Transgenic goats producing an improved version of cetuximab in milk

Abstract

Therapeutic monoclonal antibodies (mAbs) represent one of the most important classes of pharmaceutical proteins to treat human diseases. Most are produced in cultured mammalian cells which is expensive, limiting their availability. Goats, striking a good balance between a relatively short generation time and copious milk yield, present an alternative platform for the cost-effective, flexible, large-scale production of therapeutic mAbs. Here, we focused on cetuximab, a mAb against epidermal growth factor receptor, that is commercially produced under the brand name Erbitux and approved for anti-cancer treatments. We generated several transgenic goat lines that produce cetuximab in their milk. Two lines were selected for detailed characterization. Both showed stable genotypes and cetuximab production levels of up to 10 g/L. The mAb could be readily purified and showed improved characteristics compared to Erbitux. The goat-produced cetuximab (gCetuximab) lacked a highly immunogenic epitope that is part of Erbitux. Moreover, it showed enhanced binding to CD16 and increased antibody-dependent cell-dependent cytotoxicity compared to Erbitux. This indicates that these goats produce an improved cetuximab version with the potential for enhanced effectiveness and better safety profile compared to treatments with Erbitux. In addition, our study validates transgenic goats as an excellent platform for large-scale production of therapeutic mAbs.

Keywords: EGFR; biobetter; biosimilar; cetuximab; monoclonal antibody.

Figure 1

gCetuximab expression construct outline and detection. A, Schematic of the transgenes encoding the cetuximab light chain (LC) and heavy chain (HC) under control of the 5’ (dark blue) and 3’ (light blue) gCSN2 regulatory sequences used for the co‐transfection of primary fetal goat cells. The blue boxes depict exons 1, 8, 9 and partial exons 2 and 7 of gCSN2 that flank the LC and HC coding regions. The transgenes also included tandem chicken β‐globin insulators (black arrows) and at the 3’ end of the HC transgene an antibiotic selection marker (Puro) driven by the 3‐phosphoglycerate kinase (PGK) promoter. The position of primers used for PCR genotyping and the Southern hybridization probe (black bar) are indicated. B, Southern analysis of transgenic cell clones. Depicted are the resulting hybridization signals of genomic DNA from a selection of different cell clones probed with a gCSN2 fragment. CN: signal from the endogenous gCSN2 copies; HC: signal of the heavy chain encoding transgene; LC: signal of the light chain encoding transgene; WT: non‐transgenic wild type control; WT+5/10: wild type genomic DNA spiked with the equivalent of 5/10 transgene copies

Figure 2

Milk and gCetuximab production by founder goats from a hormonally induced lactation. A, Daily milk yields for individual founder goats for the medium copy number transgenic lines GN97, GN99, and GN100 (left panel), for the founder goats of the low transgene copy number lines GN304 and GN451 (center panel) and founders of the low transgene copy number line GN388 (right panel). B, gCetuximab expression in milk. Western results of induced milk samples, produced by goats of the indicated transgenic lines, probed with an anti‐human IgG antibody detecting gCetuximab light chain (LC) and heavy chain (HC). hPlasma: milk sample spiked with human plasma; ‐ve control: milk sample, not containing gCetuximab; +ve control: milk sample known to contain gCetuximab; standard: monoclonal antibody at 15 µg/µL

Figure 3

Transgene insertion sites. A, schematic representation of the functional elements of the LC and HC transgenes. 5’CN, 3’CN: gCSN2 5’ and 3’ regulatory sequences; HC, LC: heavy and light chain encoding sequences; puro: puromycin selection marker; black arrowheads: chicken β‐globin insulators. B, schematic outline of the transgene insertions in line GN388. The transgenes at the breakpoints (BP1, BP2) of the endogenous locus are depicted. The corresponding endogenous (black) and transgene sequences (highlighted in dark blue) at the breakpoints are shown below the insertion site outline. x TG copies: cluster of additional transgene copies. C, schematic outline of the transgene insertions in line GN451. The sequence of the junction between endogenous locus and transgene sequences (highlighted in blue) are shown for BP1. Location of PCR primers for amplification of the breakpoint junctions are indicated

Figure 4

Transgene copy number determination in multiple generations. A, Assessment of transgene insertions by Southern analysis. Shown are the results for the founders (F0), two F1 (F1‐1, F1‐2), two F2 (F2‐1, F2‐2), and two F3 (F3‐1, F3‐2) offspring for the transgenic lines GN388 and GN451 as indicated. WT: non‐transgenic wild type control; MW: molecular weight marker; WT+LC: wild type genomic DNA spiked with the equivalent of two LC copies; WT+HC: wild type genomic DNA spiked with the equivalent of two HC copies; CN: signal from the endogenous gCSN2 copies; HC: signal of the heavy chain encoding transgene; LC: signal of the light chain encoding transgene. Quantitation results for the LC and HC transgene copies are summarized in the table below the Southern blot. B, Analysis by ddPCR. Shown are representative 2D plots (channel 1 FAM‐target vs. channel 2 Hex‐reference) for the ddPCR copy number variation assay results for the LC and HC transgenes in lines GN388 and GN451. Gray droplets: double negative (no DNA); blue droplets: LC/HC positive, green droplets: LGB reference positive; orange droplets: double positive for LC/HC and LGB. Tables below the 2D plots summarize the calculated transgene copy numbers for the F0‐F3 generations as indicated

Figure 5

Milk yields of F1 and F2 generation goats. A) Daily natural lactation milk yields for different F1 and F2 goats of line GN388. B) Depicts natural daily milk yields for the F1‐2 goat, line GN451. Please note the different scales for A and B

Figure 6

HPLC analysis of α‐Gal on commercial and gCetuximab

Figure 7

Comparative binding efficiencies of gCetuximab and Erbitux to EGFR and CD16 and their potential for enhanced ADCC. A, Binding efficiency of gCetuximab (black) and Erbitux (red) to EGFR was assessed using flow cytometry by binding serial dilutions of antibody (as indicated) to EGFR‐expressing HTB‐30 cells. Shown are MFI measurements plotted against antibody concentration to determine binding efficiency. B, CD16 binding on Jurkat‐CD16 cells for three different batches of gCetuximab produced by F0 and F1 goats from different transgenic lines (black, as indicated), compared to commercial Erbitux (red). The MFI observed are expressed in percent of CD16 binding, with 100% being the value obtained with the R297 anti‐D mAb (EMABling, LFB; blue) and 0% the value in the presence of Rituxan. C, Anti‐EGFR antibody‐bound target cell mediated IL‐2 secretion from activated Jurkat‐CD16 cells. Shown are the results of two independent measurements of the IL‐2 levels in cell culture supernatants for co‐cultures of Hep‐2 cells incubated with the indicated concentrations of antibodies in the presence of Jurkat‐CD16 cells. Unrelated antibody (blue): mAb that does not recognize EGFR; Erbitux (red): commercial cetuximab; gCetuximab (black): purified from F1‐2, line GN451