Research led by Weill Cornell Medicine provides new evidence that most colorectal cancers begin with the loss of intestinal stem cells, even before cancer-causing genetic alterations appear. The results, published on May 29 in Developmental Cell, overturn the prevailing theory for colorectal tumor initiation and suggest new ways to diagnose the disease before it has a chance to become established.
“Colorectal cancer is very, very heterogeneous, which has made it difficult for many years to classify these tumors in order to inform therapy,” said senior author Dr. Jorge Moscat, Homer T. Hirst III Professor of Oncology in Pathology and Vice-Chair for Cell and Cancer Pathobiology in the Department of Pathology and Laboratory Medicine at Weill Cornell Medicine. This heterogeneity, the diverse characteristics of colorectal tumor cells in different patients and also within the same tumor, makes treatment particularly challenging.
Colorectal tumors can arise from two types of pre-cancerous polyps: conventional adenomas and serrated adenomas. Conventional adenomas were thought to develop from mutations in the normal stem cells that lie at the bottoms of intestinal crypts, pit-like structures in the lining of the intestine. Serrated adenomas, on the other hand, are associated with a different type of stem-like cell with fetal characteristics that appears mysteriously at the tops of the crypts. Scientists in the field have described these apparently distinct tumor-forming processes as “bottom-up” and “top-down.”
“We wanted to determine how those two routes really start and how they progress, so we can better understand their heterogeneity as the cancer progresses,” said co-senior author Dr. Maria Diaz-Meco, Homer T. Hirst Professor of Oncology in Pathology in the Department of Pathology and Laboratory Medicine at Weill Cornell Medicine and a member of the Meyer Cancer Center at Weill Cornell Medicine. That’s particularly important for serrated tumors, which doctors sometimes miss because of their initial flat shape, and which can become aggressive cancers later.
The co-first authors are Dr. Hiroto Kinoshita and Dr. Anxo Martinez-Ordoñez, postdoctoral associates in the Department of Pathology and Laboratory Medicine at Weill Cornell Medicine.
Getting to the Bottom of Colorectal Cancer
The researchers previously found that many human colorectal tumors of both origins have abnormally low levels of proteins called atypical protein kinase C (aPKC). The new study investigated what happens when the aPKC genes are inactivated in animal models and cultured intestinal organoids.

“We approached this project with the bottom-up and top-down theories, but we were surprised to find that both tumor types showed loss of intestinal stem cells after aPKC genes were inactivated,” said Dr. Moscat, who is also a member of the Sandra and Edward Meyer Cancer Center at Weill Cornell Medicine.
The characteristic top-side stem cells on serrated adenomas only arise after the normal stem cells at the bottom of the crypt die, throwing the structure of the entire crypt into disarray. “So, the conventional cancer is bottom-up, and the serrated cancer is also bottom-up,” said Dr. Moscat.
The findings suggest a new unified model for the initiation of colorectal cancer where damage to the intestinal crypts causes a decrease in aPKC protein expression, followed by loss of the normal stem cells at the bottom of the crypt. Without those stem cells, the crypt cells can’t regenerate. To survive, the structure can spawn either a replacement population of regenerative stem cells at the bottom, or more fetal-like stem cells at the top. These replacement cells may then lead to cancer.
“If we can better understand how aPKC protein expression is regulated, we could control and prevent tumor development, and also better understand the progression of tumors,” said Dr. Diaz-Meco. The team is now looking at aPKC expression patterns in human tumors at different stages, with hopes of developing molecular tests that could be used to detect tumors earlier, classify tumors in patients and develop better treatments.

A new method of drug delivery using proline, an amino acid found in chicken feathers and skin tissue, could be used to limit the side effects of chemotherapy and repair important enzymes, new research suggests.
Published in the journal Chem, researchers have designed a cage (a box made of single molecules) from biologically compatible peptides, short amino acids that form the basis of proteins. These cages can house drugs of different sizes and transport them in the body with high levels of precision.
The negative side effects associated with chemotherapy, such as hair loss and nerve damage, are a result of ‘off-site toxicity’, where the treatment kills healthy cells surrounding tumours as well as the tumour itself. By creating a nano-sized cage to house the drug and carry it into the tumour before releasing it, this effect can be channelled more directly to the tumour, shielding healthy cells.
The cage can be tuned to different sizes, enabling different payloads of drugs. This flexible structure allows for chemotherapy drugs, antibiotics, and antivirals to potentially be delivered. Previously, cages of this kind could only be made using hydrocarbon molecules found in tar, which can often be toxic to humans.
This structure, researchers believe, also opens the door for faulty enzymes to be replaced within the body, which has previously not been possible. Historically, enzymes, which are composed of proteins and perform important functions in the body, could only have their activities blocked by drugs. The blocking of this functionality would then have an impact in the body, like reducing inflammation. Now, the cages could replace this function which may lay the groundwork for a new form of treatment.
Principal author Dr Charlie McTernan, Lecturer in Chemistry at King’s College London and Group Leader at the Francis Crick Institute, said “What we’ve created is essentially a biologically compatible molecular teabag. We can fill this teabag, or cage made from widely available proline and collagen, with several different medicines and deliver them in a much more targeted way than we could before.”
“In time, we hope that this could mean that we can limit the hair loss, nausea, and other unpleasant side effects of chemotherapy. We might even be able to repair malfunctioning enzymes that have an influence on the development of cancer. The best part is we can do this sustainably and at scale.”
Proline is very straight and rigid in shape, while also being soluble in water, which makes it uniquely suited for drug delivery, as water makes up roughly 60% of the human body. By binding the peptide to small amounts of metal such as palladium, the researchers could create a tuneable structure they could rapidly increase or decrease in size.
As proline and collagen are widely available and don’t rely on chains of hydrocarbons like previous methods, the team hope to sustainably scale up their current production in the lab.

The average age at menarche — the first menstrual period — has been decreasing among younger generations in the U.S., especially those belonging to racial minorities and lower socioeconomic statuses, according to a new study led by researchers at Harvard T.H. Chan School of Public Health. It also found that the average time it takes for the menstrual cycle to become regular is increasing.
The study will be published on May 29 in JAMA Network Open. It is the latest publication from the Apple Women’s Health Study, a longitudinal study of menstrual cycles, gynecological conditions, and overall women’s health conducted by Harvard Chan School, the National Institute of Environmental Health Sciences, and Apple.
“Our findings can lead to a better understanding of menstrual health across the lifespan and how our lived environment impacts this critical vital sign,” said co-principal investigator Shruthi Mahalingaiah, assistant professor of environmental, reproductive, and women’s health at Harvard Chan School.
While previous studies have shown trends towards earlier menarche over the past five decades, data has been limited on how these trends present within different racial groups and socioeconomic statuses. Additionally, few studies have had sufficient data to identify any trends regarding time to menstrual cycle regularity.
The researchers used the Apple Women’s Health Study’s large, diverse dataset to fill this research gap. Participants who enrolled in the study between November 2018 and March 2023 — 71,341 in total — self-reported the age at which they first began menstruating and their race and socioeconomic status. The researchers divided the participants into five age brackets: born between 1950-1969, 1970-1979, 1980-1989, 1990-1999, and 2000-2005. Ages of menarche were defined as early (younger than 11 years old), very early (younger than 9), and late (ages 16 and above). A subset of participants (61,932) self-reported the time it took for their menstrual cycle to become regular and were divided into five categories: up to two years, between three and four years, longer than five years, hasn’t become regular, or became regular with use of hormones. Another subset (9,865) provided their body mass index (BMI) at their age of menarche.
The study found that as birth year increased (meaning younger participants), average age at menarche decreased and time from menarche to menstrual cycle regularity increased. Among participants born from 1950-1969, the average age at menarche was 12.5 years, and the rates of early and very early menarche were 8.6% and 0.6%, respectively. Among participants born from 2000-2005, the average age of menarche was 11.9 years, and the rates of early and very early menarche were 15.5% and 1.4%, respectively. Across the two groups, the percentage of participants who reached menstrual cycle regularity within two years of menarche decreased from 76% to 56%. The researchers observed that these trends were present among all sociodemographic groups but were most pronounced among the participants who identified as Black, Hispanic, Asian, or mixed race, and who rated themselves as belonging to a low socioeconomic status.
The findings showed that BMI at age of menarche could explain part of the trend toward periods starting earlier — in other words, that childhood obesity, a risk factor for early puberty and a growing epidemic in the U.S., could be a contributing factor to earlier menarche. Other possible factors that might explain the trend include dietary patterns, psychological stress and adverse childhood experiences, and environmental factors such as endocrine-disrupting chemicals and air pollution.

“Continuing to investigate early menarche and its drivers is critical,” said corresponding author Zifan Wang, postdoctoral research fellow in Harvard Chan School’s Department of Environmental Health. “Early menarche is associated with higher risk of adverse health outcomes, such as cardiovascular disease and cancer. To address these health concerns — which our findings suggest may begin to impact more people, with disproportionate impact on already disadvantaged populations — we need much more investment in menstrual health research.”
The authors noted some limitations to the study, including that it relies heavily on retrospective self-reporting.
Other Harvard Chan School authors included Gowtham Asokan, Jukka-Pekka Onnela, Michelle Williams, Russ Hauser, and Brent Coull.
The study was made possible by funding from Apple, Inc. and the National Institutes of Health (grant Z01ES103333).

Researchers have developed a new antibiotic that reduced or eliminated drug-resistant bacterial infections in mouse models of acute pneumonia and sepsis while sparing healthy microbes in the mouse gut. The drug, called lolamicin, also warded off secondary infections with Clostridioides difficile, a common and dangerous hospital-associated bacterial infection, and was effective against more than 130 multidrug-resistant bacterial strains in cell culture.
The findings are detailed in the journal Nature.
“People are starting to realize that the antibiotics we’ve all been taking — that are fighting infection and, in some instances, saving our lives — also are having these deleterious effects on us,” said University of Illinois Urbana-Champaign chemistry professor Paul Hergenrother, who led the study with former doctoral student Kristen Muñoz. “They’re killing our good bacteria as they treat the infection. We wanted to start thinking about the next generation of antibiotics that could be developed to kill the pathogenic bacteria and not the beneficial ones.”
Numerous studies have found that antibiotic-related disturbances to the gut microbiome increase vulnerability to further infections and are associated with gastrointestinal, kidney, liver and other problems.
“Most clinically approved antibiotics only kill gram-positive bacteria or kill both gram-positive and gram-negative bacteria,” Muñoz said.
Gram-positive and gram-negative bacteria differ in the composition of their cell walls. Gram-negative bacteria have a double layer of protection, making them more difficult to kill, Muñoz said.
The few drugs available to fight gram-negative infections also kill other potentially beneficial gram-negative bacteria. For example, colistin, one of the few gram-negative-only antibiotics approved for clinical use, can cause C. difficile-associated diarrhea and pseudomembranous colitis, a potentially life-threatening complication. The drug also has toxic effects on the liver and kidney, and “thus colistin is typically utilized only as an antibiotic of last resort,” the researchers wrote.

To tackle the many problems associated with indiscriminately targeting gram-negative bacteria, the team focused on a suite of drugs developed by the pharmaceutical company AstraZeneca. These drugs inhibit the Lol system, a lipoprotein-transport system that is exclusive to gram-negative bacteria and genetically different in pathogenic and beneficial microbes. These drugs were not effective against gram-negative infections unless the researchers first undermined key bacterial defenses in the laboratory. But because these antibiotics appeared to discriminate between beneficial and pathogenic gram-negative bacteria in cell culture experiments, they were promising candidates for further exploration, Hergenrother said.
In a series of experiments, Muñoz designed structural variations of the Lol inhibitors and evaluated their potential to fight gram-negative and gram-positive bacteria in cell culture. One of the new compounds, lolamicin, selectively targeted some “laboratory strains of gram-negative pathogens including Escherichia coli, Klebsiella pneumoniae and Enterobacter cloacae,” the researchers found. Lolamicin had no detectable effect on gram-positive bacteria in cell culture. At higher doses, lolamicin killed up to 90% of multidrug-resistant E. coli, K. pneumoniae and E. cloacae clinical isolates.
When given orally to mice with drug-resistant septicemia or pneumonia, lolamicin rescued 100% of the mice with septicemia and 70% of the mice with pneumonia, the team reported.
Extensive work was done to determine the effect of lolamicin on the gut microbiome.
“The mouse microbiome is a good tool for modeling human infections because human and mouse gut microbiomes are very similar,” Muñoz said. “Studies have shown that antibiotics that cause gut dysbiosis in mice have a similar effect in humans.”
Treatment with standard antibiotics amoxicillin and clindamycin caused dramatic shifts in the overall structure of bacterial populations in the mouse gut, diminishing the abundance several beneficial microbial groups, the team found.
“In contrast, lolamicin did not cause any drastic changes in taxonomic composition over the course of the three-day treatment or the following 28-day recovery,” the researchers wrote.
Many more years of research are needed to extend the findings, Hergenrother said. Lolamicin, or other similar compounds, must be tested against more bacterial strains and detailed toxicology studies must be conducted. Any new antibiotics also must be assessed to determine how quickly they induce drug resistance, a problem that arises sooner or later in bacteria treated with antibiotics.
The study is a proof-of-concept that antibiotics that kill a pathogenic microbe while sparing beneficial bacteria in the gut can be developed for gram-negative infections — some of the most challenging infections to treat, Hergenrother said.

Researchers at Duke University have developed an assistive machine learning model that greatly improves the ability of medical professionals to read the electroencephalography (EEG) charts of intensive care patients.
Because EEG readings are the only method for knowing when unconscious patients are in danger of suffering a seizure or are having seizure-like events, the computational tool could help save thousands of lives each year. The results appear online May 23 in the New England Journal of Medicine AI.
EEGs use small sensors attached to the scalp to measure the brain’s electrical signals, producing a long line of up and down squiggles. When a patient is having a seizure, these lines jump up and down dramatically like a seismograph during an earthquake — a signal that is easy to recognize. But other medically important anomalies called seizure-like events are much more difficult to discern.
“The brain activity we’re looking at exists along a continuum, where seizures are at one end, but there’s still a lot of events in the middle that can also cause harm and require medication,” said Dr. Brandon Westover, associate professor of neurology at Massachusetts General Hospital and Harvard Medical School. “The EEG patterns caused by those events are more difficult to recognize and categorize confidently, even by highly trained neurologists, which not every medical facility has. But doing so is extremely important to the health outcomes of these patients.”
To build a tool to help make these determinations, the doctors turned to the laboratory of Cynthia Rudin, the Earl D. McLean, Jr. Professor of Computer Science and Electrical and Computer Engineering at Duke. Rudin and her colleagues specialize in developing “interpretable” machine learning algorithms. While most machine learning models are a “black box” that makes it impossible for a human to know how it’s reaching conclusions, interpretable machine learning models essentially must show their work.
The research group started by gathering EEG samples from over 2,700 patients and having more than 120 experts pick out the relevant features in the graphs, categorizing them as either a seizure, one of four types of seizure-like events or ‘other.’ Each type of event appears in EEG charts as certain shapes or repetitions in the undulating lines. But because these charts are rarely steadfast in their appearance, telltale signals can be interrupted by bad data or can mix together to create a confusing chart.
“There is a ground truth, but it’s difficult to read,” said Stark Guo, a Ph.D. student working in Rudin’s lab. “The inherent ambiguity in many of these charts meant we had to train the model to place its decisions within a continuum rather than well-defined separate bins.”
When displayed visually, that continuum looks something like a multicolored starfish swimming away from a predator. Each differently colored arm represents one type of seizure-like event the EEG could represent. The closer the algorithm puts a specific chart toward the tip of an arm, the surer it is of its decision, while those placed closer to the central body are less certain.

Besides this visual classification, the algorithm also points to the patterns in the brainwaves that it used to make its determination and provides three examples of professionally diagnosed charts that it sees as being similar.
“This lets a medical professional quickly look at the important sections and either agree that the patterns are there or decide that the algorithm is off the mark,” said Alina Barnett, a postdoctoral research associate in the Rudin lab. “Even if they’re not highly trained to read EEGs, they can make a much more educated decision.”
Putting the algorithm to the test, the collaborative team had eight medical professionals with relevant experience categorize 100 EEG samples into the six categories, once with the help of AI and once without. The performance of all of the participants greatly improved, with their overall accuracy rising from 47% to 71%. Their performance also rose above those using a similar “black box” algorithm in a previous study.
“Usually, people think that black box machine learning models are more accurate, but for many important applications, like this one, it’s just not true,” said Rudin. “It’s much easier to troubleshoot models when they are interpretable. And in this case, the interpretable model was actually more accurate. It also provides a bird’s eye view of the types of anomalous electrical signals that occur in the brain, which is really useful for care of critically ill patients.”
This work was supported by the National Science Foundation (IIS-2147061, HRD-2222336, IIS-2130250, 2014431), the National Institutes of Health (R01NS102190, R01NS102574, R01NS107291, RF1AG064312, RF1NS120947, R01AG073410, R01HL161253, K23NS124656, P20GM130447) and the DHHS LB606 Nebraska Stem Cell Grant.

New research out of VCU Massey Comprehensive Cancer Center — recently published in Drug Resistance Updates — revealed a previously unknown biological process through which breast tumor cells develop resistance to standard treatment, which could open the door for cancer scientists around the world to further target this vulnerability in hopes of creating more effective therapies for disease.
Additionally, the research team tested a promising drug in combination with an existing therapy that achieved total remission in one breast cancer model that was resistant to the standard of care, and reduced cancer growth by nearly 70% in other models of advanced disease.
“Overcoming drug resistance is very important; we’re talking about developing new drugs that will provide hope to cancer patients who have exhausted all options available to them,” said study author Yuesheng Zhang, M.D., Ph.D., the Harrigan, Haw and Luck Families Chair in Cancer Research and a member of the Developmental Therapeutics research program at Massey.
HER2-positive breast cancers represent nearly one-fifth of all breast tumors and often grow and spread much faster than tumors that are HER2-negative, according to the American Cancer Society.
In the clinic, there are three classes of drugs approved to treat this type of cancer: 1) monoclonal antibodies (trastuzumab, sold under the brand name Herceptin, is the most commonly used drug to treat HER2-positive breast cancer), 2) tyrosine kinase inhibitors and 3) antibody small molecule conjugates.
These drugs are all designed to target the HER2 protein receptor, but Zhang’s team found that none of them are actually strong enough to completely eliminate it, offering the tumor cells more freedom to partner with other proteins within the cells to initiate cancer signaling, continue to multiply and spread.
“It’s actually quite striking. It turns out the efficacy of the standard of care is very limited, and not only do a small percentage of patients respond to treatment, even the ones who are responding will very quickly develop resistance,” said Zhang, who is also a professor in the Department of Pharmacology and Toxicology at the VCU School of Medicine.

This paper shines a light on the clinical need to more completely rid the cancer cells of HER2, but it also points a finger at the EGFR gene family being just as important in these tumors, suggesting that truly effective drugs need to target both the HER2 and EGFR receptors to eradicate disease.
“This study shows that an agent that targets the degradation of both HER2 and EGFR is highly effective in overcoming drug resistance in this disease,” Zhang said. “The findings provide new insights and innovations for advancing treatment of drug-resistant HER2-positive breast cancer that remains an unmet problem.”
In addition to unveiling a key flaw hindering current treatment options for breast cancer, Zhang and his research team also tested a novel agent that showed promising results against these tumors.
In one model, the drug achieved complete remission in combination with garadacimab, another type of monoclonal antibody. In other models where the tumors had spread to the brain, the researchers found their drug inhibited tumor growth by up to 68%.
“This means that this drug can actually cross the blood-brain barrier to attack the tumor in the brain,” Zhang said.
Zhang added they are currently in talks with the National Cancer Institute for them to manufacture the drug in an effort to gain FDA approval and move it into clinical trials.
While the team observed encouraging results with the drug they tested in the study, Zhang said the most important impact of this research is to inform other investigators about the underlying mechanism that can be targeted in HER2-positive breast cancer to improve patient outcomes.
“This paper will move the field forward because cancer scientists can develop other drugs to target this vulnerability; our agent is only one of those,” Zhang said.
Collaborators on this study include Lu Yang, Ph.D., Arup Bhattacharya, Ph.D., Yun Li, M.S., and Valentina Robila, M.D., Ph.D., of the VCU School of Medicine; Darrell Peterson, Ph.D., of the VCU School of Pharmacy; Xiaozhuo Liu, Ph.D., of Roswell Park Comprehensive Cancer Center; and Elisabetta Marangoni, Ph.D., of the Institute of Curie in France.

What makes chocolate taste and smell so delicious? Chemistry, of course! A variety of molecules work together to create that unmistakable aroma, but those same molecules might carry some unwanted health effects if there are too many around. According to research published in ACS’ Journal of Agricultural and Food Chemistry, while many of the compounds appeared in chocolate in low enough concentrations to be safe, higher amounts were found in some baked sweet treats.
When making chocolate, cocoa beans are roasted to help their chocolatey flavors shine. During this process, new molecules like α,β-unsaturated carbonyls are formed when they react with other ingredients under high temperatures. This class of carbonyls is highly reactive and potentially genotoxic, or able to cause damage to DNA when consumed. Though naturally found in many foods, these carbonyls are also used as flavoring additives, and some have been banned in the European Union, including the buttery-tasting furan-2(5H)-one. To better understand how these molecules form naturally in foods, and whether or not they are present in levels that could pose a health concern, Alexandre Dusart and colleagues tested chocolates and other sweet treats for 10 different α,β-unsaturated carbonyls — some of which have been confirmed as safe by the European Food Safety Authority, while others are still under evaluation.
The team created its own chocolates and found that α,β-unsaturated carbonyls formed during roasting and after the addition of cocoa butter; however, their concentrations remained too low to pose any health concerns from consuming the chocolates. Next, researchers screened 22 commercially available desserts, including crepes, waffles, cakes and biscuits, either with or without chocolate. In these packaged treats, they found even lower concentrations of nine of the 10 carbonyls compared to the chocolates.
The remaining carbonyl — genotoxic furan-2(5H)-one — appeared in much higher concentrations in the crepe and cake samples, reaching up to 4.3 milligrams per kilogram. Considering that the recommended threshold for genotoxic substances is only 0.15 micrograms per person per day, consuming these desserts could exceed that limit, though additional studies are needed to accurately assess the potential health risk.
Researchers concluded that the furan-2(5H)-one molecule likely formed during the baking process and did not seem to correlate with the amount of chocolate present in the packaged desserts. The team says that this work helps to better understand where these carbonyls come from in chocolate and highlights the importance of monitoring flavorings in food to keep consumers informed and safe.
The authors acknowledge funding from the Belgian Federal Public Service of Health, Food Chain Safety and Environment.

A team led by Melanie Köhler and Veronika Somoza from the Leibniz-Institute for Food Systems Biology has presented a new research approach in the journal Nature Food. The perspectives article focuses on different ways to study the mouthfeel of food using atomic force microscopy to better understand the biophysical mechanisms that contribute to taste sensations in a broader sense. New findings in this area could drive the development of health-promoting products that contain less salt, fat, sugar and calories but still have a convincing mouthfeel.
The mouthfeel of a food plays a crucial role in its acceptance. For example, many people prefer the creamy consistency of quark and yogurt. Apples, on the other hand, should be juicy and crunchy and bread crusts crispy. This diversity shows that the optimal mouthfeel is highly dependent heavily on the type of food and is not uniformly defined.
Researching complex interactions
In addition, the interplay between the constituents, the texture and the temperature of food and various sensor molecules and cell types in the mouth is extremely complex. Junior research group leader Melanie Köhler says: “Mechanoreceptors in particular, which react to pressure or stretching, have been under-researched so far with regard to the optimal mouthfeel and their contribution to the flavor of a food.”
Veronika Somoza, Director of the Leibniz Institute in Freising, adds: “In our current perspectives article, we present various experimental approaches that can be used to tackle the many unanswered questions surrounding mouthfeel from a biophysical point of view perspective. We have focused on biological atomic force microscopy.”
The atomic force microscope is a suitable tool for scanning surfaces at an atomic level or investigating interactions between molecules such as food constituents and receptor proteins. However, it can also be used to apply mechanical pressure to cells, thereby activating mechanoreceptors for identifying and characterizing their cellular signal response.
Rethinking the traditional definition
According to Melanie Köhler, a fundamental biophysical and functional understanding of the diverse mechanosensory key players in oral and extra-oral tissue as well as their responses to food constituents is important. It enables constructing new hypotheses about the contribution of mechanosensors to the overall sensory impression of a food for answering many questions that are still pending in the molecular field.
“With regard to food research, we expect that future results will lead to a revision of our traditional definition of flavor, i.e. the overall sensory impression of a food, by including mechanical perception as an additional factor alongside taste and smell,” explains the young scientist. “In terms of food production, our pioneering research approach opens up promising perspectives for the design of future, enjoyable and health-conscious nutritional options,” Melanie Köhler continues.

Tumor cells often hijack normal physiological processes to support their growth, exploiting proteins that are in charge of essential cell functions. It is therefore important to block the activity of these proteins only in cancer cells without affecting their crucial roles in healthy tissues. For this reason, classical approaches using small molecules that induce systemic inhibition across all cells in the body can lead to severe side effects.
An example of essential proteins hijacked by cancer cells are the cathepsins, a family of enzymes that is responsible for breaking down other proteins and remodeling the body’s tissues. Cathepsins are implicated in various cancers, osteoporosis, and autoimmune diseases. However, clinical trials with small molecule inhibitors of cathepsins have failed due to either lack of efficacy or toxicity.
Now, a team of scientists led by Elisa Oricchio and Bruno Correia at EPFL has developed a novel approach to overcome these limitations. They created a modular drug platform that conjugates non-natural peptide inhibitors (NNPIs) with antibodies, creating antibody-peptide inhibitor conjugates (APICs). This method ensures that the inhibitors are delivered specifically to cancer cells, thereby reducing systemic side effects and increasing therapeutic efficacy.
The researchers began by designing NNPIs that covalently bind to and inhibit cathepsins. They modified peptide sequences to include a Michael acceptor, a chemical moiety that makes it easier to form a stable bond with cathepsins.
The Michael acceptor reacts with the cysteine residue in the cathepsin’s active site (the part that of the enzyme responsible for its main job), creating a stable, covalent linkage that effectively inhibits the cathepsin. To further optimize the peptides’ specificity and potency, the team used saturation mutagenesis screening — a method that systematically changes each amino acid in a protein to find the best variants with desired properties.
The researchers identified several strong inhibitors against four different cathepsins, namely cathepsin S, B, K and L. By attaching these inhibitors to antibodies that recognize CD22, CD79, HER2, and Siglec15, the researchers could precisely deliver the NNPIs to lymphoma cells, breast cancer cells and osteoclasts. This leverages the natural ability of antibodies to be internalized by target cells, precisely directing the inhibitors where they are needed.
Then, it was time to test the APICs: in both cell lines and animal models, they showed significant therapeutic effects. For example, in lymphoma models, treatment with APICs that target cathepsin S led to tumor reduction and activation of the immune response against cancer cells. In breast cancer models, APICs targeting cathepsin B hindered tumor invasiveness and cell migration, underscoring the potential of APICs to prevent metastasis.
By delivering inhibitors specifically to cancer cells, the APIC approach can avoid or minimize the side effects commonly associated with other treatments, such as chemotherapies. Moreover, the modular nature of the APIC design means it can be adapted to target various proteases implicated in different diseases, potentially revolutionizing the treatment landscape for conditions beyond cancer.
The APIC project is now extending beyond the lab and is taking its first steps towards becoming a clinical reality. “We filed two patent applications based on this project,” says Elisa Oricchio. “Aaron Petruzzella, the PhD student who led the project, recently received the support of the SNF Bridge Proof of Concept fellowship to continue working on these inhibitors, build the foundations of a start-up and attract the attention of potential investors.”

According to the U.S. Centers for Disease Control and Prevention, 80% of those living with dementia receive informal care from family members or friends. This equates to 16 million family caregivers in the U.S. However, caring for family members with dementia is often associated with increased caregiver burden (which includes emotional, physical, and financial strain), stress, and worse physical health for the caregiver.
A recent study published in the Journal of Applied Gerontology, led by George Mason University researchers, found that a 9-week online stress management intervention program for family caregivers reduced burden scores by 15% for 97 family caregivers of older adults living with dementia. The Stress-Busting Program for Family Caregivers TM, intervention was specifically designed to help family caregivers with managing their own stress when caring for older adults living with dementia or a chronic illness
“In this study, we found evidence of a range in average caregiver burden levels based on the dementia severity category of care recipients. The findings show that an online Zoom intervention in a peer group setting can be beneficial for family caregivers of older adults with mild, moderate, or severe dementia,” said Catherine Tompkins, principal investigator, professor of social work, and associate dean of faculty and staff affairs in the College of Public Health.
The intervention provided family caregivers with education and strategies to manage stress when caring for someone living with dementia. Examples of self-care techniques included breathing and meditation; troubleshooting behaviors associated with dementia; and peer-to-peer support within a virtual group setting.
“Reducing caregiver burden and managing stress are critical to the well-being of families. These findings show that effective stress management interventions for family caregivers can be facilitated through online peer groups,” said Gilbert Gimm, first author and associate professor of health administration and policy.
“Mason Caregivers Aiming for Resilience, Empowerment, and Support Study: Assessing Family Caregiver Burden Post-Intervention” was published online in April 2024. Co-authors include George Mason Associate Professor Megumi Inoue, Professor Emily Ihara, Mason CARES Project Manager Shannon Layman, and Master of Social Work alumna graduate Harveen Pantleay. This study was supported by a grant (#2021048) from the Retirement Research Foundation (RRF).
The study is part of a larger project, entitled Mason CARES (Caregivers Aiming for Resilience, Empowerment, and Support), that implemented and assessed interventions for family caregivers.