IL-4 and TGF-β regulate inflammatory cytokines and cellular infiltration in the lung and systemic IL-6 in mouse-adapted SARS-CoV-2 infection

Abstract

The pathology of severe COVID-19 is due to a hyperinflammatory immune response persisting after viral clearance. To understand how the immune response to SARS-CoV-2 is regulated to avoid severe COVID-19, we tested relevant immunoregulatory cytokines. Transforming growth factor β (TGF-β), interleukin (IL)-10, and IL-4 were neutralized upon infection with mouse-adapted SARS-CoV-2 (CMA3p20), a model of mild disease; lung inflammation was quantified by histology and flow cytometry at early and late time points. Mild weight loss and lung inflammation including consolidation and alveolar thickening were evident 3 d postinfection (dpi), and inflammation persisted to 7 dpi. Coinciding with early monocytic infiltrates, CCL2 and granulocyte colony-stimulating factor were transiently produced 3 dpi, while IL-12 and CCL5 persisted to 7 dpi, modeling viral and inflammatory phases of disease. Neutralization of TGF-β, but not IL-10 or IL-4, significantly increased lung inflammatory monocytes and elevated serum but not lung IL-6. Neutralization of IL-4 prolonged weight loss and increased early perivascular infiltration without changing viral titer. Anti-IL-4 reduced expression of Arg1, a gene associated with alternative activation of macrophages. Neutralizing TGF-β and IL-4 had differential effects on pathology after virus control. Lung perivascular infiltration was reduced 7 dpi by neutralization of IL-4 or TGF-β, and periairway inflammation was affected by anti-TGF-β, while alveolar infiltrates were not affected by either. Anti-IL-4 prolonged IL-12 to 7 dpi along with reduced IL-10 in lungs. Overall, the immunoregulatory cytokines TGF-β and IL-4 dampen initial inflammation in this mouse-adapted SARS-CoV-2 infection, suggesting that promotion of immunoregulation could help patients in early stages of disease.

Introduction

Infection with SARS-CoV-2 causes COVID-19, which has caused over 7 million deaths worldwide in 4 yr.¹ The immune response to SARS-CoV-2 infection in has 2 phases: a viral control phase and a hyperinflammatory phase, which can cause severe disease if not well regulated. The severity of COVID-19 is dependent on SARS-CoV-2 viral load, variant tropism, and the intensity and duration of the host immune response. The virus initially causes tissue damage in the upper respiratory tract, which can progress into the lungs.^2,3 After an illness lasting 1 to 2 wk, most patients resolve the infection; however, up to 40% of admitted patients require intensive care due to severe pulmonary complications, which can lead to death.⁴ During this period after acute virus infection, high levels of systemic inflammatory markers, such as interleukin (IL)-6 and complement-reactive protein, correlate with severity and prolonged hospital stays.^3,5 While there is evidence of persistent virus, the importance of prolonged inflammation in severe disease is supported by the first therapeutic successes, namely immunosuppressants, including corticosteroids and the Jak1/2 inhibitor baricitinib, which are still therapeutic after hospital admission.^6,7 This prolonged inflammatory phase of disease has unique aspects compared with influenza or respiratory syncytial virus (RSV), suggesting unique mechanisms of pathogenesis that remain poorly understood at the immunoregulatory level.⁸ Here, we used a mouse-adapted virus first described by Muruato et al.⁹ that peaks at day 2 of infection and is live in the lungs at day 4 but is undetectable by plaque assay or polymerase chain reaction (PCR) by day 7. This virus causes mild disease, suggesting that sufficient immunoregulation during the peak of infection regulates the potential for persisting immune-mediated damage.¹⁰

The immune response to SARS-CoV-2 infection begins with a limited interferon (IFN)-I response due to successful viral antagonism^11–13 in the presence of an early burst of tumor necrosis factor (TNF) and IL-1. This systemic inflammatory response coincides with lung infiltration consisting of neutrophils, monocytes, and dendritic cells, as occurs in humans.¹⁴ Severe disease corresponds to elevated production of inflammatory cytokines from vascular endothelium, respiratory epithelium, and immune cells.^3,15 Serum IL-6 is reliably associated with disease severity,¹⁶ though it is unclear what regulatory cytokines modulate IL-6 in vivo.^17,18 The correlation of IL-6 with disease severity led to early clinical trials studying potential COVID-19 therapies; however, anti-IL-6 antibodies were not successful as therapy once patients were admitted, suggesting that the role of IL-6 in monocytic inflammation occurs before patients arrive at the hospital.^19–21 Novel populations of monocytes are recruited to the lungs in severe COVID-19.²² CCL2 and its receptor CCR2 are essential for monocyte trafficking from the bone marrow under inflammatory conditions,²³ while CCL5 is thought to locally recruit inflammatory monocytes to tissues through their expression of CCR5.^24–26 CCL5 and CCL2 are implicated in COVID-19 pathogenesis, and monocyte-derived cells play a protective role against mouse-adapted SARS-CoV-2 (maSARS-CoV-2).^27,28

As it is so crucial to understand how to limit inflammation in the response to SARS-CoV-2, we targeted immunoregulatory cytokines transforming growth factor β (TGF-β), IL-10, and IL-4 to assess their roles in pathology in maSARS-CoV-2 infection. While this is not a model of severe disease, there is evidence of monocyte recruitment and lung inflammation.^14,23 TGF-β and IL-10 are the most well understood immunoregulatory cytokines, known to be essential in very inflammatory conditions such as extracellular parasite infections and inflammatory bowel disease.²⁹ We also tested IL-4 due to evidence that T helper 2 (Th2)–type cytokines and alternatively activated macrophages are involved in human and mouse SARS-CoV-1 and -2 infections.^30–37

Evidence of Th2-type immune response to SARS-CoV-2 includes evidence that eosinophilia and IL-13 correlate with severity in COVID-19,³¹ and that helminth infection reduces prevalence of COVID-19.³² In addition, IL-5 and IL-13 are expressed in patients and mice infected with SARS-CoV-2.³³ However, IL-5 and IL-13 are considered Th2 effector cytokines, and neither has been shown to regulate IFN-I or Th1-type inflammation.³⁴ On the other hand, the signature Th2 cytokine IL-4 can inhibit the synthesis of the proinflammatory mediators TNF-α, IL-1, IL-6, IL-12, and prostaglandin E2, and the chemokines IL-8 and MIP-1,³⁵ as well as the differentiation of Th1 lymphocytes.³⁶ As Th2 cytokines are also involved in wound repair,³⁷ we are interested in the role of these cytokines in the phase of disease that determines the early severity and then the later phase of disease in which lung tissue homeostasis is restored. Therefore, in this study we studied the role of immunoregulatory cytokines during the peak of infection, as well as after virus is largely cleared.

The data indicate regulatory roles for IL-4 and TGF-β infection with maSARS-CoV-2. Neutralization of TGF-β throughout infection dramatically increased cellular infiltration in all lung spaces at 3 dpi and increased systemic IL-6 and IL-22. Inhibiting IL-4 also increased early cell infiltration, prolonged weight loss, and increased late IL-12p40 and CCL5. We showed upregulation of Arg1 in lung tissue, suggesting alternative activation of macrophages in maSARS-CoV-2 infection, as previously shown in SARS-CoV–infected lungs.³⁰  Arg1 expression was dependent on IL-4, which was present in the lung at low levels before infection. Neutralizing either TGF-β or IL-4 changed the pattern of inflammation seen in the lungs at 7 dpi, with IL-4 limiting infiltration specifically at perivascular locations at the later time point. Regulatory mechanisms will need to be further investigated to understand their potential to contribute to limiting the severity of COVID-19.

Materials and methods

Mice and in vivo antibody injection

C57BL/6J mice were purchased from the Jackson Laboratory, maintained, and aged with ad libitum access to food and water under specific pathogen–free conditions in the Animal Resource Center facility at University of Texas Medical Branch (UTMB). Male mice 38 to 41 wk old were infected under animal biosafety level 3 conditions at UTMB or Rutgers New Jersey Medical School in accordance with institutional biosafety approvals. Mice were treated intraperitoneally with either 400 μg anti-mouse IL-4 (Clone 11B11; BP0045; Bio X Cell), 200 μg anti-mouse TGF-β (1D11.16.8; BE0057; Bio X Cell), or mixed isotype controls (400 μg rat IgG1 for IL-4, TNP6A7; BE0290, and 200 μg mouse IgG1 for TGF-β, MOPC-21; BE0083; Bio X Cell) every other day starting the day of infection (days 0, 2, and 4) under anesthesia. Isotype controls were also given to mock-infected animals. All animal experiments were carried out in accordance with the Institutional Animal Care and Use Committee guidelines and have been approved by both the UTMB and Rutgers Institutional Animal Care and Use Committees.

Virus infection

Mouse-adapted SARS-CoV-2 (CMA3p20) was grown in Vero E6 cells, as described by Muruato et al.⁹ Mice were anesthetized with 5% isoflurane and infected intranasally with 50 μL of media containing 1 × 10⁶ plaque-forming units of SARS-CoV-2 (CMA3p20). Mice in the mock group were given only phosphate-buffered saline intranasally. Health checks and body weight were recorded daily. Lungs and serum were collected for downstream analysis.

Single-cell suspension and flow cytometry

Lungs were collected in RPMI (11875-093) 10% v/v fetal bovine serum (FBS) (S12450H) and 1% v/v penicillin-streptomycin (both from Thermo Fisher Scientific), and washed in phosphate-buffered saline, followed by enzymatic digestion with collagenase D and DNase I (both from Roche) in serum-free RPMI for 30 min in a humidified 5% CO₂ incubator at 37 °C. After incubation, FBS was added to stop the enzymatic activity. The single-cell suspension was collected by passing the cells through 70 μm strainer; and red blood cells were lysed with RBC lysis buffer (Life Technologies). A total of 1 × 10⁶ cells were stained using the following antibodies: Anti-CD4-BV421 (clone RM4-5), CD25-BV785 (PC61.5), MHCII-BV605 (M5/114,15.2), CD163-BV421 (S15049I), CD206-PE (MR6F3), CD45-PerCP-Cy5.5 (30-F11), CD11c-PE (N418), Siglec-F-BV421 (S17007L), Ly6C-BV650 (HK1.4), and CD11b-BV780 (M1/70). Live/Dead fixable blue stain (Life Technologies) was used to exclude dead cells. After staining, samples were inactivated and fixed with 4% ultrapure formaldehyde (18814; PolySciences) for 24 h, and then the cells were resuspended in 2% ultrapure formaldehyde for another 24 h and acquired on the LSR II Fortessa Custom 2/5/3/5/2 (Becton Dickinson) and analyzed with FlowJo version 10.8.1 (Becton Dickinson), similar to Gonzalez-Orozco et al.¹⁰

Virus quantification

Infectious virus in lung tissue was quantified by plaque assay. Tissue was collected in 1 mL of Dulbecco’s Modified Eagle Medium (11965092; Thermo Fisher Scientific) and 10% FBS. Tissue samples were weighed and homogenized, and then quickly frozen. Six-fold serial dilutions of homogenates were prepared in duplicate and used to inoculate 4 × 10⁵ Vero E6 cells grown in a 6-well tissue culture plate containing 1 mL Dulbecco’s Modified Eagle Medium and 10% FBS. Following 1 h incubation at 37 °C 5% CO₂, the cells were overlayed with 1.6% Tragacanth solution (G1128; Sigma-Aldrich) supplemented with 2% FBS and 1× penicillin-streptomycin antibiotics and incubated for 2 d. Cells were then fixed in 10% formalin and plaques were visualized by staining with 1% crystal violet (V5265; Sigma-Aldrich) diluted in 10% ethanol (1070172511; Sigma-Aldrich). Viral infectivity titers were expressed as plaque-forming units per milliliter in Vero E6 cells and were calculated according to the Behrens-Karber method.

Real-time quantitative PCR

Total RNA was isolated from lung using Direct-zol RNA MiniPrep (Zymo Research) according to the manufacturer’s protocol. Reverse transcription was conducted using high-capacity cDNA reverse transcription kit (Applied Biosystems) according to the manufacturer’s protocol. Primers and SYBR Green mix were purchased from Bio-Rad and some primers were from Origene. DNA was amplified with QuantStudio 6 Flex real-time PCR System (Applied Biosystems).

Cytokine bead assay

Luminex Intelliflex (Thermo Fisher Scientific) using MAGPIX beads was run according to the manufacturer’s protocol. High-sensitivity (IFN-γ, IL-2, IL-4, IL-6, and TNF-α) and custom cytokine (IL-1α, IL-10, IL-12-p40, IL-13, IL-18, IL-22, IL-25, IL-27, IFN-β, CCL2, CCL5, CXCL2, TSLP, and granulocyte colony-stimulating factor [G-CSF]) panel kits (Bender MedSystems) were processed according to the manufacturer’s instructions. The analysis was performed using the Luminex Intelliflex analysis software.

Histology

The right inferior lobe of the lung was fixed in 10% neutral buffered formalin (HT501128; Sigma-Aldrich) for 7 d. Tissues were cut, paraffin blocks were made, and hematoxylin and eosin (H&E) staining performed by the Anatomic Pathology Laboratory of the Pathology Department of University of Texas Medical Branch. The pathology scoring was conducted in a blinded manner. Pathology score was determined using an observational scale developed from the sections. Interstitial thickening was scored 1 if yes plus 1 if <20% of lung affected, 2 if 25% to 50%, and 3 if 60% to 75%. Peribronchial infiltrates were scored 1 if yes plus scored 1 if yes plus 1 if 1 or 2 foci, 2 if 2 or 3, and 3 if 3 or more. Perivascular infiltrates were scored 1 if yes plus scored 1 if 1 or 2 foci, 2 if 2 or 3, and 3 if 3 or more. All infiltrates were scored 1 if consolidation was present, or 1 if 1 or 2 areas and 2 if 3 or more areas were present. Additionally, one was added for each of fibrolytic tissue, edema, or debris or alveolar fluid, if present. The maximum score observed was 19. Perivascular, periairway, and nuclei counting for thickness were done manually in a blinded manner.

Statistical analysis

FlowJo 10.8.1 and GraphPad Prism version 9.1.1 (GraphPad Software) were used to perform data and statistical analysis. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Intergroup comparisons were performed using nonparametric analysis. All data are expressed as the mean ± SEM. The statistical details of the experiments are provided in the respective figure legends. A P value <0.05 was considered statistically significant.

Results

Viral ORF1ab RNA was detected in the lung by real-time quantitative PCR (qPCR) at 2 dpi, but not at 7 dpi (Fig. 1A), and has been reported to remain detectable by PCR and plaque assay through day 4.³⁸ To characterize the immune response to this mouse-adapted virus infection, cytokines and chemokines were measured in lung homogenate and serum of infected mice by multiplex bead-based assay at 3 and 7 dpi. Cytokines and chemokines related to differentiation and recruitment of monocytes present in the serum were IL-6, G-CSF, and the chemokines CCL2 (MCP1) and CCL5 (RANTES).^39–41 The infection induced G-CSF and CCL2 on 3 dpi, while CCL5 increased compared with mock infected by 7 dpi (Fig. 1B). IL-10 was not increased above background at 3 dpi (Fig. 1C), while IL-12 p40 was increased by infection and remained elevated in isotype-treated mice until 7 dpi (Fig. 1D). IL-33, an epithelial and endothelial cell damage alarmin, was induced by the infection at 3 dpi but returned to baseline by 7 dpi (Fig. 1E).

Figure 1.

Weight loss and cytokine production in aged B6 mice infected with maSARS-CoV-2. C57BL/6 mice were infected at 10 mo of age with maSARS-CoV-2, or mock infected and treated with isotype control antibody on alternate days and weighed daily. Lung homogenate was taken for measuring cytokines on day 3 postinfection. Graphs show (A) viral load by qPCR; (B) monocyte-related cytokines and chemokines G-CSF, CCL2, and CCL5; (C) IL-10; (D) IL-12p40; and (E) IL-33. Results are representative of at least 2 experiments. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice in mock and isotype at 0, 3, and 7 dpi. A P value <0.05 was considered statistically significant. *P = 0.01 to 0.05; **P = 0.01 to 0.001. d, day.

Histopathological evaluation of the lungs on 3 dpi showed that mice infected with maSARS-CoV-2, that also received the isotype control antibody, had more leukocytes infiltration into the alveoli and around airways and vessels than mock group animals (Fig. 2A). There was also alveolar thickening and consolidation seen in the infected lungs at 3 dpi. The infiltration was determined microscopically to be highly monocytic, consistent with findings in humans.²² The pathology score shows that both isotype and anti-IL-4 had more severe lung damage than mock group at 3 dpi (Fig. 2B). No differences in lung pathology were observed between anti-IL-10– and isotype control–treated animals (not shown); therefore, not all of the samples from this group were processed for all assays.

Figure 2.

Lung pathology of maSARS-CoV-2–infected mice. C57BL/6 mice were infected at 10 mo of age with maSARS-CoV-2, or mock infected and treated with isotype control or TGF-β–neutralizing antibodies on alternate days and weighed daily. H&E staining of lungs of mock (top), isotype control (middle), and TGF-β (bottom). Results are representative of at least 2 experiments. Cellular infiltration, alveolar thickening, and consolidation were identified, and representative fields are shown. Scale bars represent 100 μm.

TGF-β regulates cellular infiltration into the lung

To determine the effects of TGF-β on acute and persistent inflammation in maSARS-CoV-2 infection, we studied the pathology of the lungs after neutralization of TGF-β. H&E staining of lungs from infected mice treated with anti-TGF-β showed an increase in leukocytes in the alveolar walls 3 dpi, with alveolar thickening and some consolidation, or filling of some alveolar airspaces (Fig. 2A). Thickening was consistent throughout the infected lung section, while the fraction of the lung consolidated in each animal was variable.

To determine the cellular composition of the infiltrate, flow cytometry was performed on lungs from infected mice 3 dpi. Neutralization of TGF-β increased infiltration of CD45⁺ leukocytes in the lungs. Three populations of CD11b⁺ cells were identified using side scatter characteristics (Fig. 3A). The least granular (SSC^lo) CD11b⁺ cells contained the Ly6C^hi (CD11c⁻) inflammatory monocytes, while the most granular, or vacuolated, cells (SSC^hi) contained the CD11c^+/int alveolar macrophages. Quantifying these results, we found that anti-TGF-β treatment increased infiltration of CD45⁺ leukocytes, with a specific increase in inflammatory monocytes (CD11b⁺SSC^loLy6C⁺) (Fig. 3B). The SSC^int population contained Ly6C^int cells and may be regulatory, as it decreased in the absence of IL-4. SSC^intCD11b⁺ cells were not further identified, and Ly6G was not included in the panel to identify neutrophils as they were rarely identified in H&E tissue sections. The number of alveolar macrophages (CD11b⁺SSC^hi) was decreased in the anti-TGF-β group.

Figure 3.

TGF-β regulates cellular infiltration in the lung. C57BL/6 mice were infected with maSARS-CoV-2 and treated with isotype or cytokine-neutralizing antibodies on alternate days and weighed daily. (A) Flow cytometry plots showing gating on CD11b⁺ live cells with 3 levels of side scatter in the lungs at 3 dpi and Ly6C and CD11c staining in those gates for identification; (B) quantification of infiltrating leukocytes (CD45⁺), inflammatory monocytes (CD11b⁺SSC^lo), undefined monocytic cells (CD11b⁺SSC^int), and alveolar macrophages (CD11b⁺SSC^hi). Results are representative of at least 2 experiments. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice per group. A P value <0.05 was considered statistically significant. *P = 0.01 to 0.05. ns, not significant.

TGF-β was detected in the lungs but was not induced by infection at 3 or 7 dpi (Fig. 4A). Neutralization of TGF-β did not change live virus in the lung tissue significantly compared with the isotype treated-infected group, though there was a significant trend downward in multigroup comparison (Fig. 4B). IFN-γ was induced 3 dpi but did not vary between infected groups (Fig. 4C). As IL-6 varies considerably between patients, correlates with disease severity,⁴² and also plays a role in monocyte differentiation;^20,43 we measured this critical cytokine in lung homogenate and serum by bead-based ELISA assay. IL-6 was induced in the lung upon infection at 3 dpi but did not vary significantly among groups (Fig. 4D). Although we did not see an increase in serum IL-6 between mock-infected and infected isotype control animals in these experiments, IL-6 was significantly higher in the serum of infected mice treated with anti-TGF-β than in the isotype group (Fig. 4E). IL-12 p40 was increased in the lung on 3 dpi (Fig. 4F). These data support a significant role for TGF-β in the regulation of both acute lung infiltration and systemic cytokine storm.

Figure 4.

TGF-β regulates IL-6 and modulates viral clearance. Aged C57BL/6 mice were infected with maSARS-CoV-2 and treated with isotype or cytokine neutralizing antibodies on alternate days and weighed daily. On day 3 postinfection, the lung homogenate and serum were tested for virus and cytokines. Cytokine bead assay shows cytokine levels of (A) TGF-β; (B) viral quantification; (C) IFN-γ in serum; IL-6 in (D) lung homogenate and (E) serum; and (F) IL-12p40 in lung homogenate. Results are representative of at least 2 experiments. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice. A P value <0.05 was considered statistically significant. *P = 0.01 to 0.05; **P = 0.01 to 0.001; ***P = 0.001 to 0.0001. PFU, plaque-forming units.

Neutralizing IL-4 prolongs pathology

The effect of immunoregulatory cytokines on weight loss, which is a sign of clinical disease, was assessed. While infected, isotype control–treated mice showed significant weight loss that had a small peak 3 dpi. Neutralization of IL-4 led to weight loss that was prolonged compared with isotype control mice, with significantly more weight loss at 4 dpi than isotype control mice (Fig. 5A). Virus quantification by plaque assay was similar in anti-IL-4–treated mice to the isotype control group by qPCR (Fig. 5B). IL-4 was found to be expressed even in the lungs of mock-infected mice when tested with a high-sensitivity bead-based assay, suggesting low homeostatic levels (Fig. 5C). IL-13 was also detected in the isotype control group on both 3 and 7 dpi and was not higher than the mock-infected group in these experiments (Fig. 5D).

Figure 5.

Neutralization of IL-4 prolongs weight loss without change in viremia. Aged C57BL/6 mice were infected with maSARS-CoV-2 and treated with isotype or cytokine-neutralizing antibodies on alternate days and weighed daily. (A) Percent weight loss over the course of infection, and statistics at 4 dpi. (B) Quantification of virus at 3 dpi by relative levels of ORF1ab by qPCR. Cytokine levels in lung homogenate for (C) IL-4 at 3 dpi and (D) IL-13 are shown in bar graphs. Results are representative of at least 2 experiments. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice per group. A P value <0.05 was considered statistically significant. *P = 0.01 to 0.05; ****P < 0.0001. d, day; ns, not significant.

Neutralizing IL-4 has an effect on lung inflammation. Anti-IL-4–treated mice showed increased infiltration in perivascular spaces and around airways 3 dpi, as shown in Fig. 6A. Alveolar thickening, related to alveolar infiltration, and consolidation were similar to isotype control mice. All of these observations are quantified as an overall pathology score in Fig. 6B. Anti-IL-4 but not anti-TGF-β showed these effects on periariway and perivascular infiltration as quantified in Fig. 6C. There was no significant difference in alveolar cellularity between isotype control– and anti-IL-4– or anti-TGF-β–treated groups 3 dpi (Fig. 6D). Collectively, these data support a role for IL-4 in regulating lung pathology in this infection. As IL-4 can drive alternatively activated macrophage differentiation in infection,³¹ we tested for genes correlated with this phenotype in RNA from infected lungs on day 3 postinfection: arginase 1 (Arg1), an amino acid metabolism (AAM) effector gene, was expressed in infected lungs, and significantly more in isotype than anti-IL-4-treated animals (Fig. 6E). However, expression of resistin-like molecule a1 (Fizz1), another AAM transcript, was not increased in infection 3 dpi (Fig. 6F).

Figure 6.

Neutralization of IL-4 leads to increased cellular infiltration on 3 dpi. Aged C57BL/6 mice were infected with maSARS-CoV-2 and treated with isotype or cytokine-neutralizing antibodies on alternate days and weighed daily. (A) H&E staining of lungs 3 dpi from mice treated with anti-IL-4. Results are representative of at least 2 experiments. Cellular infiltration, alveolar thickening, and consolidation were identified, and representative fields are shown. Scale bars represent 100 μm. (B) Quantification of overall pathology score from 3 dpi. (C) Quantification of perivascular and periairway infiltration using ratio of inflamed to total. (D) Alveolar infiltration represented by a score based on average number of nuclei in the thickness of the alveoli in each field. (E) Arg1 and (F) Fizz1 expression in the lungs by real-time PCR. Results are representative of at least 2 experiments. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice. A P value <0.05 was considered statistically significant. *P = 0.01 to 0.05; **P = 0.01 to 0.001; ***P = 0.001 to 0.0001. mRNA, messenger RNA.

As we had observed elevated inflammatory cytokines CCL5 and IL-12p40 7 dpi, after virus became undetectable, we evaluated lung histology at 7 dpi (Fig. 7A). All infected groups still had some signs of inflammation at this time point. Significant interstitial thickening with 60% to 75% of the lung affected was observed in most infected mice treated with the isotype control (n = 3 of 5). Peribronchial and perivascular infiltrates were present in all mice, with half of the mice having more than 3 foci of inflammation in the lobe section studied. Consolidation was present in all mice, and half of the mice exhibited consolidation in 3 or more areas. Half of the mice had detectable necrotic cells in areas of consolidation and 1 mouse had alveolar edema. However, treatment with anti-IL-4 and anti-TGF-β changed the intensity of infiltration as well as the spatial distribution of cells and damage. The histopathology and pathology score showed signs of recovery by 7 dpi in the lungs of the anti-IL-4–treated group; However, the pathology score showed that the infected, isotype-treated group still had more lung damage than mock-infected animals 7 dpi (Fig. 7B). Perivascular infiltration was specifically reduced in mice given neutralizing antibodies to the immunoregulatory cytokines TGF-β and IL-4, compared with isotype-treated groups, while periairway infiltration was specifically reduced by anti-TGF-β at this late time point, suggesting that they may inhibit resolution of pathology (Fig. 7C). Alveolar infiltration, the most likely feature to be associated with reduced lung function, remained observable at 7 dpi and was not significantly affected by neutralization of regulatory cytokines (Fig. 7D). These quantitative studies did not include lungs from the anti-IL-10–treated mice, which appeared grossly similar to the isotype control. Combining observations at the 2 time points, it appeared that neutralizing IL-4 shifts the peak of infiltration earlier, at least around vessels, suggesting early regulation by this cytokine, which is supported by prolonged weight loss at 4 dpi separating the events in the pathological cascade into 2 mechanistic phases, as in humans.

Figure 7.

Lung pathology of maSARS-CoV-2–infected mice 7 dpi. C57BL/6 mice were infected at 10 mo of age with maSARS-CoV-2 and treated with isotype control or cytokine-neutralizing antibodies on alternate days and weighed daily. (A) H&E staining of lungs of isotype (top), IL-4 (middle), and TGF-β (bottom). Results are representative of at least 2 experiments. Cellular infiltration, alveolar thickening, and consolidation were identified, and representative fields are shown. Scale bars represent 100 μm. (B) Quantification of overall pathology score 7 dpi. (C) Quantification of perivascular and periairway infiltration using the ratio of inflamed to total, and (D) alveolar infiltration represented by a score based on average number of nuclei in the thickness of the alveoli in each field on 7 dpi. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice per group. A P value <0.05 was considered statistically significant. *P = 0.01 to 0.05; **P = 0.01 to 0.001; ****P < 0.0001.

To test potential causes of persistent inflammation 7 dpi, we measured the monokines CCL5 (Fig. 8A), IL-12p40 (Fig. 8B), and IL-10 (Fig. 8C), as well as IFN-γ (Fig. 8D). Several appeared upregulated, but only IL-12 p40 reached significance in the lung of anti-IL-4–treated mice when compared with the isotype control–infected group at 7 dpi. Strikingly, in the anti-IL-4–treated group, IL-10 was reduced compared with the isotype at 7 dpi, reversing the IL-12/IL-10 ratio during the resolution phase, a potential driver of cellular infiltration. While IL-22 was not present in this mild disease, production of the tissue protective cytokine IL-22 was increased in the serum 7 dpi by neutralization of TGF-β but not IL-4 (Fig. 8E). These data suggest that while various overlapping regulatory cytokines inhibit cellular inflammation in the lungs during peak viremia, potentially both IL-4 and TGF-β also inhibit resolution of inflammation around the vessels, but less so around airways and alveoli in this model of mild COVID-19 disease. For IL-4, the persistent inflammation may be through regulation of the IL-12 to IL-10 ratio, though further study is required to determine causality.

Figure 8.

Late lung and serum cytokines. Aged C57BL/6 mice were infected with maSARS-CoV-2, treated with isotype or cytokine-neutralizing antibodies on alternate days, and weighed daily. On day 7 postinfection, the lung homogenate and serum were tested for cytokines by cytokine bead assay or ELISA (IL-33). Graphs show (A) CCL5, (B) IL-12p40, (C) IL-10, and (D) IFN-γ in lung homogenate, and (E) serum IL-22 of infected mice. Results are representative of at least 2 experiments. Statistical analysis was performed by 1-way analysis of variance followed by Student’s t test. Data were presented as mean ± SEM for groups of 5 mice. A P value at <0.05 was considered statistically significant. *P = 0.01 to 0.05; **P = 0.01 to 0.001.

Discussion

In summary, regulatory cytokines were tested for their ability to affect lung inflammation, weight loss, cytokines, and chemokines during infection of aged mice with maSARS-CoV-2. Several studies have been done with this virus and show many similarities with human disease.^28,44,45 Lungs from mice infected with maSARS-CoV-2 showed alveolar thickening and consolidation up to and including pneumonia, as previously published by others in human autopsies and mice.^9,46–48 COVID-19 autopsy studies suggest that deceased patients’ lungs fall into 3 pathological categories, which can overlap: (i) acute respiratory distress syndrome characterized by progressive diffuse alveolar damage, (ii) bronchopneumonia resulting from secondary infection with marked neutrophilia, and (iii) tissue thrombosis.^46,48 This infection recapitulated the first category displaying excess deposition of extracellular matrix and septal thickening, though we have not quantified fibrosis here. Compared with Chrabańska et al.,⁴⁹ hyperplastic type II pneumocytes, hyaline membranes, and thrombosis were present, but variable, similar to their findings.

Cytokines that correlate with disease severity during human COVID-19 are upregulated in mouse models mirroring quite well those seen in SARS-CoV-2 infection in humans.⁵⁰ We found that maSARS-CoV-2–infected mice had increased production of G-CSF, IL-12 p40, CCL2, and CCL5, which are often linked to monocyte recruitment, a prominent feature of this infection. CCL2 correlated with infiltration of leukocytes seen in lung sections at 3 dpi, while CCL5 and IL-12 p40 also remained elevated at 7 dpi, suggesting a mechanism for prolonged inflammation seen in humans and mice. Vanderheiden et al.²⁸ showed a role for inflammatory monocytes in the lungs of animals infected with this same virus. They showed that CCR2 deficiency led to prolonged viremia indicating a role of monocyte-derived cells in viral killing.²⁸ They observed large increases in CD86⁺ parenchymal, as distinct from CD86⁻ inflammatory monocyte population. cDC2 (XCR1⁻ CD172a⁺) were also increased in that study, as in human samples. Another study using infection of adeno-associated virus-hACE2 transduced mice also found an expansion of pulmonary infiltrating myeloid-derived inflammatory cells characterized by early Ly6C^hi monocytes and inflammatory monocyte–derived macrophages (CD64⁺ CD11c⁻ CD11b⁺ Ly6C⁺).⁴⁴ While CCL2 is used by monocytes to leave the bone marrow and enter the tissue from the blood, CCL5 likely attracts maturing monocytes to inflamed niches within the tissue.⁵¹ As IL-12 induces both the chemokines and their receptors, and could promote weight loss, it is a very good candidate for causing the prolonged inflammation seen in the infected mice with IL-4 neutralized, though we have not tested that prediction here.⁵² TGF-β regulates the strength or duration of the immune response, including macrophage activation,⁵³ and is essential for lung homeostasis; however, its role in COVID-19 has been difficult to ascertain due to low expression levels. TGF-β was reported to increase furin expression leading to enhanced susceptibility to SARS-CoV-2 and disease severity due to increased infection levels, which we did not see here.^54,55 Most research on TGF-β in lung pathology has focused on fibrosis; however, fibrosis appears to be limited during acute severe COVID-19, and not a likely cause of death from COVID-19. The most common cause of death is respiratory failure due to diffuse alveolar damage, with most patients in the exudative or organizing phase of damage, while later in disease, up to 40% of patients show fibrosing damage upon autopsy.⁵⁶ To study immune regulation by TGF-β during SARS-CoV-2 infection, we neutralized TGF-β in mice infected with maSARS-CoV-2. Neutralization of TGF-β in mice infected with maSARS-CoV-2 led to the accumulation of Ly6C^hi monocytes and decreased the number of CD11c⁺ alveolar macrophages. During infection in the lung, the loss of alveolar macrophages can be compensated by monocyte-derived macrophages potentially explaining the opposite trends.^28,45 Furthermore, we detected significantly more IL-6 in the serum of anti-TGF-β–treated mice than isotype-treated group, indicating that IL-6 production is regulated by TGF-β systemically. It would be important to determine the specific steps of monocyte recruitment affected by TGF-β and IL-6, as well as the sources of systemic IL-6 in future studies.

The strongest effects of neutralizing IL-4 was prolonged weight loss and lung cytokines, suggesting a role for IL-4 in the regulation of lung pathology, which was indeed observed. While IL-4–induced type 2 inflammation is rare in viral disease, it is not unprecedented that IL-4 plays a role in viral inflammation for example in RSV infection in which RSV G protein preferentially induces a Th2 response contributing to asthma exacerbations.⁵⁷ IL-4 is also known as counter-regulatory to type 1 inflammation,⁵⁸ and can inhibit the synthesis of the proinflammatory mediators TNF-α, IL-1, IL-6, IL-12, prostaglandin E2, and the chemokines IL-8 and MIP-1³⁵ as well as the differentiation of Th1 lymphocytes.³⁶ In our experiments, neutralization of IL-4 led to increased acute lung inflammation, as evidenced by cellular infiltration and consolidation in the airways of treated mice. Neutralization of IL-4 specifically increased proinflammatory cytokine IL-12p40, and reduced regulatory cytokine IL-10, a combination that could promote prolonged inflammation in the lung of maSARS-CoV-2–infected mice. Because IFN-γ was expressed but not significantly affected by neutralization of regulatory cytokines at either time point tested, perhaps the effect is more likely to be on monocytic cells than Th1 differentiation. While IL-4–deficient mice have been shown to have an impairment in controlling influenza and other infections due to a reduction in CD8 T cell involvement,⁵⁹ the virus level was not changed in our studies, supporting a primarily immunoregulatory role of IL-4 here. Supporting our conclusions, patients with asthma or helminth infection, type 2 inflammatory diseases of the lung, overall, are less likely to have severe disease and eosinophilia associated with less severe disease.⁶⁰ Eosinophilia, with or without lymphopenia,³¹ is also a common finding in patients admitted for COVID-19, and correlates with severity but not with mortality.

IL-13, a type 2 cytokine that shares the same IL-4 receptor alpha (IL-4Rα) chain with IL-4, has been the subject of several studies in mice and human SARS-CoV-2 infection. IL-13 can be expressed in COVID-19 patients, and levels correlate with the likelihood of requiring a ventilator once hospitalized.⁶¹ Our data showing that IL-13 is downregulated in this model confirm expression but do not relate to any role in severity. To study the protective nature of asthma, IL-13 was added to infected human airway epithelial cultures and caused reduced viral infection. IL-5, a driver of eosinophil differentiation, also correlates with a prolonged time to SARS-CoV-2 virus elimination.³⁴ An interesting analysis over time in severe versus mild patients in Rwanda suggests that patients who died had, on average, more IL-4 and IL-13, as well as more IFN-γ, from the earliest time points than severely affected patients who lived. Plasma IL-13 and IL-4 are more relevant to severity later in chronic infection, opposite to IL-9, IL-10, and TGF-β, and corresponding with potential counter-regulation of IFN-γ later.⁶² Interestingly, dupilumab, the commercially available anti-IL-4Rα, an antagonist biological drug that blocks both IL-13 and IL-4 activity, reduces disease both in mice and in patients.⁶³ In a small clinical trial, asthma patients taking dupilumab at the time of SARS-CoV-2 infection experienced lower mortality.⁶¹ Overall, the data suggest a complex interplay of type 2 cytokines in COVID-19 severity, though little is known about IL-4, as distinct from IL-13, or the kinetics of the 2 cytokines effects in the context of early viremia and longer inflammatory processes. A role for AAM differentiation in immune-pathology has been elegantly verified in SARS-CoV infection using STAT6- and STAT1-deficient animals.³⁰ While IL-4 and IL-13 have been detected as increased in some mice, in some experiments, particularly on day 2,¹⁰ this increase was not observed in this study. This may be related to the extensive use of anesthesia for repeated injections, which did not occur in the other study, but controls to confirm that are lacking. Some mild inflammation was observed in some animals in the mock infected group (not shown).

In this study, we found Arg1 expression in the lung, suggestive of the presence of alternatively activated monocytic cells. Arg1 can be associated with repair of the lung damage in response to damaging lung parasites like Nippostrongilus brasiliensis,⁶⁴ liver fibrosis, and allergic airway inflammation.^65,66 As seen here, in other disease states, induction of Arg1 is dependent upon IL-4Rα signaling, and the presence and activation of STAT6. Interestingly, quickly discharged COVID-19 patients have a strong STAT signaling response in peripheral blood mononuclear cells, including pSTAT6,⁶⁷ the STAT that is specific to IL-4 signaling. This is not surprising, as in SARS-CoV infection, Arg1+ AAM was shown to be important in regulating immune-mediated pathology and prevention of progression to fibrotic lung disease.⁶⁸ The effect of basal IL-4 on Arg1 expression and pathology suggests that additional signals derived from the virus infection, or the immune response to it, are likely required to combine with IL-4 for the effects seen. Similarly, basal levels of IFN-I were shown to be crucial during Listeria infection.⁶⁹ It will also be interesting to determine if AAM is responsible for moderating the recruitment or location of infiltrating cells that are acutely increased in their absence here. It will be important to separate the roles of immunoregulatory cytokines in the earliest phases, on the immune response to virus, and later on, in restoring homeostasis. Overall, our findings suggest multiple novel mechanisms of immunoregulation of the immunopathology underlying the disease process of COVID-19.

Author contributions

Conceptualization: R.S., J.J.E., S.M.D., S.T.S., L.P.-C.; Investigation: S.T.S., L.P.-C., M.G.-O., T.R.S., H.-C.T., H.G.F., M.B.H., Y.B.M.-M., K.G.; Methodology: S.T.S., L.P.-C., M.G.-O., T.R.S., H.-C.T., H.G.F., M.B.H., Y.B.M.-M., K.G.; Project Administration: R.S., R.R., J.J.E., V.D.M., M.A.E.; Resources: V.D.M., M.A.E., R.R., R.S., J.J.E., S.M.D.; Funding Acquisition: R.S., R.R., J.J.E., S.M.D., V.D.M.; Supervision: K.G., V.D.M., M.A.E., R.R., R.S., J.J.E., S.M.D.; Formal Analysis: S.T.S., L.P.-C., M.G.-O., T.R.S., R.R., R.S.; Visualization: L.P.-C., S.T.S., R.S., M.J.E.; Writing—Original Draft Preparation: S.T.S., R.S.; Writing—Review & Editing: S.T.S., R.S., M.J.E.

S.T.S. (Formal analysis [Equal], Investigation [Equal], Writing—original draft [Lead], Writing—review & editing [Supporting]), L.P.-C. (Conceptualization [Supporting], Formal analysis [Equal], Investigation [Lead], Methodology [Lead], Project administration [Supporting]), M.G.-O. (Investigation [Supporting], Methodology [Supporting]), M.J.E. (Formal analysis [Supporting], Writing—review & editing [Supporting]), T.R.S. (Formal analysis [Supporting], Investigation [Supporting]), H.-C.T. (Investigation [Supporting]), Y.B.M.-M. (Investigation [Supporting], Methodology [Supporting]), M.B.H. (Investigation [Supporting], Methodology [Supporting]), H.G.F. (Investigation [Supporting]), K.G. (Supervision [Supporting]), V.D.M. (Resources [Supporting]), M.C.S. (Methodology [Supporting], Supervision [Supporting]), M.A.E. (Conceptualization [Supporting], Project administration [Supporting], Resources [Supporting], Supervision [Supporting]), S.M.D. (Conceptualization [Supporting], Funding acquisition [Supporting], Project administration [Supporting], Supervision [Supporting]), J.J.E. (Conceptualization [Supporting], Formal analysis [Supporting], Funding acquisition [Supporting], Investigation [Supporting], Methodology [Supporting], Project administration [Supporting], Resources [Supporting], Supervision [Supporting], Writing—review & editing [Supporting]), R.R. (Conceptualization [Supporting], Methodology [Supporting], Project administration [Supporting], Resources [Supporting], Supervision [Supporting], Writing—review & editing [Supporting]), and R.S. (Conceptualization [Lead], Data curation [Equal], Formal analysis [Equal], Project administration [Lead], Resources [Lead], Supervision [Lead], Writing—review & editing [Lead]).

Funding

This work was supported by the University of Texas Medical Branch Institute for Human Infection and Immunity, the McLaughlin Fund, the Sealy Foundation at the University of Texas Medical Branch, the CII at Rutgers, and National Institutes of Health R01AI135061 (R.S.), R01AI166668 (R.R.), 1R01AI153602-01 (V.D.M.), and R21AI145400-A1 (V.D.M.).

Conflicts of interest

The authors have no financial conflicts of interest.

Data availability

Data supporting the findings of this study are available from the corresponding author upon request.