James Weger-Lucarelli, an associate professor in the Department of Biomedical Sciences and Pathobiology at the Virginia-Maryland College of Veterinary Medicine, received a five-year, $2.01 million grant from the National Institute of General Medical Sciences in July. The funding will support his investigation into how virus-infected cells live and die, and what chemical messages they send to their neighbors in the process. The research could eventually inform the development of better antiviral drugs and more effective virus-based cancer therapies.

The award is an R35 Maximizing Investigators' Research Award, or MIRA, a funding mechanism that differs fundamentally from the standard research grant. Rather than supporting a single, narrowly defined hypothesis, a MIRA funds an investigator's entire research program over a long horizon. The National Institute of General Medical Sciences, the institute within the National Institutes of Health charged with supporting basic research that underlies all of medicine, reserves these awards for scientists it considers worth backing long-term, giving them latitude to pursue unexpected findings wherever they lead.

"An R01 is more like you're saying exactly what you're going to do, almost experiment by experiment," Weger-Lucarelli said. "This is more conceptual. You have a little more freedom to be exploratory, more innovative."

The distinction is more than administrative. Under a conventional R01 award, researchers commit in advance to a specific set of aims and experiments, and significant deviations typically require written approval from the funding agency. That structure works well for projects where the science is predictable, but it can penalize investigators whose most interesting results are the ones they did not anticipate. By contrast, the MIRA mechanism was designed around the recognition that scientific progress often comes from following surprise rather than plan. For an early-stage program like Weger-Lucarelli's, the difference can determine whether an unexpected observation becomes a dead end noted in a report or the seed of an entirely new research direction.

That flexibility matters for the kind of questions his lab is asking. When a virus infects a cell, the cell can trigger its own shutdown, a form of programmed self-destruction that halts the virus before it can replicate and escape to infect other cells. This is one of the most ancient defenses in biology, a form of cellular altruism in which the individual sacrifices itself to protect the collective. Viruses, in turn, have evolved countermeasures to interfere with that decision, deploying viral proteins that block the cell's death machinery or disable the alarm systems that would otherwise alert the immune system, keeping their host cells alive long enough to produce new viral particles.

And when an infected cell does die, it can release chemical signals into the surrounding tissue — a kind of biological warning broadcast that either rallies the immune response to contain the infection or, if a virus has found a way to redirect those signals, accelerates the spread instead. Dying cells can release inflammatory molecules that recruit immune cells to the site of infection, or they can release signals that dampen inflammation, depending on how they die and what the virus permits them to release. The distinction can mean the difference between a localized, quickly controlled infection and a systemic one.

"The cell is trying to stop the virus, and the virus is trying to stop the host," Weger-Lucarelli said. "It's kind of like a tennis match between the two."

The metaphor captures an arms race that has been running for hundreds of millions of years. Every virus that successfully infects humans today has already survived countless rounds of selection in which cells that died promptly eliminated viral lineages, and viruses that kept cells alive spread more widely. What remains in circulation is a snapshot of that ongoing negotiation, and the molecular details of it remain surprisingly poorly mapped, particularly in tissue environments that resemble the body rather than the laboratory.

Understanding that back-and-forth requires studying it in conditions that approximate a real body, and a central thrust of the funded work involves moving experiments out of flat laboratory dishes and into three-dimensional tissue models. In these 3D systems, multiple cell types can interact with one another the way they do in living tissue, rather than in the artificial uniformity of a standard cell culture. In a conventional dish, a single cell type grows in a thin layer, bathed in nutrient-rich medium, with unlimited access to oxygen and no structural architecture. In three-dimensional models, cells pack together, exchange nutrients through diffusion gradients, and communicate across cell-type boundaries much as they do in an organ.

The difference, it turns out, is not subtle. In Weger-Lucarelli's laboratory, viruses kill cells in conventional two-dimensional culture within two days. In three-dimensional models, the very same cells survive a week or more under otherwise identical conditions.

"If we were doing it in 2D, we would have missed that entirely," he said.

That observation carries implications well beyond his own lab. Cell culture in flat dishes has been a cornerstone of virology for decades, and countless antiviral strategies have been evaluated on the assumption that what happens in a dish reflects what happens in tissue. If infected cells behave so differently when embedded in a three-dimensional architecture — surviving several times longer, signaling to a more diverse cast of neighboring cells — then some conclusions drawn from traditional cultures may need re-examination. The extended survival of infected cells in 3D models suggests that the live-or-die decision Weger-Lucarelli studies may unfold on a different timeline, and through different molecular conversations, than the field has generally appreciated.

It also suggests a possible explanation for clinical observations that have long puzzled virologists, such as why some infections linger in tissues for extended periods or why antiviral drugs that perform well in cell culture sometimes underperform in patients. A drug tested against cells that die within days in a dish may look highly effective against a process that, in a real tissue, plays out over weeks. Better models of the infected tissue environment could help close that translational gap, reducing the attrition that has historically plagued antiviral development.

The grant also strengthens an existing collaboration with Samy Lamouille at the Fralin Biomedical Research Institute at VTC focused on oncolytic viruses — viruses engineered to seek out and kill tumor cells — as a potential treatment for glioblastoma, the aggressive brain cancer that remains one of the most difficult malignancies to treat. Glioblastoma is the most common malignant brain tumor in adults, and despite decades of research, median survival after diagnosis remains measured in months rather than years. The tumor's characteristic diffuse infiltration into surrounding brain tissue makes complete surgical removal nearly impossible, and the blood-brain barrier complicates drug delivery, leaving a pressing need for fundamentally different therapeutic approaches.

The logic of oncolytic therapy inverts the usual virological arms race: instead of trying to help cells survive infection, researchers deploy a virus designed to push cancer cells decisively toward death. Many cancer cells, including glioblastoma cells, have acquired defects in their antiviral defenses precisely because those defenses can otherwise trigger cell death — a vulnerability that oncolytic viruses are engineered to exploit. But the strategy is only as good as the cell-death and signaling programs the virus activates. A tumor cell killed quietly may release signals that suppress the immune response, while one killed in a way that broadcasts alarm may provoke a broader immune attack on the tumor. Understanding how infected cells decide their own fate, and what they signal when they die, is directly relevant to making such therapies work more reliably.

Positioned within a veterinary college gives the glioblastoma work an unusual advantage. Dogs are among the few animals that develop glioblastoma naturally, and their tumors develop in ways that closely resemble the human disease. That creates a clinical testing pathway that would be far harder to access from a conventional research setting. A laboratory mouse implanted with a human tumor offers a controlled but artificial system; a pet dog presenting at a veterinary hospital with a spontaneously arising brain tumor brings the full complexity of the disease — its genetic heterogeneity, its interaction with a functioning immune system, its growth in a real brain of real size — into the research program.

"There's a huge advantage of being at a vet school," Weger-Lucarelli said, "especially with brain cancer. Dogs have naturally occurring brain cancer that develops very similar to humans. Much easier than going into human clinical trials."

Comparative oncology of this kind has gained traction in recent years precisely because spontaneous tumors in pets recapitulate the biological complexity of human cancers — the heterogeneous cell populations, the three-dimensional tissue architecture, the immune environment — that laboratory models often strip away. Veterinary clinical trials can enroll more patients than early-phase human trials, proceed more quickly through regulatory pathways, and generate data on dosing, toxicity, and response in an outbred population that is far more representative of human patients than inbred laboratory mice. Findings from canine patients can inform the design of eventual human trials, while the dogs themselves may benefit from experimental treatments that would otherwise be unavailable.

The fit between the two halves of the program is deliberate. The same fundamental question — how a cell responds to viral infection, and what message that response sends — governs whether an antiviral defense succeeds in normal tissue and whether an oncolytic virus succeeds in a tumor. Insights from the 3D tissue models can feed directly into the glioblastoma collaboration, and observations from canine patients can point the fundamental research toward the biology that matters most in disease.

The five-year funding window also shapes the human side of the laboratory. Weger-Lucarelli is currently recruiting a postdoctoral researcher and plans to work at the bench alongside whoever joins the lab, rather than managing from a distance. The long grant cycle allows him to recruit and retain graduate students and postdocs without the uncertainty that accompanies shorter, annually renewed funding, giving early-career scientists time to pursue questions that may take years to answer fully. In a research environment where many trainees live grant cycle to grant cycle, that stability is itself a scientific resource.

The research, carried out by Benjamin Minch and Mohammad Moniruzzaman, set out to address a long-standing puzzle in virology. Nucleocytoviricota, the phylum that includes many of the so-called giant viruses, possess exceptionally large and mosaic genomes whose composition can differ dramatically even among close relatives. These viruses, which infect a wide range of eukaryotic hosts including algae, amoebae, and other protists, routinely defy conventional expectations of what a virus should look like. While typical viruses carry only a handful of genes, giant viruses can harbor hundreds or even more than a thousand, sometimes including genes involved in DNA repair, protein folding, and metabolic processes once thought to be exclusive to cellular organisms. Yet the mechanisms that generate and maintain this plasticity have remained poorly understood. Genomic islands, which are dynamic stretches of DNA that act as major engines of diversification and adaptation in bacteria, seemed like an obvious candidate, but their contribution to giant virus evolution had never been systematically explored.

The concept of the genomic island is well established in bacterial genetics. In prokaryotes, these are discrete regions of the chromosome, often acquired through horizontal gene transfer and frequently flanked by mobile genetic elements such as tRNAs, integrases, or repeat structures, where genes of foreign origin cluster together. Bacterial genomic islands commonly encode traits with clear ecological value: pathogenicity factors, antibiotic resistance genes, metabolic pathways for exploiting unusual nutrients, or secretion systems for interacting with other organisms. Because islands concentrate functionally important novelty in mutable corners of the genome, they allow bacterial lineages to adapt quickly without destabilizing the core genes needed for basic survival. Whether viruses, and particularly viruses with genomes as large and complex as those of Nucleocytoviricota, employ an analogous strategy has been an open question.

To fill that gap, the researchers assembled a dataset of 369 high-quality giant virus genomes. Crucially, this collection included both cultured isolates, which can be studied under laboratory conditions, and long-read metagenome-assembled genomes, which are reconstructed directly from environmental samples using sequencing technologies that produce long, continuous DNA reads. Long-read approaches are particularly valuable in this context because they reduce the fragmentation that has historically made it difficult to assemble the repetitive and structurally complex regions where genomic islands tend to reside. Short-read sequencing, which dominated the genomics era for years, generates millions of brief reads that must be stitched together computationally; repetitive sequences longer than a single read cannot be resolved unambiguously, often collapsing into truncated or scrambled assemblies. Because genomic islands are precisely the kinds of regions rich in repeats, duplications, and recently inserted foreign DNA, earlier short-read-based surveys may have systematically underestimated their presence. By combining the two types of genomes, the team was able to survey genomic islands across a broad taxonomic and ecological range of Nucleocytoviricota, spanning viruses collected from marine, freshwater, and other environments.

The scale of the finding exceeded even the authors' expectations. Across this dataset, the analysis identified 307 genomic islands distributed among more than half of the genomes examined. This prevalence demonstrates that genomic islands are not rare anomalies confined to a few unusual lineages but a widespread and recurring feature of giant virus genome architecture. Wherever the researchers looked within the phylum, they found these dynamic regions punctuating otherwise more stable stretches of viral DNA. That consistency across diverse lineages suggests that the underlying biology, whatever mechanisms create and maintain islands, is deeply embedded in the evolution of the group rather than being a peculiarity of one branch of the viral family tree.

What sets genomic islands apart from the rest of the genome is their tendency to vary. The study found that these regions are frequently associated with genomic hypervariability, meaning that their gene content and organization change rapidly compared with neighboring parts of the genome. In practical terms, two giant virus genomes that are otherwise highly similar may differ substantially within their islands, with genes appearing, disappearing, or shuffling position. The comparative analyses revealed frequent gains and losses of genes within islands, as well as rearrangements, even among genomes that were nearly identical elsewhere. This pattern marks the islands as hotspots of genome diversification, places where evolutionary change concentrates while the rest of the genome remains comparatively conserved. Such a division of labor, between a stable genomic core and a volatile accessory periphery, mirrors patterns long recognized in bacterial pangenomics, where core genes maintain essential functions while accessory genes, often in islands, mediate ecological specialization.

The functional profile of the genes inside these islands offers a clue to why the viruses maintain such volatile real estate. The researchers found that genes involved in host interaction were enriched within the islands, with surface adhesion proteins standing out as a particularly prominent category. Surface adhesion proteins are molecules that help viruses attach to their host cells, a critical step in infection. For viruses that infect eukaryotic hosts, the initial recognition of a compatible cell surface is often the decisive determinant of host range: a virus that cannot bind cannot infect, no matter how well its genome is suited to replication once inside. In the context of the ongoing arms race between viruses and their hosts, rapid diversification of attachment machinery could allow a virus to recognize new host surfaces or evade host defenses that have evolved to block familiar binding strategies. The enrichment of such genes in hypervariable regions suggests that genomic islands serve as testing grounds for host adaptation, generating variation in exactly the functions most likely to influence infection success.

Beyond host adaptation, the study uncovered evidence pointing to an unexpected source of the island gene repertoire: bacteria. Many of the genomic islands were enriched in bacterial homologs, meaning they contained genes whose closest relatives are found in bacterial genomes rather than in other viruses. More strikingly, several islands exhibited synteny, conservation of gene order and organization, with genomic regions recovered from bacterial genomes found in the same environments as the viruses. Synteny of this kind is difficult to explain by chance. Genes tend to be reshuffled over evolutionary time, so two DNA segments drawn at random from unrelated organisms would almost never preserve the same gene order. The researchers interpret the observed synteny as support for large-scale genetic exchange between bacteria and giant viruses. In other words, stretches of DNA may have moved wholesale between co-occurring bacterial and viral genomes, seeding the islands with bacterial-derived genes arranged in the same order as in their bacterial counterparts.

This proposed exchange has significant implications for how scientists think about the mosaic nature of giant virus genomes. The mosaicism of Nucleocytoviricota has long been noted, with different genes in a single genome suggesting wildly different evolutionary origins, some resembling bacterial genes, others eukaryotic, and others related to other viruses. Competing hypotheses about the origin of giant viruses, whether they descended from an ancient viral lineage that grew by accumulating genes or acquired much of their complexity by harvesting DNA from hosts and other microbes, have fueled debate for years. If genomic islands are conduits for gene flow from bacteria, they could be one of the principal mechanisms by which viruses acquire foreign DNA, incorporating it into their genomes and, over time, reshaping their metabolic and interaction capabilities. The findings thus provide a framework for understanding genome plasticity, mosaicism, and the adaptive potential of giant viruses, positioning islands as the engines driving much of this change.

The evolutionary scenario that emerges is one in which giant virus genomes are not static blueprints but dynamic mosaics, continuously renovated at designated hotspots. Genes involved in host interaction are recruited into islands, where turnover is rapid and rearrangement common, allowing viral lineages to experiment with new surface proteins and other interaction factors. Meanwhile, genetic exchange with co-occurring bacteria supplies fresh raw material, some of which may prove useful in the virus-host arms race. This combination of internal volatility and external gene acquisition helps explain both the sheer size of giant virus genomes and their bewildering diversity of gene content. It also suggests a functional logic to the architecture: by concentrating volatility in islands, giant viruses can innovate aggressively while preserving the core machinery that replication requires, an arrangement that balances stability and change.

The study's methods merit attention for how they strengthen these conclusions. By using long-read metagenome-assembled genomes alongside cultured isolates, the researchers mitigated a key limitation of earlier work, which relied heavily on short-read assemblies that can break apart or misassemble the very repetitive regions where islands concentrate. Characterizing islands across more than 50 percent of a large and diverse set of genomes also lends statistical weight to the claims of pervasiveness, while the comparative approach of examining closely related genomes allowed the team to detect gain, loss, and rearrangement events that would be invisible when comparing only distant relatives. Comparisons among close relatives function like snapshots taken at short intervals: differences that accumulate between them reveal recent evolutionary activity, whereas comparisons across deep divergence wash such signals out.

As with any study, there are caveats and boundaries to the conclusions. Metagenome-assembled genomes, even those generated with long reads, are reconstructions rather than complete, verified sequences, and some uncertainty about assembly accuracy at the finest scale can remain. The evidence for bacterial-viral gene exchange, while compelling in its synteny signal, is correlational rather than a direct observation of DNA transfer; demonstrating the mechanism and direction of exchange in the laboratory would be a logical next step. Additionally, the functional enrichment findings identify categories of genes, such as surface adhesion proteins, whose roles in host adaptation are suggested by their location and variation patterns, but experimental tests of how specific island genes affect infectivity would be needed to confirm causation.

Even with these limitations, the work reshapes the picture of giant virus evolution in a way that resonates with decades of bacterial genomics. In bacteria, genomic islands acquired through horizontal gene transfer have long been recognized as keys to ecological innovation, encoding traits like antibiotic resistance and metabolic capabilities that allow lineages to colonize new niches.

Published in npj Viruses, the study examines the viruses in mice lacking the type I interferon receptor, known as IFNAR−/− mice. Type I interferons, including interferon-alpha and interferon-beta, are among the body’s earliest antiviral defenses. They activate hundreds of genes that restrict viral replication and help coordinate innate and adaptive immunity. By removing the receptor required for cells to respond to these signals, researchers create an animal model that is highly vulnerable to CCHFV infection and can support the development of severe disease.

The use of IFNAR−/− mice is particularly important for studying CCHFV because ordinary laboratory mice often resist infection or develop disease that does not reproduce the rapid progression seen in humans. The model does not replicate every feature of human Crimean-Congo hemorrhagic fever, but it allows researchers to compare viral isolates under controlled conditions. A side-by-side design also reduces the risk that differences in housing, timing, animal age or experimental handling will be mistaken for genuine differences between viruses.

Rohde and colleagues evaluated the two isolates using a combination of clinical and virological measurements. Such assessments typically include changes in body weight, temperature or activity, the onset and progression of disease signs, survival, viral RNA levels in blood and organs, and microscopic evidence of tissue injury. Together, these measurements distinguish between viruses that replicate efficiently, viruses that spread to particular organs, and viruses that cause severe disease even when their overall quantities are similar.

The comparison demonstrates why the isolate used in an animal experiment can be a decisive variable. Closely related CCHFV isolates may differ in the speed at which they establish infection, the extent to which they disseminate through the body and the severity of the resulting pathology. Differences can arise from mutations affecting viral replication, interactions with host proteins, the ability to counter innate immune responses or the balance between viral growth and inflammatory injury. These effects may not be apparent when experiments are conducted with only one strain.

CCHFV is an enveloped, negative-sense RNA virus in the family Nairoviridae. Its genome is divided into three segments that encode structural and non-structural proteins, including the nucleoprotein, the surface glycoprotein precursor and an RNA-dependent RNA polymerase. The segmented genome creates opportunities for genetic reassortment when different viruses infect the same cell, while the error-prone nature of RNA replication generates additional diversity. This biological flexibility helps explain why isolates collected in different regions or hosts can show distinct phenotypes in laboratory systems.

The study’s findings have consequences beyond the immediate comparison. Animal models are widely used to evaluate vaccines, antibody treatments, antiviral compounds and supportive-care strategies. If two isolates produce different disease trajectories in the same mouse background, a treatment that appears effective against one virus may not perform identically against another. Conversely, a model based on an especially aggressive isolate could make an intervention appear less effective than it would be against a broader range of circulating viruses. Careful strain selection and transparent reporting are therefore essential for reproducible CCHFV research.

The work also highlights the limitations of relying on a single laboratory model. IFNAR−/− mice lack a central component of antiviral immunity, and their response to infection cannot be directly equated with the response of people, whose disease is influenced by age, genetics, prior immune activation, coagulation pathways and other factors. The model is nevertheless valuable because it provides a consistent framework for comparing viruses and identifying mechanisms that can later be tested in more complex systems, including immune-competent animals, organoid cultures and human clinical samples.

For public-health researchers, the study reinforces the importance of treating CCHFV as a genetically diverse threat rather than as one standardized pathogen. The virus is maintained in nature through cycles involving ticks and animal hosts, and human infections occur across a broad geographic range. Surveillance programs that sequence viruses and link genetic data with clinical information may help determine whether particular viral lineages are associated with altered transmissibility or disease severity. The side-by-side approach used in this study offers a practical foundation for connecting viral genotype with biological behavior.

By placing two isolates under identical experimental conditions, the researchers provide a clearer framework for interpreting virulence studies and for designing future countermeasure trials. The broader message is that model systems are only as informative as the viral strains selected for them. As CCHFV research expands, comparisons across multiple isolates, host backgrounds and immune conditions will be critical for identifying results that are truly generalizable. The study therefore contributes not only to understanding these two viruses, but also to a more rigorous strategy for investigating one of the world’s most serious tick-borne viral diseases.

Subject of Research: Comparative pathogenicity and disease biology of two Crimean-Congo hemorrhagic fever virus isolates in IFNAR−/− mice.

Article Title: Side-by-side evaluation of two Crimean-Congo hemorrhagic fever virus isolates in IFNAR−/− mice.

Article References: Rohde, C., Werner, AD., Gellhorn Serra, M. et al. Side-by-side evaluation of two Crimean-Congo hemorrhagic fever virus isolates in IFNAR^−/− mice. npj Viruses 4, 35 (2026). https://doi.org/10.1038/s44298-026-00216-2

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s44298-026-00216-2

Keywords: Crimean-Congo hemorrhagic fever virus, CCHFV, IFNAR−/− mice, viral isolates, viral pathogenesis, animal models, interferon signaling, hemorrhagic fever, antiviral research.