Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
Nature Partner Journals is an online-only, open access series of journals, collaborating with preeminent scientists and global partners to publish high-quality open access research across the applied and basic science disciplines, as well as multidisciplinary fields.
Our editors invite you to access articles from our range of journals publishing research in cancer and oncology. Browse the collection, featuring content from:
Glioblastoma is an aggressive form of brain cancer and effective immunotherapeutics are limited, with treatment currently based on chemotherapy and radiotherapy. A recent phase II clinical trial tested a personalized, targeted dendritic cell-based immunotherapy but there was no observed improvement in patient survival or progression-free survival compared to standard-of-care therapy. Here, Carmen Visus and colleagues have used tumor tissue samples from glioblastoma patients involved in this trial and receiving immunotherapy. Using a combination of mass spectrometry-based proteomics, microRNA sequencing and RT-qPCR they identified factors associated with survival or poor prognosis. Proteomics associated poor prognosis with various proteins including focal adhesion kinase 2 (FAK2), whilst microRNAs, miR-216b, miR-216a, miR-708 and let-7i, were associated with longer survival. Focussing on one pathway, FAK2, they integrated the proteomic and microRNA datasets and saw a negative association with overall survival across all patients. To test this, they added an FAK inhibitor to glioblastoma cell lines, including cells isolated from trial patients, and observed inhibition of gliomaspheres in treated cells, providing insights into potential immunotherapy targets.
Patients with the most common form of breast cancer stand to benefit from taking a drug that blocks two cell cycle–regulating proteins in addition to hormonal therapy. In a phase III clinical trial, Stephen Johnston from the Royal Marsden NHS Foundation Trust in London, UK, and colleagues randomly gave 493 postmenopausal women with HR+/HER2− metastatic breast cancer a nonsteroidal aromatase inhibitor plus either a placebo or a drug called abemaciclib, a targeted inhibitor of CDK4 and CDK6. A planned interim analysis previously showed that abemaciclib was safe and effective. Johnson’s team now reports that abemaciclib nearly doubled the time women live without disease recurrence, from 15 months on placebo to 28 months on the drug. Additionally, responses were more common and lasted longer among women who received both abemaciclib and the hormonal therapy.
An artificial intelligence tool that analyzes the morphology of cell nuclei can help pathologists predict whether a breast cancer sample expresses the estrogen receptor (ER) or not. David Agus from the University of Southern California in Los Angeles, USA, and colleagues designed a machine-learning algorithm to correlate ER status — which is usually determined via immunohistochemistry assays — with visual patterns of shape, orientation and other nuclear features that a pathologist normally sees on a stained biopsy specimen. The researchers trained the algorithm on samples taken from 57 women with untreated invasive ductal carcinoma. They then tested the model’s accuracy on a separate set of 56 patient samples. The algorithm could predict ER status with reasonable precision and accuracy, suggesting that, with improvements, it could form the basis of a diagnostic aid for guiding treatment decisions.
Artificial intelligence can accurately predict the histological subtype and molecular marker status of breast tumors from pathology slide pictures. Heather Couture from the University of North Carolina at Chapel Hill and coworkers developed a deep learning algorithm to analyze various molecular features found within pathology slide images from breast cancer tissue samples. They trained the algorithm on a set of 571 tumors with defined grade, subtype and hormone receptor status. They then tested the model on another set of 288 breast tumors, finding that the deep-learning method predicted assorted molecular features with an accuracy exceeding 75%. Artificial intelligence applied to breast tumor images in this way could thus help clinicians identify those patients who stand to benefit most from further RNA-based genomic testing, an expensive but more definitive diagnostic tool for disease classification.
A series of clinical factors may help identify those patients with hormone receptor-positive breast cancer who are most likely to benefit from a newly approved cell cycle–blocking drug. Angelo Di Leo from the Hospital of Prato, Italy, and colleagues retrospectively analyzed data from two phase III trials that evaluated the CDK4 & 6 inhibitor abemaciclib against a placebo administered in combination with either fulvestrant endocrine therapy or an aromatase inhibitor. Pooling the results, the researchers found that patients with certain indicators of poor prognosis—including liver metastases, progesterone receptor negativity, high-grade tumors, or short treatment-free intervals following the completion of adjuvant endocrine therapy—benefited most from abemaciclib, as measured by progression-free survival. The findings, if confirmed in prospective trials, could help inform personalized drug choices for patients.
Triple negative breast cancer (TNBC) is a heterogeneous disease that easily develops drug resistance. To achieve more effective clinical responses, synergistic drug combinations that inhibit multiple survival pathways of cancer are urgently needed. However, pinpointing these drug combinations is difficult, as the number of possible combinations grows exponentially. Tang and co-workers from the University of Helsinki, University of Copenhagen, and University of Heidelberg developed a network pharmacology modeling approach to predict synergistic drug combinations and their underlying target interactions. With dynamic simulation of signaling pathways, they identified a synergistic target interaction that involved Aurora B and ZAK that play a critical role in regulating the survival of TNBC cells. These new combinatorial drug targets warrant further exploration of clinical benefits in treating TNBC.
The identification of robust and reproducible molecular markers is one of the biggest challenges in personalised cancer medicine. The increasing use of systems biology approaches has prompted researchers to integrate heterogeneous data into existing knowledge bases whose incorporation into the biomarker discovery workflow may adjust for data heterogeneity and limitation, and offer more precise, robust and consistent biomarkers. In this study, we have sought to determine network-based miRNA biomarker signatures from the plasma of colorectal cancer patients that hold prognostic utility. We performed miRNA profiling and then constructed an miRNA-mediated gene regulatory network and developed a multi-objective optimisation-based computational framework to identify miRNA biomarkers using both the miRNA expression profile and knowledge from this miRNA-mediated regulatory network. We have demonstrated the ability of the proposed approach in identifying robust, accurate and reproducible biomarkers.
Many studies have found a link between gut microbes and bowel cancer, the third most common cancer worldwide. The details of the association, however, have remained elusive. Researchers in the USA and Malaysia, led by Dr. Cynthia Sears at John Hopkins School of Medicine in Maryland, examined mucosal biofilm status by fluorescence microscopy and performed a meta-analysis of bacterial genetic associations in stool and colon tissues to clarify the connection. They found that bowel cancers were enriched in invasive bacterial biofilms as well as several specific gut and oral species, including one - Fusobacterium nucleatum - known to promote tumorigenesis in mouse models. Analyzing gut microbial populations might help assess bowel cancer risk. Further research is needed, however, to determine if these bacteria directly contribute to disease causality.
Differences in the microbial populations in the gut may help predict the likelihood of adverse reactions to a drug used to treat bowel cancer. Libusha Kelly, Leah Guthrie, and colleagues at Albert Einstein College of Medicine in New York examined the undesirable reactivation of the chemotherapy drug irinotecan by microbial enzymes in the gut. They identified an association between specific forms of microbial metabolic activity and drug metabolism. Sampling the microbial population of a patient’s gut may therefore offer a relatively non-invasive way to identify biomarkers predicting the likelihood of adverse reactions due to microbial metabolism. The research also suggests that using drugs to inhibit the activity of specific microbial enzymes in the gut might improve the outcome of some treatments. Modifying the microbial population prior to treatment may be another option.
A diagnostic test for short regulatory RNA molecules contained within tiny secreted vesicles in the bloodstream can accurately pick up signs of glioblastoma brain cancer. Researchers in Australia led by Michael Buckland and Kim Kaufman from the Royal Prince Alfred Hospital and the University of Sydney isolated circulating vesicles, called exosomes, from patients with glioblastoma or lower-grade brain cancers known as gliomas as well as healthy controls. Next-generation sequencing revealed a panel of 26 microRNAs contained within the exosomes that were differentially expressed in glioblastoma samples relative to healthy controls. (A different but partially overlapping set of 23 microRNAs also helped distinguish patients with a mutant subtype of glioma.) The researchers narrowed down the list to the seven microRNAs with the most predictive power. Testing for just these microRNAs reliably diagnosed glioblastoma with greater precision than previously reported panels of “free-circulating” microRNAs.
Examining events leading to the diagnosis of non-small cell lung cancer (NSCLC) in Australia yields insights to guide further research and perhaps improve the pathways to diagnosis. NSCLC is by far the most common form of lung cancer. Researchers, led by the Cancer Institute New South Wales, investigated clinical care contacts leading to diagnosis, using a descriptive cohort study of 894 patients diagnosed between 2006 and 2012. The researchers quantified contact with GPs and lung specialists, hospital admissions and diagnostic imaging procedures. Living in remote locations had little influence on access to services. More than half of the patients did not see a lung specialist during the pathway to diagnosis, while a quarter received their diagnosis in an emergency setting. Further research should investigate whether there are barriers preventing timely diagnosis.
Landscape reveals the underlying mechanisms of the interplay between EMT and cancer metastasis. Professor Chunhe Li at Fudan University and his collaborators constructed an EMT-metastasis gene regulatory network model and quantified the potential landscape of cancer metastasis system. Four attractors emerge on the landscape, characterizing antimetastatic (A), metastatic (M), and two intermediate cell states, respectively. For metastatic progression, cells need to first undergo EMT, and then become metastatic. This suggests that temporal order is critical for the activation of cellular programs in biological systems, and provides a possible mechanism of EMT-promoting cancer metastasis. By optimizing the transition actions between the A and M attractors, the authors identified some optimal combinations of regulatory targets for preventing cancer progression. Combining the landscape with transition action optimization provides a general way to investigate stochastic dynamics of gene regulatory systems.
The main cause of disease worldwide is aging, a process long suspected to result from damage accumulation. However, new work from a team led by David Gems at University College London describes a different type of cause: quasi-programs, or the futile run-on biological programs in later life, supporting a recently proposed, alternative theory about aging. The study focuses on the short-lived worm C. elegans, which during aging develop very large tumors in the uterus. The authors show how such tumors develop due to futile attempts by unfertilized eggs to develop by switching on programs of embryogenesis. In mammals, similar runaway programs cause teratomas, a type of benign tumor of grotesque appearance. This work provides new insights about the nature of aging as a disease process, suggesting that it is to an extent teratoma-like.
Around half of all tumor-derived DNA strands found in the bloodstream of cancer patients are too short for detection by the most commonly used diagnostic sequencing methods. Alain R. Thierry from the Montpellier Cancer Research Institute, France, and colleagues quantified the size distribution of circulating cell-free DNA (cfDNA) using two genomic protocols that analyze single-stranded DNA—and can thus pick up small DNA fragments that are missed by conventional double-stranded approaches. Looking in the blood of patients with tumors of the colon, lung, breast and liver, the researchers showed that nearly half of all stretches of cfDNA are shorter than 120 nucleotides long. These sequences likely result from tightly packed DNA that’s dynamically degraded into smaller and smaller fragments. The findings highlight the need for whole-genome sequencing or other single-stranded methods to get an accurate read on cfDNA profiles.
Estimating patients’ mutation load from a small set of genes can accurately predict their response to cancer immunotherapy. Harnessing patients’ immune response to target tumor cells is an effective treatment approach in some cases but not others. A patient’s number of deleterious genetic mutations across all their protein-coding genes has been shown to correlate with their responsiveness to immunotherapy. However, whole-exome sequencing is time-consuming and costly. Yu-Chao Wang at the National Yang-Ming University, Taiwan, and colleagues have developed cancer-specific mutation load estimation models for adenocarcinoma, melanoma and colorectal cancer that require sequencing only a small number of genes. They show that the mutation load in lung adenocarcinoma patients can be estimated from 24 genes and that they can predict immunotherapy responsiveness with similar accuracy to that obtained using whole-exome sequencing.
The loss of entire chromosomes seems to be a fundamental driver of tumors arising from the hormone-producing cells of the pancreas. A team led by Cristin Print and Ben Lawrence from the University of Auckland, New Zealand, performed genomic and pathological analysis of 57 pancreatic neuroendocrine tumors, a rare form of cancer caused by the abnormal growth of hormone-producing islet cells within the pancreas. The researchers observed two distinct patterns of chromosome loss, with 26% of the samples missing one copy of 10 specific chromosomes and another 40% lacking a copy of chromosome 11. In both groups, the abnormal chromosome count prompts abnormal gene activity patterns, with recessive mutations unleashed and expressed unopposed. Single gene mutations seem to play only a minor role, suggesting that single gene-targeted drugs will provide little benefit in this disease setting, with more nuanced approaches required.
Wearable activity monitors provide an objective and continuous measure of general wellbeing and physical function among patients with advanced cancer. Gillian Gresham from Cedars-Sinai Medical Center in Los Angeles, California, USA, and colleagues enlisted 37 patients with metastatic or inoperable cancer to wear wristbands that wireless measure heart rate and activity patterns over the span of 2 weeks. At weekly clinic visits, the patients also filled in health questionnaires and their healthcare providers made independent evaluations. The researchers found a reasonably high correlation between the clinician-assessed performance status and the average daily step total calculated by the wristband. More steps also translated into lower odds of serious complications, hospitalizations or death. The study shows the feasibility of using wearable technology to track the ability of patients with advanced cancer to perform certain activities of daily living.
A sensing protein that is increased in response to DNA damage can be targeted to reduce inflammation and collateral damage during anti-cancer therapy and aging. Scientists at Saini Lab at the Indian Institute of Science have identified the protein that drives sustained and detrimental inflammation when the DNA of cells are damaged, such as during normal human aging or during anti-cancer therapy. Furthermore, blocking the functions of this protein and associated pathway was able to reduce the inflammation to less harmful levels. This discovery could potentially enable safer and more effective anti-cancer therapy by protecting non-cancerous cells surrounding tumors from lethal inflammation. Further studies on this protein could also reduce age associated inflammation, allowing us to age gracefully and healthily.
To control cell behavior in physiology and disease, genes are expressed in co-ordinated networks, but the sheer number of connections between genes challenges interpretation of this structure. Care, Westhead and Tooze—University of Leeds, UK—show how retaining only the 3 nearest neighbors of any gene improves information content of expression networks. This new approach, PGCNA, outperforms alternatives and can be used to integrate multiple sets of data. With breast and colon cancer as examples the team demonstrate how such networks can be used to derive bar-codes that summarize cancer gene expression and link to underlying gene mutation states. This has potential application in precision medicine and establishes a scalable, efficient, approach to enhance network analysis.
Existing drugs have the potential to learn “new tricks” to fight cancer when working together in combinations. Researchers from The Ohio State University and Washington University in St. Louis developed a computational approach to predict synergistic drug combinations among thousands of potential pairs based upon how cancer cells alter certain genes and connections between genes. They found that drug predictions were more accurate when this method was applied to specific molecular subtypes of melanoma. The top drug combination predicted for melanoma with BRAF gene mutations included a known melanoma drug, vemurafenib, in combination with another agent not currently used for melanoma, tretinoin. This combination could kill melanoma cells and reduce tumor size in mice. Future applications include other examples of “precision medicine” to deliver the right drugs to the right patients at the right time.