IMMUNOLOGY

Cause of death matters Brigitta Stockinger

The process of programmed cell death can either induce anti-inflammatory immune responses or actively promote inflammation. Whether the dying cell is infected seems to govern which response is triggered.

To deal with invading pathogens efficiently, helper T cells of the adaptive immune system differentiate to form distinct subsets of cells with specific immune functions. Much work has been done on two such subsets, T helper type 1 and type 2 cells (TH1 and TH2 cells). But for the past few years, a subset of T cells called TH17 cells has been in the limelight because of their potential role in autoimmunity, with research being concentrated on elucidating the molecu-lar mechanisms that underlie the cells’ differen-tiation, regulation and function. For example, it has become clear that, in vitro, combined action of two immune-mediator molecules, the TGF-β and IL-6 cytokines, is required for the develop-ment of human and mouse TH17 cells (which produce the IL-17 cytokine). But what events trigger the expression of these two cytokines in vivo? On page 78 of this issue, Torchinsky et al.1 provide an answer to this question.

Whether TGF-β acts as an anti-inflamma-tory mediator or stimulates inflammation by inducing the differentiation of TH17 cells seems to be very much context-dependent. Original observations2 indicated that, in vitro, regula-tory T (Treg) cells, which suppress immune responses, can provide TGF-β for the differ-entiation of ΤH17 cells, but it is now known that virtually all cells in the body can secrete TGF-β. Nonetheless, in response to some pathogens or pathogen components, dendritic cells, which are responsible for presenting antigens to T cells, also secrete IL-6 and TGF-β, leading to TH17-cell differentiation3. In keeping with this

finding, local — but not systemic — blockade of TGF-β action prevents TH17-cell differen-tiation and so inhibits the onset of autoimmu-nity in EAE mice, an experimental model of multiple sclerosis3.

Another study4 has shown that ablation of TGF-β production by T cells compromises the differentiation of TH17 cells and results in immune abnormalities mediated by TH1 and TH2 cells. But the question remained: which cell is the most natural source of TGF-β for TH17-cell differentiation in vivo?

Torchinsky et al.1 propose that the physiolo-gical differentiation of TH17 cells is triggered by the simultaneous synthesis of IL-6 and TGF-β by dendritic cells that have engulfed — through the process of phagocytosis — infected cells undergoing programmed cell death (apoptosis). They show that when infected apoptotic B cells or neutrophils, two other types of immune cell, are maintained in culture with dendritic cells, the dendritic cells express a battery of genes involved in TH17-cell differen-tiation, including genes encoding IL-6, TGF-β and IL-23 (Fig. 1a). These dendritic cells or their secreted products promote the preferen-tial differentiation of TH17 cells. By contrast, dendritic cells that have engulfed uninfected apoptotic cells induce the differentiation of Treg cells (Fig. 1b).

It has long been known that phagocytosis of apoptotic cells leads to the secretion of various immune-inhibitory cytokines, including IL-10, TGF-β and prostaglandins5–7. It was therefore

assumed that the resolution of inflammation following phagocytosis of apoptotic cells is accompanied by the active suppression of cytokines that stimulate inflammation. Yet Torchinsky et al. show that whether the dying cells are infected or uninfected dictates whether an inflammatory situation is ameliorated by the induction of Treg cells or driven to new heights by the induction of TH17 cells.

The seemingly reciprocal development of inducible Treg cells and TH17 cells through the effect of TGF-β in the absence or presence of IL-6 was also noted in earlier studies8,9 of TH17 cells, and remains a firmly entrenched idea. But it has yet to be proven that the production of inducible Treg cells by TGF-β has a major role in immune responses in vivo. In the periph-eral immune system, most Treg cells seem to be derived from the thymus and are maintained by homeostatic mechanisms10; so whether, under normal physiological conditions, there is even a need for the differentiation of these cells in the peripheral immune system is not known. Furthermore, if an inflammatory context can so easily ‘derail’ inducible Treg-cell generation towards TH17-cell differentiation, it might be rather risky to consider basing therapeutic intervention on the premise that TGF-β secre-tion induced by phagocytosis of apoptotic cells will lead to Treg-cell induction and so immune suppression.

Torchinsky and colleagues’ proposed mecha-nism1 may not be the only route to TH17-cell differentiation. But their suggestion that bacterial infection, which results in substantial apoptosis, may preferentially drive TH17-medi-ated responses is noteworthy. The authors used the bacterium Citrobacter rodentium, which causes excessive apoptosis of intestinal epithelial cells. They find that if apoptosis in response to infection with this pathogen is blocked, there is a fall in the levels of IL-17-expressing T cells in the small intestine and colon.

The differentiation of TH17 cells has also been associated with particular pathogen-sensing receptors (such as various Toll-like receptors and dectin 1 — a receptor involved in fungal recognition11). The differential role of these receptors in initiating responses medi-ated by TH17 cells remains to be determined. Furthermore, one should remember that not all immune-defence responses mediated by the IL-17 cytokine originate from TH17 cells. It will be interesting to know if and how provision of IL-17 by the innate immune system, and in particular the recruitment of neutrophils (short-lived cells, which are themselves prone to apoptotic death), provide the link between the innate IL-17 response and the induction of an adaptive TH17 response. ■

Brigitta Stockinger is in the Division of Molecular

Immunology, MRC Institute for Medical

Research, Mill Hill, London NW7 1AA, UK.

e-mail: bstocki@nimr.mrc.ac.uk

1. Torchinsky, M. B., Garaude, J., Martin, A. P. & Blander, J. M.

Nature 458, 78–82 (2009).

2. Veldhoen, M. et al. Immunity 24, 179–189 (2006).

Figure 1 | Infection, apoptosis and immune responses. a, Torchinsky et al.1 find that when dendritic cells ingest neutrophils infected with bacteria, they secrete TGF-β, IL-6 and IL-23 cytokines, leading to the differentiation of TH17 cells, which promote inflammation. b, By contrast, ingestion of uninfected apoptotic neutrophils causes the secretion of TGF-β and IL-10 by dendritic cells, promoting in vitro differentiation of T cells to induced regulatory T cells (iTreg cells), which suppress immune responses.

Dendritic

cells

Neutrophil

Infection

No infection

Phagocytosis

a

Proinflammatory

TH17 cell

Resting T cellTGF-βIL-6

IL-23

Bacteria

Apoptosis

without infection

Apoptotic

neutrophil

b

TGF-βIL-10

Anti-inflammatory

iTreg cell

44

NATURE|Vol 458|5 March 2009NEWS & VIEWS

5.3 n&v MH IF 445.3 n&v MH IF 44 27/2/09 17:31:0027/2/09 17:31:00

© 2009 Macmillan Publishers Limited. All rights reserved

3. Veldhoen, M., Hocking, R. J., Flavell, R. A. & Stockinger, B.

Nature Immunol. 7, 1151–1156 (2006).

4. Li, M. O., Wan, Y. Y. & Flavell, R. A. Immunity 26, 579–591

(2007).

5. Chen, W., Frank, M. E., Jin, W. & Wahl, S. M. Immunity 14, 715–725 (2001).

6. Fadok, V. A. et al. J. Clin. Invest. 101, 890–898 (1998).

7. Voll, R. E. et al. Nature 390, 350–351 (1997).

8. Bettelli, E. et al. Nature 441, 235–238 (2006).

9. Mangan, P. R. et al. Nature 441, 231–234 (2006).

10. Liston, A. & Rudensky, A. Y. Curr. Opin. Immunol. 19, 176–185 (2007).

11. LeibundGut-Landmann, S. et al. Nature Immunol. 8, 630–638 (2007).

PLANETARY SCIENCE

Volatility in Martian magmasHarry Y. McSween

The geochemistry of the Martian surface has largely been determined by the eruption of magmas to form basaltic rocks. A new line of argument has chlorine as an influential agent in that process.

The volatile constituents of magmas exercise a wholly disproportion-ate influence on magma behaviour. Although they are present in only tiny quantities, these elements and compounds with low boiling points profoundly affect thermodynamic properties and crystallization pat-terns, and the characteristics of magmatic flow and eruption. When we speak of magmatic volatiles, we normally mean water and carbon dioxide. But other volatile species might well be important, especially on planets other than Earth. Hence the interest in a proposal, reported in Chemical Geology by Filiberto and Treiman1, that chlorine is one such player.

Mars is thought to be richer in volatile elements than Earth2, and has been seen as a good place for further study. The Mars Exploration Rover Spirit has encountered and analysed numerous basaltic rocks during its five-year trek across the Gusev crater. These rocks are produced by volcanic activity, and the abun-dant vesicles found in some of them (Fig. 1) testify to the presence of significant quanti-ties of volatiles. But until now, water has been assumed to be the main volatile agent, as for instance was the case in the hydrous crystal-lization experiments performed on a sample of Gusev basalt3.

Filiberto and Treiman1, however, propose that chlorine may have a central role in the generation and evolution of Martian basalts. They have carried out experiments demon-strating that the effects of chlorine on basalt phase equilibria are similar to those of water. The addition of chlorine shifts the liquidus — the stage at which crystals begin to form — to lower temperatures and enlarges the stability field of pigeonite, a form of pyroxene that is a common constituent of basalts. This is notable because the pressure under which the two min-erals olivine and pigeonite first appear together

is assumed to correspond to the depth at which magma is generated, and that depth is signifi-cantly shifted towards the surface if chlorine is the volatile species.

No chlorine-containing minerals have been identified in rocks of the Gusev crater, but that is no surprise because Spirit does not have instruments to do this. Conversely, although high levels of chlorine have been measured chemically by one of Spirit’s spectrometers, there is the possibility that those measure-ments are a reflection of the ubiquitous pres-ence of Martian dust, which contains halides including chlorine.

The broader picture must take account of the fact that much of what we infer about magmatic volatiles on Mars derives from studies of cer-tain Martian meteorites rather than rocks still on Mars. These meteorites — basaltic achon-drites, which show the distinctive features of processing by volcanic activity — are remark-ably anhydrous, leading some geochemists to speculate that the mantle sources for these

magmas must also have been dry. That view has been challenged4 by the need for dissolved water to explain the calcium abundances and soluble-light-element zoning patterns in pyroxene crystals of some Martian basaltic meteorites. In this model, the magmas lost all their water as they were ejected from Mars, aided by the planet’s low gravity.

The water contents of Martian meteorites, however, remain a contentious subject. The most persuasive argument for a role for chlo-rine is based on the absence of water in nomi-nally hydrous daughter minerals (apatite and amphibole, both of which contain chlorine) in melt inclusions trapped in some Martian meteorites5.

Filiberto and Treiman’s thought-provoking paper1 will serve a good purpose in prompt-ing renewed debate about volatiles in Mar-tian magmas. But to my mind it is doubtful

that magma-borne chlorine would dominate on Mars. The channels and valley networks on Mars were surely eroded by water, and magmas must have delivered that water to the sur-face. Volcanic rocks in the Gusev cra-ter are estimated to be billions of years old, comparable in age to some of the channels. The existence of hydrogen, which was mapped at high latitudes by Mars Odyssey’s Gamma Ray Spec-trometer6, and which presumably is in the form of subsurface water ice, also points to the outgassing of water from magmas on a global scale.

Finally, however, there is a possible terrestrial connection to Filiberto and Treiman’s line of investigation. They point out that lavas in certain tectonic environments on Earth can have a chlorine content of up to 0.7% by weight, and they propose that its effect in terrestrial magmas might be similar to that suggested for Mars.

Experimental evidence7,8 supports the idea that water is required to produce two different magmas characteristic of two different tectonic circumstances: so-called andesitic magmas in plate subduction zones, and alkaline magmas in intraplate settings. Filiberto and Treiman speculate that partial melting and fractional crystallization in chlorine-rich terrestrial sys-tems might mimic the effect of water, an idea that merits further testing. ■

Harry Y. McSween is in the Department of Earth

and Planetary Sciences, University of Tennessee,

Knoxville, Tennessee 37996-1410, USA.

e-mail: mcsween@utk.edu

1. Filiberto, J. & Treiman, A. H. Chem. Geol. doi:10.1016/

j.chemgeo.2008.08.025 (2008).

2. Dreibus, G. & Wänke, H. Icarus 71, 225–240 (1987).

3. Monders, A. G., Médard, E. & Grove, T. L. Meteorit. Planet.

Sci. 42, 131–148 (2007).

4. McSween, H. Y. et al. Nature 409, 487–490 (2001).

5. Sautter, V., Jambon, A. & Boudouma, O. Earth Planet. Sci.

Lett. 252, 45–55 (2006).

6. Feldman, W. C. et al. Science 297, 75–78 (2002).

7. Grove, T. L. et al. Contrib. Mineral. Petrol. 142, 375–396 (2002).

8. Nekvasil, H. et al. J. Petrol. 45, 693–721 (2004).

Figure 1 | Volatile traces on Mars. The vesicular nature of this basaltic rock, photographed on the plains of the Gusev crater, is evidence of the pervasive presence of gas bubbles and so of volatile agents. (Pancam image by the Spirit rover from sol 740.)

45

NATURE|Vol 458|5 March 2009 NEWS & VIEWS

5.3 n&v MH IF 455.3 n&v MH IF 45 27/2/09 17:31:0027/2/09 17:31:00

© 2009 Macmillan Publishers Limited. All rights reserved