IGF2R: a new player in the IGF2 loop sustaining adrenocortical carcinogenesis

Abstract

Background

Most of adrenocortical carcinomas (ACC) are characterized by IGF2 overexpression; therefore, several studies have focused on its two oncogenic effectors, the tyrosine-kinase receptors IGF1R and IR. However, the specific IGF2 receptor, IGF2R, due to its scavenging activity, was considered a tumour suppressor and has never been fully investigated in this context. Nevertheless, recent evidence from other tumours identified IGF2R as a pro-tumorigenic actor able to exert its function through the downstream activation of the pro-mitotic sphingosine kinases (SphK) enzymes responsible for sphingosine phosphorylation. Hence, the main aims of this study were to elucidate the role of IGF2R in ACC cells, investigate its action mechanism, and test IGF2R and SphK inhibitors as possible novel therapies for ACC.

Methods

The present study was conducted in vitro in 4 different ACC cell lines and in 3 ACC primary cultures, to reflect the heterogeneity of ACC. Target protein and transcript expression were evaluated on normal and tumoral adrenal tissues and cell lines, while the effects of IGF2R downregulation and overexpression, along with IGF2R and SphK inhibition, were tested on cell viability, proliferation, and apoptosis as well as cortisol secretion.

Results

ACC tissue analysis demonstrated the IGF2R overexpression compared to normal adrenal cortex, which was also positively correlated with IGF2 and IGF1R expression in ACC. In vitro assays proved the pro-mitotic involvement of IGF2R and its role in the downstream activation of SphK. Finally, we observed the anti-tumoral efficacy of three different inhibitors of this pathway on ACC cell lines and primary cultures: a specific neutralizing anti-IGF2R antibody was tested for its anti-proliferative action, while two SphK inhibitors, safingol (SAF) and fingolimod (FTY), were able to decrease cell proliferation, viability, and cortisol secretion, while promoting cell apoptosis.

Conclusions

Overall, this comprehensive in vitro evaluation of the role of IGF2R in ACC demonstrates its tumorigenic effect through SphK activation. Moreover, the three pharmacological strategies here employed successfully controlled in vitro ACC growth, suggesting IGF2R/SphK pathway represents a novel therapeutic target for ACC.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12964-026-03057-x.

Introduction

Adrenocortical carcinomas (ACC), rare endocrine tumours with poor prognosis, are characterized by insulin-like growth factor 2 (IGF2) overexpression in more than 90% of cases [1, 2], leading to the activation of an autocrine mitogenic loop that promotes tumour progression. This molecular peculiarity has brought interest on the whole IGF2 system, aiming to identify novel therapeutic targets. Among the system, two tyrosine kinase receptors, the IGF1 receptor (IGF1R) and the insulin receptor (IR), mainly in its mitogenic isoform IR-A, are able to bind to IGF2 and activate a downstream signalling cascade leading to enhanced survival and proliferation (Fig. 1) [3, 4]. Given the well-established molecular role of IGF1R and IR, new therapeutic strategies, aimed at inhibiting the receptors, emerged as successful in several preclinical studies [5, 6]. These findings prompted the initiation of a phase III clinical trial evaluating Linsitinib (OSI-906), a promising dual IGF1R/IR inhibitor, in patients with advanced and metastatic ACC. However, the trial failed to demonstrate a survival benefit compared to placebo [7], suggesting that new strategies are required to modulate the complex IGF system, thus prevent cancer cells growth.

Fig. 1.

Canonical IGF2 pathway in ACC, alongside supposed IGF2R-SphK interaction. Classically, upon IGF2 binding, IGF1R and IR-A, in homo or heterodimers, are activated for the transduction of a signalling cascade leading towards cell proliferation (MAPK activation) and survival (PI3K/Akt activation). Conversely, the binding of IGF2 to its receptor, IGF2R, is followed by the internalisation of IGF2 in endocytic vessels and to its further lysosomal degradation, coupled with the recycling of the receptor on cell membrane. The novel mechanism here presented is summarized in the box: upon IGF2 binding, IGF2R activates SphK, which phosphorylates sphingosine into S1P, able to stimulate cell proliferation, thus tumour growth, via the activation of S1P receptors. IGF2: insulin-like growth factor 2, IGF2R: IGF2 receptor; IGF1R: IGF1 receptor; IR-A: insulin receptor isoform A; SphK: sphingosine kinase; Sph: sphingosine; S1P: sphingosine-1-phosphate; S1P_1-5: S1P receptors 1–5

Among the less extensively investigated components of the IGF system, a very specific IGF2 receptor is the mannose 6-phosphate/IGF2 receptor (M6P/IGF2R), a multifunctional single-transmembrane glycoprotein with no intrinsic tyrosine kinase activity, structurally and functionally completely different from the other IGF receptors [8–10]. IGF2R functions as a scavenger receptor able to internalise IGF2 and to promote its lysosomal degradation (Fig. 1) [8]. Its role in IGF2 clearance, demonstrated in other cell models [11], together with the observation that loss of heterozygosity (LOH) at the IGF2R locus is a frequent event in ACC, has suggested a tumour suppressor function for IGF2R [12]. However, a high IGF2R transcript and protein expression was found in ACC [5, 13], suggesting an oncogenic role for IGF2R in this tumour type. Previous studies demonstrated a pro-mitotic role of IGF2R: in preclinical models of haemangioma, IGF2R knockdown determined a reduction in cell proliferative potential and apoptotic increase [14]; similarly, in cervical cancer models, IGF2R knockdown reduced cell viability, and its overexpression was correlated with reduced overall survival (OS) in patients [15]. Due to the previous evidences, a proliferative pathway downstream IGF2R was hypothesised and successfully identified for the first time by the group of Luttrell [16–18]: IGF2R, upon IGF2 binding, is able to recruit to the plasma membrane two isoforms of sphingosine kinase (SphK1 and 2), an enzyme deputed to the phosphorylation of the anti-tumoral sphingosine into the pro-survival sphingosine-1-phosphate (S1P). The newly-generated S1P translocates in the extracellular environment, where it binds to S1P receptors (S1P₁₋₅), 5 G-protein coupled receptors, able to activate the MAPK signalling cascade and ultimately lead to cell proliferation (Fig. 1).

The relevance of an IGF2R-SphK pathway in ACC is supported by observations that SphK1/2 expression is negatively correlated with OS in ACC patients [19], and that SphK1, which is upregulated in ACC compared with adrenocortical adenomas (ACA), correlates with ENSAT stage and tumour size in ACC [20]. Two well-known SphK inhibitors [21–23], fingolimod (FTY720) and safingol (L-threo-dihydrosphingosine), have been tested in preclinical and clinical studies against adrenocortical malignancies: fingolimod, a sphingosine analogue, was tested in a preclinical study on ACC, where it was proved able to induce cell apoptosis and control NCI-H295R-derived xenografts growth [20]. Safingol, a strong competitive inhibitor of both SphK, was successful in controlling in vitro tumoral growth and it synergised with several chemotherapeutics, thus it was used in two different phase I clinical trials on solid tumours: in one study, safingol determined limited treatment efficacy [24], however, in the other trial, the combination with cisplatin was able to reduce two distant metastases in one ACC patient and to stabilise tumour progression in another one [25].

The aim of this paper is to study the role of the IGF2R in ACC and to identify novel therapeutic strategies targeting the IGF2R-SphK pathway.

Materials and methods

Cell lines and primary cultures

NCI-H295R primary human adrenocortical carcinoma cells obtained from American Type Culture Collection [Cat #CRL-2128, RRID: CVCL_0458, ATCC, VA, US] were harvested in DMEM: F12 1:1 [Cat #D8437, Merck KGaA, Darmstadt, DE] media, supplemented with 2.5% Nu-Serum I [Cat #354352, Corning, NY, USA], 1% ITS+Premix [Cat #355100, Corning, NY, US], 100U/mL penicillin and 100 µg/mL streptomycin [Cat #17-602E, Lonza group Ltd, Basel, CH]. JIL-2266 cells, isolated from a primary ACC of an adult female patient were cultured according to an established protocol [26]. Briefly, culture media was composed of Ham’s F12 Nutrient Mix [Cat #21765, Gibco, Waltham, MA, US] 3:1 with pyruvate-enriched DMEM [Cat #41966, Gibco, Waltham, MA, US] and supplemented with 10% Foetal Bovine Serum (FBS) [Cat#10270 Gibco, ThermoFisher Scientific, Waltham, MA, US], 1% penicillin-streptomycin, 5 µg/mL insulin [Cat #I6634, Merck KGaA, Darmstadt, DE], 0.4 µg/mL hydrocortisone [Cat #H-0888, Sigma-Aldrich Merck KGaA, Darmstadt, DE], 8.4 ng/mL cholera toxin [Cat #C9903, Merck KGaA, Darmstadt, DE], 24 µg/mL adenine [Cat #A2786, Sigma-Aldrich Merck KGaA, Darmstadt, DE], and 10 ng/mL epidermal growth factor [Cat #PHG0311, EGF, Invitrogen, Thermo Fisher Scientific, Waltham, MA, US]. Two metastatic ACC lines were cultured: MUC-1 cells [27], obtained from a male patient’s neck metastasis and TVBF-7 cells [28], established from the lymph nodal metastasis of a male patient. Both cell lines were cultured in Advanced DMEM/F12 medium [Cat #12634, Gibco, Waltham, MA, US] with 10% FBS, 1% penicillin-streptomycin and, for TVBF-7 cells, supplemented with 2 mM L-glutamine [Cat #G7513, Merck KGaA, Darmstadt, DE]. All human ACC cell lines were harvested and seeded for experiments only within a range of 20 passages, moreover they were authenticated by genetic profiling with the polymorphic short tandem repeat (STR) technique [Promega, BMR Genomics Cell Profile service, Padova, IT] in May 2024.

This study was approved by Milano Area 2 Ethical Committee (554_2022bis) and the Ethical Committee of CE-AVEC (105/2017/U/Tess), and patients who underwent adrenalectomy gave informed consent to the use of tumour samples and clinical information. 2 oncocytic adrenal cortical neoplasms of uncertain malignant potential (AON-UMP) and 3 ACC tissues were dissociated to establish primary cultures; among the 3 ACC patients, 2 were males and 1 was a female. The two male patients, of age 54 and 57 years, had tumour dimensions of 7.5 and 4.5 cm respectively. The female patient, 30 years of age, bore a 20 cm mass; both AON-UMP patients were females, their age was 49 and 64 years and tumour dimensions were 6.5 and 7.5 cm, respectively. Both mechanical and enzymatic dissociations (2 h incubation in a 2 mg/mL collagenase IA solution in DMEM) were performed on fresh tumour tissues. Cell suspension was filtered with a 100 μm pores cell strainer to eliminate aggregates and undigested material, then cells were grown into Advanced DMEM-F12 supplemented with 20% FBS, 2 mM L-glutamine, and 1% penicillin-streptomycin. Contamination with non-tumoral cells was minimized: adipocytes were eliminated directly after enzymatic digestion by centrifuging the cell suspension at 1000 rpm for 10 min, and manually removing the resulting lipidic ring. Erythrocytes were removed through repeated washing steps. Moreover, to exclude fibroblasts contamination, primary cultures were daily checked with an optical microscope and experiments were performed in a short time period to prevent fibroblasts overtake.

IGF2R silencing and transfection

The day after seeding, IGF2R genetic silencing was performed using a human SMARTpool of IGF2R predesigned small interfering RNAs (siRNAs) [Cat #L-010601-00-0005, Dharmacon, GE Healthcare Life Sciences, Chicago, IL, US] at concentrations of 25–50 nM for 3–6 days, using Lipofectamine™ RNAiMAX [Cat #13778150, Thermo Fisher Scientific, Waltham, MA, US] as a transfection reagent. A negative control (C-) siRNA (a non-targeting siRNA sequence with no significant homology to any annotated human, mouse, or rat mRNA) was used for each experiment. Only experiments with knockdown efficiency ≥ 60%, evaluated via Western blot analysis, were kept into consideration.

IGF2R transient transfection was achieve though lipofection using the IGF2R-EGFP DNA plasmid [Cat #VB900124-1080jmr, pRP[Exp]-EGFP/Puro-CAG> hIGF2R[NM_000876.4], VectorBuilder Inc., Chicago, IL, US] at 1 µg/µL concentration for 48 h, using Lipofectamine™ 2000 [Cat #11668019, Thermo Fisher Scientific, Waltham, MA, US] as a transfection reagent. The negative control of the experiments was represented by an empty-EGFP plasmid, and the efficiency of transfection was either visually observed with a fluorescence microscope or evaluated via Western blot analysis; only experiments with a transfection efficiency higher than 50% of cells were kept into consideration.

Treatments

Treatments were always performed using serum-deprived media to avoid unspecific effects of exogenous IGF2; moreover, NCI-H295R cells were treated with ITS deprived media to prevent exogenous insulin effects on the IGF2 pathway. Human anti-IGF2R neutralizing antibody [Cat #AF2447, Bio-Techne, Minneapolis, MN, US], directed against the IGF2R-binding site for IGF2, was dissolved in PBS and administered at 4 µg/mL for 24-28-72 h on ACC cell lines and at 20 µg/mL for 72 h on primary cultures in starved medium. Treatment with SphK inhibitors was administered for 24–72 h in starved medium. SAF powder [Cat #ab144070, Abcam Inc., Waltham, MA, US] was dissolved in DMSO at 2.5 mM starting concentration. FTY powder [Cat #ab142065, Abcam Inc., Waltham, MA, US] was dissolved in DMSO at 50 mM starting concentration. Final DMSO concentration was always lower than 0.5% [29] to prevent unspecific effects.

Real time-qPCR

To evaluate the transcript expression of target genes, RT-qPCR was used. The first step, RNA extraction, was performed differently according to the origin sample: RNA from 5 to 50 mg tumour tissues was obtained via Trizol extraction protocol. RNA was then quantified at NanoDrop Lite [Thermo Fisher Scientific, Waltham, MA, US] and 1 ng of each sample was reverse transcribed using the RevertAid H Minus First Strand cDNA Synthesis Kit [Cat #K1632, Thermo Fisher Scientific, Waltham, MA, US]. Upon this step, cDNA was diluted at 1:10 to perform RT-qPCR.

RT-qPCR was performed to obtain RNA expression of IGF2, IGF2R, and GAPDH genes. Ad hoc forward and reverse primers were synthesized by Thermo Fisher Scientific. Comparative ΔCT analysis method was used for retrieving results, normalize on GAPDH expression of each sample, and ultimately quantify RNA expression.

Western blot

Cell pellets were incubated on ice for 10 min in lysis buffer [Cat# 9803 S, Cell Signaling Technology, Danvers, MA, US] with the addition of protease inhibitors [Cat#05892791001 Roche, Basel, CH]. To obtain proteins from frozen tumour, tissues were mechanically dissected while kept on ice, then lysed for 30 min. Homogenates were subsequently centrifuged at 13,000 rpm for 15 min to obtain a protein supernatant. Proteins were then collected and quantified via Pierce BCA assay. 100 µg of proteins from each tissue sample or 40 µg of proteins from cell pellets were loaded on an SDS/polyacrylamide gel and transferred on a nitrocellulose membrane. IGF2R [Cat #14364S, Cell Signaling Technology, Danvers, MA, US], IGF1R [Cat #9750S, Cell Signaling Technology, Danvers, MA, US], SphK1 [Cat #302714, Abcam Inc., Waltham, MA, US], and SphK2 [Cat #32346S, Cell Signaling Technology, Danvers, MA, US] antibodies were diluted 1:1000, 1:1000, 1:500, and 1:1000, respectively. 4 °C overnight incubation of primary antibodies was followed by 1 h room temperature (RT) incubation with anti-mouse or anti-rabbit secondary antibodies [Cell Signaling Technology, Danvers, MA, US] diluted at 1:2000. GAPDH antibody [Cat #AM4300, Ambion, Thermo Fisher Scientific, Waltham, MA, US], diluted at 1:8000, was incubated for 1 h at RT, and assessed as loading control. Chemiluminescence was detected via ChemiDoc™ Imaging System [Bio-Rad, Hercules, CA, US], and densitometrical analysis was performed using NIH ImageJ software. To ensure a proper quantification of ACC, ACA, and NAG tissue protein expression, the same protein sample of NCI-H295R was used as internal technical calibrator across different gels, thus excluding inter-gel variability.

Cell proliferation

Proliferative activity was evaluated via 5-bromo-2-deoxyuridine (BrdU) incorporation assay [Cat#11647229001 Roche, Basel, CH]. Following silencing, transfection, or drug treatments, BrdU was incorporated for 2 h (cell lines) or 24 h (primary cultures) and the kit was performed according to manufacturer’s instructions. VICTOR^® Nivo™ [PerkinElmer, Shelton, US] plate reader set at 450 nm was used to measure the absorbance of samples: results were expressed as mere absorbance values (individual primary cultures) or percentage of the ratio between the mean absorbance of each sample and the relative control.

Cell viability

MTT assay for cell viability was performed after silencing, transfection, or drug treatments. Cells were incubated for 3 h at 37 °C in a 5 mg/mL MTT solution [Cat#475989 Merck KGaA, Darmstadt, DE] diluted 1:10 in DMEM without Phenol Red. MTT was removed once formazan salt formation was verified through microscopy, and DMSO revealed the colorimetric reaction. VICTOR^® Nivo™ plate reader set at 560 nm was used to detect the absorbance of samples: results were expressed as mere absorbance values (individual primary cultures) or percentage of the ratio between the mean absorbance of each sample and the relative control.

Cell apoptosis

Caspase 3/7 activation was assessed as marker of apoptotic processes via ApoONE^® Homogeneous Caspase-3/7 Assay [Cat #, Promega Corporation, Madison, WI, US]. After drug treatments of cells seeded in black plates, a solution 1:100 ApoONE substrate, containing a profluorescent rhodamine, in ApoONE Buffer was added 1:1 to the wells for a 2 h incubation in the dark at RT to allow caspases to cut the ApoONE substrate and generate fluorescence. VICTOR^® Nivo™ plate reader set at 485 nm was used to detect fluorescence: results were expressed as mere fluoresce values (individual primary cultures) or percentage of the ratio between the mean fluorescence of each sample and the relative control.

SphK activity assay

The activity of both SphK was measured through an enzymatic activity assay, able to detect the amount of ATP consumed after sphingosine conversion into S1P. After silencing, transfection of drug treatments, cell pellets were collected and lysed using a specific Mammalian Cell Lysis Kit [Cat #MCL1, Merck KGaA, Darmstadt, DE], then quantified via Pierce BCA assay. The assay was performed following instructions from the manufacturer [Cat #K3500, Echelon Biosciences Inc., Salt Lake City, UT, US]. Sample wells, assayed in duplicate, were prepared by adding 10 µL of sample lysate, 10 µL of ATP solution, and 20 µL of sphingosine solution. Moreover, standard wells were prepared by adding 10 µL of reaction buffer, 10 µL of ATP standard curve, and 20 µL of sphingosine solution. Subsequently, 2 h incubation at RT are needed for the sphingosine kinase in the samples to convert sphingosine into S1P, consuming the supplemented ATP. At last, 40 µL of K-LUMa, an ATP detector, were added to sample and standard wells and incubated for 10 min in the dark prior to reveal its luminescent signal using VICTOR^® Nivo™ plate reader set at 500 ms. The mean signal of each sample was interpolated on the ATP standard curve to calculate the remaining amount of ATP, then, its opposite was calculated to quantify the sphingosine kinase activity.

Cortisol secretion

Cortisol levels were measured on samples of supernatant media of cells after drug treatments. Supernatant samples were centrifuged 10 min at 13,000 rpm at 10 °C to eliminate debris, then measured with the specific immunoassay Elecsys Cortisol II [Cobas, Roche, Basel, CH]. Cortisol levels were expressed as ng/mL in case of individual primary cultures or percentage of the ratio between the mean of each sample and the relative control (cell lines).

Data analysis and statistics

Data from the ACC cohort available on the TCGA were analysed via GEPIA platform to retrieve Spearman correlation analyses. Given the limited number of biological replicates in the in vitro experiments, formal assessment of normality and homogeneity of variance was considered unreliable. Therefore, a conservative nonparametric approach was adopted. For comparisons between two paired experimental conditions, the Wilcoxon matched-pairs signed-rank test was used. For comparisons between two unpaired groups, the Mann–Whitney U test was applied. For analyses involving more than two experimental conditions, the Kruskal–Wallis test followed by appropriate post hoc multiple-comparison testing was used. Statistical analyses were performed using GraphPad Prism Software 10.5.1 [GraphPad Software Inc., San Diego, CA, US], p value < 0.05 was considered statistically significant. Additional details regarding sample sizes, statistical tests, and significance are provided in the figure legends.

Data availability

The data generated in this study are available from the corresponding author upon request.

Results

IGF2R expression in adrenocortical tumours and ACC cell lines

The analysis on frozen tissues derived from adrenalectomies demonstrated a significant overexpression of IGF2R protein level in ACC and ACA, compared to normal adrenal gland (NAG) (Fig. 2A). Interestingly, a strong positive correlation was observed between IGF2R and IGF1R in tumoral tissues (ρ=+0.77 [95% confidence interval (CI): 0.19–0.95] in ACC, *p = 0.019 and ρ=+0.76 [95% CI: 0.10–0.96] in ACA, *p = 0.037), but not in NAG samples at protein level (Fig. 2B, Suppl. Figure 2 C).

Fig. 2.

IGF2R protein expression across different adrenocortical tissues and ACC cell lines. A) Western blot analysis of IGF2R protein level, normalized on GAPDH housekeeping levels, in 12 ACC, 13 ACA, 13 NAG tissues. Results are expressed as individual dots of the IGF2R/GAPDH ratio using NCI-H295R protein expression as inter-gel calibrator, alongside with median and IQR. Representative immunoblots of IGF2R and GAPDH bands are showed alongside. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. **p < 0.01, ***p < 0.001. B) Correlation analysis of IGF2R and IGF1R protein expression in 8 ACC, ACA, and NAG tissues. Statistical analysis of correlation was performed via nonparametric Spearman correlation. C) Correlation analysis of IGF2 and IGF2R transcript expression in our ACC cohort (n = 11) and in TCGA dataset (n = 77). Statistical analysis of correlation was performed via nonparametric Spearman correlation. D) Western blot analysis of IGF2R protein level, normalized on GAPDH housekeeping levels, in NCI-H295R, JIL-2266, MUC-1, and TVBF-7 cells. Results of n > 5 replicates are expressed as median and IQR of the IGF2R/GAPDH ratio. Representative immunoblots of IGF2R and GAPDH bands are showed alongside. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.05, **p < 0.01, ***p < 0.001

The observed IGF2R transcript was similar in the three groups (Suppl. Figure 1 A), and, as expected IGF2 transcript resulted overexpressed in the majority of ACC compared to NAG (Suppl. Figure 1B). Interestingly, IGF2R positively correlated with IGF2 transcript in both our ACC cohort (ρ=+0.65 [95% CI: 0.10–0.89], *p = 0.026) and in the TCGA dataset (ρ=+0.26, *p = 0.02) (Fig. 2C).

Among ACC cell lines, IGF2R protein expression was variable (Fig. 2D): NCI-H295R (median and IQR of IGF2R/GAPDH 0.79 (0.47)), and MUC-1 cells have a comparable IGF2R expression (0.84 (0.74)); IGF2R was significantly overexpressed in TVBF-7 cells (2.04 (1.86), **p = 0.003 vs. NCI-H295R, **p = 0.005 vs. MUC-1, and ***p < 0.001 vs. JIL-2266) and downregulated in JIL-2266 (0.22 (0.41), *p = 0.02 vs. NCI-H295R and MUC-1, ***p < 0.001 vs. TVBF-7) compared to all the other cell lines. This pattern of expression was not associated with derivation of cell line from a primary tumour (NCI-H295R and JIL-2266) or a metastasis (MUC-1 and TVBF-7), sex (NCI-H295R and JIL-2266 derived from a female patient, and MUC-1 and TVBF-7 originated from a male), nor hormonal secretion.

Investigation of the downstream effects of IGF2R downregulation

To investigate IGF2R function in regulating ACC proliferation, we used two different strategies: IGF2R genetic silencing and IGF2R blockade by a neutralizing IGF2R antibody able to bind to the IGF2-binding site on the receptor and prevent further interaction with the endogenous IGF2. Upon IGF2R genetic silencing (Fig. 3A), cellular proliferation, measured as BrdU incorporation, was significantly reduced in all IGF2R-expressing cell lines (NCI-H295R -10.25% (8.41) vs. C- siRNA, ***p < 0.001, MUC-1 -12.55% (17.23), *p = 0.017, TVBF-7 -16.28% (26.42), **p = 0.002). This effect was only cytostatic, as no alteration in cell viability was observed (Suppl. Figure 2 A).

Fig. 3.

Effect of IGF2R knockdown and inhibition on ACC cell lines proliferation. Cell proliferation was measured as BrdU incorporation after genetic silencing or antiIGF2R neutralizing antibody treatment. A) % of cell proliferation after 72 h (NCI-H295R, JIL-2266, and MUC-1) or 144 h (TVBF-7) IGF2R genetic silencing at 25 (NCI-H295R) or 37.5 nM (JIL-2266, MUC-1, and TVBF-7) siRNA concentration. Results of n ≥ 5 experiments were plotted as median and IQR. Representative immunoblots of IGF2R and GAPDH are shown under corresponding graph. Statistical analysis was performed via Mann-Whitney test. *p < 0.05, **p < 0.01, ***p < 0.001 vs. C- siRNA. B) Time course of cell proliferation, expressed as %, after 4 µg/mL antiIGF2R neutralizing antibody treatment in starved medium. Results of n ≥ 4 experiments were plotted as median and IQR. Statistical analysis was performed via Kruskal-Wallis test coupled with Dunn’s post hoc analysis. *p < 0.05 vs. bas

In agreement, as shown in Fig. 3B, incubation with 4 µg/mL of IGF2R neutralizing antibody significantly reduced cell proliferation in all IGF2R-expressing ACC lines (NCI-H295R -28.24% (20.89) vs. bas, *p = 0.041, MUC-1 -25.77% (12.89), *p = 0.012, TVBF-7 -26.44% (17.93), *p = 0.046).

In accordance with their low-to-null IGF2R protein expression, JIL-2266 were not responsive to IGF2R knockdown (Fig. 3A) nor IGF2-site specific inhibition (Fig. 3B) in terms of proliferation rate.

These data were confirmed in 4 primary cultures derived from two ACC (Fig. 4A&C) and two oncocytic adrenal cortical neoplasm of uncertain malignant potential (AON-UMP, Fig. 4B&D) as defined by the Lin-Weiss-Bisceglia criteria for oncocytic adrenal tumours [30, 31]. AON-UMP defines a lesser common category of adrenocortical tumours, featured exclusively by minor criteria of malignancy. In these cell models we observed a tendency to reduction of proliferation (-18.55% and -39.29% respectively, Fig. 4A&B), even though not significant, possibly due to the small number of replicates, in IGF2R-knockdown samples compared to control. Accordingly, antiIGF2R neutralizing antibody was able to reduce cell proliferation in one ACC (Fig. 4C) and one AON-UMP (Fig. 4D) (-28.13%, p = 0.057 and -16.01%, *p = 0.032 respectively).

Fig. 4.

Effect of IGF2R knockdown and inhibition on the proliferation of ACC and AON-UMP primary cultures. Cell proliferation was measured as BrdU incorporation after genetic silencing or antiIGF2R neutralizing antibody treatment. A-B) Cell proliferation after 72 h IGF2R genetic silencing at 50 nM siRNA concentration (silencing efficiency: 73% ACC #2, 61% AON-UMP #2). Results of one experiment (n = 4 technical replicates) were plotted as median and IQR. Representative immunoblots of IGF2R and GAPDH are shown under corresponding graph. C-D) Cell proliferation after 72 h of 20 µg/mL antiIGF2R neutralizing antibody treatment in starved medium. Results of one primary culture (n ≥ 3 technical replicates) were plotted as median and IQR. Statistical analysis was performed via Mann-Whitney test. *p < 0.05 vs. bas

Investigation of the downstream effects of IGF2R upregulation

To test the pro-mitotic role of IGF2R, a transient transfection was pursued in both ACC cell lines (Fig. 5A) and primary cultures (Fig. 5B). Consistently with the former observations, cell proliferation was significantly increased due to IGF2R upregulation (NCI-H295R +27.07% (31.54), ***p < 0.001, JIL-2266 +98.01% (38.68), **p = 0.002, MUC-1 +10.15% (23.55), **p = 0.006, TVBF-7 +58.75% (62.92), ***p < 0.001). Remarkably, JIL-2266 was the cell line most affected by IGF2R overexpression in agreement with their low basal expression of this receptor. Accordingly, cell viability was significantly increased in all cell lines (Suppl. Figure 2B).

Fig. 5.

Effect of IGF2R overexpression on the proliferation of ACC cell lines and primary cultures. Cell proliferation was measured as BrdU incorporation after 48 h transient transfection of IGF2R-EGFP at 1 µg/µL. A) % of cell proliferation in ACC cell lines, after IGF2R-EGFP transfection. Results of n ≥ 6 experiments were plotted as median and IQR. Representative immunoblots of IGF2R and GAPDH are shown under corresponding graph. Statistical analysis was performed via Mann-Whitney test. *p < 0.05, **p < 0.01, ***p < 0.001 vs. EMPTY-EGFP. B) Cell proliferation of two individual ACC primary cultures. Results of n = 5 in ACC #1 and n = 5 in ACC #2 technical replicates were plotted as median and IQR. Representative transfection of EMPTY-IGF2R and IGF2R-EGFP samples is shown alongside the graphs. Statistical analysis was performed via Mann-Whitney test. *p < 0.05, **p < 0.01 vs. EMPTY-EGFP

IGF2R mitogenic effect was also proved in two different ACC primary cultures, where proliferation significantly rose upon IGF2R transient transfection (ACC #1 +194%, **p = 0.008 and ACC #2 +54.1%, *p = 0.029 respectively).

SphK mediate IGF2R oncogenic action

To prove the involvement of SphK1/2 in IGF2R pro-mitogenic effects, we first tested their protein expression in ACC cell lines. SphK1 is significantly overexpressed in JIL-2266 (1.08 (2.93) vs. 0.06 (0.34) TVBF-7, *p = 0.012) and MUC-1 (1.77 (4.63) vs. 0.14 (0.15) NCI-H295R, **p = 0.009; vs. 0.06 (0.34) TVBF-7, ***p < 0.001) cells, while all cell lines display similar levels of SphK2 (Fig. 6A-C).

Fig. 6.

SphK expression and IGF2R-dependant activation in ACC cell lines. A-C) Western blot analysis of SphK1 (A) or SphK2 (B) protein level, normalized on GAPDH housekeeping levels, in NCI-H295R, JIL-2266, MUC-1, and TVBF-7 cells. Results of n = 5 replicates are expressed as median and IQR of the SphK/GAPDH ratio. Representative immunoblots of SphK1, SphK2, and GAPDH bands are showed alongside (C). Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.05, **p < 0.01, ***p < 0.001. D) % of SphK activity after either 144 h IGF2R genetic silencing at 50 nM siRNA concentration or 48 h transient transfection of IGF2R-EGFP at 1 µg/µL in TVBF-7 cells. Results of n = 7 (silencing) or n = 6 (transfection) replicates are expressed as median and IQR. Statistical analysis was performed via Wilcoxon test. *p < 0.05 vs. control. E) % of SphK activity after 72 h of 4 µg/mL antiIGF2R neutralizing antibody treatment in starved medium in MUC-1 cells. Results of n = 6 replicates are expressed as median and IQR. Statistical analysis was performed via Wilcoxon test. *p < 0.05 vs. bas

In order to test the ability of IGF2R in regulating SphK activation, we performed a SphK activity assay upon IGF2R alterations (Fig. 6D). Our data showed that the enzymatic activity of SphK was significantly reduced after IGF2R knockdown (-14.93% (20.4), *p = 0.031) and increased after IGF2R upregulation (+35.76% (70.7), *p = 0.031) in TVBF-7 cells, the most responsive cell line to IGF2R alterations (Figs. 3A and 5A). In agreement, we observed a significant decrease of SphK activity (-65.25% (64.28), *p = 0.031) after incubation with neutralizing IGF2R antibody (Fig. 6E). Overall, these results confirmed the presence of an IGF2-IGF2R-SphK axis and its relevance in ACC cell growth.

Effect of SphK inhibition in ACC in vitro models

We then tested the effects of two different inhibitors of SphK: safingol (SAF), a pan-SphK inhibitor, and fingolimod (FTY), a SphK1-selective inhibitor in ACC cell lines. IC50 related to cell viability was assessed for each drug (Fig. 7A, Suppl. Figure 3 A&B). Interestingly, SphK1, but not SphK2, protein level of ACC cell lines was negatively correlated, even though not significant, with both SAF (ρ= -0.80 95% CI: -0.10-0.73) and FTY IC50 (ρ= -0.80, 95% CI: -0.10-0.73) (Fig. 7B), suggesting that high SphK1 expression levels may predict a better responsiveness to these inhibitors.

Fig. 7.

Functional assays on ACC cell lines. A) Table of IC50 values calculated on cell viability after 72 h of SAF/FTY treatment in starved medium. B) Correlation analysis of IC50 viability values of SAF/FTY with SphK1 protein expression in 4 ACC cell lines. Statistical analysis of correlation was performed via nonparametric Spearman correlation. C-D) Evaluation of cell viability, apoptosis, cortisol secretion, and proliferation after SAF (C) or FTY (D) treatment. Results of n ≥ 4 experiments were plotted as median and IQR. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis (VIABILITY and APOPTOSIS) or via Mann-Whitney test (CORTISOL and PROLIFERATION). *p < 0.05, **p < 0.01, ***p < 0.001 vs. bas

SphK activity assay confirmed the inhibitory effect of these drugs on SphK enzymatic activity in MUC-1 cells (52.87% (52.95) reduction of activity after 72 h SAF incubation), in 3 ACC primary cultures (SAF - 39.66% (37.19), *p = 0.034, and FTY  -22.06% (27.83), *p = 0.049), and in 2 AON-UMP primary cultures (-21.92% (14.98) at 10 µM SAF treatment, Suppl. Figure 3 C-E).

Furthermore, the efficacy of both drugs was evaluated in functional assays on all ACC cell lines, as shown in Fig. 7C and D. Both SAF and FTY were successful in significantly impacting on cell viability, stimulating apoptosis, through caspase-3/7 activation, and hampering the proliferation of all ACC cell lines assessed. Interestingly, cortisol secretion was measured in NCI-H295R, the only ACTH-dependent cortisol-secreting cell line used in the other functional assays. After drug treatment cortisol secretion was significantly hampered (Fig. 7C&D, SAF  -60.9% (20.18), **p = 0.002, and FTY  -85.98% (42.81), **p = 0.002).

To corroborate the previous findings, 2 ACC primary cultures were treated with either drug to evaluate their efficacy (Fig. 8). In ACC #1, SAF and FTY, at 10 µM, displayed a similar significant efficacy in reducing cell proliferation (Fig. 8A), viability (Fig. 8C), and in increasing apoptosis (Fig. 8B). Similarly, ACC #2 cells proliferation was significantly decreased (Fig. 8E) and apoptosis stimulated (Fig. 8F) by either SAF and FTY. Both ACC primary cultures were derived from cortisol-secreting tumours, thus secreted cortisol was assessed upon drug stimulation. Figure 8D shows a marked cortisol reduction, induced by both SAF and FTY on ACC #1, while ACC #2 secreted cortisol was detectable only in control samples (0.63 ng/mL) and undetectable after treatment (< 0.54 ng/mL, data not shown).

Fig. 8.

Functional assays on ACC and AON-UMP primary cultures. A-E-G-I) Cell proliferation, measured as BrdU incorporation, after 72 h of SAF or FTY treatment at 7.5 or 10µM in starved medium of individual ACC or AON-UMP primary cultures. Results of n ≥ 4 technical replicates were plotted as median and IQR. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.05, **p < 0.01, ***p < 0.001 vs. bas. B-F-H-J) Cell apoptosis, quantification of caspase-3/7 activity, after 24 h of SAF or FTY treatment at 7.5 or 10µM in starved medium of individual ACC or AON-UMP primary cultures. Results of n = 5 technical replicates were plotted as median and IQR. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.05, **p < 0.01, ***p < 0.001 vs. bas. C) Cell viability, measured via MTT assay, after 72 h of SAF or FTY treatment at 7.5 or 10µM in starved medium of one individual ACC primary culture. Results of n = 5 technical replicates were plotted as median and IQR. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.05, **p < 0.01, ***p < 0.001 vs. bas. D) Cortisol secretion, measured as ng/mL, after 72 h of SAF or FTY treatment at 7.5 or 10µM in starved medium of one individual ACC primary culture. Results of n = 2 technical replicates were plotted as mean and SD

Similarly, 2 AON-UMP primary cultures were assessed in functional assays with SAF or FTY. AON-UMP #1, due to the low number of available cells, was assessed only for SAF effect, obtaining a significant dose-response reduction of cell proliferation (Fig. 8G) and a corresponding increase of apoptosis (Fig. 8H). AON-UMP #2 showed a significant reduction of cell proliferation after both SAF and FTY treatments (Fig. 8I) and a strong induction of apoptosis, particularly high after SAF 10 µM (Fig. 8J).

Anti-proliferative effects of SphK inhibition are IGF2R-mediated

To test the involvement of IGF2R, we evaluated the anti-mitotic effect of SphK inhibitor SAF upon IGF2R downregulation or overexpression. As shown in Fig. 9A, in NCI-H295R cells, the anti-proliferative action of SAF in basal settings (-36.05% (14.59) vs. basal, *p = 0.021) was abolished upon IGF2R silencing (-23.09% (53.84) vs. IGF2R siRNA). Coherently, JIL-2266 treated with 3 µM SAF displayed a significant drop in cell proliferation only after IGF2R upregulation (-19.25% (29.91) vs. EMPTY-EGFP; -47.17% (14.68) vs. IGF2R-EGFP, **p = 0.009, Fig. 9B), suggesting an increased efficacy of SphK inhibitors when the IGF2R/SphK pro-mitotic pathway is active.

Fig. 9.

The anti-mitotic effect of SphK inhibitors is IGF2R-dependent. Cell proliferation was measured as BrdU incorporation. A) % of cell proliferation after 72 h IGF2R genetic silencing at 50nM siRNA concentration and 72 h 5µM SAF treatment in NCI-H295R cells. Results of n = 4 experiments were plotted as median and IQR, both untreated conditions (C- siRNA and siIGF2R siRNA) were expressed as 100% proliferation and treated samples were normalized accordingly. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.05 vs. C- siRNA. B) % of cell proliferation after 48 h transient transfection of IGF2R-EGFP at 1 µg/µL and 72 h 3µM SAF treatment in JIL-2266 cells. Results of n = 3 experiments were plotted as median and IQR, both untreated conditions (EMPTY-EGFP and IGF2R-EGFP) were expressed as 100% proliferation and treated samples were normalized accordingly. Statistical analysis was performed via Kruskal-Wallis test, coupled with Dunn’s post hoc analysis. *p < 0.01 vs. IGF2R-EGFP

Discussion

In the present study, we demonstrated the pro-mitotic role of the IGF2-specific receptor IGF2R in the context of adrenocortical carcinoma. In our in vitro models, 4 ACC cell lines as well as 3 ACC and 2 AON-UMP primary cell cultures, we described the downstream action mechanism of IGF2R inducing an increase of cell proliferation and viability through the activation of SphK1 and 2. Furthermore, we successfully demonstrated the anti-tumoral properties of a neutralizing antibody against IGF2R, along with two SphK inhibitors, safingol and fingolimod.

Several literature evidences identified the IGF2 autocrine proliferative loop as pivotal for ACC growth. Indeed, more than 90% of ACC are characterized by IGF2 overexpression, that exerts its action transducing a signalling cascade through the two tyrosine kinase receptors, IGF1R and IR [1, 2]. Although the promising preclinical results obtained with IGF1R and IR inhibitors [5, 6], clinical trials have not succeeded in significantly improving patients’ OS [7]. However, little is known about the scavenger receptor IGF2R, which is also involved in this pathway.

Our initial results show that adrenocortical tumours (ACT), including both adenomas and carcinomas, exhibit higher levels of IGF2R compared to NAG, suggesting a potential tumorigenic role for this protein. This result was supported by literature, De Martino et al. found intermediate-to-high intensity staining of IGF2R protein in all IHC samples of their ACC cohort [5]. Conversely to our evaluation, IGF2R overexpression in ACC was retrieved at transcript level by both Velázquez-Fernández et al. and De Martino et al. who compared it versus ACA and NAG, respectively [5, 13].

Interestingly, a positive correlation between IGF2R and IGF2 transcripts was retrieved from both the TCGA and our ACC cohort, and in addition in our ACT patients we found that IGF2R protein levels correlate with those of IGF1R, suggesting that IGF2R could contribute in transducing the pro-tumorigenic signal of IGF2 together with IGF1R.

In agreement with these expression data, in vitro functional assays support an oncogenic role of IGF2R in ACC. Particularly, we assessed BrdU incorporation, as marker of proliferation, as well as cell viability on 4 different cell lines: NCI-H295R and JIL-2266, derived from primary tumours, and MUC-1 and TVBF-7, derived from metastases. Genetic silencing and transient transfection of IGF2R in our in vitro and ex vivo models, 2 ACC and one AON-UMP primary cultures, confirmed that IGF2R retains a proliferative function. This feature is an IGF2 downstream effect, since after the specific inhibition of the IGF2-binding site on the receptor, achieved through an anti-IGF2R antibody, cell proliferation was successfully decreased. Interestingly, JIL-2266 cells, featured by low IGF2R protein levels, were not affected by IGF2R downregulation and inhibition, while after IGF2R overexpression we observed the doubling of cellular proliferation. This result strengthens the hypothesis of an active IGF2R pro-mitotic involvement. However, literature findings are controversial: IGF2R onco-suppressive connotation was largely suggested in ACC [32], as supported by the frequent LOH of its gene [12] and by its IGF2-scavenging function, as demonstrated in other cell models [11, 33]. Nevertheless, our data are corroborated by proteomics analysis in laryngeal cancer and in vitro findings on haemangioma, myeloma, non-small cell lung cancer, and on cervical cancer cells that proved IGF2R as actively implicated in oncogenic mechanisms, with anti-apoptotic, pro-survival, pro-invasive, and pro-mitotic features [14, 15, 34–36].

The relevance of IGF2R in ACC preclinical models was poorly investigated, however, in vitro studies of the Luttrell group using HEK293 cells proved that the proliferative action of IGF2R was exerted through the downstream activation of SphK [16–18]. Indeed, sphingosine kinases, enzymes deputed to sphingosine phosphorylation, contribute to the progression of several tumours by producing S1P and activating MAPK cascade. Our data confirmed the results of Xu et al., who demonstrated SphK1 involvement in ACC tumorigenesis, and, for the first time, we were able to prove a connection between IGF2R and SphK in ACC [20]. Particularly, upon IGF2R genetic silencing and IGF2-binding site inhibition, we observed a reduced SphK activity, and coherently, upon overexpressing IGF2R, SphK activity was enhanced. Similarly, the anti-proliferative action of SphK inhibitors resulted to be IGF2R-dependent, furtherly confirming the IGF2R-SphK interaction.

This novel pathway, represented in Fig. 1 and identified here for the first time in ACC, is particularly interesting in the light of several pharmacological studies on SphK in other pathological conditions. Even though to date no clinical use of SphK inhibitors was approved for cancer, several preclinical and clinical trials were conducted towards various malignancies, as reviewed by Companioni et al. in 2021 [21]. In particular, in our study we selected safingol (SAF) and fingolimod (FTY), the only two SphK inhibitors involved in preclinical or clinical trials on ACC [20, 24, 25]. Our results confirmed the efficacy of either drug on these tumour types; indeed, we were able to successfully hamper in vitro the main hallmarks of tumorigenesis: both SAF and FTY significantly reduced cell proliferation and viability, while promoting cell apoptosis both in primary and in metastatic-deriving ACC cell lines. Interestingly, correlation analyses demonstrated that cell lines featured by higher SphK1 levels, JIL-2266 and MUC-1, were responsive to both SAF and FTY at lower concentrations compared to NCI-H295R and TVBF-7. These results were supported by similar antitumoral effects observed in all ACC and AON-UMP primary cultures tested, a significant impairment of cell proliferation and increase of caspases activation, obtained by each inhibitor treatment. Moreover, cortisol secretion was significantly hampered by both SAF and FTY either in NCI-H295R, the only ACTH-inducible cortisol producing cell line, and in 2 primary ACC cultures; this finding is in line with previous observations reporting the active involvement of sphingolipids in steroidogenesis [37, 38].

The present comprehensive analysis of SAF and FTY efficacy on ACC in vitro models constitutes a preliminary investigation for the future clinical use of these drugs on ACC patients. Furthermore, literature data reported both SphK inhibitors as strong chemosensitizers in other tumours, as they were efficiently paired in several preclinical and clinical studies with doxorubicin, cisplatin, and etoposide [39–41]. Indeed, the combination of either one with one or more components of the EDP-M scheme could result in a beneficial effect for ACC patients.

Despite the robust data presented above, the low number of ACC primary cultures is a downside of this paper. This limitation did not allow to address the heterogeneity of results to possible confounding factors as sex, Weiss score, hormonal hypersecretion, or else. To further corroborate the results obtained here, additional preclinical investigation, possibly using in vivo models, as cell- or patient-derived xenografts, will be needed.

Conclusions

The in vitro results obtained in this paper highlight for the first time the oncogenic connotation of the specific IGF2 receptor IGF2R in the context of ACC. Firstly, we demonstrated the active pro-mitotic involvement of IGF2R in 4 different ACC in vitro cell lines and (several) primary cultures. Moreover, we investigated IGF2R mechanism of action, that exerts its tumorigenic effect through the activation of SphK. Furthermore, we proved the anti-proliferative effect of the neutralizing antiIGF2R antibody and the anti-tumoral action of the two different SphK inhibitors, SAF and FTY. Altogether these data provide the rationale for further preclinical and clinical investigations on IGF2R and sphingosine kinase as therapeutic targets in ACC.

Abbreviations

ACA

Adrenocortical adenoma

ACC

Adrenocortical carcinoma

ACT

Adrenocortical tumour

AON-UMP

Oncocytic adrenal cortical neoplasm of uncertain malignant potential

BrdU

5-bromo-2-deoxyuridine

CI

Confidence interval

FBS

Foetal bovine serum

FTY

Fingolimod, FTY720

IGF1R

Insulin-like growth factor 1 receptor

IGF2

Insulin-like growth factor 2

IGF2R

Mannose 6-phosphate/IGF2 receptor, M6P/IGF2R

IQR

Interquartile range

IR

Insulin receptor

NAG

Normal adrenal gland

OS

Overall survival

SAF

Safingol, L-threo-dihydrosphingosine

SD

Standard deviation

siRNA

Small interfering RNA

SphK

Sphingosine kinase

STR

Short tandem repeat

S1P

Sphingosine-1-phosphate

S1P1-5

S1P receptors

Authors’ contributions

Conceptualization: EN, GM, EP; Data curation: EN, EE, RC, GM, EP; Formal analysis: EN, RC, EP; Funding acquisition: GM, EP, RC; Investigation: EN, EE, SDB, RC, GDM, AMB; Methodology: EN, EE, EC, MT, CH, LSL, EP; Project administration: GM, EP; Resources: EC, SP, AM, VM, MT, CH, LSL, GDD, EF, GM, EP; Supervision: GM, EP; Validation: EN, EP; Writing - Original draft: EN, EE, SDB, RC, EP; Writing – review & editing: EN, EE, RC, EC, SP, AM, VM, MT, CH, LSL, GDD, EF, GM, EP. All authors read and approved the final manuscript.

Funding

This research project was funded by an AIRC (Associazione Italiana Ricerca Cancro) grant to EP (IG 2021–25920), by two Progetti di Ricerca di Interesse Nazionale (PRIN) grants to EP (P20227KXJK and 2022CZR88M), by the MUR under the FIS2 Call for proposal- Fondo Italiano per la Scienza (FIS2-2023-01290, CUP: G53C25000090001) grant to RC, and by Ricerca Corrente Funds from the Italian Ministry of Health to Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico.

Data availability

The data generated in this study are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki and approved by Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico Ethics Committee (Comitato Etico Milano Area 2, protocol code 554_2022bis, date of approval 9 June 2022).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.