High-performance oxygen carriers for cultivated meat

Current challenge

While animal cells are capable of metabolizing nutrients either aerobically or using lactic acid fermentation (used when oxygen levels are limited), the aerobic pathway is vastly more efficient, theoretically producing 19x the energy per molecule of glucose and avoiding excess lactic acid production. Efficient use of media will be essential in achieving cost-effective and environmentally friendly cultivated meat production. Therefore, cells must be supplied with adequate oxygen so they can take advantage of the efficiencies of aerobic respiration. However, approximately 45 times the amount of dissolved oxygen can be carried by blood versus cell culture media, making this a challenge for culturing cells outside the body (Martin and Vermette 2005).

Cellular metabolism produces CO₂ as a waste product, and the rates of both O₂ delivery and CO₂ removal represent important constraints on the achievable cell density (Humbird 2021, Figure 1). Because oxygen is highly reactive, cells can also suffer negative consequences from oxidative stress if dissolved oxygen concentrations are inappropriately high.

Insufficient oxygenation can also be a challenge in hollow fiber bioreactors (HFBs) (Chen and Palmer 2009; Chen and Palmer 2010). Variation in dissolved oxygen concentration along the length of an HFB can lead to cells at the inlet suffering from oxidative stress associated with hyperoxic conditions, while those near the outlet of the same bioreactor suffer from severe hypoxia (Chen and Palmer 2009).

A secondary consequence of cells’ need for O₂ delivery and CO₂ removal is that the methods used to oxygenate and mix the culture media can harm the cells. Both mixing and gas sparging can lead to cell death, necessitating a compromise between the need to mix and aerate cultures and the need to limit these forms of hydrodynamic stress (Singh and Al-Rubeai 1998). The mechanism of cell death due to high sparge rates is thought to relate to shear stress on the cells when gas bubbles burst at the surface of the culture (Cherry and Hulle 1992). A modeling study of a stirred tank-based bioprocess designed with cultivated meat in mind showed a strong increase in cell death at high rotor speeds (Camphuijsen et al. 2022).

The need for efficient and controllable gas exchange is, therefore, a substantial challenge for efficient cultivated meat production. Cultivated meat producers need to maximize the proliferation rate, media use efficiency, viability, and density of their cells. This can be hampered by too much or too little oxygen, too much CO₂, or the shear stress associated with mixing and sparging.

Proposed solution

Because whole organisms also face the need for reliable gas transport to and from metabolically active tissues, this challenge has already been elegantly addressed by evolution. Hemoglobin is an oxygen carrier found in almost all vertebrates. It has properties that likely make it well-suited for use in a bioprocessing context. Multiple other oxygen-binding proteins are found in various animal, plant, and microbial species. The use of oxygen-carrying proteins is already being explored both in cell culture and as a therapeutic option for human patients. To our knowledge, only one study (Simsa et al. 2019) has explored the use of oxygen-binding proteins for cultivated meat. Their use has not been investigated experimentally or computationally in large-scale cultures of myogenic or adipogenic cells, where oxygen transport is more likely to be limiting in comparison to lab-scale cultures. Though we focus on protein-based oxygen carriers here, various other oxygen-releasing biomaterials (Camci-Unal et al. 2013) have been tested for use in tissue engineering contexts, and could also theoretically be applied to cultivated meat.

Oxygen-binding proteins are well-adapted to the job of transporting oxygen from a high-oxygen environment to a low-oxygen one. In vivo in most vertebrates, hemoglobin transports O₂ from the lungs or gills to the tissues and carries CO₂ in the opposite direction. The vast majority (98%) of the oxygen carried by blood is carried by hemoglobin, while the remainder is simply dissolved in the blood (Ferenz 2020). The same system could handle gas transport between a medium recycling unit and a dense cellular tissue within a bioreactor (Chen and Palmer 2010; Montagne et al. 2011, Figure 2) or could be used in a stirred-tank reactor under conditions where O₂ and CO₂ transport is limiting. The use of such oxygen carriers could help meet the metabolic needs of dense, rapidly-proliferating cells while reducing the need for high rates of mixing and gas sparging that could decrease viability. However, which specific carrier is used will influence the effectiveness of this solution. Evolution has produced a vast array of hemoglobins and other oxygen-binding proteins well-adapted to the physiology and needs of specific organisms. The oxygen-binding properties vary both between oxygen carriers and depending on the conditions (e.g., pH or temperature) a given protein is subjected to.

As a first step in implementing this solution, it will be helpful to computationally model (Cantarero Rivera and Chen 2022; Li et al. 2019; Zhang et al. 2021) the effects of oxygen carriers on the performance of various hypothetical cultivated meat bioprocesses. First, such models can help test the hypothesis that oxygen carriers could substantially improve the productivity of these systems under conditions where growth is limited by oxygen transfer rates or factors such as shear stress that result from existing oxygen delivery systems. Second, these models will make it possible to estimate the properties of an “ideal” oxygen carrier for a given system. The predictions resulting from such models must then be validated experimentally in relevant contexts.

Desired characteristics

Hemoglobin proteins from different species vary substantially in their oxygen-binding properties. While the ideal characteristics of a hemoglobin-based oxygen carrier for cultivated meat will vary depending on the details of the bioprocess, several variables should be considered in making that calculation. These include the oxygen affinity (p50), cooperativity, Bohr effect, temperature effect, and degradation rate (More details on this below). 

Mild hypoxia as a cue

It is worth noting that the typical partial pressure of oxygen in skeletal muscle tissue (15-76 mmHg) is substantially lower than that in the atmosphere (160 mmHg at sea level), and that the level of oxygenation can serve as an important physiological signal (Pircher et al. 2021). Studies have generally found positive effects of mild hypoxia on proliferation and satellite cell renewal, though this effect is not entirely consistent (Pircher et al. 2021; Beaudry et al. 2016), perhaps due to differences in experimental conditions across studies. On the other hand, studies have generally found negative effects of hypoxia on myogenic differentiation (Pircher et al. 2021; Beaudry et al. 2016). However, under certain conditions, hypoxia (1% or 10% O₂) has been demonstrated to promote differentiation (Kook et al. 2008; Sakushima et al. 2020). For a given cell type, it will be crucial to optimize the culture conditions (including extracellular O₂ concentration) to achieve the desired result, whether that be proliferation or differentiation. Because mild-to-moderate hypoxia may be a useful cue under certain conditions (Elashry et al. 2022), the goal of an oxygen delivery system should not be simply to maximize oxygen delivery but rather to tune oxygen concentrations to the current needs of the cells. Selecting oxygen carriers with appropriate oxygen-binding and release properties—and the use of such carriers at appropriate concentrations—may be a useful tool in achieving this. Using different carriers during the proliferation and differentiation phases or modulating their properties by pH or temperature (if compatible with the needs of the cells) may be helpful.

Diversity of naturally-occurring oxygen carriers

Evolution has produced a vast array of oxygen carrier proteins, each one with properties such as the strength of its cooperative binding and Bohr effect adapted to the needs of the corresponding organism (Rapp and Yifrach 2019). This presents an opportunity for bioprocess engineers to select hemoglobins as oxygen carriers with oxygen-binding properties well-suited to a given bioprocess. Some of those available options are discussed here, though other types of oxygen-binding proteins may also be worthy of investigation. High-throughput screening approaches, such as those leveraged by Ginkgo Bioworks, may be helpful in identifying the most promising candidates for a given bioprocess.

Mammalian hemoglobin & myoglobin

The hemoglobin proteins found in the red blood cells of mammals and other vertebrates contain four protein subunits, each of which is associated with one heme prosthetic group. Mammalian hemoglobins exhibit both cooperative binding and the Bohr effect, as described above. Generally speaking, hemoglobins from smaller animals tend to show a stronger Bohr effect, possibly due to their faster metabolic rates (Riggs 1960). An exception is large marine mammals such as whales, which have a strong Bohr effect similar to that of a guinea pig, which is hypothesized to act as an evolutionary adaptation favoring tissue oxygenation during long dives. In addition, adult and fetal forms of human hemoglobin show differences in their oxidative reactivity, suggesting likely advantages and disadvantages as oxygen carriers (Simons et al. 2018). A related protein, myoglobin, is expressed by muscle fibers and serves as an oxygen storage protein within muscles. Unlike hemoglobin, myoglobin is monomeric and thus does not exhibit cooperative binding. It also has a substantially higher oxygen affinity.

Mammalian hemoglobins have shown mixed results in cell culture contexts. Bovine hemoglobin was tested as an oxygen carrier for human liver cells in an HFB, and it was observed that cells were more metabolically efficient and proliferated more during the culture period when in the presence of hemoglobin (Chen and Palmer 2010). Additional experiments with bovine and porcine hemoglobin have shown inconsistent results (Stančić et al. 2020; Stančić et al. 2021), though it was suggested that residual lipids from the lysed red blood cells might have contributed to some of the negative effects (Stančić et al. 2020). Adding equine myoglobin to bovine satellite cells cultured in 2D had a generally positive effect on cell proliferation, while the effects of bovine hemoglobin were neutral or negative (Simsa et al. 2019). In the same study, the metabolic activity of small-scale (1-3 mm) 3D constructs was increased by the presence of myoglobin but not hemoglobin. Another reason for these mixed results might be the potential for free hemoglobin to cause oxidative stress (Schaer & Buehler 2013).

Research into the use of mammalian hemoglobins as blood substitutes has included the investigation of strategies for fine-tuning the properties of these proteins. For example,  variants of human hemoglobin have been designed for lower oxygen affinity and greater stability (Looker et al. 1992) or lower susceptibility to oxidation (Cooper et al. 2019). Hemoglobin differs substantially in its oxygen-binding properties depending on whether it is contained within erythrocytes or not, largely due to the effects of intracellular 2,3-bisphosphoglycerate (Ferenz 2020). Therefore, it may be worthwhile to encapsulate hemoglobin within a semipermeable membrane and to test the effects of additives such as 2,3-bisphosphoglycerate and carbonic anhydrase on the overall performance of the system. Encapsulation strategies have been demonstrated for both hemoglobin (Montagne et al. 2011; Centis et al. 2011) and myoglobin (Kishimura et al. 2007), and could be applied to other oxygen carriers. 

Annelid giant hemoglobins (erythrocruorin and chlorocruorin)

Many annelid species contain an extracellular version of hemoglobin that may have beneficial properties as an oxygen carrier for cultivated meat. Annelid giant hemoglobins, primarily those from Arenicola marina, have been commercialized by the company Hemarina for a diverse set of applications including food fermentation (HEMBoost) and mammalian mesenchymal stem cell (MSC) culture (HEMOXCell, derived from Alitta virens (Le Pape et al. 2015)). Like mammalian hemoglobins, annelid giant hemoglobins exhibit Bohr and temperature effects (Rousselot et al. 2006).

Whereas the fully-assembled giant hemoglobin has a molecular weight of 3682 kDa, its modular structure means that smaller subunits are also functional. For example, under human plasma-like conditions, Arenicola marina giant hemoglobin degraded rapidly into 205 kDa dodecamers that were both functional and stable (Rousselot et al. 2006). The fact that smaller subunits are functional bodes well for both the feasibility of recombinant production (Zal et al. 2006) and the carrier’s stability in long-term culture, though the presence of multiple forms of the protein may make modeling efforts more complicated.

CHO cells cultured in 125 mL shake flasks showed up to a 4.6-fold increase in cell density when supplemented with HEMOXCell as compared to non-supplemented controls (Le Pape et al. 2015). Measurements of oxygen concentrations in 96-well plates containing culture medium, MSCs, and/or HEMOXCell revealed that the presence of MSCs reduced dissolved oxygen concentrations and that this was largely reversed by the presence of the oxygen carrier (Le Pape et al. 2017). MSCs cultured in 6-well plates showed 25% higher growth rates than controls while maintaining their differentiation potential (Le Pape et al. 2017).

Plant phytoglobins

Plants also express oxygen-binding proteins with homology to animal hemoglobin and myoglobin, which are sometimes referred to as phytoglobins. The best known of these is leghemoglobin, which is expressed in the root nodules of plants colonized by symbiotic nitrogen-fixing bacteria, but many other plants also express non-symbiotic hemoglobin genes. However, most phytoglobins exhibit a much higher oxygen affinity than vertebrate hemoglobins. This might make them inefficient oxygen carriers for many cultivated meat bioprocesses, though this assumption may need to be validated by modeling or experimental approaches. However, some non-symbiotic phytoglobins in sugar beets have a somewhat lower oxygen affinity comparable to vertebrate myoglobin (Leiva Eriksson et al. 2019). Impossible Foods uses precision fermentation to produce soy leghemoglobin as an ingredient, further demonstrating the potential for recombinant phytoglobins to be produced at scale.

Expression systems

Precision fermentation of recombinant oxygen carriers in microbial hosts (Zhao et al. 2020) may be a practical strategy for producing these proteins at the scales needed for use as a cultivated meat cell culture supplement. Human hemoglobin has been successfully produced in E. coli (Looker et al. 1992; Simons et al. 2018; Cooper et al. 2019) and yeast (Ishchuk et al. 2021). The stability of recombinant hemoglobin can be further increased by deletion or overexpression of specific genes in the host organism (Ishchuk et al. 2021). Bovine myoglobin (Motif Foodworks and Luyef), soy leghemoglobin (Impossible Foods), and heme proteins from a variety of animal species (Paleo) have been produced in yeast for food applications. While the requirements for heme-containing proteins intended for use in cultivated meat may be somewhat different (e.g., purity, sterility, post-translational modifications) from those for use directly in food, the existence of these products is a positive sign when it comes to the scalability of recombinant heme protein production. Sugar beet non-symbiotic phytoglobin has been produced in E. coli (Hugner 2020; Leiva-Eriksson et al. 2019), and phytoglobins from several plant species have been produced in C. glutamicum (Wang et al. 2023).

Plant molecular farming is another alternative, as human hemoglobin production has already been demonstrated in Nicotiana benthamiana (Kanagarajan et al. 2021). Motif Foodworks has announced that they are exploring this possibility for their bovine myoglobin ingredient via a partnership with the plant molecular farming startup IngredientWerks. Moolec is using plant molecular farming to produce an unspecified—presumably heme-containing—pig protein in soybeans for use as a plant-based meat ingredient.

Unmodified plants or microorganisms (Buisson and Labbe-Bois 1998; Frey and Kallio 2003) could also serve as sources of oxygen carrier proteins for cultivated meat production, provided that the properties of their naturally-occurring oxygen carriers are compatible with the intended bioprocesses and that their expression levels are sufficiently high. Alternative protein startup tHEMEat produces a heme product known as Veme, which is derived from an unmodified vegetable source. Similarly, alternative protein startups HN Novatech and Yemoja have developed “heme-like” compounds from seaweed and microalgae, respectively. It is unclear whether these compounds are in fact true heme proteins with oxygen-carrying capacity and whether their oxygen-carrying properties—if any—are compatible with cultivated meat bioprocesses. If so, they could represent a promising natural source of heme. Plants such as sugar beets, whose phytoglobins show both high expression (Leiva-Eriksson et al. 2014) and a relatively low oxygen affinity (Leiva Eriksson et al. 2019), might be worth investigating for this purpose. The utility of plant or microbial hemoglobins in mammalian cultures has not been extensively tested, though Vitreoscilla hemoglobin was shown to increase recombinant protein production by CHO cells (Pendse and Bailey 1994).

Finally, cultured erythrocytes (Bernecker et al. 2019) could, at least in theory, serve as either oxygen carriers themselves or a source of extracellular hemoglobin. However, the benefits of such a system would need to be carefully weighed against the added cost and complexity associated with culturing an additional vertebrate cell type.

Learn more about the key variables distinguishing oxygen carrier proteins

p50: The affinity of hemoglobin for oxygen can be expressed according to its p50 value, defined as the partial pressure of oxygen (pO2) at which 50% of the oxygen-binding sites are occupied. A higher p50 thus corresponds to a lower affinity of hemoglobin for oxygen. The selection of a hemoglobin with an appropriate p50 for the intended application will ensure that oxygen is picked up efficiently under conditions of higher oxygen partial pressure and released under conditions of lower oxygen partial pressure. The rate of oxygen release should be rapid enough to support the cells’ metabolic needs, but slow enough that oxygen is delivered throughout the tissue or the bioreactor while minimizing oxidative stress. If the p50 value is too high (i.e., the hemoglobin does not bind oxygen strongly enough), the hemoglobin will not become saturated with oxygen when exposed to an oxygen source, or it will become fully desaturated too rapidly, leading to uneven oxygenation. If the p50 value is too low (i.e., the hemoglobin binds very strongly to oxygen), the hemoglobin’s  oxygen-binding sites may be overoccupied, meaning that the delivery of oxygen to the issue where it is needed will be less efficient. One useful illustration of this effect is the work of Chen and Palmer (2009), in which pO2 profiles were calculated along the length of a modeled HFB supplemented with four different forms of hemoglobin whose p50 values ranged from 5-38 mmHg. Under the modeled conditions, the use of hemoglobins with higher p50 values resulted in good oxygenation across the entire length of the HFB, while HFBs supplemented with low-p50 hemoglobins suffered from hypoxia nearer the outlet end.

Cooperativity: The cooperativity of an oxygen carrier describes the extent to which the binding of one or more oxygen molecules increases the affinity for additional oxygen molecules. One feature of the hemoglobin molecule that makes it especially well-adapted as a physiological oxygen transporter is its cooperative binding to oxygen. Vertebrate hemoglobin is a tetramer that binds up to four molecules of oxygen. The binding of each molecule of oxygen changes the conformation of the tetramer in a manner that favors the binding of additional oxygen molecules. This results in a sigmoidal relationship between pO2 and oxygen saturation (Figure 3), which ultimately makes hemoglobin more effective at both picking up oxygen in the lungs or gills and releasing it in the lower-oxygen environment of the muscles and other tissues. The degree of cooperativity can be described by the Hill coefficient (n). A Hill coefficient of 1 describes non-cooperative binding, negative values indicate negative cooperativity, and positive values indicate positive cooperativity. Cooperative binding is likely to be similarly helpful in a bioprocessing context, with the oxygen carrier becoming saturated in high-oxygen contexts and releasing all or most of that oxygen in areas of high cell density which are typically lower in oxygen.

Bohr effect & temperature effect: The Bohr effect describes the effect of CO₂ concentration and pH on the oxygen affinity of hemoglobin (Figure 4). In addition to cooperative binding, theBohr effect further enhances the ability of hemoglobin to pick up oxygen in the lungs or gills and release it in the tissues. Carbon dioxide decreases the pH (i.e., increases the concentration of free protons) in metabolically active tissues like muscle. The binding of CO₂ and protons to hemoglobin favors the dissociation of oxygen. Therefore, oxygen delivery is increased in metabolically active tissues, which tend to be high in CO₂ and low in pH. This effect is reversed in the lungs or gills as CO₂ and protons are released, meaning that the hemoglobin protein can pick up a full set of four oxygen molecules to carry back to other tissues. A related effect in fish hemoglobin is known as the Root effect.

Similarly, the oxygen dissociation curve can be shifted by changes in temperature, with higher temperatures leading to lower affinities and easier oxygen dissociation. It might be possible to capitalize on this effect in a bioprocessing context by lowering the temperature at the site where the oxygen is loaded onto the carrier.

Degradation rate: As with any protein, oxygen carrier proteins added to the culture media will degrade at a certain rate. To reduce the costs associated with protein production and the need for continual additions of ingredients during the culture period, it would be advantageous to select carriers with degradation rates sufficiently slow that they do not need to be replenished frequently during the culture period.

Anticipated impact

Designing a cultivated meat bioprocess is undeniably a daunting challenge, in part due to the challenges associated with gas exchange and the need for mixing of stirred tank bioreactors (Humbird 2021). By mitigating these challenges, oxygen carrier proteins may increase the achievable efficiencies and yields, bringing the industry closer to the point where price parity with conventional meat is achievable. As a secondary benefit, these same ingredients might have positive impacts on the organoleptic properties or the iron content and bioavailability of the end product.

Because the achievable cell densities and doubling times are simultaneously constrained by multiple factors, the effectiveness of any one solution may depend on the presence or absence of others. For example, oxygen carriers might have little effect on cell performance if ammonia concentrations are the primary limiting factor. However, in the presence of an effective media recycling system, oxygen might become the limiting factor, making the need for carriers much more salient. 

In a sense, the proposed solution is quite simple and tractable. However, two sources of uncertainty may make successful implementation somewhat more difficult. First, it will be necessary to take into account a number of variables, from the properties of the carrier itself to the conditions under which it is used (concentration, temperature, encapsulation, etc.). Selection of an appropriate carrier and appropriate conditions for its use will be essential. Second, because the problem to be solved by the use of oxygen carriers is associated primarily with larger-scale cultures, early-stage tests of this strategy may not reflect the expected behavior at larger scales.

It will be necessary to carefully consider the system in which small-scale tests are conducted, as many lab-scale culture systems may be primarily limited by factors other than gas transport. For example, Montagne et al. (2011) found dramatic effects on the morphology and bioactivity of liver cells cultured under perfusion conditions in a flat plate bioreactor when liposome-encapsulated hemoglobin was added to the culture media, but no effect of the same carriers on the same cells in static culture in 12-well plates. It will be essential when designing these experiments to understand 1) the oxygen-related needs of the cells in question, 2) their oxygen status under control conditions, and 3) the predicted and actual effects of a given oxygen carrier on oxygen availability in different parts of the bioreactor. Modeling approaches, together with direct measurements of dissolved oxygen concentrations to verify the effects of the carriers, will be extremely helpful in ensuring that experiments are truly testing what they are designed to test and increasing the chances that the data collected at lab-scale will be able to inform scale-up efforts.

Related efforts

• Hemarina sells several oxygen carriers derived from marine worms. The most applicable to testing for cultivated meat applications are likely to be HEMBoost and HEMOXCell.
• Non-animal sources of heme-containing proteins have been created by Motif Foodworks, Impossible Foods, Paleo, Moolec, tHEMEat, HN Novatech, and Yemoja.

GFI resources

Cultivated meat LCA/TEA report analysis

Recent studies show cultivated meat could have reduced environmental impacts and be cost-competitive with some forms of conventional meat.

The science of cultivated meat

Learn about the science of cultivated meat and the challenges that must be addressed for commercial production.