Cancer Research: Past, Present and Future

PERSPECTIVES

Moreover, scientists are exploring the pos- THE NEXT 10 YEARS — VIEWPOINT sibility of targeting pathogens, such as the Epstein–Barr virus-encoded oncoproteins, Cancer research: past, present to improve radiation therapy in patients. Notably, with the rapid development of molecular imaging methods that target and future receptors, metabolic enzymes, apoptosis and blood vessels, imaging is becoming a power- Ya Cao, Ronald A. DePinho, Matthias Ernst and Karen Vousden ful tool to understand the development and progression of cancer. Improved imaging Abstract | Research into cancer over the past 10 years has diverged enormously, techniques will also aid in the development partly based on the large number of new technologies that are now at our finger of anticancer drugs and help to evaluate tips. With areas of cancer research so disparate, it is not always easy to identify therapeutic effects in cancer patients. where the next new findings and therapies might come from. With this in mind, we asked four leading cancer researchers from around the world what, in their opinion, Ronald A. DePinho. The past 10 years have we have learnt over the past 10 years and how we should progress in the next brought many advances across a broad front. It is not specific advances that excite me 10 years. but rather the collective whole. The cancer research community has reached a point of In your opinion, what have been the molecular mechanisms of carcinogenesis conceptual and technical maturity to mount most important findings in cancer from the genome-wide level4. On the basis a decisive assault on cancer. This optimism research in the past 10 years? of results obtained from the HGP regard- is fuelled by multiple important advances ing single-nucleotide polymorphisms — a deeper understanding of the biological Ya Cao. Cancer has been identified as a (SNPs) and genome-wide associated stud- processes of cancer; comprehensive multi- chronic disease1. In an attempt to battle ies (GWAS), numerous changes in SNPs dimensional cancer genome profiles; robust this disease the World Health Organization and gene signatures have been identified5. functional validation technologies; faith- have developed three principles on cancer Notably, these studies were carried out by ful genetically engineered mouse models; prevention: they have estimated that cur- combining large-scale population data and quantitative analysis of clinically annotated rent prevention strategies could prevent up in vitro findings to address the molecular biospecimens; enabling nanotechnology; to one-third of new cancers; they have sug- mechanisms of carcinogenesis. and other major discoveries. It is anticipated gested that improved early screening could The study of epigenetic changes that that cancer death rates will continue their result in the detection of one-third of cancers occur during carcinogenesis is rapidly steady decline through prevention, more at an early stage; and they have proposed developing into an important field of study. effective early detection, genotype-informed that a comprehensive treatment strategy Epigenetic silencing that occurs through the drug application and the combined use of could improve survival and quality of life for CpG island methylator phenotype (CIMP), targeted therapies, including those harness- another one-third of patients with advanced for example, embodies a novel viewpoint in ing the immune system. The increased use of cancer2. These strategies offer the most cost- cancer diagnosis and therapy6. The dynamic target-engagement and response-prediction effective, long-term control of cancer. interplay between epigenetic signatures and biomarkers in clinical trials holds promise Cancer is a vastly complex disease post-translational modifications, including for reducing the staggeringly high rate of exhibiting a plethora of changes in multiple phosphorylation, acetylation and ubiqui- failure in cancer drug development that has genes. More and more attention is now tylation, corresponds to the complicated been seen in recent years, often occurring in being focused on the relationship between regulation of carcinogenesis. In addition, late-stage clinical testing at an enormous cost infection and cancer. For example, about non-coding RNAs, such as microRNAs, are to the pharmaceutical industry and to the 200,000 women die every year from cervi- also involved in the signalling networks that patient. There is also a sense that the combi- cal carcinoma, which is closely associated are associated with carcinogenesis. nation of existing and emerging therapeutic with human papillomavirus (HPV) infec- Translational medicine is regarded as a agents will elicit more prolonged responses tion. Importantly, a vaccine against HPV switch from concept to practice. The change that currently elude at least one-third of was the first cancer vaccine to be devel- from 1‑B (from bench to bed) to 2‑B (from cancer patients. oped and should substantially reduce the bed to bench) has allowed scientists and incidence of cervical cancer. Besides HPV, clinicians to focus more on patients in order Matthias Ernst. Much of the cancer Helicobacter pylori infection is linked with to directly address the three most common research carried out in the past decade has gastrohelcosis, which is the precancerous issues in cancer prevention: diagnosis, ther- been preoccupied with assigning the vast stage of gastric cancer3. Clearly, the rela- apy and healing7. Drugs that target specific genomic information generated through tionship between H. pylori and gastric molecules, such as the epidermal growth The Cancer Genome Atlas (TCGA)9 and cancer is an important basis for developing factor receptor (EGFR), are being used more similar projects, and to the cancer hall- an efficient therapeutic strategy for con- and more to treat cancer patients8. Currently, mark concepts introduced by Hanahan and trolling H. pylori infection and, ultimately, clinicians are combining radiation therapies Weinberg1. We can now better distinguish gastric cancer. with pharmacological interventions, such as the few driver mutations from the many In the post-Human Genome Project checkpoint inhibitors and molecules target- passenger mutations in individual cancer (HGP) era, researchers and scientists are ing signal transduction pathways and the genomes, and evidence indicates that tailor- again recognizing the importance of the cancer microenvironment, to treat cancer. ing treatment to molecular signatures rather

than to the histological grade of tumours High-throughput sequencing of cancer is undermined by an endogenous safeguard increases response rate10. However, we now genomes has revealed that somatic muta- response: the induction of apoptosis12,13 and also appreciate a mutational landscape with tions are clustered in a small number of oncogene-induced senescence14. The reliance few ‘mountains’ that represent common signalling circuits and that such mutations of these intracellular stress mechanisms on genes mutated in many cancers, and a large often circumvent the need for physiologi- intact p53 has highlighted the therapeutic number of ‘hills’ that represent the majority cal engagement by ligand-occupied growth potential of the large proportion of tumours of infrequently mutated genes that are found factor receptors. Although the discovery of that have not yet lost wild-type p53 and that in a small number of cancers11. In light of the non-coding RNAs has added a new layer of therefore remain susceptible to therapeutic vast discordance between these hills, even complexity to the molecular drivers of cancer activation of p53 pro-apoptotic pathways. among cancers of the same tissue origin, we development, they frequently affect the same Meanwhile, our molecular understanding of now understand the incremental advantages recurring pathways. We predicted that the the balance between anti-apoptotic BCL‑2 they individually confer to tumour progres- ever-increasing number of mutations in these proteins and their pro-apoptotic BH3‑only sion. As a corollary, we now face the more pathways would follow the oncogene addic- counterparts heralded the first clinical trials daunting prospect of having to therapeuti- tion paradigm and would equate to increased with BH3‑mimetics to curb the excessive cally target many hills rather than, as origi- proliferation and tumour growth. Instead, survival signals received by cancer cells. nally anticipated, focusing solely on the few and in particular for pathways including RAS Autophagy is another example of a physi- mountains. or MYC, we have realized that this concept ological response hijacked by cancer cells that enables them to survive conditions of 15 The contributors* prolonged cellular stress . The induction of autophagy in cancer cells can also contribute Ya Cao is the Vice Director of the Cancer Research Institute of Xiangya School of Medicine, Central to the paradoxical cytoprotective effect that South University and the Dean of the Key Laboratory of Carcinogenesis and Cancer Invasion, occurs in response to radiotherapy and some Ministry of Education, Changsha, Hunan, China. She worked as a visiting scientist on six occasions chemotherapy, and this explains why drugs at the National Cancer Institute at Frederick, Maryland, USA, between 1989 and 1997. She is involved in many of the grant-awarding bodies in China, including the National Natural Science that inhibit autophagy are now being investi- Foundation, the State Key Development Program for Basic Research in China and the China gated in the clinic. Medical Board. Her research focuses on the mechanism and biological importance of the Inflammation has long been linked to Epstein–Barr virus (EBV) encoded latent membrane protein 1 (LMP1) and whether targeting LMP1 cancer, but only in the past decade have we can improve the response of EBV-driven nasopharyngeal carcinoma (NPC) to radiotherapy. Her understood how inflammatory cells, and research also focuses on the epigenetic changes in NPC that is driven by EBV infection. other components of the tumour stroma, Ronald A. DePinho is the new President of the MD Anderson Cancer Center, Houston, Texas, USA. fuel tumour progression by creating a His current position is the Director for the Belfer Institute for Applied Cancer Science, at the Dana microenvironment that is enriched for the Faber Cancer Institute, Boston, USA, and the American Cancer Society Research Professor and same cytokines that are the protagonists of Professor of Medicine and Genetics at Harvard Medical School, Boston, USA. He received a the chronic inflammation associated with bachelor’s degree in biological sciences from Fordham College, New York, USA, and received his MD liver, stomach and colon cancers. Indeed, with distinction from Albert Einstein Medical College, USA, in 1981. He later completed his research inflammation is a tumour-enabling charac- training at Columbia Presbyterian Hospital, USA, and in 1998, joined the Dana Faber Cancer teristic that affects and converges on almost Institute. His research has focused on key genetic events in the life-history of cancer cells, including all types of solid cancers to enable many of those that affect cell cycle control, cellular survival, genome stability, host–tumour interactions and tumour maintenance and the translation of such knowledge into clinical end points. the hallmarks of cancer and to foster the progression of incipient neoplasias into Matthias Ernst is currently the Interim Director of the Ludwig Institute Melbourne-Parkville Branch, full-blown malignant tumours. Melbourne, Australia. He graduated from the Swiss Institute of Technology (ETH) in Zurich, Sophisticated knockout studies in mice Switzerland, and trained as a postdoctoral fellow at the Merck Sharp and Dohme Research and zebrafish have provided us with a daz- Laboratories in the United States where he focused on bone biology. He initially joined the Ludwig Institute in Melbourne in the laboratory of Ashley Dunn to extend his studies into the molecular zling view of the plasticity of many systems mechanism of intracellular signal transduction. In 1996 he was recruited as a deputy head for the and their inbuilt redundancies. This is now bone biology department at Novartis, Switzerland, but returned to the Ludwig Institute recognized as a major reason why even the Melbourne-Parkville Branch in 1998 as the joint-laboratory head of the Colon Molecular and Cell most efficacious targeted drugs fail to act as Biology Laboratory. His research focuses on reverse genetics in the mouse to dissect molecular long-lasting ‘magic bullets’ — and why even mechanism regulating epithelial homeostasis in the gastrointestinal tract in health and disease. the best personalized approaches mostly Karen Vousden is the director of Cancer Research UK’s Beatson Institute of Cancer Research delay disease progression but rarely eradicate in Glasgow, UK, and is the Director of the Institute of Cancer Sciences at Glasgow University, it. We could not anticipate how short-lived Glasgow, UK. She gained a BSc. degree in genetics and microbiology and a Ph.D. degree in the therapeutic responses would be owing to genetics in 1982, both from Queen Mary College, University of London, London, UK. Her early drug resistance, which seems to emerge with postdoctoral positions were with Chris Marshall at the Institute of Cancer Research, London, UK, the predictability of a ‘whac‑a-mole’. The and Douglas Lowy at the National Cancer Institute, Bethesda, USA. Following 7 years as a group recent discovery that cancer cells express leader at the Ludwig Institute, London, UK, she returned to the National Cancer Institute in 1995, different protein isoforms that limit the use this time in Frederick, USA, serving successively as Head of the Molecular Carcinogenesis section of the initially promising BRAF inhibitors of the ABL-Basic Research Program, Director of the Molecular Virology and Carcinogenesis 16–18 Laboratory, Interim Director of the ABL-Basic Research Program and Chief of the Regulation of serves as a prominent example . Cell Growth Laboratory, Division of Basic Sciences. She is interested in the mechanisms of tumour suppression, particularly the role of p53. Karen Vousden. The past 10 years have seen substantial advances in both our under- *The contributors are listed in alphabetical order. standing of cancer and the technologies that

are available to study it. There have also been Massive-scale sequencing has also of personalized or individualized medicine28. many exciting breakthroughs in the develop- allowed us to identify the likely common Clearly, a change from an emphasis on treatment of rationally designed therapies that driver mutations that underpin malignant ment to an emphasis on the prevention of offer us new ways of treating or preventing development4. There have been some cancer and on combining both treatment cancer. I’ve highlighted a sample of these astounding successes using this approach. and prevention will be instrumental for findings here — but there are many others of The identification of BRAF mutations in an effective strategy for eliminating this equal impact and importance. melanoma and the speed of development chronic disease. The development and implementation of and use of BRAF inhibitors in clinical trials Answering basic scientific questions the vaccine against genital HPV infection is is one example24. Neomorphic mutations in in cancer will continue to be a key point an incredible success story19. The potential the metabolic enzymes isocitrate dehydro- of emphasis in life science research. advantages of cancer prevention as opposed genase 1 (IDH1) and IDH2 are another25, Molecular profiles from all types of to cancer therapy are obvious, but many but there are many more. These and other tumours will be developed, and con- prevention strategies are difficult to put into large-scale information-gathering exercises ventional oncological pathology will be practice — especially those that involve long- are delineating the underlying networks and integrated into the discipline of molecular term lifestyle changes. The International pathways that are commonly and recur- pathology. The interplay between patho- Association for Research on Cancer (IARC) rently mutated in different cancer types, logical and morphological changes and monograph that concluded “HPV types 16 information that will be used to develop molecular signatures will guide diagnosis and 18 are carcinogenic to humans” was pub- and improve targeted therapies. A pleasing and therapy in cancer1. Metabolic repro- lished in 1995 (REF. 20), and the progress since extension to this is the concept of synthetic gramming is proving to be widespread in then has been remarkable. The HPV vac- lethal approaches26, as demonstrated so ele- cancer cells and is regarded as an emerging cine has yet to be proven to prevent cervical gantly by the use of poly(ADP ribose) poly- hallmark of cancers. Studying and elucidat- cancer, as this will take decades to ascertain, merase (PARP) inhibitors in BRCA-deficient ing the relationship of metabolic disorders but the implementation of a wide-scale vac- breast cancers. Although each of these and cancer should provide new ideas for cination programme in so many countries therapies may hit subsequent problems with molecular intervention mechanisms and — despite significant objections — is a major resistance and other unanticipated issues27, may also help to promote a new field that step forwards. This is a shining example of the proof of principal that they provide is will provide metabolism-based targets what can be achieved when epidemiologists, hugely satisfying. for cancer patients. In addition, ‘omics’ in molecular biologists, immunologists and cancer together with systems biology will public health clinicians collaborate. Where do you expect progress to be effectively answer numerous key scientific Unprecedented advances in our ability to made in the next 10 years? questions in carcinogenesis4. sequence genomes, measure the expression The increasing global migration patterns of genes and non-coding RNAs (the involve- Y.C. In the future, chemoprevention needs are challenging in epidemiology. Clearly, ment of which in cancer has been another to be an important part of cancer research epidemiological studies in populated areas massively exciting development21) across and will need to be promoted as a lifestyle that exhibit a high incidence of cancer may the entire genome, and analyse epigenetic, choice. Strategies to popularize knowledge need to be revisited and perhaps reinvesti- proteomic and metabolomic profiles on a of cancer prevention through health educa- gated. Samples acquired from susceptible similarly large scale are already having a big tion and tobacco control should help to populations and cases need to be stored impact on the direction of cancer research ensure that cancer prevention is promoted for future study. Moreover, the develop- and the vision of how cancer patients should as a key public health task to governments ment of molecular epidemiology will be treated22. It seems that these technolo- in many different countries. Currently, at provide some new evidence in aetiology, gies are putting the personalized analysis of least one-half of all cancers are diagnosed in diagnosis and therapy of cancer from the tumours tantalizingly within our grasp, and developing countries, and ways of targeting whole-population level. are helping to define and identify biomarkers and treating cancer in these countries are Developing anticancer drugs that are that should allow us to direct cancer therapy desperately needed. Vaccination is one pos- based on proven molecular targets will still based on an understanding of the pathways sible route, and the Chinese government has be an important topic. Cancer drug devel- that operate in an individual tumour, rather recently planned widespread immunization opment will move from validation and test- than crudely treating all malignancies from using the hepatitis B vaccine. This should ing in the traditional cancer clinical trial to a particular organ site as one disease. This help to reduce the incidence of liver cancer. biomarker-driven and hypothesis-testing concept of personalized medicine is not a Moreover, as much of this programme is trials29. These types of trials will specifi- new idea, of course. For example, George focused on controlling infection, the inci- cally address and focus on proof of mecha- Beatson of Glasgow, UK, first suggested the dence of infection- nism, proof of concept and individualized concept of hormonal control of cancers in a associated cancers should greatly decrease medicine with pharmacogenomics (that Lancet paper in 1896 (REF. 23), and differen- over the next 10 years. is, the ‘3Ps’). Biological hypothesis-testing tiating hormone-sensitive from hormone- Meanwhile, based on the success of trials will provide new opportunities in resistant breast cancers is now standard care screening for specific tumour protein personalized therapy for tumour patients. for patients with this disease. However, the markers, such as α-fetoprotein (AFP) and The trials will be based on new biomark- progress over the past decade in our ability prostate-specific antigen (PSA), several ers selected from laboratory studies and to characterize the inner workings of each new cancer biomarkers will be included in moved into translational trials, which tumour is raising the hope that this rational routine medical examinations. Moreover, should lead to the development of a unique approach can be applied to a much greater screening of specific molecular targets will model of molecular biomarker-based depth and to all cancer types. guide and enhance the effective development patient selection.

Translational medicine in cancer must be It is vital that we are not target and identification of those targets based on a close relationship between basic tempted to divert funding with the highest importance (not just any science research and clinical trials. Strategies validated target) for drug discovery. In in cancer prevention and therapy, as well away from basic research in addition, securing the knowledge needed as basic research, will provide the evidence the erroneous belief that we for the major advances in cancer drug needed for implementing personalized, somehow know enough. development will require a creative but predictive, preventive (or pre-emptive) and directed culture, it will take patience, com- participatory medicine (that is, the ‘4Ps’). prehensiveness and resources, and it will Furthermore, cooperation between multi- validation that is sufficient to attract require full unencumbered access to extant ple disciplines and maintaining a tight link substantial financial investments from models and technology. It is my view that between basic and clinical medicine will pharmaceutical companies that are needed academia is the only environment that is in establish a new mechanism for collabora- for late-stage clinical testing and commer- a position to take on the challenge of build- tive actions against cancer, the theme of this cialization. This model, which has yielded ing this construct of applied cancer science. years Annual World Congress on Cancer, marginal clinical and economic success, Such an applied cancer science programme which took place in China. In addition, the has experienced major changes that have would be responsible for the timely and sys- continued input of substantial finances from moved us farther away from a solution. tematic large-scale genetic and functional non-governmental organizations, such as Over the past few years, many large phar- analyses that are needed to define coopera- the Bill & Melinda Gates Foundation, should maceutical companies have downsized tive relationships among all known and help to further support research focused their internal research programmes with unprecedented cancer targets that govern on major health problems, such as cancer the goal of preserving capital to license the diverse biological hallmarks of cancer1. and HIV. drugs at clinical proof-of-concept from It is worth emphasizing that the issue is biotechnology firms. Coincident with this not whether a gene is important in cancer R.A.D. The TCGA project30–32 and func- trend, the recent economic crisis severely but in what genotypic or cell biological tional genomic screens33 are revealing limited the availability of venture capital, context its extinction would lead to tumour many new cancer genes, providing new resulting in historic lows in the creation regression. The ultimate goal of such efforts potential therapeutic points of attack. of new companies, particularly companies would be to generate rational co-extinction Although encouraging, it is worth empha- with early stage ‘lead’ drugs or drug targets strategies for specific cancers: targeting sizing that these genomic discoveries are that require a longer period of investment single elements is generally not sufficient just a beginning. The conversion of vali- before generating any financial return: a to elicit durable clinical responses34 — the dated targets into highly effective drugs route that has traditionally delivered the only goal that patients care about. with durable clinical responses requires lion’s share of our novel innovative drugs. As drugs are developed for these coop- a clear clinical path hypothesis for drugs On the academic front, although trans- erative targets, this organization would be entering the pipeline. Realization of such lational efforts have increased, the pre- responsible for testing these drugs, alone a programme necessitates an in-depth eminence of hypothesis-driven research and in combination, in a diverse array of preclinical effort to fully define the biol- and the emphasis on trainee development model systems in order to pressure-test the ogy and mechanism of a target, as well has not been conducive to the creation of clinical hypothesis to the maximum extent as a clear understanding of the genotypic multidisciplinary teams that are driven by possible at the preclinical stage. There is and cell biological context in which the the mandate to perform all of the critical a growing appreciation that if a drug does new target serves a rate-limiting function preclinical studies needed to fully under- not work in today’s refined in vivo models, in tumour maintenance. So, in the com- stand the importance of a novel target and then the chance for clinical success is low. ing decade, as the TCGA and functional to test new drugs against such targets in It is also my view that model system test- genomics deliver a comprehensive com- the best model systems available. The types ing should encompass the full range of pendium of the genetic events driving the of ‘applied’ experiments needed to make hundreds of genome-annotated cell lines major cancers, it is fair to ask whether the timely go or no-go decisions in cancer drug in two-dimensional (2D) and 3D cultures, cancer drug development ecosystem, span- development are often not aligned with primary human cancer tumour cells main- ning discovery to clinical proof-of-concept, the generation of high-profile manuscripts tained in vivo and genetically engineered is optimally structured to harness the full that support the aspirations of graduate mouse models of cancer, as each possesses clinical potential of these breathtaking and postdoctoral trainees for top careers in strengths and weaknesses and each provides scientific advances. I think not. Serious academia and industry. vital information on drug action alone and operational deficiencies prevent systematic So, in the decade ahead, what is needed in combination in specific genotypic and and reliable execution along the multitude is a new organizational construct. This tumour-type contexts. Beyond these pre- of preclinical activities that are needed construct should be embedded with the clinical efforts, target validation must con- to illuminate an optimal path forwards. rich, innovative culture of academic sci- tinue in early stage clinical testing for which Traditionally, academic institutions cele- ence but should also be populated by goal- pre-imaging, post-imaging and biopsies are brate the individual and serve as the major oriented, milestone-driven staff scientists, essential to leverage our emerging capa- engine for the discovery of novel targets. whose primary mission is to systematically bilities in tumour genotyping, assessment In a rather random process, a handful of enlist the entire compendium of emerging of target engagement and on-mechanism such early stage discoveries might then fuel cancer targets into a full array of quality- responses to therapy to guide genomics- venture-backed biotechnology companies, controlled in vitro and in vivo validation based medicine. Collectively, these invest- which would in turn drive these pro- assays that are designed to achieve full ments are modest relative to the cost of a grammes to a level of preclinical or clinical verification of the cancer relevance of a failed Phase III clinical trial.

The implementation of an applied cancer cells. This might be more straightforward researchers, and are likely to generate sub- science paradigm will require a cultural shift in tumours in which cancer stem cells func- stantial advances in cancer treatment over in academia to one that rewards team work, tion in a hierarchical fashion and possess the next 10 years. Examples include the applied science and the achievement of spe- phenotypes that are stable traits. Tumours importance of altered metabolism in the cific goals and milestones. The realization in which cells with stem cell-like proper- development and maintenance of malignan- of maximal progress in this coming decade ties arise in a stochastic manner and retain cies43 and the indisputable role of tumour will come with the confluence of knowledge, phenotypic plasticity, however, could prove cell interactions with the rest of the host technology and the organizational con- more challenging37. Understanding the in the form of stromal cells, angiogenesis, structs capable of driving discovery science nature of tumour cell plasticity is there- immune responses and so on44. Crucially, to clinical end points. Such an effort would fore rapidly becoming a key goal in cancer most of these discoveries came from basic deliver higher quality programmes for our research. molecular, developmental and cell research, patients in clinical trials, attract private The clinic will benefit from the use of and such fundamental science continues to sector investments to produce marketed existing (targeted) drugs for a wider range inform many of the rationally designed ther- products and ultimately mark this decade of tumours that may arise from fortuitous apies that are currently in clinical trials. It is as the one that dethroned this emperor of off-target effects, as has been illustrated vital that we are not tempted to divert fund- maladies. with the used of the BCR–ABL inhibitor ing away from basic research in the errone- imatinib against KIT-driven gastrointes- ous belief that we somehow know enough. M.E. With existing large international cancer tinal stromal tumours38. In turn, this will One unanticipated consequence of our consortia, the next decade will see unprec- encourage the pharmaceutical industry to recent technological advances — in par- edented opportunities arising from an explo- dedicate resources for the development of ticular the relative ease with which data on sion of DNA-sequencing capabilities for ever drugs for rarer types of cancer that express the genome, transcriptome, epigenome, decreasing costs, and to enable personalized a shared, prototypical cancer signature or proteome and metabolome can be gathered genome and expression data to become a a common mutated pathway. However, — is that we are in danger of drowning in reality. Supplemented with information on if RNA interference technologies can be information. It seems clear that biology must epigenetic modification, the cancer proteome, effectively harnessed and delivered to embrace mathematics, computing, informat- metabolome and all the other ‘omes’ that are patients, we may see alternatives to today’s ics and data analysis in order to make sense yet to come, this compendium will allow us drug development approaches by global of it all30. Above all, such analytical tools to reconstruct a timeline of progression for pharmaceutical companies, by enabling must always bear in mind that cancers, like individual tumours and provide novel oppor- many more players to explore an unprec- all living systems, have not been designed but tunities for personalized treatment. Equally edented number of targets. As we more have evolved. Making sense of things in biol- important, this information will define and completely understand the various molec- ogy is challenging because cause and effect validate stage-specific surrogate markers for ular requirements of any given cancer, we relationships can be difficult to discern or diagnosing disease and for monitoring treat- can more comprehensively focus on its dissect, and because things can have function ment response or relapse. Integrating all these various Achilles heels. Therapies will arise without an obvious purpose. There is a grow- data is the key to developing more effective by homing in on the augmented reliance of ing concern that the extent and nature of het- personalized and, most likely, multipronged the cancer cell on normal pathways, a phe- erogeneity within cancers may simply defy treatments that help to eradicate the entire nomenon that is referred to as non-onco- rational explanation and explication. How tumour, rather than merely interfering with gene addiction39, or on the need of a cancer can we make sense of data that not only doc- its growth. The ultimate success, however, will cell to alter its metabolome so that energy ument differences between individual can- depend on how well we can understand and production, macromolecule biosynthesis cers of the same type, but also changes within predict resistance, by combining the insight and redox reactions are balanced40. regions of the same cancer? The application gained from therapeutic investigations using Our emerging capacity to replicate parts of new technologies will help us to identify complimentary in vivo, in vitro and in silico of the cancer ecosystem in genetically tai- genes and epigenes that have an influence cancer models. lored and complementary in vitro and in vivo on cancer development, but integrating all Much effort will need to go into deci- model systems will see a more systematic of this information in a way that will be use- phering metastatic signatures, as metastases approach to comprehensively take advantage ful for making therapeutic decisions could are what ultimately threaten the patient’s of synthetic lethal interactions. The power prove difficult or perhaps impossible. One life, and examples from pancreatic cancer of this approach, most recently illustrated route through all of this information is to make a clear case that metastasis-initiating with PARP inhibitors in BRCA-mutant ask what is common to cancers, rather than cells require driver mutations beyond cancers41,42, has proved to be very success- what makes them all individual and unique. those required within primary tumours35,36. ful when it has been empirically exploited Finding common vulnerabilities to target for Epigenetic mechanisms, including DNA in the past. However, with the next decade the treatment of multiple cancer types would methylation and histone modifications, will generating an unprecedented number of new make for an extremely attractive and alterna- need to be factored in more vigorously, as targets and experimental compounds, as well tive way forwards. they not only confer heritable phenotypes as treatment modalities, we may need to re- Another area in which progress will hope- (such as the silencing of tumour suppressor evaluate the current Phase III paradigm to fully be made is in expanding the repertoire genes), but are also constant companions maximize benefit for the patient. of targets for new drug discovery. Although for the incremental steps of the neoplastic the concept of rationally designed therapies cell’s journey to metastatic disease. We also K.V. Several areas of biology that have has been amply vindicated, the range of need to better understand whether and been underappreciated for years are now mechanisms by which such drugs function how to therapeutically target cancer stem permeating the consciousness of all cancer has been limited. Many of the most attractive

and validated targets, such as transcription 4. Stratton, M. R. Exploring the genomes of cancer cells: 28. Workman, P. & de Bono, J. Targeted therapeutics progress and promise. Science 331, 1553–1558 for cancer treatment: major progress towards factors, have been deemed undruggable — (2011). personalised molecular medicine. Curr. Opin. surely this needs to change? New approaches 5. Broderick, P. et al. Deciphering the impact of Pharmacol. 8, 359–362 (2008). common genetic variation on lung cancer risk: a 29. Carden, C. P., Banerji, U., Kaye, S. B., Workman, P. & to developing drugs that do not depend on genome-wide association study. Cancer Res. 69, de Bono, J. S. From darkness to light with biomarkers the established rules are beginning to bear 6633–6641 (2009). in early clinical trials of cancer drugs. Clin. Pharmacol. 6. Teodoridis, J. M., Hardie, C. & Brown, R. CpG island Ther. 85, 131–133 (2009). fruit, and the concept that protein–protein methylator phenotype (CIMP) in cancer: causes and 30. Chin L, H. W., Getz G &Meyerson M. Making sense of interactions can be successfully modulated is implications. Cancer Lett. 268, 177–186 (2008). cancer genomic data. Genes Dev. 25, 534–555 7. de Bono, J. S. & Ashworth, A. Translating cancer (2011). 45 gaining credence . The ability to build drugs research into targeted therapeutics. Nature 467, 31. The Cancer Genome Atlas Research Network. from small fragments will hopefully open 543–549 (2010). Comprehensive genomic characterization defines 8. Quatrale, A. E. et al. EGFR tyrosine kinases inhibitors human glioblastoma genes and core pathways. Nature the door to attacking the most logical weak in cancer treatment: in vitro and in vivo evidence. 455, 1061–1068 (2008). spots in cancers. Furthermore, the discovery Front. Biosci. 16, 1962–1972 (2011). 32. The Cancer Genome Atlas Research Network. 9. Collins, F. S. and A. D. Barker. Mapping the cancer Integrated genomic analyses of ovarian carcinoma. of microRNAs and other non-coding RNAs genome. Pinpointing the genes involved in cancer will Nature 474, 609–615 (2011). may lead to completely novel ways of think- help chart a new course across the complex 33. Schreiber, S. L. et al. Towards patient-based cancer landscape of human malignancies. Sci. Am. 296, therapeutics. Nature Biotech. 28, 904–906 (2010). ing about how to design new cancer therapies. 50–57 (2007). 34. Stommel, J. M. et al. Coactivation of receptor tyrosine Conversely, our expanding knowledge of can- 10. Slamon, D. J. et al. Use of chemotherapy plus a kinases affects the response of tumor cells to targeted monoclonal antibody against HER2 for metastatic therapies. Science 318, 287–290 (2007). cer cell biology — such as the role of changes breast cancer that overexpresses HER2. N. Engl. 35. Campbell, P. J. et al. The patterns and dynamics of in metabolism — may allow us to use existing J. Med. 344, 783–792 (2001). genomic instability in metastatic pancreatic cancer. 11. Wood, L. D. et al. The genomic landscapes of human Nature 467, 1109–1113 (2010). drugs that are currently used to treat other breast and colorectal cancers. Science 318, 36. Yachida, S. et al. Distant metastasis occurs late during diseases for the treatment of cancer. 1108–1113 (2007). the genetic evolution of pancreatic cancer. Nature 12. Fanidi, A. et al. Cooperative interaction between c‑myc 467, 1114–1117 (2010). There is cause to hope that the use of and bcl‑2 proto-oncogenes. Nature 359, 554–555 37. Shackleton, M. et al. Heterogeneity in cancer: cancer rational combinations of targeted therapies (1992). stem cells versus clonal evolution. Cell 138, 822–829 13. Druker, B. J. et al. Efficacy and safety of a specific inhibitor (2009). will become routine in cancer care over the of the BCR-ABL tyrosine kinase in chronic myeloid 38. Tuveson, D. A. et al. STI571 inactivation of the next 10 years, bringing more effective and less leukemia. N. Engl. J. Med. 344, 1031–1037 (2001). gastrointestinal stromal tumor c‑KIT oncoprotein: 14. Collado, M. & Serrano, M. Senescence in tumours: biological and clinical implications. Oncogene 20, toxic treatment options. But maybe the best evidence from mice and humans. Nature Rev. Cancer 5054–5058 (2001). prediction is that new advances will be unpre- 10, 51–57(2010). 39. Solimini, N. L. et al. Non-oncogene addiction and the 15. Degenhardt, K. et al. Autophagy promotes tumor cell stress phenotype of cancer cells. Cell 130, 986–988 dictable, and in 10 years time we can again survival and restricts necrosis, inflammation, and (2007). look back and wonder at the unimaginable tumorigenesis. Cancer Cell 10, 51–64 (2006). 40. Cairns, R. A. et al. Regulation of cancer cell 16. Hatzivassiliou, G. et al. RAF inhibitors prime wild-type metabolism. Nature Rev. Cancer 11, 85–95 (2011). progress that has been made. RAF to activate the MAPK pathway and enhance 41. Bryant, H. E. et al. Specific killing of BRCA2‑deficient growth. Nature 464, 431–435 (2010). tumours with inhibitors of poly(ADP-ribose) Ya Cao is at the Cancer Research Institute of Xiangya 17. Heidorn, S. J. et al. Kinase-dead BRAF and oncogenic polymerase. Nature 434, 913–917 (2005). School of Medicine, Central South University, RAS cooperate to drive tumor progression through 42. Farmer, H. et al. Targeting the DNA repair defect in Changsha, Hunan 410078, China. CRAF. Cell 140, 209–221 (2010). BRCA mutant cells as a therapeutic strategy. Nature 18. Poulikakos, P. I. et al. RAF inhibitors transactivate RAF 434, 917–921 (2005). Ronald A. DePinho is at the University of Texas MD dimers and ERK signalling in cells with wild-type BRAF. 43. DeBerardinis, R. J., Sayed, N., Ditsworth, D. & Anderson Cancer Center 1400 Pressler, Nature 464, 427–430 (2010). Thompson, C. B. Brick by brick: metabolism and 19. Villa, L. L. HPV prophylactic vaccination: the first tumor cell growth. Curr. Opin. Genet. Dev. 18, 54–61 Unit 1491 Houston, Texas 77030, USA. years and what to expect from now. Cancer Lett. 305, (2008). 44. McAllister, S. S. & Weinberg, R. A. Tumor-host Matthias Ernst is at the Melbourne-Parkville branch of 106–112 (2011). 20. Human papillomaviruses. IARC Monogr. Eval Carcinog. interactions: a far-reaching relationship. J. Clin. Oncol. the Ludwig Institute for Cancer Research, 6th floor, Risks Hum. 64, 1–378 (1995). 28, 4022–4028 (2010). Centre for Medical Research, Royal Melbourne Hospital, 21. Garzon, R., Marcucci, G. & Croce, C. M. Targeting 45. Wilson, A. J. Inhibition of protein-protein interactions Royal Parade, Parkville, Victoria VIC 3050, Australia. microRNAs in cancer: rationale, strategies and using designed molecules. Chem. Soc. Rev. 38, challenges. Nature Rev. Drug Discov. 9, 775–789 3289–3300 (2009). Karen Vousden is at the Beatson Institute for Cancer (2010). Research, Garscube Estate, Switchback Road, 22. McDermott, U., Downing, J. R. & Stratton, M. R. Acknowledgements Bearsden, Glasgow G61 1BD, UK. Genomics and the continuum of cancer care. N. Engl. M.E. is a research fellow of the National Health and Medical J. Med. 364, 340–350 (2011). Research Council Australia. Y.C. thanks Z. Dong and A.M. Bode, e-mails: [email protected]; rdepinho@ 23. Beatson, G. On the treatment of inoperable cases of The Hormel Institute, University of Minnesota, USA. carinoma of the mamma: suggesitons for a new mdanderson.org; [email protected]; method of treatment with illustrative cases. Lancet Competing interests statement [email protected] 148, 104–107 (1896). The authors declare no competing financial interests. doi:10.1038/nrc3138 24. Vultur, A., Villanueva, J. & Herlyn, M. Targeting BRAF in advanced melanoma: a first step toward manageable FURTHER INFORMATION Published online 15 September 2011 disease. Clin. Cancer Res. 17, 1658–1663 (2011). Ya Cao’s homepage: http://www.csu.edu.cn/chinese/ academy/yanjiusuo/mirc/mirc.htm 1. Hanahan, D. & Weinberg, R. A. Hallmarks of 25. Yen, K. E., Bittinger, M. A., Su, S. M. & Fantin, V. R. Ronald A. DePinho’s homepage: http://www.mdanderson. cancer: the next generation. Cell 144, 646–674 Cancer-associated IDH mutations: biomarker and org/about-us/president-ronald-depinho-m-d-/index.html (2011). therapeutic opportunities. Oncogene 29, 6409–6417 2. Ngoma, T. World Health Organization cancer priorities (2010). Matthias Ernst’s homepage: http://www.ludwig.edu.au/ in developing countries. Ann. Oncol. 17, viii9–viii14 26. Brough, R., Frankum, J. R., Costa-Cabral, S., Lord, parkville/research/colon-biology.htm (2006). C. J. & Ashworth, A. Searching for synthetic lethality in Karen Vousden’s homepage: http://www.beatson.gla.ac.uk/ 3. Moore, P. S. & Chang, Y. Why do viruses cause cancer. Curr. Opin. Genet. Dev. 21, 34–41 (2011). Regulation-of-Cancer-Cell-Death-and-Survival/Karen- cancer? Highlights of the first century of human 27. Poulikakos, P. I. & Rosen, N. Mutant BRAF melanomas- Vousden-Tumour-Suppression.html Cancer Cell 19 tumour virology. Nature Rev. Cancer 10, 878–889 dependence and resistance. , 11–15 ALL LINKS ARE ACTIVE IN THE ONLINE PDF (2010). (2011).