Interaction of a viral insulin-like peptide with the IGF-1 receptor produces a natural antagonist

Abstract

Lymphocystis disease virus-1 (LCDV-1) and several other Iridoviridae encode viral insulin/IGF-1 like peptides (VILPs) with high homology to human insulin and IGFs. Here we show that while single-chain (sc) and double-chain (dc) LCDV1-VILPs have very low affinity for the insulin receptor, scLCDV1-VILP has high affinity for IGF1R where it can antagonize human IGF-1 signaling, without altering insulin signaling. Consequently, scLCDV1-VILP inhibits IGF-1 induced cell proliferation and growth hormone/IGF-1 induced growth of mice in vivo. Cryo-electron microscopy reveals that scLCDV1-VILP engages IGF1R in a unique manner, inducing changes in IGF1R conformation that led to separation, rather than juxtaposition, of the transmembrane segments and hence inactivation of the receptor. Thus, scLCDV1-VILP is a natural peptide with specific antagonist properties on IGF1R signaling and may provide a new tool to guide development of hormonal analogues to treat cancers or metabolic disorders sensitive to IGF-1 without affecting glucose metabolism.

Fig. 1. Structural homology of LCDV1-VILP with insulin and IGFs.

a Sequence alignment of LCDV1-VILP with the human insulin and IGF-1. Strictly conserved cysteine residues are in red. The yellow square represents the C-domain, while the A- and the B-domains are in blue and green respectively. Cleavage sites composed of dibasic residues are underlined. Conserved amino acids are represented by an asterisk (*), and conservatively substituted residues are noted as a dot (.). b The LCDV1-VILP sequence was aligned with the 10 sequences showing the highest sequence identity. c Three-dimensional representation of insulin (PDB: 3I40), IGF-1 (AlphaFold2 prediction, pLDDT 78.23) and LCDV1-VILP as a single- (sc) and double chain (dc) molecules (AlphaFold2 predictions, pLDDT scores: scLCDV1 = 71.47, dcLCDV1 = 80.58). The A- and B-chains are represented in blue and green, respectively, and the C-peptide is in yellow. Residues involved in binding to site 1 and site 2 of the insulin/IGF-1 receptor are respectively in red and blue. The percentage of conserved binding residues with the sites 1 and 2 are indicated in red and blue, respectively.

Fig. 2. LCDV1-VILP as a single chain peptide is more potent to bind on IGF1R and on IR.

Competition binding assays on human IGF1R (a) and human IR (b) ectodomains. Data are expressed as the mean ± SEM, normalized to the baseline and expressed as % of the maximal binding of ¹²⁵I-IGF-1 (a) and ¹²⁵I-Insulin (b) incubated alone (n = 3 (IGF1R) and 4 (IR) independent wells). Western blot detection of IGF1R, IR, AKT, IRS1, ERK1, and ERK2 phosphorylation in lysates of murine preadipocytes overexpressing the human IGF1R (c) or the human IR (d). Quantitative analysis of phosphorylated IGF1R, IR and AKT. Data, expressed as mean ± SEM (n = 3 independent experiments), were normalized by phosphorylation intensity induced by 10 nM of IGF-1 (e) or insulin (f).

Fig. 3. scLCDV1-VILP is a potent IGF1R antagonist.

Western blot detection of IGF1R, IR, AKT, ERK1 and ERK2 phosphorylation in lysates of murine preadipocyte overexpressing the human IGF1R (a) or the human IR isoform B (b) as described in Methods. Quantitation of phosphorylated IGF1R, IR, AKT and ERK1. Scan intensities were normalized to total protein, and data expressed as mean ± SEM (two-way ANOVA followed by a Šídák’s multiple comparison test; Graphpad Prism V.9; n = 4 (IGF1R) and 6 (IR) independent experiments; data were normalized by phosphorylation intensity induced by 10 nM of IGF-1 (c) or insulin (d).

Fig. 4. scLCDV1-VILP reduces cell proliferation and growth in mice oversecreting bovine growth hormone.

a Murine preadipocytes overexpressing the human IGF1R were plated at a density of 20,000 cells/well then counted after a 48-h incubation with the indicated concentrations of IGF-1, scLCDV1-VILP and dcLCDV1-VILP. Data are expressed as the mean ± SEM (Two-way ANOVA followed by a Dunnett’s multiple comparison test (Graphpad Prism V.9); n = 3 independent experiments). b The number of murine preadipocytes expressing the hIGF1R were counted after a 48-h incubation with various concentrations of scLCDV1-VILP and dcLCDV1-VILP and a fixed concentration of IGF-1 (10 nM). Data are expressed as the mean ± SEM (two-way ANOVA followed by a Šídák’s multiple comparison test (Graphpad Prism V.9); n = 3 independent experiments). c Viable murine preadipocytes overexpressing the hIGF1R were stained with crystal violet after a 48-h incubation with various concentrations of IGF-1, scLCDV1-VILP and dcLCDV1-VILP. Data are express as mean ± SEM (ns, not significant; Kruskal-Wallis test followed by a Dunn’s multiple comparisons test (Graphpad Prism V.9); n = 5 (scLCDV1-VILP) and 6 (dcLCDV1-VILP) independent wells). d Relative expression of the LCDV-1 and IGF-1 mRNA expression in liver samples from mice that received either an AAV control (n = 6 mice) or the AAV LCDV-1 (n = 8 mice). Data are expressed as the mean ± SEM (ns, not significant; two-sided Mann–Whitney test; Graphpad Prism V.9). e Body weight gain at week 2 and 6 following AAV injection. Data are normalized by the basal weight measured at t = 0 and expressed as mean ± SEM (two-sided Mann–Whitney test; Graphpad Prism V.9; AAV control = 6 mice, AAV LCDV-1 = 8 mice). f The increment (delta) of body weight gain was measured weekly for seven weeks following the AAVs injection. Data are express as mean ± SEM (AAV control=6 mice, AAV LCDV-1 = 8 mice). g Fasting blood glucose was assessed 2 and 5 weeks after the AAVs injection. Data are express as mean ± SEM (ns, not significant; two-way ANOVA followed by a Šídák’s multiple comparison test; Graphpad Prism V.9; AAV control=6 mice, AAV LCDV-1 = 8 mice).

Fig. 5. Relative conformations of the apo- and ligand-bound IGF1R ectodomain.

a CryoEM density plus the atomic model of scLCDV-1VILP bound to IGF1R.zip shown in ribbon representation (map contour level 0.149). b CryoEM density associated with scLCDV1-VILP, domain FnIII-1′, αCT´ and domain L1 (map contour level 0.149). c CryoEM structure of IGF1R ectodomain in complex with two copies of scLCDV1-VILP with one receptor monomer in low-resolution surface representation and the other in ribbon representation. Note domains FnIII-3 and FnIII-3´ and upstream regions of the insert domains are positioned based on weak potential density from an initial consensus cryoEM map; these entities are unmodelled in the final deposited structure as they are not discerned in the final potential map. Domains are labeled L1, L2: first and second leucine-rich repeat domains; FnIII-1, -2, -3: first, second and third fibronectin Type III domains; αCT: C-terminal α-helical segment of IGF1R α-chain (prime symbol ´ denotes domains from alternate monomer). d Apo IGF-1R ectodomain crystal structure (PDB: 5U8R). Domain representation and nomenclature is as in panel c. e CryoEM structure of a single IGF-1 bound to IGF1R (PDB: 6PYH). Domain representation and nomenclature as in c. f, g Relative disposition of scLCDV1-VILPs and (h) IGF-1 upon binding to receptor, illustrating (*) the comparative lack of interaction of the L1´-bound scLCDV1-VILP with domain FnIII-1.