operation following robustness studies.

Replacing the Membrane Dialysis Purification Step

Most of the limitations of the process used to manufacture product for the POC studies resulted from the membrane-based dialysis purification step, which was designed to remove the free conjugation reagent and other chemical substances before final formulation. The main drawbacks of this step were extended process time, limited scalability, and the inclusion of several manual operations in the process.

To overcome these difficulties, a purification step based on crossflow membrane ultrafiltration–diafiltration (UF–DF) was introduced as an alternative to the dialysis membrane.

This UF-DF procedure reduced processing time and significantly reduced the risk of microbial contamination. It also made it possible to perform clean-in-place operations. Furthermore, this purification alternative could be scaled up in a linear fashion to adapt to varying batch volumes.

This purification step also made it possible to remove excess autologous protein (low molecular weight conjugates) from the vaccine preparation by using higher cut-off membranes, with the aim of producing a more homogeneous vaccine drug substance. The rationale behind this purification of the mixture of conjugation species was that the excess of autologous protein (rEGF), either in form of polymers or as free molecules, does not contribute to the immuno-genicity of the vaccine preparation, and instead dilutes the immunological action of the chemical conjugated rEGF-rP64k.

Figure 2 shows the anti-rEGF immunological response in mice immunized with a) the vaccine preparation for POC trials, b) the vaccine preparation for the ASD process, or c) fractions of rEGF polymers and free rEGF molecules. Mice immunized with free rEGF or rEGF polymers did not show anti-rEGF antibody responses, whereas mice immunized with both vaccine preparations did. It was also observed that the vaccine preparation for the ASD process showed higher anti-rEGF antibody responses in mice when compared with the vaccine preparation for POC trials. These results confirmed that it made sense to carry out ultrafiltration of the conjugation mixture with higher cut-off membranes, to remove excess autologous protein from the vaccine preparation.

Figure 2. Anti-rEGF immunological response in mice immunized with a) the vaccine preparation for proof-of-concept (POC) trials; b) the vaccine preparation from advanced stages of development (ASD); c) fractions of rEGF polymers and free rEGF molecules.

The UF–DF purification step thus provided a more homogeneous vaccine composition, rich in immunologically relevant species (rEGF–rP64k). Figure 3 shows the molecular exclusion chromatographic profile of the products obtained by both processes. As can be observed, the product made for POC trials yielded a chroma-tographic profile in which 42% of the total area belonged to the rEGF-rP64k conjugated species (Figure 3A, peak 1). On the other hand, the vaccine preparation for the ASD process resulted in a chromatographic profile in which the conjugated species (rEGF–rP64k) corresponds to 86% of the area (Figure 3B, peak 1).

Figure 3. The molecular exclusion chromatographic profiles of the products obtained by the two different manufacturing processes. The product prepared for proof-of-concept trials yielded a chromatographic profile in which 42% of the total area corresponded to the rEGF-rP64k conjugated species (chromatogram A, peak 1). The vaccine preparation for advanced stages of development resulted in a chromatographic profile in which the conjugated species rEGF-rP64k corresponded to 86% of the area (chromatogram B, peak 1).

In addition, the UF–DF step is used to concentrate the protein mixture, allowing a wider range of potency and dose adjustments for the vaccine preparation for different clinical settings.

Introducing Disposable Technology into the Process

Disposable technology was introduced into the process to provide flexibility for scale-up and facilitate GMP compliance. Disposable filters and bioprocess containers (bags) were introduced into the manufacturing process to support operations such as mixing, buffer preparation, and buffer and bulk storage.

The steps of final sterile filtration and aseptic filling also were improved by using disposable technology. A list of single-use products incorporated into various steps in the vaccine manufacturing process is shown in Table 1.

Table 1. Disposable products introduced into the manufacturing process

Improved Process and Product Characterization

During translation of the POC vaccine formulation into the ASD product, a set of analytical tests was developed for product characterization and for quality control release of the purified bulk and final product (Figure 4). For antigenic identity, the presence of conjugated species (rEGF–rP64k) was detected by SDS/Western blot. For this purpose, we developed a specific anti-EGF antibody.

Figure 4. Process flow chart for the CIMAvax-EGF manufacturing process for advanced stages of product development. The in-process controls and the quality control test for final product and biological starting materials (rEGF and rP64k proteins) are included in the corresponding steps.

The content of immunogenic species (rEGF–rP64k) was determined by HPLC–GF. Quantification of total protein concentration was performed using the Lowry method. To evaluate residual glutaraldehyde content, an HPLC–RP assay was developed using a C-8 column.

Immunogenic activity tests are a significant challenge in human cancer vaccines based on a self protein. The difficulty is greater for in vivo tests, because what is self for humans is not self for other animal species. In our case, we had the advantage that the outbred mice strain (NMRI) only developed anti-rEGF responses when immunized with the conjugated rEGF–rP64k, and did not do so when immunized with rEGF alone. We then developed a test by immunizing NMRI mice, in which immune responses demonstrated the immunogenicity not only of the antigen, but also of the conjugation of both species (rEGF and rP64k) .

Evaluating Process Consistency

The analytical data (Table 2) showed that the process was able to produce consecutive vaccine lots with consistent performance. They also showed the equivalence between the products manufactured for POC trials and ASD.

Table 2. Comparative analytical results of the final product tests for the two vaccine preparations

Evaluating the Equivalence of the Vaccine Preparations

The equivalence of products produced for POC trials and for ASD was evaluated using molecular characterization, biological activity in mice, and clinical data from administration to patients with advanced lung cancer.

To demonstrate equivalence through molecular characterization, the biologically active moieties (rEGF–rP64k conjugated species) of both products (for POC and ASD) were analyzed by peptide mapping, following digestion of the conjugated rEGF-rP64k fractions with the endoprotease Glu-C. Each fraction of the resulting maps was identified and sequenced by mass spectrometry. Figure 5 shows the peptide maps of the rEGF–rP64k conjugated species from A) the product made for POC and B) the product made for ASD. A high degree of similarity is seen in the peptide profiles and their relative amounts, indicating the equivalence of the products in terms of the structure of their biologically active moieties (rEGF–rP64k conjugated species).

Figure 5. Peptide maps of chromatographic fractions corresponding to the conjugated rEGF-rP64k (active moieties) from (A) product made for proof-of-concept trials and (B) product made for advanced stages of development.

Figure 6 shows biological activity using measurements of anti-rEGF antibody response in mice for the POC and the ASD vaccine preparations. No statistically significant differences were observed.

Figure 6. Comparison of the immunogenic activity test results obtained after immunizing mice with vaccine preparations for proof-of-concept (POC) trials and for advanced stages of development (ASD). Each bar represents the mean optical density from an ELISA assay performed on eight lots of vaccine. Different sera dilutions were tested (1:100, 1:1,000 and 1:10,000).

The clinical performance of both products was also evaluated in advanced lung cancer patients. Clinical data from eight patients vaccinated with each vaccine preparation were compared (Figure 7). As can be observed, there were no significant differences in the anti-rEGF antibody titers obtained when immunizing patients with the vaccine for POC and with the vaccine for ASD, demonstrating an equivalent immunological effect in humans.

Figure 7. Anti-EGF antibody responses in patients vaccinated with the preparations made for proof-of-concept (POC) trials and advanced stages of development (ASD). Y axis represents the geometric means of anti-EGF antibody titers; bars represent the geometric mean of antibody titers in eight patients vaccinated with the two different products. The line represents the geometric mean of anti-EGF antibody in time for control sera (unvaccinated patients).

Conclusions

The initial limitations of the manufacturing process used for proof of concept studies of the CIMAvax-EGF vaccine were overcome by a process development strategy that focused on scalability and GMP requirements. The new in-process tests and end-product testing showed acceptable consistency and comparability. The vaccine preparation obtained following the manufacturing improvements showed similar clinical performance compared to the product used during the proof-of-concept trials.

GRYSSELL RODRIGUEZ is the CIMAvax-EGF manufacturing manager, AIRAMA ALBISA is a process development scientist, LISEL VIÑA is the CIMAvax-EGF quality control manager, ARIADNA CUEVAS is the CIMAvax-EGF regulatory affairs manager, BEATRIZ GARCIA is a clinical researcher, AURORA TAMARA GARCIA is the CIMAvax-EGF quality assurance manager, ALEJANDRO PORTILLO is a quality control specialist, LOANY CALVO is a quality control specialist, TANIA CROMBET is the clinical research director, and ERNESTO CHICO is the technical director, all at the Centro de Inmunología Molecular (Center for Molecular Immunology), Havana, Cuba, grisel@cim.sld.cu, +53-7 2716810. GISELA GONZALEZ is the business development manager at CIMAB SA, Havana, Cuba.

References

1. González G, Lage A. Cancer vaccines for hormone/growth factor immune deprivation: A feasible approach for cancer treatment. Curr Cancer Drug Targets. 2007;7(3):191–201.

2. Rosenberg S. Progress in human tumor immunobiology and immunotherapy. Nature. 2001;411:380–384.

3. Lage A, Crombet T, González G. Targeting epidermal growth factor receptor signaling: Early results and future trends in oncology. Annal Med. 2003;35:327–336.

4. Fukuoka M, Yano S, Giaccone G, Tamura T, Nakagawa K, Douillard JK. Final results from a phase II trial of ZD 1839. ("Iressa") for patients with advanced non-small-cell lung cancer (IDEAL 1). Proc Am Soc Clin Oncol. 2002;21:292a Abs1188.

5. Mateo C, Moreno E, Amour K, Lombardero J, Harri W, Perez R. Humanization of a mouse monoclonal antibody that block the EGF-R: recovery of antagonistic activity. Immunotechnology. 1997;3:71–81.

6. Slamon DJ, Leyland-Jones B, Fuchs H, Platon V. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast that overexpresses Her2. N Engl J Med. 2001;344:783–792.

7. Lage A, Pérez R, Fernández L.E. Therapeutic cancer vaccines—At midway between immunology and pharmacology. Curr Cancer Drug Targets. 2005;5:611–627.

8. Yu Z, Restifo NP. Cancer vaccines: Progress reveals new complexities. J Clin Invest. 2002;110:289–294.

9. Patent EP 0 657 175 B1 and US 5 894 018: Vaccine comprising human autologous epidermal growth factor and use thereof.

10. González G, Sánchez B, Suárez, E, Beausoleil, I, Perez, O, Lastre, M, Lage, A. Induction of immune recognition of self epidermal growth factor (EGF): Effect on EGF biodistribution and tumor growth. Vaccine Res.1996;5(4):233–244.

11. González G, Pardo O.L, Sánchez B, García JL, Beausoleil I, Marinello P, González Y, Domarco A, Guillén G, Pérez R., Lage A. : induction of immune recognition of self epidermal growth factor II: Characterization of the antibody response and use of a fusion protein. Vaccine Res. 1997;6(4):91–100.

12. González G, Crombet T, Catalá M, Mirabal V, Hérnandez J.C, González Y, Marinello P, Guillén G, Lage A. A novel cancer vaccine composed of human recombinant epidermal growth factor linked to a carrier protein: Report of a pilot clinical trial. Ann Onc. 1996;9:431–435.

13. González G, Crombet T, Torrea F, Catalá M, Alfonso L, Osorio M, Neninger E, García B, Mulet A, Pérez R, Lage A. Epidermal growth factor-based cancer vaccine for non-small-cell lung cancer therapy. Ann Onc. 2003;14:461–466.

14. Crombet T, Neninger E, Catalá M, Garcia B, Leonard I, Gonzalez G, Perez R, Lage A. Treatment of NSCLC patients with an EGF-based cancer vaccine: Report of a phase I trial. Cancer Biol Ther. 2006;5(2):145–149.

15. González G, Lage A. Cancer vaccines for hormone immune deprivation: The EGF vaccine approach. Leading Top Cancer Vacc. 2007; Chapter 11, Ed NOVA Publishers.

16. González G, Lage A. Cancer vaccines for hormone/growth factor immune deprivation: A feasible approach for cancer treatment. Curr Cancer Drug Targets. 2007;7(3):191–201.

17. González G, Crombet T, Neninger E, Viada C, Lage A. Therapeutic vaccination with epidermal growth factor/EGF) in advanced lung cancer: Analysis of pooled data from three clinical trials. Human Vacc. 2007;3(1):8–13.

18. García B, Neninger E, De la Torre A, Leonard I, Martínez R, Viada C, González G, Mazorra Z, Lage A, Crombet T. Effective inhibition of the epidermal growth factor / epidermal growth factor receptor binding by anti-epidermal growth factor antibodies is related to better survival in advanced non-small-cell lung cancer patients treated with the epidermal growth factor cancer vaccine. Clin Cancer Res. 2008;14(3):840–846.

19. Neninger E, De la Torre A, Osorio M, Catalá M, Bravo I, Mendoza M, Abreu D, Acosta S, Rives R, Carrillo C, González M, Viada C, García B, Crombet T, González G, Lage A. Phase II randomized controlled trial of an epidermal growth factor vaccine in advanced non-small-cell-lung cancer. J Clin Onc. 2008;26(9):1452–1458.