Hypoxia-dependent modification of collagen networks promotes sarcoma metastasis

Abstract

Intratumoral hypoxia and expression of hypoxia-inducible factor-1α (HIF-1α) correlate with metastasis and poor survival in patients with sarcoma. We show here that hypoxia controls sarcoma metastasis through a novel mechanism wherein HIF-1α enhances expression of the intracellular enzyme procollagen-lysine, 2-oxoglutarate 5-dioxygenase 2 (PLOD2). We show that loss of HIF-1α or PLOD2 expression disrupts collagen modification, cell migration, and pulmonary metastasis (but not primary tumor growth) in allograft and autochthonous LSL-Kras(G12D/+); Trp53(fl/fl) murine sarcoma models. Furthermore, ectopic PLOD2 expression restores migration and metastatic potential in HIF-1α-deficient tumors, and analysis of human sarcomas reveals elevated HIF1A and PLOD2 expression in metastatic primary lesions. Pharmacologic inhibition of PLOD enzymatic activity suppresses metastases. Collectively, these data indicate that HIF-1α controls sarcoma metastasis through PLOD2-dependent collagen modification and organization in primary tumors. We conclude that PLOD2 is a novel therapeutic target in sarcomas and successful inhibition of this enzyme may reduce tumor cell dissemination.

Significance: Undifferentiated pleomorphic sarcoma (UPS) is a commonly diagnosed and particularly aggressive sarcoma subtype in adults, which frequently and fatally metastasizes to the lung. Here, we show the potential use of a novel therapeutic target for the treatment of metastatic UPS, specifi cally the collagen-modifying enzyme PLOD2.

Figure 1. HIF1α is an important regulator of metastasis in an autochthonous, genetic model of UPS potentially via PLOD2 modulation

(A) (Left and Middle Panels) Relative gene expression in human metastatic (N=5) and non-metastatic (N=8) UPS and fibrosarcoma in patients treated as Massachusetts General Hospital (32). HIF1α (P=0.0312) and PLOD2 (P =0.0011) were significantly upregulated in metastastic sarcomas. (Right Panel) qRT-PCR analysis of 10 human UPS patient samples treated at the University of Pennsylvania; P=0.044. (B) Mouse models of sarcoma. LSL-Kras^G12D/+;Trp53^fl/fl (KP) and LSL-Kras^G12D/+;Trp53^fl/fl;Hif1α^fl/fl (KPH) genotyping showed effective recombination of Hif1α^fl alleles in Adeno-Cre initiated tumors. (C) Mice remained tumor free for roughly 40 days, by 90 days all of the mice had developed palpable tumors (volume= 200mm³) KP; n=30, KPH; n=20; (P=0.3755). (D) Primary tumor size. Two weeks after tumors were palpable they had grown 2-8 fold larger, but there was no difference between KP; n=7 and KPH; n =9 tumor growth (P=0.7342) (E) Metastasis free survival in KP; n=33 and KPH; n=28 mice; P=0.0456. Lung metastases were confirmed histologically. (F) Masson's Trichrome and picrosirius red staining of tumor nest areas and blood vessels in primary KP and KPH tumors. Deletion of HIF1α alters collagen in KPH tumors. Masson's Trichrome stains collagen fibers blue. Cells were counterstained in red with Weigert's hematoxylin. (Scale bar;50 μm). (G) Western blot analyses of sarcoma cells derived from KP and KPH tumors. Expression of HIF1α and PLOD2 proteins is hypoxia inducible and abolished when HIF1α is deleted. (H) qRT-PCR analysis of 2 individually derived KP and KPH cell lines. Plod2 mRNA transcription is induced under hypoxic conditions in control cells (KP1; P =0.0284 and KP2; P =0.0391). Deletion of HIF1α abolished hypoxia-induced Plod2 mRNA levels. (I) Western blot of PLOD2 expression in KIA cells and (J) HT-1080 cells. (K) qRT-PCR analyses of KIA cells. Expression of Hif1α and Plod2 is hypoxia inducible (Plod2 qRT-PCR: P =0.0284) and is abolished when HIF1α is deleted (Plod2 qRT-PCR: P =0.0403). (L) HT-1080 cells were evaluated by qRT-PCR as in (K). Expression of Hif1α and Plod2 is hypoxia inducible (Plod2 qRT-PCR: P =0.0006) and is abolished when HIF1α is deleted (Plod2 qRT-PCR: P =0.0210).

Figure 2. HIF1α and PLOD2 are dispensable for primary sarcoma formation but essential for metastasis

(A) Tumor allograft using 1×10⁶ Scr, HIF1α-deficient, or PLOD2-deficient KIA cells subcutaneously injected into flanks of nude mice. n=5 mice per group, 2 tumors per mouse, i.e. 10 tumors per shRNA treatment. (B) Tumor weight was determined upon dissection of euthanized animals. PLOD2-deficient tumors were slightly larger than Scr tumors (P=0.0495) and HIF1α-deficient tumors (P =0.0131). (C) Lungs from mice bearing KIA transplanted subcutaneous tumors. H&E staining revealed the presence of numerous metastases in control tumors (large purple areas) but very few in lungs from animals bearing PLOD2- and HIF1α-deficient tumors. (D) % Tumor burden was evaluated with ImagePro7 software; (P< 0.39) (E). Loss of HIF1α in the primary tumor significantly reduced the total number of sarcoma foci in lungs (P <0.0001) (left panel) calculated over multiple experiments and the average number of sarcoma foci/lung (P =0.0500) (middle panel). Loss of PLOD2 also significantly reduced the total number of sarcoma foci in lungs (P <0.0001) and the average number of sarcoma foci/lung (P=0.0453). HIF1α and PLOD2 depletion in primary tumors also decreased the % of lungs with sarcoma foci. (right panel) (F) H&E, HIF1α, Hypoxyprobe, and Lectin staining were used to characterize control KIA tumors using a 50x objective (scale bar; 200 μm) and a 200x objective (scale bar; 100 μm). Black boxes indicate enlarged areas; arrow (left panel) indicates HIF1α-positive cells. Arrow (right panel) indicates lectin-positive cells of a blood vessel. (G) Picrosiruis stain was used to characterize collagen organization in subcutaneous KIA tumors. Deletion of HIF1α and PLOD2 altered collagen organization. Scale bar=50 μm. (H) 4-12% NuPage Bis-Tris (top panel) and 3-8% Tris-Acetate (middle panel) gradient gels were used to characterize collagen structure in KP cells treated with scr, HIF1α, or PLOD2 shRNA. *monomers, **dimers, and ***trimers of collagen I were detected. Confirmation of PLOD2 and HIF1α deletion was established using standard 10% SDS-PAGE gels (bottom panel). Hypoxia exposure lasted 16 hours. (I) Hydroxyproline modifications on collagen were measured using acid hydrolyzed tumor tissue. Deletion of HIF1α (P= 0.0536) and PLOD2 (P=.0102) increased hydroxyproline levels compared with scramble control.

Figure 3. HIF1α and PLOD2 mediate sarcoma cell migration via a cell extrinsic mechanism

(A) Scratch migration assays of confluent, and therefore oxygen and nutrient-limited, KIA cells stably expressing Scr, HIF1α, or PLOD2 specific shRNAs and either copGFP (HIF1α, PLOD2) or dsRed (Scr). Green and Red cells were mixed 1:1. (B) Quanitfication of recovery in (A) (all P values are ≤ 0.0105). (C) Western blot analyses of KIA cells treated as in (A) and (B). ShRNA-mediated knockdown of HIF1α and PLOD2 shown here also reflects knockdown occurring in panels (A) and (B) as cell lines generated for these assays were then transduced with copGFP lentivirus of dsRed lentivirus. (D) Proliferation of KIA cells expressing Scr, HIF1α, or PLOD2 specific shRNAs under hypoxic conditions. Cells were counted daily.

Figure 4. PLOD2 expression rescues sarcoma migration but not invasion

(A) Quantification of migration assays of normoxic and hypoxic KIA cells expressing Scr or HIF1α specific shRNAs and ectopically expressing control or wild type human PLOD2 cDNA. All P values are ≤ 0.0432) (B) Quantification of HT-1080 migration assays performed as in (A) except murine Plod2 was ectopically expressed. All P values are ≤ 0.0431. (C) Representative images of boyden chamber migration assay using HT-1080 cells treated as in (B). (scale bar; 50 μm). (D) Quantification of invasion assays of normoxic and hypoxic KIA cells expressing Scr or HIF1α specific shRNAs and ectopically expressing control or wild type human PLOD2 cDNA using matrigel coated transwell invasion chambers. All P values are ≤ 0.0084. (E) Quantification of HT-1080 invasion assays performed as in (D) except murine Plod2 was ectopically expressed. All P values are ≤ 0.0017. (F) Representative images of matirgel coated chamber invasion assay using KIA cells treated as in (D) (scale bar; 50 μm).

Figure 5. PLOD2 control of cell migration and metastasis is dependent upon its lysyl hydroxylase activity

(A) Quantification of migration assay of normoxic and hypoxic KIA cells expressing Scr or HIF1α specific shRNAs and ectopically expressing control or mutant human PLOD2 D668A cDNA (all P values < 0.0001). (B) Quantification of migration assay of normoxic and hypoxic HT-1080 cells expressing Scr or HIF1α specific shRNAs and ectopically expressing control or mutant murine Plod2 D689A cDNA (all P values ≤ 0.0106). (C) Scratch migration assays of HT-1080 cells stably expressing copGFP in the presence or absence of 0.5 mM Minoxidil pretreatment for 48 hrs. (D) Quantification of recovery from (C) (P values are ≤ 0.0058). (E) Western blot analyses of HT-1080 and KIA cells treated as in (C,D). (F) Tumor allograft growth using 1x10⁶ KIA cells subcutaneously injected into flanks of nude mice. n=10 mice per group, 2 tumors per mouse, i.e. 20 tumors per treatment with vehicle or Minoxidil. (G) Lungs from mice bearing KIA transplanted subcutaneous tumors treated with PBS or Minoxidil. H&E staining revealed the presence of numerous metastases in control tumors (large purple areas) but very few in lungs from animals treated with Minoxidil. (H) Intra-peritoneal Minoxidil treatment reduced the average number of sarcoma foci/lung. (I) Picrosirius red staining of KIA subcutaneous tumors from (F). Minoxidil treatment altered collagen organization in the primary tumors, (scale bar; 50 μm).

Figure 6. Sarcoma cell migration, access to vasculature, and metastasis are dependent on HIF1α/PLOD2-mediated production of disorganized collagen in vivo

(A) Masson's Trichrome staining and SHG of collagen in Scr and HIF1α-deficient KIA tumors. Images of various tumor areas were taken, including areas of significant collagen deposition (collagen deposit). Scale bars for 50x images represent 200 μm and scale bars for 400x images represent 50 μm. SHG: collagen (red); tumor cells (green). Arrows indicated areas where tumor cells are elongated and adhere to collagen fibers. (B) Masson's Trichrome staining and SHG of collagen in Scr and HIF1α-deficient KIA tumors. Images show areas lacking large amounts of collagen (tumor nest), and tumor vasculature (blood vessel). (C) Quantification of collagen deposition in Scr and HIF1α-deficient KIA tumors. ImagePro7software subtracted red hues from images, leaving only blue (collagen) stain to be measured. Collagen to be quantified is outlined in green, and ImagePro7 software calculated these areas P <0.0001 (lower left panel). Number of collagen intersects/blood vessel was counted manually from 12 images and 4 separate primary tumors P =0.0016. (D) Masson's Trichrome staining and SHG of collagen in Scr and PLOD2-deficient KIA tumors. Scale bars for 50x images represent 200 μm and scale bars for 400x images represent 50 μm. Arrows indicated areas where tumor cells are elongated and adhere to collagen fibers. (E) Masson's Trichrome staining of control and Minoxidil treated KIA tumors. Arrow indicates the presence of collagen and tumor cells in the vasculature. Scale bars for top row 50x images represent 200 μm and scale bars for remaining row 400x images represent 50 μm.

Figure 7. Expression of PLOD2 restores metastasis in animals bearing HIF1α-deficient sarcomas

(A) Tumor volume from Scr, and HIF1α-deficient, as well as HIF1α-deficient tumors that stably express the wild-type PLOD2 expression vector (rescue) n=6 mice per group, 2 tumors per mouse; i.e. 12 tumors per shRNA treatment. (B) H&E staining of lungs from Scr, HIF1α-deficient, and rescue tumor groups (HIF1α shRNA+ PLOD2). Metastases are stained dark purple. (C) Quantification of lung metastases from tumors in (B). (left panel) Total number of sarcoma foci in lungs from HIF1α-deficient tumors is decreased compared to Scr (P=0.0083) and to HIF1α + PLOD2 cDNA (P =0.0199). (middle panel) Average number of sarcoma foci/lung is decreased in HIF1α-deficient tumors P =0.0132. (right panel) % of total lungs containing sarcoma foci in all three groups. (D) Model of hypoxia-dependent effects on collagen and metastasis in sarcomas.