Cytocentric Visionaries: Romain Vuillefroy de Silly
Part 1: Is Physiologic Oxygen Better for CAR-T Cells In Vitro?
Dr. Romain Vuillefroy de Silly is a post-doc at the University Ludwig center for Cancer Research of the UNIL in Lausanne, Switzerland. He has published original research and several compelling reviews on the role of oxygen in T cell function. Here is a written interview Dr. Alicia Henn, Chief Scientific Officer of BioSpherix, conducted with him about his most recent review in Oncoimmunology.  This interview was edited for length.
Alicia Henn: We see you as a Cytocentric Visionary because you are working to truly understand the T cell environment from the cells’ point of view. How did you first get interested in the effects of oxygen on T cells?
Romain Vuillefroy de Silly: During my PhD, I developed a profound interest in fundamental immunology, and in how metabolism and nutrients can impact the immune response. T cells are unique in the way that they navigate between virtually all sites of the body, facing a wide spectrum of chemical and physical properties (including oxygenation) where they still have to exert their function.
Therefore, when I had the opportunity to study how one of the most basic, yet crucial, components of species evolution (i.e. oxygen) influences T cells during my post-doc, I did not hesitate! Given the contradictions found in literature, I wanted to have a clear idea about the answer.
AH: In the literature you can see that in the 1970s and 1980s there was a lot of interest in controlling oxygen for cells. Interest faded after that. Now the interest is coming back again. Do you have an opinion as to why scientists decided that oxygen wasn’t important for awhile?
RV: I believe immunology (and certainly science in general) can be compared with fashion. Indeed, there are periods where a dogma is extensively studied, or included in unrelated studies. After this period, interest is lost and it is considered old-fashioned.
For oxygen, I also think that the available tools limited researchers. The interest came back again now that more tools are available (e.g. pimonidazole/hypoxyprobe, hypoxia workstation, oxygen monitoring). Also, the discovery of HIF transcription factors together with their potent functions has added interest. Finally, we could certainly include oxygen in the so-called field of immunometabolism, which has gained a considerable interest in the past ten years.
AH: In your recent review, you point out that low oxygen levels enhance the priming of T cells into cytolytic CD8+ T cells, even though proliferation was reduced. Do you think that lower oxygen might affect activation/transfection efficiencies included in many CAR-T protocols?
RV: Most of the CAR T cell protocols start with an activation step before infection or transduction because it promotes expression of receptors needed for retrovirus and lentivirus entry. Moreover, integration of the retroviral genome needs DNA replication, which is obtained when T cells are activated and proliferate.
Since modulation of oxygen levels can alter the phenotype of T cells, it is conceivable that it could affect infection efficiencies. However, it might also have no impact since most infection protocols are carried out 24h post T-cell activation, which is a relatively early time-point. Whether oxygen can modulate expression of molecules involved in virus entry into the cell still has to be determined.
AH: You point out that control of oxygen levels in vitro for adoptive transfer may enhance effector capacity. CAR-T therapy efficacy has been correlated with the in vivo expansion of the cells after administration. Do you think that the CAR-T cells might need the variable physiologic oxygen levels available in vivo for optimal expansion and activity?
RV: This is an interesting question. Indeed, we could imagine that T cells cultured under lower oxygen fractions in vitro could lead to better antitumor response, given that they are more cytolytic. Further, it may be possible that they express adhesion molecules, integrins and chemokine receptors, allowing a better entry and retention in the tumor.
However, it is becoming clear that an enhanced effector capacity is associated with a decreased self-renewal ability, and vice-versa. Actually, the more differentiated/effector the cells are, the closer they should be to an exhausted phenotype. Therefore, we may conclude that lowering oxygen fraction in vitro might not be desirable if we want T cells with a less differentiated/memory phenotype, which are described to possess long-lasting persistence capacity in vivo.
Apart from this paradigm, it is also possible that culturing cells in vitro in conditions that are closer to the ones they will encounter in vivo could help them to metabolically adapt to the “battlefield” (to lower oxygen, glucose,…), leading to better survival and function in vivo.
AH: You point out in the review that in situ hypoxia may inhibit tumor cell cytolysis by its effect on the tumor cells rather than the T cells. This is a very important point. How often do we blame the T cells in the assay when the tumor cells might be modified by the conditions?
RV: Indeed, it might occur frequently. I do not think this is really important when one is trying to depict what is happening in vivo, since the final result integrates both the effect on T cells and the effect on tumor cells. However, when trying to determine whether a condition affects directly the T cell, I think this is of crucial importance.
This is why in our previous research paper, showing the impact of hypoxia on T cells, we decided to do most of our work using T cells activated with anti-CD3 and anti-CD28 antibodies, allowing us to be sure that any effect observed was due to a direct impact on T cells themselves. However, CTL-induced tumor cell cytolysis cannot be assessed without the presence of both T cell and tumor cells; thus yielding an integrated view that gathers the impact on T cells and on tumor cells. Consequently, one needs to analyze in detail which cell type is affected.
AH: How does T cell expansion, with increasing cell densities, affect pericellular oxygen levels in culture?
RV: We never tried to measure pericellular levels in culture. However, based on previous studies, it is more than likely that the more the cell density increases, the more it affects pericellular oxygen levels in culture. Actually, confirming this idea, we observed that the more a glioma cell line was seeded at high densities, the more HIF-1 alpha was stabilized under room air oxygen.
This is why I tried to keep the cells at a relatively low density in order to keep oxygenation "relatively" stable during culture. However, T cells are a relatively unique subtype, since they produce clusters when they get activated, potentially creating their own low oxygen environment.
AH: Do you think that T cells themselves might have a role in modulating local oxygen levels in vivo?
RV: Indeed, I think that T cell accumulation might lead to consumption of oxygen in situ. If so, and if they participate mostly to the modulation of in situ oxygen levels, the question that would arise would be: is it an undesired consequence or an effect done on purpose?
AH: You say that the tumor microenvironment is hostile to T cells. How hostile is the room air research environment to T cells?
RV: The answer probably depends in which context you are. In a fundamental immunology context where you want to explain some in vivo results by dissecting it in vitro, it might probably be hostile. Indeed, extrapolation from in vitro data might be biased since T cells would not behave exactly in the same way in vivo since oxygen levels are much lower (even if in vitro conditions differ in many ways to in vivo conditions independently of oxygen levels).
For example, it is known that, sometimes, when trying to analyze T cell immunosuppression induced by a specific cell type or molecule, it is better to give a suboptimal TCR activation stimulus in order to identify an effect. When T cells are maximally activated, you might not detect an immunosuppressive effect.
Coming back to oxygen, since culture under physiological oxygen levels in vitro leads to a diminished proliferation of T cells, it might be possible that some molecules or immunosuppressive cells have an observable effect using physiological normoxia.
In a translational context, when one wants to know the cytotoxic potential of CTLs, results might be misleading under room air since their potential is diminished as compared to CTLs pre-cultured under physiological normoxia. Actually, one might underestimate their in vivo cytotoxic potential. However, room air might also be beneficial since it allows optimal expansion of T cells for adoptive cell transfer in a cancer context.
Our conversation with Dr. Vuillefroy de Silly is continued in Part Two: What’s in a Word? If It Describes the Oxygen Level, Quite a Bit.
1. Vuillefroy de Silly, R., P.Y. Dietrich, and P.R. Walker, Hypoxia and antitumor CD8+ T cells: An incompatible alliance? Oncoimmunology, 2016. 5(12): p. e1232236.
About the Author
Alicia D Henn, PhD, MBA
Alicia Henn has been the Chief Scientific Officer of BioSpherix, Ltd for two years. Previously, she was a researcher at the Center for Biodefense Immune Modeling in Rochester, NY. Alicia obtained her PhD in molecular pharmacology and cancer therapeutics from Roswell Park Cancer Institute in Buffalo, NY and her MBA from the Simon School at University of Rochester in Rochester, NY.