Cell Stem Cell Letter - Applied BioPhysics

1 Cell stem CellLetterPO2 Matters in stem Cell CultureDidier Wion,1,*Thomas Christen,1 Emmanuel L. Barbier,1and Jonathan A. Coles21 INSERM U836 and Universite Joseph Fourier, Grenoble Institute of Neuroscience, 38043 Grenoble, France2 Centre for Biophotonics, Strathclyde Institute for Pharmacy and Biomedical Sciences, University of Strathclyde, 27 Taylor Street,Glasgow G4 0NR, a century ago, conditions wereworked out for maintaining growing tissueand cells outside the body. From thebeginning, care was taken to maintaincultures at a physiological temperature,and to include precise concentrations ofsalts and other compounds, but the ox-ygen concentration in the culture mediumwas simply the result of letting the me-dium equilibrate with air. This approachwas a reasonable first approximation,given that values of partial pressure ofoxygen (PO2) in animal tissues were notmeasured until over a decade later, andall that mattered seemed to be to providecells with enough oxygen.

2 However, it isnow firmly established that, in vivo, moststem cells are in environments with low,or very low, PO2(for review, seeSimonand Keith 2008; Lin et al., 2008; bibliog-raphy in theSupplemental Dataavailableonline). Additionally, in vitro, it has beenshown that oxygen concentration issensed by stem cells and low PO2canradically modify their phenotypes. Thesefacts suggest that we should think morecarefully about PO2in cell ambient air has a PO2of 150mm Hg corresponding to a concentrationof 21%, PO2in arterial blood is 95 mm Hg,and in normal tissues it is generally con-siderably lower, with the amount between50 and 5 mm Hg (7% ). The PO2sexperienced in vivo by blastocysts in theuterine fluid, hematopoietic stem cells ,and cancer cells in poorly vascularizedtumors are at the lower end of this range(Supplemental Data, sections S3, S4,and S7). In vitro, low PO2is known toaffect stem cell phenotype in an in-creasing number of ways (reviewed inCsete, 2005).

3 For example, low PO2canreduce spontaneous differentiation andenhance clonality of human embryonicstem cells (hESCs). PO2can also influ-ence the subsequent fate of stem of neural stem cells in lowoxygen increases the percentage of do-paminergic neurons obtained after differ-entiation, an effect partly mediated by anautocrine loop involving the expressionand secretion of erythropoietin. In hypoxicregions of tumors, low PO2promotesdedifferentiation and could be a determi-nant of the cancer stem cell phenotype(seeKeith and Simon, 2007). Now, in thisissue ofCell stem Cell, Yoshida et that lowering cell culturePO2enhances the generation of inducedpluripotent stem cells (Yoshida et al.,2009). Some of the mechanisms by whichlow PO2influences stem cell behaviorhave been elucidated. Activation of theNotch signaling pathway by the hypoxia-inducible transcription factor HIF-1amain-tains the undifferentiated state, whereasexpression of Oct-4, which controls stemcell renewal and pluripotency, is inducedby HIF-2a.

4 It is likely that many other path-ways are affected by PO2given that whenhESCs are cultured in 4% O2, rather than20%, 149 genes are either upregulated ordownregulated (Westfall et al., 2008; seealsoForsyth et al., 2008). In sum, PO2isimportant for stem cell , accurately controllingpericellular PO2is technically lowering the PO2in a standardlaboratory incubator goes only a shortway toward achieving a defined, low PO2at the cells themselves. The first problemis the time required for equilibrating acell-culture medium previously exposedto atmospheric PO2with a gas phase ata lower PO2. Routine laboratory practiceis to change the medium and to passagethe cells in a laminar flow hood underatmospheric PO2. The gas trapped insidethe culture vessel is initially at the atmo-spheric PO2, and equilibration with thelow PO2gas phase can take several hours(Westfall et al., 2008). Most of this pro-blem can be overcome by equilibratingthe culture medium at the desired PO2before adding it to the cells (and, ideally,by using a closed hypoxia workstation).

5 If these precautions are not taken, notonly is there an oxygen shock associ-ated with changing the medium but alsothe time during which the cells were actu-ally exposed to the low oxygen tension isshorter than the time they were in theincubator. The future value of currentwork might be increased if authors wereto give precise details of the proceduresthey may also be an error in the otherdirection: when the oxygen tension in thebulk medium in the culture vessel finallyfalls to the value in the gas phase, the peri-cellular PO2will be even lower becauseoxygen consumption by the cells them-selves reduces PO2(seeSupplementalData, section S9 for more complete refer-ences). Hence, when a PO2value is indi-cated in a publication, it is important tomention if this value corresponds to theincubator gas phase, to the bulk medium,or to the pericellular space. If the cells aregrown as a uniform monolayer, then theywill all experience roughly the same PO2(except at the edges of the vessel), somonolayer cultures are to be preferredwhen the aim is to subject all the cells tothe same conditions.

6 If the aim is toapproximate in vivo conditions, spheroidsare a useful model for embryonic orcancer stem cells . However, in a spheroid,oxygen consumption by peripheral cellsreduces diffusion of oxygen into thecenter of the sphere and it is virtuallyimpossible to impose a uniform PO2onall the cells . Little is known about theconsequences of contiguous stem cellsbeing in a PO2gradient, but new technol-ogies offer hope that such questionsmight be investigated. For example, asystem developed by Michel Maharbizand colleagues is designed to allow thecreation of gradients of PO2in cell mono-layers (Park et al., 2006). Dissolvedoxygen is locally generated at a controlledrate by electrolysis, and PO2can beimaged optically at the base of the culturechamber. Almost any desired 2D PO2mi-crogradient can be produced with stem Cell5, September 4, 2009 2009 Elsevier aim of much stem cell research is toproduce stem cells for therapies, and ex-isting evidence suggests that cells cul-tured in 20% oxygen will change theirproperties when injected into the lowPO2environment in the patient.

7 Equally,if the aim is to understand the behaviorin vivo of stem cells , then experimentscarried out at low, physiological PO2sare likely to be necessary. In any case, itis desirable that reporting of how oxygenis provided to stem cells in culture shouldbe more DATAS upplemental Data include Supplemental Classi-fied Bibliography and can be found with thisarticle online (09) , M. (2005). Ann. N Y Acad. , 1 , , Kay, A., Hampson, K., Downing, A.,Talbot, R., and McWhir, J. (2008). Regen ,817 , B., and Simon, (2007). Cell129, 465 , Q., Kim, Y., Alarcon, , and Yun, Z. (2008).Gene Regul. Syst. , 43 , J., Bansal, T., Pinelis, M., and Maharbiz, (2006). Lab Chip6, 611 , , and Keith, B. (2008). Nat. Rev. , 285 , , Sachdev, S., Das, P., Hearne, ,Hannink, M., Roberts, , and Ezashi, T.(2008). stem cells , 869 , Y., Takahashi, K., Okita, K., Ichisaka, T.,and Yamanaka, S. (2009). Cell stem Cell5, thisissue, 237 stem CellLetterCell stem Cell5, September 4, 2009 2009 Elsevier