Regions of hypoxia arise in 90% of solid tumors as cancer cells quickly proliferate, and scarce, newly formed vasculature fails to supply sufficient oxygen [1]. In patients with breast cancer, the mean partial pressure of oxygen (PO₂) in breast tumors ranges from 2.5 to 28 mmHg, with a median value of 10 mmHg (1% O₂), while normal human breast tissue has a median value of 65 mmHg (8% O₂) [2]. Several studies have demonstrated that patients with hypoxic tumors have an increased risk of metastasis and mortality [3,4,5]. Hypoxia has been reported to be an adverse prognostic indicator, independent of clinical stage, at the time of diagnosis [6]. Cancer cells can adapt and survive under oxygen deprivation, and the most well-reported mechanism involves the hypoxia-inducible factors (HIFs) [7]. In an O₂-rich environment, prolyl hydroxylases (PHDs) hydroxylate HIF-1α and HIF-2α. The von Hippel–Lindau (VHL) E3 ubiquitin ligase ubiquitinates hydroxylated HIF-1α and HIF-2α causing its proteasomal degradation [8]. In contrast, under hypoxia, HIF-1α and HIF-2α subunits become stabilized and bind to the HIF-1β subunit [9]. Both HIF-1 and HIF-2 heterodimers recognize and bind to the 5′-ACGTG-3′ enhancer sequence, resulting in the transcriptional regulation of more than a thousand genes [10,11]. The abundance and activity of both HIF-1α and HIF-2α can also be enhanced due to post-translational modifications [12]. HIF-regulated genes have been associated with angiogenesis, apoptosis, cell proliferation, cell survival, metabolism, invasion, metastasis, altered pH, and chemoresistance [5].

Hypoxia has been detected using both direct and indirect methods in cells cultured in the laboratory and animal and human tumors. Direct O₂ measurements have been made in solid tumors of cancer patients using needle-type O₂ electrodes [2,13,14,15]. Indirect methods, such as the immunolabeling of HIF-1/2α or downstream HIF-targets, have been used to detect potential regions of hypoxia in fixed tissue after surgical resection or biopsy [16]. Moreover, exogenous 2-nitroimidazole probes, such as pimonidazole, can be delivered to animals or humans and incorporate into hypoxic adducts that can be immunolabeled once the tissue is harvested and fixed [17]. Immunolabeling-based methods are insightful but limited by protein turnover and fixation artifacts, and they cannot be used for real-time assessment. For preclinical studies, several groups have used DNA constructs that cause the cell to express reporters, such as fluorescent proteins or luciferase, in a HIF-dependent manner as an indirect real-time readout of hypoxia [18,19]. In preclinical and clinical approaches, intratumoral O₂ levels are measured in real-time, albeit indirectly, using techniques such as Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) [5].

2. Current Insights

Intratumoral hypoxia has a frequent incidence in solid primary tumors and greatly contributes to tumorigenesis and metastasis. In the early 1990s, multiple studies investigated the O₂ levels in solid tumors utilizing polarographic probes, such as Clark probes, that are now commercially available from companies such as Eppendorf. They established that intratumoral hypoxia is an adverse indicator for patients with cancer [5]. Since then, in vitro models that progressed from 2D monolayer cell culture to 3D cellular clusters have been used to model physiological O₂ and nutrient gradients found in vivo. In vivo models faithfully reproduce the lack of vasculature and subsequent hypoxic regions in solid tumors. Although mouse models are a standard in cancer research, larger animals, such as rats [124] or rabbits [125], better resemble human scale and physiology and have been used to test O₂ detection methods. Furthermore, non-human primate models have been employed to assess biosafety parameters such as the biodistribution of radioactive hypoxia tracers for PET imaging [126] or to determine the effect of BOLD signal on neural activity [127]. As the relevance of intratumoral hypoxia increased, methods to assess O₂ levels both in vitro and in vivo became critical.

The selection of adequate methods to monitor O₂ levels in experimental set-ups must take several factors into account, namely whether they can be performed in live animals or cells versus fixed tissues and the scale and sensitivity of the measurement. To facilitate comparison across the different methods described in this review, we categorized the techniques according to a series of features that we consider relevant and their applicability in different experimental settings (Table 1). A method can be selected to fit the scientific question or the available resources.

Table 1. Compilation of methods to detect hypoxia. Partially adapted from [79,128]. Y = Yes; N = No; NA = Not Applicable; +/− = positive-negative readout. LM = Light Microscopy; FM = Fluorescent Microscopy; FC = Flow Cytometry; Temp = Temporal; Res = Resolution; Non-Inv = non-invasive.

Method	Detection	Live	Direct	Readout	Scale	Single Cell Res	Non -Inv	Dyna-mic	Temp. Res	In Vitro	Animal	Human	Processing
Endogenous markers	LM, FM, FC	N	N	+/−	μm [26]	Y	NA	N	NA	Y	Y	Y	Fixation Staining
Exogenous markers	LM, FM, FC	N	N	+/−	μm [66]	Y	NA	N	NA	Y	Y	Y	Fixation Staining
Fluorescent Reporter	FM, FC, Fluorescent imager	Y	N	+/− [79]	μm [79]	Y	Y	Y	ms [79]	Y	Y	N	Fixation Dissociation
NTR-sensitive Fluorescence	FM, FC, Fluorescent imager	Y	N	+/− [79]	μm [73]	Y	Y	Y	ms [79]	Y	Y	N	Fixation Dissociation
Phosphorescence CELI	FM, Fluorescent imager	Y	Y	pO2 [79]	μm [79]	N	Y	Y	s [128]	Y	Y	N	Pre-exposure
PAI	Ultrasound	Y	N	sO2 [79]	μm [79,129]	N	Y	Y	ms [79]	Y	Y	N	Pre-injection
BLI	Luminescent imager	Y	N	Intensity Gradient [130]	mm [77]	N	Y	Y	min [129]	Y	Y	N	Pre-injection
MRI	MRI machine	Y	N	B: deoxyHb [90] T: [O2(s)] [90]	mm [131]	N	Y	Y	s-min [128]	N	Y	Y	Pre-injection
EPRI	EPRI machine Spin tracers	Y	Y	pO2 [99]	mm [129,131]	N	Y	Y	min-hr [128,129]	N	Y	N	Pre-injection
PET	Radiolabeled Tracers	Y	N	radiotracer [129]	mm [131,132]	N	Y	Y	min-hr [133]	N	Y	Y	Pre-injection
Clark Electrode	Current meter	Y	Y	pO2 [121]	μm [121]	N	N	Y	s [121]	N	Y	Y	Implant/ Insertion
Invasive optical probes	Optical Detector	Y	Y	pO2 [121]	μm [121]	N	N	Y	ms [121]	Y	Y	Y	Insertion

While O₂-electrode or optical probes provide an accurate readout in situ, alternative methods have been established to overcome the need for an invasive approach to assess deeper tumors and preserve tissue integrity. Immunolabeling of hypoxia-regulated proteins in tissue sections has been extensively optimized to be performed in vitro and has also been employed in the clinic by staining human pathological specimens [134,135]. This particular approach is limited by rapid protein turnover and potential cross-regulation of these target proteins via mechanisms other than hypoxia. Another multifaceted approach requires the delivery of 2-nitroimidazole probes such as pimonidazole, which can be detected exclusively in hypoxic cells by immunolabeling. This method can be applied to basic cell culture or animal and human tissue sections [136]. However, immunolabeling techniques can only allow imaging of hypoxic regions when the tissue is resected and processed. Moreover, the staining is usually performed in a small portion of harvested and fixed tissue, which might not represent the entire tumor, particularly in the case of a human biopsy.

Phosphorescent and nitroreductase-regulated fluorescent probes are an easy-to-implement alternative to investigate dynamic changes in hypoxia. These probes can be added to cell culture medium or delivered to the animal intravenously to be immediately measured. This feature makes them readily marketable, and several options are available commercially. Basic scientists have taken advantage of transcriptionally regulated fluorescent, and luminescent reporters to monitor the dynamics of hypoxic and reoxygenated cancer cells. These approaches require extensive design and optimization, which can be time-consuming. However, fluorescent markers have allowed us to isolate hypoxic and reoxygenated cells to begin to unravel their contribution to tumor progression and metastatic spread.

Recent advances in medical imaging have optimized imaging modalities such as MRI and PET to specifically detect O₂ levels in a tumor. Basic scientists have kept up with these advances by utilizing these methods in preclinical models. Furthermore, in murine models, whole-body imaging has been extensively employed by utilizing BLI and PAI to enable time course studies in live animals. Although whole-body imaging allows dynamic monitoring of intratumoral hypoxia throughout long time-course experiments and better resembles clinical methods, the low spatial resolution prevents single-cell tracing.

Recently, promising efforts have been made to improve O₂ resolution and measure actual PO₂ levels using PET. Furthermore, both CELI and EPRI are high O₂-resolution (PO₂) emerging technologies with clinical potential that are currently being optimized to obtain FDA approval. In summary, whole-body imaging methods allow dynamic monitoring of O₂ levels in mice and are the most promising path for clinical use, but lower spatial and oxygen resolution or access to specialized high-cost equipment are limiting factors in preclinical research. Currently, PET offers the safest and cost-efficient option to be implemented in the clinic.

Hypoxia is a condition studied across multiple research fields, particularly in the context of developmental biology with focus on embryogenesis [137]. Aside from cancer, hypoxia also plays pathophysiological roles in myocardial ischemia, metabolic diseases, chronic heart and kidney diseases, and in reproductive disorders [138]. Although significant efforts have been conducted to detect and monitor hypoxic O₂ levels experimentally, it is still challenging. Many of the described methods are only a proxy for actual levels of O₂, can be challenging to adapt, or require access to specialized equipment. However, current techniques have been critical in establishing findings that have impacted the field of cancer research, namely: (1) the prognostic value of intratumoral hypoxia; (2) the hypoxic-upregulation of cancer hallmarks such as metabolism, migration, and chemoresistance; and (3) the significant contribution of cells that experienced intratumoral hypoxia to metastasis. As the role of hypoxia in cancer progression and resistance becomes more defined, multiple therapeutic approaches are being investigated pre-clinically and in clinical trials [139,140]. Thus, there is an urgent need to standardize methods to detect intratumoral hypoxia to (1) aid treatment decision and (2) assess whether hypoxia-targeted therapies perform as designed. The ultimate goal of this review is to aid researchers in selecting a suitable method to detect hypoxia in their studies by providing a side-by-side comparison across the most well-established and most recently optimized methods.