Imaging Tumoral Hypoxia: Oxygen Concentrations and Beyond
Imaging Tumoral Hypoxia: Oxygen Concentrations and Beyond
The presence of hypoxic cells within tumors and the implications of this situation for cancer therapy were first noted over 50 years ago by Thomlinson and Gray. Since that time, a plethora of studies have confirmed the existence of hypoxic cells within a variety of tumor types as well as the prognostic importance of hypoxia in the clinical management of human cancers. The role of hypoxia as a key determinant of outcome for human cancers has encouraged efforts to noninvasively detect and localize regions of poor oxygenation in tumors. In this review, we will summarize existing and developing techniques for imaging tumoral hypoxia.
Biology of Tumor Hypoxia
Most tumors possess (1) lower oxygen levels than their corresponding tissue of origin, (2) significant intra- and intertumoral variation in oxygen concentrations, and (3) lower oxygenation at the time of recurrence than the corresponding primary tumor.[2,3] Hypoxic tumors are more refractory to therapy and are associated with an aggressive tumor phenotype. For example, the "oxygen effect" has been well studied and dictates that the radiation dose required to kill a given fraction of cells in a cell population fully deprived of oxygen is 2.5 to 3 times greater than that required to achieve an equal amount of cell kill in an aerobic cell population.[4-6] Hypoxia-induced resistance to chemotherapy including cyclophosphamide, carboplatin, and doxorubicin,[7,8] have also been noted. Tumor hypoxia has been associated in both laboratory and clinical studies with a more aggressive neoplastic phenotype as well as with increased potential for invasion, growth, and metastasis.[9-16]
A cascade of alterations in cellular gene expression are induced by the absence of oxygen. A central player in the ability of tumor cells to adapt to a low oxygen environment is the hypoxia-inducible factor 1 (HIF-1). HIF-1 is a heterodimeric protein consisting of an oxygen-regulated alpha subunit and a constitutively expressed beta subunit. In the presence of oxygen, a family of prolyl hydroxylases are activated and hydroxylate two proline residues of HIF-1-alpha (P402 and P564).[18-20] Following hydroxylation, HIF-1-alpha is then further modified by the addition of ubiquitin molecules through association with a complex consisting of the von-Hippel Lindau (VHL) protein and elongins B and C, after which the protein is rapidly degraded by the proteasome. In cells with wild-type expression of all components of this pathway, HIF-1-alpha therefore only persists under hypoxic conditions where the oxygen-dependent degradation pathway is inactive due to inhibition of the prolyl hydroxylases that require molecular oxygen. In these situations, HIF-1-alpha combines with HIF-1-beta to generate the composite transcription factor HIF-1, and selectively activate transcription of genes bearing promoters containing HIF-1-binding hypoxia-response elements (HREs).
A number of clinical studies have shown that increased expression of HIF-1-alpha is a significant negative prognostic indicator for many types of cancer, including brain,[22,23] breast,[24-26] cervix,[27-29] esophagus, head and neck,[31,32] and uterus. These observations have encouraged the development of a variety of anti-HIF-1 cancer therapeutics.[34,35] It has furthermore been observed that HIF-1-alpha can be found in oxic as well as hypoxic conditions in many types of tumor cells. This finding provides evidence of oxygen-independent regulation of HIF-1 in tumor cells, a process that is partly mediated by loss and/or gain of function mutations in a number of genes relating to the degradation pathway described above, such as VHL.[34,35]
A variety of growth factors including insulin and insulin-like growth factor 1 (IGF-1) have also been shown to stimulate oxygen-independent accumulation of HIF-1 through the phosphatidylinositol 3-kinase (PI3K)/Akt/mTOR pathway.[37-39] Overexpression of HIF-1 and its downstream genes, including vascular endothelial growth factor (VEGF), glucose transporter 1 (GLUT-1), lysyl oxidase (LOX), and carbonic anhydrase IX (CA IX), appears to be at least partly responsible for the phenotype of increased aggressiveness, invasion, metastasis, and resistance to therapy associated with hypoxic tumor cells.[35,40]
Gatenby and Gillies have speculated that hypoxic conditions in premalignant lesions may select for tumor cells overexpressing HIF-1 because of the survival advantage it provides. This could produce adult tumors with elevated HIF-1 levels under both oxic and hypoxic conditions that would therefore constitutively exhibit an aggressive phenotype. Two recent experimental studies have suggested that the absence of HIF-1-alpha confers therapeutic sensitivity independent of tissue oxygen levels, indicating that HIF-1-alpha is an effector and not merely a surrogate of hypoxia-mediated therapeutic resistance.[42,43] Thus, HIF-1 and hypoxia represent two related but independently significant aspects of cancer biology. Correspondingly, imaging methods and therapies targeting HIF-1 and/or HIF-1-alpha in tumors, while not necessarily specific to oxygenation status, may nonetheless be directed toward neoplastic cells with a more malignant phenotype.
Measuring Tumoral Oxygenation
A variety of techniques have been advanced for the identification and quantification of tumoral hypoxia. A compilation of these methods is given in Table 1. The past gold standard for measurement of oxygen levels in tumors is the Eppendorff needle electrode, which applies a voltage to cause the reduction of molecular oxygen to hydroxide. The current produced is then measured and related to the concentration of oxygen within the measurement volume. Alternative probe-based oxygen measurement strategies have been developed, including systems based on oxygen-sensitive fluorophores such as the OxyLite.
In addition, immunohistochemical methods have been applied to detect markers of hypoxia in biopsy samples taken from cancer patients. These markers include both exogenous, systemically delivered probes that are administered to a patient, localize in hypoxic regions, and are detected by antibody techniques in tissue specimens (pimonidazole,[46,47] EF5), as well as endogenous proteins that are overexpressed under hypoxic conditions (HIF-1-alpha, CA IX). These methods bear restrictions to their applicability and utility because of their invasive nature, the possibility of sampling error in relating discrete and/or microscopic observations to a macroscopic, potentially heterogenous tumor, and the inability to perform serial measurements needed to assess changes in oxygenation during treatment.