Inceptor binds to and directs insulin towards lysosomal degradation in β cells

SC-line generation

The Ethics Committee of the Technical University Munich positively voted on non-commercial research on human iPS cells (219/20 S, 290/20 S). The human iPS cell line HMGUi001-A³³ (here labelled IIR^+/+) has been used to generate the SC lines analysed. Both the IIR^+/+ line and its derivative HMGUi001-A-8 (ref. ³⁴) with a C-PEP-Cherry knock-in (here labelled IIR^+/+; C-PEP-Cherry) were targeted by CRISPR–Cas9 to generate the respective IIR^−/− lines according to a previously published protocol³⁹. In brief, iPS cells were transfected with a targeting vector (pU6-(BbsI)-sgRNAs-CAG-Cas9-Venus-bpA; Addgene no. 86986) encoding four single-guide RNAs (sgRNAs) (Supplementary Table 1) with Lipofectamine Stem Transfection Reagent (Invitrogen) according to the manufacturer’s instructions. Individual clones derived from single transfected, Venus⁺ cells were genotyped by a forward (fwd) primer (ATGGCGACCCGCAGGTGAGC) and a reverse (rev) primer (CTGGCCTCCTCATGACGCCAGAC). A homozygous 50 bp deletion in exon 1 was confirmed by Sanger sequencing from the U6 promoter (Eurofins). Furthermore, karyotyping was performed on colcemid-treated cells in the logarithmic growth phase. A minimum of 20 metaphases per cell line were analysed by the standard G banding technique followed by categorizing using the International System for Human Cytogenic Nomenclature.

iPS cell maintenance and differentiation

iPS cells were maintained in StemMACS iPS-Brew XF culture medium (Miltenyi Biotec) under standard culture conditions (37 °C, 5% CO₂ and 95% humidity) and on Geltrex (Gibco) or Matrigel (Corning)-coated dishes. Cultures were regularly checked for mycoplasma contamination with the MycoAlert PLUS Mycoplasma Detection Kit (Lonza) according to the manufacturer’s instructions. Planar iPS cell cultures were split by detaching the cells with 0.5 mmol l⁻¹ EDTA in PBS and seeding them in StemMACS iPS-Brew XF and Y-27632 (10 µM; SantaCruz), which was changed to StemMACS iPS-Brew XF without Y-27632 after 24 h. To generate aggregate suspension cultures, iPS cells were detached with Accutase (Sigma-Aldrich) and seeded in StemMACS iPS-Brew XF supplemented with Y-27632 to a 30 ml spinner flask (Reprocell) (0.6 to 0.8 × 10⁶ cells per ml), which was placed on an incubator-based magnetic stirrer (Cultistir, Able) set to 60 rpm. The aggregates were split every 3–4 days by dissociating in Accutase and seeding into StemMACS iPS-Brew XF supplemented with Y-27632. Then, 48 h after seeding, the medium was changed to StemMACS iPS-Brew XF without Y-27632. The iPS cell lines were differentiated towards SC islets with a six-stage protocol as previously described (Supplementary Tables 2 and 3)³². At the beginning of S6 (day 19), cells were reseeded to six-well ultra-low binding plates (Corning) at a density of 1 × 10⁶ cells per ml. Plates were placed on an incubator-based rotational shaker (MaxQ 2000 CO₂ Plus; Thermo Fischer Scientific) set to 100 rpm. Experiments were carried out after 3 weeks of S6 culture (day 40). For the analysis of SC islets after long-term S6 culture, SC islets were cultured until day 61. Morphology and C-PEP–Cherry expression were monitored with an EVOS M5000 microscope (Invitrogen).

Human donors

Human islets

Human islets for research were provided by the Alberta Diabetes Institute IsletCore at the University of Alberta in Edmonton, Alberta, Canada (www.bcell.org/adi-isletcore), with the assistance of the Human Organ Procurement and Exchange (HOPE) programme, Trillium Gift of Life Network (TGLN) and other Canadian organ procurement organizations. Islet isolation was approved by the Human Research Ethics Board at the University of Alberta (Pro00013094). All donors’ families gave informed consent for the use of pancreatic tissue in research. The Ethics Committee of the Technical University Munich positively voted on research on human material (557/16 S). The batches used and donor characteristics are summarized in Supplementary Table 4. Islets were derived from two female and two male donors. Donors had a body mass index of between 23.5 and 35.8. The HbA1c was unknown for one donor and ranged from 3.8 to 5.8 for the other three donors.

Human pancreatic tissue

Human foetal pancreatic tissues (11–12 weeks post-conception) were provided by the INSERM HuDeCA Biobank. Maternal written consent was obtained, along with approval from the French agency for biomedical research (Saint-Denis La Plaine, France). Human adult pancreases were removed from adult brain-dead organ donors before cardiac death with prior consent for research use at the Hôpital Saint-Louis (Paris, France)⁴⁰. Human pancreatic tissue was processed in accordance with the French bioethics legislation and INSERM guidelines (French bylaw, published 29 December 1998).

Molecular cloning

The IIR-HaloTag N-terminal fusion and IIR-HaloTag or IIR-Venus C-terminal fusion plasmids were obtained from Genescript as pGenlenti backbone inserts (NCBI reference sequence NM_020775.5). The HaloTag sequence was inserted in the IIR open reading frame between amino acids 43 and 44 for N-terminal tagging. The inserted sequence was flanked at the 5′ end by an EcoRI cutting site, a stop codon and a Kozak sequence, and at the 3′ end by a NheI cutting site. Silent mutations were introduced at amino acids P527 (TTC to TTT) and N815 (AAT to AAC). For C-terminal tagging, the HaloTag was inserted at the 3′ end of the IIR open reading frame. The resulting HaloTag constructs were subcloned into a pCAG vector (Addgene vector no. 79009 modified with an SV40 ori /pA) by EcoRI and NheI digestion.

The pLenti IIR-Venus construct (C-terminally tagged) was synthesized as described above and subcloned into a pLenti6 backbone⁴¹ by BamHI restriction.

Immunostaining and image analysis

Differentiating aggregates were fixed with 4% PFA, dehydrated in a sucrose gradient (10–30% in PBS), frozen in tissue freezing medium (Leica) and cut into 15 µm-thick sections. Sections were rehydrated in PBS, permeabilized with 0.1% Triton-X-100 and blocked with blocking solution (10% FBS, 0.1% BSA, 3% donkey serum and 0.1% Tween-20 in PBS). Primary antibodies were diluted in blocking solution and incubated with the sections at 4 °C overnight. Slides were washed with PBS, incubated with secondary antibodies (1:500) for 4 h and stained with 2 µg ml⁻¹ DAPI in PBS for 30 min. All antibodies are summarized in Supplementary Tables 5 and 6. Slides were mounted in Elvanol (25% glycerol, 10% Mowiol, 2% DABCO, 100 mmol l⁻¹ Tris-HCl pH 8.0) and imaged at the Zeiss LSM 880 confocal microscope with or without Airyscan FAST. Image quantification and processing were performed with Zen Blue (v.3.9; Zeiss) or Fiji (ImageJ v.1.53o)⁴². Colocalization was determined with the Zen Blue (v.3.9) software, and the ‘colocalization coefficient’ expressed in percentage was used. For each multi-Z stack image, the values from images (numbers indicated in figure legends) were averaged and denoted below the corresponding panel.

Human foetal pancreatic sections (5 μm thickness) were prepared and processed from paraffin-embedded tissues as previously described⁴⁰. Nuclei staining was performed with Hoechst 33342 (0.3 mg ml⁻¹, Invitrogen). Images were acquired using a Leica DM 4000B microscope running Wasabi software (v.1.5) (Hamamatsu Photonics). The images were processed using Fiji⁴².

Cell culture

MIN6 K8 cells, a gift from Jun-ichi Miyazaki, were cultured in high-glucose DMEM supplemented with 10% FBS, 50 µmol l⁻¹ β-mercaptoethanol and penicillin (100 units per ml)–streptomycin (100 µg ml⁻¹). INS-1 (C0018007, AddexBio) and INS-1E (C0018009, AddexBio) cells were cultured in RPMI-1640 (Gibco) supplemented with 10% FBS, 50 µmol l⁻¹ β-mercaptoethanol, 1 mmol l⁻¹ sodium pyruvate, 10 mmol l⁻¹ HEPES, 1 mmol l⁻¹ glutamine and penicillin–streptomycin. HEK293 (ATCC CRL-1573) and C2C12 (ATCC CRL-1772) cells were maintained in high-glucose DMEM supplemented with 10% FBS and penicillin–streptomycin.

INS-1 Iir
^−/− generation

INS-1 cells (AddexBio, C0018007) were targeted by CRISPR–Cas9 to generate the INS-1 Iir^−/− cell line according to a previously published protocol³⁹. In brief, INS-1 cells were transfected with two targeting vectors (pU6-(BbsI)-sgRNAs-CAG-Cas9-Venus-bpA; Addgene no. 86986) encoding for sgRNAs directed against exon 2 (GTGACAGCACAGGTTCCAGG) and exon 3 (GTCCTGCAAGCCGTGTGCGG) with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Individual clones derived from single transfected cells were genotyped by three primers (TGTTGATGTTTGATCTCTGATGATGGACCG, CTCCCCAAGTGCTGGGATTAAGGT and GAAGTGCAGTTCTCGGTGGACTCT). The KO was confirmed by the lack of protein expression in immunofluorescence and western blot.

FRET experiments

INS-1E cells were transfected using Lipofectamine 3000 (Thermo Fisher, L3000008) according to the manufacturer’s instructions with a plasmid encoding the inceptor–HaloTag fusion protein. After 48 h, the cells were incubated with the HaloTag ligand tetramethylrhodamine (Promega, G8251; 1:5,000) for 20 min. After extensive washing steps, 100 nmol l⁻¹ of INS-630 and 0, 10, 100 and 1,000 nmol l⁻¹ proinsulin were added to the medium for 40 min before imaging with a Zeiss LSM 880. To calculate the corrected FRET (cFRET) value, crosstalk was subtracted as previously described⁴³; cFRET efficiency was calculated as the ratio of cFRET to donor fluorescence.

Insulin–biotin and INS-630 synthesis

Insulin–azide

Human insulin (100 mg, 17.2 µmol) was dissolved in H₂O and dimethylformamide (H₂O/DMF) (2:1, 3.3 ml) and cooled to 0 °C under argon atmosphere. Triethylamine (48 µl, 17 µmol) was added to adjust the pH to 11. A solution of 2,5‑dioxopyrrolidin‑1‑yl 4‑azidobutanoate (5.45 mg, 24.1 µmol) in a mixture of DMF (496 µl) and 5% aqueous H₂SO₄ (4 µl) was added gradually in portions of 100 µl. Upon the start of the formation of disubstituted product, the reaction was stopped by adjusting the pH to 3 using 1 M aqueous HCl. The product was lyophilized, dissolved in a water and acetonitrile mixture (80:20, 3 ml) and purified by high-performance liquid chromatography (HPLC) using an automated Reveleris PREP Plant system (Buchi) on an RP-peptide column (Jupiter 10 µm Proteo 90 Å, LC column 250 ×30 mm; Phenomenex). This provided the B29-azide-modified insulin as a colourless lyophilizate (45 mg, 7.60 µmol, 45%).

INS-630

BDP630/650-alkyne (1.06 mg, 2.17 µmol) in DMF (100 µl) was added to a solution of azide-modified insulin (10.8 mg, 1.82 µmol) in deionized H₂O (5 ml). A mixture of CuSO₄ (1.14 mg, 4.56 µmol), Tris((1-hydroxy-propyl-1H-1,2,3-triazol-4-yl)methyl)amine (THPTA; 7.93 mg, 18.3 µmol), sodium ascorbate (7.23 mg, 36.5 µmol) and aminoguanidine hydrochloride (4.04 mg, 36.5 µmol) in H₂O (433 µl) was added to the solution. The solution was allowed to stir for 27 h and the product was purified by HPLC (RP-peptide column Jupiter 4 μm Proteo 90 Å, LC column 250 ×10 mm; Phenomenex). After lyophilization, BDP630/650-labelled insulin (3.21 mg, 0.50 µmol, 28 %) was obtained as a bluish solid.

Insulin–alkyne

Human insulin (300 mg, 0.052 mmol) was dissolved in H₂O/DMF (60 ml, 2:1) and cooled to 0 °C under argon atmosphere. Triethylamine (144 μl, 1.039 mmol) was added to the solution to adjust the pH to 10. A solution of hept-6-ynoic acid NHS ester (10.81 mg, 0.052 mmol) in a mixture of DMF (2.79 ml) and 5% aqueous H₂SO₄ (21 μl) was added gradually in portions of 300 μl. After observing the start of formation of the disubstituted product, the reaction was stopped by adjusting the pH of the reaction mixture to 3 using 1 M aqueous HCl. The reaction mixture was lyophilized. The lyophilizate was dissolved in a water and acetonitrile mixture (3 ml, 80:20) and purified by HPLC to yield B29-alkyne-modified insulin.

Biotin-PEG(9)-azide

To a solution of biotin N-hydroxysuccinimide ester (100 mg, 0.293 mmol) in DMF (2.5 ml), a PEG linker (NH₂-PEG(9)-azide) (122 mg, 0.293 mmol) was added. The reaction was stirred for 44 h at 20–25 °C. The solvent was removed under reduced pressure and the residue was diluted in 1 ml dichloromethane (DCM). Upon adding diethyl ether, a colourless solid precipitated, which was filtered off, washed with diethyl ether (three times) and re-dissolved in DCM. DCM was removed under reduced pressure to produce a white solid (0.145 g, 0.218 mmol, 75%).

Insulin–biotin

A solution of biotin-PEG(9)-azide (8.43 mg, 12.7 μmol) in tert-butanol (100 μl) was added to a solution of alkyne-modified insulin (15 mg, 2.54 μmol) in deionized H₂O (1 ml). A mixture of CuSO₄ (1.58 mg, 6.34 μmol), THPTA (11.02 mg, 25.37 μmol), sodium ascorbate (10.05 mg, 50.74 μmol) and aminoguanidine hydrochloride (5.61 mg, 50.74 μmol) in H₂O (603 μl) was added to the solution. The solution was allowed to stir for 41 h and the reaction mixture was purified using Amicon-15 centrifuge filter units (cutoff, 3 kDa). After lyophilization of the residue, the insulin–biotin (4.73 mg, 0.719 μmol, 29 %) was obtained as a white solid.

Pulse-chase live imaging of cell surface inceptor

INS-1E cells were transfected using Lipofectamine 3000 with a plasmid encoding N-terminally tagged inceptor–HaloTag fusion protein. Then, 48 h after transfection, the cells were incubated on ice for 30 min with the cell-impermeable HaloTag ligand Alexa Fluor 488 (Promega, G1001; 1:1,000) and 50 μg ml⁻¹ Human Transferrin CF640R (Biotium, BT00085). The ice-treated cells were released from the cold shock directly in the incubation chamber of a Zeiss LSM 880 (Airyscan FAST mode) and imaged or fixed at the indicated time points. Alternatively, transfected INS-1E cells were incubated for 30 min with the HaloTag ligand Alexa Fluor 488 and SiR-lysosome (Cytoskeleton CY-SC012, 1:1,000) at 37 °C and imaged after washing.

PLA

INS-1 Iir^+/+ or Iir^−/− cells were seeded into microwell plates (Ibidi) and treated the next day with 1 µmol l⁻¹ insulin for 1 h or 1 µg ml⁻¹ isotype control or mAB against the extracellular portion of inceptor overnight. The samples were processed according to the Duolink PLA Orange kit protocol (Sigma-Aldrich, DUO92102). The PLA reaction was performed between endogenous inceptor and proinsulin (Supplementary Table 5), followed by nuclear counterstain (Hoechst 33342; Thermo Fisher Scientific, 62249) and an immunofluorescence staining. Images were acquired using the Airyscan FAST mode on a Zeiss LSM 880, and segmentation was performed using Fiji to identify the nuclei, followed by Voronoi segmentation and quantification of PLA fluorescence intensity (thresholded with the Moments method) per Voronoi-identified cell.

Flow cytometry

Differentiating aggregates were dissociated into single cells by Accutase. For live analysis of fluorescence, the cells were resuspended in PBS and analysed immediately. For fixation, the cells were resuspended in 4% PFA, washed with 5% FBS in PBS, permeabilized and blocked in FACS blocking solution (10% heat-inactivated FBS, 0.1% BSA, 3% donkey serum, 0.1% Tween-20 and 0.2% Triton-X-100 in PBS), incubated with primary antibodies at 4 °C overnight, washed with PBS, incubated with secondary antibodies (1:500) for 1 h at room temperature and washed with PBS. All antibodies were diluted in FACS blocking solution (Supplementary Tables 5 and 6). In negative controls, the primary antibody incubation was replaced by incubation in FACS blocking solution. For analysis, all cells were filtered and loaded onto a BD FACS Aria III flow cytometer. The selected subpopulations were analysed for their percentage or for median intensity by FlowJo (v.10.7.1).

Perifusion assay for dGSIS

The flowthrough of 50–70 SC islets or isolated human islets incubated in different glucose concentrations was collected with the perifusion system PERI-4.2 (Biorep). The samples were pre-treated with Krebs-Ringer Buffer (KRB) (25.8 mmol l⁻¹ NaCl, 0.96 mmol l⁻¹ KCl, 0.24 mmol l⁻¹ KH₂PO₄, 0.24 mmol l⁻¹ MgSO₄, 0.4 mmol l⁻¹ CaCl₂, 4.8 mmol l⁻¹ NaHCO₃, 10 mmol l⁻¹ HEPES pH 7.4) supplemented with 0.1% BSA and 2 mmol l⁻¹ glucose for 1 h and loaded into the perifusion system with a flow rate of 100 µl min⁻¹. Flowthrough was collected in 2 min steps. The samples were perifused with KRB supplemented with 0.1% BSA and 2 mmol l⁻¹ glucose during steps 1–4, 20 mmol l⁻¹ glucose during steps 5–10 and 2 mmol l⁻¹ glucose during steps 11–14. After the last step, the islets were collected and lysed in 0.2% SDS, 0.1 mol l⁻¹ NaCl, 5 mmol l⁻¹ EDTA, 0.1 mg ml⁻¹ Tris-HCl pH 8.0 and 0.5 mg ml⁻¹ proteinase K. DNA was precipitated with isopropanol and washed with ethanol. DNA concentration was measured by a Nanodrop 2000c (Thermo Fisher Scientific) and used for the normalization of insulin content in the flowthrough.

Insulin and proinsulin ELISA

For proinsulin stability assays, SC islets were treated with CHX (100 µg ml⁻¹; Merck) or lysosomal inhibitors pepstatin A (10 μg ml⁻¹; Torcis) and E64d (10 μg ml⁻¹; Torcis) or maintained in growth conditions for 6 h and processed as other untreated samples. The proinsulin secretion assay was performed with SC islets differentiated from SC-derived endocrine progenitors, cryopreserved at day 19 and recovered as previously described⁴⁴. Cells were cultured for 2.5 weeks in standard S6 media and were then transferred in either 5.5 mM or 20 mM glucose containing S6 medium (Supplementary Table 3). After 24 hours, the proinsulin content of the supernatant was measured directly, and the cellular proinsulin content was measured as described below.

SC islets were dissociated into single cells with Accutase. Single cells were counted in a haemocytometer and part of the sample was used to determine the SC β cell percentage by flow cytometry. The remaining sample was sonicated in 50 µl water and lysed in 150 µl of 96% ethanol with 0.18 mol l⁻¹ of HCl. Perifusion assay flowthrough samples, cell culture medium or cell lysates were analysed for insulin and proinsulin content by insulin and proinsulin ELISA (Mercodia 10-1113-01 and 10-1118-01) according to the manufacturer’s instructions. The measured concentrations were normalized to the total sample volume and the DNA content or cell count. For calculation of the proinsulin to insulin ratio, insulin content was converted from µU to nmol by multiplying by a conversion factor of 6.945.

Lentiviral production and transduction

Human 293T (ATCC CRL-3216) cells were maintained in high-glucose DMEM (Gibco) supplemented with 10% FBS and geneticin (2 mg ml⁻¹; Gibco). HEK293T cells cultured without antibiotics were co-transfected with pLenti-Puro6-inceptor-venus (generated as described above) or pLKO.1-puro-CMV-TurboGFP (Sigma-Aldrich), psPAX2 and pMD2.G (gift from Didier Trono; Addgene plasmid nos. 12260 and 12259) with Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions with molar ratios 1:1:1; 16 h later, the medium was changed to Opti-MEM (Gibco) supplemented with glutaMAX (Gibco), 2% FBS and 1 mmol l⁻¹ sodium pyruvate (Gibco). The supernatant was collected 48 h after transfection, filtered through a 45 µm filter. To transduce SC β cells, 500 µl of supernatant was added to the cell suspension in one well of a six-well ultra-low binding plate directly after the reaggregation step on day 19. After 5 days, SC islets were dissociated, fixed and analysed by flow cytometry.

Animal experiments

Animal experiments were carried out in compliance with the German Animal Protection Act and with the approved guidelines of the Society of Laboratory Animals (GV-SOLAS) and the Federation of Laboratory Animal Science Associations (FELASA). The animals were housed under a 12:12 light–dark cycle at 22 ± 1 °C, 45–55% humidity and fed chow diet (Altromin 1314) ad libitum. Inceptor full-body KO mice were generated on a C57BL/6J background as described previously²⁸. For TEM analysis of Iir^+/+ and Iir^−/− murine islets, tissues of 4-month-old male and female mice were used.

scRNA-seq

SC islets were dissociated by TripLE Select Enzyme (Gibco), and single cells, sorted by flow cytometry, were used for scRNA-seq library preparation with a target recovery of 10,000 cells. Libraries were prepared using the Chromium Next GEM Single Cell 3′ Kit (v.3.1) (10× Genomics, PN-1000121) according to the manufacturer’s instructions.

Libraries from IIR^−/−; C-PEP-Cherry and IIR^+/+; C-PEP-Cherry cell lines were pooled and sequenced according to 10× Genomics’ recommendations on an Illumina NovaSeq 6000 system with a target read depth of 50,000 reads per cell. Demultiplexed reads were aligned to the GRCh38 human genome and pre-processed using the CellRanger software (v.3.1.0) (10× Genomics) for downstream analyses. Ambient genes were estimated based on expression in empty droplets using DropletUtils⁴⁵ (v.1.14.2). Genes with an ambient expression score larger than 0.0007 were considered ambiently expressed genes. To filter low-quality cells, droplets with fewer than 1,500 genes and with less than 3,500 or more than 40,000 unique molecular identifier counts were excluded. Cells with a mitochondrial count fraction smaller than 0.025 or greater than 0.4 were excluded. To obtain robust doublet estimates (expected doublet rate of 0.1), we used scrublet⁴⁶ (v.0.2.3), DoubletDetection⁴⁷ (v.4.2), scds⁴⁸ (v.1.10.0), scDblFinder⁴⁹ (v.1.11.4), DoubletFinder⁵⁰ (v.2.0.3) and SOLO⁵¹ (implemented in scvi-tools v.0.17.1)⁵² to detect doublets and counted the number of times any cell was detected as a doublet. Cells consistently detected by four or more methods as a doublet were considered to be doublets and subsequently excluded from further analysis. Gene counts were normalized using sctransform⁵³ (v.0.3.3) with v2 regularization and default parameters in Seurat (v.4.1.1)⁵⁴. Corrected counts were used for subsequent analysis steps and visualizations unless stated otherwise. Samples were integrated using Harmony⁵⁵ implemented in harmonypy (v.0.0.6). The latent spaces produced by Harmony were used for uniform manifold approximation and projection embeddings and Leiden clustering. To ensure optimal integration and embedding of the endocrine cell fraction, integration was repeated on only endocrine cells (that is, cells with high CHGA expression). Clustering and annotation were performed separately on endocrine (high CHGA expression) and non-endocrine cells (low CHGA expression), as with the integration. Final clusters were then annotated using known cell type marker genes (INS, GCG, TPH1, SST, ARX, GAP43, MKI67, EPCAM, VIM).

Gene expression analysis by quantitative real-time PCR

SC islets were dissociated by Accutase on day 40, FAC-sorted for the C-PEP–Cherry signal and lysed in QIAzol lysis reagent (Qiagen). RNA was extracted by the RNeasy Micro Kit (Qiagen). For cDNA preparation, 50 ng RNA was processed by the SuperScript VILO cDNA Synthesis Kit (Invitrogen) according to the manufacturer’s instructions. The qPCR was performed on ViiA 7 Real-Time PCR System (Applied Biosystems), using the TaqMan Universal Master Mix II, no UNG (Applied Biosystems), with TaqMan primers for insulin Hs02741908_m1 and for HPRT Hs02800695_m1 (Thermo Fisher Scientific). Negative ΔCt values were calculated against HPRT.

TEM

SC islets were dissociated by Accutase on day 33, and SC β cells with high C-PEP–Cherry were sorted and re-aggregated as described above. Murine pancreases were isolated from 4-month-old mice and dissected in 4% PFA. Re-aggregated SC β enriched SC islets, human islets or murine pancreases were fixed in 4% PFA in 100 mmol l⁻¹ phosphate buffer (pH 7.4) and stored in 4% PFA or 1% PFA at 4 °C for Epon embedding into epoxy resin and for Tokuyasu cryosectioning, respectively.

For epoxy resin embedding, the samples were further processed according to a modified protocol for serial block face scanning electron microscopy⁵⁶ using osmium tetroxide (OsO₄), thiocarbohydrazide (TCH) and OsO₄ again to generate enhanced membrane contrast^57,58. In brief, samples were post-fixed overnight in modified Karnovsky fixative (2% glutaraldehyde and 2% formaldehyde in 50 mmol l⁻¹ HEPES pH 7.4), post-fixated in a 2% aqueous OsO₄ solution containing 1.5% potassium ferrocyanide and 2 mmol l⁻¹ CaCl₂ (30 min on ice), washed in water, 1% TCH in water, rinsed in water and incubated a second time in 2% OsO₄. Samples were washed and en-bloc contrasted with 1% uranyl acetate (UA), washed in water and dehydrated in a graded series of ethanol in water, followed by three changes in pure ethanol on a molecular sieve. Samples were infiltrated into the Epon substitute EMBed 812 (resin in ethanol mixtures: 1:3, 1:1 and 3:1 for 1 h each, followed by pure resin overnight and for 5 h), embedded into flat embedding moulds and cured at 65 °C overnight. Ultrathin sections (70 nm) were prepared with a Leica UC6 ultramicrotome (Leica Microsystems) and a diamond knife (Diatome), collected on formvar-coated slot grids and then stained with lead citrate⁵⁹ and UA.

For Tokuyasu cryosectioning and immunogold labelling, the samples were processed as previously described^58,60,61. In brief, they were washed in phosphate buffer, infiltrated stepwise into 10% gelatine at 37 °C, cooled down on ice, cut into small blocks, incubated in 2.3 mol l⁻¹ sucrose for 24 h at 4 °C, mounted on pins (Leica, no. 16701950) and plunge-frozen in liquid nitrogen. Then, 70–100 nm sections were cut on a Leica UC6 + FC6 cryo-ultramicrotome (Leica Microsystems) and picked up in methyl cellulose (MC) and sucrose (one part 2% MC, Sigma-Aldrich, M-6385, 25 cP; one part 2.3 mol l⁻¹ sucrose) using a perfect loop. Ultrathin sections were stained with primary antibodies (Supplementary Table 5) for immunogold labelling^58,60. Grids were washed with PBS, 0.1% glycine in PBS, blocked with 1% BSA in PBS and incubated with the primary antibodies for 1 h. For single labelling, the sections were washed in PBS and incubated with bridging antibodies⁶⁰ (1:100), except for the proinsulin antibody, which was detected directly with protein A gold, followed by washes in PBS and incubation with protein A conjugated to 10 nm gold (1:25, UMC, Utrecht) for 1 h. For double labelling, the primary antibodies were applied simultaneously followed by washes in PBS and incubation with a combination of anti-rat 12 nm IgG gold anti-mouse 6 nm IgG gold in 1% BSA in PBS. Immunogold-labelled grids were washed in PBS, post-fixed in 1% glutaraldehyde (5 min), thoroughly washed in water to remove the PBS, contrasted with neutral uranyl oxalate (2% UA in 0.15 mol l⁻¹ oxalic acid, pH 7.0) for 5 min, washed in water and incubated in MC containing 0.4% UA. Grids were looped out with a perfect loop and the MC–UA film was reduced to an even, thin film and air-dried. All sections were analysed on a JEM 1400Plus transmission electron microscope (JEOL) at 80 kV, and images were taken with a Ruby digital camera (JEOL).

TEM image analysis

Immunogold-labelled inceptor and proinsulin densities in TEM images acquired from fixed human islets (24 cells) and mouse pancreatic sections of Iir^+/+ (eight cells) and Iir^−/− (eight cells) animals were analysed using QuPath software (v.0.4.3)⁶². Immunogold particles were detected automatically by a trained pixel classifier. SGs were annotated using a pixel classifier and manual corrections. The density values were obtained by dividing the number of immunogold particles by the area of each organelle. SGs were divided into mSGs (low proinsulin levels, characterized by a halo) or iSG (highly proinsulin levels, enriched without a halo). The automatically detected immunogold labels were quantified in each annotated organelle. Data were analysed and plots were generated in R⁶³ and GraphPad Prism (v.10.1.2).

Co-immunoprecipitation, pulldowns and western blot

Proinsulin and inceptor co-immunoprecipitation

HEK293 cells were transfected with human proinsulin overexpressing plasmid (pTARGET-hProinsulin, gift from Peter Arvan⁶⁴) or C-terminally tagged inceptor–HaloTag with Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. HEK293 or INS-1E cells were lysed in 2% NP-40, 1% Triton-X-100, 10% glycerol and 1% protease inhibitor cocktail (Sigma-Aldrich, P8340) in PBS for 30 min at 4 °C. The lysate was centrifuged for 10 min at 21,000g on a tabletop centrifuge and the supernatant was collected. Then, 20 μl SureBeads Protein G Magnetic Beads (Bio-Rad) were coupled to either an inceptor antibody and a corresponding isotype control (rat IgG2b) or a proinsulin antibody and a control mouse antibody (see Supplementary Table 5). The beads were washed with PBS-T and incubated in PBS-T with 2% BSA for 30 min then washed with lysis buffer. Next, 100 μg lysate (measured by Pierce BCA Protein Assay Kit; Thermo Fisher Scientific) was loaded onto the beads and incubated at 4 °C overnight. Then, the beads were thoroughly washed in PBS-T and the proteins were eluted in 2× Laemmli buffer at 100 °C for 5 min.

Adaptor protein co-immunoprecipitation

INS-1 cells were homogenized in a Potter–Elvehjem homogenizer in 125 mmol l⁻¹ KCl, 1 mmol l⁻¹ EDTA, 50 mmol l⁻¹ sucrose, 20 mmol l⁻¹ HEPES pH 7.4 and 1% protease inhibitor cocktail (Sigma-Aldrich) and then centrifuged at 2,000g for 10 min. The supernatant was used for co-immunoprecipitation with AP1M1, AP2B1 and AP3D1 as described above.

Insulin–biotin-pulldown with purified inceptor ECD-His

Insulin–biotin was synthesized as described above and immobilized on Pierce Streptavidin Magnetic Beads (Thermo Fisher Scientific, 88817). His-tagged inceptor ectodomain (pCDNA3-inceptor ECD-6×His) was expressed and purified from HEK293 cells as previously described²⁹ and added onto the beads in pH-adjusted lysis buffer (2% NP-40, 1% Triton-X-100, 10% glycerol, 1% protease inhibitor in PBS) at 4 °C overnight. The beads were washed with the pH-adjusted lysis buffer and eluted into Laemmli buffer as described above.

Insulin–biotin pulldown from MIN6 cells and C2C12 cells

MIN6 or C2C12 cells were incubated in growth medium or starvation medium (114 mmol l⁻¹ NaCl, 4.7 mmol l⁻¹ KCl, 1.2 mmol l⁻¹ KH₂PO₄, 1.16 mmol l⁻¹ MgSO₄, 2.5 mmol l⁻¹ CaCl₂, 25.5 mmol l⁻¹ NaHCO₃, 20 mmol l⁻¹ HEPES pH 7.2) for 1 h, followed by incubation with 100 nmol l⁻¹ insulin–biotin or 100 nmol l⁻¹ insulin in the growth or starvation media. Cells were lysed with 2% NP-40, 1% Triton-X-100, 10% glycerol and 1% protease inhibitor in PBS for 20 min and centrifuged at 21,000g for 10 min; then 100 μg of the supernatant was loaded onto Pierce Streptavidin Magnetic Beads at 4 °C overnight. The beads were washed with lysis buffer and eluted in Laemmli buffer as described above.

Western blot

Samples were boiled in Laemmli buffer, loaded on a precast 4–20% gradient SDS-polyacrylamide gel (Bio-Rad), separated by electrophoresis and transferred onto a PVDF membrane using a Trans-Blot Turbo (Bio-Rad). The membrane was blocked in 5% milk in TBS-T, incubated with primary antibodies at 4 °C overnight, washed with TBS-T, incubated with HRP-coupled secondary antibodies (1:5,000) for 1 h and then with the Clarity Western ECL substrate (Bio-Rad). Images were acquired with a ChemStudio SA2 (Analytic Jena) and quantified using the Fiji GelAnalyzer (ImageJ v.1.53o)⁴².

In vitro dimerization and binding assays

Purified inceptor ECD-His peptide was equilibrated to different pH in a mass-spectrometry-compatible buffer made of 150 mM ammonium acetate (from a 7.5 mol l⁻¹ stock solution; Sigma-Aldrich) and pH-adjusted to 3, 4, 5, 6 and 7 with formic acid. Then, 25 µg of purified inceptor ECD-His was subjected to buffer exchange in 500 µl spin filters with a 30 kDa cutoff (Vivaspin 500 PES, Sartorius) over five rounds of centrifugation (15 min, 3,000g at 4 °C) and resuspension in fresh buffer. Incubation with 0, 30 and 100 µmol l⁻¹ human insulin (recombinant; Sigma-Aldrich) or bovine proinsulin (Novo BioLabs) was done in 35 µl final volume (6.5 µmol l⁻¹ inceptor) by adding 2 mg ml⁻¹ insulin or proinsulin stock solution dissolved in 0.1% formic acid (~pH 2.7). Dilution of the initial 25 µl volumes with 0.1% formic acid did not alter the desired pH (as tested on larger volumes with a pH meter). After incubation for 30 min at 37 °C, samples were directly measured, either by shotgun mass spectrometry or CDMS, or stored for a few days at 4 °C to test binding stability.

For the differential alkylation experiment, inceptor ECD-His solubilized in PBS at pH 7.2 was alkylated with N-ethylmaleimide (NEM) for 1 h at room temperature in the dark. After buffer exchange with PBS at pH 7.2, disulphide bonds were reduced by the addition of 8 mM dithiothreitol for 30 min at 54 °C, followed by alkylation with 16 mM IAA for 1 h at room temperature in the dark. The mass spectrometry data were recorded on the Fusion Orbitrap platform connected online to an Ultimate 3000 RSLC nano system through an analytical column (Poroshell EC-C18, 2.7 µm, 50 cm × 75 µm). Peptides were separated over a 60 min gradient with standard shotgun settings applied to the mass spectrometry platform. The resulting data were analysed with MaxQuant⁶⁵ (v.2.0.3.0) with standard settings applied and searched against the full human proteome supplemented with the sequence of inceptor. The intensities were extracted from the NEM and carbamidomethyl site-specific tables and a final ratio was calculated. Peptides in which one C was annotated with NEM and another with carbamidomethyl were excluded from the analysis.

For the insulin and proinsulin binding CDMS assays, samples containing inceptor (5 µM) alone or inceptor with insulin at a ratio of 1:6 and 1:20 and at varying pH values (pH 3–7) were incubated at 37 °C for 30 min before analysis. Free and insulin-bound inceptor samples were introduced into an Orbitrap Q Exactive UHMR mass spectrometer with glass capillaries produced in-house, using the following settings: capillary voltage, 1.5-2.0 kV; ion mode, positive; collision gas, nitrogen; in-source trapping voltage, 50–100 V; noise level, 3.64; number of microscans, 1; injection time, 1–10 ms. The HCD direct eV setting was optimized to ensure sufficient desolvation and cooling of ions and to minimize the number of split peaks observed in the raw CDMS data. The resolution was set to 280,000 at 200 Th (1 s transient time) for all measurements. The calibration of the Orbitrap detector was performed using CsI in the m/z range of interest (350–12,500 Th). The acquisition was typically performed for 10–20 min for the recording of 1,000–10,000 single ions. The resulting raw data were processed to remove dephasing ions (detected as split peaks) for accurate charge determination⁶⁶. The intensity value cutoff for noise rejection was set to approximately eight to ten elementary charges. All remaining intensities of centroid peaks were normalized to 1 s injection time, and a factor of 12.55 was applied to convert from intensity to charges based on the previous study. Such conversion results in charge determination accuracy within 1.6 elementary charges, which holds true for up to 250 charges and allows for calculation of the final mass. The amount of insulin bound to the monomer and dimer of inceptor was calculated by subtracting the masses of free inceptor (averaged for all pH values) from the masses of inceptor incubated at 1:6 and 1:20 ratios with insulin.

Preparation and evaluation of inceptor mAB

The purified ectodomain of human inceptor (KIAA1324, Uniprot entry Q6UXG2, residues 1–910) was generated as described previously²⁹. A monoclonal rat antibody was generated by immunization with the purified human inceptor ectodomain and validated by immunostaining⁶⁷. The purified ectodomain and rat antibody were kindly provided by Ünal Coskun. The humanized version of the anti-inceptor ectodomain antibody was generated by Yumab. As a control antibody, palivizumab was used.

SC islets were incubated with humanized inceptor mAB or control antibody throughout S5 and S6. The antibody-containing medium was replaced every other day. The mAB uptake was quantified by flow cytometry and immunofluorescence using an anti-human secondary antibody. For the antibody washout experiment, SC islets were washed and the medium was exchanged for fresh S6 medium without mAB. The SC islets were fixed after 0, 2, 6, 24, 48 and 168 h. Whole fixed SC islets were cryo-sectioned for immunostaining and fixed single cells were stained for flow cytometry analysis for insulin and the human antibody.

Data reporting

No statistical methods were used to pre-determine sample sizes. The experiments were not randomized, and the investigators were not blinded to allocation during experiments and outcome assessment. For the PLA and FRET analysis, the ROUT method was used with the GraphPad Prism software to exclude outliers (Q = 1%).

Statistics and reproducibility

A value of P < 0.05 was considered statistically significant. All statistical tests, sample sizes and their P values are provided in the figure legends. Unless otherwise specified, statistical tests to compare two groups were performed as two-tailed, and the minimum sample size for representative experiments was n = 3. Data distribution was assumed to be normal but this was not formally tested. Unless differently specified, statistics were performed using GraphPad Prism (v.9 or v.10.1.2; GraphPad Software).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s42255-024-01164-y