Intratumor HIF-1α modulates production of a cachectic ligand to cause host wasting

Abstract

Tumor-host interactions play critical roles in cancer-associated cachexia. Previous studies have identified several cachectic proteins secreted by tumors that impair metabolic homeostasis in multiple organs, leading to host wasting. The molecular mechanisms by which malignant tumors regulate the production or secretion of these cachectic proteins, however, still remain largely unknown. In this study, we used different Drosophila cachexia models to investigate how malignant tumors regulate biosynthesis of ImpL2, a conserved cachectic protein that inhibits systemic insulin/IGF signaling and suppresses anabolism of host organs. Through bioinformatic and biochemical analysis, we found that hypoxia-inducible factor HIF-1α/Sima directly binds to the promoter region of ImpL2 gene for the first time, promoting its transcription in both tumors and non-tumor cells. Interestingly, expressing HphA to moderately suppress HIF-1α/Sima activity in adult yki ^3SA gut tumors or larval scrib ¹ Ras ^V12 disc tumors sufficiently decreased ImpL2 expression and improved organ wasting, without affecting tumor growth. We further revealed conserved regulatory mechanisms conserved across species, as intratumor HIF-1α enhances the production of IGFBP-5, a mammalian homolog of fly ImpL2, contributing to organ wasting in both tumor-bearing mice and patients. Therefore, our study provides novel insights into the mechanisms by which tumors regulate production of cachectic ligands and the pathogenesis of cancer-induced cachexia.

Graphical abstract

Fig. 1

Sima regulates ImpL2 transcription. (A) Heatmap showing up-regulated Sima target genes in the midguts of Con and yki^3SA flies after tumor induction for 8 days. (B) Expression levels of the ImpL2 (left) and sima (right) isoforms in guts of flies bearing yki^3SA-gut tumors at day 8 from RNA-seq data (GSE113728, NCBI-GEO) (n = 3). (C) Regions containing putative HRE sites (5'-(A/G)CGTG-3′) within the ImpL2 promoter. (D) Representative images of ImpL2 (red) and Sima.GFP (green) expression in both control and yki^3SA-tumor midgut. (E) Expression levels of ImpL2 and putative Sima targets in S2R + cells with Sima overexpression under normoxia (N, 5% O₂) or hypoxia (H, 1% O2) for 24 h (n = 3). (F-G) Relative luciferase (Luc) activities of the indicated vectors that contain normal (F, ImpL2-10K/5.5K/2.8K/1.4K) and mutated (G, 10K.HRE1-7 mutation) Sima binding sites with Sima overexpression or/and desferrioxamine (DFO, 100 μM) treatment for 24 h in S2R + cells (n = 3). (H) S2R + cells were transfected with the indicated vectors, followed by lysis and immunoprecipitation for subsequent in vitro ChIP assays. Sima occupancy at the indicated ImpL2 promoter regions was quantified by qPCR analysis relative to input DNA. Data were presented as means ± SEM. Unpaired Student's t-test and one-way ANOVA followed by post hoc test were performed to assess differences. ∗p < 0.05.

Fig. 2

Sima-ImpL2 axis is required fortumor-inducedhost wasting. (A)-(G) Bloating phenotypes (A, up), gut tumors (A, middle, GFP) and ovaries (A, bottom), bloating rates (D, n = 3, 20 flies/replicate) and TAG and trehalose levels (E, n = 3, 5 flies/replicate), ovary sizes (F, n > 10), and climbing rates (G, n = 3, 20 flies/replicate) of Con and yki^3SA flies under yki^3SA-tumor induction for 8 days with or without HphA overexpression. (B)-(C) Gene expression in midguts (B, n = 3, 10 flies/replicate) and phosphorylated Akt (pAkt) in thoraces (C) of flies bearing yki^3SA-gut tumors with or without HphA overexpression at day 8. (H) Lifespans of these tumor-bearing adult flies (n = 100). Data were presented as means ± SEM. Unpaired Student's t-test and one-way ANOVA followed by post hoc test were performed to assess differences. Lifespan was assessed using the log-rank test. ∗p < 0.05.

Fig. 3

HIF-1α removal in tumors alleviates host organ wasting. (A)-(B) HIF-1α deficiency, as indicated by immunoblot (A) and HIF-1α targets expression (B) in shGFP or shHif1α C26 cells in response to normoxia or hypoxia. (C)-(E) Body weights (C), weights of gastrocnemius (Gas) and tibialis anterior (TA) muscles (D, left) and inguinal (iWAT) and epididymal (eWAT) adipose tissues (D, right), and expression of genes associated with muscle atrophy in TA muscles (E) of mice with PBS or C26 cells subcutaneous (sc) injection for 25 days (n = 6). (F) Initial s.c. injection of shGFP (1.6 × 10⁶ cells) or shHif1α C26 (5 × 10⁶ cells/mouse) in CD2F1 mice. (G)-(H) Tumor morphologies (G) and sizes (H) of mice bearing with shGFP or shHif1α C26 cells for 25 days (n = 6). (I)-(M) Gene expression in tumors (I) and TA muscles (J), body weights (K), muscle (L, left) and fat weights (L, right) and morphologies (M) of mice bearing shGFP or shHif1α C26 tumors for 25 days (n = 6). Data were presented as means ± SEM. Unpaired Student's t-test was performed to assess differences. ∗p < 0.05.

Fig. 4

Tumor-derived IGFBP-5 causes organ wasting. (A) Immunoblots indicating IGFBP-5 expression in 3T3-L1 adipocytes, C2C12 myotubes, and C26 cells. (B) Differentiation (MHC, red) of C2C12 myoblasts treated with 1 μg/mL IGFBP-5 (BP5) in the differentiation medium for 5 days. (C)-(E) The widths and morphologies (C), gene expression (D, n = 3), and protein degradation rates indicated by ³H-Tyrosine release (E, n = 3) of C2C12 myotubes that were differentiated for 5 days and subsequently treated with fresh differentiation medium containing 1 μg/mL IGFBP-5 (BP5), C26-conditioned differentiation medium (CM), or CM containing 100 ng/mL anti-IGFBP-5 antibodies (Ab) for 2 days. (F) Lipolysis rates indicated by released glycerol in differentiated 3T3-L1 adipocytes that were treated 2 μg/mL IGFBP-5 (BP5), or C26-conditioned medium (CM) with or without anti-IGFBP-5 antibodies (Ab) (right, n = 3) for 24 h. (G)-(I) IGF signaling responses, as indicated by pAkt, in differentiated 3T3-L1 adipocytes that were treated with 100 ng/mL IGF-1 plus 1 μg/mL IGFBP-5 (BP5) in growth medium for 1 h (G), or 1 μg/mL IGFBP-5 (BP5) in growth medium at different time points (H), or C26-conditioned growth (CM) medium containing anti-IGFBP-5 antibodies (Ab) for 1 h (I). (J)-(K) Immunoblots indicating IGFBP-5 expression in C26 cells (K) with frameshift-associated IGFBP5 knockout (BP5KO) caused by a single-base deletion in the Igfbp5 coding region (J). (L)-(N) Tumor weights (L), initial injection and post-dissection body weights (M), muscle (N, left) and fat (N, right) weights of mice bearing control or BP5KO C26 tumors for 25 days (n = 6). Data were presented as means ± SEM. Unpaired Student's t-test and one-way ANOVA followed by post hoc test were performed to assess differences. ∗p < 0.05.

Fig. 5

HIF-1α-IGFBP-5 axis is associated with cachexia in pancreatic cancer patients. (A) Immunofluorescence staining indicating HIF-1α (green) and IGFBP-5 (red) levels in tumors from patients with benign pancreatic diseases (up, n = 10) and malignant pancreatic tumors with (n = 25) or without (n = 17) weight decline. (B)-(C) Quantification of fluorescence intensities of intratumor HIF-1α (B) and IGFBP-5 (C). (D) Positive correlation between fluorescence intensities of HIF-1α (B) and IGFBP-5 (C) in all pancreatic cancer patients. Data were analyzed by Prism 10.0 software using Pearson correlation analysis. Data were presented as means ± SEM. Unpaired Student's t-test and one-way ANOVA followed by post hoc test were performed to assess differences. ∗p < 0.05.