Modifications of lipid pathways restrict SARS-CoV-2 propagation in human induced pluripotent stem cell-derived 3D airway organoids

Abstract

Background: Modifications of lipid metabolism were closely associated with the manifestations and prognosis of coronavirus disease of 2019 (COVID-19). Pre-existing metabolic conditions exacerbated the severity of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection while modulations of aberrant lipid metabolisms alleviated the manifestations. To elucidate the underlying mechanisms, an experimental platform that reproduces human respiratory physiology is required.

Methods: Here we generated induced pluripotent stem cell-derived airway organoids (iPSC-AOs) that resemble the human native airway. Single-cell sequencing (ScRNAseq) and microscopic examination verified the cellular heterogeneity and microstructures of iPSC-AOs, respectively. We subjected iPSC-AOs to SARS-CoV-2 infection and investigated the treatment effect of lipid modifiers statin drugs on viral pathogenesis, gene expression, and the intracellular trafficking of the SARS-CoV-2 entry receptor angiotensin-converting enzyme-2 (ACE-2).

Results: In SARS-CoV-2-infected iPSC-AOs, immunofluorescence staining detected the SARS-CoV-2 spike (S) and nucleocapsid (N) proteins and bioinformatics analysis further showed the aberrant enrichment of lipid-associated pathways. In addition, SARS-CoV-2 hijacked the host RNA replication machinery and generated the new isoforms of a high-density lipoprotein constituent apolipoprotein A1 (APOA1) and the virus-scavenging protein deleted in malignant brain tumors 1 (DMBT1). Manipulating lipid homeostasis using cholesterol-lowering drugs (e.g. Statins) relocated the viral entry receptor angiotensin-converting enzyme-2 (ACE-2) and decreased N protein expression, leading to the reduction of SARS-CoV-2 entry and replication. The same lipid modifications suppressed the entry of luciferase-expressing SARS-CoV-2 pseudoviruses containing the S proteins derived from different SARS-CoV-2 variants, i.e. wild-type, alpha, delta, and omicron.

Conclusions: Together, our data demonstrated that modifications of lipid pathways restrict SARS-CoV-2 propagation in the iPSC-AOs, which the inhibition is speculated through the translocation of ACE2 from the cell membrane to the cytosol. Considering the highly frequent mutation and generation of SARS-CoV-2 variants, targeting host metabolisms of cholesterol or other lipids may represent an alternative approach against SARS-CoV-2 infection.

Keywords: Airway organoid; Angiotensin-converting enzyme 2; Induced pluripotent stem cell; Severe acute respiratory syndrome coronavirus 2; Single cell RNA-sequencing.

Graphical abstract

Fig. 1

Ipsc-derived airway organoids (ao) exhibit ace2 expression and vulnerable to sars-cov-2 pseudovirus infection. (A) Schematic illustration of the differentiation process of iPSCs to AO. (B) Brightfield image of the iPSC-derived AO at different stages of differentiation spanning ≧ 30 days. Scale bar = 200 µm. (C) Immunofluorescence staining of AO showed marker expression of basal cell (KRT5), secretory cell (SCGB3A2 and CFTR), and ciliated cell (Act-α-TUB). Nuclei and cellular actin filaments were counterstained with DAPI (blue) and ACTIN (green), respectively. Scale bar = 20 µm. (D) Cluster map showing the assigned identity for each airway component (Multiciliated, Secretory, Basal, AT1, and AT2 cells). (E) t-SNE projection showing major airway marker CDK1, CEACAM6, KRT17, OBSL1, TTYH1, and ACE2 expressions in the AO. (F) Immunoblotting analysis showed high ACE2 protein expression in iPSC-derived AO. (G) iPSC-derived AO expressed GFP signals after 48 h infection with GFP-conjugated SARS-CoV-2 pseudovirus. Scale bar = 100 µm. (H) Quantitative real-time PCR analysis showed increased mRNA expression level of GFP; n = 3. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Human iPSC-derived airway organoids (iPSC-AO) are vulnerable to SARS-CoV-2 infection. (A) Schematic illustration of SARS-CoV-2 virus infection in AO. (B) Time-lapse images of SARS-CoV-2-infected AO. The white arrow indicated the same organoid in the mock group, and black arrow indicated the same organoids in the infection group. (C) Immunofluorescence staining of N protein and S protein in SARS-CoV-2-infected AO. Nuclei were counterstained with DAPI (blue). Scale bar = 20 µm. (D) Transmission electron microscope analysis detected successful SARS-CoV-2 virus infection in AO, indicated by arrow. Scale bar = 200 nm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Next-generation sequencing analysis of SARS-CoV-2 infection in airway organoid. (A) Flowchart of RNA sequencing data analysis. The major analysis components consist of data quality control and preprocessing, differential expression gene analysis, gene ontology and pathway analysis, and isoform detection and splice junction analysis. (B) The left-hand-side and right-hand-side volcano plots display the results in a comparison analysis between the mock sample and SARS-CoV-2 infected sample at 24 h and 96 h, respectively. Volcano plot representing the DEGs satisfying the criteria of log 2 (fold-change) value > 2 or < -2 and p < 0.05. Red and green symbols indicate significantly upregulated and downregulated genes, respectively. (C) Venn diagram shows that 279 genes were commonly identified at 24 h- and 96 h-exposure to viral infection. Purple circle indicates the number of DEGs identified at 24 h exposure to viral infection compared to the mock sample; Orange circle indicates the number of DEGs identified at 96 h exposure to viral infection compared to the mock sample. (D) The bubble plot showed the gene ontology of the 24 h and 96 h infection-affected genes. (E) The heatmap of the semantic similarity matrix from top-ranking random GO terms in 24 h infection-affected group. (F) Heatmap presents the significance of biological processes in the infected cells after 24 h SARS-CoV-2 infection, which are compared to that after 96 h SARS-CoV-2 infection. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Dysregulation of immune response and lipid biogenesis in airway organoids after SARS-CoV-2 infection. (A) The flowchart of gene ontology analysis. (B) Temporal DEGs were classified according to temporal dynamic patterns of gene expression at mock, 24 h, and 96 h samples. According to the expression from the RNA sequencing, we classified the pattern into five major patterns. In each pattern of gene expression, we examined the enrichment of the identified DEGs and list the three major activated biological processes. (C) ClueGO pathway analysis shows the overrepresented functional categories under SARS-CoV-2 infection (** p < 0.001, * p < 0.01). (D) The presentative GO terms were labeled in bold fonts. The bubble size reflects the frequency of the GO term in the underlying gene ontology annotation database. (E) The top 10 significant functions which were annotated by BioPlanet 2019 database. (F) The scattering plot of p-value vs. Odds Ratio of all DEGs after 24 h SARS-CoV-2 infection. (G) The transcript level of all associated genes of lipoprotein metabolism and interferon alpha/beta signaling from the RNA sequencing.

Fig. 5

Alternative splicing expression and junction analysis after SARS-CoV-2 infection. (A) the gene ontology of all upregulated genes with significantly upregulated isoform expression. (B) Venn’s diagram showed the isoform lower than 200 counts (20 genes, isoform counts 〈2 0 0) and higher than 200 counts (13 genes, isoform counts > 200). These isoforms intersected with the significantly upregulated genes (mRNA FC > 5 fold). (C) The transcript expression of identified isoform post SARS-CoV-2 infection. The information in the box showed the top ten genes with abundant isoform transcripts and their coordinates. (D) Differential spicing junction analysis identified new spicing isoforms. Differential expression spicing junction was found in five genes. DMBT1, RPS8, AC060780.1, and APOA1 were found in a comparison of mock vs. 24 h. and HSPA5 was found in a comparison of mock vs. 96 h. Asterisk indicates the condition that differential expression spicing junction was found compared to the uninfected mock samples. (E) Sashimi plots of gene APOA1. This plot focuses on the human genome region chr11:116837001–116837950. An alternative splicing junction was discovered at the 24 h infected sample compared to the uninfected mock samples. The transformed expression levels at the two replicates of the 24 h infected samples in the splicing junctions are . (F) Sashimi plots of gene DMBT1. This plot focuses on the human genome region chr10:122585271–122615306. Three alternative splicing junctions were discovered at the 24 h infected sample compared to the uninfected mock samples. The transformed expression levels at the two replicates of the 24 h infected samples in the three splicing junctions are (425,281), , , and (705,708), respectively.

Fig. 6

Statin alters ACE2 localization in SARS-CoV-2 infected airway organoids. (A) The scheme of Statin analogs treatment to the SARS-CoV-2 infected AO. (B) We used immunoblotting analysis and (C) immunofluorescence staining to assay the inhibition effect on the N protein production in SARS-CoV-2-infected AO after 10 µM Fluvastatin or Simvastatin treatment. We detected the ACE2, TMPRSS2, and N protein levels in the indicated viruses/Statin combinations. GAPDH protein was used as a loading control. Immunofluorescence staining of N protein in the AO. The AO cells were stained for N protein (green) or EpCAM (red) and counterstained with DAPI (blue). Scale bar = 10 μm. (D, E) We used immunofluorescence staining to trace the ACE2 internalization process after indicated concentration of Fluvastatin or Simvastatin treatment. The airway organoid cells were stained for ACE2 (green) and counterstained with DAPI (blue). Scale bar = 10 μm. (F) Geographical drawing of amino acid changes in the spike protein of Wuhan, Alpha, Delta, and Omicron SARS-CoV-2 variants. The differences are shown in reference to SARS-CoV-2 Wuhan-1 genome, labeled with triangle. NTD, N-terminal domain; RBD, receptor binding domain; RBM, receptor binding motif; SD1, subdomain 1; SD2, subdomain 2; N, nucleocapsid. (G) Luciferase assay for quantification of SARS-CoV-2 pseudovirus infection (one MOI) in the 10 µM Fluvastatin or Simvastatin treated HEK293-ACE2 cells. Until 48 h later, we harvested the cell lysate for detecting the luciferase activity. The luciferase activity of each sample was normalized to the untreated control as 100%. Significance was determined by two-way ANOVA (*, p < 0.001). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)