Evaluation of anti-insulin receptor antibodies as potential novel therapies for human insulin receptoropathy using cell culture models

Culture media for Chinese hamster ovary (CHO) Flp-In cells (Invitrogen, Carlsbad, CA, USA) and 3T3-L1 pre-adipocytes (Zenbio, Raleigh, NC, USA) are shown in electronic supplementary (ESM) Table 1. Cell lines were all mycoplasma negative by PCR. 3T3-L1 pre-adipocytes were grown to confluence and differentiation was induced by differentiation medium 1 for 72 h then differentiation medium 2 for a further 72 h. Adipocytes were maintained in adipocyte medium containing 1 μmol/l insulin ±1 μg/ml doxycycline (DOX). Experiments were undertaken at day 14 or 16 of differentiation.

Mutation numbering refers to mature hINSR ex11+ (GenBank M1005.1), which was amplified from pDNR-Dual (Clontech, Mountain View, CA, USA) using primers incorporating a C-terminus myc-tag. Sub-cloning is detailed in ESM Table 2. Mutations were generated with the Quickchange II XL kit (Stratagene, La Jolla, CA, USA). CHO Flp-In cells were transfected with pCDNA5/FRT/TO/hINSR and pOG44 using Lipofectamine 2000 (Invitrogen). The population surviving hygromycin B was used for experiments.

Target sequences, primers, vectors and sub-cloning steps are detailed in ESM Table 3. Virus was packaged and concentrated as described by Shin et al [21]. 3T3-L1 pre-adipocytes were infected with the lowest multiplicity of infection (MOI) of virus needed to confer hygromycin B resistance. Several clones per line were characterised for endogenous Insr knockdown and adipocyte differentiation by Oil Red O staining [22]. For hINSR re-expression studies, 3T3-L1 murine Insr-knockdown (MmINSRKD) cells were infected with virus containing myc-tagged hINSR transgenes at the lowest MOI needed to confer G418 resistance to generate polyclonal populations. hINSR expression was confirmed by cDNA sequencing.

CHO Flp-In hINSR cells were blocked by 5% (vol./vol.) FCS/FACS buffer (ESM Table 4) before incubation with primary antibodies for 1 h at 4°C. Bound antibodies were detected using FITC-conjugated anti-mouse IgG and a BD FACSCalibur Flow Cytometer (530 nm/30 nm bandwidth filter, Becton Dickinson, Franklin Lakes, NJ, USA). Stacked histograms were visualised with FCS Express 6 Plus (DeNovo Software, Glendale, CA, USA).

CHO Flp-In hINSR cells were washed twice and serum starved (16 h) before stimulation with insulin, antibody or both for 10 min at 37°C/5% CO₂ and lysed on ice in lysis buffer (ESM Table 4). Receptors were captured overnight at 4°C on anti-myc antibody 9E10-coated white Greiner Lumitrac 600 96 well plates. Phosphotyrosines on immunocaptured receptors were detected with biotin-conjugated 4G10 platinum phospho-tyrosine antibody and europium-labelled streptavidin. DELFIA enhancement solution was added and time-resolved fluorescence measured (excitation 340 nm/emission 615 nm).

3T3-L1 adipocytes were washed twice in DMEM, serum starved for 16 h in DMEM/0.5% BSA/1 μg/ml DOX and treated for 10 min at 37°C/5% CO₂ with 10 nmol/l insulin, 10 nmol/l antibody or both in DMEM/0.5% (wt/vol.) BSA. Cells were washed, snap frozen and lysed on ice before centrifugation twice at 4°C for 15 min to pellet insoluble material and separate lipids prior to western blotting.

Lysate, 10 μg, was resolved on NuPAGE 4–12% bis-tris gels or E-PAGE 48 8% gels (Life Technologies, Carlsbad, CA, USA) and transferred to nitrocellulose by iBlot (Life Technologies). Membranes were blocked in 3% BSA (wt/vol.)/tris-buffered saline with Tween 20 (TBST) before overnight incubation at 4°C with primary antibodies (ESM Table 5). Horseradish peroxidase (HRP)-conjugated secondary antibodies and Immobilon Western Chemiluminescent HRP substrate (Millipore, Darmstadt, Germany) were used to detect protein–antibody complexes, and grey-scale 16 bit tag image file formats (TIFFs) captured with an ImageQuant LAS4000 camera system (GE Healthcare Lifesciences, Marlborough, MA, USA). Each immunoblot in Fig. 4 and ESM Fig. 2 contained a sample of 3T3-L1 MmINSRKD hINSR wild type (WT) treated with 10 nmol/l insulin.

Pixel density of grey-scale 16 bit TIFFs was determined in ImageJ 1.47v (NIH, Bethesda, MD, USA). The rectangle tool was used to select lanes and the line tool to enclose the peak of interest and subtract background. The magic-wand tool was used to select the peak area and obtain the raw densitometry value. Mean band intensities of total INSRβ, myc-tagged INSRβ, Akt, extracellular signal-regulated kinase 1/2 (ERK1/2), glycogen synthase kinase 3 (GSK3)α/β, ribosomal protein S6 kinase β 1 (p70S6K) and calnexin were used to normalise raw densitometry values for p-INSRβ, p-Akt, p-ERK1/2, p-GSK3α, p-p70S6K and p-Akt substrate of 160 kDa (p-AS160). Normalised values for phosphorylated targets were scaled to the mean WT INSR response to insulin.

3T3-L1 adipocytes were washed twice (DMEM), serum starved for 16 h in low-glucose DMEM/0.2% BSA/1 μg/ml DOX, washed twice in PBS and then stimulated for 30 min at 37°C/5% CO₂ with 10 nmol/l insulin, 10 nmol/l antibody or both in KRPH/0.2% BSA buffer (ESM Table 4). Cells were incubated with 1 mmol/l 2-deoxy-d-glucose for 5 min at 37°C/5% CO₂ before washing (PBS), lysing with 0.1 mol/l NaOH, and snap freezing. Glucose uptake was measured by the fluorescence method of Yamamoto et al [23].

One-way ANOVAs with Tukey’s multiple comparisons test were performed with GraphPad Prism 6 (GraphPad software, San Diego, CA, USA). Error bars represent SEM or SD as indicated. All experiments were performed at least three times.

Eleven INSR mutations were selected for study (ESM Table 6). Eight were chosen based on evidence of cell surface expression, prioritising mutations identified in multiple reports to maximise potential availability of participants for future trials. A previously unpublished mutation, F248C, that we identified in a child with Rabson–Mendenhall syndrome, was included opportunistically. The well-studied P1178L tyrosine kinase mutation [24, 25] and the L62P mutation, which severely impairs processing [26], were added as controls. Figure 1a displays the extracellular INSR mutations mapped onto the crystal structure of the INSR [27]. Four mouse monoclonal anti-human INSR antibodies were used, which had all previously been shown to have partial agonist activity at WT receptors, but different effects on kinetics and affinity of insulin binding (Table 1). As Fab fragments of 83-7 and 83-14 were used in determining the crystal structure of insulin-bound INSR [28], their binding epitopes are known (Fig. 1b).

Mutations were introduced into the B isoform of the INSR, believed to be the more important isoform for the metabolic actions of insulin [29]. To enable discrimination of endogenous INSR and human INSR mutants, a C-terminal myc-tag was used in the mutant constructs. Tagged mutants were expressed in CHO cells using the Flp-In system, ensuring differences in protein expression are due to differential processing or stability of receptor protein rather than differential mRNA expression. The mutants were well processed to mature β subunits, with the exception of L62P, for which β subunit was barely detectable. More modest reductions were seen for the previously unstudied F248C and for the P1178L mutation (Fig. 1c).

Cell surface expression and antibody binding of mutant INSR was assessed by flow cytometry (Fig. 1d–g). All INSR antibodies bound each mutant INSR, as shown by right-shifted peaks relative to control IgG, indicating no gross changes in receptor morphology. Poor expression of L62P was in keeping with prior reports [30], and L62P was not studied further. The rightward shift for mutants corresponded to expression of mature β subunits seen by immunoblotting, suggesting that relative shifts reflected differences in receptor expression rather than antibody affinities. Although some mutations are close to the epitope for antibody 83-7, none of the affected residues provides critical antibody contacts. Indeed, no difference in binding of 83-7 (Fig. 1d) to the mutant panel was seen compared with 83-14 (Fig. 1e), which binds to a surface unaffected by the mutations (Fig. 1a, b). Antibody 83-7 demonstrated cross-reactivity with endogenous CHO INSR, as evidenced by positive staining of CHO Flp-In parent cells, while the other antibodies did not detectably cross-react.

Trans-autophosphorylation of tyrosines in the intracellular INSR is the first detectable signalling event after insulin binding, so the ability of insulin and antibodies to induce tyrosine phosphorylation of mutant INSR was next examined using anti-myc immunoprecipitation and europium-based immunoassay. Most mutant receptors (P193L, F248C, R252C, S323L, F382V, D707A, P1178L) demonstrated diminished maximal autophosphorylation response to insulin, ranging from 0 to 27% WT (Fig. 2a–j, Table 2, data not shown for non-responsive P1178L). However, R118C, I119M and K460E showed autophosphorylation comparable with WT, and so were not studied further. Altered insulin EC₅₀ was discernible only for S323L (Table 2), although the insulin concentration range tested and the small magnitude of responses precluded precise determinations.

Antibodies 83-7 and 83-14 alone also elicited autophosphorylation of WT and all mutant INSRs except F248C and P1178L. In most instances, antibody response was lower than insulin response (Fig. 2a–j, Table 2); however, for S323L the maximal autophosphorylation response to 83-7 and 83-14 was similar to that with insulin (Fig. 2g), while D707A was activated by antibodies but not insulin (Fig. 2j).

We next evaluated responses to insulin +10 nmol/l antibody, based on evidence that this concentration elicits the maximal response [18, 20]. In the presence of antibodies 83-7 and 83-14, the maximal response of WT and mutant INSRs to insulin was increased without affecting potency (although EC₅₀ values were not precisely determined) (Fig. 2k–t, Table 2). This was observed across all mutant receptors except the kinase-dead P1178L [24, 25]. Antibodies 18-44 and 18-146 elicited smaller effects than 83-7 and 83-14, and for clarity of presentation data for these antibodies are shown in the ESM (ESM Results, ESM Figs 1, 2, ESM Table 7).

To assess antibody-induced signalling downstream from the INSR, an adipocyte model of insulin receptoropathy was generated. A tetracycline (tet)-responsive microRNA (miR)-short hairpin (sh)RNA selectively targeting murine Insr was transduced into 3T3-L1 pre-adipocytes to generate a stable clone (Fig. 3a). This was transduced with lentiviruses encoding C-terminal myc-tagged WT or mutant hINSR, also controlled by tet-responsive elements (Fig. 3b), generating cells in which DOX simultaneously knocked down endogenous murine Insr and induced overexpression of myc-tagged human INSR. This system permitted pre-adipocyte differentiation uncompromised by mutant receptor expression before induction of Insr knockdown/hINSR re-expression in adipocytes (Fig. 3c, d). The DOX concentration producing maximal Insr knockdown resulted in overexpression of hINSR transgenes (Fig. 3c, e); however, the receptor-processing defects observed in CHO cells (Fig. 1c) were preserved. The C-terminal myc-tag enabled discrimination of endogenous mouse and ectopic human INSR by size shift of the INSR β subunit on immunoblotting, or by anti-myc antibodies (Fig. 3e). The pre-adipocyte cell lines generated differentiated efficiently into mature adipocytes, as evidenced by Oil Red O staining (Fig. 3f).

Plasma insulin concentration in human insulin receptoropathies lies between 0.3 and 3 nmol/l in the fasting state (ESM Table 6), and at least an order of magnitude higher when fed. We used an insulin concentration of 10 nmol/l, mimicking the fed disease state. WT INSR autophosphorylation was strongly induced by insulin (Fig. 4a, b), but was undetectable after receptor knockdown alone (ESM Fig. 2c, d). Otherwise, the pattern of autophosphorylation of overexpressed receptors in response to insulin and/or antibody was similar to that seen in CHO cells. Thus, antibodies 83-7 and 83-14 alone induced WT receptor autophosphorylation on Y1162/Y1163, while antibodies 18-44 and 18-146 were less effective (ESM Fig. 2, ESM Results). Insulin-stimulated autophosphorylation was reduced by 75-100% in mutant INSRs compared with WT. Although antibodies alone induced low-level phosphorylation of mutant INSRs (<10%), the responses of S323L and D707A to antibodies 83-7 and 83-14 were equal to or greater than those to insulin (Fig. 4g–j). Combined insulin and antibody treatment enhanced phosphorylation of each mutant INSR in the case of 83-7 and 83-14, likely synergistically (Fig. 4, ESM Fig. 2).

Akt2/PKBβ transduces metabolic actions of insulin after phosphorylation of T308 and S473. p-Akt2 phosphorylates substrates including glycogen synthase kinase (GSK3α/β), which regulates glycogen synthesis, p70 S6 kinase (p70S6K), which stimulates protein synthesis, and AS160, which encodes a GTPase-activating protein that restrains GLUT4 vesicle translocation until phosphorylated. Insulin treatment of WT INSR induced strong Akt phosphorylation at both sites (Fig. 4a, b), and this was severely attenuated by knockdown of endogenous Insr or by knockdown with re-expression of the kinase-dead P1178L mutant (Fig. 4k, l). Attenuation of signalling was also apparent downstream of Akt, with phosphorylation of p70S6K and GSK3 only modestly impaired, and AS160 phosphorylation unaffected.

Several patterns were seen across the panel of mutants studied. In D707A receptor-expressing cells (Fig. 4i, j), insulin-induced phosphorylation of Akt and its substrates was severely attenuated, while a progressive ‘escape’ from signalling impairment was seen in mutants S323L (Fig. 4g, h), F248C (ESM Fig. 2g, h) and F382V (ESM Fig. 2m, n), with lesser impairment of Akt phosphorylation than of receptor autophosphorylation, and only partial inhibition at downstream substrates. P193L (Fig. 4c, d) and R252C (Fig. 4e, f) demonstrated similar insulin-induced Akt and Akt substrate phosphorylation to WT receptor.

Antibodies alone stimulated Akt and Akt substrate phosphorylation in all cells except those overexpressing the kinase-dead P1178L mutant (Fig. 4k, l). For S323L and D707A mutants, antibodies 83-7 and 83-14 stimulated greater phosphorylation than insulin alone, by virtue of the low response of those mutants to insulin (Fig. 4g–j). Co-treatment of cells with insulin and antibodies 83-7 and 83-14 enhanced Akt and Akt substrate phosphorylation with respect to insulin alone, without evidence of synergy. The additive effects of insulin and antibody co-stimulation were generally less than those observed for receptor autophosphorylation in CHO cells (Fig. 2k–t).

Activation of the INSR by insulin stimulates not only phosphoinositide 3-kinase (PI3K)/Akt, but also RAS/RAF/mitogen-activated protein kinase kinase (MEK)/ERK signalling, through both IRS-dependent and IRS-independent mechanisms [31]. Activation of this pathway is a surrogate for mitogenicity of insulin analogues [32], which is important in view of concerns about long-term cancer risks of analogues with pro-proliferative activity. Insulin treatment of WT INSR induced robust phosphorylation of ERK1/2 at Y204/Y187, with each mutant INSR displaying reduced phosphorylation in response to insulin compared with WT (Fig. 4). Antibody treatment of mutant or WT INSR did not induce ERK1/2 phosphorylation, while dual stimulation with antibody + insulin did not increase ERK1/2 phosphorylation compared with insulin alone. Higher basal ERK1/2 phosphorylation was observed in cells with Insr knockdown alone (ESM Fig. 2c, d), or with Insr knockdown and P1178L receptor overexpression (Fig. 4k, l), but this did not change with any treatment.

Glucose uptake is a key outcome of INSR activation and was assessed in the 3T3-L1 model. Parent 3T3-L1 cells and cells harbouring the Insr-knockdown construct but not treated with DOX displayed similar high levels of insulin-stimulated glucose uptake (ESM Fig. 3a, b), but insulin did not stimulate uptake in conditional Insr-knockdown cells treated with doxycycline (Fig. 5i). Cells with endogenous mouse Insr knockdown and WT hINSR re-expression, in contrast, demonstrated only a 1.8-fold increase in glucose uptake on insulin stimulation (ESM Fig. 3a). The apparently poor response to insulin was due to increased basal glucose uptake in WT receptor-overexpressing cells (ESM Fig. 3b). Basal uptake among mutant receptor-expressing cell lines reflected mutant receptor function (ESM Fig. 3c).

Despite reduced dynamic range in the assay, insulin stimulated glucose uptake via WT, P193L, F248C, R252C and F382V receptors (Fig. 5a, b, d, f, respectively). Insulin-stimulated uptake was similar in cells expressing P193L, R252C or F382V receptor and those expressing WT receptor, but was reduced in cells expressing the F248C mutant. No stimulation of glucose uptake was seen in cells expressing S323L, D707A or P1178L receptors (Fig. 5e, g, h, respectively).

Antibodies 83-7 and 83-14 alone stimulated glucose uptake via P193L, S323L, F382V and D707A receptors (Fig. 5), while antibodies 18-44 and 18-146 were again less effective across the full range of mutants (ESM Fig. 4). While the magnitude of antibody-stimulated uptake was less than that seen with insulin via WT, P193L, F248C, R252C and F382V receptors, antibodies 83-7, 83-14 and 18-44 were more effective than insulin at stimulating glucose uptake via D707A. Dual treatment with antibodies + insulin did not enhance glucose uptake compared with insulin alone acting via WT, P193L, F248C, R252C and F382V receptors, or antibody alone when acting via S323L and D707A receptors.