Construction and application of a human scFv phage display library based on Cre‑LoxP recombination for anti‑PCSK9 antibody selection

Abstract

A large human natural single‑chain fragment variable (scFv) phage library was constructed based on Cre‑LoxP recombination, and used to successfully identify antibodies against proprotein convertase subtilisin/kexin type 9 (PCSK9). The library was derived from 400 blood samples, 30 bone marrow samples, and 10 cord blood samples from healthy donors. Lymphocytes were isolated from each sample and cDNA was synthesized using reverse transcription‑quantitative PCR. Two‑step overlap PCR was then used for scFv synthesis using a LoxP peptide as the linker. The scFv gene was inserted into the phagemid vector pDF by enzymatic digestion and ligation, and then transformed into Escherichia coli (E. coli) SS320 to establish a primary antibody library in the form of scFvs. A primary antibody library consisting of 5x107 peripheral blood and umbilical cord blood sources, as well as a primary antibody library of 5x107 bone marrow samples were obtained. By optimizing the recombination conditions, the primary phage library was used to infect E. coli BS1365 strain (which expresses the Cre enzyme), and a human scFv recombinant library with a size of 1x1011 was obtained through Cre‑LoxP enzyme‑mediated heavy and light chain replacement and recombination. This constructed recombinant library was employed to screen for antibodies against recombinant PCSK9. After four rounds of selection, a fully human antibody (3D2) was identified with a binding affinity of 1.96±1.56ⅹ10‑10 M towards PCSK9. In vitro, the PCSK9/low‑density lipoprotein receptor (LDLR) pathway of Hep‑G2 cells was inhibited by 3D2 treatment, thereby increasing LDL uptake in these cells. In addition, combination treatment with 3D2 and statin was more effective at increasing LDLR levels than treatment with 3D2 or statin alone. Furthermore, 3D2 resulted in a 3‑fold increase in hepatic LDLR levels, and lowered total serum cholesterol by up to 61.5% in vivo. Taken together, these results suggest that the constructed human Cre‑LoxP scFv phage display library can be used to screen fully human scFv, and that 3D2 may serve as a candidate hypolipidemic therapy.

Keywords: proprotein convertase subtilisin-like/kexin type 9; low-density lipoprotein; receptor; fully human single-chain fragment variable phage display antibody library; Cre-LoxP; hypercholesterolemia.

Figure 1

Diagram of stepwise PCR amplification design. RNA was extracted from lymphocytes acquired from different human samples, and reverse tran-scribed into cDNA, which was used as the template for stepwise PCR. Firstly, VH and VL (Vκ/Vλ) were amplified using the cDNA template. PCR products ranging around 400 bp were recovered and further separately amplified for VH- and VL-linkers with a BssHII/NheI recognition site and LoxP511 sequence. Finally, the VH- and VL-linkers were joined by overlap PCR of scFvs. VH, heavy chain variable region; VL, light chain variable region; scFv, single-chain fragment variable.

Figure 2

PCR amplification of scFvs by stepwise PCR. (A) 1st Round PCR with amplicons ranging around 400 bp of the VH, Vκ and Vλ genes (lane M, DNA marker 2000; lane 1-11, Vκ; lane 12-18, Vλ; lane 19-22, VH). (B-D) 1st Round products were gel-purified and used as the templates for 2nd round PCR. 2nd Round PCR with amplicons ranging around 400 bp of the Vκ-linker, Vλ-linker (B, lane M, DNA marker 2000; lane 1, Vκ-linker; lane 2, Vλ-linker) and VH-linker (C, lane M, DNA marker 2000; lane 1-4, VH-linker) fragments were also gel-purified and used as the templates for overlap PCR of the scFvs (D, Lane M, DNA marker 2000; lane 1, scFv). Amplicons ranging around 800 bp were recovered and stored at -80°C. scFv, single-chain fragment variable; VH, heavy chain variable region; VL, light chain variable region.

Figure 3

(A) After recombination of the library in E. colistrain BS1365, the library size was investigated by inoculating culture plates with 10 µl BS1365 in triplicate, and counting the number of colonies; the library was estimated to be 1×10¹¹ cells/ml. (B) Determination of the accuracy and diversity of the library based on PCR of 60 colonies PCR (Fig. 3B) confirmed that >75% of the sequences were different from each other.

Figure 4

Phage screening of anti-PCSK9 antibodies. (A) After purification, human PCSK9 protein in the cell culture supernatant was separated by SDS-PAGE under reducing conditions. Lane 1 and 3, purified PCSK9 protein; lane 2, protein molecular weight marker. (B) Phagemid vector pDF was used for the con-struction of the scFv display library. (C) Ratio of positive clones during the four rounds of selection. (D) Identification of the binding selectivity of 24 clones detected by phage ELISA. PCSK9, proprotein convertase subtilisin/kexin 9; scFv, single-chain fragment variable.

Figure 5

Generation and characterization of fully anti-PCSK9 antibodies. (A) Flow diagram for the construction of expression plasmids. (B) Reducing and non-reducing SDS-PAGE detection of 3D2 fully anti-PCSK9 antibody (lane 1, non-reducing conditions; lane 3, reducing conditions) and normal human IgG control antibody (lane 2, non-reducing conditions; lane 4, reducing conditions). (C) Western blotting and indirect binding ELISA were performed to assess the specificity of 3D2 against PCSK9. (D) Affinity of 3D2 was determined by biolayer interferometry technology. PCSK9, proprotein convertase subtilisin/kexin 9.

Figure 6

Human 3D2 antibody inhibits PCSK9 function. (A) 3D2 inhibits the binding of PCSK9 to LDLR in a dose-dependent manner, with an IC₅₀ of 2.25±1.23 nM. (B) Western blotting indicated that after 24-48 h of incubation with 1, 3 or 10 µg/ml 3D2, Hep-G2 cell LDLR levels were increased 1.3- to 2.5-fold, compared with those in untreated cells. GAPDH served as the loading control. (C) Hep-G2 cells were treated with mevinolin (0.5 or 10 g/ml) and/or 3D2 (5 or 10 µg/ml) for 48 h, and the levels of LDLR protein in the cell lysates were assessed by western blotting. (D) C57BL/6J mice (n=7 per group) were administered a single i.v. injection of 3D2 antibody (10 mg/kg) and serum TC levels were determined at 24, 48 and 72 h. (E) Hepatic LDLR protein levels at 8, 24, 48 and 72 h post-injection were analysed by western blotting; the fold change was calculated as the ratio of LDLR in the presence or absence of 3D2 antibody for each time point after normalization of the LDLR to GAPDH in each lane. GAPDH served as the loading control. (F) Mice (n=7 per group) were injected with a single dose of 3, 6 or 10 mg/kg 3D2 antibody, and serum TC levels were determined at 3, 6 and 9 days, respectively. Results are expressed as the mean ± SEM. **P<0.05 and ^***P<0.01 vs. control. PCSK9, proprotein convertase subtilisin/kexin 9; LDLR, low-density lipoprotein receptor; TC, total cholesterol.