Hypoxia-inducible factor (HIF)–mediated effects of the hypoxic niche in bone cancer

Introduction

Oxygen is essential for cellular respiration and efficient energy production. Hypoxia can be defined as a reduction in the normal level of oxygen available to cells or tissues that arises when oxygen demand exceeds oxygen supply. It is frequently associated with pathological states such as cancer, ischemia, arthritis, and inflammation.

Polarographic needle electrode measurements estimate the normal range of oxygenation in most tissues as 24–66 mmHg (3%–8% O ₂ ) [ ]. The oxygen tension (pO ₂ ) is considerably lower in solid tumors including breast cancer (0–30 mmHg), cervical cancer (median 20 mmHg at stage 0, <5 mmHg at stage 2), and head-and-neck cancer (median pO ₂ 14.6 mmHg) [ ], rendering these tumors hypoxic. Most solid tumors have areas where the pO ₂ falls below 5 mmHg (0.6% O ₂ ) [ , ].

Due to the physical inaccessibility of the bone microenvironment, fewer measurements of oxygenation in human bone have been performed than in other tissues. The normal bone pO ₂ measured with polarographic needle electrodes in cancellous human mandibular bone is 71.4 mmHg (8.6% O ₂ ) [ ]. Oxygen levels in the bone marrow are lower. Human bone marrow aspirates measured in a blood gas analyzer record mean oxygenation levels of 54.9 mmHg (6.6% O ₂ ) [ ]. In comparison with normal bone, diseased bone is hypoxic. Diseased human mandibular bone records 32.3 mmHg (4% O ₂ ) in osteoradionecrosis and 28.4 mmHg (3.5% O ₂ ) in chronic osteomyelitis [ ].

This is supported by work in animal models. Recently, Spencer et al. used two-photon phosphorescence lifetime microscopy (2PLM) following injection of an oxygen-sensitive probe into live mice to confirm the lower pO ₂ in bone marrow. The average pO ₂ was approximately 50 mmHg (6% O ₂ ) in the periosteum and 30 mmHg (3.5% O ₂ ) in cortical bone. Bone marrow oxygenation ranged from 5 to 30 mmHg (0.6%–3.5% O ₂ ), with the lowest concentrations in the deep perisinusoidal regions distant from bone while the endosteal region exhibited higher oxygen levels [ ]. Diseased bone also becomes hypoxic in animal models of disease. In rabbits with osteoarthritic synovitis, the pO ₂ is lower in the synovial fluid (17.7 vs. 39.4 mmHg) and juxta-articular bone (38.2 vs. 60.3 mmHg) than in normal tissue [ ]. MRI-based noninvasive imaging techniques have identified regions of hypoxia in rat models of osteosarcoma [ ] and correlated tumor necrosis with increased risk of metastasis in Ewing's sarcoma [ ].

In summary, the bone microenvironment has a normal to low tissue pO ₂ to which resident bone cells are adapted. The higher oxygenation of the bone itself could be due to its low cellularity and therefore low oxygen consumption, whereas the considerable metabolic demand in the highly cellular bone marrow causes a relative reduction in oxygen availability. Currently, the most effective method used to demonstrate hypoxia in bone tumors is immunohistochemical detection of molecular markers of hypoxia. Hypoxia triggers complex adaptive responses and activates the expression of numerous genes, which are involved in pathways including erythropoiesis, angiogenesis, metabolic reprograming, cell cycle regulation, and tumorigenesis. The primary driver of the transcriptional response to hypoxia is hypoxia-inducible factor (HIF) and most studies investigating hypoxia in bone tumors to date have used HIF-1α or downstream targets of HIF as markers of hypoxia.

Hypoxia-inducible factor

HIF is a heterodimeric transcription factor comprising an inducible alpha subunit (HIF-α) and a constitutive beta subunit (HIF-β/aryl hydrocarbon receptor nuclear translocator [ARNT]). Under standard conditions, HIF-α is posttranslationally hydroxylated at two conserved proline residues in its oxygen-dependent degradation (ODD) domain by three HIF prolyl-4-hydroxylase enzymes (PHD1–3). This targets it for interaction with the von Hippel–Lindau protein (VHL), which polyubiquitinates HIF-α and targets it for proteasomal degradation. Separately, hydroxylation of an asparagine residue in the C-terminal transactivation domain (C-TAD) by factor-inhibiting HIF-1 (FIH) regulates the transcriptional activity of HIF-α [ ].

HIF-α accumulates under conditions causing either reduced PHD enzyme activity (e.g., hypoxia [ , ], oncogenic mutation [ ]) or enhanced translation of HIF-α that exceeds the substrate capacity of the PHD enzymes (e.g., exposure to insulin [ ] or hepatocyte growth factor (HGF) [ ]). HIF-α protein then translocates to the nucleus, dimerizes with HIF-β and other cofactors, and binds the hypoxia-response element (HRE) in the promoter of HIF target genes to initiate transcription [ ]. HIF transcriptional activity is increased via reduced asparaginyl hydroxylation by FIH.

Humans have three distinct HIF-α genes. Knowledge of HIF is mostly derived from studies on HIF-1α, which is the main focus of this chapter. HIF-2α regulates different but overlapping genes and modulates the HIF transcriptional response in a tissue-specific manner [ ]. Less is known about HIF-3α, partly because the gene encodes many different proteins with vastly different functions. Full-length HIF-3α is an oxygen-regulated activator of transcription, HIF-3α4/IPAS is a dominant-negative regulator of HIF-1/2α, and other variants inhibit HIF-1/2α by competing for HIF-β [ ].

Bone cancer

Bone cancer comprises two broad categories: cancers that arise in bone (primary bone tumors, known as bone sarcomas) and those that metastasize to bone from other primary tumor sites (commonly breast, lung, or prostate cancer). Here we will focus on osteosarcoma, the most common primary malignant tumor of bone, briefly mention highlights from other bone sarcomas, and then discuss effects of hypoxia and HIF in metastatic bone disease.

Osteosarcoma

Osteosarcoma mainly arises in the metaphysis of the long bones of children and adolescents. Approximately 90% of osteosarcomas are aggressive high-grade tumors requiring neoadjuvant chemotherapy to address systemic spread already present at diagnosis. Currently only 60%–70% of patients survive, mainly due to metastatic and/or recurrent disease [ ].

HIF is a prognostic factor in osteosarcoma

Yang et al. were the first to report that HIF-1α predicts poor clinical outcome in osteosarcoma. Nuclear expression of HIF-1α protein occurred in 79% of osteosarcomas, correlated with surgical stage, and was indicative of shorter overall and disease-free survival; adjacent noncancerous tissue was HIF-1α-negative [ ]. Expression of HIF-1α in osteosarcoma has since been associated with high tumor grade [ , ], tumor recurrence [ ], distant metastases [ ], and poor overall survival [ , ]. Several meta-analyses confirmed that HIF-1α is an effective prognostic biomarker predicting overall survival and poor clinical outcome in osteosarcoma [ ]. Additionally, levels of HIF-regulated vascular endothelial growth factor (VEGF) are higher in the tumor and serum of patients who subsequently relapse, tumor VEGF being predictive of pulmonary metastasis and poor prognosis [ , ].

Only one study describes the prognostic significance of HIF-2α. Li et al. observed overexpression of HIF-2α mRNA in osteosarcoma tissue that correlated with tumor size, advanced clinical stage, distant metastases, and reduced overall and disease-free survival [ ].

HIF affects tumor cell proliferation and apoptosis

Osteosarcoma cell lines stabilize HIF-1α and HIF-2α in hypoxia, both of which transcribe HIF target genes such as VEGF and phosphoglycerate kinase 1 (PGK1) [ ]. HIF target genes promote the hypoxic proliferation of osteosarcoma cells, including angiopoietin-like 4 (ANGPTL4) [ ], ANGPTL2 [ ], carbonic anhydrase IX (CAIX) [ ], platelet-derived growth factor-BB (PDGF-BB) [ ], arginase II [ ], and the long noncoding RNAs urothelial carcinoma associated 1 (UCA1) [ ] and FOXD2-AS1 [ ] Fig. 25.1 . Reciprocal positive feedback loops between HIF-1α and both SUMO-specific protease 1 (SENP1) [ ] and the serine/threonine protein kinase AKT [ ] promote proliferation. Activation of AKT is common to many reports of HIF-regulated osteosarcoma cell proliferation: AKT, ERK1/2 and STAT3 signaling is induced by PDGF-BB [ ]; UCA1 promotes phosphorylation and activation of AKT while suppressing expression of the tumor suppressor PTEN [ ]; and HIF-1α promotes upregulation of cyclin D1 protein via AKT [ ].

Antiapoptotic effects of hypoxia in osteosarcoma are often mediated by similar signaling pathways, including the HIF-1α/SENP1 [ ] and HIF-1α/AKT/cyclin D1 [ ] positive feedback loops. Additionally, FOXD2-AS1 interacts with enhancer of zeste homologue 2 (EZH2) to epigenetically silence the transcription of p21 [ ], which normally drives apoptosis following caspase activation.

Fewer reports relate to HIF-2α, describing opposing effects on hypoxic cell proliferation in different osteosarcoma cell lines [ , ] and potential proapoptotic effects due to modulation of MAPK-p38 signaling [ ].

Proangiogenic effects of HIF

Tumor growth and metastasis depend on angiogenesis, the formation of new blood vessels from preexisting vessels. HIF drives the expression of proangiogenic factors by tumor cells including VEGF, angiopoietin-2 (Ang2), basic fibroblast growth factor (bFGF), ANGPTL4, and PDGF [ ]. However, there is limited specific evidence for proangiogenic effects of HIF in osteosarcoma.

Expression of HIF-1α and VEGF correlates with mean vessel density in human osteosarcoma [ ]. Saos-2 osteosarcoma cells treated with HIF-1α siRNA downregulate VEGF to cause reduced blood vessel formation in inoculated tumors in vivo [ ]. Tetrahydrocurcumin (THC) inhibits HIF-1α expression in MG-63 and U2OS osteosarcoma cells by suppressing Akt/mTOR and p38 MAPK signaling, inhibiting hypoxic expression of VEGF and matrix metalloproteinases (MMPs), and reducing tube formation by endothelial cells [ ]. HIF-regulating PHD3 is itself induced by hypoxia. Silencing of PHD3 does not directly affect HIF-1α or HIF target genes in osteosarcoma cells but causes phenotypic changes in the tumor vasculature following subcutaneous injection of PHD3-depleted LM8 osteosarcoma cells in mice (reduced vessel density, increased vessel size) that are reversed by silencing of PDGF-C [ ].

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here