Background

Oxygen availability can impact on many signaling pathways and small changes to the level of pericellular oxygen can alter growth rate, metabolism, free radical production, and general cell behavior (Place et al., 2017; Bordt, 2018). Therefore, to preserve normal cellular function, oxygen levels need to be maintained within a relatively narrow ‘normoxic’ range. This applies both to cells in vivo and to cell lines used in cell culture (Hunyor and Cook, 2018).

In most cell culture models, experimental ‘normoxic’ values are based on the oxygen levels supplied in cell culture incubators. In dry air at sea level, atmospheric oxygen is ~20.9% v/v (Hunyor and Cook, 2018). In standard humidified cell culture incubators maintained at 5% CO₂, oxygen levels are equal to 18.6% O₂ (v/v), equivalent to an oxygen tension or oxygen partial pressure (pO₂) of 138 mmHg (Wenger et al., 2015). However, cells are maintained at much lower pO₂ levels in vivo (Carreau et al., 2011). Under normal ventilatory conditions, pO₂ in the arterial blood ranges from 80-100 mmHg (10.8-13.5% O₂ v/v) and this is further lowered as oxygen molecules diffuse through tissue and are consumed by cells (Carreau et al., 2011). Moreover, different tissues have unique ranges of physiological normoxia, depending on the specific oxygen demands of individual tissues. Using a single oxygen concentration to represent physiological normoxia for a range of cell types is not representative of the in vivo environment (Carreau et al., 2011). When studying pathological oxygen conditions (such as in ischemia or in tumor hypoxia) cells should be exposed to oxygen tensions that appropriately model the specific pathological state of the tissue. In brain ischemia, pO₂ levels can decrease below 10 mmHg (1% v/v) (Hoffman et al., 1999). In tumors, pO₂ levels can range from 15 mmHg (2.1% v/v) to as low as 2 mmHg (0.3% v/v) and this can also depend on the tissue of origin (Byrne et al., 2014).

The concentration of oxygen used to represent hypoxia is important because oxygen-sensing proteins have optimal pO₂ ranges for which they operate. Hypoxia inducible factor (HIF) is a transcription factor activated by low oxygen. HIF is thought to control the expression of up to 1,000 genes and it has important roles in health and disease including embryogenesis (Dunwoodie, 2009), angiogenesis (Cook and Figg, 2010), cancer (Cook et al., 2009), diabetes (Thangarajah et al., 2010), inflammation and immune function (Palazon et al., 2014). HIF-1α and HIF-2α have been found to be stabilized under different oxygen tensions (Holmquist-Mengelbier et al., 2006; Lin et al., 2011). In neuroblastoma, HIF-2α is stabilized in 5% O₂ and can be strongly expressed even in well-vascularized areas, while HIF-1α is stabilized below 1% O₂ in extremely hypoxic regions of the tumor (Holmquist-Mengelbier et al., 2006). A small change in oxygen levels can therefore alter the balance of HIF-1α and HIF-2α, which affects what genes are expressed and ultimately how the cells behave. As we build better in vitro systems to accurately model in vivo conditions, oxygen levels are a key experimental variable to consider.

A common method used to modulate oxygen conditions in cell culture involves the use of a CO₂ incubator with variable oxygen control (Wu and Yotnda, 2011). The major drawback of this method when using traditional cell culture techniques is that long exposure periods are required for the oxygen present in the gas phase to diffuse to cells attached at the bottom of the culture dish. Depending on the volume and depth of media, this can take hours due to poor oxygen diffusivity in aqueous media (Allen et al., 2001). Oxygen conditions at the pericellular level therefore do not immediately match the oxygen conditions supplied in the incubator air. One method to counteract this is to pre-equilibrate media/reagents to the desired pO₂ levels prior to culturing. The major issue with this is that unless the cell culture is handled within a closed compartment (such as in an airtight glove box), oxygen from the ambient air is constantly diffusing into media and reagents, preventing the precise control over the oxygen concentrations being delivered to cells (Byrne et al., 2014). Many laboratories are not readily equipped with oxygen-controlled incubators or glove boxes due to their high costs.

As a more cost-effective alternative, we have designed a cell culture system that utilizes commercially available oxygen-permeable cultureware. The culture dishes are composed of a thin membrane surface that is highly permeable to oxygen. Cells adhere to the membrane and are covered in media as with normal cell culture methods. However, due to the extremely high oxygen permeability of the membranes, cells have access to oxygen almost instantly from ambient air through the membrane rather than through the media (Figure 1). This enables faster oxygen equilibration rates to be achieved at the cellular level. These fast rates of oxygen cycling therefore make this model particularly useful for protocols aimed at studying rapid changes in oxygenation or intermittent hypoxia, relevant to diseases such as cancer (Michiels et al., 2016), obstructive sleep apnea (Minoves et al., 2017; Martinez et al., 2019), cardiovascular disease and ischemia (Savla et al., 2018).