Search This Blog

Monday, July 4, 2022

How antibiotics may shut down cancer metastasis

 Metastasis makes cancer hard to treat. As tumor cells migrate and colonize at new sites, they rely on energy-producing mitochondria to adapt to the changing microenvironment. But the exact mechanisms of how these cellular power plants support the invasive spread of tumors remain unclear.

Now, an international team of scientists led by the German Cancer Research Center (DKFZ) has fingered certain modifications in mitochondrial RNA as culprits behind cancer metastasis.

Interestingly, certain available antibiotics are able to inhibit problematic mitochondrial mRNA’s translation into proteins. Treatment with the antibiotics reduced cancer metastases in mice, the team reported in a new study published in Nature.

As cancer cells invade into surrounding tissue and travel to distant organs, they need to navigate adverse conditions such as shortages of oxygen or nutrients. To survive, cancer cells adjust their energy production pattern through mitochondria.

In the current study, the DKFZ-led team found that a specific chemical modification process found in mitochondrial tRNA, known as mitochondrial cytosine-5 RNA methylation (m5C), drives the translation of mitochondrial mRNA to support metastasis. Modifications of tRNAs ensure the accurate and optimal rate of protein synthesis in mitochondria.

In contrast, cancer cells that lack m5C instead turn to a relatively less efficient energy production mechanism called glycolysis, the team found.

An enzyme called NSUN3 is required for the m5C RNA modification. Depleting NSUN3 led to a reduction of m5C in mitochondrial tRNA—and hence mitochondrial activity. In mice bearing head and neck cancer, NSUN3-deficient cancer cells showed decreased metastases into both the lymph nodes and lungs.

An analysis of The Cancer Genome Atlas also showed that in patients with head and neck cancer, higher levels of NSUN3 were correlated with genes that encode regulators of metastasis and response to low oxygen levels. The NSUN3 levels were also predictive of disease stage and the presence of lymph node metastasis at the time of diagnosis, the team found.

Because mitochondrial and bacterial protein synthesis machineries are highly similar, the scientists proposed that certain antibiotics might be able to curb mitochondria activity without affecting other critical protein productions in human cells.

Indeed, treating head and neck cancer cells with the antibiotics tigecycline and doxycycline reduced the number of invading leader cells in tumor-like organoids. In mice, treatment with the two antibiotics decreased the number of lymph node metastases from 80% to 20%, the team reported. 

In contrast, the antibiotic amoxicillin, which target the bacterial cell wall, didn’t have any effect on mitochondrial translation or cancer metastasis. 

Certain antibiotics are already used as adjuvant therapy in the treatment of cancers, but their mode of action and effect on cancer therapies is controversial, the DKFZ-led team noted. The current findings showed that different classes of antibiotics could have distinct effects on tumors, the team added.

Once the tumor metastasizes, the cancer becomes difficult to treat, and metastasis is the major cause of cancer deaths. So finding ways to stop the dissemination of cancer cells from primary tumors has been an important research area.

In a 2021 Science study, scientists from the Massachusetts Institute of Technology, the University of California, Berkeley and the University of California, San Francisco used the CRISPR gene-editing tool to map out the development route of tumor metastasis. They uncovered genes that either contributed to or suppressed the cells’ metastatic ability, as well as physicial metastasis “hubs” they believe could be targeted with new cancer therapies. One gene called KRT17, for example, was most strongly correlated with reduced metastatic ability.

The DKFZ-led team now suggests that the inhibition of mitochondrial RNA modifications could be a promising approach to stop the spread of cancer cells.

“We propose that mitochondrial RNA-modifying enzymes should be added to the growing list of RNA-modifying anticancer drug targets,” the scientists wrote in the study. “NSUN3 is a highly promising drug target because as a stand-alone enzyme, it is solely responsible for mitochondrial m5C formation.”

https://www.fiercebiotech.com/research/how-antibiotics-may-shut-down-cancer-metastasis

U.S. drug distributors prevail in $2.5 billion West Virginia opioid case

 McKesson Corp, AmerisourceBergen Corp and Cardinal Health Inc are not responsible for fueling an opioid epidemic in a part of West Virginia, a federal judge ruled on Monday.

The decision from U.S. District Judge David Faber came in a $2.5 billion case brought by the city of Huntington and Cabell County, which at trial sought to show the three largest U.S. drug distributors caused a surge in opioid prescriptions in their communities.

But in a long-awaited, 184-page ruling, Faber said the volume of prescription painkillers shipped to those communities was due to doctors' "good faith" prescribing decisions and that the companies did not cause any oversupply of opioids.

"The opioid crisis has taken a considerable toll on the citizens of Cabell County and the city of Huntington," he wrote. "And while there is a natural tendency to assign blame in such cases, they must be decided not based on sympathy, but on the facts and the law."

Cardinal Health in a statement applauded the decision, which it said recognized that it only provided a "secure channel to deliver medications of all kinds." McKesson said it maintains strong programs to prevent the diversion of opioids to illicit channels.

Lawyers for the plaintiffs had no immediate comment.

More than 3,300 lawsuits have been filed by local and tribal governments over the opioid abuse and overdose epidemic. They accuse drugmakers of downplaying the risks of the addictive pain medicines and distributors and pharmacies of ignoring red flags that they were being diverted into illegal channels.

U.S. officials have said that by 2019, the health crisis led to nearly 500,000 opioid overdose deaths over two decades.

The distributors, along with drugmaker Johnson & Johnson, in July agreed to pay up to $26 billion to resolve the thousands of lawsuits brought against them by state and local governments around the country. [nL1N2V01ZA

But communities in hard-hit West Virginia opted against joining a national opioid settlement in favor of seeking a bigger recovery.

https://www.yahoo.com/news/u-drug-distributors-prevail-2-221813127.html

How the coronavirus attacks the heart

 This occurs mainly in patients with comorbidities such as obesity, diabetes and hypertension. The International Journal of Cardiology from 26 May 2022 published an article on these findings.

Virus detected in heart cells

In order to track down the new entry mechanism, the research team at the university hospital used histochemical methods and microscopy to analyse heart tissue structures from patients suffering from Covid-19 and those who died from or with the disease. In a first step, they provided evidence that the virus can indeed be detected directly in the cells of the heart muscle. “Our observations show that the virus exerts pressure on the heart muscle, attacks and weakens the contractile force, i.e. the pumping function of the heart,” says Hamdani.

But how does the virus enter the heart? What mechanisms facilitate the penetration of the virus into the heart muscle cells? The Bochum team showed that one possible mechanism of cardiac muscle cell dysfunction in Sars-Cov-2 patients is the activation of certain enzymes that degrade proteins. In fact, the team detected an increased so-called proteolytic activity.

This suggests that Sars-Cov-2 enters cells as a result of the activation of the spike protein by enzymes responsible for the degradation of proteins, and that its entry into cells depends on these degradation enzymes. In addition, Hamdani’s group investigated proteins that are responsible for apoptosis, i.e. cellular suicide. The team showed that while the apoptotic proteins had increased activity, their expression was drastically reduced. “This indicates that the proteins are cleaved and apoptosis is activated,” explains Nazha Hamdani. “The results imply that apoptosis contributes to the deterioration in cardiac contractility observed in Sars-Cov-2 patients.”

The key role of inflammation and oxidative stress

In the next step, the team set out to explore what promotes the increased proteolytic activity and apoptosis of cardiac myocytes. The study showed that oxidative stress and a pro-inflammatory environment exacerbate the damage associated with Sars-Cov-2. The focus here was on the so-called neutrophils. Neutrophils are one of the primary cell types that release proteolytic enzymes. They play an essential role during an inflammatory response. They are rapidly mobilised from the bloodstream into the damaged tissue. Since proteolytic enzymes are released more frequently in Sars-Cov-2 patients, Hamdani’s team analysed the signalling pathway, more specifically the interleukin-6-driven neutrophil traffic. The researchers found that inflammatory signalling pathways in cardiac myocytes were highly regulated – i.e. interleukin-6 was highly elevated – suggesting a key role for these white blood cells in Covid-19 and associated inflammatory pathologies.

Alternative gateways

Furthermore, the Bochum-based researchers have backed the existing findings that the virus also uses the protein neuropilin-1 (NRP-1) as a gateway into the cells. Hamdani’s research shows that the coronavirus thus has several mechanisms at its disposal to spread in human organs. “Sars-Cov-2 is able to spread in the infected heart in a receptor-dependent and receptor-independent manner. We also examined another mechanism by which the virus can gain access to the heart muscle cells, thus contributing to endothelial dysfunction. We will soon be able to publish these results,” concludes Hamdani.

Variants BA.2.12.1, BA.4 and BA.5 are inhibited less efficiently by antibodies

 Emerging Omicron subvariants BA.2.12.1, BA.4 and BA.5 are inhibited less efficiently by antibodies

Infections with the "old" omicron subvariants BA.1 and BA.2 provide little protection against the SARS-CoV-2 subvariant BA.5 causing a “summer wave” of cases in Germany

The Omicron subvariants BA.1 and BA.2 of SARS-CoV-2 have dominated the COVID-19 pandemic in early 2022. In many countries, these viruses are now outcompeted by emerging subvariants, with BA.5 being responsible for the current uptick of cases in Germany. However, it is at present largely unclear whether the “new” Omicron subvariants BA.2.12.1, BA.4, and BA.5 acquired biological traits that allow for more efficient transmission or whether they are less efficiently blocked by antibodies compared to the “old” Omicron subvariants BA.1 and BA.2. A study by researchers at the German Primate Center (DPZ) - Leibniz Institute for Primate Research together with colleagues from Hannover Medical School and Friedrich-Alexander-University Erlangen-Nürnberg shows that most of the therapeutic antibodies available for treatment of COVID-19 patients do not inhibit BA.2.12.1, BA.4, and BA.5 at all or only inhibit with reduced potency. The antibody Bebtelovimab constitutes the only exception since this antibody blocked all tested variants with high efficiency. Furthermore, the study shows that the Omicron subvariants BA.2.12.1 and especially BA.4 and BA.5 are inhibited worse than their predecessors BA.1 and BA.2 by antibodies generated after vaccination or inoculation followed by infection. Thus, BA.2.12.1, BA.4, and BA.5 are immune escape variants. A pass-through infection with "old" Omicron subvariants confers only limited protection against infection with "new" subvariants (The Lancet Infectious Diseases).

New SARS-CoV-2 variants emerge because of errors during viral genome replication. Thus, the virus acquires mutations that change the viral proteins, including the surface protein, spike, the central target of the antibody response. In case the mutations reduce recognition of the spike protein by antibodies, such variants become more adept at spreading among people with preexisting immunity due to vaccination or past infection.

Infection researchers at the German Primate Center are specialized in the analyses of SARS-CoV-2 inhibition by antibodies. Jointly with colleagues at Hannover Medical School and Friedrich-Alexander-University Erlangen-Nürnberg they have investigated inhibition of the SARS-CoV-2 Omicron subvariants BA.2.12.1, BA.4, and BA.5 by antibodies. BA.2.12.1, BA.4, and BA.5 (the spike protein of the latter two subvariants is identical) are becoming dominant in several countries and BA.5 is largely responsible for the recent uptick of cases in Germany.

The team, including Prerna Arora, Markus Hoffmann and Stefan Pöhlmann at the German Primate Center, found that out of ten therapeutic antibodies studied only two were able to at least partially inhibit BA.2.12.1, BA.4, and BA.5 and that only one antibody, Bebtelovimab (LY-CoV1404), efficiently blocked infection by all Omicron subvariants. “These results confirm a trend that we have already seen in previous studies: Omicron subvariants are not appreciably inhibited by most therapeutic antibodies and the few antibodies that inhibit frequently do so in a subvariant-specific fashion. Therefore, it is important to develop new antibodies in order be prepared for future subvariants,“ says Prerna Arora, first author of the study.

Antibodies from unvaccinated individuals that were infected with Omicron subvariants BA.1 or BA.2 in spring 2022 neutralized BA.2.12.1 with similar efficiency but were much less potent against BA.4 and BA.5. Therefore, it is likely that a previous BA.1 or BA.2 infection provides little protection against a subsequent infection with BA.4 or BA.5. Antibodies induced by three immunizations with the mRNA vaccine of BioNTech/Pfizer blocked all Omicron subvariants. However, inhibition was less efficient as compared to that measured for a virus that circulated early during the pandemic, and inhibition of BA.2.12.1, BA.4, and BA.5 was less efficient as compared to BA.1 and BA.2. Similar results were obtained for antibodies induced upon vaccination plus breakthrough infection. Although this so-called hybrid immunity conferred overall higher neutralizing activity against all variants tested, inhibition of BA.2.12.1, BA.4 and BA.5 was significantly reduced.

“BA.2.12.1 and particularly BA.4 and BA.5 are antibody evasion variants. Vaccination will still protect against severe disease induced by these variants but protection might be somewhat less efficient as that measured for previously circulating variants,“ concluded Markus Hoffmann, senior author of the study. “Our future studies must show whether BA.2.12.1, BA.4 and BA.5 are not only less efficiently inhibited by antibodies but are also better at infecting lung cells. If this is the case, then an uptick in hospitalizations might be the consequence, although it should be stated that this has so far not been observed in South Africa, where BA.4 and BA.5 were first detected,“ comments Stefan Pöhlmann, who headed the study jointly with Markus Hoffmann.

 

The German Primate Center GmbH (DPZ) - Leibniz Institute for Primate Research conducts biological and biomedical research on and with primates in the fields of infection research, neuroscience and primate biology. The DPZ also maintains four field stations in the tropics and is a reference and service center for all aspects of primate research. The DPZ is one of the 97 research and infrastructure facilities of the Leibniz Association.

Signal pathway in the brain that controls food intake discovered

 A group of researchers has developed an entirely novel approach to treating eating disorders. The scientists showed that a group of nerve cells in the hypothalamus (so-called AgRP, agouti-related peptide neurons) control the release of endogenous lysophospholipids, which in turn control the excitability of nerve cells in the cerebral cortex, which stimulates food intake. In this process, the crucial step of the signalling pathway is controlled by the enzyme autotaxin, which is responsible for the production of lysophosphatidic acid (LPA) in the brain as a modulator of network activity. The administration of autotaxin inhibitors can thereby significantly reduce both excessive food intake after fasting and obesity in animal models. The article 'AgRP neurons control food intake behaviour at cortical synapses via peripherally-derived lysophospholipids' has now appeared in Nature Metabolism.

Eating disorders and especially obesity are one of the most common causes of a variety of diseases in industrialized societies worldwide, especially cardiovascular diseases with permanent disabilities or fatal outcomes such as heart attacks, diabetes, or strokes. The Robert Koch Institute reported in 2021 that 67 per cent of men and 53 percent of women in Germany are overweight. 23 per cent of adults are severely overweight (obese). Attempts to influence eating behaviour with medication have so far proved ineffective. A novel therapy that modulates the excitability of networks that control eating behaviour would be a decisive step towards controlling this widespread obesity.

The research team found an increased rate of obesity and the attendant type II diabetes in people with impaired synaptic LPA signalling. A group led by Professor Johannes Vogt (Faculty of Medicine, University of Cologne), Professor Robert Nitsch (Faculty of Medicine, University of Münster) und Professor Thomas Horvath (Yale School of Medicine, New Haven, USA) has now shown that control of the excitability of neurons in the cerebral cortex by LPA plays an essential role in the control of eating behaviour: AgRP neurons regulate the amount of lysophosphatidylcholine (LPC) in the blood. Through active transport, LPC reaches the brain, where it is converted by the enzyme autotaxin (ATX) into LPA, which is active at the synapse. Synaptic LPA signals stimulate specific networks in the brain, thus leading to increased food intake.

In the mouse model, after a period of fasting an increase in LPC in the blood led to an increase in stimulating LPA in the brain. These mice showed typical food-seeking behaviour. Both could be normalized by administrating autotaxin inhibitors. Obese mice, on the other hand, lost weight when these inhibitors were administered continuously. Johannes Vogt explained: 'We saw a significant reduction in excessive food intake and obesity through gene mutation and pharmacological inhibition of ATX. Our fundamental findings on the LPA-controlled excitability of the brain, which we have worked on for years, therefore also play a central role for eating behaviour.' Robert Nitsch sees the findings as an important step towards new drug development: 'The data show that people with a disturbed synaptic LPA signalling pathway are more likely to be overweight and suffer from type II diabetes. This is a strong indication of a possible therapeutic success of ATX inhibitors, which we are currently developing together with the Hans Knöll Institute in Jena for use in humans.'

These findings on the excitation control of neuronal networks in eating behaviour through lysophospholipids and the new therapeutic possibilities they suggest could in future contribute not only to treating eating disorders, but also neurological and psychiatric illnesses.


Story Source:

Materials provided by University of CologneNote: Content may be edited for style and length.


Journal Reference:

  1. Heiko Endle, Guilherme Horta, Bernardo Stutz, Muthuraman Muthuraman, Irmgard Tegeder, Yannick Schreiber, Isabel Faria Snodgrass, Robert Gurke, Zhong-Wu Liu, Matija Sestan-Pesa, Konstantin Radyushkin, Nora Streu, Wei Fan, Jan Baumgart, Yan Li, Florian Kloss, Sergiu Groppa, Nils Opel, Udo Dannlowski, Hans J. Grabe, Frauke Zipp, Bence Rácz, Tamas L. Horvath, Robert Nitsch, Johannes Vogt. AgRP neurons control feeding behaviour at cortical synapses via peripherally derived lysophospholipidsNature Metabolism, 2022; 4 (6): 683 DOI: 10.1038/s42255-022-00589-7

Off-shelf glucose monitors prove accurate for dialysis patients

 In what is believed to be the first study of its kind, new UVA Health research reveals that a factory-calibrated continuous glucose monitor (CGM) may be sufficiently accurate for use by people on dialysis, a group often plagued by dangerous swings in blood-sugar levels.

The findings suggest that factory-calibrated blood glucose monitors could offer an important diabetes-management tool for patients on dialysis and those suffering end-stage renal disease [ESRD], the researchers conclude.

"Patients with end-stage renal disease are often excluded from clinical research trials, as they are medically complex. Therefore, these CGM devices -- often considered 'game changers' for patients with diabetes to monitor their sugars -- are not yet FDA approved for patients with ESRD on dialysis," said researcher Meaghan M. Stumpf, MD, an expert on diabetes and diabetes-management technology at UVA Health. "However, ESRD patients and their physicians may still benefit from their use. Our research team conducted this pilot study so that we could begin to understand the accuracy of these devices for patients with ESRD on hemodialysis. This study is not large enough to lead to FDA approval, but it is important to take the first step."

Continuous glucose monitors are becoming widely used by patients with diabetes. These devices allow patients to track their blood-sugar (glucose) levels automatically, helping them prevent their blood-sugar levels from getting dangerously high or low.

Managing blood-sugar levels is a particular challenge for patients on dialysis, which is a procedure to filter blood for patients whose kidneys can no longer do so adequately. Patients on dialysis often suffer from "hypoglycemia," or low blood sugar, which is potentially deadly. That means that these patients need highly reliable, accurate ways to track their blood sugar. Until now, physicians and researchers did not have accuracy data for factory-calibrated continuous glucose monitors, so it was unclear if these devices would be up to the job.

To determine this, Stumpf and her colleagues enlisted 20 volunteers receiving hemodialysis at UVA to test such a device, the Dexcom G6-Pro. Most of the participants were male, African-American, and on insulin, with an average age of 61. The participants were asked to wear the CGM for 10 days and to take four to seven fingerstick blood-sugar readings per day with a home glucometer. Venous blood samples were also collected during their hemodialysis sessions. The researchers compared the CGM glucose results with the blood-sugar results collected by the patients and during the patients' thrice-weekly dialysis sessions.

The researchers determined that the continuous glucose monitor, overall, showed "clinical reliability," meaning that they were sufficiently accurate for estimating blood-sugar levels. Almost 99% of the readings were accurate enough to be used without confirmatory fingerstick blood sugar readings.

When the devices erred, they tended to overestimate, rather than underestimate, blood-sugar levels, prompting the researchers to note that additional research is warranted, especially considering that people on dialysis tend to be at elevated risk for low blood sugar.

"Although we certainly need larger studies, I am encouraged that these factory- calibrated continuous glucose monitors may be reasonably accurate for patients on hemodialysis therapy," Stumpf said. "CGM use for these patients could lead to improved glucose control, improved safety from life-threatening hypoglycemia and, very importantly, improved quality of life."


Story Source:

Materials provided by University of Virginia Health SystemNote: Content may be edited for style and length.


Journal Reference:

  1. Orianne Villard, Marc D. Breton, Swati Rao, Mary K. Voelmle, Morgan R. Fuller, Helen E. Myers, Ryan K. McFadden, Zander S. Luke, Christian A. Wakeman, Mary Clancy-Oliveri, Ananda Basu, Meaghan M. Stumpf. Accuracy of a Factory-Calibrated Continuous Glucose Monitor in Individuals With Diabetes on HemodialysisDiabetes Care, 2022; DOI: 10.2337/dc22-0073

Mechanism controlling spread of pancreatic cancer discovered

 Scientists have shown it is possible to reverse a key process that allows pancreatic cancer cells to grow and spread around the body.

These findings, published in Nature,show that a protein called GREM1 is key to regulating the type of cells found in pancreatic cancer -- and manipulating its levels can both fuel and reverse the ability of these cells to change into a more aggressive subtype.

The researchers believe this fundamental discovery could ultimately pave the way for new pancreatic cancer treatments.

Researchers from The Institute of Cancer Research, London, studied pancreatic cancer with the gene that makes the GREM1 protein switched off in mice, and in pancreatic 'mini-tumours', which are also known as organoids.

Switching off GREM1 caused the tumour cells to rapidly change shape and develop new properties that help them invade new tissues and migrate around the body. Within just 10 days, all the tumour cells changed their identity into a dangerous, invasive cell type.

Switching off the gene also made tumours in mice more likely to spread. The researchers studied a mouse model of pancreatic ductal adenocarcinoma (PDAC) -- the most common and aggressive form of the disease. Around 90 per cent of mice without functioning GREM1 developed tumours which had spread to their liver, compared to 15 per cent of mice where GREM1 was working normally.

Crucially, the scientists, who were largely funded by The Institute of Cancer Research (ICR), which as well as being a research institute is also a charity, then showed that boosting GREM1 levels could reverse this process and cause invasive cell types to revert into a less dangerous form. Researchers hope, in the future, to use this knowledge to find ways to reverse more advanced pancreatic cancer into a less aggressive form, which is easier to treat.

The researchers, who work in the Breast Cancer Now Toby Robins Research Centre at the ICR, stress that the science is early stage, and significant amounts of research would be required to discover and develop treatments that change PDAC cell fates and make the tumour respond better to therapies. However, fundamental discoveries such as this are crucial in directing efforts to find new cancer drugs and treatments.

Pancreatic cancer has the lowest survival rates of common cancers. Less than seven per cent of people will survive for five years or more. More than 10,000 people are diagnosed with pancreatic cancer in the UK each year, and more than 9,000 will die from it.

The researchers also discovered that another protein, called BMP2, is involved in regulating GREM1, and that these two proteins regulate the form PDAC cells ultimately take, according to a mathematical model first proposed by Alan Turing in 1952. These 'Turing patterns' are found in nature -- from the patterns on the skin of the giant puffer fish to seashells -- and strikingly the same sort of patterns are seen in the different types of cells found in pancreatic cancer. Further studies are needed to determine whether this model is also applicable in other forms of cancer.

Professor Axel Behrens, Leader of the Cancer Stem Cell Team at The Institute of Cancer Research, London, and senior author of the study, said:

"This is an important and fundamental discovery that opens up a new avenue for uncovering treatments for pancreatic cancer. We have shown that it is possible to reverse cell fate in pancreatic cancer in the lab -- turning back the clock on aggressive tumours and switching them to a state that makes them easier to treat.

"By better understanding what drives the aggressive spread of pancreatic cancer, we hope to now exploit this knowledge and identify ways to make pancreatic cancer less aggressive, and more treatable."

Professor Kristian Helin, Chief Executive of The Institute of Cancer Research, London, said:

"Pancreatic cancer is one of the most devastating of all cancers -- the most common form of the disease spreads aggressively, making it hard-to-treat and a terrifying diagnosis for patients and their loved ones.

"This new finding has broadened our understanding of the molecular basis of how pancreatic cancer gains the ability to grow and spread around the body. Although more work is required, this type of fundamental research is essential for developing concepts for new and more effective treatments for cancer."


Story Source:

Materials provided by Institute of Cancer ResearchNote: Content may be edited for style and length.


Journal Reference:

  1. Linxiang Lan, Theodore Evan, Huafu Li, Aasia Hussain, E. Josue Ruiz, May Zaw Thin, Rute M. M. Ferreira, Hari Ps, Eva M. Riising, Yoh Zen, Jorge Almagro, Kevin W. Ng, Pablo Soro-Barrio, Jessica Nelson, Gabriela Koifman, Joana Carvalho, Emma L. Nye, Yulong He, Changhua Zhang, Anguraj Sadanandam, Axel Behrens. GREM1 is required to maintain cellular heterogeneity in pancreatic cancerNature, 2022; DOI: 10.1038/s41586-022-04888-7