Regulatory approval and a first-in-human phase I clinical trial of a monoclonal antibody produced in transgenic tobacco plants.

Although plant biotechnology has been widely investigated for the production of clinical-grade monoclonal antibodies, no antibody products derived from transgenic plants have yet been approved by pharmaceutical regulators for clinical testing. In the Pharma-Planta project, the HIV-neutralizing human monoclonal antibody 2G12 was expressed in transgenic tobacco (Nicotiana tabacum). The scientific, technical and regulatory demands of good manufacturing practice (GMP) were addressed by comprehensive molecular characterization of the transgene locus, confirmation of genetic and phenotypic stability over several generations of transgenic plants, and by establishing standard operating procedures for the creation of a master seed bank, plant cultivation, harvest, initial processing, downstream processing and purification. The project developed specifications for the plant-derived antibody (P2G12) as an active pharmaceutical ingredient (API) based on (i) the guidelines for the manufacture of monoclonal antibodies in cell culture systems; (ii) the draft European Medicines Agency Points to Consider document on quality requirements for APIs produced in transgenic plants; and (iii) de novo guidelines developed with European national regulators. From the resulting process, a GMP manufacturing authorization was issued by the competent authority in Germany for transgenic plant-derived monoclonal antibodies for use in a phase I clinical evaluation. Following preclinical evaluation and ethical approval, a clinical trial application was accepted by the UK national pharmaceutical regulator. A first-in-human, double-blind, placebo-controlled, randomized, dose-escalation phase I safety study of a single vaginal administration of P2G12 was carried out in healthy female subjects. The successful completion of the clinical trial marks a significant milestone in the commercial development of plant-derived pharmaceutical proteins.

Profound structural conservation of chemically cross-linked HIV-1 envelope glycoprotein experimental vaccine antigens.

Chemical cross-linking is used to stabilize protein structures with additional benefits of pathogen and toxin inactivation for vaccine use, but its use has been restricted by the potential for local or global structural distortion. This is of particular importance when the protein in question requires a high degree of structural conservation for inducing a biological outcome such as the elicitation of antibodies to conformationally sensitive epitopes. The HIV-1 envelope glycoprotein (Env) trimer is metastable and shifts between different conformational states, complicating its use as a vaccine antigen. Here we have used the hetero-bifunctional zero-length reagent 1-Ethyl-3-(3-Dimethylaminopropyl)-Carbodiimide (EDC) to cross-link two soluble Env trimers, selected well-folded trimer species using antibody affinity, and transferred this process to good manufacturing practice (GMP) for experimental medicine use. Cross-linking enhanced trimer stability to biophysical and enzyme attack. Cryo-EM analysis revealed that cross-linking retained the overall structure with root-mean-square deviations (RMSDs) between unmodified and cross-linked Env trimers of 0.4-0.5 Å. Despite this negligible distortion of global trimer structure, we identified individual inter-subunit, intra-subunit, and intra-protomer cross-links. Antigenicity and immunogenicity of the trimers were selectively modified by cross-linking, with cross-linked ConS retaining bnAb binding more consistently than ConM. Thus, the EDC cross-linking process improves trimer stability whilst maintaining protein folding, and is readily transferred to GMP, consistent with the more general use of this approach in protein-based vaccine design.

Live attenuated influenza viruses produced in a suspension process with avian AGE1.CR.pIX cells.

BACKGROUND: Current influenza vaccines are trivalent or quadrivalent inactivated split or subunit vaccines administered intramuscularly, or live attenuated influenza vaccines (LAIV) adapted to replicate at temperatures below body temperature and administered intranasally. Both vaccines are considered safe and efficient, but due to differences in specific properties may complement each other to ensure reliable vaccine coverage. By now, licensed LAIV are produced in embryonated chicken eggs. In the near future influenza vaccines for human use will also be available from adherent MDCK or Vero cell cultures, but a scalable suspension process may facilitate production and supply with vaccines. RESULTS: We evaluated the production of cold-adapted human influenza virus strains in the duck suspension cell line AGE1.CR.pIX using a chemically-defined medium. One cold-adapted A (H1N1) and one cold-adapted B virus strain was tested, as well as the reference strain A/PR/8/34 (H1N1). It is shown that a medium exchange is not required for infection and that maximum virus titers are obtained for 1 × 10⁻6 trypsin units per cell. 1 L bioreactor cultivations showed that 4 × 106 cells/mL can be infected without a cell density effect achieving titers of 1 × 108 virions/mL after 24 h. CONCLUSIONS: Overall, this study demonstrates that AGE1.CR.pIX cells support replication of LAIV strains in a chemically-defined medium using a simple process without medium exchanges. Moreover, the process is fast with peak titers obtained 24 h post infection and easily scalable to industrial volumes as neither microcarriers nor medium replacements are required.

Dichotomy in Neutralizing Antibody Induction to Peptide-Conjugated Vaccine in Squalene Emulsion Contrast With Aluminum Hydroxide Formulation.

W614A-3S peptide is a modified 3S motif of the HIV-gp41 (mutation W614A). We previously detected the presence of natural neutralizing antibodies directed against W614A-3S peptide (NAbs) in long-term non-progressor HIV+ patients. Here, we compared the efficacy of W614A-3S peptide formulated in either squalene emulsion (SQE) or in aluminum hydroxide (Alum) in inducing broadly-NAbs (bNAbs). Rabbit and mouse models were used to screen the induction of bNAbs following 4 immunizations. SQE was more efficient than Alum formulation in inducing W614A-3S-specific bNAbs with up to 67%-93% of HIV strains neutralized. We then analyzed the quality of peptide-specific murine B cells by single-cell gene expression by quantitative reverse transcription-PCR and single-cell V(D)J sequencing. We found more frequent germinal center B cells in SQE than in Alum, albeit with a different gene expression profile. The V(D)J sequencing of W614A-3S-specific BCR showed significant differences in BCR sequences and validates the dichotomy between adjuvant formulations. All sixteen BCR sequences which were cloned were specific of peptide. Adjuvant formulations of W614A-3S-peptide-conjugated immunogen impact the quantity and quality of B cell immune responses at both the gene expression level and BCR sequence.

Interplay of diverse adjuvants and nanoparticle presentation of native-like HIV-1 envelope trimers.

The immunogenicity of HIV-1 envelope (Env) trimers is generally poor. We used the clinically relevant ConM SOSIP trimer to compare the ability of different adjuvants (squalene emulsion, ISCOMATRIX, GLA-LSQ, and MPLA liposomes) to support neutralizing antibody (NAb) responses in rabbits. The trimers were administered as free proteins or on nanoparticles. The rank order for the adjuvants was ISCOMATRIX > SE > GLA-LSQ ~ MPLA liposomes > no adjuvant. Stronger NAb responses were elicited when the ConM SOSIP trimers were presented on ferritin nanoparticles. We also found that the GLA-LSQ adjuvant induced an unexpectedly strong antibody response to the ferritin core of the nanoparticles. This "off-target" effect may have compromised its ability to induce the more desired antitrimer antibodies. In summary, both adjuvants and nanoparticle display can improve the magnitude of the antibody response to SOSIP trimers but the best combination of trimer presentation and adjuvant can only be identified experimentally.

Live cold-adapted influenza A vaccine produced in Vero cell line.

The African green monkey kidney (Vero) cell line was used as a substrate for the development of a live cold-adapted (ca) reassortant influenza vaccine. For that purpose, a new master strain was generated by an adaptation of the wild type (wt) A/Singapore/1/57 virus to growth at 25 degrees C in a Vero cell line. The resulting cold-adapted (ca) muster strain A/Singapore/1/57ca showed temperature sensitive (ts) phenotype and was attenuated in animal models and protective in the challenge experiments in ferrets. Two vaccine candidates of influenza A(H1N1) and A(H3N2) subtypes (6/2 reassortants) inheriting six genes coding internal proteins from the new master strain and the surface antigens hemagglutinin (HA) and neuraminidase (NA) from the epidemic viruses were obtained by a standard method of genetic reassortment. All steps of the vaccine preparation were done exclusively in Vero cells, including the isolation of the epidemic viruses. Both vaccine strains were used for immunization of young adult volunteers in a limited clinical trial and appeared to be safe, well tolerated and immunogenic after intranasal administration.

Ch14.18 antibody produced in CHO cells in relapsed or refractory Stage 4 neuroblastoma patients: a SIOPEN Phase 1 study.

PURPOSE: This study aimed to assess the safety, pharmacokinetic and activity profiles of the human-mouse chimeric monoclonal anti-disialoganglioside GD2 antibody ch14.18 produced in Chinese hamster ovary (CHO) cells (ch14.18/CHO). METHODS: Sixteen children with recurrent/refractory neuroblastoma (median age 7.6 y) were enrolled in this Phase 1 dose-finding study. Patients received ch14.18/CHO courses of 10, 20 or 30 mg/m (2)/day as an eight-hour infusion over five consecutive days. Three courses at the same dose level were allowed unless disease progressed. Clearance and biodistribution of radiolabelled ch14.18/CHO in Balb/c and A/J mice were analyzed. RESULTS: A total of 41 ch14.18/CHO courses were given (10 × 3 courses, 5 × 2 courses, 1 × 1 course). Side effects were similar in expectedness, frequency and magnitude to those reported for ch14.18/SP2/0. The dose level of 20 mg/m(2)/day was confirmed. Toxicity was reversible and no treatment-related deaths occurred. In children, the peak plasma concentration was 16.51 µg/ml ± 5.9 µg/ml and the half-life was 76.91 h ± 52.5 h. A partial response following ch14.18/CHO was observed in 2/7 patients with residual disease. In mice, the half-lives were 22.7 h ± 1.9h for ch14.18/CHO and 25.0 h ± 1.9 h for ch14.18/SP2/0. The biodistribution of (125)I-ch14.18/CHO in mice with neuroblastoma was identical to (125)I-ch14.18/SP2/0, indicating GD 2 targeting activity in vivo. Ch14.18 produced in CHO cells showed an unchanged toxicity profile and pharmacokinetics in neuroblastoma patients compared with ch14.18 produced in SP2/0 cells, and evidence of clinical activity was observed. In mice, analysis of pharmacokinetics and biodistribution showed comparable results between ch14.18/CHO and ch14.18/SP2/0. Based on these results, ch14.18/CHO was accepted for prospective clinical evaluation.

Sustained release of proteins from a modified vaginal ring device.

A new vaginal ring technology, the insert vaginal ring (InVR), is presented. The InVR overcomes the current shortfall of conventional vaginal rings (VRs) that are generally ineffectual for the delivery of hydrophilic and/or macromolecular actives, including peptides, proteins and antibodies, due to their poor permeation characteristics in the hydrophobic polymeric elastomers from which VRs are usually fabricated. Release of the model protein BSA from a variety of insert matrices for the InVR is demonstrated, including modified silicone rods, directly compressed tablets and lyophilised gels, which collectively provided controlled release profiles from several hours to beyond 4 weeks. Furthermore, the InVR was shown to deliver over 1 mg of the monoclonal antibody 2F5 from a single device, offering a potential means of protecting women against the transmission of HIV.

Distinct host range of influenza H3N2 virus isolates in Vero and MDCK cells is determined by cell specific glycosylation pattern.

Influenza A viruses were isolated in Vero, MDCK cells and chicken embryos. In contrast to MDCK-derived variants all H3N2 isolates obtained in Vero cells neither agglutinated chicken erythrocytes nor grew in chicken eggs. These host range differences of H3N2 Vero and MDCK isolates were noticed even in the absence of amino acid substitutions in the HA1 molecule. Evaluation of HA glycosylation pattern by treatment with endoglycosidases H and F revealed that Vero-variants contained more oligosaccharides of the high mannose type than did the corresponding MDCK-isolates. Removal of some mannose residues from the non-reducing termini of the carbohydrates by exomannosidase treatment resulted in the ability of Vero-isolates to agglutinate chicken erythrocytes. Glycosylation pattern and properties of H3N2 viruses grown in Vero cells were close to those of viruses grown in human kidney epithelial cells, whereas the H1N1 variants isolated from Vero, MDCK cells or eggs did not differ in agglutination properties, carbohydrate composition or ability to infect eggs.

Receptor-binding properties of modern human influenza viruses primarily isolated in Vero and MDCK cells and chicken embryonated eggs.

To study the receptor specificity of modern human influenza H1N1 and H3N2 viruses, the analogs of natural receptors, namely sialyloligosaccharides conjugated with high molecular weight (about 1500 kDa) polyacrylamide as biotinylated and label-free probes, have been used. Viruses isolated from clinical specimens were grown in African green monkey kidney (Vero) or Madin-Darby canine kidney (MDCK) cells and chicken embryonated eggs. All Vero-derived viruses had hemagglutinin (HA) sequences indistinguishable from original viruses present in clinical samples, but HAs of three of seven tested MDCK-derived isolates had one or two amino acid substitutions. Despite these host-dependent mutations and differences in the structure of HA molecules of individual strains, all studied Vero- and MDCK-isolated viruses bound to Neu5Ac alpha2-6Galbeta1-4GlcNAc (6'SLN) essentially stronger than to Neu5Acalpha2-6Galbeta1-4Glc (6'SL). Such receptor-binding specificity has been typical for earlier isolated H1N1 human influenza viruses, but there is a new property of H3N2 viruses that has been circulating in the human population during recent years. Propagation of human viruses in chicken embryonated eggs resulted in a selection of variants with amino acid substitutions near the HA receptor-binding site, namely Gln226Arg or Asp225Gly for H1N1 viruses and Leu194Ile and Arg220Ser for H3N2 viruses. These HA mutations disturb the observed strict 6'SLN specificity of recent human influenza viruses.