of material into the stratosphere, blocking out sunlight: Krakatoa (Indonesia) in 1883, Santa Maria (Guatemala) in 1902, Agung (Indonesia) in 1963–64, El Chichón (Mex-ico) in 1982, and Pinatubo (Philippines) in 1991. These jumps exist in both the land and ocean data. But one shift remained a puzzle: a significant drop in SSTs from 1945 to 1946 that was not replicated in the land data. This shift is also present, but not as obvious, in the un filtered data. In an effort to explain this change, Thompson et al. looked to the meta-data — in particular, to the provenance of SST measurements from around 1945.

Figure 1 hints at the explanation; it shows the total number of SST observations from the various national temperature archives. What’s striking is that both the relative fractions and the total numbers of observations vary consid-erably from year to year. These changes per-vade the record, but unsurprisingly the two World Wars (1914–19 and 1939–45) represent the most significant shifts, both in source and in the total number of observations. And here, Thompson et al. suggest, we have the clue to the jump exposed in 1945: whereas during the preceding war years, 80% of measurements came from ships flying the US flag, a resump-tion of UK measurements at the end of the war saw their proportion jump to some 50%. At that time, unlike their American counterparts who took engine-intake measurements, the British relied primarily on uninsulated-bucket measurements.

So, what are the implications? Most imme-diately, a further correction to the global temperature series will be necessary, albeit of a magnitude yet to be assessed. There are many wider ramifications to consider, but one should be handled directly: should we doubt the rise in global mean temperatures during the twentieth century as a result of this or other hidden, and as yet undiscovered, biases in the record?

The answer is no. According to the filtering of natural variabilities that Thompson and col-leagues have done, the only major discontinu-ity in SSTs is the one in 1945 (although other, insignificant shifts are dotted through the record). The shift from upwards-biased engine-intake measurements to downwards-biased bucket measurements demands a correction; naively speaking, temperatures between 1942 and 1945 would shift downwards by a mag-nitude of, say, 0.3 °C. Global warming would remain a reality — it would just be a bit more than previously thought.

How does this anticipated correction ripple through to climate models? Global mean surface temperatures are the most widely used data for evaluating the predictive capabilities of models on continental and larger scales6–8. They are also crucial for evaluating two other principal uncertainties in climate predictions: the factors forcing climate change (primarily, levels of aerosol particles in the atmosphere) during the twentieth century and thus in the

future6–11; and the rate of heat uptake from the atmosphere to the ocean (Fig. 1 in ref. 7).

The SST adjustment around 1945 is likely to have far-reaching implications for model-ling in this period. One particularly striking example can be found in a figure prominently displayed on page 11 of the 18-page ‘Sum-mary for Policymakers’12 of the IPCC’s Fourth Assessment Report, produced last year. In this, the observed decadal mean temperatures of the 1940s — those that contain the anomalous 3–4-year interval dominated by (upwardly biased) US engine-room-intake measurements — are the only ones to lie above model predictions. Although we don’t know exactly how the tem-perature record prior to 1946 will be affected by the awaited correction, it is a safe bet that temperatures in this particular decade will be lower.

The 1940s just happen to fall at the end of what seemed to be a warming trend from the 1910s that was followed by a weak decline in global temperatures into the 1960s. Climate modellers have explained the warming as a response to natural forcings13, and the cooling as due to an increase in tropospheric aerosols, principally sulphates, as a result of increased economic activity in the decades following the Second World War. This tem po rarily offset the effects of man-made warming. Data analysts, on the other hand, have considered the maximum in the 1940s to be the expression of a natural fluctuation14. In light of the new finding1, each interpretation will need to be reconsidered — the first of many implications that will need to be explored. ■

Chris E. Forest is in the Joint Program on the Science and Policy of Global Change, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, USA. Richard W. Reynolds is in the US National Climatic Data Center, National Oceanic and Atmospheric Administration, 151 Patton Avenue, Asheville, North Carolina 28801, USA.e-mails: ceforest@mit.edu; richard.w.reynolds@noaa.gov

1. Thompson, D. W. J., Kennedy, J. J., Wallace, J. M. & Jones, P. D. Nature 453, 646–649 (2008).

2. www.ipcc.ch/ipccreports/assessments-reports.htm3. www.climatescience.gov/Library/sap/sap1-1/default.php4. Kent, E. C. & Taylor, P. K. J. Atmos. Ocean. Technol. 23,

464–475 (2006).5. Kent, E. C. & Kaplan, A. J. Atmos. Ocean. Technol. 23,

487–500 (2006).6. Forest, C. E. et al. Science 295, 113–117 (2002).7. Stott, P. A. & Forest, C. E. Phil. Trans. R. Soc. A 365,

2029–2052 (2007). 8. Knutti, R., Stocker, T. F., Joos, F. & Plattner, G.-K.

Clim. Dynam. 21, 257–272 (2003).9. Andronova, N. G. & Schlesinger, M. E. J. Geophys. Res. 106,

22605–22612 (2001).10. Hegerl, G. C. et al. in Climate Change 2007: The Physical

Science Basis (eds Solomon, S. et al.) 663–746 (Cambridge Univ. Press, 2007).

11. Meehl, G. A. et al. in Climate Change 2007: The Physical Science Basis (eds Solomon, S. et al.) 747–846 (Cambridge Univ. Press, 2007).

12. www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf (2007).

13. Stott, P. A. et al. Science 290, 2133–2137 (2000).14. Schlesinger, M. E. & Ramankutty, N. Nature 367, 723–726

(1994).15. Woodruff, S. D., Slutz, R. J., Jenne, R. L. & Steurer, P. M.

Bull. Am. Meteorol. Soc. 68, 1239–1250 (1987). 16. Kent, E. et al. Bull. Am. Meteorol. Soc. 88, 559–564 (2007).17. Slutz, R. J. et al. Comprehensive Ocean–Atmosphere Data Set;

Release 1 (NOAA Environmental Research Laboratories, Climate Research Program, Boulder, CO, 1985).

IMMUNOLOGY

Soothing intestinal sugars Marika C. Kullberg

The gut is a new frontier in microbiology, offering many opportunities for innovative investigation. The finding of one such study is that intestinal inflammation in mice can be tamed by bacterial sugars.

The human ‘gut flora’ consists of between 300 and 1,000 microbial species, and some 1014 microorganisms in total (about ten times the number of cells of the human body). We usually live in harmony with these microbes, and would be less healthy without them. For example, they synthesize essential vitamins and amino acids, and also degrade otherwise indigestible plant material, as well as certain drugs and pollutants.

On page 620 of this issue, Mazmanian et al.1 report that Bacteroides fragilis, a com-mon bacterium of the lower gastrointestinal tract in mammals, can prevent intestinal inflammation in mice. Specifically, the authors show that polysaccharide A (PSA) of B. fragilis prevents gut inflammation induced by another

bac terium, Helicobacter hepaticus, or by the chemical compound TNBS (2,4,6-trinitro-benzene sulphonic acid).

This is an exciting finding, not least given that the incidence of human intestinal inflam-mation and inflammatory bowel disease has increased steadily in the Western world since the early 1950s. These conditions, which include Crohn’s disease and ulcerative colitis, are believed to stem in part from inappropri-ate immune responses to the gut microbiota2. In the healthy intestine, immune balance is regulated by different types of white blood cells called CD4+ T lymphocytes. These cells include CD4+ effector T lymphocytes (which help us fight pathogens by secreting various immune mediators called cytokines) and CD4+

602

NATURE|Vol 453|29 May 2008NEWS & VIEWS

Dendritic cell

a

b

c

Activateddendritic cell

+ PSA

+ PSA

Hh

Hh

Hh

IL-23, IL-12,IL-6, TNF-α, IL-1β

Inflammatory TH1/TH17CD4+ effector T cell

ImmunosuppressiveCD4+ regulatory T cell

CD4+ T cell

IFN-γIL-17

↓TNF-α, IL-1β↑IL-10

↓TNF-α, IL-1β↑IL-10 IL-10

Sub-optimal effectorT-cell activation

regulatory T lymphocytes (which, through their production of the cytokines IL-10 and TGF-β, dampen the effector T cells when their action is no longer needed). When the balance between these types of T cell is disturbed, the immune response goes awry and intestinal inflammation occurs2.

Evidence in support of the theory that bacteria trigger gut inflammation came from the discovery that mutations in NOD2, a host immune-cell receptor involved in detecting bacterial peptidoglycan, are associated with an increased risk of Crohn’s disease3,4. Fur-thermore, treatment with broad-spectrum antibiotics or probiotics (beneficial microbial species5) can improve health in patients with inflammatory bowel disease. Probiotics might inhibit the growth or invasion of pathogenic bacteria, or strengthen the gut-wall barrier. They may also stimulate the production of IgA (antibodies that are secreted into the gut lumen), and of IL-10 and TGF-β (ref. 5).

Mazmanian et al.1 used two experimental approaches. In the first, colitis was induced in mutant (Rag–/–) T-cell-deficient mice by infect-ing the animals with H. hepaticus and giving them CD4+ effector T cells. These effector cells start to respond to H. hepaticus, but in the absence of counterbalancing host regulatory T cells intestinal inflammation develops within a few weeks. The cytokine IL-23 plays a key role in this inflammatory response, and the disease is associated with a TH1/TH17 effector T-cell response to H. hepaticus6 (Fig. 1a). The second approach involves a chemical-induced colitis, in which the administration of TNBS to normal mice leads to acute inflammation within a few days.

In their new paper, Mazmanian et al.1 demonstrate that giving mice B. fragilis at the same time as the colitis-inducing agents improves the animals’ health. A mutant of B. fragilis that lacks PSA could not prevent inflammation, implicating PSA in helping to maintain immune balance. Finally, colitis did not develop when purified B. fragilis PSA was administered together with H. hepaticus and effector T cells to the Rag–/– mice, or when this polysaccharide was given to the TNBS recipients. Together, these results provide direct evidence that PSA prevents intestinal inflammation in both model systems.

Bacteroides fragilis synthesizes at least eight distinct surface polysaccharides as part of its capsule, or coat7, PSA being the most abundant. Although CD4+ T cells normally recognize and respond to peptide fragments of proteins, PSA can be taken up by so-called antigen-present-ing cells, such as dendritic cells, and presented to CD4+ T cells, resulting in T-cell activation7. Mazmanian et al.1 show that IL-10 secreted by T cells is essential for PSA to protect against colitis. But it is not yet clear whether PSA has induced ‘true’ IL-10-secreting regulatory T cells in the mice protected from colitis. Thus, although the authors demonstrate that PSA cannot prevent colitis when T cells

cannot produce IL-10, this could be due to a more pathogenic nature of such IL-10-deficient effector T cells.

What about the molecular mechanism by which PSA prevents H. hepaticus-induced colitis? One possibility is that this sugar acts on dendritic cells, thereby altering their capacity to trigger an efficient effector-T-cell response (Fig. 1b). Support for this hypothesis comes from studies showing that T cells isolated from PSA-treated animals are hyporesponsive8.

Alternatively, is PSA inducing IL-10-secret-ing regulatory T cells (Fig. 1c)? Treatment with filamentous haemagglutinin (FHA) from Bordetella pertussis, the causative agent of whooping cough, can protect Rag–/– mice from colitis induced by CD4+ effector T cells9. T cells isolated from these FHA-treated mice produced IL-10 in vitro following polyclonal stimulation9, supporting the model shown in Figure 1c. It will be interesting to see whether the same observation (that IL-10 is produced by T cells isolated from disease-protected mice) applies in the H. hepaticus colitis model in which PSA prevents intestinal inflammation.

Moreover, do CD4+ T cells isolated from disease-free mice infected with B. fragilis plus H. hepaticus respond to B. fragilis and/or H. hepaticus antigens? If so, what cytokines

do such bacterium-specific T cells produce? Has PSA skewed the T-cell immune response to H. hepaticus antigens away from a pro-inflam-matory TH1/TH17-type response (Fig. 1a) towards IL-10 secretion (Fig. 1c)?

There is also the question of whether PSA can act against H. hepaticus-induced colitis not only when administered from the start of an experiment, but also when used to treat established disease. The long-term aim, of course, is to develop drugs to cure intestinal inflammation in humans. In this regard, treat-ment of patients suffering from inflammatory bowel disease with live eggs from the porcine whipworm, Trichuris suis, has produced promising results10. These effects are believed to be due to the induction by T. suis of regu-latory T cells and factors such as IL-10, TGF-β and prostaglandin E2 that help main-tain immune balance10. One view is that the apparent boons of modern life — anti-biotics, vaccines and improved sanitation — have reduced the incidence of parasitic worms and other microbes, and therefore also of disease-protective molecules such as ‘T. suis-like antigens’ or PSA. That in turn may have altered the way our immune system responds to challenges, leading to the increased incidence of inflammatory diseases.

Figure 1 | Intestinal inflammation, and how B. fragilis PSA may prevent it in mice1. a, In mice that develop colitis, dendritic cells activated by H. hepaticus (Hh) produce pro-inflammatory cytokines (such as IL-23, IL-12, IL-6, TNF-α and IL-1β). They also present Hh antigen to CD4+ T cells that then differentiate into inflammatory TH1/TH17 CD4+ effector T cells; these cells are specific for Hh antigen, and produce the inflammatory cytokines IFN-γ and IL-17 (ref. 6). b, c, Dendritic cells jointly stimulated with Hh and B. fragilis PSA are less responsive, with reduced secretion of the pro-inflammatory cytokines TNF-α and IL-1β, and enhanced production of anti-inflammatory IL-10 (ref. 1). They present both Hh antigen and PSA to CD4+ T cells, and may either (b) produce only sub-optimal effector T-cell activation or (c) induce the differentiation of IL-10-producing CD4+ regulatory T cells. The antigen specificity (Hh antigen and/or PSA) of such regulatory T cells is unknown.

603

NATURE|Vol 453|29 May 2008 NEWS & VIEWS

Ever-improving molecular techniques are providing tantalizing glimpses of the gut ecosystem11. With the launch of the Human Microbiome Project12, which plans to charac-terize the human microbiota and analyse its role in human health and disease, we are set to see considerable advances in understand-ing how host–microbial interactions may affect human health. When such information will translate into new therapeutic approaches is, however, anyone’s guess. ■

Marika C. Kullberg is in the Immunology and Infection Unit, Department of Biology,University of York, and The Hull York Medical School, PO Box 373, York YO10 5YW, UK. e-mail: mk512@york.ac.uk

malignancy7. Postovit and colleagues found that metastatic tumour cells do not express Lefty. Their results strongly support stromal regula-tion of malignancy and indicate that Lefty has a suppressive effect on cancer cells. The authors’ work also suggests that factors secreted by the tumour stroma, and their derivatives, could be used as treatments to ‘reprogramme’ the dif-ferentiation of malignant cells, suppressing tumour development and growth.

Although modulating tumour properties in invasive cancers — as discussed in the Finak and Postovit papers3,4 — could reduce the associated morbidity and mortality, early diag-nosis and prevention are even more effective means of preventing cancer-associated death. To address the clinical problems of cancer at these early stages, understanding the molecular processes underlying cancer initiation and pro-gression is crucial. A paper by Hu et al.5 pub-lished in Cancer Cell addresses the mechanism of breast-cancer transition from a localized (in situ) lesion to an invasive form.

The authors used a cell line that, when injected into mice, mimics aspects of an early, non-malignant form of human breast cancer called ductal carcinoma in situ (DCIS)8, by forming non-invasive lesions in the animals’ mammary gland. They next studied the role of myoepithelial cells in these lesions in sup-pressing the transition of DCIS to malignancy. (Myoepithelial cells separate the basement membrane of the duct from the epithelial cells that face the duct lumen.)

Hu and colleagues’ functional analysis of cell-type-specific gene expression identified several pathways that could be essential for interactions between stromal fibroblast cells and myoepithe-lial cells in controlling the integrity of a tissue’s basement membrane. These pathways, which modulate myoepithelial-cell differentiation, are mediated by essential signalling molecules such as TGF-β, Hedgehog, cell-adhesion mol-ecules and the gene transcription factor p63. Malfunction of these signalling pathways leads to the loss of myoepithelial cells and subsequent invasion of the basement membrane by their adjoining epithelial cells, which respond to sig-nals originating from fibroblasts. Determining whether the loss of myoepithelial cells is a cause or a consequence of the transition from in situ disease to invasive cancer will help to dictate therapeutic strategies.

Myoepithelial cells secrete a protein called maspin, which inhibits degradation of the extracellular matrix, an event thought to be essential for the transition from in situ cancer to an invasive form9. Moreover, this crucial tumour-suppressor protein is postulated to affect tissue invasion, programmed cell death and blood-vessel development. Hu and col-leagues5 identified several extracellular-matrix metalloproteins that are implicated in cancer transition to the invasive state, but maspin is not one of them. Perhaps distinct subtypes of pre-malignant tumours use different pathways for the transition. Recent characterization of

CANCER

Whispering sweet somethings Thea Tlsty

That genetic mutations contribute to cancer is undisputed. What now emerges is that a cancer cell’s microenvironment has a much stronger hand in the course a cancer takes than previously thought.

The goal of personalized medicine is to tailor a treatment to a specific cellular target at the appropriate stage of a disease, thus ‘defusing’ the disease process. Cancer is an example of the way in which multifaceted approaches to attaining this goal are emerging. We have come to appreciate that a tumour is a collection of diverse cells — cells carrying cancer-causing mutations and the cells of its immediate micro-environment — that act in concert towards disease progression1,2. Three studies3–5 illus-trate how these cells collude, and focus on the contribution of non-tumour cells.

Within tissues, epithelial cells are supported by a connective framework called the stroma, which itself consists of specific cells, such as fibroblasts, endothelial cells and immune cells, as well as the extracellular matrix. Physiologi-cal processes occurring in this compartment, for example the development of new blood vessels in response to oxygen shortage, and host immune responses, could dictate cancer progression.

Writing in Nature Medicine, Finak et al.3 set out to examine how gene-expression profiles in cells of the stroma are affected by cancer. Com-paring morphologically normal and tumour stroma from the breast tissue of patients with breast cancer, they identify gene-expression patterns that are strongly associated with a specific outcome of the disease and that could be used as predictors of disease progression.

One specific predictor, a group of 26 genes that the authors call the stroma-derived prog-nostic predictor (SDPP), stratifies the risk of breast-cancer progression using molecular markers that are independent of — but add

power to — both standard clinical prognos-tic factors, such as the presence or absence of tumour cells in adjacent lymph nodes, and the more recently described6 predictors based on gene expression. SDPP identifies stromal subtypes that have gene-expression profiles relating to a good or poor outcome of breast cancer.

The clinical significance of work such as that of Finak et al. is twofold. First, discerning the subtleties of cell–cell interactions within the microenvironment of a malignant lesion (a localized, disease-associated change in a tissue) will indicate which particular therapy might be most effective for the specific biology of that tumour. Second, such insights could provide targets for developing new therapies. Finak et al. find that SDPP is not affected by treat-ment, suggesting that existing therapies do not target host responses that affect SDPP genes.

Reporting in Proceedings of the National Academy of Sciences, Postovit et al.4 use a con-temporary approach to address the question of the stromal contribution to cancer malignancy. In this exciting study, the authors use an in vitro three-dimensional model that exposes cancer cells to the microenvironment to which human embryonic stem cells are normally exposed; they were hoping to identify conditions in the stroma that suppress malignant characteristics of cancer cells.

Stromal cells surrounding embryonic stem cells secrete a protein factor called Lefty, which inhibits the Nodal protein. Nodal, which during embryonic development prevents stem-cell differentiation, is abnormally expressed in human tumour cells, causing

1. Mazmanian, S. K., Round, J. L. & Kasper, D. L. Nature 453, 620–625 (2008).

2. Coombes, J. L., Robinson, N. J., Maloy, K. J., Uhlig, H. H. & Powrie, F. Immunol. Rev. 204, 184–194 (2005).

3. Hugot, J.-P. et al. Nature 411, 599–603 (2001).4. Ogura, Y. et al. Nature 411, 603–606 (2001).5. Sartor, R. B. Curr. Opin. Gastroenterol. 21, 44–50 (2004).6. Kullberg, M. C. et al. J. Exp. Med. 203, 2485–2494

(2006).7. Mazmanian, S. K. & Kasper, D. L. Nature Rev. Immunol. 6,

849–858 (2006).8. Stingele, F. et al. J. Immunol. 172, 1483–1490 (2004).9. Braat, H. et al. Gut 56, 351–357 (2007).10. Elliott, D. E., Summers, R. W. & Weinstock, J. V. Int. J.

Parasitol. 37, 457–464 (2007).11. Dethlefsen, L., McFall-Ngai, M. & Relman, D. A. Nature

449, 811–818 (2007).12. http://nihroadmap.nih.gov/hmp

See Editorial, page 563.

604

NATURE|Vol 453|29 May 2008NEWS & VIEWS