122 NATURE MATERIALS | VOL 13 | FEBRUARY 2014 | www.nature.com/naturematerials

news & views

dominated by the interactions between the DNA ligands, the surface energies of individual DNA-capped nanoparticles and their connectivity could be designed so as to obtain a particular (macroscopic) crystal shape. That is, from knowledge of the local DNA interactions and thermodynamic considerations (such as the Wulff construction) one should be able to predict both the interactions among the individual nanoparticles and the macroscopic crystallization outcome. This connection between the microscopic interactions and the macroscopic behaviour in DNA-mediated nanoparticle crystallization is consistent with conventional crystal-growth theory; hence, concepts of crystallization theory may be extended from atoms and molecules to ‘nanoparticle atoms’ connected by programmable DNA bonds.

What’s more, the analogies between atomic, molecular and DNA-mediated nanoparticle crystallization suggest that one could think of a multidimensional ‘periodic table’ for nanoparticles by drawing from the already massive library of both DNA motifs with varied topologies and nanoparticles with unique size- and shape-dependent properties9. In fact, nanoparticle atoms and DNA motifs can already be combined to generate unique DNA-based nanoparticle building blocks10. The possibility of applying the theoretical framework of atomic crystallization to these building blocks will enhance the control of the assembly of a wider variety of macroscopic crystal shapes and sizes with programmable properties. In this regard, Mirkin and colleagues’ observations and insights set the stage for more to come. ❐

Shogo Hamada1, Shawn J. Tan2 and Dan Luo1,3 are at 1Kavli Institute at Cornell for Nanoscale Science, Cornell University, Ithaca, New York 14853, USA, 2Institute of Materials Research and Engineering, 3 Research Link, Singapore 117602, Singapore, 3Department of Biological and Environmental Engineering, Cornell University, Ithaca, New York 14853, USA. e-mail: dl79@cornell.edu

References1. Seeman, N. C. J. Theor. Biol. 99, 237–247 (1982).2. Winfree, E., Liu, F., Wenzler, L. A. & Seeman, N. C.

Nature 394, 539–544 (1998).3. Zheng, J. et al. Nature 461, 74–77 (2009).4. Mirkin, C. A. et al. Nature 382, 607–609 (1996).5. Park, S. Y. et al. Nature 451, 553–556 (2008).6. Nykypanchuk, D., Maye, M. M., van der Lelie, D. &

Gang, O. Nature 451, 549–552 (2008).7. Cheng, W. et al. Angew. Chem. Int. Ed.

49, 380–384 (2009).8. Auyeung, E. et al. Nature 505, 73–77 (2014). 9. Zhang, C. et al. Nature Mater. 12, 741–746 (2013).10. Tan, S. J., Campolongo, M. J., Luo, D. & Cheng, W.

Nature Nanotech. 6, 268–276 (2011).

CANCER IMAGING

Lighting up tumoursDetection of a wide range of tumours remains a challenge in cancer diagnostics. By exploiting changes in the tumour microenvironment, a pH-responsive polymeric nanomaterial enables ultrasensitive tumour-specific imaging in many types of cancer.

Daishun Ling, Michael J. Hackett and Taeghwan Hyeon

The ultimate goal for cancer diagnosis is the development of an imaging probe that is sensitive enough to

differentiate tumours from normal tissues in the early stages of disease. Tumours, however, usually contain heterogeneous cell populations of diverse genotypes and phenotypes, which present great challenges for specific cancer detection1. At the cellular level, no two cancers are identical, so it is impossible to establish a universal strategy for tumour detection by targeting specific cancer biomarkers. Conversely, at the macroscopic level, some commonalities emerge within the local tumour environment. Now, writing in Nature Materials, Gao and colleagues report an imaging strategy that is ultrasensitive to the tumour microenvironment: they have developed a polymeric, micelle-based nanoprobe that is highly responsive to both the angiogenic tumour vasculature and the extracellular pH (pHe; Fig. 1)2.

Tumours attempt to grow uncontrollably, which comes at great energetic expense, leading to a very high rate of glycolysis under both aerobic and anaerobic conditions. This causes a build-up of

lactic acid, which is excreted by tumours leading to a decreased pHe of ~6.5–6.8, compared with 7.4 for normal tissue and blood3. Although the pH itself is variable from tumour to tumour, it is always acidic in nature. Extrapolating this idea further, sustained tumour growth requires a sustainable supply of nutrients. To achieve this, tumours increase their share of blood flow by releasing a deluge of angiogenic growth factors that initiate the generation of many new blood vessels4. These vessels tend to grow too quickly and, like any job completed too fast, the process is sloppy. The result is the emergence of large gaps or ‘fenestrae’ on the order of several hundred nanometres5. These gaps allow tumours to filter the blood in a similar manner to the kidneys and the liver; both of which also have fenestrated endothelia. As a consequence, nanoparticles that are normally constrained to the systemic circulation find a new site of accumulation: filtration into the tumour6.

Results from bioresponsive polymeric materials have been promising for cancer imaging, drug delivery and cancer therapy because of their capability to differentiate

diseased and healthy tissue7. Indeed, directing fluorescent nanoparticles to the tumour microenvironment seems to be a promising strategy for broadly applicable tumour detection, but it is only half the battle. Problems arise because the tumour pHe is not drastically different from blood, and it is very difficult to design a pH-dependent chemical probe that is sensitive enough to realize a fast and sharp signal amplification over such a modest pH drop. As such, usually only very weak signals are observed in tumours.

The nanoprobe reported by Gao and colleagues, however, is ultra-pH-sensitive with a sharp and tunable response to pH change. The pH-sensitive nanoprobes comprise an ultra-pH-sensitive core, a cell-specific targeting moiety and a series of conjugated fluorophores. These components work synergistically to greatly improve the imaging functionality of the nanoprobe.

First, Gao and colleagues introduced tertiary amines with controlled hydrophobic substituents as ionizable hydrophobic blocks for the pH-sensitive core. Nanoprobes with different transition pH values can thus be achieved by using

© 2014 Macmillan Publishers Limited. All rights reserved

NATURE MATERIALS | VOL 13 | FEBRUARY 2014 | www.nature.com/naturematerials 123

news & views

tertiary amino groups that protonate at a different pH. Subsequently the polymers were derivatized with hydrophobic fluorophores. At physiological pH, the monomers spontaneously assemble with the fluorophores oriented into the centre of the particle and in close proximity with each other. As a result of this, the fluorophores tend to silence each other by a process called fluorescence resonance energy transfer8. Thus the particle, despite being full of fluorophores, exhibits very little fluorescence (Fig. 1). As the particle navigates the body and enters the tumour site, the drop in pH begins to protonate the amines. The protonation of even a few monomers causes the particle to destabilize as a consequence of electrostatic repulsion and it dissociates into monomers. Because the fluorophores are not in close proximity, they no longer suppress each other, leading to a sharp increase in fluorescence. In fact, over the pH drop from 7.4 to 6.7, the nanoprobes show a 102-fold increase in fluorescence intensity; this compared with theoretically a 2- to 3-fold increase over the same range for small-molecule pH sensors, which acidify slowly according to the Henderson–Hasselbalch equation. Moreover, because of the tunable transition pH values from diverse tertiary amine groups, the nanoprobe can be tuned to target either the extracellular tumour (pHe, 6.5–6.8) or the more acidic endocytic organelles (pHi, 5.0–6.0).

Second, a polyethylene glycol (PEG) shell provides stealth-like characteristics and long-term plasma stability. The PEG is further modified with a targeting unit, cyclic RGD (cRGD), which binds to αvβ3 integrins overexpressed on the tumour neovasculature. The increased circulation time allows the nanoprobe to find the integrins in the tumour vasculature and the cRGD moiety promotes uptake into the cells via endocytosis. The nanoprobes within the late endosome experience a more significant drop in pH, which is then exploited to amplify the fluorescent signal of the particles, and hence to highlight the tumour vasculature.

Last, a variety of fluorophores ranging from green to near infrared could be conjugated to the ionizable block of the copolymer to enable colour-dependent, simultaneous imaging of both the tumour extracellular space as well as the neovasculature. The results show both the tumour vessels and interstitial space in the tumour parenchyma were illuminated, whereas no observable fluorescence was observed within the plasma of the tumour vasculature. This indeed demonstrates that the nanoprobes are highly quenched

in the blood circulation, then become dramatically activated in response to the low extracellular pH or neovascular uptake in tumours. In fact, the tumour-to-blood fluorescence ratio is very high with >300-fold increase in fluorescence after particle uptake.

To determine the generality of these nanoprobes, Gao and colleagues tested their fluorescent ones in a variety of cancer models (including transgenic, orthotopic and subcutaneous) and they showed impressive results across the board. Based on this, the researchers suggest the pH-activatable nanoprobe can be a robust and universal tumour imaging agent. Moreover, even small tumours exhibit

these properties, making these probes applicable to both the early detection of primary tumours as well as the detection of metastases. This should also prove useful in observing the size of tumours over time, both for preclinical drug evaluation and potentially as a clinical method to observe treatment progress in a rapid, reliable way.

Although this strategy overcomes many difficulties in targeted cancer imaging and holds great promise for imaging-guided surgery, several issues must be addressed before translating this technique to the clinic. Because fluorescence can only penetrate up to few millimetres of tissue9, a radioactive or magnetic resonance imaging agent, despite a lower sensitivity, would be

OFF

pH = 7.4

ON

Bloodflow

Angiogenictumour vessels

Nor

mal

vas

cula

ture

Homo FRET Fluorescence emission

Cancer cells

hv1hv1 hv2

H+

pHe(6.5–6.8)activation

pHi(5.0–6.0)activation

e

Figure 1 | A schematic representation of pH-sensitive imaging of the tumour microenvironment using ultra-pH-sensitive nanoprobes2. The nanoprobes are latent particles when in circulation in the bloodstream but dissociate into highly fluorescent monomers within the acidic milieu of tumours (pHe, 6.5–6.8) or endocytic vesicles (pHi, 5.0–6.0) of the tumour endothelium. FRET, fluorescence resonance energy transfer.

© 2014 Macmillan Publishers Limited. All rights reserved

124 NATURE MATERIALS | VOL 13 | FEBRUARY 2014 | www.nature.com/naturematerials

news & views

more compatible with clinically accepted methods of cancer imaging10. Additionally, Gao and colleagues suggest the preliminary animal studies showed negligible toxicities with regard to the tertiary-amine-based polymeric micelles. However, the literature is rife with many toxic tertiary amino polycations that have been known to induce cell necrosis11. More systemic and long-term toxicological studies are required before the translation of this powerful nanoprobe to clinical adaptation. Despite the fact that much work remains to be done, the extremely high sensitivity makes

these nanoprobes promising candidates for clinical tumour diagnosis. The purpose of this research, namely proof-of-concept for exploiting several general commonalities between tumours, has been clearly demonstrated. In this regard, ultra-pH-sensitive nanoprobes may have a bright future in oncodiagnostics. ❐

Daishun Ling, Michael J. Hackett and Taeghwan Hyeon are in the Center for Nanoparticle Research, Institute for Basic Science (IBS), Seoul 151-742, Korea, and School of Chemical and Biological Engineering,

Seoul National University, Seoul 151-742, Korea. e-mail: thyeon@snu.ac.kr

References1. Gerlinger, M. et al. N. Engl. J. Med. 366, 883–892 (2012).2. Wang, Y. et al. Nature Mater. 13, 204–212 (2014).3. Trédan, O., Galmarini, C. M., Patel, K. & Tannock, I. F. J.

Natl Cancer Inst. 99, 1441–1454 (2007).4. Wang, R. et al. Nature 468, 829–833 (2010).5. Roberts, W. G. & Palade, G. E. Cancer Res. 57, 765–772 (1997).6. Matsumura, Y. & Maeda, H. Cancer Res. 46, 6387–6392 (1986).7. Mura, S., Nicolas, J. & Couvreur, P. Nature Mater.

12, 991–1003 (2013).8. Zhou, K. et al. J. Am. Chem. Soc. 134, 7803–7811 (2012).9. Ntziachristos, V. et al. Nature Biotechnol. 23, 313–320 (2005).10. Kim, J., Piao, Y. & Hyeon, T. Chem. Soc. Rev.

38, 372–390 (2009).11. Dagmar, F. et al. Biomaterials 7, 1121–1131 (2003).

MATERIA

L WITN

ESS

Materials scientists are well accustomed to basing their materials choices on a careful balancing of performance criteria, for example trading off toughness, hardness and cost. It’s less common to have to base those choices on geopolitical criteria. But such stark realities are not so unfamiliar, perhaps most notably in recent years in concerns about the Chinese near-monopoly on rare-earth elements, and continuing fluctuations in the availability of tantalum for semiconductor electronics due to political instability in the Congo region. When abrupt materials shortages have occurred, substitutes have sometimes been found. The isolation of southeast Asian colonial rubber plantations during the Second World War prompted seminal work in Europe and the USA on synthetic rubber. And when civil war in Zaire in the 1970s impaired the supply of cobalt, an important component of magnets, cobalt-free designs were developed.

It might be tempting to suppose, then, that whenever shortages of materials components occur, alternatives will soon emerge. That would be complacent, according to a recent report by Graedel and colleagues1. They have considered as many as possible of the major uses for 62 different metals and metalloids, and whether there are known alternatives that could substitute for them if supplies become scarce. In many cases there are, but for 12 of the

elements examined, substitutes are either inadequate or non-existent at present. What’s more, none of the 62 elements have good replacements for all of their major uses.

Although scarcity of important materials is an ancient issue — one famous, although controversial, case was the British navy’s wood shortage in the seventeenth and eighteenth centuries due to deforestation — it is rendered more critical today by several factors. The market for materials is more global, and so more susceptible to international affairs. And the complexity of materials usage has increased. In engineering alloys, for example, a steady enhancement of performance by an accumulation of ingredients means that the formulations of the superalloys used for high-temperature applications may include a dozen or more elements. Environmental considerations now play a much bigger role in resource exploitation, even while the rapid growth of some emerging economies has increased the demand on non-renewable sources.

It’s for such reasons that in 2006 the US National Research Council performed an audit of economically important materials, developing a framework for assessing their ‘criticality’ on the basis of both the importance of their uses and the security of their sources2. Several metals, including rhodium, manganese, platinum and niobium,

were deemed by these criteria to be ‘at risk’. Graedel et al. seek to extend that work, concurring, for example, with the assessment for the former two metals while offering more reassurance for the latter.

What are we to do? Simply recognizing the problem (and abandoning a naive faith in the ability of markets to produce substitutes) is a start. A systematic enumeration of the risks — which applications of ‘at risk’ materials would be hit first, for example? — would provide a framework for assigning priorities. But arguably this problem might compel a realignment of some of the materials community’s research objectives: instead of obsessing over improved performance, more attention might be given to maintaining current performance by other means. ❐

References1. Graedel, T., Harper, E. M., Nassar, N. T. & Reck, B. K.

Proc. Natl Acad. Sci. USA http://dx.doi.org/10.1073/pnas.1312752110 (2013).

2. National Research Council Minerals, Critical Minerals, and the U. S. Economy (National Academies, 2008).

NO SUBSTITUTE

PHILIP BALL

© 2014 Macmillan Publishers Limited. All rights reserved