The starting point for the study was CD388, a long-acting drug–Fc conjugate that has recently emerged as a new class of antiviral against both influenza A and influenza B. CD388 links a small-molecule inhibitor to the crystallizable fragment of an antibody, extending the drug’s residence time in the body while retaining its ability to bind the viral target. What remained unclear, however, was how exactly the Fc conjugation reshapes the function of the small molecule it carries. The researchers set out to answer that question systematically, and in doing so they uncovered a mechanism that had not been appreciated before.

To do this, the team built a platform based on click chemistry, a family of highly reliable bioconjugation reactions that allow molecular building blocks to be joined in a modular, almost Lego-like fashion. The platform makes it possible to flexibly combine different valencies, meaning different numbers of drug copies, with different conjugated drugs on the same Fc scaffold. Using the neuraminidase inhibitor zanamivir as a model drug, the researchers prepared mono-, di- and trivalent versions, each carrying one, two or three zanamivir molecules attached to the Fc fragment. This controlled series allowed them to interrogate, for the first time in a rigorous way, how valency alone governs antiviral behavior.

The results were striking. The trivalent zanamivir–Fc conjugate, designated AVC-Z3, exhibited higher antiviral activity than the mono- and divalent versions. But the more important discovery was mechanistic. Rather than merely blocking the enzymatic activity of viral neuraminidase more strongly, AVC-Z3 did something physically dramatic: it induced inter-virion aggregation. Because each conjugate carries three zanamivir ligands, a single AVC-Z3 molecule can bind neuraminidase proteins on more than one virus particle at the same time, cross-linking the virions into aggregates. The Fc fragment, with its own multivalent binding capacity, further supports this clustering effect.

This cross-linking produces a dual-stage inhibition that represents a genuine mechanistic shift. In the traditional view, neuraminidase inhibitors such as zanamivir and oseltamivir work by suppressing the enzyme that influenza uses to cleave sialic acid receptors, a step essential for newly formed virions to detach from an infected cell and spread. With AVC-Z3, the blockade of viral entry is coupled to the restriction of viral release. Virions that are cross-linked into aggregates cannot move freely, attach efficiently to new cells, or escape from the surfaces where they are produced. In effect, the drug no longer just disables a viral enzyme; it physically immobilizes the virus particles themselves.

The implications of this shift are considerable. A mechanism based on physical cross-linking is less dependent on the precise chemistry of the neuraminidase active site, which is where most resistance mutations to conventional inhibitors arise. By converting a single-target enzyme inhibitor into a multivalent cross-linking agent, the platform could in principle make it harder for the virus to escape through the usual mutational routes. The authors describe ligand valency as a determinant of antiviral mechanism, a framing that elevates valency from a simple potency knob to a design parameter that controls what kind of antiviral activity a conjugate delivers.

Encouraged by the zanamivir results, the team applied the same logic to a second neuraminidase inhibitor, peramivir. They synthesized a trivalent peramivir–Fc conjugate, AVC-P3, and found that it showed stronger antiviral activity than AVC-Z3 in vitro. Just as importantly, AVC-P3 displayed a much longer half-life in vivo than AVC-Z3 and than CD388 carrying the YTE mutation, an Fc engineering change often used to extend antibody half-life by improving binding to the neonatal Fc receptor. The combination of high potency and long persistence is exactly what is wanted in a long-acting prophylactic or therapeutic agent, particularly for vulnerable populations or outbreak settings where dosing logistics are difficult.

The in vivo performance of AVC-P3 was tested in mouse models of infection with four different influenza challenges: H1N1, H3N2, H5N1 and influenza B. Across all of these models, the efficacy of AVC-P3 surpassed that of the other AVCs, of baloxavir, a cap-dependent endonuclease inhibitor that represents one of the newest classes of influenza drugs, and of CD388 itself. Covering both influenza A subtypes, including the highly pathogenic H5N1, and influenza B in a single agent speaks to the breadth that the neuraminidase-targeting, Fc-conjugated approach can offer. H5N1 remains a pandemic concern given its ongoing circulation in birds and sporadic human infections, so a long-acting agent active against it has obvious strategic value.

The study also carries a broader message for antiviral drug design. Small-molecule antivirals and biologics have traditionally occupied separate worlds, with small molecules prized for potency and manufacturability and antibodies valued for specificity and long half-life. Drug–Fc conjugates blur that boundary, and the present work demonstrates that the design space within that hybrid class is richer than assumed. By making valency programmable, the click-chemistry platform turns the number of drug copies per Fc into an experimentally tunable variable, allowing researchers to map how valency reshapes mechanism, potency and pharmacokinetics. The finding that trivalency unlocks cross-linking suggests that other viral targets with dense surface display of ligands, such as hemagglutinin or surface glycoproteins of unrelated viruses, might be amenable to the same strategy.

Caveats remain, as they always do at this stage. The efficacy data come from mouse models, and the pharmacokinetic advantages observed in animals will need to be confirmed in higher species before the approach can be considered for clinical development. Questions about manufacturing consistency, immunogenicity of the conjugates and the behavior of virion aggregates in the complex environment of the human respiratory tract will all require attention. Nevertheless, the study defines a clear principle: ligand valency is not merely a matter of avidity but a switch that can convert a conventional enzyme inhibitor into a physically immobilizing, dual-stage antiviral. If that principle holds up through further development, the multivalent conjugate platform could offer a durable new line of defense against a virus that has repeatedly shown its capacity to surprise.

Subject of Research: Multivalent drug–Fc conjugates that cross-link influenza virions to enhance antiviral activity

Article Title: Multivalent antiviral conjugates enhance activity against Influenza via cross-linking viruses

Article References: Feng, X., Yuan, L., Zhao, L., Liu, L., Xie, F., Zhou, M., Cheng, G., Guan, Y., Xia, N., Cheng, T., Zhong, W., Xiao, D., Zhou, X., & Li, S. (2026). Multivalent antiviral conjugates enhance activity against Influenza via cross-linking viruses. Nature Communications. https://doi.org/10.1038/s41467-026-78126-3

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78126-3

Keywords: influenza, antivirals, drug–Fc conjugates, multivalency, neuraminidase inhibitors, zanamivir, peramivir, CD388, cross-linking, viral aggregation, H5N1, drug development

The review frames viral emergence as a cascade of linked molecular gates rather than a single event. Viral surface proteins determine attachment and entry, but compatibility between viral glycoproteins and host receptors does not guarantee productive infection. Intracellular host factors, polymerase activity, codon usage, replication organelles, interferon sensitivity and immune evasion proteins collectively determine whether progeny virus can be produced and transmitted. For SARS-CoV-2, the convergence of ACE2 engagement, spike priming by TMPRSS2, furin-site biology, replication-complex efficiency and immune evasion created a biologically favourable foundation for pandemic expansion. The same logic explains why a virus may be detected frequently at the animal-human interface yet fail to establish sustained human transmission, whereas another virus with modest ecological exposure may spread explosively once critical within-host barriers are crossed.

At the family level, the review identifies distinct molecular architectures that shape emergence risk. Coronaviruses combine large RNA genomes, recombination, broad mammalian reservoirs and spike-mediated receptor flexibility. Orthomyxoviruses such as influenza A possess segmented genomes that permit reassortment, enabling rapid antigenic and host-range shifts. Filoviruses and paramyxoviruses have non-segmented negative-sense RNA genomes, but their glycoproteins and immune-antagonist proteins facilitate severe disease in susceptible hosts. Flaviviruses and alphaviruses are shaped by vector ecology, climate-sensitive transmission and envelope-protein adaptation. Orthopoxviruses demonstrate that large DNA viruses can still undergo epidemiologically meaningful evolution under transmission pressure, as mpox has shown under new contact networks. This family-level view is not merely taxonomic; it provides a practical map for prototype-pathogen research, in which studying one coronavirus polymerase, protease or spike architecture can inform assay development and therapeutic design across related coronaviruses.

Entry is the first molecular gate through which an emerging virus must pass. Enveloped viruses use surface glycoproteins to bind receptors or attachment factors and then undergo fusion at the plasma membrane or within endosomal compartments. For henipaviruses, ephrin-B2 and ephrin-B3 receptor usage helps explain endothelial and neuronal tropism. For filoviruses, macropinocytosis, glycoprotein processing and NPC1-dependent endosomal entry are central determinants of permissiveness. However, tropism is also determined by host proteases, attachment factors, glycan presentation and innate-state context. A receptor may be expressed in a tissue yet fail to support efficient infection if the proteolytic environment or replication cofactors are absent. Conversely, viruses can exploit multiple entry routes or evolve altered receptor-binding surfaces that expand their host range. This explains why entry studies should be integrated with organoid models, primary cells, pseudovirus assays and animal-interface surveillance rather than being interpreted from receptor-binding data alone.

Replication machinery constitutes a second determinant of emergence. RNA-dependent RNA polymerases, viral proteases, nucleocapsid proteins and replication organelles determine not only viral yield but also mutation rate, recombination potential and sensitivity to nucleoside analogues. Coronaviruses are unusual among RNA viruses because their replication complex includes proofreading activity, enabling large genomes while still permitting adaptive diversity. This balance of fidelity and flexibility is one reason SARS-CoV-2 could accumulate meaningful mutations without losing genome integrity. Assembly mechanisms differ across viral families but often expose vulnerabilities: coronavirus assembly occurs at ER-Golgi intermediate compartments through coordinated interactions among M, E, S and N proteins; influenza assembly requires packaging of eight genome segments, creating opportunities for reassortment but also imposing strict compatibility constraints; filovirus budding depends heavily on matrix protein VP40 and host ESCRT machinery. Host-factor screens and proteomics have expanded the field from single viral proteins to systems-level replication networks. A SARS-CoV-2 protein-interaction map identified host pathways that could be repurposing targets, while CRISPR screens revealed TMEM41B and other host factors as pan-coronavirus dependencies.

Innate immunity imposes one of the strongest barriers to emergence, and successful emerging viruses frequently encode proteins that mask pathogen-associated molecular patterns, block adaptor proteins, degrade signalling molecules, suppress STAT activation, remodel stress responses or antagonise antigen presentation. These mechanisms are not incidental virulence accessories; they are often prerequisites for efficient replication in a new host. Convergent immune-evasion logic can be seen across unrelated viruses: influenza A NS1 inhibits antiviral RNA sensing and interferon responses; Ebola virus VP35 suppresses RIG-I-like receptor signalling while VP24 interferes with interferon-induced nuclear import; coronaviruses deploy several antagonists, including nsp1-mediated host translation suppression and accessory proteins that disrupt interferon signalling; henipavirus V, W and C proteins impair interferon production or response; and orthopoxviruses encode multiple soluble and intracellular immunomodulators. Pathogenesis results from both direct viral injury and dysregulated host response, with severe disease reflecting tropism for endothelial, respiratory, neurological or immune cells alongside excessive inflammation, vascular leakage, coagulopathy or immunopathology.

Evolutionary dynamics determine whether control measures remain effective over time. SARS-CoV-2 demonstrated that variants can accumulate spike mutations that alter transmissibility, antigenicity and neutralisation while maintaining receptor engagement. Omicron showed that immune escape can become a dominant population-level advantage in settings of widespread infection- or vaccine-derived immunity, yet escape must remain compatible with entry and replication fitness. Influenza viruses exemplify the risk of segmented genomes: reassortment between avian, swine and human viruses can generate new constellations of internal genes and surface antigens, while point mutations in HA, PB2 and other segments influence receptor preference and mammalian adaptation. Highly pathogenic H5N1 clade 2.3.4.4b has increased concern because infection in mammals creates additional opportunities for adaptation, even though sustained human-to-human transmission has not been established. The review argues that genomic surveillance should evaluate variants as phenotypic hypotheses rather than as lists of mutations, because a mutation in a receptor-binding domain, polymerase, protease, matrix protein or immune antagonist may be important only in a specific genomic background.

Rapid diagnostics are central to containment, but performance is conditional. RT-qPCR is sensitive and scalable, yet primer mismatch, sampling time, specimen quality and infrastructure can reduce effectiveness. Programmable CRISPR assays can shorten redesign cycles, whereas metagenomic next-generation sequencing offers agnostic detection when the causative agent is unknown, divergent or coinfecting; neither replaces prospective clinical validation, contamination control or external quality assurance. Outbreak genomics has transformed reconstruction for Ebola, Zika, yellow fever, SARS-CoV-2 and mpox by linking sequence data to transmission dynamics, revealing introductions, hidden transmission, superspreading and international dissemination. The value of genomic data depends on representative sampling, metadata quality, rapid deposition, analytic capacity and equitable participation by affected regions. On the countermeasure side, the COVID-19 pandemic demonstrated both the power and fragility of platform approaches: mRNA, adenoviral-vector and protein-based vaccines can be developed quickly when antigen design, manufacturing capacity, clinical trial networks and regulatory pathways are prepared in advance, but antigenic evolution can reduce neutralisation, requiring booster strategies, updated antigens or broader immunogen design. Direct-acting antivirals are most useful when they target conserved enzymatic functions, have acceptable safety profiles and can be delivered early in infection, as remdesivir and nirmatrelvir-ritonavir showed in selected clinical settings, while molnupiravir demonstrated more modest clinical utility and raises different resistance and mutagenesis considerations.

For filoviruses, the PALM trial established that monoclonal antibody therapies can improve outcomes for Ebola virus disease when delivered with supportive care, shifting the field from empiricism towards evidence-based outbreak therapeutics. For mpox, tecovirimat has been widely used because of orthopoxvirus rationale and animal-rule approval, but recent human trial evidence for clade I mpox did not demonstrate faster lesion resolution, highlighting the need for outbreak-embedded randomised trials rather than reliance on plausible mechanism alone. The evidence base remains structurally uneven: SARS-CoV-2 offers abundant structural, clinical and population data, while Lassa virus, Crimean-Congo haemorrhagic fever virus, hantaviruses and Rift Valley fever virus remain comparatively under-characterised. This asymmetry can make publication volume appear equivalent to biological importance. Lassa-virus diversity and immune evasion complicate vaccine generalisation; CCHFV combines tick ecology with incomplete correlates of protection; hantavirus pathogenesis centres on endothelial dysfunction despite limited specific therapy; and RVFV risk spans livestock, vectors and human disease. Conflicting evidence commonly arises from model choice, because receptor overexpression can exaggerate permissiveness, pseudoviruses isolate entry while omitting replication and immune antagonism, organoids improve tissue relevance but incompletely reproduce systemic immunity, and animal models may not capture human transmission.

The review proposes a deliberately modular sequence-to-decision standard in which a novel signal passes through five stages: detection and metadata review; family and structural triage; functional testing of entry, replication and innate responses; explicit evidence grading; and a proportionate decision about surveillance, assay updating, countermeasure screening or clinical research. Artificial intelligence can accelerate protein-structure prediction, variant prioritisation, epitope screening and surveillance anomaly detection, yet confidence scores do not establish biological activity, host range or clinical relevance, and training-data bias, poor representation of disordered complexes and limited prospective validation require that AI outputs remain auditable hypotheses tested against experimental and epidemiological data. Across virus families, three implementation priorities recur: assay standardisation so that variant phenotyping does not depend on isolated laboratories using incomparable methods; predefined connections between molecular evidence and public-health thresholds, such as repeated detection of a high-risk virus in a new host or convergent mutations at receptor-binding sites; and sustainable capacity in genomic sequencing, biobanking, bioinformatics, clinical trials and vaccine delivery that must exist before a crisis, especially in regions where spillover and neglected outbreaks are most likely to occur. Molecular systems analysis cannot predict emergence with certainty, but evidence-graded integration can shorten the interval between a credible signal and a proportionate response while reducing both false reassurance and unjustified alarm.

Subject of Research: Molecular mechanisms of viral emergence, immune evasion and cross-species adaptation across major virus families, with a sequence-to-decision framework for precision countermeasures

Article Title: Molecular systems of emerging viruses: replication, host immune evasion, cross-species adaptation and precision countermeasures

Article References: Hejran, A. B., Niazi, P., & Sangin, M. Y. (2026). Molecular systems of emerging viruses: replication, host immune evasion, cross-species adaptation and precision countermeasures. Frontiers in Virology, 6, Article 1926007. https://doi.org/10.3389/fviro.2026.1926007

Image Credits: AI Generated

DOI: 10.3389/fviro.2026.1926007

Keywords: emerging viruses, systems virology, SARS-CoV-2, immune evasion, cross-species transmission, genomic surveillance, CRISPR diagnostics, vaccine platforms, antivirals, One Health, zoonotic spillover, viral replication

The scale of the problem justifies the effort. The most recent estimates cited by the authors suggest that roughly six percent of United States adults and half a percent of children and adolescents are still living with new or persistently worsened symptoms that began shortly after an acute SARS-CoV-2 infection. For clinical purposes, the authors define Long COVID as new or worsened symptoms developing after confirmed or suspected infection and persisting for more than three months, a deliberately broad formulation that sidesteps the stricter laboratory-confirmation requirements of research definitions. The dominant symptoms cluster around fatigue, post-exertional malaise, cognitive impairment often called brain fog, and orthostatic intolerance, the lightheadedness and palpitations that appear or worsen when a patient stands upright.

What makes the review distinctive is its use of clinical vignettes that mirror how patients actually describe their lives. In one case, a 32-year-old woman with a history of anxiety reports four months of fatigue that is qualitatively different from anything she has felt before. After minimal activity, she describes crashing, as though wearing a lead coat, with profound exhaustion and difficulty thinking that can persist for hours to days. Even preparing meals or answering emails has become a hurdle, and she has cut back her work hours. Her symptoms began about a month after a mild COVID-19 infection, a temporal link that might easily be missed if a clinician never asks about preceding illness.

A second vignette captures the dysautonomia pattern: a previously athletic 19-year-old develops palpitations, lightheadedness, and near-fainting on standing, symptoms that improve when she lies down. She recalls an upper respiratory illness two weeks before onset, and although she never tested herself, multiple family members had confirmed COVID-19 at the time. A third describes a 45-year-old aerospace engineer whose forgetfulness, word-finding difficulty, and slowed processing speed have cost him deadlines at work, all despite a normal basic laboratory evaluation. Together, the cases illustrate the review’s central message: Long COVID frequently announces itself through recognizable symptom clusters, and related conditions such as myalgic encephalomyelitis/chronic fatigue syndrome and postural orthostatic tachycardia syndrome often ride alongside it.

Beneath the clinical patterns, the authors propose a three-part conceptual framework for what SARS-CoV-2 actually leaves behind. The first category is a hit-and-run phenomenon, in which acute infection causes direct end-organ tissue injury with lasting consequences, such as pulmonary fibrosis or vascular events. The second is a stressor effect, in which the infection reduces physiological reserve and unmasks or accelerates chronic diseases like diabetes, hypertension, or rheumatoid arthritis in the following weeks and months. The third and most scientifically contested category is ongoing biological dysregulation, encompassing persistent viral reservoirs, immune dysregulation, gut dysbiosis, mitochondrial dysfunction, and microvascular clotting, processes repeatedly identified in basic science studies but largely invisible to current clinical diagnostic tools. Management of the first two categories follows standard practice, the authors note, while the third currently rests on rehabilitation and supportive care.

Diagnostically, the review resists exotic testing. Because no clinically available biomarker can confirm or exclude Long COVID, and severe symptoms can coexist with entirely normal conventional results, the workup serves mainly to rule out competing diagnoses and addressable sequelae. The authors recommend a focused physical examination guided by the patient’s symptoms, with particular attention to orthostatic vital signs, which are often skipped and frequently abnormal in this population. Targeted laboratory and imaging studies, including complete blood count, metabolic panel, thyroid function, and inflammatory markers, follow the same logic a primary care clinician would apply to any undifferentiated complaint. A validated ten-item questionnaire for post-exertional malaise, the DSQ-PEM, can help document a symptom that patients often struggle to articulate.

On treatment, the message is sobering but constructive: there is no cure, and no FDA-approved medication, yet validation and education alone can meaningfully improve quality of life. The authors emphasize that mental health conditions are not the cause of Long COVID, but the illness frequently produces new or worsened depression and anxiety, partly because patients do not feel believed by clinicians or loved ones. Nonpharmacologic strategies form the backbone of care. Activity pacing, energy conservation, and task modification, such as spreading grocery shopping and meal preparation across separate days or using a scooter and a seated chopping station, help patients live within their energy envelope. Referral to therapists experienced with chronic fatiguing illnesses, typically in neurorehabilitation rather than orthopedic settings, can be transformative: physical therapy for fatigue and activity intolerance, occupational therapy for self-care and job demands, and speech therapy for attention, memory, and breathing discomfort despite normal pulmonary tests.

One warning stands out with unusual force. While some patients tolerate gradual, supervised exercise progression, graded exertion must be approached with extreme caution in patients with post-exertional malaise. The authors cite evidence that even seemingly minor overexertion can trigger delayed symptom flares and has been associated with amyloid deposition, reduced mitochondrial enzyme activity, and signs of severe skeletal muscle damage in affected patients. For those with orthostatic intolerance, simpler measures carry real benefit: increasing daily fluid intake to 2.5 to 3.5 liters and sodium to 2.8 to 4 grams, equivalent to 7 to 10 grams of salt, alongside compression garments such as waist-high stockings or an abdominal binder, provided no contraindications exist.

Pharmacologic management remains symptom-based, borrowing indications from other contexts: beta blockers or ivabradine for POTS, stimulants for attention-dominant brain fog, gabapentinoids for neuropathic pain, and SSRIs for new or worsened mood symptoms. Two off-label options receive special mention. Low-dose naltrexone, an opioid receptor blocker, has been associated in some patients with improved pain, fatigue, and cognition, and a 2026 systematic review and meta-analysis examined the evidence base. Patients with suspected mast cell activation syndrome triggered by Long COVID sometimes report relief from combined H1 and H2 antihistamines. The authors also address the longer arc of illness: symptoms persist beyond one year in roughly a quarter of patients and beyond two years in about ten percent, making documentation of functional impact, workplace accommodation letters, and disability paperwork an essential, billable-worthy part of primary care.

Perhaps the most practical contribution is a set of principles for delivering all of this within a standard visit. The authors advise screening for preceding illness and symptom duration in the history, asking which symptom is most bothersome right now, and tackling only one or two symptoms per visit while leveraging longitudinal follow-up. Medication trials should be stepwise, one intervention at a time, starting low and titrating slowly, with de-prescribing of drugs that may worsen fatigue or cognition. Written after-visit summaries matter disproportionately for patients with cognitive symptoms, and telehealth or first-appointment scheduling can spare fatigued patients the exertion of a waiting room. On prevention, the review is equally concrete: vaccination shows a dose-dependent reduction in Long COVID risk, early outpatient antivirals offer modest protection, with nirmatrelvir/ritonavir associated with a 26 percent relative risk reduction, molnupiravir 14 percent, and ensitrelvir 45 percent, while metformin was linked to a 41 percent reduction in incidence in a randomized trial. Indoor air filtration, N95 or KN95 masking, and isolation during infectivity round out the defensive toolkit. Until reliable biomarkers and curative therapies arrive, the authors argue, the frontline of Long COVID medicine will remain the primary care office, and clinicians equipped with a structured approach can already change what the illness means for the patients living with it.

Subject of Research: Clinical evaluation and management of Long COVID in primary care

Article Title: Long COVID in Primary Care: A Practical Approach to Evaluation and Management

Article References: Chopra, A., Bonilla, H. F., & Cheng, A. L. (2026). Long COVID in Primary Care: A Practical Approach to Evaluation and Management. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10901-w

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10901-w

Keywords: Long COVID, primary care, SARS-CoV-2, post-exertional malaise, dysautonomia, POTS, brain fog, ME/CFS, diagnosis, rehabilitation, antivirals, prevention

The rationale for the combination rests on the complementary profiles of the two drugs. Azvudine, known chemically as FNC, is a nucleoside reverse transcriptase inhibitor approved in China for HIV treatment and later authorized for COVID-19. As an analogue, it is incorporated by viral polymerases and terminates RNA synthesis, acting directly on the SARS-CoV-2 RNA-dependent RNA polymerase. Carrimycin, by contrast, is a macrolide derived from the engineered antibiotic compound 4″-O-isovalerylspiramycin. Macrolides are best known as antibacterials, but several members of the class display immunomodulatory and anti-inflammatory properties, and Carrimycin has previously shown antiviral activity against coronaviruses. Pairing a direct-acting polymerase inhibitor with an agent that modulates host responses is an appealing strategy precisely because evolving SARS-CoV-2 variants have begun to erode the effectiveness of replication-targeted monotherapy.

To test the concept, the team turned to rhesus macaques, which remain among the most faithful animal models of human COVID-19 because they develop viral pneumonia, radiographic abnormalities, and immune responses that closely mirror clinical disease. Animals were inoculated intratracheally with SARS-CoV-2 and then treated orally for seven days with Carrimycin at 9 milligrams per kilogram, Azvudine at 0.07 milligrams per kilogram, or the two drugs together. The researchers then tracked viral loads, lung pathology, immune responses, and gut microbiota across the treatment groups, conducting the work in an animal biosafety level 3 facility under protocols approved by the institutional animal care committee.

The antiviral results were clear-cut. Both monotherapies significantly inhibited SARS-CoV-2 replication, but the combination produced additive effects, driving viral loads down further in both throat swabs and lung tissue. That pattern matters because viral burden in the lower respiratory tract correlates with disease severity and onward transmission risk. Notably, the authors reported no significant side effects in the treated animals, an important consideration for a combination intended for clinical use, since drug-drug interactions and cumulative toxicity often limit real-world combination regimens.

Lung protection was assessed by chest imaging and histopathology, and here the combination again outperformed either drug alone. Monotherapy reduced ground-glass opacities, the hazy areas on chest scans that reflect fluid and inflammatory infiltration of lung tissue, and mitigated focal interstitial pneumonia, the inflammation of the tissue between air sacs that characterizes severe COVID-19. The combination treatment demonstrated superior efficacy on both measures, suggesting that blunting viral replication while simultaneously damping host inflammation preserves lung architecture more effectively than either intervention alone.

To understand how the combination reshaped the lung environment, the researchers performed transcriptomic profiling and immunofluorescence staining of lung tissue. Gene expression analysis revealed enhanced immunomodulation under combination therapy, including suppression of neutrophil degranulation, a process in which activated neutrophils release enzymes and antimicrobial proteins that, when excessive, damage surrounding tissue and contribute to acute lung injury. The combination also increased populations of CD19-positive B cells and CD3-positive T cells, the cornerstones of adaptive humoral and cellular immunity, while decreasing CD68-positive macrophages, whose accumulation in the lung is associated with sustained inflammation.

At the molecular level, combination therapy downregulated pro-inflammatory and apoptotic markers, pointing to reduced inflammatory signaling and less programmed cell death in infected lung tissue. Taken together, these findings suggest the combination does more than lower viral numbers; it shifts the lung immune landscape away from the maladaptive innate hyperinflammation seen in severe COVID-19 and toward a more effective adaptive response. That distinction is clinically significant, because mortality in severe COVID-19 is driven as much by immunopathology as by the virus itself, and therapies that address both axes have been sought since the beginning of the pandemic.

The study also uncovered a difference between the two drugs that monotherapy comparisons alone would have missed. Carrimycin, but not Azvudine, increased gut bacteria that produce short-chain fatty acids, metabolites with well-documented anti-inflammatory and immune-regulatory effects, and decreased harmful pathogens previously associated with COVID-19 progression. The gut-lung axis, the bidirectional communication between intestinal microbes and respiratory immunity, has emerged as a contributor to COVID-19 outcomes, and dysbiosis has been linked to worse disease. A drug that simultaneously acts on the lung and restores a favorable microbiome composition could therefore offer benefits beyond direct antiviral activity, potentially supporting recovery and reducing secondary complications.

The authors conclude that the Carrimycin-Azvudine combination showed superior efficacy in suppressing viral replication, alleviating lung injury, and correcting immune and inflammatory dysregulation, with Carrimycin providing additional gut microbiota benefits. They position the pairing as a promising therapeutic strategy against COVID-19 and its complications, particularly in a landscape where ongoing viral mutation, emerging resistance mutations, inadequate immunomodulation, and suboptimal organ protection limit existing antiviral options. The work was supported by the National Natural Science Foundation of China, the National Key Research and Development Program of China, and the CAMS Innovation Fund for Medical Sciences.

As with any preclinical animal study, the path from macaques to patients requires clinical validation, including pharmacokinetic confirmation of the doses used, human safety data for the combination, and efficacy trials in infected populations. Nevertheless, the study offers a mechanistically grounded template for next-generation COVID-19 therapy: a direct-acting antiviral paired with an immunomodulatory and microbiome-friendly partner, evaluated with the full toolkit of modern virology, imaging, transcriptomics, and metagenomics. If the additive antiviral and lung-protective effects observed here translate to humans, the combination could address precisely the shortcomings, resistance, inflammation, and organ damage, that have constrained single-drug approaches throughout the pandemic.

Subject of Research: Combination antiviral therapy with Carrimycin and Azvudine against SARS-CoV-2 infection in rhesus macaques

Article Title: Better antiviral and lung-protective effects of Carrimycin in combination with Azvudine in rhesus macaques infected with SARS-CoV-2

Article References: Li, H., Liu, Y., Xue, J., Li, J., Xu, J., Dong, B., Wang, T., Peng, X., Che, Y., Jiang, J., & Peng, Z. (2026). Better antiviral and lung-protective effects of Carrimycin in combination with Azvudine in rhesus macaques infected with SARS-CoV-2. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-08933-1

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08933-1

Keywords: SARS-CoV-2, Carrimycin, Azvudine, combination therapy, rhesus macaques, lung injury, immunoinflammatory regulation, gut microbiota, antivirals, macrolides, COVID-19, translational medicine