Interleukin-11 promotes lung adenocarcinoma tumourigenesis and immune evasion

Abstract

Rationale: Interleukin-11 (IL-11) has emerged as a significant player in tumourigenesis, with implications across various cancer types. However, its specific role in driving tumour progression in lung adenocarcinoma (LUAD) remains elusive. IL-11's multifaceted impact on both tumour cells and the tumour microenvironment underscores its potential as a therapeutic target in LUAD. This study aims to unravel the involvement of IL-11 in LUAD progression and its influence on the tumour microenvironment.

Methods: Here, we used transcriptomic and digital spatial profiling analyses together with clinic data from two retrospective LUAD patient cohorts. LUAD cell lines genetically engineered to overexpress or to silence IL-11 or its receptor (IL-11RA) were used for in vitro functional analysis and for in vivo experiments. Additionally, we used three different in vivo models: patient-derived xenografts (PDXs), tobacco-exposed mice and genetically engineered mouse models. A neutralising monoclonal antibody against IL-11RA was produced and tested.

Results: Our findings revealed a pivotal role for IL-11 in driving tumourigenesis across various mouse models, highlighting its capacity to modulate tumour immunity towards an immunosuppressive microenvironment. Moreover, we observed a correlation between IL-11 expression and poorer patient outcomes in LUAD. Notably, therapeutic targeting of IL-11RA with a neutralising antibody demonstrated significant anti-tumour efficacy in a PDX model.

Conclusion: The IL-11/IL-11RA axis emerges as a critical driver of LUAD tumourigenesis, exerting its effects through enhanced tumour cell proliferation and remodelling of the tumour microenvironment. Our study highlights the therapeutic potential of disrupting this axis, suggesting that patients exhibiting elevated IL-11 levels may benefit from therapies targeting the IL-11/IL-11RA pathway.

Keywords: IL‐11/IL‐11RA axis; immunosuppressive tumour microenvironment; lung adenocarcinoma; therapeutic target; tobacco smoke.

FIGURE 1

Augmented expression of IL‐11 is associated with poorer outcomes in LUAD patients. (A) Representative microphotographs of IL‐11 immunohistochemistry (IHC) in two independent lung adenocarcinoma patients. Scale bar 200 µm. (B) and (C) Kaplan–Meier curves of overall survival (OS) were plotted for the LUAD patients from: cohort 1; H12O n = 106 and cohort 2; TCGA n = 527, respectively. (B) For cohort 1, patient samples were divided in positive immunohistochemistry staining which corresponded to samples with weak, moderate or strong IL‐11 expression, and negative staining that included samples without any IL‐11 expression. (C) For cohort 2, LUAD patient samples were divided into high or low IL‐11 mRNA expression levels, establishing the third quartile as cut‐off. The Kaplan–Meier method (*p < .05, **p < .01) was used for survival analysis, using the Log‐Rank model to obtain p values. The Cox proportional hazards model was used to assess hazard ratio (HR) values. Overall survival (OS) was defined as the period between diagnosis and the last clinical review or death. N = number of patients, HR = hazard ratio.

FIGURE 2

IL‐11 signalling promotes lung cancer tumourigenesis. (A and B) Single cell suspensions of IL‐11‐ or IL‐11RA‐overexpressing lung adenocarcinoma cell lines (H358 and H3122) and a 1:1 mixture of IL‐11 and IL‐11RA variants were subcutaneously engrafted into nude mice. The tumours generated were measured weekly. (A) Xenograft relative tumour volumes of IL‐11‐ or IL‐11RA‐overexpressing cell lines. Five mice per group were included. (B) Quantification of pSTAT3‐ and cyclin D1‐positive cells in the tumour samples were measured by immunohistochemistry. Statistical analysis for (A) was performed by Kruskal–Wallis test; ** (p < .01), *** (p < .001). EV = cell line with empty vector, IL‐11 = cell line over‐expressing IL‐11, IL‐11RA = cell line over‐expressing IL‐11RA. In (B), p values were obtained by Mann–Whitney U test ** (p < .01). (C) Schematic layout of the experiments in D and E. A LUAD patient‐derived xenografts (PDX, TP60) were engrafted in nude mice. When tumours reached a volume of 400 mm³, mice were intraperitoneally injected with rhIL‐11 or vehicle for 9 consecutive days. (D) Relative tumour growth of PDX treated with rhIL‐11 or vehicle (n = 5). (E) Quantification of pSTAT3 and cyclin D1 positive cells assessed by IHC from tumour sections of rhIL‐11‐treated mice or control mice. (F) Schematic representation of the experiments in G to I. LUAD were developed in the KRas^LSLG12Vge/+o;P53^LoxP/LoxP model through intranasal inoculation of adeno‐Cre viruses. Tumour development was monitored with microCT photographing for 6–9 months post virus inoculation. Micro‐CT scans on day 0 were used to stratify mice into treatment groups and to quantify relative response. RhIL‐11 (n = 7) or vehicle (n = 6) were administered intraperitoneally three times a week for 4 weeks. (G) Bronchoalveolar lavages were collected 3 h after the last rhIL‐11 dosage (n = 4) or vehicle (n = 2), and hIL‐11 concentration was quantified by ELISA. (H) Size of tumours was measured as the percentage of tumour volume change compared with the baseline tumour volume at day 0. (I) Representative images of H&E staining of lungs from mice treated with vehicle or rhIL‐11 at day 28 post‐virus inoculation. Top panels are 1.25× magnification and bottom panels are 4× magnification. Scale bar 200 µm. Arrows point to tumours. (J) The increase in tumour volume was calculated as the difference between the tumour volume at the end of the experiment and its initial volume. Data and error bars are indicated as mean ± SD. Statistical significance was determined with Mann–Whitney U test (*p < .05, **p < .01). H&E = haematoxylin–eosin, V = vehicle, rhIL‐11 = recombinant human IL‐11.

FIGURE 3

Genetic ablation of IL‐11 reduces pro‐tumourigenic properties in vitro and in vivo. (A) IL‐11 production by IL‐11‐KO and EV genetically modified A549 and H1975 cell lines quantified by ELISA. (B and C) Representative Western blot analysis of pSTAT3 (Tyr705), STAT3, pSTAT1 (Tyr701), STAT1, pMAPK (Thr202/Tyr204), MAPK, pAKT (Ser473), AKT (B) and Cyclin21 and BCL2 (C) on protein extracts of IL‐11‐silenced A549 and H1975 cell lines. (D–G) In vitro surrogate assays were performed to analyse tumourigenic properties: (D) growth curves, (E) colony formation assays (CFA) in 2D, (F) transwell migration assays and (G) 3D soft agar colony formation assay. All experiments were reproduced a minimum of three times in the laboratory and three technical replicates were obtained for each experiment. (H–K) A549^IL‐11KO and H1975^IL‐11KO and control cell lines were injected subcutaneously into athymic nude mice. (H and J) Relative tumour growth was measured (n = 5 mice per group). (I and K) Expression levels of pSTAT3, cyclin D1 and Ki67 were analysed by immunohistochemistry, and we represented the quantification of positive cells of pSTAT3, cyclin D1 and Ki67 tumour sections of xenografts. For growth curves, a representative figure is shown. Means and SDs for the technical replicates are shown on the growth curves. For E to K, data are given as mean + SD. The statistical analysis used was the Mann–Whitney U test (*p < .05, **p < .01, ***p < .001). EV = cell line with empty vector, IL‐11‐KO = cell line with IL‐11 silencing.

FIGURE 4

Therapeutic inhibition of IL‐11/IL‐11RA signalling with a neutralising anti‐IL‐11RA antibody reduces tumour growth in a PDX model. (A) Three‐dimensional organoids were generated from a lung adenocarcinoma PDX model (PDXDO from TP57) and were treated with rhIL‐11 and anti‐IL‐11RA Ab (1000 µg/mL). Representative Western blot analysis of pSTAT3 (Tyr705) on protein extracts from PDXDO of TP57 pre‐incubated with the neutralising antibody against hIL‐11RA for 24 h before stimulation with rhIL‐11. (B) Densitometry quantification of the reactive bands for pSTAT3 (Tyr705) and STAT3 and relative expression were normalised to the total amount of β‐actin. Data are shown as mean. (C) Schematic layout of the experiments in D to E. PDX tumours were engrafted in both flanks of athymic mice (PDX 57). When a volume of 200 mm³ was reached, mice were treated with anti‐IL‐11RA mAb (300 µg/mice), anti‐GST mAb (300 µg/mice) or vehicle by intraperitoneal injection, three times a week for 3 weeks (n = 5). (D) Relative tumour growth of PDX TP57 normalised to the tumour volume at baseline (day 0). (E) Positive cells expressing pSTAT3, Ki67 and Cyclin D1 in tumour sections were analysed by immunohistochemistry and quantified. Representative microphotographs and quantifications are shown. Scale bar 200 µm. Data are given as mean + SD. Statistical significance was determined with the Kruskal–Wallis test (*p < .05, **p < .01, ***p < .001).

FIGURE 5

Effect of IL‐11 in the tumour microenvironment in a tobacco‐exposure lung cancer mouse model. (A) Schematic layout of the experiments in b to c. The air control group was exposed to a normal air environment. Tobacco smoke group belongs to A/J mice intraperitoneally injected with NNK (4‐(Methylnitrosamino)‐1‐(3‐Pyridyl)‐1‐Butanone) 1 week before initiation of MTS (mainstream tobacco smoke) exposure. On the fifth month after exposure, half of mice were injected with rhIL‐11, three times per week for 3 weeks. Tumour development was analysed 9 months after tobacco‐exposure initiation. (B) Percentage of mice per group with lung tumours. Histology (H&E stain) in A/J mice 9 months after initiation of the tobacco‐exposure protocol. Scale bar 1000 and 200 µM. (C) Lungs from IL‐11 treated mice were analysed by flow cytometry for the frequencies of infiltrating lymphocytes: CD3 T cells (Live/Dead⁻ CD45⁺ CD3⁺); CD4 T cells (Live/Dead⁻ CD45⁺ CD3⁺ CD4⁺); CD8 T cells (Live/Dead⁻ CD45⁺ CD3⁺ CD8⁺); B cells (Live/Dead⁻ CD45⁺ CD3⁻ CD20⁺); alveolar macrophages (Live/Dead⁻ CD45⁺ B220⁻ Ly6G⁻ SiglecF⁺); non‐alveolar macrophages (Live/Dead⁻ CD45⁺ B220⁻ Ly6G⁻ CD11b^high F4/80⁺) and neutrophils (Live/Dead⁻ CD45⁺ B220⁻ Ly6G⁺). Data are given as mean + SD. Representative dot plots are shown for groups. Statistical significance was determined with the Kruskal–Wallis test ** (p < .01), *** (p < .001).

FIGURE 6

IL‐11 abundance is associated with a skewed tumour microenvironment in LUAD patients’ samples. (A) IL‐11 expression determined by RT‐qPCR was compared between the patient groups in cohort 3: those with null CD4 infiltration N = 39 and those with high CD4 infiltration N = 25, as assessed by immunohistochemistry. (B) Heatmap of changes in expression of a group of genes analysed by targeted RNASeq (OIRRA panel). The expression profile from this set of 37 genes, segregates the samples according to a high (red) or low (green) expression of IL‐11 in patients with lung adenocarcinoma (cohort 3; HU12O, n = 67). FDR ≤ .1. (C) Over‐representation analysis (ORA) associating onco‐immune functions. Each coloured sector represents a functional category whose size indicates over‐ or‐under representation of genes in that cluster within a specific function. (D) Example of a selected region of interest (ROI) (Å∼200 µm) from a representative patient with advanced lung adenocarcinoma. A mask for immune cells was generated based on CD45 staining (red) and another mask for tumour was created based on PanCytokeratine staining (green) using GeoMx DSP. (E) Volcano plot of differentially expressed markers obtained by unpaired test with a Benjamini–Hochberg correction from the tumour compartment (PanCK+) of high‐IL‐11 versus low‐IL‐11 LUAD patients (cohort 4; HU12O, n = 20). Tumour samples were separated based on IL‐11 expression detected in tumour cells by immunohistochemistry and classified as low‐IL‐11 [IL‐11≤10% (N = 12)] or high‐IL‐11 [(IL‐11 > 10% (N = 8)]. (F) Total counts of immune cell markers of high‐IL‐11 and low‐IL‐11 expressing tumour samples patients. p Values were obtained by Mann–Whitney U test (*p < .05, **p < .01, ***p < .001).