Supplementation of αKG along with TGFβ/HIF1α antagonist rescues lung fibrosis in mice with severe SARS-CoV-2 infection

Abstract

Background

Pulmonary fibrosis and acute respiratory distress syndrome (ARDS) are the clinical challenges among patients with post-acute sequelae of COVID-19 (PASC), affecting ~ 10% of the survivors worldwide. PASC manifestations include pulmonary fibrosis, impaired respiratory function, and long-term morbidity. Recent WHO reports from 83 countries also describe severe morbidity among COVID-19 survivors associated with PASC, indicating an unmet need to address the issue. We developed a mouse model showing the above disease pathophysiology, and investigated the related signalling molecules, with a focus on developing targeted treatment strategies against it.

Methods

We developed the mild and severe SARS-CoV-2 infection mice models. We took the opportunity to transiently express high/low levels of hACE2 in the airways and lungs of the mice, to manipulate infection severity and longer survivability. The severe infection model developed ARDS pathogenesis and pulmonary fibrosis, limiting respiratory functions (assessed by dual-chamber plethysmography), and survived beyond 30 DPI. The signalling molecules of these pathogenic pathways were identified from related tissue/blood samples of the mice using mass spectrometry-based proteome analysis, histopathology, immunoblotting and flow cytometry.

Results

The proteomic analysis revealed key pathways driving epithelial-mesenchymal transition (EMT) and pulmonary fibrosis in mice lungs. Elevated expression of P-AKT, TGFβ and HIF1α supported the severity of ARDS and fibrosis in lungs at 15 as well as 30 DPI. We developed a treatment strategy by targeting the above signalling molecules. Treatment with dietary alpha-ketoglutarate (1% αKG daily), known suppressor of P-AKT, along with inhibitor to TGFβ (SB431542,10 mg/kg BW) or HIF1α (CAY10585, 10 mg/kg BW/daily, delivered intraperitoneally) till 7 DPI, significantly reduced EMT and fibrosis by suppressing the PAKT: TGFβ:HIF1α signalling, and finally improved mice survivability (50%) beyond 50 DPI with a restored pulmonary function.

Conclusion

The study thus demonstrated the upregulation signalling molecules like P-AKT, TGFβ and HIF1α and their association with pulmonary fibrosis in mice after severe SARS-CoV-2 infection. Further, the inhibition of the above pathophysiology by the co-administration of αKG along with HIF1α/TGFβ antagonist highlighted a targeted therapeutic strategy against it. Importantly, for the first time, our study described the HIF1α axis as the therapeutic target in rescuing post-acute sequelae such as ARDS and fibrosis in mice with severe SARS-CoV-2 infection.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-08700-2.

Introduction

The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has placed an immense burden on global health [1]. According to the WHO, an estimated 7.1 million deaths have been reported globally between January’2020 till May’2026. While most infections are mild, severe cases can progress to acute respiratory distress syndrome (ARDS), a life-threatening condition and a major cause of intensive care admission and mortality [2]. ARDS is characterized by pulmonary inflammation, alveolar injury, and fluid accumulation [3]. Beyond the acute phase, post-acute sequelae of COVID-19 (PASC) or long COVID have become a significant concern, affecting ~ 10% of documented cases worldwide, with higher incidence in hospitalized patients and notable prevalence even among vaccinated individuals [4]. Among PASC manifestations, pulmonary fibrosis is particularly alarming, defined by irreversible lung scarring, impaired respiratory function, and increased long-term mortality [5]. Clinical studies have shown that many patients exhibit persistent dyspnea, dry cough, fatigue, and reduced diffusion capacity of the lungs for carbon monoxide (DLCO, a test indicating that lungs are not transferring oxygen from inhaled air to the bloodstream) months after infection, even when pulse oximetry and spirometry appear normal. Longitudinal cohorts report a decrease in DLCO in up to 42% of patients at 60–100 days post-discharge, with those having severe acute illness showing the greatest impairment [6]. Notably, imaging findings often parallel these functional deficits, with residual ground-glass opacities and fibrotic changes on CT scans, while exercise capacity remains compromised in some patients despite rehabilitation [7, 8]. These observations highlight the lasting impact of post-COVID pulmonary fibrosis and emphasize the urgent need for further research into its prevalence, pathophysiology, and targeted anti-fibrotic interventions [9].

Emerging evidence suggests that SARS-CoV-2 induced upregulation of epithelial–mesenchymal transition (EMT) or dysregulation of TGF-β signalling and mesenchymal to epithelial transition (MET) play central roles in this fibrotic remodelling. SARS-CoV-2 infection upregulates EMT-related genes such as ZEB1, AXL [10], SNAIL2, and vimentin, while downregulating epithelial markers like EPCAM, leading to loss of junctional integrity and enhanced motility [11–13]. Concurrently, oxidative stress [14] and epithelial injury triggers the overproduction of TGF-β [15] by infected epithelial cells and macrophages, which clears apoptotic debris [16–18]. Apart from TGF-β, a wide range of factors, including cytokines (TNF-α, IL-6), growth factors, and transcription factors such as HIF-1α, can influence the progression of EMT and fibrosis [14, 19, 20]. Elevated TGF-β drives downstream PI3K/AKT, ERK [21], and SMAD signalling, inducing TGF-β-dependent EMT and upregulation of profibrotic mediators such as β-catenin and type II collagen, ultimately promoting collagen deposition and alveolar remodelling [22, 23]. Notably, heightened TGF-β expression has been reported in COVID-19 patients, underscoring its potential as a therapeutic target. Interventions aimed at modulating EMT and TGF-β signalling may thus hold promise for mitigating post-COVID pulmonary fibrosis and improving long-term respiratory outcomes.

COVID-19 remains endemic in all countries around the globe, contributing significant morbidity and mortality. According to the WHO report for February-March 2025, 23 countries reported mortality, and 83 countries reported severe morbidity associated with new SARS-CoV-2 variants. Building on the response to SARS-CoV-2 outbreaks, the WHO’s R&D blueprint has facilitated a coordinated, accelerated response to COVID-19, including an unprecedented approach to develop vaccine and potential pharmaceutical treatments. In an attempt, we focused on investigating the above pathogenesis and developing targeted treatment strategies against post-acute sequelae such as ARDS and fibrosis using a mouse model of severe SARS-CoV-2 infection.

Materials and methods

Cells and virus

Vero E6 (CRL-1586, ATCC, USA) cells were cultured in Dulbecco’s modified eagle’s medium (DMEM; Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (Sigma-Aldrich, USA). All cell lines were authenticated before use and monitored for mycoplasma contamination, and the experiments were performed in mycoplasma free cells. The SARS-CoV-2 strain USA-WA-1/2020 (obtained from the University of Texas Medical Branch, USA Table S1) was utilized at a multiplicity of infection (MOI) of ~ 0.1, as detailed in our previous work [24–26].

Virus preparation

SARS-CoV-2 was propagated in Vero E6 cells cultured in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin. The cells were infected with SARS-CoV-2 at an MOI of ~ 0.1 and incubated at 37 °C in a 5% CO₂ for 1 h, with gentle rocking to ensure uniform infection. After the infection period, the inoculum was removed, and the cells were overlaid with DMEM containing 2% FBS. The cultures were incubated at 37 °C with 5% CO₂ for 48 h. Post-incubation, the supernatant was collected and clarified by centrifugation at 2000 rpm for 10 min at 4 °C, followed by filtration through a 0.2 μm membrane. The purified viral supernatant was aliquoted and stored at -80 °C. The viral titter was determined by plaque-forming units (PFU) [24–26].

Plaque-forming assay

The SARS-CoV-2 virus stock was serially diluted to infect Vero E6 cells for 1 h. Following infection, cells were overlaid with 2% CMC and incubated at 37 °C with 5% CO₂ for 2 days. After removing the CMC, cells were washed with PBS and fixed with 4% paraformaldehyde for 20 min. Cells were then stained with 1% crystal violet for 10 min and washed with tap water. Plaques were counted to determine the viral titter [24–26].

Mouse infection model

We took the opportunity of transiently expressing high/low levels of hACE2 in the airway and lungs of the mice, to manipulate the SARS-CoV-2 infection severity with a longer survival. Balb/c mice were housed under pathogen-free conditions in the Experimental Animal Facility (EAF) at the Regional Centre for Biotechnology (RCB), Faridabad, India. For transient expression of human ACE2 (hACE2) in the respiratory tract, a replication-deficient adenovirus expressing hACE2 (AD5CMV-hACE2) was administered through intranasal and intratracheal routes as mentioned in our previous work [27]. Intratracheal administration involved anesthetizing the mice with ketamine (100 mg/kg) and xylazine (10 mg/kg), followed by carefully removing the hair from the neck region. A 5 mm incision was made using a sterile scalpel to expose the trachea, and 50 µL of 2 × 10⁸ PFU of hACE2 adenovirus was administered using a 24-gauge IV cannula. The incision was then sealed with tissue adhesive glue (Vetbond, 3 M Corp, USA) to ensure proper recovery [28, 29]. In parallel, another group of mice received 25 × 10⁵ PFU/mL of hACE2 adenovirus intranasally. Five days post-adenovirus transduction, the mice were infected with 1 × 10⁶ PFU/mL of SARS-CoV-2 (strain USA-WA-1/2020). Post-infection, 1% α-ketoglutarate (αKG) was administered daily via oral gavage in drinking water until day 5 days post-infection (DPI) or 7DPI. Respiratory dysfunction, a key feature of ARDS, was monitored at two-day intervals using a dual-chamber plethysmograph to assess respiratory parameters till 15DPI and 30DPI [27]. Animals were euthanized at 5, 15, and 30DPI, and lung tissues were collected for downstream analyses to evaluate the pathological and molecular alterations induced by the infection. The experimental workflow for the development of a severe SARS-CoV-2 model is illustrated in Fig. 1a.

Fig. 1.

SARS-CoV-2-infected mice exhibited respiratory distress, inflammation, and thrombosis at 15 DPI. a, Schematic represents the experimental design of developing mild and severe SARS-CoV-2 mice infection model. Details are mentioned in the methods section. b, The body weight was decreased persistently in the severe infection as compared to the mild group. c, SARS-CoV-2 viral load was measured using Real time-PCR at 5 and 15 DPI. d-f, The respiratory parameters were measured using a dual-chamber plethysmograph at 0, 4, 10, and 14 DPI. Increased airway conductance (Sraw) and respiratory rate (RR), along with reduced tidal volume (TV), indicating diffuse alveolar damage (DAD) was observed in the severe group as compared to mild. Two-way ANOVA with Tukey’s multiple comparison was used. g, Images of hematoxylin and eosin (H&E) staining of the lung tissue at 15 DPI arrows indicating leukocyte infiltration h, Images of masson’s trichrome (MT) staining red colour indicates thrombus formation, blue staining denoted collagen deposition in the lungs at 15 DPI i, Histology score of tissue leucocyte infiltration, j, thrombus formation, and k, collagen deposition. The data are mean ± SEM. One-way ANOVA and Sidak’s multiple comparison test was used. l, the levels of pro-inflammatory cytokines (TNF-α, IL-6, IL-4, and IL-1β) in the lungs were up-regulated in the severe infection group as compared to mild at 15 DPI. Measured using cytometric bead array. The data are mean ± SEM. Two-way ANOVA with Tukey’s multiple comparison was used. For all figures, ns = non-significant, *P < 0.05, **P < 0.01, ***P < 0.001 & **** P < 0.0001

Assessing respiratory parameters using dual-chamber plethysmography (EMKA technologies)

Respiratory parameters were measured in mice using a dual-chamber plethysmograph (EMKA Technologies). Mice were acclimatized to the plethysmography chambers for 10 min prior to data collection to minimize stress-related artifacts. The system recorded parameters such as tidal volume (TV), respiratory rate (RR), specific airway resistance (Sraw), and specific airway conductance (SGaw) in unrestrained and conscious mice. The test was performed in a temperature-controlled environment (22–24 °C) with the chambers connected to a data-acquisition system. The respiratory signals were monitored for 20 min, and stable recordings were used for analysis. Data were processed using EMKA iox2 software to calculate respiratory dynamics, and results were expressed as mean values for each parameter [30, 31].

Real-time PCR

Total RNA was extracted from mouse tissues or cell pellets using RNAiso (Takara Bio, Japan), followed by phenol-chloroform treatment. First-strand cDNA was synthesized from 1 µg of RNA using a cDNA synthesis kit (BioRad, USA) according to the manufacturer’s protocol. The cDNA was then used for real-time PCR with SYBR Green Supermix (BioRad) on an Applied Biosystems Quant Studio™ 6 Flex Real-Time PCR System. Primer sets for gene detection are listed in Table S2.

Western blotting

Cells and mouse tissues were lysed using RIPA buffer (Sigma Aldrich) with 1X protease-phosphatase inhibitor (Thermo Scientific). Proteins separated by SDS-PAGE were transferred to a PVDF membrane and immunoblotted with primary antibodies against HIF1α, ACE-2, P-SMAD3, SMAD3, TGF-β, ZEB-1, E-cadherin, Vimentin, and β-Actin (mentioned in Table S3). Secondary HRP-conjugated anti-mouse and anti-rabbit IgG antibodies (Table S3) were used for blot development. Standard protocol is described in our previous work [24–26].

Flow cytometry

For immunophenotyping, lung tissues were harvested at 15DPI and 30DPI. Lung tissues were enzymatically digested using Liberase and DNase (Sigma) to obtain single-cell suspensions. Initially, the lungs were minced into small pieces and incubated with the enzymatic solution at 37 °C on an agitator hot plate until the tissue was fully dissociated. The cells were centrifuged at 1200 rpm for 10 min, and the pellet was washed with 1xPBS. The cell pellet was resuspended in PBS, counted, and stained with fluorophore-conjugated antibodies against CD45, CD11b, F4/80, and CD206 (Table S3) for 45 min at room temperature. Following staining, cells were fixed using a fixation buffer (Invitrogen), washed, and analyzed using a BD FACS Symphony A5SE flow cytometer (BD Biosciences). Data were processed and analyzed using FlowJo software (FlowJo LLC, Oregon, Table S5) [24–26].

ELISA

The levels of TGF-β proteins were measured in lung tissue lysate using an ELISA kit (Finetest, Table S4) according to the manufacturer’s protocol.

Cytometric bead array

The cytometric bead array (CBA) was performed to measure pro-inflammatory and anti-inflammatory cytokines TNF-α, IL-6, IL-10, IL-1β, and IL-4 (Table S4) from the lung tissue lysate collected from SARS-CoV-2 infected mice of different treatment groups and different time points, as described in the results, and analyzed by CBA analysis software (BD Biosciences) [24–26, 32].

Histopathology and analysis of inflammation, thrombosis, and fibrosis

The lungs of the mice were harvested and fixed in 4% paraformaldehyde (PFA), followed by paraffin embedding. Thin sections of 2.5 μm thickness were prepared using a semi-automatic microtome (Histocore Multicut, Leica Biosystem, Germany) and mounted on glass slides. These sections were stained with hematoxylin and eosin (H&E) and Masson’s trichrome (MT) according to the standard protocols of the facility. Leukocyte accumulation, serving as an inflammation marker, was evaluated by quantifying cellularity in lung sections. Images were captured at 10X and 20X magnifications using a Nikon Eclipse Ti-E inverted stage microscope. The cellularity score was determined using ImageJ software, in which the percentage of nucleated cell area relative to the total tissue section area was calculated as described. Additionally, the lung sections were subjected to immunohistochemistry staining for Vimentin (Cell Signaling Technology, USA) and Fibronectin (Abclonal). The analysis was performed using the deconvolution function of ImageJ software as per the established method [24–26].

Immunofluorescence staining

The lungs of the mice were harvested and fixed in 4% paraformaldehyde (PFA), followed by paraffin embedding. Thin sections, 2.5 μm in thickness, were prepared using a semi-automatic microtome (Histocore Multicut, Leica Biosystem, Germany) and mounted on charged glass slides. Slides were deparaffinized using xylene and rehydrated through a gradient of ethanol (100%, 95%, 70%, and 50%) for 5 min each, followed by a rinse with cold running water. Antigen retrieval was performed using a 10 mM citrate buffer (0.1 M citric acid and 0.1 M tri-sodium Sigma.) in an antigen retrieval chamber. To quench endogenous peroxidases, 3% H₂O₂ (EMPARTA) was applied dropwise onto the slides and incubated in the dark for 10 min, followed by two washes with 1xPBS. Blocking was carried out with 5% goat serum (Sigma) for 1 h at room temperature. The primary antibody, surfactant protein-C (Abcam), was diluted at 1:500 and applied to the slides, which were then incubated overnight at 4 °C. The next day, excess antibody was removed by washing the slides three times with 1xPBS. An anti-mouse secondary antibody (Invitrogen), diluted at 1:500, was applied and incubated for 2 h at room temperature. Excess secondary antibody was removed with two to three washes using 1xPBS. Finally, nuclei were stained with DAPI. The slides were visualized using a Stellaris confocal microscope, allowing detailed imaging of the tissue sections.

Targeted mass spectrometry

Targeted mass spectrometry was carried out in scheduled multiple reaction monitoring (MRM) mode on the QTRAP 6500+ (Sciex, MA, USA) mass spectrometer coupled with ExionLC (Sciex, MA, USA). The transition list for the targeted proteins was made using the SRMatlas Database. Initially, all the available transitions were scanned in an unscheduled MRM mode using a pooled protein sample of lung tissue lysate to optimize the retention time (RT) of the ions. The transitions with a proper gaussian peak shape and intensity of more than 100 counts per second (cps) were further selected and acquired in a scheduled MRM mode for final method optimisation. The optimized method consisted of 240 analytes corresponding to 42 proteins. The data was acquired with the target scan time of 1 s with a total number of 4202 scan cycles and declustering potential of 80. Peptide amount of 8 µg was injected on column name and were eluted at a flow rate of 0.2 ml/min in a linear gradient of 2% solvent B (100% ACN (v/v) with 0.1% formic acid) to 95% solvent B in 65 min with a total run time of 70 min. The raw MRM data were analyzed using the Analytics software integrated with Sciex OS platform (v 3.1.6.44, Sciex, MA, USA). The extracted peaks were selected based on expected RT, pre-processed by smoothing on a moving average, and filtered based on the following criteria: minimum peak height of 100 cps and minimum signal/noise ratio of 2. All selected spectral peaks were checked manually to ensure correct peak detection and accurate integration. The areas under the peak of selected transitions were used for all downstream processing. The raw area under the curves were pre-processed in Perseus (v 2.0.7.0) and statistical evaluation was performed using two-sided student’s t test. The differentially expressed transitions were filtered using the fold change cut off 1.3 for upregulation and 0.76 for downregulation with p-value < 0.05. The pathway enrichment for differentially expressed transitions was performed with the ShinyGO (v 0.82) online tool using the reactome database and represented using a chord plot. Further the heatmaps were generated using the heatmap package in RStudio (v 4.4.1) and principal component analysis was done using SRplot online platform.

Statistical analysis

All in vivo data are presented as scatter plots with mean ± SEM or column bar graphs with mean ± SEM. One-way ANOVA followed by multiple comparisons with Sidak’s test was used for data having normal distribution with one independent variable, and a two-way ANOVA multiple-comparison with Turkey’s test was used to compare grouped data with two independent variables, as indicated in the figures. All statistical analyses were performed using GraphPad Prism 10. The p-value < 0.05 was considered statistically significant. For all Figures, ns = non-significant, *P < 0.05, **P < 0.01, ***P < 0.001 & **** indicates P < 0.0001.

Results

Mice with severe SARS-CoV-2 infection showed high inflammation, injury, and EMT pathogenesis in the lungs, limiting respiratory function at 15 DPI

We developed a mouse model for severe SARS-CoV-2 infection. Wildtype BALB/c mice were given hACE2 adenovirus via intranasal (in case of mild infection model) or intratracheal (severe model, as mention in our previous work [27]) route, and mice were kept for 5 days to allow hACE2 expressing transiently, and given one-time infection with SARS-CoV-2 (~ 1 × 10^6 PFU/ml via intranasal route) to both mild and severe models (Fig. 1a). The severe model showed a persistent decline in body weight (Fig. 1b) and presence of viral load in the lungs (Fig. 1c). High inflammation and injury were observed in the lungs (Fig. S1a) at 15 days post-infection (DPI) as compared to the mild model. Respiratory function was measured using a plethysmograph as described (Fig. S1b). Airway resistance (Sraw) vs. conductance (SGaw, Fig. S1c), tidal volume (TV), and respiratory rate (RR) were measured (Fig. 1d-f). The respiratory rate was higher in the severe infection group than in the mild group at 14 DPI (Fig. 1f). The inflammation [defined by infiltration of inflammatory immune cells (Fig. 1g & i) and cytokine levels (Fig. 1l)] and thrombosis [defined by thrombus around blood vessels and collagen deposition (Fig. 1h & i-j)] were higher in the severe group at 15 DPI. Elevated lung inflammation was also measured in these severe infection groups at 5 DPI, as described in Fig. S1d-h.

Further, the proteomic analysis revealed the upregulation of EMT associated proteins (ZEB1, SNAI, TWIST, VIM, and CDH2), inflammatory cytokines (IL-12, IL-13 A, IL-17 A and IL-33) and fibrosis associated proteins (fibronectin (FN1), thrombospondin (THBS1), MMP3, MMP14, PDGF and FGF2) with a dysregulation in the expression of and surfactant-metabolism associated proteins (Sftb-B, Sftb-C and Sftb-D) in the severe infection group as compared to the mild group (Fig. 2a-c). The GO analysis described the major pathways, including cytokine response, extracellular matrix organization, surfactant metabolism, and mesenchymal to epithelial transition (MET) dysregulation in the severe group (Fig. 2b). The pathway enrichment analysis also highlighted the crucial involvement of crosstalk between TGFβ, AKT, and HIF1α in mediating inflammation, epithelial to mesenchymal transition (EMT) pathogenesis, and fibrosis in severely infected mice (Fig. 2c). We measured elevated TGF-β, a major driver of the EMT pathway, in the lungs in the severe SARS-CoV-2 infection group (Fig. 2g). EMT and fibrosis were confirmed by IHC in the lung tissues with elevated expression of vimentin and fibronectin in the severe group (Fig. 2d-f). The signalling cascade of EMT and fibrosis including HIF1α, P-SMAD3, ZEB1, vimentin, E-cadherin and Sftb-C was found elevated in the lungs of severely infected mice (Fig. 2h-i & Fig. S2a-g). Thus, indicating the progression of the EMT pathogenesis leading to pulmonary fibrosis in the severe infection group.

Fig. 2.

Mass spectrometry analysis showed epithelial-mesenchymal transition drives pulmonary fibrosis in severe SARS-CoV-2 infection group at 15 DPI. a, Heatmap showing differential protein expression (targeted mass spectrometry measurement) in lung lysate between mild vs. severe SARS-CoV-2 infection groups. The analysis highlighted the peptide transition of EMT/fibrosis/inflammatory pathways. b, The chord diagram illustrates the interaction between key proteins and signalling pathways, including extracellular matrix remodelling, mesenchymal-epithelial transition (MET), immune system activation, surfactant metabolism, dysregulation of extracellular matrix, platelet degranulation, and apoptosis, etc. c, Pathway enrichment analysis showing AKT, TGF-β, and HIF-mediated regulation of EMT, fibrosis, surfactant-associated alveolar functions, and inflammatory cytokines between mild vs. severe SARS-CoV-2 infection groups. d-f, Image showing immunohistochemistry of increased Vimentin (left panel) and fibronectin expression (right panel), and the quantification for both at 15 DPI in severe SARS-CoV-2-infected mice lungs. The data are mean ± SEM. One-way ANOVA and Sidak’s multiple comparison test was used. g, TGF-β levels, measured by ELISA, were elevated in severe group as compared to mild at 15 DPI. The data are mean ± SEM. One-way ANOVA and Sidak’s multiple comparison test was used. h, The expression of EMT signalling proteins, measured by western blotting, indicated upregulation of HIF-1α, P-SMAD3, ZEB1, vimentin, and decreased expression of E-cadherin in severe group as compared to mild at 15 DPI (Densitometry analysis is mentioned in Fig. S2 a-f). i, Immunofluorescence of Surfactant protein-C showed a decreased expression in the lungs of severe group at 15 DPI. For all figures, ns = non-significant, *P < 0.05, **P < 0.01, ***P < 0.001 & **** P < 0.0001

The above pathogenesis was even more pronounced and persistent in severe infection at 30 DPI

The extension of the above experiment till 30 DPI (Fig. 3a) even showed the persistent inflammation and EMT pathogenesis more prominently in the severe group. Although no viral load was detected in the lungs of these mice (Fig. S3a), the respiratory parameters, including timing of inspiration and expiration, airway resistance vs. conductance, and tidal volume, were impaired. The respiratory time was higher in the severe group than in the mild group (Fig. 3b-d, Fig. S3b). The lung injury and inflammation persisted, with elevated TNFα, IL4, and TGFβ levels (Fig. 3e, j), along with increased infiltration of inflammatory immune cells, thrombus formation around blood vessels, and collagen deposition (Fig. 3f-i) in the severe infection group at 30 DPI. Persistent fibrotic lesions were evidenced by elevated fibronectin deposition in the lung tissues of these mice (Fig. 3l-m). Further, elevation in EMT signalling cascades including HIF1α and TGFβ indicated the progression of chronic fibrosis in the severe infection group (Fig. 3k & Fig S3c-i).

Fig. 3.

Elevated HIF-1α and TGF-β, and epithelial-mesenchymal transition (EMT) and fibrosis in the lungs of severe infection group at 30 DPI. a, Schematic represents experimental design. b-d, The respiratory parameters were measured using a dual-chamber plethysmograph at 18, 22, 26, and 30 DPI. Severe infection group exhibited an increased airway conductance (Sraw) and respiratory rate (RR), alongside reduced tidal volume (TV), indicating diffuse alveolar damage (DAD) as compared to mild group. Two-way ANOVA with multiple comparisons was used. e, The levels of pro-inflammatory cytokines (TNF-α, IL-6, IL-4, and IL-1β) in the lungs were upregulated in the severe infection group as compared to mild at 30 DPI. The data are mean ± SEM. Two-way ANOVA with multiple comparison was used. f, Images of hematoxylin and eosin (H&E) staining (left panel) and Masson’s trichrome (MT) staining (right panel) of the lung tissue at 30 DPI, with arrows indicated leukocyte infiltration and thrombus formation. Blue staining indicated collagen deposition in the lungs of severe group at 30 DPI. g-i, Histology score of tissue leucocyte infiltration, thrombus formation, and collagen deposition. The data are mean ± SEM. One-way ANOVA with multiple comparison was used. j, TGF-β level was higher in severe group at 30 DPI as compared to mild. The data are mean ± SEM. One-way ANOVA with multiple comparison was used. k, The expression of EMT signalling proteins was measured by western blotting, which indicated upregulation of HIF-1α, TGF-β, P-SMAD3, ZEB1, vimentin, and decreased expression of E-cadherin in severe group as compared to mild at 30 DPI. Densitometry analysis is mentioned in Fig. S3a-f. l-m, Image showing immunohistochemistry of increased fibronectin expression in the lungs of severe group at 30 DPI. For all figures, ns = non-significant, *P < 0.05, **P < 0.01, ***P < 0.001 & **** P < 0.0001

Combined treatment with αKG supplementation along with antagonist to either TGFβ or HIF1α rescued EMT and fibrosis, and restored respiratory function, finally improved mice survivability in severe SARS-CoV-2 infection

To develop a therapeutic strategy against COVID-19, we used inhibitors/antagonists of major signaling molecules AKT, HIF1α, and TGFβ (Fig. 4a). We used the Krebs cycle metabolite αKG as an inhibitor of P-AKT and TGFβ (SB431542) or HIF-1α antagonists (CAY10585). Our previous work described the rescue effects of dietary αKG in SARS-CoV-2-infected hamsters/mice. We described that dietary αKG decreased lung inflammation in SARS-CoV-2 infected hamster by suppressing the P-AKT signalling [24–26]. We never assessed the long-term rescue effect of αKG treatment on a severe infection model before. As observed, αKG alone did not show a significant rescue effect in a severe model of SARS-CoV-2 infection in mice. However, the co-administration with dietary αKG along with the antagonist to either TGF-β (SB431542) or HIF-1α (CAY10585, for both ~ 10 mg/kg/daily, IP route) till 7DPI (Fig. 4a) improved the survivability (50%) of SARS-CoV-2 infected mice beyond 50 DPI (Fig. 4b). Either combination therapy significantly improved respiratory function by reducing respiratory time (Fig. 4c-e, Fig. S4a). The TGFβ level and the percentage of F4/80 + + and CD206 + + M2 macrophages, one of the major cell types that produce TGFβ in the lungs, were significantly reduced by either combination treatment (Fig. 4f-h). Inflammation [measured by infiltration of inflammatory immune cells, (Fig. 5a&b), and cytokine IL-1β, IL-4, IL-6 and TNF-α levels (Fig. 5g)] and thrombosis [defined by thrombus around blood vessels (Fig. 5a & d)]; and EMT pathogenesis and fibrosis [defined by collagen deposition (Fig. 5a & c), and expression of HIF1α, P-SMAD 3, ZEB-1, E-cadherin, vimentin, fibronectin and surfactant protein-C (SFTB-C) (Fig. 5e-f and h, Fig. S4 b-g)] were significantly altered in mice with combination treatment, indication a reduction in EMT pathogenesis and fibrosis. Data for 15 DPI, along with these inhibitor treatments, are described in supplementary Fig. S5, S6, and S7. Finally, combination treatment improved respiratory function and survival in these mice.

Fig. 4.

Combined treatment of αKG along with an antagonist to either TGF-β or HIF-1α rescued EMT and restored respiratory functions, and improved mice survivability in severe infection. a, Schematic representation of experimental plan. b, αKG supplementation along with antagonist to either TGF-β (SB431542) or HIF-1α (CAY10585) improved mice survivability. Both treatments improved 50% mice survival beyond 50 DPI as compared to 100% death in case of severe or severe + αKG (only) treated groups by 30–40 DPI. c-e, Plethysmograph data described significant reduction in airway resistance (Sraw), and respiratory rate (RR) along with increased tidal volume (TV) in both combination-treatment groups at 18, 22, 26 and 30 DPI. Data are represented as a box-and-violin plot. Two-way ANOVA with multiple comparison was used. f, the levels of TGF-β in the lungs, quantified by ELISA, were significantly reduced in the combination treatment with antagonists. Each dot represents one mouse. g, FACS image of the above analysis showed the F4/80 + + and CD206 + + M2 macrophage population in the lung single cell suspension of lung tissues of the mice. h, The percentage of F4/80 + + and CD206 + + M2 macrophages were significantly reduced upon combination treatment. The data are mean ± SEM; one-way ANOVA with Sidak’s multiple comparison was used. For all Figures, ns = non-significant, *P < 0.05, **P < 0.01, ***P < 0.001 & **** P < 0.0001

Fig. 5.

Combined treatment of αKG along with an antagonist to either TGF-β or HIF-1α rescued EMT and restored respiratory functions, and improved mice survivability in severe infection. In continuation of the Fig. 4, a, Microscopy images of H&E and Masson’s trichrome staining, Immunohistochemistry (IHC) for fibronectin, and Immunofluorescence (IF) staining of surfactant protein-C (SFTB-C) of the lung tissue at 30 DPI (~100 µm). Arrows indicated immune cell infiltration. b, The percentage cellularity score was calculated from 10 different fields of different animals. c-d, Quantification of thrombus and collagen scores at 30 DPI. e-f, Quantification of fibronectin and Sftb-C at 30 DPI. The data are mean ± SEM. One-way ANOVA and Sidak’s multiple comparison test were used. g, The levels of pro-inflammatory cytokines (IL-4, IL-6, and TNF-α) in the lungs were reduced significantly upon combination treatment. Data are presented as medians, and a two-way ANOVA and Tukey’s multiple-comparison test were used. h, Western blot analysis showed that the combination treatment decreased the expression of HIF-1α, P-SMAD3, ZEB-1, and vimentin, and increased E-cadherin, indicating a reduction in EMT pathophysiology (Densitometry analysis is shown in Figure S4 b-g). For all figures, ns= non-significant, *P < 0.05, **P < 0.01, ***P < 0.001 & **** indicates P < 0.0001

Thus, the study described that the co-administration of αKG along with either TGFβ or HIF1α inhibitor decreased inflammation, EMT pathogenesis, and fibrosis in the lungs of severe SARS-CoV-2 infected mice via suppression of the PAKT: HIF1α: TGFβ axis (Fig. 6).

Fig. 6.

Schematic representation of the study. The co-administration of αKG along with either TGFβ inhibitor (SB431542) or the HIF1α inhibitor (CAY10585) till 7 DPI, decreased PAKT expression, and reduced inflammation including immune cell infiltration, cytokine storm. The combination treatments decreased HIF1α stabilization and TGFβ levels alongside TGFβ-producing M2 macrophages, and attenuated EMT pathogenesis and lung fibrosis. Improved mice survivability. Thus, highlighting either of the combination treatments as the therapeutic regimen against SARS-CoV-2 induced lung fibrosis

Discussion

COVID-19 infection-related lung complications remain a substantial and unresolved health concern among ICU survivors [33]. A significant number of these survivors continue to experience symptoms like fatigue, dyspnea, cognitive dysfunction, and chest pain, resulting in persistent pulmonary abnormalities [34], including reduced diffusion capacity of the lungs (DLCO test), and developing symptoms like ARDS [2]. Studies describe fibrosis-like changes in the lung tissues [35], such as chronic inflammation, immune dysregulation [36], and impaired healing of injured endothelial tissues [37, 38]. Emerging evidence implicates the elevation of EMT and dysregulation of MET as a key pathological process driving this aberrant repair response and thereby contributing to progressive fibrotic remodelling in the lungs of COVID survivors [11, 13, 39, 40]. These observations underscore the unmet need for an effective therapeutic strategy to address COVID-associated pulmonary complications.

COVID-19 is now transitioning to an endemic globally, including India. A growing population of survivors continues to experience persistent and often unrecognized morbidity, despite advances in vaccines and therapeutics; the treatment regimen against long COVID remains unanswered. Notably, there are currently no approved treatment regimens that effectively prevent or reverse the progression of long COVID. Although extensive studies have reported the involvement of several pathways and signaling molecules in driving the above pathogenesis in patients, the gap between understanding disease pathogenesis and developing targeted treatment strategies remains unclear. To address this gap, we developed the mice model showing above disease pathogenesis in severe infection group as compared to mild group. That helped us to investigate the pathogenic molecules and targeted therapeutics against these.

To develop the mice model, we followed our previous protocol [27], by expressing the high/low level of hACE2 transiently in the airway and lungs of the mice, to manipulate the severity of SARS-CoV-2 infection. Unlike in K18 mice, this model provided the flexibility of inducing severe SARS-CoV-2 infection and a longer survivability [41–43]. This allows us to study the post-infection sequelae in this model. To develop this infection model, we used the SARS-CoV-2 Wuhan strain (USA-WA-1/2020), which is well standardized in our previous protocol [25–27]. Although the mutation landscape in current variants like BA3.2 [44] and KP.2 [45], and their pathogenicity is different from the above original strain, but we developed this model with a focus on investigating the signalling mechanism of severe pathogenesis and related post-infection sequelae [46].

The severe infection model showed a clear ARDS pathophysiology with a higher airway resistance/lower conductance, increased tidal volume, and decreased respiratory time at 15 and 30 DPI. These mice showed upregulation of EMT-associated proteins such as ZEB1, SNAI, TWIST, VIM, and CDH2; inflammatory cytokines; and fibrosis-associated proteins such as FN1, THBS1, MMP3, MMP14, PDGF, and FGF2, alongside dysregulation of surfactant metabolism-associated proteins Sftb-B, Sftb-C, and Sftb-D. These mice also exhibited elevated TGFβ and HIF1α, highlighting their roles in EMT pathogenesis and fibrosis. Several studies indicated that SARS-CoV-2 infection mediated the induction of EMT in mice [11, 12]. The literature also reported that SARS-CoV-2 induced EMT progression via TGFβ1 and ZEB1 signaling [11, 39, 40]. The TGFβ-mediated stimulation of alveolar macrophages, in turn, elevates inflammatory cytokines IL-4, IL-6, and IL-13 [22, 23]. The progression of lung fibrosis in SARS-CoV-2-infected mice was mediated partially by SPP1⁺ macrophages and enhanced TGFβ signalling [47]. Importantly, reports also suggested exacerbation of these pro-fibrotic immune and epithelial cells due to prolonged stabilization of HIF1α in patients with PASC-associated persistent lung injury [48].

To develop a targeted therapeutic strategy against EMT and fibrosis in the lungs, we supplemented SARS-CoV-2-infected mice with inhibitors of P-AKT, HIF1α, and TGFβ. Previously, we described that dietary αKG treatment decreased lung inflammation in SARS-CoV-2-infected hamsters/mice by suppressing the P-AKT signaling [24, 25]. The αKG supplementation increased the enzyme activity of prolyl hydroxylase 2 (PHD2) and suppressed AKT phosphorylation and HIF1α stabilization, thereby improving type I and II interferon (IFN) synthesis in hamsters/mice with a moderate SARS-CoV-2 infection [26]. As observed, αKG alone did not show a significant rescue effect in a severe model of SARS-CoV-2 infection in mice. However, co-administration of dietary αKG along with the antagonist to either TGFβ (SB431542) or HIF1α (CAY10585) till 7 DPI, improved EMT and fibrosis, and improved mice survivability (50%) beyond 50 DPI. The interplay between TGF-β and HIF1-α has been suggested to form a feed-forward loop in the regulation of pathophysiological conditions, including EMT and fibrosis [36–40, 47–49]. Unlike αKG, both small-molecule SB431542 (TGF-β inhibitor) and CAY10585 (HIF1-α inhibitor) are limited to laboratory and research use. Using these inhibitors, our study has validated the involvement of a specific signaling pathway, such as AKT: HIF1α:TGFβ, in the pathogenesis of EMT and fibrosis in the lungs of severely SARS-CoV-2-infected mice. The study also suggested that disruption of TGFβ-mediated extracellular matrix accumulation or HIF1α-mediated inflammatory surge might have contributed to the rescue effects observed in mice, including pulmonary inflammation and fibrosis, following SARS-CoV-2 infection.

Conclusion

We described the involvement of EMT pathogenesis leading to fibrosis in the lungs, and limiting respiratory functions in the severe SARS-CoV-2 infection group. We identified the key signalling axis P-AKT: HIF1α:TGFβ in the regulation of EMT-mediated lung remodelling. Based on these mechanistic insights, we designed a targeted therapeutic strategy against EMT and fibrosis. We demonstrated that co-administration of αKG along with HIF1α/ TGFβ antagonist significantly rescued the above pathogenesis in severely SARS-CoV-2 infected mice, and improved their survivability. Thus, indicating the above signalling axis as a potential therapeutic target against severe SARS-CoV-2 infection and related post-infection sequelae including pulmonary inflammation and fibrosis. For the first time, our study described the HIF1α axis as a potential therapeutic target in rescuing lung injury, inflammation and fibrosis in mice with severe SARS-CoV-2 infection.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: Garima Joshi, Prasenjit Guchhait; Methodology: Garima Joshi, Gulistan Parveen, Simran Kaur, Naman Kharbanda, Tejeswara Rao Asuru, Anupam Chawla; Validation: Garima Joshi, Tushar K Maiti, Prasenjit Guchhait.; Formal analysis: Garima Joshi, Naman Kharbanda; Investigation: Garima Joshi, Gulistan Parveen, Simran Kaur, Tejeswara Rao Asuru, Anupam Chawla; Data curation: Garima Joshi, Gulistan Parveen, Simran Kaur, Naman Kharbanda; Writing–original draft: Garima Joshi, Prasenjit Guchhait; Writing – review & all; Supervision and data review: Prasenjit Guchhait, Tushar K Maiti; Project administration: Prasenjit Guchhait; Funding acquisition: Prasenjit Guchhait.

Funding

This work was supported by the Department of Biotechnology (DBT), Government of India, under Mission COVID Suraksha (Project Ref. No. BT/CS0094/06/22). The funding agency did not influence the study design, data collection and analysis, interpretation of results, or the preparation of this manuscript.

Data availability

All data are mentioned in the mail and supplementary figures. Raw data of all figures will be available from the corresponding author on request.

Declarations

Ethics statement

Small animals: The Institutional Animal Ethics Committee (IAEC) of the Regional Centre for Biotechnology (RCB) approved the animal experiment protocol (RCB/IAEC/2023/155), and experiments using the Balb/c mouse strain (RRID: IMSR_JAX_000651) were conducted within the IAEC guidelines in our institute's Experimental Animal Facility (EAF). Biosafety: The Institutional Biosafety Committee (IBSC RCB/IBSC/22-23/463) of RCB and RCGM (BT/IBKP/053/2020) approved the protocol. All the COVID-19-related in vitro and animal experiments were performed in the BSL3 facility.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.