NEWS & VIEWS NATURE|Vol 436|28 July 2005

468

and radioactive contamination in the detector(around 28%). Moreover, geoneutrinos pro-duced from K decays are not — yet — detectedat KamLAND, because their energies arebelow the threshold of 1.8 MeV required totrigger the existing detector system.

The data reported by Araki et al.1 are theresults from their first experiment, whichcomprised just over two years of counting.Future observations at KamLAND, and at the Borexino detector under the Gran Sassomountain in central Italy, which begins opera-tion in 2006, will generate more data and provide greater sensitivity in testing the nature and sources of geoneutrinos. A crucialadvance will be to confirm that the geo-neutrino heat flux moving radially outwardsfrom the Earth is directly proportional to theradiogenic heat flux. This will, however,require an exact knowledge of the abundanceand distribution of K, Th and U in the Earth.

To this end, a first detailed assessment hasbeen made11 of the predicted geoneutrino flux relative to the distribution of radioactive elements in the regional crust and underlyingmantle near KamLAND, and throughout theEarth’s interior. Further in the future, combin-ing angle-integrated geoneutrino fluxes at dif-ferent detector sites12 with element distributionmaps will enable us to construct geoneutrinotomographic maps of the Earth that will tell usmore about the planet-wide distribution of K,Th and U. Proposed sites for future (anti)neu-trino detectors must therefore be sure toinclude areas beneath both continental regionsrich in K, Th and U and oceanic regions wherethe three radionuclides are depleted.

The pioneering results from KamLANDpresented by Araki et al.1, along with data fromfuture work, will provide a fundamental con-straint for the Earth’s U and Th budget (and, itis to be hoped, shortly for that of K), and definethe fractional contribution of radioactive heat-ing to the total energy budget. Later this year,particle physicists and Earth scientists willgather to discuss these exciting and commonareas of research at a meeting on Hawaii13. ■

William F. McDonough is in the Department ofGeology, University of Maryland, College Park,Maryland 20742, USA.e-mail: mcdonoug@geol.umd.edu

1. Araki, T. et al. Nature 436, 499–503 (2005).2. Eguchi, K. et al. Phys. Rev. Lett. 90, 021802 (2003).3. Ahmad, Q. R. et al. Phys. Rev. Lett. 89, 011301 (2002).4. Pollack, H. N., Hurter, S. J. & Johnson, J. R. Rev. Geophys.

31, 267–280 (1993).5. Hofmeister, A. M. & Criss, R. E. Tectonophysics 395,

159–177 (2005).6. McDonough, W. F. in Treatise on Geochemistry Vol. 2

(ed. Carlson, R. W.) 547–568 (Elsevier, Oxford, 2003).7. Palme, H. & O’Neill, H. St C. in Treatise on Geochemistry

Vol. 2 (ed. Carlson, R. W.) 1–38 (Elsevier, Oxford, 2003).8. Rama Murthy, V., van Westrenen, W. & Fei, Y. Nature 423,

163–165 (2003).9. Lee, K. K. M. & Jeanloz, R. Geophys. Res. Lett. 30, 2212 (2003).10. Labrosse, S. Phys. Earth Planet. Inter. 140, 127–143 (2003).11. Fiorentini, G., Lissia, M., Mantovani, F. & Vannucci, R.

preprint at www.arxiv.org/hep-ph/0501111 (2005).12. Field, B. D. & Hochmuth, K. A. preprint at

www.arxiv.org/hep-ph/0406001 (2004).13. www.phys.hawaii.edu/~sdye/hnsc.html

before the vasculature shut down, and allowedthe staged release of the two drugs. Morespecifically, the delivery of the anti-angiogenicfactor could lead to a collapse of the vascularnetwork and imprison the vehicle — still bear-ing its second payload of chemotherapeuticdrug — in the tumour. The subsequent releaseof the latter drug within the tumour would killthe cancer cells.

The authors exploited the fact that the bloodvessels of tumours are ‘leaky’7, so tumour tissue can take up larger particles than cannormal tissues, promoting selectivity. Theycreated composite vehicle particles of 80–120nm, consisting of a solid biodegradable poly-mer core surrounded by a lipid membrane(Fig. 1). The anti-angiogenic drug combreta-statin was dissolved in the lipid layer, fromwhich it rapidly escaped. This drug attacks theinternal skeleton of cells, and quickly disruptsblood vessels. The chemotherapeutic drugdoxorubicin was bound chemically to theinner core of the particle, and so was releasedmore slowly as the bond holding the drug tothe polymer broke down. Doxorubicin is acommon chemotherapeutic agent, and itsstructure consists of chemical groups that areamenable to attachment to polymers.

Sengupta et al. examined the effects of thedrugs on two types of tumour in mice, andshowed that, unsurprisingly, either drug aloneslowed tumour growth, and that when thedrugs were delivered simultaneously there wasan additive effect. Strikingly, however, thestaged release of the two drugs using the newdelivery vehicle improved the outcome stillfurther — survival time increased fromapproximately 30 days when the drugs weredelivered simultaneously to more than 60 dayswhen they were released sequentially. Thedelivery vehicles tended to accumulate in thetumours, rather than in other body tissues,and the drugs they transported killed bothendothelial and cancer cells.

The effect of the sequential delivery of thesetwo drugs on tumour growth is dramatic, butwe cannot assume a quick translation of theseresults to therapy for humans. The biologicaldifferences between mice and humans preventdirect comparison between the systems, and itwill also be important to extend these studiesto longer time periods. Moreover, it has beenspeculated that anti-angiogenic drugs mayactually promote the spread of tumours toother tissues, owing to a complex feedbackloop, although there is no evidence of this in

CANCER

One step at a timeDavid Mooney

Traditional chemotherapy kills tumour cells directly; some newer drugswork instead by cutting the tumour’s blood supply. An innovative approachcombines these strategies sequentially to pack a double whammy.

In 1971, Judah Folkman proposed that theprogression of cancer might be halted by preventing tumours from recruiting newblood vessels (a process called angiogenesis) to provide them with oxygen and nutrients. Lastyear, this theory bore fruit with the approvalby the US Food and Drug Administration ofthe first anti-angiogenic cancer treatment,Avastin (also known as bevacizumab)1. Sen-gupta and colleagues (page 568 of this issue)2

advance this concept by designing a drug-delivery vehicle that sequentially releases ananti-angiogenic drug and a traditionalchemotherapeutic drug at high concentrationsspecifically into a tumour. They report thattheir strategy can slow tumour growth in micemore than can either drug alone or the twodrugs delivered at the same time.

Traditional chemotherapeutic agents kill allrapidly growing cells in the body — both can-cer cells and other cells that divide quickly (forexample, blood, hair and cells lining the intes-tine). This leads to the distressing side effectsof chemotherapy, and limits the practical doseand frequency of application of the drugs. Onetactic to avoid these effects is to target the drugspecifically to the tumour, and approachesbeing tested include the incorporation ofdrugs into materials or complexes that caneither be placed in, or directed to, tumours3.

A second issue, however, is that sometumours develop resistance to a particulardrug, so efforts to identify targets that are not prone to developing resistance continue.Endothelial cells, which line blood vessels,may provide an attractive target, as they arethought to be genetically more stable than can-cer cells and so less likely to develop mutationsthat might promote resistance. A number ofdrugs that kill endothelial cells or prevent theirgrowth are proving effective in phase III clinical trials for treating colon, kidney andlung cancer, and gastrointestinal stromaltumours1,4–6. These drugs can be useful alone,but they are commonly combined with tradi-tional chemotherapy to prevent blood-vesselgrowth while also killing cancerous cells.

Simultaneous delivery of chemotherapeuticand anti-angiogenic drugs is clearly beneficial,but because chemotherapy is blood-borne,shutting down the tumour’s blood supply withanti-angiogenic drugs may decrease the deliv-ery of drugs designed to kill the tumour cells.Sengupta et al.2 hypothesized that a moreeffective strategy would be to use a deliveryvehicle that became concentrated in tumours

28.7 News & Views 467 new2 MH 25/7/05 3:04 PM Page 468

Nature Publishing Group© 2005

© 2005 Nature Publishing Group

NATURE|Vol 436|28 July 2005 NEWS & VIEWS

469

humans8. It is promising, in this regard, thatSengupta and colleagues’ system produced noincrease in the expression of a factor (HIF-1�)that can link the low oxygen levels resultingfrom reduced blood flow with potential resis-tance to drug therapy and tumour invasive-ness. Finally, in contrast to combretastatin,many anti-angiogenic drugs require pro-longed tissue exposure to shut down the vasculature, and so may not be amenable tothe particular approach described by Senguptaand colleagues.

The general concept of timing the availabil-ity of drugs aimed at specific stages or targets in cancer is widely applicable, however, and is consistent with similar efforts to promoteblood-vessel formation in diseases involvinginsufficient blood flow9. Appropriate design ofdrugs will allow targeting of cancer cells orother specific cell types10, and the deliverydevice described by Sengupta et al. could read-ily be modified for this. It may also be necessaryto target multiple aspects of angiogenesis, eitherby using several drugs or by using a drug thatinterferes with several pathways (for example,MAPK inhibitors)11, to prevent tumours fromswitching on alternative angiogenesis path-ways. Ultimately, combining the developmentof advanced drug-delivery systems with theidentification of early markers of cancer mayallow early and highly effective intervention,and help to accomplish the US National CancerInstitute’s stated goal of eliminating the suffer-ing and death from cancer by 2015. ■

David Mooney is in the Division of Engineeringand Applied Sciences, Harvard University,Cambridge, Massachusetts 02138, USA. e-mail: mooneyd@deas.harvard.edu

1. Hurwitz, H. et al. N. Engl. J. Med. 350, 2335–2342 (2004).2. Sengupta, S. et al. Nature 436, 568–572 (2005).3. Moses, M. A. et al. Cancer Cell 4, 337–341 (2003).4. Escudier, B. et al. Annu. Meet. Am. Soc. Clin. Oncol.

abstr. LBA4510; www.asco.org/ac/1,1003,_12-002636-00_18-0034-00_19-0032211,00.asp (2005).

5. Sandler, A. B. et al. Annu. Meet. Am. Soc. Clin. Oncol.abstr. LBA4; www.asco.org/ac/1,1003,_12-002643-00_18-0034-00_19-0033325,00.asp (2005).

6. Demetri, G. D. et al. Annu. Meet. Am. Soc. Clin. Oncol.abstr. 4500; www.asco.org/ac/1,1003,_12-002643-00_18-0034-00_19-0034169,00.asp (2005).

7. Jain, R. K. Adv. Drug Deliv. Rev. 46, 149–168 (2001).

Figure 1 | Step-by-step in fighting cancer. The delivery system of Sengupta et al.2 causes the sequential loss of blood vessels and the death of tumour cells. a, Nanometre-scale particles have an outer lipid layer (blue) and an inner core (yellow). b, Once injected into the bloodstream, the particle is selectively takenup into tumour tissues, where the lipid layer rapidly releases a drug that kills endothelial cells and disrupts blood vessels. c, The inner core gradually releasesa chemotherapeutic drug to destroy the cancer cells (d).

CARBON CYCLE

The age of the Amazon’s breathPeter A. Raymond

The inorganic carbon carried in rivers of the Amazon basin seems to originatelargely from the decomposition of young plant material — a finding thatimproves our understanding of the role of rivers in the carbon cycle.

Increases in atmospheric carbon dioxide fromthe burning of fossil fuels have unknowneffects on the global climate and economy. Sci-entists aim to understand more about theseeffects by studying the mechanisms that con-trol the exchange of carbon between land, theatmosphere and the oceans. Processes thatremove CO2 from the atmosphere, where itcould cause global warming, and move it intolong-term storage on land or in the oceans, areof particular interest in this context. Becausethey connect land and sea, rivers are a vitallink in these processes (Fig. 1, overleaf). Onpage 538 of this issue, Mayorga et al.1 provideinsights into how this river linkage works forthe world’s largest river system — the Amazon.

There are two main forms of carbon: organic(such as the biomass within a tree) and inor-ganic (CO2 in the atmosphere, for example).These forms are intimately coupled throughphotosynthesis in plants, which creates organicfrom inorganic carbon, and decomposition,which returns plant-produced carbon to itsinorganic form. In rivers, organic and inorganic

carbon exist in approximately equal propor-tions, and originate mainly when rainfall hitscontinental surfaces and either dissolves car-bon, or carries it to rivers in particulate form.

A single river can drain a landscape that hasa wide array of plant species, land uses, soilsand climatic zones. This complexity has madeit difficult to pin down exactly where mostriver carbon originates, how long it existed onland before being carried to a river, and howreactive it might be once in a river and, later, inthe coastal ocean. Most of the carbon in riversultimately comes from atmospheric CO2 andtherefore represents mobile ‘greenhouse car-bon’ that either cycles back to the atmosphereand contributes to global warming, or enters astorage compartment that is not in contactwith the atmosphere (coastal sediments, forexample; Fig. 1).

The Amazon basin is a central player in theglobal carbon balance because it stores largeamounts of carbon in biomass above ground,and this carbon is being returned to the atmos-phere by slash-and-burn agriculture2. But the

8. Blagosklonny, M. V. Cancer Cell 5, 13–17 (2004).9. Richardson, T. P., Peters, M. C., Ennett, A. & Mooney, D. J.

Nature Biotechnol. 19, 1029–1034 (2001).10. Nori, A. & Kopecek, J. Adv. Drug Deliv. Rev. 57, 609–636

(2005).11. Kim, D. W., Lu, B. & Hallahan, D. E. Curr. Opin. Invest. Drugs

5, 597–604 (2004).

28.7 News & Views 467 new2 MH 25/7/05 3:04 PM Page 469

Nature Publishing Group© 2005

© 2005 Nature Publishing Group