Hyaluronan in COVID-19 morbidity, a bedside-to-bench approach to understand mechanisms and long-term consequences of hyaluronan

Morphological differences induced by COVID-19 visualized by light sheet fluorescent microscopy

We have previously shown that the lungs of fatal COVID-19 patients are filled with HA (8). However, it is not clear which consequences HA has on the structural integrity of the lung and the alveolar volume. To address this, we set out to morphologically determine the three-dimensional organization within lung biopsies with light sheet fluorescence microscopy (LSFM). These biopsies originated from three deceased COVID-19 patients, four healthy donors who underwent lung resection, and one patient who underwent lung resection ten weeks post intensive care treatment due to severe COVID-19 infection, here referred as recovered.

The lung biopsies were processed, cleared and subjected to LSFM visualizing the autofluorescence (figure 1A). Major differences in the morphology of the lung were detected by maximum intensity projection (figure 1B-D, supplementary figure 1, and supplementary movie 1-3). The lungs of fatal COVID-19 patients were considerably denser compared to the healthy controls and recovered lung. Optical sections of the tissue revealed thin alveolar walls in the controls and recovered, whereas the distances between the alveoli in COVID-19 lungs were larger (figure 1E-G). The size distribution of the alveoli diameter was also smaller in the COVID-19 compared to control and recovered lung (figure H). The number of empty alveoli was fewer in the COVID-19 lungs as visualized in the surface rendering of the alveolar space (figure 1I-K). Next, the surface of the alveoli was quantified to better understand the capacity of the gas exchange in the samples. The isosurfaces of the alveoli were determined and showed a major loss of alveoli surface in the COVID-19 lungs compared to the healthy controls (figure 1L). Worth noting is that the lung morphology found in the recovered COVID-19 patient showed high similarity to the healthy controls.

• Download figure
• Open in new tab

Figure 1. Light sheet fluorescent microscopy (LSFM) of biopsies from deceased COVID-19 patients displays a reduced alveolar surface.

A) Schematic overview of LFSM procedure with tissue biopsies from healthy donors (n=4), deceased COVID-19 patients (n=3) and recovered COVID-19 patient (n=1). B-D) Maximum intensity projections of one representative biopsy from each patient group. Scale bar in D is 500 µm. E-G) LSFM sections and magnifications showing the alveolar structure in each patient group. Scale bar in G is 500 µm and scalebar in G’ is 100 µm. H) Size distribution of alveolar diameter measured in LSFM sections, ten random diameter measurements in five different z planes from each patient. I-K) Iso-surfaced empty space in the lung biopsies as a proxy for alveolar volume. Scalebar in K is 400 µm and scalebar in K’ is 100 µm. L) Quantification of the surface volume surrounding the empty space as a proxy for alveolar surface volume. Statistical significance was calculated by unpaired t-test (***p < 0.001).

To confirm the presence of HA in alveoli and to see if there is a difference in accumulation between the upper and lower lung lobes, histochemistry was performed on five COVID-19 necropsies (figure 2A-B). We did not observe a major difference between patients or lung regions, but extensive accumulation of HA was seen in COVID-19 necropsy compared to healthy controls (figure 2C). Collagen coils, a marker for irreversible lung damage, were also found in large areas of the necropsy where no remains of alveolar walls could be detected (figure 2C).

• Download figure
• Open in new tab

Figure 2. Lung necropsies and lung aspirates from severe COVID-19 patients show large amounts of hyaluronan (HA) and a high degree of HA fragmentation.

A) Schematic overview of collection of patient samples for HA analysis. B) Lung biopsies from the upper and lower left lobe of five deceased COVID-19 patients were stained for HA. Scale bar is 1000 µm. C) Collagen and HA staining in lung biopsies from healthy lung tissue (scale bar is 200 µm) and from a deceased COVID-19 patient (scale bar is 100 µm). The COVID-19 lung biopsy shows areas with no alveoli structures. D) HA was isolated from the lung biopsies in B (n=5) and a size distribution analysis was performed with gas-phase electrophoretic mobility molecular analysis (GEMMA). Shown is an average of all five patients. High molecular weight HA with a known molecular size of 30 kDa was included as control. HA was also isolated and subjected to size analysis by GEMMA from E) nasopharyngeal aspirate and F) endotracheal aspirate collected from three severe COVID-19 patients and displayed individually for each patient.

Severe COVID-19 induces fragmentation of HA in lung tissue and aspirates

As fragmented low molecular HA increases at sites of active inflammation and display a proinflammatory activity, HA fragmentation likely contributes to disease progression of COVID-19. HA was extracted from necropsies derived from the upper, and lower left lobe of five COVID-19 patients to investigate this. The isolated HA was subjected to size determination analysis by gas-phase electrophoretic mobility molecular analysis (GEMMA). Both the upper and lower lung lobes displayed highly fragmented HA (figure 2D). Although we did not observe an apparent difference in the abundance of HA with histochemistry between the upper and lower lung lobes (figure 2B), we found higher levels of both total HA and fragmented HA in the lower parts of the lung with GEMMA (figure 2D, area under the curve). Since fragmented HA induces a proinflammatory state and initiates migration of immune cells (16), we analysed the presence of neutrophils in lungs by staining for elastase. We detected major neutrophil infiltration in COVID-19 lungs compared to uninfected controls (supplementary figure 2). Next, nasopharyngeal aspirates (NPA) and endotracheal aspirates (ETA) collected from three COVID-19 patients were analysed to investigate the fragmentation process of HA during ongoing severe disease. All patients displayed fragmented HA, however, in contrast to the fragmentation in the necropsy samples, there were large individual variations between the total amount and degree of fragmentation (figure 2E-F). The finding that HA accumulated in the lungs and aspirates of severely diseased patients prompted us to measure the systemic levels of HA. We performed this in a cohort of mild and severely ill patients during acute disease and convalescence. Our objective was to investigate if plasma concentrations of HA could be used as a biomarker for disease severity.

Hyaluronan levels in plasma increase in severe COVID-19 and correlate with a reduction in lung function at follow up

Patients with COVID-19 infection classified as severe, based on requirement of high-flow nasal oxygen treatment and/or admission to the intensive care unit (ICU) corresponding to WHO-CPS 6-10 (17), were matched according to age to patients with mild COVID-19 (WHO-CPS 1-5). In total 103 individuals were selected; 37 patients were classified as severe and 66 patients as mild. The demography and clinical characteristics are presented in table 1. The groups differed significantly in BMI, obesity, minimum levels of hemoglobin, maximum levels of C-reactive protein, white blood cell count, and neutrophil count. No significant differences were observed when it came to comorbidities at baseline. Thirty-two (86.5%) patients in the severe group and four (6.1%) in the mild group received corticosteroid treatment (p < 0.001). Most of these patients, 28 in the severe and two in the mild group, received the first dose before the first sampling timepoint.

The concentration of HA in plasma samples was determined using ELISA in order to assess the association between HA and disease severity. A general increase in HA concentrations was observed in samples taken during the acute phase of the disease in both patients with mild and severe COVID-19, as compared to healthy controls (figure 3A). The HA concentrations were even further increased in patients with severe COVID-19. Interestingly, although HA concentrations declined in the convalescent phase (mild p ˂ 0.0001, severe p ˂ 0.0001), they remained elevated compared to healthy controls, especially in patients with severe disease. A sex comparison of HA plasma concentrations in the acute phase showed a similar increase in HA in severe COVID-19 in women and men (figure 3B). However, the HA levels remained higher among women compared to men (p=0.043, data not shown) in the convalescent phase of mild COVID-19. Analyses assessing the impact of age showed an increase in HA concentrations in severe compared to mild COVID-19 in both older (60-89 years) and younger (18-59 years) patients, with a general trend towards higher HA concentrations in the older age group (mild disease: p=0.01, severe disease: p=0.17) (figure 3C). As we have seen that HA accumulated in the lungs during COVID-19, and systemic levels of HA remains high even after 12 weeks, we next wanted to investigate how the systemic HA levels would correlate to actual lung function. HA levels of acutely infected patients were associated with the percentage of predicted diffusion capacity (DL_CO%pred) three to six months after infection (figure 3D-G). Lung function measurements were available in 70 out of 103 patients. The analysis was adjusted for sex, chronic lung disease, cardiovascular disease, hypertension, diabetes, smoking (current or previous), obesity (BMI ≥30), severity of COVID-19 (mild or severe), and age (20-59 and 60-89 years). Our results showed a negative correlation between HA levels in both the convalescent phase and acute phase with DL_CO%pred (supplementary table 1 and 2).

• Download figure
• Open in new tab

Figure 3. High concentrations of hyaluronan (HA) in plasma are associated with COVID-19 severity and long-term lung impairment.

A) HA concentrations in plasma from patients with severe and mild COVID-19 compared to controls (Ctrl). Samples were taken during the acute phase (0-4 weeks from disease onset) and again during the convalescent phase of the disease (≥12 weeks). HA concentrations were determined by ELISA. B-C) HA concentrations in plasma during the acute phase grouped based on severity and B) sex or C) age. Each dot represents one patient, and the line represents the median. Severity was based on WHO Clinical Progression Scale (WHO-CPS) with patients requiring high-flow nasal oxygen treatment and/or admission to the intensive care unit during the acute phase of illness classified as “severe”, corresponding to WHO-CPS 6-9, and all other patients as “mild”, WHO-CPS 1-5. Statistical significance was calculated by Mann-Whitney U-test (*p < 0.05, **p < 0.01, ***p < 0.001). D-E) The percentage of predicted values of diffusion capacity (DL_CO) related to HA concentrations in plasma in D) the acute phase or E) the convalescent phase divided by sex and disease severity. F-G) The percentage of predicted values of DL_CO related to HA concentrations in plasma in D) the acute phase or E) the convalescent phase divided by age and disease severity.

In summary, plasma concentrations of HA were substantially increased in COVID-19 patients and correlated to disease severity and reduced diffusion capacity. We then performed infection experiments in a human 3D-lung model to further study the potential mechanisms behind the increase of HA in COVID-19.

SARS-CoV-2 infection causes an inflammatory response, which is counteracted by corticosteroids in a human 3D-lung model

We used a lung model that closely resembles the human respiratory tract based on primary human bronchial epithelial cells (HBECs) isolated from human donors. This model was used to characterize the cellular pathways affected by SARS-CoV-2 infection and the effects of corticosteroid treatment with respect to inflammation and HA metabolism. The cells were differentiated at an air-liquid interface (ALI) to form a polarized epithelium containing an apical layer of fully functional secretory and ciliated cells and an underlying layer of basal cells (18). The lung cultures were treated with, or without, the corticosteroid betamethasone, and infected with SARS-CoV-2 (figure 4A). The course of infection was monitored daily by collection of apical secretions containing released progeny virus and the viral load was quantified by qPCR. A distinct increase in viral RNA was observed over time, indicating an active viral replication in the HBEC ALI-cultures (figure 4B). Betamethasone-treated cultures showed reduced levels of released progeny virus. The infected HBEC ALI-cultures were harvested at 96h post infection and subjected to total RNA sequencing, to identify the mechanisms and pathways regulated by SARS-CoV-2 infection and betamethasone treatment. Statistical comparison of genes expressed in uninfected vs. infected HBEC ALI-cultures (figure 4C, supplementary table 3) identified 167 differentially expressed genes (DEGs) that were upregulated upon SARS-CoV-2 infection, with a strong enrichment of genes involved in the immune response, especially type I interferon (IFN-I) signaling. A similar statistical comparison of infected HBEC ALI-cultures with or without betamethasone treatment was performed to evaluate the effect of betamethasone treatment. This identified 102 genes that were upregulated, and 229 genes that were downregulated in the betamethasone-treated cultures (figure 4D, supplementary table 3). Interestingly, 73 of the 167 genes upregulated by infection (figure 4C) were downregulated by betamethasone treatment (figure 4D). A heatmap displaying the mean expression values of the individual overlapping genes showed a clear upregulation of these genes by SARS-CoV-2 infection, which was counteracted by betamethasone treatment (figure 4E). Several of the affected genes were involved in inflammatory responses, correlating well with the anti-inflammatory activity of betamethasone. We also collected basolateral samples from the HBEC ALI-cultures at 96h post infection and analyzed these with a targeted cytokine panel based on proximity extension assay, allowing a quantitative concentration measurement of each cytokine (figure 4F). Of the ten cytokines that showed significant changes, eight were downregulated by betamethasone treatment, but not affected by the infection. CXCL10 and CXCL11 were both upregulated by the SARS-CoV-2 infection and subsequently downregulated by steroid treatment.

• Download figure
• Open in new tab

Figure 4. Inflammatory response and effect of corticosteroid treatment in a SARS-CoV-2 infected 3D-lung model.

A) Schematic overview of the primary 3D-lung model that was infected with SARS-CoV-2 in the presence or absence of corticosteroids to investigate the effect on inflammation and hyaluronan metabolism. B) The lung cultures based on differentiated primary human bronchial epithelial cells at an air-liquid interface were pretreated with/without betamethasone (Beta) in the basal media for 20 h prior to infection with SARS-CoV-2 (multiplicity of infection=0.5). The accumulated viral release from the apical side of the cultures was quantified by qPCR at indicated timepoints. C-D) Volcano plot showing differentially expressed genes between C) infected vs uninfected (mock) lung cultures and D) infected lung cultures with betamethasone vs without betamethasone treatment. The statistical p-value (-log₁₀) is plotted against the fold change in gene expression (log₂). Dotted lines highlight the significance cut off corresponding to a fold change of 1.3 and p-value=0.05. E) Heatmap displaying the mean expression (log₂) of the overlapping genes in uninfected, infected and infected + betamethasone-treated lung cultures. F) Cytokine levels in basolateral samples from HBEC ALI-cultures collected at 96 h post infection analyzed by Proximity Extension Assay (Olink). The gene expression levels of HA synthases G) HAS1, H) HAS2 and I) HAS3 along with the hyaluronidases J) HYAL1 and K) HYAL2 were determined by qPCR 120 h post SARS-CoV-2 infection of lung cultures from two different donors (D1 and D2). L) HA concentrations were determined by ELISA in apical secretions from lung cultures 120 h post infection. Mean values and SEM are shown, statistical significance was calculated by unpaired t-test (*p < 0.05, **p < 0.01***p < 0.001).

SARS-CoV-2 affects HA metabolism by upregulation of HA synthases and downregulation of hyaluronidases

To further investigate the direct effect of SARS-CoV-2 on HA synthesis and degradation, expression levels of the three human hyaluronan synthases, HAS1, HAS2 and HAS3 and the two major hyaluronidases, HYAL1 and HYAL2, were determined in HBEC ALI-cultures from two different donors five days post infection. No effect was seen on HAS1 expression, but SARS-CoV-2 infection increased HAS2 in donor 2 and HAS3 expression in both donors (figure 4G-I). Betamethasone treatment of infected cultures resulted in a reduction of the upregulated HAS2 expression in donor 2 and HAS3 in donor 1 (figure 4H and I). Additionally, SARS-CoV-2 infection decreased expression of hyaluronidase HYAL1 in both donors and HYAL2 in donor 2 (figure 4J and K). The hyaluronidase genes were not significantly affected by betamethasone treatment (figure 4J and K). Both upregulation of HA synthases and downregulation of degrading hyaluronidases can induce an increase in HA concentrations. Indeed, measurements of HA concentrations in the apical secretions from HBEC ALI-cultures confirmed the transcriptional changes with an increase in HA in cultures infected by SARS-CoV-2, which were kept at baseline concentrations when simultaneously treated with betamethasone (figure 4L).

Taken together, our results identified genes involved in viral defense and inflammation affected by SARS-CoV-2 infection. Betamethasone treatment demonstrated a general counteraction against the viral transcriptional effect. SARS-CoV-2 infection and betamethasone treatment also affected genes with potential impact on HA production and degradation. In addition to the effects on HA synthases and hyaluronidases, we observed an upregulation of transcription factors early growth response 1 and 2 (EGR1, 2) upon SARS-CoV-2 infection and a downregulation of lactate dehydrogenase A (LDHA) and TP53 induced glycolysis regulatory phosphatase (TIGAR) after betamethasone treatment, all of which are involved in the glycolysis and therefore affect HA metabolism (supplementary table 3). Based on our results, we present a model where the positive action of betamethasone in severe COVID-19 patients is a combined action of a reduced inflammatory response (figure 5A, supplementary table 3) and a reduction of the pathological overproduction of HA upon SARS-CoV-2 infection (figure 5B, supplementary table 3).

• Download figure
• Open in new tab

Figure 5. The effect of SARS-CoV-2 and betamethasone on cellular pathways regulating IFN-I responses and hyaluronan (HA) production in a human in vitro 3D-lung model.

Schematic presentation of a cell infected with SARS-CoV-2 with or without betamethasone treatment and affected cellular pathways regulating A) the IFN-I response and B) HA production. The left side of the gene shows the effect on gene expression upon SARS-CoV-2 infection, and the right side shows effect of betamethasone treatment in infected lung cultures. Blue color corresponds to an increase in gene expression and yellow indicates a decrease in gene expression. Viral genes are represented by dotted lines.

Tissue pre-processing for optical 3D-imaging

Lung biopsies from four healthy controls, three deceased COVID-19 patients and one patient that recovered from COVID-19 were washed in PBS, fixed in formalin overnight, dehydrated successively into methanol (MeOH) (25%, 50%, 75%, 100%, 15 min each step) and stored at - 20°C until use. Prior to clearing, the lung samples were treated in bleaching solution (H₂O₂: dimethyl sulfoxide: MeOH, 3:1:2) overnight at room temperature (RT). The bleaching solution was refreshed, and the samples were incubated for an additional 6 hours at RT. Following bleaching, the samples were rehydrated into PBS (25%, 50%, 75%, 100%, 15 min each step) and mounted in 1.5% low-melting point agarose (Lonza™ SeaPlaque™ Agarose, Lonza, USA). After solidifying overnight, the agarose pieces were cut into cuboids and dehydrated into MeOH (25%, 50%, 75%, 100%, 15 min each step). The samples were subsequently washed in MeOH for 2x24 hours on rotation in order to remove any residual water. Following MeOH washes, the samples were optically cleared in BABB (benzyl alcohol, benzyl benzoate, 1:2) changing solution every 12 hours until the biopsies were completely cleared.

Light sheet fluorescence microscopy and 3D image processing

High resolution 3D images of the lungs were acquired using a UltraMicroscope II (Miltenyi Biotec, Germany) fitted with an 1x Olympus objective (Olympus PLAPO 2XC) and a lens corrected dipping cap MVPLAPO 2x DC DBE objective attached to an Olympus MVX10 zoom body with a 3000 step chromatic correction motor. The lung regions of interest were captured at a magnification of 1.6x with a scan depth of 1000 µm, a dynamic focus range across the specimen capturing 10 images per section, and a step-size of 5 µm yielding a voxel size of 1.89 x 1.89 x 5 µm. The samples were scanned for autofluorescence with the filters: Ex 470/40, Em: 525/50 using an exposure time of 300ms. Optical sections were saved in *ome.tif format native to the ImspectorPro software (version 7.0.124.9 LaVision Biotex GmbH, Germany). The *ome.tif files were converted into 3D projection *ims files using the Imaris file converter (version 9.9.1, Bitplane, UK) 3D volumes were generated using the built in surfacing method of Imaris (version 9.8.0, Bitplane, UK). To quantify the “empty space volume”, roughly translating to the alveolar volume, the threshold was set to only include hypointense space using absolute intensity for surfacing, with a surface grain size set to 0.8 µm. Further, an exclusion filter removing objects consisting of less than 10 voxels was applied to remove general noise. The anatomy surface was generated and set to include all signal above the hypointense signal with a surface grain size of 3.78 µm and a 10 voxel exclusion filter. Surfaces generated from hypointense regions outside the tissue volume were manually excluded when possible. Volumes and area of the segmented surfaces were extracted from Imaris as Excel (Microsoft, Office 365, version 2301) *.XML file format for quantification.

Isolation and fragmentation of HA from lung biopsies and aspirates

Study design and study population

Data and clinical samples were obtained from the CoVUm study, a prospective, multicenter observational study of COVID-19 including patients from Umeå and Örebro, and coordinated from Umeå University, Sweden (www.clinicaltrials.gov identifier NCT 04368013). The study protocol and cohort have been described in detail earlier (32). Non-hospitalized patients aged ≥15 years and hospitalized patients aged ≥18 years with a positive PCR-test for SARS-CoV-2 were enrolled in the study. Written informed consent was obtained from all participants or their next of kin before the first sampling timepoint. At data export on May 4, 2022, a total of 543 participants, enrolled between 27 April 2020 and 28 May 2021, had been registered in the CoVUm database. Seven of the participants were excluded since they did not fulfill the inclusion criteria (false positive PCR-tests for SARS-CoV-2). Out of the remaining 536 participants, all patients classified as severely ill during the acute phase of the disease, and with available blood samples, were selected for this study. In addition, 66 patients with mild COVID-19 were selected from the remaining study cohort. Participants were classified as “severe” if they required high-flow nasal oxygen treatment (HFNO) and/or was admitted to the intensive care unit (ICU) during the acute phase of illness, corresponding to WHO Clinical Progression Scale (WHO-CPS) 6-10 (17). All other participants were classified as “mild”, corresponding to WHO-CPS 1-5. Blood samples were obtained at acute phase (0-4 weeks after onset) and convalescent phase (≥12 weeks). For study outline, see figure 2A. A healthy control group, consisting of plasma samples collected from anonymous blood donors was included for reference. The study was performed according to the Declaration of Helsinki and approved by the Swedish Ethical Review Authority.

Data collection and clinical samples

All clinical metadata including age, sex, Charlson Comorbidity Index (CCI) (33), tobacco use, medication at enrollment, body mass index, type of respiratory support and medical treatment were collected and managed using the REDCap electronic data capture tools hosted at Umeå University (34, 35). Clinical chemistry data of conventional inflammatory markers were extracted retrospectively from the patients’ electronic medical records. The highest and lowest values of each biomarker from each individual, during the first 180 days after symptom onset were extracted for further analysis. Plasma samples used for HA analysis were collected at each timepoint in 6 ml EDTA tubes (BD Diagnostics).

Measurement of hyaluronan concentration in plasma samples

Plasma hyaluronan (HA) concentrations were measured with a competitive HA-binding protein-based enzyme-linked immunosorbent assay (ELISA)-like concentration measurement kit (K-1200; Echelon Biosciences Inc., Salt Lake City, UT, USA), according to the manufacturer’s instructions. Samples were run in duplicate and a coefficient of variation (CV) < 10% was considered acceptable. Absorbance was measured on a ThermoMultiskan Ascent (Thermo Fisher Scientific, MA, USA) and plotted by polynomial regression against the concentration of the standard curve.

Lung function tests

Lung function tests were conducted 3-6 months after study enrolment for non-hospitalized patients, or discharge from hospital for patients hospitalized during the acute phase of illness, as previously described (15). Reference values for DL_CO were calculated using The Global Lung Function Initiative (GLI) Network guidelines.

Assessment of hyaluronan metabolic pathways upon SARS-CoV-2 infection in an in vitro 3D-lung model

Transcriptomics, total RNA sequencing

At 96 h post infection the mock treated (n=3), infected (n=3) and infected + betamethasone treated (n=3) HBEC ALI-cultures were washed three times on both sides with PBS. The HBEC ALI-cultures were then lysed and RNA extraction was done using the NucleoSpin RNA II kit (Macherey-Nagel) following the manufacturer’s instructions. RNA-seq libraries were prepared using the Smart-seq2 method, and sequenced on an Illumina NextSeq 500 (75PE, v2.5 High Output kit). STAR 2.7.1a was used to align the reads against a reference genome consisting of GRCh38 and Sars_cov_2.ASM985889v3. A gene expression table was produced using featureCounts (36). Differential expression analysis was performed using DESeq2 (37).

Cytokine quantification

At 96 h post infection, samples were collected from the basal chambers of each of the HBEC ALI-cultures. The samples were inactivated with Triton X-100 at a final concentration of 1% followed by incubation at room temperature for 3 h. The levels of 45 cytokines were quantified by Proximity Extension Assay (Olink Target 48 Cytokine panel at Affinity Proteomics Uppsala, SciLifeLab Sweden), which gives absolute (pg/mL) and relative (normalized protein expression, NPX) concentration measurements of 45 pre-selected cytokines.

Quantification of HA synthases and hyaluronidases by qPCR

Total RNA from cells were extracted using the Nucleo-Spin RNA II kit (Macherey-Nagel). 1000 ng of RNA was used as input for cDNA synthesis using High-capacity cDNA Reverse Transcription kit (Thermo Fisher). Cellular HA synthases (HAS1, 2 and 3) and hyaluronidases (HYAL1 and 2) were quantified using qPCRBIO SyGreen mix Hi-ROX (PCR Biosystems) and QuantiTect primer assay (Qiagen, HAS1;QT02588509, HAS2;QT00027510, HAS3;QT00014903, HYAL1;QT01673413, HYAL2;QT00013363) with actin as a housekeeping gene (QT01680476) and run on a StepOnePlus^TM Real-Time PCR System (Applied Biosystems).

Measurement of HA concentration in apical secretions from HBEC ALI-cultures

Apical secretions from HBEC ALI-cultures were collected at 120 h post infection and HA concentrations were measured with a Hyaluronic Acid AlphaScreen Assay (K-5800; Echelon Biosciences Inc., Salt Lake City, UT, USA), according to the manufacturer’s instructions. AlphaScreen beads from the Histidine (Nickel Chelate) Detection Kit (PerkinElmer, MA, USA) was used. Chemiluminescent emission was measured on a SpectraMax i3x (Molecular Devices, CA, USA) and plotted by polynomial regression against the concentration of the standard curve.

Statistical analysis

Statistical analysis was performed with Graphpad Prism 9 and Jamovi version 2.2.5 (The jamovi project (2021)). jamovi (Version 1.6). Retrieved from https://www.jamovi.org). Descriptive variables of the patient cohort were analyzed by Mann-Whitney U-test (continuous variables) and Fisher’s exact test or Chi2 test (dichotomous variables). Missing data were handled by complete case analysis. No correction for multiple testing was performed. Multiple linear regression was used to explore the association between DL_CO and HA levels in plasma during the acute illness and the convalescent phase.Independent variables were sex, chronic lung disease, cardiovascular disease, hypertension, diabetes, smoking (current or previous), obesity (BMI ≥30), severity of COVID-19 (mild or severe), and age (20-59 and 60-89 years). The log transformation was used to address skewed data in HA.

Data availability

The CoVUm data cannot be made publicly available, according to Swedish data protection laws and the terms of ethical approval that were stipulated by the Ethical Review Authority of Sweden. Access to data from the CoVUm database is organized according to a strict data access procedure, to comply with Swedish law. For all types of access, a research proposal must be submitted to the corresponding authors for evaluation. After evaluation, data access is contingent on vetting by the Ethical Review Authority of Sweden, according to the Act (2003:460) concerning the Ethical Review of Research Involving Humans.