Translational insights into Long-COVID: evaluation of preclinical animal models along the lung-brain-immune axis with focus on Golden Syrian Hamsters Authors: Narendra Kumar, Navya Khandavalli, Sean N. Avedissian, Hitendra S. Chand, Arpan Acharya, Siddappa N. Byrareddy Published: 2026 ABSTRACT Long-COVID, also referred to as post-acute sequelae of COVID-19 (PASC), is a heterogeneous disorder encompassing more than 200 reported symptoms that commonly affect the respiratory and nervous systems. Emerging clinical evidence indicates that unresolved lung inflammation, vascular injury, and immune dysregulation drive sustained neuroinflammation and impaired neurocognitive function in Long-COVID patients. Given the ethical and logistical constraints of human studies, biologically relevant animal models are essential for understanding the mechanisms and for evaluating therapeutic strategies against Long-COVID. In this review, we synthesize current evidence from preclinical animal models of Long-COVID, with a particular emphasis on the Golden Syrian Hamsters. Golden Syrian Hamsters are naturally susceptible to SARS-CoV-2 infection without the need for genetic modification and recapitulate key features of human disease, including robust viral replication, pulmonary pathology, and inflammatory response during acute infection. Importantly, accumulating evidence demonstrates that Golden Syrian Hamsters develop persistent post-acute abnormalities along the lung–brain-immune axis, including impaired alveolar repair, fibrotic lung remodeling, neuroinflammation, viral or antigen persistence, and behavioral alterations that parallel core features of Long-COVID. We compare the strengths and limitations of Golden Syrian Hamsters with other commonly used pre-clinical animal models including mice, and non-human primates, highlighting differences in translational relevance, feasibility, and ability to model chronic lung-brain-immune axis dysfunction. While there are limitations, particularly regarding limited availability of immunological reagents and validated cognitive and behavioral assays, the Golden Syrian Hamsters offers a balanced and accessible platform for mechanistic studies of PASC. Overall, this review positions Golden Syrian Hamster as a robust translational model for investigating lung–brain–immune axis pathology in Long-COVID and for advancing the development of targeted therapeutic interventions. Background The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has resulted in substantial global morbidity and mortality with approximately 779 million confirmed cases and more than 7.1 million deaths worldwide [1, 2]. Since its emergence in Wuhan, China, SARS-CoV-2 has continued to evolve, with new viral variants altering transmission dynamics, and host immune responses, alongside persistent and heterogeneous long-term health consequences, underscoring ongoing gaps in our understanding of disease pathogenesis and recovery [3, 4]. These challenges highlight the continued need for mechanistic research to inform therapeutic strategies and vaccine development. A significant subset of individuals infected with SARS-CoV-2 develop persistent symptoms collectively termed Long-COVID or post-acute sequelae of COVID-19 (PASC). The World Health Organization (WHO) defines Long-COVID as the continuation or emergence of symptoms three months after the initial infection, lasting for at least two months and not explained by alternative diagnoses [5, 6]. Clinical manifestations are diverse and frequently multisystemic, commonly including chest pain, palpitations, cough, dyspnea, fatigue, sleep disturbances, gastrointestinal symptoms, and autonomic complaints [7]. Neuropsychiatric features such as anxiety, depression, brain fog, headache, and persistent fatigue are also prominent and often fluctuate over time [8]. These symptoms can progress to chronic conditions, including cardiovascular and cerebrovascular disease, type II diabetes, myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), and postural orthostatic tachycardia syndrome (POTS) [9]. Epidemiological estimates from the American Medical Association suggest that around 10–30% of individuals diagnosed with COVID-19 develop Long-COVID, with higher prevalence among those hospitalized during acute SARS-CoV-2 infection [10]. Emerging evidence suggests that many hallmarks of Long-COVID arise from coordinated, multi-organ dysfunction rather than isolated, organ-specific pathology. For example, persistent pulmonary injury is characterized by impaired gas exchange, exertional dyspnea, and incomplete alveolar repair, which frequently coexists with neurological symptoms such as fatigue, cognitive dysfunction, and autonomic instability [11]. This clinical coupling supports the concept of a lung–brain–immune axis, in which chronic respiratory inflammation, vascular injury, and immune dysregulation propagate sustained neuroinflammatory and neurodegenerative processes [12, 13]. Therefore, understanding how unresolved lung pathology propagates signals that drives downstream brain dysfunction represents a critical unmet need in Long-COVID research, particularly for elucidating mechanisms linking persistent pulmonary inflammation, vascular injury, and neuroinflammatory and neurodegenerative processes. However, logistical and ethical constraints continue to significantly limit the translation of these clinical observations into mechanistic insight in humans. Access to tissues and the ability to carry out long-term, controlled interventions are necessary for analyzing the temporal and causal linkages underpinning lung-brain-immune interactions; such approaches are limited in clinical settings [14, 15]. From a neurological standpoint, restricted access to brain tissue during life, the impracticality of repeated cerebrospinal fluid sampling, and the inability to perform targeted experimental manipulations in patients hinder efforts to investigate whether CNS dysfunction arises from viral persistence, chronic immune activation, blood–brain barrier disruption, or a combination of these events. Similarly, investigating the progression of post-acute lung diseases requires longitudinal assessments using imaging, bronchoalveolar lavage, and pulmonary function testing approaches that are invasive, challenging to repeat over time, and ethically constrained in human subjects. Moreover, the inability to implement controlled immune modulation or staged therapeutic interventions in humans, precludes precise dissection of immune contributions, including the roles of specific cell populations and temporal dynamics of immune responses. Collectively, these limitations underscore the critical need for preclinical animal models to examine the integrated lung-brain-immune pathways that drive Long-COVID pathophysiology. Preclinical animal studies enable controlled investigation of SARS-CoV-2 transmission, replication dynamics, immune responses, endothelial injury, and long-term tissue remodeling, as well as evaluation of vaccines, antivirals, and other host directed interventions including anti-inflammatory therapies [16]. Among available models, Golden Syrian Hamsters are uniquely suited for SARS-CoV-2 research because they are naturally susceptible to infection without genetic manipulation, and recapitulate key clinical, pathological, and immunological features observed in humans [17–20]. Importantly, Golden Syrian Hamsters develop robust pulmonary pathology during acute infection and exhibit persistent post-acute lung remodeling, neuroinflammation, and behavioral changes, positioning them as a powerful model for studying Long-COVID linked pathophysiology along lung–brain and immune axis. Several host-related risk factors have been associated with increased susceptibility to Long-COVID, which include advanced age, biological sex, and pre-existing comorbidities such as asthma, diabetes, cardiovascular disease, chronic lung disease, and kidney disease [21, 22]. Understanding the molecular mechanisms of the COVID-19 disease pathogenesis may help in the effective management of Long-COVID, which is crucial for elucidating how these variables interact with viral persistence, immune dysregulation, and neuroinflammation, and is essential for identifying therapeutic targets and stratifying disease endotypes. In this context, the use of pre-clinical animal models provides an indispensable platform for interrogating COVID-19-associated immune dysregulation, vascular injury, gut-lung-brain interactions, autoimmunity, and chronic neuroinflammatory responses [7]. In this review, we provide a comprehensive overview of available animal models used to study Long-COVID, with a particular emphasis on the Golden Syrian Hamsters. We focus on how host-virus interactions across the lung-brain and immune axis contribute to persistent post-acute disease and highlight the translational relevance, strengths, and limitations of each model in advancing mechanistic understanding of Long-COVID. Methods of literature search This review was conducted using a structured search of peer-reviewed literature describing animal models of SARS-CoV-2 infection and Long-COVID or PASC. The literature search covered publications available up to December 10, 2025, and was performed using PubMed, Scopus, and Google Scholar databases. Search strategies combined controlled vocabulary (MeSH terms, where applicable) and free-text keywords related to SARS-CoV-2 infection, Long-COVID, and experimental animal models. Keywords included SARS-CoV-2, COVID-19, Long-COVID, PASC, Animal Model, Mouse, Rat, Hamster, Golden Syrian Hamster, Syrian Hamster, Ferret, Non-human Primate (NHP), Macaque, and Rhesus. For PubMed, the following Boolean strategy was used: (“Post-Acute COVID-19 Syndrome“[MeSH] OR “Long-COVID“[tiab] OR “Long-COVID“[tiab] OR “Post-Acute Sequelae of SARS-CoV-2“[tiab] OR PASC[tiab] OR “Post-acute COVID“[tiab] ) AND ( “COVID-19“[MeSH] OR “SARS-CoV-2“[MeSH] OR COVID-19[tiab] OR SARS-CoV-2[tiab] ) AND ( “Animal Models“[MeSH] OR “Disease Models, Animal“[MeSH] OR “Animals“[MeSH] OR animal*[tiab] ) AND ( “Mice“[MeSH] OR mouse[tiab] OR mice[tiab] OR “Rats“[MeSH] OR rat[tiab] OR rats[tiab] OR “Hamsters“[MeSH] OR hamster*[tiab] OR “Golden Syrian Hamster“[tiab] OR “Syrian Hamster“[tiab] OR “Ferrets“[MeSH] OR ferret*[tiab] OR “Macaca mulatta“[MeSH] OR macaque*[tiab] OR rhesus[tiab] OR “Primates, Non-Human“[MeSH] OR NHP[tiab] ). For Scopus, searches were conducted using the following query applied to titles, abstracts, and keywords: TITLE-ABS-KEY( (“long covid” OR “long-covid” OR PASC OR “post-acute sequelae of sars-cov-2” OR “post-acute covid” OR “chronic covid”) ) AND TITLE-ABS-KEY( (“COVID-19” OR “SARS-CoV-2”) ) AND TITLE-ABS-KEY( (“animal model*” OR “disease model*” OR animal*) ) AND TITLE-ABS-KEY( (mouse OR mice OR rat OR rats OR hamster* OR “Golden Syrian hamster” OR “Syrian hamster” OR ferret* OR macaque* OR rhesus OR “non-human primate*” OR NHP) ). Additional relevant studies were identified through manual screening of reference lists from selected articles. The following inclusion criteria were applied: (1) studies employing animal models of SARS-CoV-2 infection, such as transgenic, wild-derived, humanized, or mouse-adapted virus models; (2) evaluation of post-acute, persistent, or long-term outcomes following infection, defined as outcomes assessed at ≥ 28 days post-infection; (3) reporting of at least one feature associated with Long-COVID, such as pulmonary pathology, immune dysregulation, viral persistence, or multi-organ involvement; (4) primary research articles published in English. The following exclusion criteria were applied: (1) studies limited to using in vitro, ex vivo, or in silico systems; (2) studies reporting only acute-phase outcomes (< 28 days post-infection) without post-acute follow-up; and (3) non-primary literature, including review articles, editorials, or commentaries. Clinical features of Long-COVID patients Long-COVID or the PASC is being increasingly recognized as a multisystemic syndrome wherein symptoms persist or emerge weeks to months after the resolution of initial acute infection [7, 23]. The estimated prevalence of Long-COVID varies widely with an estimated frequency between 33% and 40%, depending on factors such as disease severity, hospitalization status, vaccination status, and the follow-up duration [24]. Longitudinal studies have identified several risk factors, including female sex, metabolic and cardiopulmonary comorbidities, elevated BMI, impaired acute immune responses, and latent viral reactivation [22, 25–27]. Long-COVID is now recognized as a syndrome associated with chronic immune dysregulation, viral or antigen persistence, microvascular injury, and is associated with increased risk of new-onset conditions, including cardiovascular disease, diabetes, myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), and dysautonomia [7, 27, 28]. Despite these advances, human studies remain largely observational and have not established causal relationships between underlying mechanisms and organ specific pathology [7, 29]. The critical unresolved question is which biological pathways act as primary versus secondary drivers of Long-COVID, and why only a subset of COVID-19 patients develops persistent symptoms. Equally important is identifying the interventions capable of preventing or reversing specific sequalae. Addressing these knowledge gaps requires experimental systems in which key variables, such as viral strains, host genetic background, sex, comorbidities, and timing of interventions, can be precisely controlled, conditions that are not feasible in human studies. The following subsections outline the clinical and pathological benchmark in Long-COVID patients identified across three key domains to evaluate each animal model in this review. Lung pathology in Long-COVID patients The most prevalent and clinically significant manifestation of Long-COVID is persistent pulmonary abnormalities, which affects about 57% of hospitalized patients even after 8-10-month post infection [30, 31]. Months after the initial infection, radiographic and histologic evidence shows poor alveolar repair, bronchiolization, and fibrosis with a significant percentage of survivors showing subpleural fibrosis and ground-glass opacities (GGO) on CT scans in humans [30, 32]. Further, Long‑COVID syndromes are frequently marked by impaired gas diffusion, exertional dyspnea, reduced exercise capacity, with radiographic or histologic evidence of incomplete alveolar repair, bronchiolization, and fibrosis months after acute infection [33–35]. The pathological substrate is the inability of alveolar epithelial progenitors, such as AT2-, club-, and basal-like-derived cells, to correctly differentiate into alveolar type 1 (AT1) cells. This results in the accumulation of CK8⁺ damage-associated transient progenitor (DATP) cells with a cell-cycle arrest phenotype, which hinders gas exchange, alveolar regeneration, and promotes progressive fibrosis [36, 37]. With the loss of native alveolar macrophages and their replacement by proinflammatory monocyte-derived macrophages that promote innate immune-dominant inflammation, macrophages are crucial in COVID-19 lung damage [38]. By interacting spatially with epithelial progenitor cells (such as DATP/alveolar differentiation intermediate cells (ADI)), these macrophages maintain a defective epithelial–immune interface, which prolongs lung damage and hinders repair. Fibroblast recruitment and the development of pulmonary fibrosis may be further aided by persistent activation of these macrophage populations during the subacute and Long-COVID [38]. Collectively, these structural and cellular alterations manifest clinically as persistent exertional dyspnea, reduced diffusion capacity, and exercise intolerance that may last for months to years [39]. Importantly, these chronic pulmonary abnormalities are increasingly recognized as a source of sustained systemic inflammation, with potential to propagate downstream effects along the lung-brain and immune axis through circulating immune mediators and neuroimmune signaling pathways, including vagal afferent mechanisms [40]. Studying post-acute lung disease in humans requires serial assessment using pulmonary function testing, imaging (CT scans), bronchoalveolar lavage analysis, approaches that are either invasive or minimal invasive, logistically challenging, and ethically constrained for repeated use. These limitations hinder longitudinal investigation of the molecular drivers of fibrosis and impede the evaluation of targeted interventions aimed at restoring alveolar repair in living patients. Brain pathology and neurocognitive consequences in Long-COVID patients The most prevalent and disabling manifestation of Long-COVID is neurological and neurocognitive symptoms, affecting approximately 13–27% of recovered individuals, with some studies reporting persistent neurological symptoms in up to 60% of patients [41, 42]. The most common presentations include cognitive impairment, anxiety, depression, fatigue, anosmia, sleep disturbances, and autonomic dysfunction [41, 42]. Neuroimaging studies have revealed evidence of accelerated brain aging, reduced grey matter volume in olfactory and hippocampal regions, microstructural changes, and hypometabolism in the brainstem/olfactory gyrus on positron emission tomography (PET) imaging [43–48]. Chronic neuroinflammation has been implicated as a key mechanism underlying these changes, supported by elevated biomarkers such as neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), and other glial activation markers detected in cerebrospinal fluid (CSF) and plasma of Long-COVID patients [49–52]. The COVID-19 anosmia, reported in up to 60% people during the acute phase is primarily attributed to infection of sustentacular cells within the olfactory epithelium rather than the direct viral invasion of olfactory sensory neurons. Since these cells supports neuronal function, their disruption can be linked with impaired smell without the direct viral infection of neurons, however long-term changes may involve chronic neuroinflammation [53, 54]. In addition, increasing evidence links the severity of cognitive impairment to blood-brain barrier (BBB) disruption, suggesting a vascular and neurological component to PASC that may occur independently of direct viral neuroinvasion [55]. Despite these advances, the relative contributions of the viral persistence, chronic immune activation, BBB disruption, and their interactions in driving CNS dysfunction remain unclear. Elucidating these mechanisms in humans is inherently limited by restricted access to brain tissue during life, the impracticality of serial CSF sampling and the inability to perform controlled, targeted interventions in the CNS. These constraints underscore the need for preclinical models to dissect the mechanisms underlying neurocognitive dysfunction in Long-COVID. Immune dysregulation in Long-COVID patients Although heterogenous, the immune landscape of Long-COVID is increasingly being defined, with distinct patterns of persistent immune dysregulation observed in affected individuals compared to fully recovered patients [56, 57]. Key immune signatures include exhaustion of CD4 + and CD8 + T cells, a reduction in central memory CD4 + T cells, expansion of non-classical and intermediate monocyte populations, persistent elevation of pro-inflammatory cytokines (such as IL-6, TNF-α, CXCL10, and IL-1β), and prolonged upregulation of interferon-stimulated genes (ISGs) reported months after apparent viral clearance [27, 56–59]. Autoantibody generation targeting G-protein coupled receptors has been implicated in a subset of patients with autonomic dysfunction [60, 61]. Other contributing factors include microbiome dysbiosis, complement and coagulation pathway disruption, mast cell activation and reactivation of latent viruses such as EBV and HHV-6 [62–66]. There have been noteworthy sex-dependent immunological differences: males are more likely to have more severe acute illness, whereas females exhibit stronger early IFN responses and a higher prevalence of Long-COVID [67]. A central unresolved question is whether persistent immune activation is driven by replicating viral reservoirs, non-replicating residual viral antigens, cross-reactive autoimmunity, or a self-sustaining inflammatory loop independent of viral persistence [68, 69]. Mechanistically, several non-mutually exclusive pathways have been proposed to link pulmonary immune activation with CNS dysfunction (a) systemic immune signaling, whereby lung-derived cytokines (e.g., IL-6, TNF- α, IFN-γ, CXCL10) reaching the brain via circulation or through a compromised BBB, leading to microglial and other CNS cell activation, (b) vagus afferent signaling, in which the inflammatory signals from lung are transmitted to brainstem autonomic and sensory centers independent of viremia, and (c) viral antigen dissemination or immune cell trafficking, involving the migration of activated immune cells or viral antigens to CNS-associated lymphoid tissue or across a compromised BBB [27, 29, 68, 70–72]. Dissecting the relative contributions of these pathways or evaluating therapeutic interventions at defined stages of disease, or targeted immune alterations at specific timepoints, are not feasible in human studies, underscoring the critical need for well-characterized preclinical animal models. Preclinical animal models for Long-COVID Animal models have been indispensable for elucidating mechanisms underlying Long-COVID and its multi-organ sequelae, including pulmonary, cardiovascular, neurological, and immunological dysfunction. Studies of SARS-CoV-2–infected animals provide critical insights into host-virus interactions, immune dysregulation, and tissue-specific pathology that may persist beyond viral clearance. These models have been extensively employed to investigate molecular and cellular mechanisms driving both acute disease and post-recovery phenotypes [73, 74]. Importantly, disease severity and post-acute outcomes vary considerably across models and are influenced by host factors such as biological sex, age, genetic background, and pre-existing comorbidities [21, 22, 73]. Given the impact of these variables on disease trajectories, comparative evaluation across animal systems is essential for identifying conserved mechanisms relevant to Long-COVID pathogenesis. Animal models used to study COVID-19, including those relevant to Long-COVID, are broadly categorized into small or large animal model systems. Small animal models, such as mice, rats, ferrets, and Golden Syrian Hamster are widely used due to their cost-effectiveness, ease of handling, availability, and suitability for longitudinal studies, facilitated by shorter lifespans [17, 75]. These models are particularly advantageous for investigating post-acute and chronic disease phenotypes. However, key biological differences between small animals and humans often limit their translational relevance, frequently necessitating genetic modification (e.g., human angiotensin-converting enzyme 2 [ACE2] expression) to more accurately recapitulate SARS-CoV-2 infection and disease manifestations [76, 77]. Rat models of SARS-CoV-2 infection have primarily relied on genetically engineering to express human ACE2, enabling limited viral replication and acute pulmonary pathology [78, 79]. Human ACE2-expressing (ACE2) rats on the Fischer 344 (F344) or Sprague Dawley (CD) genetic backgrounds were created using a random transgenesis approach [79]. Infected hACE2 rats develop lung lesions characterized by thickened alveolar septa and lymphocytic infiltration, with detectable sub-genomic viral RNA indicating active viral replication [78]. These features make rats a suitable model for studying acute lung injury and toxicological or pharmacological aspects of disease pathogenesis [79]. However, viral replication in rats remains comparatively restricted, limiting their utility for studying antiviral efficacy or long-term disease progression [78]. Importantly, rat models have not been employed for Long-COVID research, and there is currently no evidence supporting persistent pulmonary pathology, neuroinflammation, or behavioral changes indicative of lung-brain-immune axis dysfunction. Ferrets are well-established models for respiratory viral infections and have been extensively used to study SARS-CoV-2 due to similarities in airway anatomy, ACE2 expression, and viral replication kinetics relative to humans [80, 81]. Infected ferrets exhibit human-like acute respiratory symptoms, including fever and cough, with viral replication in the nasal turbinate, trachea, and lungs. However, severe lung disease is quite rare in ferrets, and overall illness severity is typically mild. Because pulmonary inflammation and alveolar damage are key drivers of downstream neuroinflammatory signaling in Long-COVID, this represents a significant limitation of studying lung-brain-immune axis [80–82]. In ferret SARS-CoV-2 models, the magnitude of lung-derived cytokine release, immune cells activation, and vagal afferent signaling –processes that spread to CNS neuroinflammation –may not consistently achieve the threshold required to model these pathways [80–82]. Studies using ferrets shows presence of viral RNA in multiple tissues, including the brain, for up to 21 dpi. However, ferrets do not exhibit overt neurobehavioral abnormalities such as fatigue, cognitive dysfunction, or anxiety, and post-acute follow‑up periods remain relatively short [83, 84]. Further ferret models have not been employed extensively for Long-COVID study. In contrast, large animal models, most notably NHPs, more closely mimic human physiology, immune responses, and disease pathogenesis. NHPs offer important advantages for translational studies, including the ability to evaluate complex immune-neurovascular interactions. However, their use is constrained by high cost, limited availability, ethical considerations, and the need for specialized infrastructure and expertise, which collectively restrict sample size and statistical power [75, 85, 86]. To enable direct and consistent cross model comparisons, this review applies unified evaluative framework across all the species. Specifically, three core domains are assessed: (a) Lung pathology, (b) Brain pathology, including neurocognitive and behavioral consequences and (c) Immune features and systemic dysregulation. Figure 1 and Table 1 provide a comparative overview and limitations across commonly used model systems in Long-COVID studies. Table 2 summarizes key experimental parameters and reported Long-COVID outcomes. Fig. 1Distinct and shared pathophysiologic manifestations in small and large animal models of Long-COVID. The central panel summarizes core features observed across commonly used experimental models, including persistent viral RNA/antigen, chronic pulmonary injury and remodeling, immune dysregulation, and neurocognitive alterations. The animal model‑specific strengths, limitations, and translational relevance are highlighted in individual panels Distinct and shared pathophysiologic manifestations in small and large animal models of Long-COVID. The central panel summarizes core features observed across commonly used experimental models, including persistent viral RNA/antigen, chronic pulmonary injury and remodeling, immune dysregulation, and neurocognitive alterations. The animal model‑specific strengths, limitations, and translational relevance are highlighted in individual panels Table 1Comparative overview of animal models of Long-COVID: Advantages and LimitationsModelModel detailsKey AdvantagesKey LimitationsReferencesMouseK18-hACE2 mice: Express hACE2 under human k18 promoter resulting in transgene expression in airway epithelial cellsPost-acute fibrotic scarring; CNS transcriptomics; long-term behavioral follow-up; recapitulate clinical symptoms reported in long-COVID patientsRequires hACE2 transgene; aberrant hACE2 expression in brain may confounds neurotropism; exaggerated disease severity[87, 88]BALB/c mice infected with mouse-adapted SARS-CoV-2May be used for testing monoclonal antibody-based therapies for acute and post-acute sequelae of COVID-19; develop fibrotic lung disease post recovery from acute phase of infectionRequires mouse-adapted virus[89]MISTRG6-hACE2 mice: Humanized mice transduced with AAV-CMV-hACE2 in lungsRecapitulate SARS-CoV-2 mediated innate and adaptive immune responses up to 28 days post-infection as observed in humansTechnically complex (HSPC engraftment, AAV delivery), variable humanization between animals, few neutrophils of human origin, non-physiologic ACE2 expression[90]C57BL/6 mice infected with SARS-CoV-2 (B.1.351)No genetic modification required; human strain infection; Develop cognitive impairments similar to long-COVID patientsLimited to select human strains; mild disease; not yet fully characterized for PASC spectrum[91]Non-human primatesRhesus Macaques and African green monkeysSimilar to humans, develop neurobehavioral impairments post recovery from acute illness; These symptoms are associated with T-cell dysfunction and inflammatory responses during acute illness; Closest immunological approximation to humans; CSF access; causal IL-6-to-cognition link established; obesity-comorbidity modeling, Metabolic PASC modelingSmall cohorts (n = 3–15); high cost; long study duration; limited scalability, minimal CNS characterization of Long-COVID[92–94]HamstersGolden Syrian HamstersNo genetic modification; natural infection; robust replication; post-acute lung fibrosis; brainstem viral reservoir; neurobehavioral phenotyping; sex-specific patterns; BSL-2/3 compatibleLimited validated cognitive battery: behavioral assays adapted from mice; precise CNS reservoir not fully resolved[20, 35, 95, 96] Table 1Comparative overview of animal models of Long-COVID: Advantages and LimitationsModelModel detailsKey AdvantagesKey LimitationsReferencesMouseK18-hACE2 mice: Express hACE2 under human k18 promoter resulting in transgene expression in airway epithelial cellsPost-acute fibrotic scarring; CNS transcriptomics; long-term behavioral follow-up; recapitulate clinical symptoms reported in long-COVID patientsRequires hACE2 transgene; aberrant hACE2 expression in brain may confounds neurotropism; exaggerated disease severity[87, 88]BALB/c mice infected with mouse-adapted SARS-CoV-2May be used for testing monoclonal antibody-based therapies for acute and post-acute sequelae of COVID-19; develop fibrotic lung disease post recovery from acute phase of infectionRequires mouse-adapted virus[89]MISTRG6-hACE2 mice: Humanized mice transduced with AAV-CMV-hACE2 in lungsRecapitulate SARS-CoV-2 mediated innate and adaptive immune responses up to 28 days post-infection as observed in humansTechnically complex (HSPC engraftment, AAV delivery), variable humanization between animals, few neutrophils of human origin, non-physiologic ACE2 expression[90]C57BL/6 mice infected with SARS-CoV-2 (B.1.351)No genetic modification required; human strain infection; Develop cognitive impairments similar to long-COVID patientsLimited to select human strains; mild disease; not yet fully characterized for PASC spectrum[91]Non-human primatesRhesus Macaques and African green monkeysSimilar to humans, develop neurobehavioral impairments post recovery from acute illness; These symptoms are associated with T-cell dysfunction and inflammatory responses during acute illness; Closest immunological approximation to humans; CSF access; causal IL-6-to-cognition link established; obesity-comorbidity modeling, Metabolic PASC modelingSmall cohorts (n = 3–15); high cost; long study duration; limited scalability, minimal CNS characterization of Long-COVID[92–94]HamstersGolden Syrian HamstersNo genetic modification; natural infection; robust replication; post-acute lung fibrosis; brainstem viral reservoir; neurobehavioral phenotyping; sex-specific patterns; BSL-2/3 compatibleLimited validated cognitive battery: behavioral assays adapted from mice; precise CNS reservoir not fully resolved[20, 35, 95, 96] Comparative overview of animal models of Long-COVID: Advantages and Limitations Table 2Experimental parameters and Long-COVID outcomes across animal modelsModelVirus strainRoute & DoseKey Long-COVID OutcomesFollow-up (dpi)ReferencesBALB/c miceMouse adapted BA.5 MA and BA.2 MAIntranasal; 1 × 10^4 or 1 × 10^5 PFU BA.5 MA, or 1 × 10^5 PFU Dose-dependent acute weight loss and lung dysfunction; survivors recover weight but develop chronic lung changes, modeling post-acute sequelaePost-acute subpleural and interstitial lung fibrosis with organizing pneumonia and tertiary lymphoid structures in survivors up to ≥ 100 days post-infection107[89]K18-hACE2 miceDelta and Omicron B.A.2Intranasal; 6-8 x 10^2 PFUSublethal Delta causes acute weight loss, neuroinflammation, partial mortality; survivors show persistent neuropsychiatric and motor abnormalities up to 4 months, with ataxia-like deficitsAcute brain and lung inflammation with encephalitis; post-acute phase shows viral clearance, but persistent CNS immune-pathway expression changes linked to cognitive and motor dysfunction120[88]MISTRG6 humanized miceUSA-WA1/2020Intranasal; 1–3 × 10^6 PFUChronic weight loss, persistent illness ≥ 28–35 days, sustained viral RNA and lung dysfunction despite immune responses, modeling chronic/post-acute COVID-19Fibrotic lung phases with collagen deposition and architectural remodeling, persistent inflammatory human macrophages, prolonged ISG signature, T cell lymphopenia, and SLE-like extrafollicular B cell responses28[90]C57BL/6J miceSARS-CoV-2 B.1.351 (Beta variant)Intranasal; 5 × 10^5 PFUSARS-CoV-2 induces hippocampal IL-1β–driven neuroinflammation that impairs neurogenesis and causes memory deficits without direct brain infection, while vaccination prevents this30[91]African green monkeys2019-nCoV/USA-WA1/2020Intranasal; 10^6 TCID50SARS‑CoV‑2-infected African green monkeys develop early hyperglycemia that remains elevated for at least 18 weeks, with higher and more persistent glucose in unvaccinated animals than in those given BNT162b2 on day 4 post‑infectionWeek 1 plasma signature dominated by chemokines (especially CCL25) and GDNF correlates with higher glucose across months, alongside hyper‑responsive cytokine‑producing CD8 T cells and increased hepatic glycogen without long‑term virus in liver or pancreas126[92]Rhesus macaques2019-nCoV/USA-WA1/2020Intranasal and intratracheal; 1.1 × 10^6 PFUAdult rhesus macaques with mild WA1 infection develop persistent loss/alteration of taste-smell, reduced cognitive flexibility, sleep disruption, and altered heart-rate variability for up to ~ 4.5 months, with generally stronger effects in femalesEarly CSF and plasma cytokine shifts (e.g., IL‑6, IL‑4, IL‑12p70, IL‑15, TARC) and intermediate monocyte/T‑cell changes correlate with later chemosensory and cognitive deficits, alongside increased IBA1⁺ microglia in insula and olfactory cortex and low‑level, long‑lasting viral RNA/protein in lung and lymph nodes150[93]Rhesus macaquesIsolate hCoV-19/USA/MND-HP05647/2021 B.1.617.2; delta variantIntranasal and intratracheal; 1 × 10^6 PFUPersistent long COVID–like metabolic, inflammatory, and lung abnormalities lasting at least 6 months, especially in obese animals, even lean animals developed some obesity-like metabolic changes, Obesity does not change acute viral loads but associates with greater chronic lung pathology, distinct neutralizing antibody decay, body weight loss–driven insulin-sensitivity “improvement,” and vRNA/lung damage correlating with fat mass195[94]Golden Syrian HamstersSARS-CoV-2 Wuhan BetaCoV/France/ IDF00372/2020,SARS-CoV-2 Delta/2021/ I7.2 200, SARS-CoV-2 Omicron/B1.1.529Intranasal; 6 × 10^4 PFUAfter acute weight loss and clinical illness, infected hamsters show persistent depression‑like behavior, recognition‑memory deficits, and late‑onset anxiety up to 80 days post‑infectionPersistent SARS‑CoV‑2 infection in the brainstem (infectious virus detectable at 80 days), astroglia activation, and a neurodegenerative-like brainstem transcriptomic signature affecting synapses, energy metabolism80[35]Golden Syrian HamstersUSA-WA1/2020Intranasal; 1000 PFUSARS‑CoV‑2–infected hamsters show more sustained weight loss, reduced appetite, and delayed recovery versus IAV, plus persistent behavioral changes (altered marble‑burying) about one month after infection, consistent with neuropsychiatric‑type PASC featuresAfter lung viral clearance, hamsters exhibit lasting lambertosis and kidney tubular atrophy, and critically prolonged IFN‑I/II and chemokine signatures with microglial/myeloid activation in olfactory bulb and epithelium, paralleling inflammatory transcriptomes in olfactory tissues from recovered human COVID‑19 donors31[95]Golden Syrian HamstersAncestral SARS-CoV-2Intranasal; 10^5 TCID50Golden Syrian hamsters develop mild to moderate disease with weight loss, lung viral replication, and delayed weight regain, especially in males, but are clinically recovered by 31 days post‑infectionDespite virological clearance, lungs show a distinct proteomic signature at 31 days. sex‑biased reductions in many proteins in males, and sustained upregulation of mucus/airway-barrier proteins (Muc5AC, Muc5B, CLCA1, BPIFB1, intelectin‑1), implying altered mucosal homeostasis consistent with respiratory PASC mechanisms31[97]Golden Syrian HamstersB.1Intranasal; 10^4 TCID50Intranasally infected golden Syrian hamsters develop mild–moderate acute disease with transient ~ 10–13% weight loss and recovery by ~ 2–3 weeks, but a subset shows persistent deficits in burrowing (daily‑living analogue), open‑field, and elevated‑zero‑maze behavior, suggesting long‑lasting anxiety‑ and motivation‑related impairments.Viral gRNA persists in nasal turbinates and lungs, with early brain viral RNA and neuroinflammation, prolonged, heterogeneous cytokine dysregulation in lung, brain, and plasma (notably IL‑1, IFN‑γ, IL‑10, TNF, chemokines), robust neutralizing and T‑cell responses, and occasional structural change in the cervical vagus nerve, paralleling proposed neuro‑PCC (Post-COVID conditions) mechanisms involving persistence, chronic inflammation, and dysautonomia60[98]Golden Syrian HamstersSARS-CoV-2 Delta variantIntranasal; 10^4 TCID50Aged male Delta‑infected hamsters have acute pneumonia that resolves, but show persistent exercise‑induced lung dysfunction and reduced treadmill tolerance for ~ 7 weeks, despite appearing well at restLungs retain subpleural/interstitial fibrosis and bronchiolization with SCGB1A1⁺ club‑cell–rich foci, M2‑like macrophages, pro‑fibrotic transcripts, and RdRp RNA up to 112 days, without chronic extra‑pulmonary lesions112[96] Table 2Experimental parameters and Long-COVID outcomes across animal modelsModelVirus strainRoute & DoseKey Long-COVID OutcomesFollow-up (dpi)ReferencesBALB/c miceMouse adapted BA.5 MA and BA.2 MAIntranasal; 1 × 10^4 or 1 × 10^5 PFU BA.5 MA, or 1 × 10^5 PFU Dose-dependent acute weight loss and lung dysfunction; survivors recover weight but develop chronic lung changes, modeling post-acute sequelaePost-acute subpleural and interstitial lung fibrosis with organizing pneumonia and tertiary lymphoid structures in survivors up to ≥ 100 days post-infection107[89]K18-hACE2 miceDelta and Omicron B.A.2Intranasal; 6-8 x 10^2 PFUSublethal Delta causes acute weight loss, neuroinflammation, partial mortality; survivors show persistent neuropsychiatric and motor abnormalities up to 4 months, with ataxia-like deficitsAcute brain and lung inflammation with encephalitis; post-acute phase shows viral clearance, but persistent CNS immune-pathway expression changes linked to cognitive and motor dysfunction120[88]MISTRG6 humanized miceUSA-WA1/2020Intranasal; 1–3 × 10^6 PFUChronic weight loss, persistent illness ≥ 28–35 days, sustained viral RNA and lung dysfunction despite immune responses, modeling chronic/post-acute COVID-19Fibrotic lung phases with collagen deposition and architectural remodeling, persistent inflammatory human macrophages, prolonged ISG signature, T cell lymphopenia, and SLE-like extrafollicular B cell responses28[90]C57BL/6J miceSARS-CoV-2 B.1.351 (Beta variant)Intranasal; 5 × 10^5 PFUSARS-CoV-2 induces hippocampal IL-1β–driven neuroinflammation that impairs neurogenesis and causes memory deficits without direct brain infection, while vaccination prevents this30[91]African green monkeys2019-nCoV/USA-WA1/2020Intranasal; 10^6 TCID50SARS‑CoV‑2-infected African green monkeys develop early hyperglycemia that remains elevated for at least 18 weeks, with higher and more persistent glucose in unvaccinated animals than in those given BNT162b2 on day 4 post‑infectionWeek 1 plasma signature dominated by chemokines (especially CCL25) and GDNF correlates with higher glucose across months, alongside hyper‑responsive cytokine‑producing CD8 T cells and increased hepatic glycogen without long‑term virus in liver or pancreas126[92]Rhesus macaques2019-nCoV/USA-WA1/2020Intranasal and intratracheal; 1.1 × 10^6 PFUAdult rhesus macaques with mild WA1 infection develop persistent loss/alteration of taste-smell, reduced cognitive flexibility, sleep disruption, and altered heart-rate variability for up to ~ 4.5 months, with generally stronger effects in femalesEarly CSF and plasma cytokine shifts (e.g., IL‑6, IL‑4, IL‑12p70, IL‑15, TARC) and intermediate monocyte/T‑cell changes correlate with later chemosensory and cognitive deficits, alongside increased IBA1⁺ microglia in insula and olfactory cortex and low‑level, long‑lasting viral RNA/protein in lung and lymph nodes150[93]Rhesus macaquesIsolate hCoV-19/USA/MND-HP05647/2021 B.1.617.2; delta variantIntranasal and intratracheal; 1 × 10^6 PFUPersistent long COVID–like metabolic, inflammatory, and lung abnormalities lasting at least 6 months, especially in obese animals, even lean animals developed some obesity-like metabolic changes, Obesity does not change acute viral loads but associates with greater chronic lung pathology, distinct neutralizing antibody decay, body weight loss–driven insulin-sensitivity “improvement,” and vRNA/lung damage correlating with fat mass195[94]Golden Syrian HamstersSARS-CoV-2 Wuhan BetaCoV/France/ IDF00372/2020,SARS-CoV-2 Delta/2021/ I7.2 200, SARS-CoV-2 Omicron/B1.1.529Intranasal; 6 × 10^4 PFUAfter acute weight loss and clinical illness, infected hamsters show persistent depression‑like behavior, recognition‑memory deficits, and late‑onset anxiety up to 80 days post‑infectionPersistent SARS‑CoV‑2 infection in the brainstem (infectious virus detectable at 80 days), astroglia activation, and a neurodegenerative-like brainstem transcriptomic signature affecting synapses, energy metabolism80[35]Golden Syrian HamstersUSA-WA1/2020Intranasal; 1000 PFUSARS‑CoV‑2–infected hamsters show more sustained weight loss, reduced appetite, and delayed recovery versus IAV, plus persistent behavioral changes (altered marble‑burying) about one month after infection, consistent with neuropsychiatric‑type PASC featuresAfter lung viral clearance, hamsters exhibit lasting lambertosis and kidney tubular atrophy, and critically prolonged IFN‑I/II and chemokine signatures with microglial/myeloid activation in olfactory bulb and epithelium, paralleling inflammatory transcriptomes in olfactory tissues from recovered human COVID‑19 donors31[95]Golden Syrian HamstersAncestral SARS-CoV-2Intranasal; 10^5 TCID50Golden Syrian hamsters develop mild to moderate disease with weight loss, lung viral replication, and delayed weight regain, especially in males, but are clinically recovered by 31 days post‑infectionDespite virological clearance, lungs show a distinct proteomic signature at 31 days. sex‑biased reductions in many proteins in males, and sustained upregulation of mucus/airway-barrier proteins (Muc5AC, Muc5B, CLCA1, BPIFB1, intelectin‑1), implying altered mucosal homeostasis consistent with respiratory PASC mechanisms31[97]Golden Syrian HamstersB.1Intranasal; 10^4 TCID50Intranasally infected golden Syrian hamsters develop mild–moderate acute disease with transient ~ 10–13% weight loss and recovery by ~ 2–3 weeks, but a subset shows persistent deficits in burrowing (daily‑living analogue), open‑field, and elevated‑zero‑maze behavior, suggesting long‑lasting anxiety‑ and motivation‑related impairments.Viral gRNA persists in nasal turbinates and lungs, with early brain viral RNA and neuroinflammation, prolonged, heterogeneous cytokine dysregulation in lung, brain, and plasma (notably IL‑1, IFN‑γ, IL‑10, TNF, chemokines), robust neutralizing and T‑cell responses, and occasional structural change in the cervical vagus nerve, paralleling proposed neuro‑PCC (Post-COVID conditions) mechanisms involving persistence, chronic inflammation, and dysautonomia60[98]Golden Syrian HamstersSARS-CoV-2 Delta variantIntranasal; 10^4 TCID50Aged male Delta‑infected hamsters have acute pneumonia that resolves, but show persistent exercise‑induced lung dysfunction and reduced treadmill tolerance for ~ 7 weeks, despite appearing well at restLungs retain subpleural/interstitial fibrosis and bronchiolization with SCGB1A1⁺ club‑cell–rich foci, M2‑like macrophages, pro‑fibrotic transcripts, and RdRp RNA up to 112 days, without chronic extra‑pulmonary lesions112[96] Experimental parameters and Long-COVID outcomes across animal models Dose-dependent acute weight loss and lung dysfunction; survivors recover weight but develop chronic lung changes, modeling post-acute sequelae Post-acute subpleural and interstitial lung fibrosis with organizing pneumonia and tertiary lymphoid structures in survivors up to ≥ 100 days post-infection Sublethal Delta causes acute weight loss, neuroinflammation, partial mortality; survivors show persistent neuropsychiatric and motor abnormalities up to 4 months, with ataxia-like deficits Acute brain and lung inflammation with encephalitis; post-acute phase shows viral clearance, but persistent CNS immune-pathway expression changes linked to cognitive and motor dysfunction Chronic weight loss, persistent illness ≥ 28–35 days, sustained viral RNA and lung dysfunction despite immune responses, modeling chronic/post-acute COVID-19 Fibrotic lung phases with collagen deposition and architectural remodeling, persistent inflammatory human macrophages, prolonged ISG signature, T cell lymphopenia, and SLE-like extrafollicular B cell responses SARS‑CoV‑2-infected African green monkeys develop early hyperglycemia that remains elevated for at least 18 weeks, with higher and more persistent glucose in unvaccinated animals than in those given BNT162b2 on day 4 post‑infection Week 1 plasma signature dominated by chemokines (especially CCL25) and GDNF correlates with higher glucose across months, alongside hyper‑responsive cytokine‑producing CD8 T cells and increased hepatic glycogen without long‑term virus in liver or pancreas Adult rhesus macaques with mild WA1 infection develop persistent loss/alteration of taste-smell, reduced cognitive flexibility, sleep disruption, and altered heart-rate variability for up to ~ 4.5 months, with generally stronger effects in females Early CSF and plasma cytokine shifts (e.g., IL‑6, IL‑4, IL‑12p70, IL‑15, TARC) and intermediate monocyte/T‑cell changes correlate with later chemosensory and cognitive deficits, alongside increased IBA1⁺ microglia in insula and olfactory cortex and low‑level, long‑lasting viral RNA/protein in lung and lymph nodes SARS-CoV-2 Wuhan BetaCoV/France/ IDF00372/2020, SARS-CoV-2 Delta/2021/ I7.2 200, SARS-CoV-2 Omicron/B1.1.529 After acute weight loss and clinical illness, infected hamsters show persistent depression‑like behavior, recognition‑memory deficits, and late‑onset anxiety up to 80 days post‑infection Persistent SARS‑CoV‑2 infection in the brainstem (infectious virus detectable at 80 days), astroglia activation, and a neurodegenerative-like brainstem transcriptomic signature affecting synapses, energy metabolism SARS‑CoV‑2–infected hamsters show more sustained weight loss, reduced appetite, and delayed recovery versus IAV, plus persistent behavioral changes (altered marble‑burying) about one month after infection, consistent with neuropsychiatric‑type PASC features After lung viral clearance, hamsters exhibit lasting lambertosis and kidney tubular atrophy, and critically prolonged IFN‑I/II and chemokine signatures with microglial/myeloid activation in olfactory bulb and epithelium, paralleling inflammatory transcriptomes in olfactory tissues from recovered human COVID‑19 donors Golden Syrian hamsters develop mild to moderate disease with weight loss, lung viral replication, and delayed weight regain, especially in males, but are clinically recovered by 31 days post‑infection Despite virological clearance, lungs show a distinct proteomic signature at 31 days. sex‑biased reductions in many proteins in males, and sustained upregulation of mucus/airway-barrier proteins (Muc5AC, Muc5B, CLCA1, BPIFB1, intelectin‑1), implying altered mucosal homeostasis consistent with respiratory PASC mechanisms Intranasally infected golden Syrian hamsters develop mild–moderate acute disease with transient ~ 10–13% weight loss and recovery by ~ 2–3 weeks, but a subset shows persistent deficits in burrowing (daily‑living analogue), open‑field, and elevated‑zero‑maze behavior, suggesting long‑lasting anxiety‑ and motivation‑related impairments. Viral gRNA persists in nasal turbinates and lungs, with early brain viral RNA and neuroinflammation, prolonged, heterogeneous cytokine dysregulation in lung, brain, and plasma (notably IL‑1, IFN‑γ, IL‑10, TNF, chemokines), robust neutralizing and T‑cell responses, and occasional structural change in the cervical vagus nerve, paralleling proposed neuro‑PCC (Post-COVID conditions) mechanisms involving persistence, chronic inflammation, and dysautonomia Aged male Delta‑infected hamsters have acute pneumonia that resolves, but show persistent exercise‑induced lung dysfunction and reduced treadmill tolerance for ~ 7 weeks, despite appearing well at rest Lungs retain subpleural/interstitial fibrosis and bronchiolization with SCGB1A1⁺ club‑cell–rich foci, M2‑like macrophages, pro‑fibrotic transcripts, and RdRp RNA up to 112 days, without chronic extra‑pulmonary lesions Mouse models and the lung–brain-immune axis Mice are the most widely used small-animal models in infectious disease research because of their well-characterized immune systems, short gestation periods, and high reproductive capacity [17]. However, majority of the laboratory strains of mice are resistant to SARS-CoV-2 infection due to incompatibility between murine ACE2 and the SARS-CoV-2 spike protein. As a result, to establish productive infection laboratory used mice strains require genetic modification of either the host or the virus, complicating interpretation of lung-brain-immune axis mechanisms relevant to human disease. Although some genetically engineered and C57BL/6 mice can be infected with SARS-CoV-2, replicating the COVID-19 disease manifestations and its progression remains challenging [91, 99]. Notably, C57BL/6 mice can support infection with select SARS-CoV-2 strains without genetic modification and have been shown to develop pulmonary infection with memory impairments [91]. However, generally mice do not reliably exhibit clinical signs or pathological features that are seen in COVID-19 patients [90]. Lung pathology In comparison to human disease, as defined by ADI cell accumulation, macrophage driven fibrosis, bronchiolization and exercise intolerance as discussed in "Lung pathology in Long-COVID patients" section, mouse models provide the following evidence of lung pathology during Long-COVID. The most extensively used murine model used for COVID-19 research is the K18-hACE2 transgenic mouse, in which human ACE2 expression is driven by the human cytokeratin 18 (K18) promoter [100, 101]. Following SARS-CoV-2 inoculation, these mice develop acute lung injury, pulmonary inflammation and weight loss. Several other transgenic mouse strains have also been developed to study the broader effects and pathogenesis of the viral infection [101]. These findings mirror non‑fatal COVID‑19, where residual lung abnormalities (ground‑glass opacities, peripheral fibrosis), vascular injury, and neuroinflammatory signatures coexist after virologic resolution [88]. Mouse-adapted SARS-CoV-2 strains further emphasize pulmonary persistence without parallel brain assessment. In aged BALB/c female mice (10–12 months) infected with the mouse-adapted Omicron BA.5 strain, animals developed chronic subpleural and parenchymal lung fibrosis lasting beyond 100 dpi, with organizing pneumonia and tertiary lymphoid structures persisting after viral clearance [89]. Females were challenged with the virus and were followed up to 107 dpi with longitudinal assessment of body weight, lung function, viral persistence, tissue histology, and neutralizing antibodies [89]. Survivors developed chronic subpleural and parenchymal lung fibrosis with organizing pneumonia and tertiary lymphoid structures that persisted for months, even after viral clearance. Serum showed strong homotypic neutralization but poor cross‑neutralization of antigenically distant variants. These features recapitulate post‑acute interstitial lung disease and exertional dyspnea reported in Long‑COVID patients. Moreover, the severity of disease contrasts with the relatively lower Long‑COVID risk associated with Omicron infections in humans [89]. SARS-CoV-2 (B.1.351 variant) infects the respiratory tract of C57BL/6J mice, resulting in transient weight loss and robust antiviral cytokine responses [91]. SARS-CoV-2 infected humanized mouse models, such as MISTRG6 mice transplanted with human hematopoietic stem and progenitor cells and transduced with AAV‑hACE2, further highlight immune-driven lung pathology with systemic consequences [90]. These mice develop prolonged lung viral persistence, progressive lung fibrotic remodeling, sustained human inflammatory macrophage recruitment, and interferon‑stimulated gene expression, closely paralleling severe human COVID-19. However, these models predominantly reflect ICU‑level disease rather than community-acquired Long‑COVID and lack defined behavioral or neurocognitive endpoints relevant to lung–brain–immune axis dysfunction. Brain pathology, neurocognitive and behavioral consequences In comparison with human neuroPASC, which encompasses brain fog, memory impairment, neuroinflammation, BBB dysfunction, and potential neurodegenerative trajectories as discussed in "Brain pathology and neurocognitive consequences in Long-COVID patients" section, the mouse model captures several key features, as outlined below. The K18-hACE2 transgenic mouse shows evidence of frequent neuroinvasion with neuronal damage alongside pulmonary injury and inflammation [101]. While K18-hACE2 transgenic mouse model enables study of viral neurotropism, disease severity is often exaggerated, and pathological features do not consistently reflect the spectrum of human COVID-19 or Long-COVID [102]. Notably, brain pathology persisted even in the absence of detectable pulmonary virus, suggesting delayed or indirect neuroinflammatory consequences downstream of lung injury [88]. Further, the short observation window and lack of measurable behavioral phenotypes limit modeling of chronic lung–brain–immune axis dysfunction, such as fatigue, cognitive impairment, or autonomic instability [88]. In aged BALB/c female mice (10–12 months) infected with the mouse-adapted Omicron BA.5 strain, lung pathology have been reported during Long-COVID while neurological and systemic manifestations were not characterized, limiting insight into the downstream brain consequences of chronic lung pathology [89]. More recently, long-term behavioral studies have extended murine lung–brain–immune axis modeling [88]. K18‑hACE2 mice infected with sub‑lethal doses of SARS‑CoV‑2 Delta or Omicron BA.2 variants exhibited persistent neuropsychiatric and motor abnormalities up to four months post‑infection, accompanied by transcriptional signatures of complement activation, phagocytosis, impaired cognition, and neuronal degeneration [88]. These changes occurred despite viral clearance and minimal residual microgliosis; early pulmonary and systemic immune activation may imprint long-lasting brain dysfunction. Nevertheless, reliance on ectopic hACE2 expression, high‑dose viral exposure, and murine behavioral surrogates limits direct extrapolation to Long‑COVID. Humanized MISTRG6 models, while recapitulating immune-driven lung pathology, lack defined behavioral or neurocognitive endpoints relevant to lung–brain–immune axis dysfunction [90]. Despite no direct brain infection, C57BL/6J mice develop hippocampal dysfunction with reduced neurogenesis, synapse loss, and memory deficits [91]. Immune features and systemic dysregulation As described in "Immune dysregulation in Long-COVID patients" section, the chronic immune dysregulation in Long-COVID includes T-cell exhaustion, monocyte expansion, ISG persistence, and sex-biased cytokine responses. In comparison, mouse models recapitulate the immune features of Long-COVID to a limited extent. Four months after infection, transcriptional markers of complement activation, phagocytosis, decreased cognition, and neuronal degeneration were observed in the brains of K18-hACE2 mice infected with sub-lethal Delta or Omicron BA.2, despite very little residual microgliosis [88]. In BALB/c mice infected with mouse-adapted Omicron BA.5 variant, viral RNA was undetectable in lung by 107 dpi, yet fibrotic pathology persisted, demonstrating that structural Long-COVID sequelae can be driven by virus-independent mechanisms, including immune dysfunctions [89]. The most human-like immune dysfunctions observed in humanized MISTRG6 mice, which exhibit prolonged ISG expression, T-cell lymphopenia, extrafollicular SLE-like B-cell responses, and a temporal macrophage switch from antiviral to pro-fibrotic activity [90]. Infection in C57BL/6J mice triggers microglial activation and infiltration of immune cells into the brain, with elevated IL-1β driven impaired neurogenesis and cognitive dysfunction [91]. Overall, while mouse models provide valuable mechanistic insights into acute pulmonary injury and immune-driven pathology, their reliance on genetic manipulation, exaggerated disease severity, and incomplete coupling of lung and brain phenotypes limit their utility for modeling chronic lung-brain-immune axis dysfunction in Long-COVID. Non-human primate models and lung-brain-immune axis integration NHPs provide the closest physiological and immunological approximation to humans and have been instrumental in studying SARS-CoV-2 pathogenesis, transmission, and vaccine/therapeutic efficacy [103–106]. While many NHP species supports productive SARS-CoV-2 replication with mild clinical disease, baboons develop severe lung pathology, and rhesus macaques (RMs) show robust viral replication with moderate pulmonary involvement [107]. NHP models are uniquely positioned to recapitulate the mechanistic crosstalk along the lung-brain-immune axis that underpins Long-COVID pathophysiology, given their close similarity to humans in immune system architecture, cytokine signaling, T- and B-cell compartments, and neuroimmune communication pathways. However, their use is constrained by high cost, limited scalability, ethical considerations, and extended study durations. Lung pathology In comparison with Long-COVID lung pathology ("Lung pathology in Long-COVID patients" section), RMs offer high-fidelity recapitulation of post-acute pulmonary pathology of Long-COVID, specifically in the context of metabolic comorbidities that exacerbate lung disease severity. Longitudinal studies demonstrate that RMs infected with the Delta variant develop persistent pulmonary pathology, including bronchioloalveolar hyperplasia, chronic interstitial pneumonia, and perivascular lymphoid aggregates detectable at six-month post infection [94]. These RMs also exhibit metabolic, and behavioral abnormalities lasting up to six months, with pre‑existing obesity markedly amplifying lung pathology, as evidenced by greater interstitial damage, inflammatory infiltration, and systemic manifestations [94]. Notably, the persistence of viral RNA in nasal turbinate of obese animals at six months post infection suggests that metabolic comorbidities may prolong pulmonary viral antigen persistence and sustain immune activation, thereby driving downstream organ dysfunction reported in Long-COVID [94]. These findings underscore how chronic lung injury interacts with metabolic state to influence systemic and downstream brain, immune and behavioral phenotypes, reinforcing the lung-brain-immune axis framework. Brain pathology, neurocognitive and behavioral consequences As discussed in "Brain pathology and neurocognitive consequences in Long-COVID patients" section, human neurocognitive PASC encompasses impairments in cognitive flexibility, sleep, autonomic function, anosmia, and potential neurodegeneration. Among preclinical models, NHPs represent the most translationally relevant platform for capturing these features. Adult RMs infected with ancestral SARS-CoV-2 (WA1/2020) exhibit long-lasting alterations in taste, smell, cognition, sleep, circadian rhythms, and autonomic function for months after infection, despite clearance of high-level viral replication; these effects are more pronounced in females [93]. Importantly, even mild infection is sufficient to induce durable neurocognitive alterations, making them the most therapeutically relevant model for studying neuro-PASC caused by mild illness. These neurocognitive aftereffects directly connected lung infection to brain neuroinflammation via the neuroimmune axis, as seen by microglial activation in the insula and nucleocapsid protein detection in the olfactory cortex [93]. The RMs infected with delta variant of SARS-CoV-2 showed behavioral changes such as sleep disturbances and metabolic abnormalities up to six month of follow-up post infection [94]. Immune features and systemic dysregulation NHP models provide the closest recapitulation of the Long-COVID immune profile as mentioned ("Immune dysregulation in Long-COVID patients" section) and uniquely enable the identification of causal links between acute immune response and chronic neurocognitive outcomes. In African green monkeys (AGMs), SARS-CoV-2 infection induces persistent hyperglycemia associated with immune-metabolic signaling pathways involving GDNF and CCL25, highlighting systemic consequences of pulmonary infection [92]. Plasma cytokines, including IP-10, TARC, IL-15, MCP-1, IL-18, and IL-1RA, are elevated, while BAL cytokine profiling reveals early peaks in IL-6, IP-10, IL-5, MIP-1α, TARC, and IL-16 at 2-4 dpi. Intermediate monocytes (CCR2⁺/CD11b⁺/Siglec-1⁺/CD163⁺) remain persistently elevated, and CD8⁺ T cells exhibit sustained proliferation and prolonged Granzyme B production up to 56 dpi, features that mirrors immune alterations in Long-COVID patients. Furthermore, NHP studies have established one of the first causal immune-to-brain mechanistic link in a PASC animal model and early acute plasma IL-6 levels predict subsequent taste and smell loss at 77 dpi, while CSF cytokine changes correlate with impairments in cognitive performance at 85 dpi [93]. These findings, together with evidence of persistent immune remodeling, viral RNA or protein retention in lymphoid and neural tissues and brain region-specific microglial activation (e.g., insula), – provide compelling evidence for lung-triggered, immune-mediated neuroinflammation underlying PASC-like phenotypes [93]. Hamster models of COVID-19 and Long-COVID Golden Syrian Hamsters are widely used to study acute SARS-CoV-2 infection because they support robust viral replication and develops pulmonary inflammation that closely resembles COVID-19, including comparable infection routes, viral replication kinetics, and respiratory pathology [101, 108, 109]. Following intranasal infection, Golden Syrian Hamsters exhibit body weight loss and high viral loads in the nasal mucosa, lungs, intestines, and lower airway epithelial cells. Acute disease in hamsters is typically mild and self-limiting, with clinical recovery and development of neutralizing antibodies by approximately 14 dpi, mirroring mild-to-moderate COVID-19 [18]. Golden Syrian Hamster models offer a uniquely powerful platform. As summarizes in Table 3 and Fig. 2, Golden Syrian Hamsters consistently recapitulate core features off Long‑COVID, including dysregulated alveolar repair, fibrotic lung remodeling, persistent neuroinflammation with brainstem involvement, olfactory dysfunction affecting feeding behavior, chronic immune activation, and viral or antigen persistence. Across studies, Golden Syrian Hamsters reveal lung-centered disease propagation to the brain and sensory systems, along with sex- and obesity-dependent phenotypes and exercise‑unmasked functional impairments that are difficult to capture clinically. Notably, recovery of infectious virus from lung and brain tissue up to 80–112 dpi provides direct experimental support for viral persistence hypotheses proposed in Long-COVID [35, 96]. Table 3Translational comparison of lung–brain–immune axis and multi-organ pathophysiology in Long-COVID across Golden Syrian Hamster model and humansPathophysiologic DomainExperimental Findings in Golden Syrian HamstersClinical and Molecular Findings in HumansTranslational AlignmentBrain & cognitive (behavior)Golden Syrian Hamsters show persistent depression‑like behavior, late‑onset anxiety, and impaired short‑term recognition memory up to ~ 80 dpi after intranasal SARS‑CoV‑2, linked to brainstem changes [35]Long‑COVID cohorts report fatigue, brain fog, memory problems, anxiety, and depression months post‑infection, often fluctuating over time [7, 35, 110]Strong behavioral parallel; hamsters lack psychosocial confounders, but only approximateBrain & cognitive (molecular)Brainstem RNA‑seq reveals a neurodegeneration-like signature (inflammasome/innate genes, synaptic and mitochondrial pathways, PD/AD‑linked genes) persisting to 80 dpi; CNS regions show long‑lasting metabolic and synaptic transcriptional changes after SARS‑CoV‑2 vs. IAV [35, 95]Neuroimaging and fluid‑biomarker studies in patients suggest accelerated brain aging, microstructural changes, and possible neurodegenerative trajectories after COVID‑19 [35, 95, 111–113]Conceptually similar molecular signatures indicating neurodegenerative risk; human evidence is still emergingOlfactory systemHamsters exhibit acute anosmia plus persistent IFN‑I/II, chemokine (CXCL10, CCL5), and microglial/myeloid activation signatures in olfactory bulb and epithelium weeks after viral clearance [95]Recovered patients can have prolonged smell loss; human olfactory bulb and epithelium show persistent interferon, complement, and T‑cell/chemotaxis programs months after COVID‑19 [95, 114]Highly similar inflammatory/transcriptomic patterns in olfactory tissues.Neuro‑PCC heterogeneityLong‑term study upto 60 dpi finds subsets of hamsters with persistent or late‑onset deficits in burrowing (motivation) and anxiety‑like behavior, alongside variable brain and plasma cytokine profiles and vagus‑nerve changes; others recover, mirroring individual variability [98]PCC is clinically heterogeneous, with some individuals severely affected and others minimally symptomatic, and variable immune/biomarker signatures over time [98, 115]Strong qualitative similarity in heterogeneityRespiratory structure (alveoli)Multiple studies show dysregulated alveolar regeneration: fibrosis and bronchiolization with SCGB1A1⁺ club‑cell–rich foci, M2‑like macrophages, pro‑fibrotic transcripts; subpleural and interstitial fibrosis persists up to 112 dpi [96, 97]Long‑COVID and post‑COVID patients exhibit impaired diffusion, exertional dyspnea, reduced exercise capacity, and CT/histologic evidence of incomplete alveolar repair, bronchiolization, and fibrosis months after infection [33, 34, 96]Strong structural and functional parallels, especially for respiratory Long‑COVID; human changes can be more variable and influenced by comorbid lung diseaseMulti‑organ injurySARS‑CoV‑2–infected hamsters show persistent lambertosis and inflammation in the lung, tubular atrophy in the kidney, and subtle heart changes; some models report systemic proteomic shifts and kidney damage persisting to 31 dpi, with no major extra‑pulmonary lesions in some focused respiratory models [95–97]Long‑COVID can involve heart (myocarditis, dysautonomia), kidney dysfunction, and other organ involvement, but patterns vary widely and are often confounded by ICU care, pre‑existing disease, and therapies [95, 96, 116, 117]Partially similar: both show multi‑organ involvement, but hamsters mainly model the lung-kidney axis alongside the brain; the human spectrum is broader and clinically more complexViral persistenceBrainstem and lung harbor genomic RNA and low‑titer infectious SARS‑CoV‑2 up to 80–112 dpi in hamsters; some models show viral RNA in olfactory bulb and other CNS regions early, with long‑term immune activation even after loss of detectable infectious virus [35, 95, 96]Human studies detect viral RNA/proteins in tissues and fluids months after acute infection in some individuals, but true persistence of infectious virus in vivo is harder to prove; persistent antigen is a leading mechanistic hypothesis for Long- COVID [35, 95, 96, 118]Mechanistically similar (antigen/possible virus persistence), but direct proof of long‑term infectious virus is stronger in hamsters than in patientsImmune & proteomic remodelingHamsters show prolonged, tissue‑specific interferon and chemokine signatures (especially in the olfactory bulb), persistent cytokine heterogeneity in lung/brain/plasma up to 60 dpi, and a distinct, sex‑biased lung proteome at 31 dpi with ongoing complement/coagulation and mucus‑barrier changes [95, 97, 98]Long‑COVID patients exhibit chronic immune activation, altered interferon and chemokine landscapes, complement/coagulation pathway dysregulation, and sex differences in risk and presentation [7, 13, 27, 95, 97, 98]Strong conceptual overlap in persistent immune/proteomic perturbations and sex effects, though human signals are more heterogeneous and influenced by prior immunity and treatmentSex differencesLung proteome study shows marked sex‑biased long‑term changes: at 31 dpi, males have many more proteins decreased and 38 longitudinally altered proteins vs. only 3 in females, despite similar acute responses [97]Epidemiology suggests females more often report PCC, while males more often develop severe acute COVID‑19; immune and proteomic sex differences are reported [13, 97, 98, 119]Both species show sex‑modulated responses, but direction and clinical expression differ, emphasizing the need to model sex as a key variable Table 3Translational comparison of lung–brain–immune axis and multi-organ pathophysiology in Long-COVID across Golden Syrian Hamster model and humansPathophysiologic DomainExperimental Findings in Golden Syrian HamstersClinical and Molecular Findings in HumansTranslational AlignmentBrain & cognitive (behavior)Golden Syrian Hamsters show persistent depression‑like behavior, late‑onset anxiety, and impaired short‑term recognition memory up to ~ 80 dpi after intranasal SARS‑CoV‑2, linked to brainstem changes [35]Long‑COVID cohorts report fatigue, brain fog, memory problems, anxiety, and depression months post‑infection, often fluctuating over time [7, 35, 110]Strong behavioral parallel; hamsters lack psychosocial confounders, but only approximateBrain & cognitive (molecular)Brainstem RNA‑seq reveals a neurodegeneration-like signature (inflammasome/innate genes, synaptic and mitochondrial pathways, PD/AD‑linked genes) persisting to 80 dpi; CNS regions show long‑lasting metabolic and synaptic transcriptional changes after SARS‑CoV‑2 vs. IAV [35, 95]Neuroimaging and fluid‑biomarker studies in patients suggest accelerated brain aging, microstructural changes, and possible neurodegenerative trajectories after COVID‑19 [35, 95, 111–113]Conceptually similar molecular signatures indicating neurodegenerative risk; human evidence is still emergingOlfactory systemHamsters exhibit acute anosmia plus persistent IFN‑I/II, chemokine (CXCL10, CCL5), and microglial/myeloid activation signatures in olfactory bulb and epithelium weeks after viral clearance [95]Recovered patients can have prolonged smell loss; human olfactory bulb and epithelium show persistent interferon, complement, and T‑cell/chemotaxis programs months after COVID‑19 [95, 114]Highly similar inflammatory/transcriptomic patterns in olfactory tissues.Neuro‑PCC heterogeneityLong‑term study upto 60 dpi finds subsets of hamsters with persistent or late‑onset deficits in burrowing (motivation) and anxiety‑like behavior, alongside variable brain and plasma cytokine profiles and vagus‑nerve changes; others recover, mirroring individual variability [98]PCC is clinically heterogeneous, with some individuals severely affected and others minimally symptomatic, and variable immune/biomarker signatures over time [98, 115]Strong qualitative similarity in heterogeneityRespiratory structure (alveoli)Multiple studies show dysregulated alveolar regeneration: fibrosis and bronchiolization with SCGB1A1⁺ club‑cell–rich foci, M2‑like macrophages, pro‑fibrotic transcripts; subpleural and interstitial fibrosis persists up to 112 dpi [96, 97]Long‑COVID and post‑COVID patients exhibit impaired diffusion, exertional dyspnea, reduced exercise capacity, and CT/histologic evidence of incomplete alveolar repair, bronchiolization, and fibrosis months after infection [33, 34, 96]Strong structural and functional parallels, especially for respiratory Long‑COVID; human changes can be more variable and influenced by comorbid lung diseaseMulti‑organ injurySARS‑CoV‑2–infected hamsters show persistent lambertosis and inflammation in the lung, tubular atrophy in the kidney, and subtle heart changes; some models report systemic proteomic shifts and kidney damage persisting to 31 dpi, with no major extra‑pulmonary lesions in some focused respiratory models [95–97]Long‑COVID can involve heart (myocarditis, dysautonomia), kidney dysfunction, and other organ involvement, but patterns vary widely and are often confounded by ICU care, pre‑existing disease, and therapies [95, 96, 116, 117]Partially similar: both show multi‑organ involvement, but hamsters mainly model the lung-kidney axis alongside the brain; the human spectrum is broader and clinically more complexViral persistenceBrainstem and lung harbor genomic RNA and low‑titer infectious SARS‑CoV‑2 up to 80–112 dpi in hamsters; some models show viral RNA in olfactory bulb and other CNS regions early, with long‑term immune activation even after loss of detectable infectious virus [35, 95, 96]Human studies detect viral RNA/proteins in tissues and fluids months after acute infection in some individuals, but true persistence of infectious virus in vivo is harder to prove; persistent antigen is a leading mechanistic hypothesis for Long- COVID [35, 95, 96, 118]Mechanistically similar (antigen/possible virus persistence), but direct proof of long‑term infectious virus is stronger in hamsters than in patientsImmune & proteomic remodelingHamsters show prolonged, tissue‑specific interferon and chemokine signatures (especially in the olfactory bulb), persistent cytokine heterogeneity in lung/brain/plasma up to 60 dpi, and a distinct, sex‑biased lung proteome at 31 dpi with ongoing complement/coagulation and mucus‑barrier changes [95, 97, 98]Long‑COVID patients exhibit chronic immune activation, altered interferon and chemokine landscapes, complement/coagulation pathway dysregulation, and sex differences in risk and presentation [7, 13, 27, 95, 97, 98]Strong conceptual overlap in persistent immune/proteomic perturbations and sex effects, though human signals are more heterogeneous and influenced by prior immunity and treatmentSex differencesLung proteome study shows marked sex‑biased long‑term changes: at 31 dpi, males have many more proteins decreased and 38 longitudinally altered proteins vs. only 3 in females, despite similar acute responses [97]Epidemiology suggests females more often report PCC, while males more often develop severe acute COVID‑19; immune and proteomic sex differences are reported [13, 97, 98, 119]Both species show sex‑modulated responses, but direction and clinical expression differ, emphasizing the need to model sex as a key variable Translational comparison of lung–brain–immune axis and multi-organ pathophysiology in Long-COVID across Golden Syrian Hamster model and humans Fig. 2Translational alignment of lung-brain-immune axis and multisystem sequelae in Long-COVID across humans and Golden Syrian Hamster model studies. The schematic highlights five principal pathological domains impacted by Long-COVID: brain and cognitive dysfunction, olfactory system impairment, lung structural and functional alterations, multi-organ and metabolic effects, and viral persistence and immune dysregulation Translational alignment of lung-brain-immune axis and multisystem sequelae in Long-COVID across humans and Golden Syrian Hamster model studies. The schematic highlights five principal pathological domains impacted by Long-COVID: brain and cognitive dysfunction, olfactory system impairment, lung structural and functional alterations, multi-organ and metabolic effects, and viral persistence and immune dysregulation Together, these cross‑species parallels establish Golden Syrian Hamster as a highly translational model for mechanistic dissection of lung–brain–immune axis dysfunction and preclinical testing of targeted therapies aimed at mitigating Long-COVID. Lung pathology With respect to the pulmonary pathological parameters described in "Lung pathology in Long-COVID patients" section, the Golden Syrian Hamster model faithfully recapitulates post-acute pulmonary trajectory observed in humans. Aged hamsters infected with the SARS-CoV-2 Delta variant develop acute broncho-interstitial pneumonia that resolves clinically yet leaves persistent structural and functional lung abnormalities up to 112 dpi, closely resembling respiratory dysfunctions observed in Long-COVID [96]. Despite clearance of replicating virus, subpleural and interstitial fibrosis persists, with increased collagen deposition in alveolar septa and widespread alveolar bronchiolization driven by expansion of airway-derived CK14 + basal cells and SCGB1A1 + club cells. Transcriptomic analyses reveal a transition from acute immune-dominated program responses to sustained activation of pro-fibrotic and extracellular matrix-remodeling pathways, indicating maladaptive lung repair [96]. Sex-specific remodeling of the lung proteome has been reported following resolution of acute infection with ancestral SARS-CoV-2 strains, reflecting epidemiologic sex differences observed in Long-COVID cohorts [97]. Physiological stress unmasks the functional relevance of this remodeling. Long-term treadmill challenge combined with plethysmography demonstrates exercise-induced impairment in lung function and reduced metabolic performance lasting approximately 7 weeks [96, 97]. In an independent ancestral-strain model, lung proteomics at 31 dpi identifies a distinct post-viral proteome signature involving complement and coagulation pathways and increased mucus-associated proteins (e.g., MUC5B, CLCA1), consistent with a chronically remodeled airway environment that predisposes to persistent respiratory symptoms [95, 96]. Together, these findings establish Golden Syrian Hamsters as a robust model of dysregulated alveolar repair and fibrotic lung disease that likely serves as an initiating node for lung–brain–immune axis dysfunction associated with Long-COVID. Brain pathology, neurocognitive and behavioral consequence Clinically, Long‑COVID is associated with fatigue, cognitive dysfunction, memory impairment, anxiety, and depression that may persist or fluctuate for months after infection [7, 110]. Neuroimaging and biomarker studies suggest accelerated brain aging, microstructural alterations, and potential neurodegenerative trajectories following COVID‑19 [111–113], implicating chronic neuroinflammation as a key mechanism as discussed in "Brain pathology and neurocognitive consequences in Long-COVID patients" section. In Golden Syrian Hamsters, intranasal SARS-CoV-2 infection leads to sustained brain pathology despite apparent systemic recovery. Viral genomic RNA and low levels of infectious virus are detectable in the brainstem up to 80 dpi, accompanied by distinct glial phenotypic changes, indicating prolonged viral persistence and delayed clearance within central autonomic and sensory circuits [35]. Brainstem transcriptomic analysis reveals a biphasic response: an early phase dominated by innate immune and interferon signaling, followed by a chronic phase characterized by persistent disruption of dopaminergic and glutamatergic synaptic signaling, mitochondrial energy metabolism, proteasome function, and multiple neurodegeneration-related KEGG pathways, including those linked to Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis or ALS [35, 95]. A similar pattern of persistent inflammation has been reported in human studies, with differentially expressed genes (DEGs) linked to neurological disorders and traits [120, 121]. Importantly, within the olfactory neuroepithelium, only sustentacular cells–not olfactory sensory neurons are susceptible to SARS-CoV-2 infection [54]. This cellular tropism provides a mechanistic explanation for anosmia without requiring direct neuroinvasion–a process congruent with human observations, because sustentacular cells maintain the function of olfactory sensory neurons but do not offer a direct pathway to the central nervous system. Blood-brain barrier (BBB) disruption was seen in both humans and hamsters in the same study, and in Long-COVID cohorts, BBB damage has been linked to cognitive impairment [55]. Thus, BBB breakdown likely represents a critical molecular intermediary linking peripheral immune activation to CNS neuroinflammation, rather than direct viral invasion of the brain. These molecular changes are accompanied by sustained astrocyte and microglial activation in brainstem regions and are temporally associated with delayed-onset of behavioral phenotypes, including depression-like behavior, anxiety, and impaired recognition memory. The coupling of persistent lung pathology with chronic brainstem neuroinflammation supports involvement of lung–brain–immune axis in developing neuropsychiatric and cognitive sequelae in Hamster model. Notably, similar transcriptional signatures have been identified in lungs of deceased COVID-19 patients, further strengthening the translational relevance of these findings [35, 95]. Long-COVID-like neurobehavioral outcomes in Golden Syrian Hamsters are assessed using longitudinal assays that capture activity, affective state, cognition, and exercise tolerance. These phenotypes align with molecular signatures of synaptic dysfunction, neurodegenerative pathway activation, and glial activation in the brainstem [35]. Respiratory-linked functional impairment is revealed through weekly treadmill running and plethysmography. Previously infected older male hamsters exhibit reduced willingness to run, lower performance scores, abnormal breathing patterns, and altered oxygen utilization for several weeks post-infection paralleling exertional intolerance and post-exertional symptom exacerbation reported in respiratory Long-COVID [35, 95, 96]. Additional behavioral studies reveal sex- and time-dependent effects. Burrowing behavior increases transiently at 25 dpi in the males, before declining below baseline by 59 dpi in several males and one female, suggesting dynamic alterations in motivation or activity. In open-field testing, a subset of animals - particularly females - spend less time and distance in the center zone at specific time points, consistent with anxiety-like behavior, although group averages may resemble controls. Overall locomotor outcomes exhibit increased variability rather than uniform impairment, reflecting the heterogeneity observed in Long-COVID. Importantly, further validation of certain behavioral paradigms in hamsters remains necessary [98]. Immune features and systemic dysregulation Long-COVID is characterized by chronic immune activation, altered interferon and chemokine landscapes, complement and coagulation dysregulation, and sex-dependent differences in risk and presentation as discussed in "Immune dysregulation in Long-COVID patients" section [7, 13, 27]. In the lung of SARS-CoV-2 infected hamsters, chronic phase of disease features persistence of M2-like macrophages, activation of pro-fibrotic programs (e.g., collagen deposition, and TGF-β signaling), extracellular matrix remodeling, and angiogenic signaling, despite minimal observable inflammation- feature that closely align with respiratory Long-COVID pathology. Systemically, SARS-CoV-2 infected hamsters display early elevations in IFN-γ, IL-1β, IL-6, MCP-1, TNF-α, and other cytokines at 4 dpi, followed by partial normalization. However, MCP-1, VEGF, and IL-1β remain variably elevated up to 60 dpi. A durable adaptive immune response is also evident, with sustained neutralizing antibody titers and detectable SARS-CoV-2-specific IFN-γ T-cell responses to spike and nucleocapsid antigens at 60 dpi [98]. Persistent immune activation extends beyond the lung to the olfactory system and brain. Even in the absence of detectable infectious virus, the olfactory epithelium and bulb show sustained myeloid and T-cell infiltration, elevated proinflammatory cytokines, and interferon signaling that correlate with behavioral alterations lasting at least one month after viral clearance. These molecular signatures mirror transcriptional changes reported in recovered COVID-19 patients with ongoing neurological symptoms, supporting a mechanistic link between respiratory infection, sensory pathway inflammation, and central nervous system dysfunction [95]. At 60 dpi, hamsters exhibit persistent viral RNA, prolonged immune dysregulation, and behavioral changes, although limitations in validated cognitive tests are acknowledged [98]. Within the nervous system and sensory tract, the brainstem retains ISGs and neuroinflammation‑related transcriptional signatures at 80 dpi. Concurrently, the olfactory epithelium and olfactory bulb exhibit prolonged myeloid and T‑cell activation, with persistent IFN-γ and chemokine signaling (e.g., ISG-15, MX genes, CXCL10, CCL5) despite the absence of culturable virus. These immune and immunometabolic alterations provide a mechanistic framework for sustained neuropsychiatric, cognitive, and sensory symptoms observed in the hamster model of Long-COVID [35]. Involvement of other organs (heart, kidney, intestine) In humans, Long-COVID manifests as a multisystem disorder, with involvement of cardiovascular dysfunction (e.g., myocarditis, dysautonomia), kidney injury, and other organ systems. However, these patterns are highly variable and frequently confounded by pre‑existing co-morbidities, and status of hospitalization, and intensive care treatment during acute illness [116, 117]. In contrast, hamster model demonstrated a more organ-specific PASC phenotype centered on lung and brain involvement. In long-term Delta infected aged hamsters show persistent viral RNA and lung pathology, whereas heart, kidney, and other organs display only sporadic or mild age-related changes without consistent evidence of chronic SARS-CoV-2-induced injury [96]. Comparative studies using ancestral SARS-CoV-2 and influenza virus demonstrate renal tubular atrophy and interstitial edema at 31 dpi, with more severe injury following SARS-CoV-2 infection, while cardiac tissue largely recovers without significant chronic myocarditis or fibrosis [96]. These findings indicate limited kidney involvement and minimal chronic cardiac pathology in current hamster models. Intestinal and other systemic organs are not consistently identified as chronic targets. Thus, available evidence supports Golden Syrian Hamster as an experimental model for lung- and brain- and immune-dominant PASC phenotype, with secondary renal effects, rather than the full multisystem spectrum observed in some Long-COVID patients [95–97]. Conclusions and future perspectives Animal models are indispensable for elucidating the pathogenesis of SARS-CoV-2 infection and the mechanisms driving its long-term sequelae. While small animal models-including mice, ferrets, and hamsters offer accessibility and cost effectiveness, and large animal models, such as NHPs, provide closer physiological similarity to humans, none of them fully captures the complexity of Long-COVID. Importantly, accumulating evidence indicates that Long-COVID is best understood through a lung-brain-immune axis framework, in which persistent pulmonary injury, immune and vascular dysregulation, and downstream neuroinflammation jointly sustain chronic symptoms. Mouse models have contributed valuable mechanistic insights into acute infection and immune responses; however, their utility for Long-COVID remains limited. Most murine models require genetic modification of the host or virus, often produce exaggerated disease phenotypes, and incompletely reproduce the pulmonary-neurological coupling characteristic of human Long-COVID. Although cost-effective and experimentally tractable, these limitations constrain their relevance for modeling chronic dysfunction along the lung-brain-immune axis. Ferret models excel in studying respiratory transmission and acute airway infection but typically exhibit mild disease and lack persistent clinical or neurobehavioral phenotypes. Short follow-up windows, higher costs, and handling complexity further limit their applicability to Long-COVID, particularly for investigating sustained pulmonary pathology and its neurological consequences. In contrast, Golden Syrian Hamsters emerge as a uniquely valuable translational model. Critically, they develop persistent post-acute lung remodeling, fibrotic repair, immune dysregulation, and neuroinflammation, enabling a mechanistic study of lung-derived signals that propagate to the brain. While differences between hamster and human immune systems and limited availability of species-specific immunological reagents remain challenges, particularly for immune function characterization, ongoing tool development is likely to mitigate these limitations. The enduring burden of Long-COVID underscores the urgent need for models that can interrogate disease heterogeneity and causality. Future work in Golden Syrian Hamsters should incorporate pre-existing comorbid conditions known to increase Long-COVID risk and severity, including biological sex, age, obesity, diabetes, cardiovascular disorders, chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), and metabolic syndromes [21, 22, 122, 123]. Modeling these pre-existing disorders will enable dissection of inflammatory states amplify lung injury and shape downstream neuroimmune outcomes. Longitudinal studies in aged hamsters are especially promising for defining how age-dependent declines in tissue repair, immune regulation, and mitochondrial function influence chronic lung–brain–immune pathology. Another critical frontier lies in neurobehavioral and psychological phenotyping. Cognitive function, anxiety, depression, sleep disturbance, and fatigue are among the most debilitating symptoms reported by Long-COVID patients, yet remain very little explored in animal models [124]. Expanding and validating behavioral assays in hamsters coupled with molecular analyses of synaptic integrity, microglial activation, and metabolic stress will be essential for linking pulmonary pathology to persistent brain dysfunction and for testing targeted interventions [107, 125, 126]. NHP models have provided persistent metabolic, pulmonary, and neurobehavioral changes following SARS-CoV-2 infection and remain the most physiologically faithful systems for studying lung-brain-immune axis dysfunction. However, their limited use in Long-COVID research reflects substantial barriers related to cost, logistics, and extended study durations. As such, NHPs are best positioned as complementary validation platforms for hypotheses generated in scalable small-animal models. In summary, Golden Syrian Hamsters represent a well-balanced and translationally relevant model for studying Long-COVID through a lung-brain-immune axis framework. Their natural susceptibility to SARS-CoV-2, capacity to recapitulate persistent pulmonary injury and neuroinflammation, experimental tractability, and suitability for longitudinal and risk-stratified studies position them as a powerful platform for defining causal mechanisms and accelerating development of targeted therapies to mitigate Long-COVID.