Differential expression of insulin-like growth factor type 1 receptor identifies heterogeneous intrahepatic regulatory T subsets in mouse hepatocellular carcinoma

Abstract

Understanding regulatory T-cell (Treg)-mediated tumor tolerance is critical for designing immunotherapy against hepatocellular carcinoma (HCC). In this study, we characterized the expression of insulin-like growth factor type 1 receptor (IGF1R) in intrahepatic Tregs in a chemical-induced mouse HCC model. We found two intrahepatic Treg subsets with differential IGF1R expression: IGF1R^hi Tregs and IGF1R^lo/– Tregs. Functional assays indicated that compared with IGF1R^lo/– Tregs, IGF1R^hi Tregs produced more TGF-β and IL-10 and were more proliferative in vivo. Furthermore, IGF1R^hi Tregs exhibited higher phosphorylation of the mammalian target of the rapamycin complex 1 (mTORC1) in vivo. However, in vitro stimulation and immunosuppression assay revealed that the immunosuppressive capacity of the two Treg subsets was equivalent, as evidenced by comparable cytokine production and immunosuppressive effect over conventional T cells. The transcriptome sequencing analysis revealed up-regulation of genes that encode proteins essential for glycolysis, oxidative phosphorylation, and electron transport chain in IGF1R^hi Tregs. Consistently, IGF1R^hi Tregs produces more adenosine triphosphate (ATP), lactate, and reactive oxygen species (ROS). Furthermore, malignant cells in the tumor nodules induced IGF1R down-regulation in Tregs at the mRNA level. In summary, we identified the heterogeneity of intrahepatic Tregs in HCC which might play significant roles in tumor immunity.

The study characterized the expression of IGF1R in intrahepatic Tregs in a chemical-induced mouse HCC model and found two intrahepatic Treg subsets with differential IGF1R expression: IGF1Rhi Tregs and IGF1Rlo/- Tregs. These Treg subsets exhibited distinct cytokine production patterns, proliferation rates, and metabolic activities.

Graphical Abstract

Graphical Abstract.

Introduction

Hepatocellular carcinoma (HCC) is one of the most common malignancies and the third most lethal cancer across the globe [1]. Mobilization of adaptive immunity particularly cellular immunity is a promising strategy to treat HCC progression. However, tumor immunity can be down-regulated or even abolished by multiple mechanisms. Regulatory T cells (Tregs) are believed to be essential for tumor tolerance [2, 3]. Tregs undermine effector cytotoxic T-cell-mediated anti-tumor responses in both cell contact-dependent manner contact-independent manner [2]. Understanding the heterogeneity and plasticity of intrahepatic Tregs is crucial for developing novel HCC immunotherapies.

By engaging with insulin-like growth factor type 1 (IGF1) with high affinity and insulin-like growth factor type 2 (IGF2) and insulin with lower affinity, the insulin-like growth factor type 1 receptor (IGF1R) plays a significant role in the development of tumors [4]. IGFs and IGF1R are also reported to influence Treg functions. IGF1 stimulates the proliferation of Tregs and maintains their suppressive properties in autoimmune diseases, whereas ablation of IGF1R specifically on Treg-cell populations abrogated the beneficial effects of IGF1 on autoimmune diseases [5]. Similarly, IGF1 induces Treg-mediated suppression of allergic contact dermatitis [6] and autoimmune diabetes [7]. IGFs produced by human mesenchymal stem cells induce Treg generation [8]. However, a recent study suggests that IGF1R signaling activates the mammalian target of rapamycin pathway, increases aerobic glycolysis, and favors Th17 cell differentiation over that of Treg cells [9]. These studies suggest that insulin/IGF1 and IGF1R have fundamental roles in the immunosuppressive property of Tregs. However, whether IGF1R signaling impacts Treg activity in HCC remains unknown. Since normal hepatocytes secret IGF1 and malignant HCC cells produce IGF2 [10–12], IGF1R signaling could modulate Treg behavior inside and outside HCC sites. Nonetheless, this hypothesis has yet to be tested.

In the current research, we analyzed the phenotype of Tregs in a chemical-induced mouse HCC model. We found two novel intrahepatic Treg subsets with differential IGF1R expression. Furthermore, these Treg subsets exhibited distinct cytokine production patterns, proliferation rates, and metabolic activities. Our data revealed the heterogeneity of intrahepatic Tregs and shed light on the role of IGF1R signaling in Treg-mediate immunosuppression in HCC.

Materials and methods

Animals

The animal study was approved by The Animal Ethical and Welfare Committee of Renmin Hospital of Wuhan University (Approval ID: WDRY2019-K085) and conducted following the Wuhan University Animal Use Guidelines. Eight-week-old male C57BL/6J mice and Foxp3-GFP transgenic mice (C57BL/6J background) were purchased from the Center for Animal Experiment at Renmin Hospital of Wuhan University.

HCC model

To induce HCC, N-nitrosodiethylamine (Weikeqi Biotech Co., Ltd, 40 μg/g body weight) was administered into the peritoneal cavity of each mouse. Three days later, 10% carbon tetrachloride (ThermoFisher, 10 µl/g body weight) was intraperitoneally administered twice every week for 4 weeks. Another single dose of 40 μg/g body weight N-nitrosodiethylamine was then administered the same way. After that, 10 µl/g body weight 10% carbon tetrachloride was intraperitoneally injected once every week for consecutive 14 weeks. HCC induction was confirmed by the formation of initiation foci, hyperplastic nodules, and culminating appearance of the liver (Supplementary Fig. S1). Mice with confirmed HCC induction were used for the study.

Enrichment of splenic and intrahepatic immune cells

One week after the end of HCC induction, mice were euthanized and spleens were ground on cell strainers (Corning) to prepare single splenocyte suspensions. Mouse livers were ground on steel mesh filters and liver cells were suspended in 30% Percoll (Sigma-Aldrich), followed by overlaying onto an equal volume of 70% Percoll. Liver cell suspensions were centrifuged at 500g for 20 min and mononuclear cells were harvested in the Percoll interface. Red blood cells were removed by incubation in the red blood cell lysis buffer (Beyotime Biotech).

Fluorescence-activated cell sorting (FACS)

The following monoclonal antibodies were ordered from BioLegend: APC/Cy7 anti-CD3 (17A2), PE/Cy7 anti-CD4 (RM4-5), PE anti-CD25 (3C7), PE anti-ICOS (15F9), PE anti-CTLA4 (UC10-4B9), PE anti-TGF-β (TW7-16B4), PE anti-IL-10 (JES5-16E3), PE/Cy7 anti-CD8b (53-5.8), PE anti-perforin (S16009A), and PE anti-Ki67 (11F6). PE anti-phospho-mTOR antibody (Ser2448, clone# MRRBY) was purchased from eBioscience. Unconjugated anti-IGF1R antibody (MAB391) was purchased from R&D Systems. Alexa Fluor 647 anti-IGF2R antibody (2G11) was purchased from Novus Biologicals. Corresponding isotype control antibodies were purchased from BioLegend. For surface marker staining, 1 × 10⁶/ml cells were incubated with 5 µg/ml each antibody in 50 µl of PBS at 4^oC for 30 min. Since the primary anti-IGF1R antibody was not conjugated to a fluorophore, cells were washed with PBS once and incubated with 5 µg/ml PE-Cy5.5 goat anti-mouse IgG (ab130784, Abcam) for 20 min. Dead cells were excluded by staining with SYTOX™ Blue (Thermo Fisher Scientific) following the vendor’s instructions. To stain intracellular proteins such as cytokines, Ki67, and phospho-mTOR, cells were fixed with 2% paraformaldehyde for 15 min, permeabilized with cold methanol (90%) for 30 min, followed by incubation with 5 µg/ml corresponding antibodies for 1 h. Cell apoptosis was determined by the Annexin V-APC Assay Kit (Abcam) following the manufacturer’s manual. A BD LSR II flow cytometer was used to analyze cells and a BD FACSAria sorter was used for cell sorting.

In vitro Treg stimulation and immunosuppression assay

For Treg stimulation, a 96-well microplate (Corning) was coated with 5 µg/ml anti-CD3 antibody (17A2, BioLegend) overnight. 1 × 10⁶/ml FACS-sorted intrahepatic Tregs were resuspended in the culture medium (RPIMI 1640 supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin). One hundred microliters of Treg suspension were seeded in each well of the microplate for 16 h in the presence of 2 µg/ml anti-CD28 antibody (37.51, BioLegend) and 100 U/ml recombinant mouse IL-2 (R&D Systems). Brefeldin A (10 µg/ml, Sigma-Aldrich) and monensin (5 µg/ml, Sigma-Aldrich) were added at the final 5 h of incubation. Tregs were then subjected to intracellular cytokine staining and FACS analysis as described above.

For the immunosuppression assay, splenic CD3⁺CD4⁺CD25^– conventional T cells were sorted from wild-type mice by FACS and labeled with the CellTrace™ Violet Cell Proliferation Kit (Thermo Fisher Scientific) following the vendor’s protocol. Briefly, 1 × 10⁶/ml conventional T cells were incubated in 5 µM CellTrace™ Violet staining solution for 20 min in a 37°C water bath. The staining was stopped by the addition of four volumes of culture medium and centrifugation for 5 min at 300g. The supernatant was discarded and the conventional T-cell pellet was resuspended in the culture medium. After that, 2 × 10⁵ intrahepatic Tregs and 2 × 10⁵ labeled conventional T cells were seeded in a 48-well microplate pre-coated with the anti-CD3 antibody. Cells were then incubated for 5 days in the presence of 2 µg/ml anti-CD28 antibody and 100 U/ml IL-2. The dilution of CellTrace™ Violet was evaluated by FACS. In some experiments, these T cells were stained with 5 µg/ml PE anti-CD25 antibody (3C7, BioLegend) on ice for 15 min to detect CD25 expression by FACS.

Enzyme-linked immunosorbent assay

Concentrations of IL-10 and TGF-β in Treg culture medium were measured using a sandwich ELISA as follows. 1 × 10⁶/ml sorted Treg subsets were cultured in vitro as described above. The cell suspensions were centrifuged for 5 min at 250g and the supernatants were collected for ELISA. The Mouse IL-10 Quantikine ELISA Kit (M1000B, R&D Systems) and Mouse TGF-β1 Quantikine ELISA Kit (DB100C, R&D Systems) were used. Wells were blocked with PBS containing 1% bovine serum albumin and 0.05% Tween 20 for 2 h at room temperature. One hundred microliters of test samples and cytokine standards were added to separate wells and the plate was incubated at room temperature for 2 h, after which the plate was washed. One hundred microliters of detection antibodies were added, and the reaction was allowed to proceed for 2 h at room temperature. The plate was washed and 100 μl of substrate solution was added to each well and incubated for 30 min at room temperature. One hundred microliters of stop solution were then applied to each well. The plates were loaded on an Infinite 200 PRO multimode plate reader (Tecan) to record OD450 and OD540. The final readouts were computed as OD450 minus OD540. Cytokine concentrations were calculated based on the data of cytokine standards.

To quantify IFN-γ secreted from activated CD3⁺CD4⁺CD25^- conventional T cells, conventional T cells were stimulated for 5 days as described above, followed by re-stimulation with 5 ng/ml phorbol 12-myristate 13-acetate (PMA, Sigma-Aldrich) plus 200 ng/ml ionomycin (Sigma-Aldrich) for an additional 5 h. IFN-γ in the culture media was measured using the Mouse IFN-γ Quantikine ELISA Kit (R&D Systems) following the vendor’s manual.

Adoptive transfer

HCC was induced in both wild-type mice and Foxp3-GFP mice. Intrahepatic Treg subsets were sorted from HCC-bearing Foxp3-GFP mice by FACS. 2.0 × 10⁶ sorted Tregs were resuspended in 100 µl of saline and retro-orbitally injected into each HCC-bearing wild-type mouse or healthy wild-type mouse, respectively. One week later, the recipients were sacrificed and intrahepatic immune cells were recovered from the recipients’ spleens and livers as described above. These intrahepatic immune cells were incubated with 5 µg/ml anti-IGF1R antibody on ice for 30 min, washed with 1 ml of PBS once, and stained with 5 µg/ml PE-Cy5.5 goat anti-mouse IgG on ice for 20 min. After that, IGF1R expression on donor-derived GFP⁺ Tregs was analyzed by FACS.

Transcriptome sequencing

Cohort 1 included Tregs pooled from 10 mice and Cohort 2 included Tregs pooled from 8 mice. Total RNAs were extracted from pooled Tregs using the RNeasy Micro Kit (Qiagen) following the manufacturer’s instructions. The high quality of the RNA samples (RNA integrity number, >7.0) was confirmed using the 4200 TapeStation (Agilent Technologies). One microgram of RNA (>200 ng/µl) from each cohort was delivered to Wuhan SeqHealth Co., Ltd. to prepare cDNA libraries using the Collibri™ Stranded RNA Library Prep Kit for Illumina (ThermoFisher). Libraries were sequenced on a HiSeq 2000 Sequencing System (Illumina). Reads were demultiplexed (bcl2fastq). Fastq files were aligned to the mm10 mouse genome (TopHat v2.1.1) and mapped to genes (HTSeq v0.12.4) using the ensemble gene annotation. Gene ontology (GO) enrichment analysis was conducted by the hypergeometric test.

Quantitative RT-PCR

cDNAs were synthesized using the PrimeScript 1st strand cDNA Synthesis Kit (Takara Bio). The SYBR Green Quantitative RT-PCR Kit (Sigma-Aldrich) was used for quantitative PCR on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad) following the standard protocol: pre-warming at 50^oC for 2 min, 94^oC for 10 min, and then 40 cycles of 30 s at 94^oC and 1 min at 62^oC. The amounts of target transcripts were normalized to β-actin transcript levels and computed by the 2^–ΔΔCt formula. Primer sequences are provided in Supplementary Table S1.

Adenosine triphosphate (ATP) assay

The intracellular ATP level was measured using the Luminescent ATP Detection Assay Kit (ab113849, Abcam) following the manufacturer’s protocol. Briefly, 1 × 10⁵ intrahepatic Tregs (in 90 μl of culture medium) were seeded into each well of a 96-well plate. Fifty microliters of detergent were added to each well, followed by shaking for 5 min in an orbital shaker at 600–700 rpm. After that, 50 μl of substrate solution was added to each well, followed by shaking for 5 min in an orbital shaker at 600–700 rpm. The plate was then placed in the dark at room temperature for 10 min. The luminescence was read on an Infinite 200 PRO multimode plate reader (Tecan). ATP was then calculated by linear regression according to the ATP standard provided by the manufacturer.

2,7-Dichlorodihydrofluorescein diacetate (H2DCFDA) staining

Reactive oxygen species (ROS) in Tregs were assessed by H2DCFDA staining. 1 × 10⁶/ml sorted Tregs were incubated in pre-warmed PBS containing 10 μM H2DCFDA (ThermoFisher) for 15 min at 37°C. After two washes with PBS, cells were suspended in PBS and the green fluorescence was measured on the flow cytometer.

Lactate assay

Intracellular l-lactate was quantified using the l-Lactate Assay Kit (ab65330, Abcam) following the manufacturer’s protocol. Briefly, 2.5 × 10⁵ Tregs were washed with cold PBS and resuspended in 200 μl of lactate assay buffer. Cells were then homogenized quickly by pipetting up and down 10 times, followed by centrifugation at 13 000g for 5 min at 4^oC. The supernatant was transferred to another tube and kept on ice. Perchloric acid (Sigma-Aldrich) was added into the homogenate at a final concentration of 1 M and the sample was incubated on ice for 5 min. The sample was centrifuged at 13 000g for 2 min at 4°C. The supernatant was transferred to another tube and potassium hydroxide was added at a final concentration of 500 mM. The sample was centrifuged at 13 000g for 15 min at 4°C and the supernatant was transferred to another tube. Fifty microliters of the reaction mixture were added to each sample and incubated at room temperature for 30 min in the dark. The signal was detected at Ex/Em = 535/587 nm on an Infinite 200 PRO multimode plate reader (Tecan). l-Lactate quantity was then calculated by linear regression according to the l-lactate standard provided by the manufacturer.

Enrichment of malignant cells from tumor nodules

One week after the end of HCC induction, mouse livers were collected. Tumor nodules in the livers were harvested using curved surgical scissors. The tumor nodules were then minced into small pieces and ground on steel mesh filters to prepare single-cell suspensions. Red blood cells were lysed with red blood cell lysis buffer (BioLegend) following the vendor’s instructions. Cells were incubated with 5 µg/ml FITC anti-CD45 antibody and FITC anti-CD31 antibody (S18009F; 390, BioLegend) for 30 min on ice. CD45⁺ leukocytes and CD31⁺ endothelial cells were excluded while CD45^–CD31^– cells (predominantly malignant cells) were sorted by flow cytometry. CD45^–CD31^– cells were either subjected to Ki67 staining or co-culture with Tregs.

Co-culture of HCC cells and Tregs

CD45^–CD31^– cells were sorted from tumor nodules and Tregs were sorted from mouse spleens as described above. The cell density of Tregs and CD45^–CD31^– cells were adjusted to 1 × 10⁶/ml in supplemented RPIMI 1640 medium. For co-culture without cell–cell contact, 1 × 10⁵ CD45^–CD31^– cells were seeded into each well of a 24-well Transwell plate (Corning) and 1 × 10⁴ Tregs were seeded into each insert with a 0.4-µm pore membrane. For co-culture with cell–cell contact, CD45^–CD31^– cells and Tregs were mixed and seeded into each well. Mouse IL-2 (R&D Systems) was added into the culture at a final concentration of 100 U/ml to maintain Treg identity. Twenty-four hours later, cells were stained with APC/Cy7 anti-CD3 antibody and IGF1R antibody followed by incubation with PE-Cy5.5 goat anti-mouse IgG. IGF1R expression on CD3⁺ cells (i.e. Tregs) was then analyzed. CD3⁺ cells were sorted for measuring IGF1R transcripts using quantitative RT-PCR.

Statistics

Each experiment was independently performed two or three times. The data were shown as mean ± standard deviation. Student’s t-test or one-way ANOVA with Tukey post hoc test was performed for statistical analysis. A P-value of <0.05 was regarded as significant.

Results

IGF1R expression is differential in intrahepatic Tregs

To characterize Treg phenotype in HCC, we evaluated a series of surface markers in both splenic and intrahepatic Tregs in Foxp3-GFP transgenic mice immediately after HCC was induced. As displayed in Fig. 1A, mononuclear cells were isolated from the spleens and livers of HCC-bearing mice. Single cells were first gated within total events, and dead cells were excluded by SYTOX™ blue staining. Lymphocytes were gated based on FSC-A and SSC-A, and CD3⁺CD4⁺GFP⁺ Tregs were then distinguished. The expression of Foxp3, CD25, inducible T-cell costimulator (ICOS), and cytotoxic T-lymphocyte-associated protein 4 (CTLA4) were all comparable between splenic and intrahepatic Tregs (Fig. 1B). Because IGF receptors have been reported to modulate Treg function, we also analyzed the expression of IGF1R and IGF2R. As indicated in Fig. 1B and C, splenic Tregs uniformly expressed IGF1R, while intrahepatic Tregs were divided into two subsets: one subset expressing high Foxp3 and low IGF1R (Hereinafter IGF1R^lo/^– Tregs), and another subset expressing moderate or low Foxp3 but high IGF1R (Hereinafter IGF1R^hi Tregs). The percentage of IGF1R^hi Tregs was higher than IGF1R^lo/– Tregs (Fig. 1D). To further characterize the expression of IGF1R, CD3⁺CD4⁺GFP⁺IGF1R^hi Tregs (Hereinafter IGF1R^hi Tregs) and CD3⁺CD4⁺GFP⁺IGF1R^lo/– Tregs (Hereinafter IGF1R^lo/– Tregs) were sorted by FACS (Supplementary Fig. S2). Quantitative RT-PCR indicated that IGF1R^hi Tregs expressed more IGF1R transcripts than IGF1R^lo/– Tregs (Fig. 1E). The transcript levels of T-bet and RORγt were comparable in the two subsets (Supplementary Fig. S3). Gata3 transcripts were too low to be detected (data not shown). The expression of Helios, a key marker of Treg functional potency, was equivalent in these subsets (Supplementary Fig. S4).

Figure 1:

Intrahepatic Treg phenotype. (A) FACS dot plots showing the gating strategy for splenic and intrahepatic CD3⁺CD4⁺Foxp3-GFP⁺ Tregs in HCC-bearing mice 1 week after the end of HCC induction. S: spleen. L: liver. (B) FACS histograms showing the expression of Foxp3 and indicated surface proteins in splenic and intrahepatic CD3⁺CD4⁺Foxp3-GFP⁺ Tregs. (C) FACS plots show the expression of IGF1R in splenic and intrahepatic CD3⁺CD4⁺Foxp3-GFP⁺ Tregs. (D) Percentages of the two Treg subsets in total intrahepatic Tregs. F^hi1R^lo/–: Foxp3^highIGF1R^low/– Tregs. F^mo/lo1R^hi: Foxp3^moderate/lowIGF1R^high Tregs. Student’s t-test. (E) IGF1R transcript levels in Foxp3^highIGF1R^low/^– Tregs and Foxp3^moderate/lowIGF1R^high Tregs sorted from livers of HCC-bearing mice. One-way ANOVA. ∗∗P < 0.01; ∗∗∗P < 0.001. N = 6 mice per group.

IGF1R^hi Tregs express more suppressive cytokines than IGF1R^lo/– Tregs in vivo

To investigate the functions of the two Treg subsets, we sorted IGF1R^hi Tregs and IGF1R^lo/^– Tregs and directly stained TGF-β and IL-10 in these cells. As shown in Fig. 2A and B, in comparison to splenic Tregs, IGF1R^lo/– Tregs produced more TGF-β and IL-10. IGF1R^hi Tregs produced even higher TGF-β and IL-10 than IGF1R^lo/– Tregs. To further check whether the two subsets had different abilities to express suppressive cytokines, sorted IGF1R^lo/^– Tregs and IGF1R^hi Tregs were stimulated in vitro with agonistic CD3 and CD28 antibodies. As shown in Fig. 2D, after stimulation, the two subsets expressed equivalent levels of TGF-β and IL-10, suggesting that their cytokine production abilities were the same and extrinsic factor(s) resulted in the difference in vivo. These results were validated by ELISA (Supplementary Fig. S5).

Figure 2:

Expression of TGF-β and IL-10 in vivo and in vitro. (A and B) Intracellular staining of TGF-β (A) and IL-10 (B) in sorted intrahepatic IGF1R^lo/^– and IGF1R^hi Tregs without any other treatment. SP: splenic total Tregs. (C and D) TGF-β (C) and IL-10 (D) in Tregs after in vitro stimulation with agonistic CD3 and CD28 antibodies for 16 h. Left panels: FACS plots. Right panels: statistics. One-way ANOVA. ∗P<0.05; ∗∗P<0.01; ns: not significant. N = 5 mice per group.

Intrahepatic IGF1R^hi Tregs were more proliferative with stronger mTORC1 signaling

To check the immunosuppressive potency of the two Treg subsets on the expansion of conventional T cells, sorted IGF1R^hi Tregs and IGF1R^lo/– Tregs were co-cultured with CellTrace™ Violet-labeled splenic CD4⁺CD25^– T cells (i.e. conventional T cells) in the presence of agonistic antibodies. We found comparable dilutions of CellTrace™ Violet in CD4⁺CD25^– T cells after co-culture with the two Treg subsets, suggesting their similar immunosuppressive function (Fig. 3A). The two Treg subsets also comparably decreased conventional T-cell-derived IFN-γ in the culture medium (Fig. 3B) and inhibited CD25 (a T-cell activation marker) expression on conventional T cells (Fig. 3C and D). The two Treg subsets had no remarkable differences in apoptosis when they were sorted from the mice (Fig. 3E). Interestingly, more Ki67⁺ cells were found in intrahepatic IGF1R^hi Tregs relative to IGF1R^lo/– Tregs, suggesting their distinct proliferation rates in vivo (Fig. 3F and G). Consistently, the phosphorylation of mTOR, which is essential for cell proliferation, was elevated in IGF1R^hi Tregs relative to IGF1R^lo/– Tregs (Fig. 3H).

Figure 3:

Immunosuppressive activity, apoptosis, and proliferation of intrahepatic Treg subsets. (A) CellTrace™ Violet dilution in splenic conventional CD4⁺CD25^- T cells after in vitro stimulation with agonistic antibodies for 5 days in the presence or absence of sorted intrahepatic Treg subsets. Unstimulated: unstimulated CD4⁺CD25^– T cells. Stimulated: CD4⁺CD25^– T cells stimulated with agonistic antibodies. Alone: CD4⁺CD25^– T cells alone. With IGF1R^lo/–: CD4⁺CD25^– T cells with IGF1R^lo/– Tregs. With IGF1R^hi: CD4⁺CD25^– T cells with IGF1R^hi Tregs. The data represent two independent experiments. (B) IFN-γ concentrations in the culture medium of CD4⁺CD25^– T cells after stimulation with agonistic antibodies for 5 days in the presence or absence of sorted intrahepatic Treg subsets. (C and D) CD25 expression on stimulated CD4⁺CD25^– T cells. (E) Staining of intrahepatic Treg subsets with Annexin V and propidium iodide to show their apoptosis immediately after sorting from HCC-bearing mice. The data represent two independent experiments. (F and G) Ki67 staining in intrahepatic Treg subsets immediately after sorting from HCC-bearing mice. FACS plots are in (F), and statistics are in (G). (H) Staining of phosphorylated mTOR (Ser2448) in intrahepatic Treg subsets immediately after sorting from HCC-bearing mice. Left panel: FACS histograms. Right panel: statistics of mean fluorescent intensities of phosphorylated mTOR. One-way ANOVA for (B) and (D). Student’s t-test for (G) and (H). ∗∗P < 0.01; ∗∗∗P < 0.001. N=5–8 samples per group.

HCC microenvironment modulates IGF1R expression in Tregs

To know whether the differential IGF1R expression was intrinsic or extrinsic, we sorted intrahepatic IGF1R^hi Tregs and IGF1R^lo/– Tregs from HCC-bearing Foxp3-GFP mice and adoptively transferred them into HCC-bearing wild-type mice, respectively. One week after transfer, GFP⁺ Tregs were detected in spleens and livers of the recipients (Fig. 4A and B). Donor-derived IGF1R^hi Tregs still expressed high IGF1R in recipients’ spleens, whereas they partially down-regulated IGF1R in recipients’ livers (Fig. 4C). In contrast, donor-derived IGF1R^lo/– Tregs still expressed low IGF1R in recipients’ livers but up-regulated IGF1R in recipients’ spleens (Fig. 4D). To test whether the healthy hepatic or splenic microenvironment could induce the same changes, we transferred intrahepatic IGF1R^hi and IGF1R^lo/– Tregs of HCC-bearing Foxp3-GFP mice into healthy mice. One week after transfer, no change of IGF1R expression was found in donor-derived IGF1R^hi Tregs in either recipients’ spleens and livers (Fig. 4E and G). However, IGF1R expression was up-regulated in donor-derived IGF1R^lo/– Tregs in recipients’ spleens but not livers (Fig. 4F and H). Therefore, the malignant liver microenvironment impaired IGF1R expression while the splenic microenvironment favored IGF1R expression.

Figure 4:

Adoptive transfer assay of Treg subsets. (A and B) Recognition of donor-derived GFP⁺ Treg subsets, which were originally IGF1R^hi (A) or IGF1R^lo/– (B) before the transfer, in spleens and livers of HCC-bearing wild-type recipients. (C) Expression of IGF1R on donor-derived GFP⁺ Tregs, which were originally IGF1R^hi before the transfer, in spleens and livers of HCC-bearing wild-type recipients. Original: Treg subsets before transfer. Spleen: transferred Tregs in recipients’ spleens. Liver: transferred Tregs in recipients’ livers. Left panels: FACS histograms. Right panels: statistics of the mean fluorescent intensities of IGF1R. (D) Expression of IGF1R on donor-derived GFP⁺ Tregs, which were originally IGF1R^lo/^– before the transfer, in spleens and livers of HCC-bearing wild-type recipients. Left panels: FACS histograms. Right panels: statistics of the mean fluorescent intensities of IGF1R. (E and F) Recognition of donor-derived GFP⁺ Treg subsets, which were originally IGF1R^hi (A) or IGF1R^lo/– (B) before the transfer, in spleens and livers of healthy wild-type recipients. (G) Expression of IGF1R on donor-derived GFP⁺ Tregs, which were originally IGF1R^hi before the transfer, in spleens and livers of healthy wild-type recipients. Left panels: FACS histograms. Right panels: statistics of the mean fluorescent intensities of IGF1R. (H) Expression of IGF1R on donor-derived GFP⁺ Tregs, which were originally IGF1R^lo/– before the transfer, in spleens and livers of healthy wild-type recipients. Student’s t-test. ∗P < 0.05; ∗∗P < 0.01. N = 3 mice per group.

Transcriptomic and metabolic differences exist in the Treg subsets

To understand the gene expression profiles of the two Treg subsets, intrahepatic IGF1R^lo/– Tregs and IGF1R^hi Tregs were sorted for transcriptome sequencing. Compared with IGF1R^lo/– Tregs, a total of 2139 up-regulated transcripts and 56 down-regulated transcripts were observed in IGF1R^hi Tregs (Fig. 5A). Since gene up-regulation was predominant in IGF1R^hi Tregs, we conducted GO enrichment analysis on these up-regulated transcripts. We found significant associations with respiratory electron transport chain, oxidative phosphorylation (OXPHOS), ATP synthesis, glycolysis, and translation-related process in the category of “biological process” (Fig. 5B). The GO enrichment analysis also showed strong associations with respiratory chain, oxidative complex, NADH dehydrogenase complex, respiratory chain complex I, and respiratory chain complex IV in the category of “cellular component” (Fig. 5C). In the category of molecular function, remarkable associations with NADH dehydrogenase (ubiquinone or quinone) activity, electron carrier activity, and oxidoreductase activity were revealed (Fig. 5D). The top 20 up-regulated genes encode proteins essential for glycolysis, OXPHOS, and electron transport chain (Fig. 6A and Table 1). Quantitative RT-PCR validated the increases of these transcripts, except for PGK1 and NDUFV1 (Fig. 6B and C). Therefore, we then quantified ATP levels in sorted intrahepatic IGF1R^lo/^– Tregs and IGF1R^hi Tregs. As exhibited in Fig. 6D, IGF1R^hi Tregs produced more ATP than IGF1R^lo/– Tregs. Evaluation of intracellular lactate, which is the end product of glycolysis, indicated that IGF1R^hi Tregs generated more lactate than IGF1R^lo/– Tregs, confirming their higher glycolysis activity (Fig. 6E). Furthermore, the ROS level, which was assessed by H2DCFDA staining, was higher in IGF1R^hi Tregs relative to IGF1R^lo/– Tregs (Fig. 6F).

Figure 5:

Transcriptome sequencing of intrahepatic IGF1R^hi Tregs and IGF1R^lo/– Tregs. (A) Volcano plot of the differentially expressed genes between intrahepatic IGF1R^hi Tregs and IGF1R^lo/– Tregs in mouse cohort 1. Black vertical dashed lines indicate log fold changes of −2 and 2, while the black horizontal dashed line represents a padj of 0.05. (B–D) GO enrichment of up-regulated transcripts in the category of biological process terms (B), cellular component terms (C), and molecular function terms (D). 1: cohort 1 including 10 mice. 2: cohort 2 including 8 mice. Red arrows indicate GO terms related to glycolysis, OXPHOS, and electron transport chain.

Figure 6:

Metabolic status of intrahepatic in sorted intrahepatic IGF1R^lo/^– Tregs and IGF1R^hi Tregs. (A) Heat map of the top 20 up-regulated genes related to glycolysis, OXPHOS, and electron transport chain in IGF1R^hi Tregs. 1: mouse cohort 1. 2: mouse cohort 2. The blue bands indicate low gene expression quantity, and the red bands represent high gene expression quantity. (B and C) Validation of differential expression of the 20 genes by quantitative RT-PCR. N = 6 mice per group. (D) ATP production in 1 × 10⁵ Treg subsets. (E) Intracellular lactate levels in 2.5 × 10⁵ Treg subsets. (F) H2DCFDA intensities in Treg subsets. Vehicle: IGF1R^hi Tregs incubated with PBS without H2DCFDA. Left panel: representative FACS histograms. Right panel: statistics of mean fluorescence intensities of H2DCFDA. N = 5 or 6 mice per group. Student’s t-test. ∗P < 0.05; ∗∗: P < 0.01; ∗∗∗P < 0.001.

Table 1:

Top 20 up-regulated DEGs related to glycolysis, OXPHOS, and electron transport chain

Gene	Product
HK2	Hexokinase 2
PGM3	Phosphoglucomutase 3
ALDOA	Aldolase, fructose-bisphosphate A
PKM	Pyruvate kinase
ENO1	Enolase 1
PGK1	Phosphoglycerate kinase 1
NDUFA1	NADH:ubiquinone oxidoreductase subunit A1
NDUFA4	NDUFA4, mitochondrial complex associated
ND4	NADH dehydrogenase subunit 4
NDUFV1	NADH:ubiquinone oxidoreductase core subunit V1
SDHAF4	Succinate dehydrogenase assembly factor 4
SDHC	Succinate dehydrogenase complex, subunit C
SDHD	Succinate dehydrogenase [ubiquinone] cytochrome b small subunit
UQCRC2	Ubiquinol-cytochrome c reductase core protein 2
UQCRB	Ubiquinol-cytochrome c reductase binding protein
UQCRH	Ubiquinol-cytochrome c reductase hinge protein
UQCRC1	Ubiquinol-cytochrome c reductase core protein 1
AFG1L	AFG1 like ATPase
MT-CO1	Cytochrome c oxidase subunit 1
COX7C	Cytochrome c oxidase subunit 7C

The frequencies of the Treg subsets are associated with HCC onset but not progression

To evaluate the relationship between Treg subsets and HCC development, we divided the HCC induction period into three stages: initiation, promotion, and progression (Supplementary Fig. S6). The frequencies of Treg subsets were quantified at each stage. As shown in Fig. 7A and B, <10% of intrahepatic Tregs were IGF1R^lo/– Tregs at the initiation stage, and the frequency increased to ~40% at the promotion stage and progression stage. The frequency of IGF1R^hi Tregs was ~85% at the initiation stage and decreased to 50% at the promotion stage and progression stage. Ki67 staining of enriched malignant HCC cells, which were the CD45^–CD31^– population among cells isolated from tumor nodules, was little at the initiation stage but increased at the promotion stage and progression stage, indicating progressive promotion of HCC growth (Fig. 7C and D). The tumor nodule number in the liver was zero at the initiation stage and was progressively increased from the promotion stage to the progression stage (Fig. 7E). Because the frequencies of each Treg subset were comparable between the promotion stage and progression stage, the increase in IGF1R^lo/– Tregs was likely associated with the onset but not the progression of HCC. To evaluate anti-HCC immunity, intrahepatic CD3⁺CD8⁺ T cells were sorted (Fig. 7F) and their perforin expression was determined by flow cytometry. Interestingly, as shown in Fig. 7G and H, perforin expression was very low at the initiation stage and was significantly promoted at the promotion stage. However, perforin expression was decreased at the promotion stage but still higher than that at the initiation stage, probably signifying CD8⁺ T-cell exhaustion. Therefore, the increase in IGF1R^lo/– Tregs seemed to be related to the initiation rather than the exhaustion of the CD8⁺ T-cell response.

Figure 7:

Association of Treg subsets with tumor growth and immunity. (A and B) Frequencies of Treg subsets at different HCC stages. Representative flow cytometry dot plots are shown in (A) and statistics are shown in (B). N = 5 mice per group. (C) Representative flow cytometry dot plots indicating Ki67⁺ cells in CD45^–CD31^– cells enriched from tumor nodules. (D) Statistics of Ki67⁺ cells. N = 10 mice per group. (E) Tumor nodule number in each liver at different stages. N = 10 mice per group. (F) Gating of intrahepatic CD3⁺CD8⁺ T cells. (G and H) Perforin expression in intrahepatic CD8⁺ T cells. Representative flow cytometry plots are shown in (G). Statistics are shown in (H). N = 5 mice per group. One-way ANOVA. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

HCC cells induce IGF1R down-regulation on Tregs

To determine the cause of IGF1R down-regulation on Tregs, we sorted HCC cells, i.e. CD45^–CD31^– cells from tumor nodules after HCC induction and co-cultured them with splenic Tregs (predominantly IGF1R^hi) in a separate manner or cell–cell contact manner. The expression of IGF1R was measured 24 h later. As exhibited in Fig. 8A and B, <3% of Tregs were IGF1R^lo when Tregs were cultured alone, whereas ~15% of Tregs were IGF1R^lo when HCC cells were present regardless of the co-culture manners. The IGF1R down-regulation was substantiated by decreases in IGF1R mRNA (Fig. 8C). Therefore, HCC cells might produce a soluble factor(s) to down-regulate IGF1R expression. However, the identity of the factor(s) remains unknown and should be revealed in the future.

Figure 8:

HCC cells induce IGF1R down-regulation on Tregs. (A) Representative flow cytometry dot plots indicating IGF1R expression on Tregs after 24-h culture. Alone: Tregs alone. Co-culture/separate: Tregs-HCC cell co-culture without cell–cell contact. Co-culture/contact: Tregs-HCC cell co-culture with cell–cell contact. (B) Statics of IGF1R^lo Treg frequencies after culture. (C) IGF1R transcript levels after culture. N = 5 mice per group. One-way ANOVA. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

Discussion

IGFs and IGF1R have been implicated in Treg induction [8], suppressive function [6], and Treg/Th17 balance in different disorders [9]. To the best of our knowledge, this is the first study reporting IGF1R as the maker of heterogeneous Treg subsets. We observed the expression of both IGF1R and IGF2R in intrahepatic Tregs. Because IGF2R has a short cytoplasmic domain and is thus regarded to have a limited signaling ability, we focused on IGF1R in this study. In our experimental setting, this IGF1R down-regulation was at least partially induced by malignant HCC cell-derived soluble factor(s). However, the identity of the factor(s) remains unknown and should be revealed in the future.

Our data demonstrated that the change of IGF1R is not associated with alterations in the expression of T-bet and RORγt. Therefore, IGF1R signaling might not bias the differentiation and plasticity of intrahepatic Tregs. In HCC livers, IGF1R^hi Tregs produced higher TGF-β and IL-10 and are more proliferative than IGF1R^lo/^– Tregs, suggesting they are more immunosuppressive and active than IGF1R^lo/^– Tregs in HCC tissue. This is consistent with previous studies stating the positive effects of IGF1 and IGF2 on Treg number and suppressive properties in autoimmune disease [5, 13]. Importantly, it is known that IGF1R signaling activates mTORC1 signaling not only in cancers but also in T cells [9, 14], and we did find higher mTOR phosphorylation in IGF1R^hi Tregs. Therefore, the data suggest that IGF1R^hi Tregs exert a stronger immunosuppressive effect in vivo. However, because the in vitro stimulation and suppression assays showed no notable difference s in either cytokine production or inhibition of conventional T cell proliferation, we concluded that IGF1R^hi Tregs and IGF1R^lo/^– Tregs have equal immunosuppressive capacities and in vitro stimulation overcomes the difference of in vivo IGF1R signaling. Whether IGF1 or IGF2 favors the suppressive function of intrahepatic Tregs under the in vitro activation condition is under investigation by our team.

Using transcriptome sequencing, we reveal that IGF1R^hi Tregs expressed more enzymes crucial for glycolysis, OXPHOS, and electron transport chain, suggesting that IGF1R^hi Tregs are more metabolically active than IGF1R^lo/- Tregs in HCC tissues. This result is consistent with the higher proliferation rate and phosphorylated mTOR in IGF1R^hi Tregs, because proliferating cells require abundant energy, and activated mOTRC1 signaling augments glycolysis and OXPHOS to fuel cells [15]. Interestingly, genes related to translation, ribosome units, and nucleoside metabolism were all up-regulated in IGF1R^hi Tregs relative to IGF1R^lo/– Tregs. This is another proof of elevated mOTRC1 signaling since mOTRC1 is a strong drive of gene translation and transcription [16, 17]. Therefore, it is very likely that IGF1R^hi Tregs bind IGFs and ignite intracellular signals to activate mOTRC1.

In conclusion, for the first time, this study identified the heterogeneity of intrahepatic Tregs in HCC and provide valuable clues for evaluating the role of IGF1R in modulating Treg function. It will be necessary to further explore the exact effect of IGF1R signaling in Tregs both in HCC and normal livers to deepen our understanding of the molecular mechanisms underlying the alterations of Treg activities in the liver.

Glossary

Abbreviations

ATP

adenosine triphosphate

CTLA4

cytotoxic T-lymphocyte-associated protein 4

FACS

fluorescence-activated cell sorting

FOXP3

forkhead box protein P3

H2DCFDA

2, 7-dichlorodihydrofluorescein diacetate

HCC

hepatocellular carcinoma

ICOS

inducible T-cell costimulator

IFN

interferon

IGF1

insulin-like growth factor type 1

IGF1R

insulin-like growth factor type 1 receptor

IGF2

insulin-like growth factor type 1

IGF2R

insulin-like growth factor type 2 receptor

IL

interleukin

mTORC1

the mammalian target of rapamycin complex 1

ROS

reactive oxygen species

TGF

transforming growth factor

Treg

regulatory T cell

OXPHOS

oxidative phosphorylation

Funding

This work was supported by the National Natural Science Foundation of China (81972673 to KD).

Conflict of Interest

The authors declare no potential conflicts of interest.

Author contributions

KD and YH: conceived and designed the projects. LH, YH, JS, and YW: performed the experiments. KD, JZ, and YH: analyzed the data, performed the statistical analysis, and wrote the manuscript. All authors read and approved the final manuscript.

Ethical approval

All animal experiments complied with the ARRIVE guidelines and were carried out in accordance with the ethical standards of The Animal Ethical and Welfare Committee of Renmin Hospital of Wuhan University.

Data availability

All data relevant to the study are included in the article or uploaded as supplementary information.