An ex vivo human precision-cut lung slice platform provides insight into SARS-CoV-2 pathogenesis and antiviral drug efficacy

ABSTRACT

Coronavirus disease 2019 (COVID-19) has claimed millions of lives since the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and lung disease appears the primary cause of death in COVID-19 patients. However, the underlying mechanisms of COVID-19 pathogenesis remain elusive, and there is no existing model where human disease can be faithfully recapitulated and conditions for the infection process can be experimentally controlled. Herein we report the establishment of an ex vivo human precision-cut lung slice (hPCLS) platform for studying SARS-CoV-2 pathogenicity and innate immune responses, and for evaluating the efficacy of antiviral drugs against SARS-CoV-2. We show that while SARS-CoV-2 continued to replicate during the course of infection of hPCLS, infectious virus production peaked within 2 days, and rapidly declined thereafter. Although most proinflammatory cytokines examined were induced by SARS-CoV-2 infection, the degree of induction and types of cytokines varied significantly among hPCLS from individual donors. Two cytokines in particular, IP-10 and IL-8, were highly and consistently induced, suggesting a role in the pathogenesis of COVID-19. Histopathological examination revealed focal cytopathic effects late in the infection. Transcriptomic and proteomic analyses identified molecular signatures and cellular pathways that are largely consistent with the progression of COVID-19 in patients. Furthermore, we show that homoharringtonine, a natural plant alkaloid derived from Cephalotoxus fortunei, not only inhibited virus replication but also production of pro-inflammatory cytokines, and thus ameliorated the histopathological changes caused by SARS-CoV-2 infection, demonstrating the usefulness of the hPCLS platform for evaluating antiviral drugs.

IMPORTANCE

Here, established an ex vivo human precision-cut lung slice platform for assessing severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, viral replication kinetics, innate immune response, disease progression, and antiviral drugs. Using this platform, we identified early induction of specific cytokines, especially IP-10 and IL-8, as potential predictors for severe coronavirus disease 2019 (COVID-19), and uncovered a hitherto unrecognized phenomenon that while infectious virus disappears at late times of infection, viral RNA persists and lung histopathology commences. This finding may have important clinical implications for both acute and post-acute sequelae of COVID-19. This platform recapitulates some of the characteristics of lung disease observed in severe COVID-19 patients and is therefore a useful platform for understanding mechanisms of SARS-CoV-2 pathogenesis and for evaluating the efficacy of antiviral drugs.

INTRODUCTION

The human respiratory tract is the primary target of infection by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that causes the pandemic coronavirus disease 2019 (COVID-19). Clinical outcomes of SARS-CoV-2 infection vary widely from asymptomatic infection to death, and its underlying mechanisms remain elusive. The innate immune system in the respiratory tract is the first line of defense against invading respiratory pathogens, and infection often results in immediate and rapid induction of inflammatory cytokines, which in turn recruit leukocytes to infected sites that cause local inflammation and immunopathology. The speed, amount, and type of the local cytokines induced upon initial contact of a virus with target cells often dictate the outcome of infection. Indeed, clinical data have shown that the severity of COVID-19 often correlates with an overwhelming production of multiple inflammatory cytokines and chemokines—the so-called cytokine storm, and widespread alveolar damage and pneumonia (1), with the resulting lung disease the primary cause of death in COVID-19 patients (2, 3). However, because many social and behavioral factors (such as social distance, personal hygiene, or mask usage) can greatly influence SARS-CoV-2 transmission and alter disease outcome, what specific factors contribute to the initiation and ultimately the progression of COVID-19 remains under intensive investigation. To better understand COVID-19 pathogenesis, it is critically important to have a model system where the earliest host–pathogen interactions during SARS-CoV-2 infection in humans can be recapitulated and the conditions for the infection process can be experimentally controlled.

Although animal models have been widely used for studying disease pathogenesis and preclinical testing for vaccines and therapeutics, there is always uncertainty as to what extent findings in animals, particularly small rodents, recapitulate host–pathogen interactions and pathogenesis in humans. Furthermore, the applicability of many animal models to human disease has been of recent concern, as countless findings have not been translated during human clinical trials (4–8). While several animal models for SARS-CoV-1 and SARS-CoV-2 have been established, including several nonhuman primates, mice, hamsters, and ferrets, each model exhibits only certain clinical manifestations and histopathological features and does not faithfully reflect the whole picture observed in humans (9–11). Thus, developing a better approach that can recapitulate SARS-CoV-2 infection in human lungs is critical for understanding its pathogenesis in humans and complementing the currently available infection models.

Advanced study of infection in human lung tissue has largely been limited to the use of human primary and established cell lines (12). Since the emergence of the COVID-19 pandemic, several reports have described the development or adaptation of cell- or organoid-based systems, such as human stem cell-based alveolospheres and lung organoids for SARS-CoV-2 infection and for antiviral drug screening (13–15). While these systems are permissive for SARS-CoV-2 infection, they lack the native lung environment, particularly the immune cells, that play important roles in COVID-19 pathogenesis. Analysis of primary human tissue has largely been limited to post-mortem analysis of samples from infected patients. We posit that engineering and fabrication of standardized platforms from viable human lungs obtained from deceased transplant donors offer a critical native context for studying critical early host–pathogen interactions for infectious diseases of the human lung. Human precision-cut lung slices (hPCLS) are slices of living human pulmonary tissue that can be maintained under standard cell culture conditions in a laboratory. The hPCLS platform maintains the 3D cellular structure present in native tissue, and therefore fills a critical gap in existing infection models. The hPCLS platform accurately reflects not only the actual lung niche, preserving ciliary beat frequency and mucous production, but also cellular viability of the entire repertoire of cells found in the lung, including alveolar epithelial cells, endothelial cells, dendritic cells, alveolar and interstitial macrophages, and type two innate lymphoid cell (16). It also elicits diverse cytokine and chemokine responses and airway hyperresponsiveness to infection (17, 18). The complexity of human lung tissue supports direct translation of results from animal to human and from in vitro to in vivo. Herein we report the establishment of hPCLS as a powerful platform for modeling early host–pathogen interactions during SARS-CoV-2 infection, and demonstrate its utility as a tool for studying pathogenicity and host innate immune responses, and for evaluating the efficacy of antiviral drugs against SARS-CoV-2.

RESULTS

Establishment of the hPCLS platform for SARS-CoV-2 infection

As human lungs are the native organ for SARS-CoV-2 infection, we sought to establish hPCLS as an ex vivo infection platform for studying the initial events that occur during SARS-CoV-2 infection. Transplant-quality lungs that were either declined for transplant or not placed for transplant for logistical reasons were obtained from anonymous donors through the Arkansas Organ Recovery Agency (Fig. 1A–a). Lungs were processed and sectioned into 600-µm-thick slices with 8.5 mm diameter. Slices were sorted into 48-well plates of airway (contain at least one major airway) and alveolar (consist of only alveoli and minor airways) hPCLS that can be maintained in the laboratory for up to 3 months (Fig. 1A–b and c). In our initial experiment, multiple hPCLS slices were infected with SARS-CoV-2 at three doses (2 × 10³, 2 × 10⁴, or 2 × 10⁵ TCID₅₀ per slice). We found that infectious virus was detected in the supernatant of hPCLS wells at 24 h post infection (p.i.) only when the highest dose of the virus (2 × 10⁵ TCID₅₀/slice) was used (Fig. 1B). To confirm if the detected virus at 24 h p.i. resulted from virus replication and not from residual virus inoculum, hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice. At 3 and 24 h p.i., culture supernatants were collected for determining virus titer. We found that SARS-CoV-2 indeed replicated in hPCLS as viral titers increased from 3 × 10² to 2 × 10⁴ TCID₅₀/mL from 3 to 24 h p.i. (Fig. 1C). Therefore, we used the virus dose of 2 × 10⁵ TCID₅₀ per slice for all subsequent experiments. We also determined viral RNAs in the hPCLS using qRT-PCR. Consistent with the virus titers, viral RNAs were increased from 108 fold to 363 fold from 3 to 24 h p.i. (Fig. 1D). Viral nucleocapsid (N) protein was also detected in hPCLS at 24 h p.i. by immunofluorescence staining but was not detected in mock-infected hPCLS (Fig. 1E). It is noted that the immunofluorescence staining was generally weak and the majority of the N protein appeared to colocalize with epithelial cells (Fig. 1E).

Fig 1.

Establishment of the hPCLS platform for SARS-CoV-2 infection. (A) Processing of hPCLS. Donated human lungs (a) were processed into hPCLS (b, c). (b) An image of hPCLS with one or more large airways (arrow) in one well of a 48-well plate. (c) Bright field image of hPCLS at 200 x magnification showing alveoli (arrow) and interstitial spaces (arrowhead). (B and C) Analysis of susceptibility of hPCLS to SARS-CoV-2 infection. (B) hPCLS (n = 3–6) were infected with SARS-CoV-2 at 2 × 10³, 2 × 10⁴, and 2 × 10⁵ TCID₅₀ per slice, and culture supernatants were collected at 24 h p.i. for determination of virus titer (in TCID₅₀/mL). (C) hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice and supernatants were collected at 3 and 24 h p.i. for determination of virus titer (TCID₅₀/mL). Data is the mean and standard deviation (SD) of 15 replicates (n = 15) from five donors. *, P < 0.05 (unpaired t test). Dashed line indicates detection limit. (D) Quantification of viral RNA in hPCLS. hPCLS (n = 6) were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice for 3 and 24 h p.i. Each RNA sample was pooled from three slices. Viral RNA was measured by qRT-PCR and expressed as mean and SD of a duplicate. (E) Detection of SARS-COV-2 N protein by immunofluorescence. hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice for 24 h or mock infected as a control. Slices were fixed with formalin and processed and embedded with paraffin. The slices were then stained with a monoclonal antibody against viral N protein and an anti-mouse IgG conjugated with FITC. Cell nuclei were stained with DAPI. (F–G) Susceptibility of hPCLS derived from individual lungs (n = 14). hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice for various periods of time as indicated for lung#248, #252, and #253 (F) or for 24 h (lung#1–8) or 48 h (lung#9–11 with asterisk) (G). The culture supernatants were harvested for determination of virus titer (TCID₅₀/mL). Data is the mean and SD of 3 replicates for each individual lung as indicated. ns, P > 0.05 (one way ANOVA).

To further assess the susceptibilities of hPCLS to SARS-CoV-2 infection, hPCLS from three donors were infected with SARS-CoV-2 as indicated for 3, 24, and 48 h. Results show that the difference in virus titers between the three donor lung tissues was relatively small, i.e., within one log₁₀ (Fig. 1F). In addition, virus titers were also similar in hPCLS from an additional 8 donors at 24 h p.i. or 3 donors at 48 h p.i. (Fig. 1G). It is noted that the ages of these donors range from 31 to 48 years. While the number of donors is small, these results suggest that lungs from adults between 30 and 50 years of age have similar susceptibility to SARS-CoV-2 infection. Collectively, these results demonstrate that the ex vivo hPCLS platform is permissive for SARS-CoV-2 infection.

Kinetics of SARS-CoV-2 replication in hPCLS

Lungs are the primary target of SARS-CoV-2 infection. Despite millions of deaths during the pandemic, little is known about the precise viral replication kinetics in the lungs of individual COVID-19 patients. This information is critical for understanding COVID-19 pathogenesis and for effectively managing COVID-19 patients. We took advantage of the hPCLS platform to determine the replication characteristics of SARS-CoV-2 in human lungs under controlled experimental conditions. Three to six hPCLS slices from each of the two donors were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice. At various time points p.i. as indicated, culture supernatants were collected for determining virus titers. As shown in Fig. 2A, virus titers rapidly increased from 3 to 24 h p.i., reached a plateau at 1.2 × 10⁴ TCID₅₀/mL at 36 h p.i., and rapidly decreased thereafter. By 96 h p.i. infectious viruses remained just above the detection limit of 100 TCID₅₀/mL (167 TCID₅₀/mL), suggesting transient reproduction kinetics for infectious viruses. However, viral RNAs isolated from infected lung slices increased continuously from 3 to 96 h p.i. (Fig. 2B). The inverse correlation between the kinetics of infectious virus reproduction and of viral RNA replication at late times of infection suggests that SARS-CoV-2 continues to replicate in the lungs during the period of 4 days p.i., but that infectious virus production is readily inhibited from 48 h p.i. onward. This result is in stark contrast to those obtained from Vero and human lung epithelial A549/ACE2 cell culture systems, in which virus replication reached and maintained high titers (≈10⁷ TCID₅₀/mL) from 24 to 72 h p.i., and decreased only slightly (≈one log₁₀) from 72 to 96 h p.i. (Fig. 2C). These results indicate that virus titers increased rapidly in both immortalized cell lines and hPCLS during the first 24 h of infection, followed by a plateau, and in the case of hPCLS decreased rapidly.

Fig 2.

SARS-CoV-2 replication kinetics in hPCLS. hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice. At indicated time p.i., supernatants were harvested for determining virus titer (A) and slices for viral RNA quantification (B). Virus titer is expressed in TCID₅₀/mL and viral RNA in log2 fold over the mock-infected control. Data is the mean and standard deviation (SD) of 3–6 replicates (n = 3–6). Dashed line indicates detection limit. (C) SARS-CoV-2 replication kinetics in cell cultures. Vero or A549/ACE2 cells were infected with SARS-CoV-2 at MOI of 1 and supernatants were harvested at various time points p.i. for determination of virus titer, which is expressed as the mean TCID₅₀/mL of a duplicate.

Induction of proinflammatory cytokines and chemokines in hPCLS by SARS-CoV-2 infection

The innate immune response in the respiratory tract is a double-edged sword: it is the first line of host defense against respiratory pathogens, but it can also trigger damaging inflammation. Initial induction of local cytokines and chemokines often dictate the outcome of an infection or disease. In an effort to identify the initial innate immune response to SARS-CoV-2 infection in the lungs that might drive progression of COVID-19, we assessed the induction of common proinflammatory cytokines and chemokines in hPCLS following SARS-CoV-2 infection. hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ and culture supernatants were collected at 3, 24, and 48 h p.i. for measuring proinflammatory cytokines (IL-8, IL-1β, IL-6, IL-10, TNF-α, and IL-12p70) and chemokines (CCL2/MCP-1, CCL5/RANTES, CXCL8/IL-8, CXCL9/MIG, and CXCL10/IP-10) by flow cytometry. While we found considerable variations in the expression of most cytokines/chemokines tested between donors and among cytokines/chemokines, IP-10, IL-8, MIG, and IL-1b were significantly induced across all donors, with IP-10 and IL-8 being at high levels (Fig. 3). In addition, IL-10 and IL-12p70 were not induced by SARS-CoV-2 infection; however, induction of IL-6 appeared nonspecific as it was induced in both virus-infected and mock-infected hPCLS (data not shown). We thus conclude that IP-10 and IL-8 are two potential inflammatory biomarkers for SARS-CoV-2 infection in the lungs as both were consistently induced in all donors at high levels. This result is consistent with the finding that the levels of IP-10 and IL-8 in COVID-19 patient sera correlate with the progression and severity of the disease (19–22).

Fig 3.

Induction of proinflammatory cytokines and chemokines in hPCLS by SARS-CoV-2 infection. hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice, and culture supernatants were harvested at 3, 24, and 48 h p.i. for quantifying the protein levels of secreted cytokines and chemokines using cytometric beads array (CBA) kits by flow cytometry. The amounts from infected samples were subtracted by the amounts from mock-infected samples and were expressed as pg/mL. Data is the mean and SD of 3 replicates (n = 3) for each cytokine and chemokine and is indicated for three individual lung donors (#248, #252, #253). Significance was calculated with Tukey’s multiple comparisons test in the GraphPad Prism program. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Induction of cytopathic effects in hPCLS by SARS-CoV-2 infection

Histopathological changes in the lungs can result from direct virus infection (i.e., cytolytic infection) or a bystander effect (i.e., through localized innate immune responses). Induction of proinflammatory cytokines often leads to inflammation that in turn results in histopathological changes in the lungs, which is a hallmark of severe COVID-19 in patients (23). While hPCLS do not allow for analysis of systemic immune responses that include innate immune infiltration and involvement of a circulatory system, they do allow for monitoring of initial inflammatory responses that can cause localized effects and initiate/drive downstream immune responses. To evaluate the utility of the hPCLS platform for studying COVID-19 pathogenesis, we determined histopathological changes of hPCLS following SARS-CoV-2 infection. hPCLS from six different donors were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀. On days 1, 3, and 5 p.i., hPCLS were fixed with formalin and processed for staining with hematoxylin and eosin (H&E). Mock-infected hPCLS were used as negative controls. As shown in Fig. 4, localized (focal) cytopathic effects as indicated by cellular debris and accumulation of cells in alveolar spaces weren’t observed in SARS-CoV-2-infected hPCLS until day 5 p.i. (marked areas). These results indicate that hPCLS are useful for evaluating the impact of SARS-CoV-2 infection on the initial cells and tissues it encounters in the lung during infection.

Fig 4.

Histopathological development in SARS-CoV-2-infected hPCLS. hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice for 1, 3, and 5 days or mock-infected as a control, and the slices were fixed and processed for H&E staining. Areas with cytopathic effects at day 5 p.i. are marked with dashed lines in 2 hPCLS samples (#45 and #65). Data are representative of the experiments from six individual lungs (n = 6). Scale bar, 50 µm.

Gene signatures identified by transcriptomic profiling shed light on COVID-19 progression in the lungs

While gene expression profiles from the lungs of post-mortem COVID-19 patients have been reported (24), these data represent only a snap shot of the disease at its latest stages, and do not reflect the early events or resulting disease progression in individual patients during infection. To assess the early host responses in the lung that drive COVID-19 progression, we evaluated the transcriptome profiles in hPCLS following infection from 24 to 96 h with SARS-CoV-2 by RNA sequencing (RNAseq), and compared them with those from uninfected hPCLS. The goal of this analysis was to identify those initial host signaling events that may be possible for driving the inflammatory responses and pathology that are seen in vivo during disease. We performed differential and pathways analyses and found that at 24 h p.i. 87 genes were upregulated while 433 genes were downregulated by SARS-CoV-2 infection with a P-value < 0.05 and a log2 fold change >1 (Fig. 5A). The number of genes that were up-regulated at 48 h, 72 h, and 96 h p.i. were 358, 744, and 822, respectively. The number of genes that were down-regulated at 48 h, 72 h, and 96 h p.i. were 237, 995, and 2233, respectively (Fig. S1; Data set S1 to S4). Of note, three of the top 10 down-regulated genes were the TNF-superfamily 11 (TNFSF11), carbonic anhydrase 4 (CA4) and oxidative stress induced growth inhibitor 1 (OSGIN1), which are involved in T cell-dependent immune response, CO₂/O₂ exchange, and NRF2-dependent antioxidant gene expression and cell death, respectively. Down-regulation of these genes may repress T cell immunity, decrease lung function, and exacerbate oxidative stress and cell death. On the other hand, up-regulation of chemokine CCL4/MIP-1β (macrophage inflammatory protein 1β), PPFIA4 (protein tyrosine phosphatase, receptor type, F polypeptide, interacting protein, alpha 4), and SFTPA2 (surfactant protein A2) that are involved in regulation of inflammation, cell adhesion, and interstitial lung disease/pulmonary fibrosis, respectively, potentially promotes inflammation and lung disease. In addition, five KEGG pathways were particularly enriched, including viral protein interaction with cytokine and cytokine receptor, rheumatoid arthritis, Rap1 signaling (CTNND1, FGF7, ITGB1, RASSF5, TLN1, VAV3 and VEGFA), leukocyte transendothelial migration (ARHGAP35, CTNND1, ICAM1, ITGB1, RASSF5, and VAV3) and chemokine signaling pathway (CCL3, CCL3L1, CCL3L3, CCL4, CXCL5, and VAV3) (Fig. 5B), indicating commencement of an inflammatory phase. By 48 h p.i., while more genes involved in cell adhesion and migration were continuously and significantly up-regulated (e.g., ITGA7, LAMB4, PRKCG), enriched genes in two additional pathways (i.e., HIF-1 and cellular senescence) began to emerge (Fig. 5C), suggesting that hypoxia and senescence have initiated following SARS-CoV-2 infection in the lungs at this stage. However, many of the genes in IL-17 signaling pathways and viral protein interaction with cytokine and cytokine receptor pathway (CCL17 CCL2 CCL7, CXCL10, CXCL13, CXCL2, CXCL7, DEFB4A, and DEFB4B) were downregulated (Fig. 5D). In particular, CCL17 was downregulated more than seven log2 folds. These results indicate a transient nature of transcriptional regulation of these chemokine genes in virus-infected lungs. At 72 h p.i., a large cluster of 35 genes related to ribosome and three other genes (CSF2, IL-12A, and JUN) were upregulated, all of which have been previously implicated in COVID-19 (Fig. 5E). Additionally, a cluster of 25 enriched genes in the protein processing in endoplasmic reticulum pathways were significantly upregulated, including many of the heat shock proteins (e.g., HSP90AB1, HSP90B1, HSPA1A, HSPA1B, HSPH1). Twelve genes enriched in the p53 signaling pathway were also upregulated (Fig. 5E). By 96 h p.i., enriched genes in three pathways were upregulated, including 13 genes in mitophagy, 10 genes in ferroptosis and 14 genes in complement and coagulation cascades (Fig. 5F). Of particular note are MAP1LC3C (>6 log2 fold increase) and SERPINA5 (>6 log2 fold increase) that are involved in cell death and blood coagulation, respectively. These results indicate that SARS-CoV-2 infection leads to profound changes of the transcriptional landscape in the lungs during the course of infection, i.e., from alteration of expression of genes associated with inflammation and regulation of lung physiology at the early stage of infection (24 h p.i.), to the genes that promote cell death and are associated with lung disease at the late stage of infection (96 h p.i.).

Fig 5.

Gene signatures in COVID-19 progression in the lungs identified by transcriptome profiling. Groups of 6 hPCLS slices each were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice or mock-infected as a control. At various time points (24–96 h) p.i., RNAs were isolated from the slices and subjected to RNAseq analysis. Differential gene expression was identified by comparing virus-infected group to mock-infected group. (A) A volcano plot showing the down- and up-regulated genes in the lungs by SARS-CoV-2 at 24 h p.i. with three representative signature genes each. (B–F) Heat maps showing signature genes in the major pathways during the disease progression in the lungs at 24 (B), 48 (C and D), 72 (E), and 96 (F) h p.i. (G) Bar plot showing the differential expression of IFN-I- and IFN-II-related genes at 24, 48, 72, and 96 h p.i. Positive and negative log2FC values indicate upregulation and downregulation, respectively.

Notably, seven type I interferon (IFN-I) related genes (IFNAR1, IFNAR2, IFI27, IFI35, IFI44L, ISG15, and ISG20) were significantly down-regulated (up to −8 log2 fold reduction) at various time points p.i., and none of the other IFN-I related genes were up-regulated at any of the four-time points p.i. (Fig. 5G; Supplementary Data set S1 to S4). In contrast, two type II interferon (IFN-II)-related genes (IRF6, IFI16) were moderately upregulated (up to 1.4 log2 fold increase). In addition, IRF2BP2, which is an interferon regulatory factor 2 (IRF2)-binding protein and acts as a co-repressor for IRF2 to repress IFN-I gene transcription (25), was also moderately up-regulated (one log2 fold increase) at 72 h p.i. (Fig. 5G). As IFN-I tends toward more antiviral and IFN-II more inflammatory, the down-regulation of IFN-I and upregulation of IFN-II may play a role in the development of severe COVID-19 in the lungs.

Proteomic analysis provides insight into activation of host molecular networks during SARS-CoV-2 infection in human lungs

To further understand the molecular basis of SARS-CoV-2 pathogenesis in human lungs, a group of 5 hPCLS samples each were infected with SARS-CoV-2 or mock-infected as control. At 48 h p.i., slices were lysed and proteins were extracted for proteomic analysis. Ingenuity Pathway Analysis (IPA) of the proteomic data predicted a number of upstream molecules, regulators, pathways, and diseases that might be associated with SARS-CoV-2 infection (Fig. 6A). Specifically, several growth factors, cytokines and chemokines in the networks appeared to be activated by SARS-CoV-2 infection (e.g., IGF1, TGFB1, GDF2, CCR2, IFNA2, IFNL1), which leads to branching of vasculature, infiltration by T lymphocytes, and cell movement of leukocytes, all indicative of the onset of an inflammatory phase of the disease, which is consistent with the findings from RNAseq (Fig. 5). Activation of these growth factors and many other transcriptional regulators (e.g., MRTFA, MRTFB, FOXM1, NPM1 and TEAD4) can also lead to formation of intercellular junctions, sprouting, and inhibition of organismal death. These functional changes are predicted to trigger several signaling pathways, e.g., (hepatic) fibrosis, wound healing, and dilated cardiomyopathy. Notably, the canonical GP6 signaling pathway is predicted to be activated by SARS-CoV-2 infection (Fig. 6A). GP6 is a member of the immunoglobulin superfamily and serves as the major signaling receptor for collagens and laminins, which leads to the platelet activation and thrombus formation. Indeed, numerous collagens (COL3A1, COL4A1, COL6A1, COL6A2, COL6A3, COL14A1, COL15A1, and COL18A1) and laminins (LAMA4, LAMA5, LAMB1, LAMB2, and LAMC1) were significantly activated in SARS-CoV-2-infected lungs (a partial list shown in Fig. 6B). The IPA also predicts potential links to several diseases, such as detachment of retina, congenital encephalopathy, and congenital malformation of brain (Fig. 6A), and a cluster of differentially expressed proteins also have links to systemic lupus erythematosus, human papillomavirus infection, and alcoholism (Fig. 6B). These results suggest that the cellular proteomic networks in the lungs that are altered by SARS-CoV-2 during the first 48 h of infection not only promote pulmonary inflammation, but may also contribute to other aspects of COVID-19, such as fibrosis, heart failure, thrombosis, and autoimmune disease.

Fig 6.

Alterations of molecular networks in human lungs during SARS-CoV-2 infection as determined by proteomic analysis. Groups of 6 hPCLS slices each were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice or mock-infected as a control. At 48 h p.i., proteins were isolated from the slices and subjected to proteomic analysis. The proteomes in virus-infected hPCLS were compared to those in mock-infected hPCLS. (A) Summary of upstream molecules, regulators, pathways and diseases that are associated with SARS-CoV-2 infection as predicted by Ingenuity Pathway Analysis (IPA). Node symbol: canonical pathway; function; cytokine; growth factor; transcription regulator; transmembrane receptor; complex; G-protein coupled receptor; enzyme; kinase; disease. Color code: orange, upregulation; blue, downregulation. (B) Major pathways and associated proteins in the lungs that are up-regulated by SARS-CoV-2 infection.

The hPCLS platform as a “clinical trial at the bench” for evaluating antiviral drugs against SARS-CoV-2

Cell cultures and small animal models have been the gold standard for pre-clinical testing of antivirals in vitro and in vivo, respectively. However, information gained from cell cultures is limited to antiviral effect and cytotoxicity of a drug. While animals are essential for in vivo testing, countless findings from animal testing have not translated during human clinical trials (4–8). These concerns prompt the scientific community to seek alternative systems. We posit that hPCLS as a native tissue may offer such an alternative platform, especially, for testing antivirals against SARS-CoV-2 that is associated with lung diseases. Moreover, host responses to infection and their impact on drug efficacy can be evaluated simultaneously, and therefore hPCLS are ideal for evaluating how antivirals may act in the lung.

To extend the utility of the hPCLS platform, we carried out antiviral drug testing in hPCLS as a “clinical trial at the bench” (Fig. 7A). We used homoharringtonine (HHT) as the first example. We previously identified HHT as one of the most potent anti-coronavirus small molecule compounds during library screening (26), and recent studies have confirmed its anti-SARS-CoV-2 activity in cell cultures (27) (Fig. S2). hPCLS from three donors were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice. At 1 h p.i., hPCLS were treated with HHT at 1 µM for 23 or 47 h (i.e., 24 or 48 h p.i.). Results show that virus titers were decreased approximately one log₁₀ or less at both time points (Fig. 7B). To determine whether the inhibition is specific, we performed a dose-response experiments. At 1 h p.i., hPCLS were treated with HHT at 1, 5, and 10 µM for 47 h, and culture supernatants were then collected for determining virus titers (i.e., at 48 h p.i.). Results show that treatment of hPCLS with HHT decreased virus titer more than one log₁₀ at 1 µM, and virus replication was completely blocked when HHT concentration increased to five or 10 µM (Fig. 7C). To further determine if the reduced virus titer resulted from the specific antiviral effect or was due to the cytoxicity of HHT, hPCLS were treated with HHT at these same concentrations for 47 h, and the culture supernatants were harvested for determining the lactate dehydrogenase (LDH) activity, an indicator for cell lysis. hPCLS treated with 0.1% DMSO or untreated were used as a vehicle control or as a negative control for background LDH level in the culture medium, respectively, while hPCLS treated with 1% triton X-100 was used as a positive control for cell lysis. As shown in Fig. 7D, the LDH activity remained at a background level when treated with HHT at one or 5 µM, or with DMSO; the LDH activity increased only slightly but statistically insignificantly when HHT concentration increased to 10 µM. In contrast, the LDH activity in triton X-100-treated hPCLS was significantly higher. These results confirm that HHT has an antiviral activity against SARS-CoV-2 in hPCLS and has no cytotoxicity at these concentrations.

Fig 7.

Establishment of the ex vivo hPCLS platform for evaluating antiviral drugs. (A) Schematic of the experimental design. HHT, homoharringtonine. Note that the image of hPCLS is being reused for illustration purposes only. (B) Inhibition of SARS-CoV-2 replication in hPCLS by HHT. hPCLS from three individual lung donors as indicated were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice and treated with HHT at 1 h p.i. at 1 µM for 24 h (24+) or 48 h (48+) or untreated for 24 h (24-) or 48 h (48-). Virus titers in the supernatants were determined and expressed as TCID₅₀/mL. Data is the mean and standard deviation (SD) of 3 replicates. (C) Determination of dose response. Six hPCLS each were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice or mock-infected, and treated with HHT at 1 h p.i. at various concentrations as indicated or treated with 0.1% DMSO as a vehicle (Veh) control. Virus titers in the supernatants were determined and expressed as TCID₅₀/mL. Data is the mean and SD of 3 replicates (n = 3). Dashed line indicates detection limit. (D) hPCLS were treated as in (C) for 48 h or untreated (-) as a negative control. HHT cytotoxicity was then determined by measuring LDH activity in the culture supernatants. Supernatants from hPCLS treated with 1% triton X-100 (+) were used as a positive control for cell lysis. Data is the mean and SD of 3 replicates (n = 3). ns, not significant. (E) Inhibition of proinflammatory cytokines and chemokines by HHT. hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice and treated with HHT (+) at 1 µM or untreated (-). Supernatants were harvested at 24 h or 48 h p.i., and the amounts of cytokines were determined with cytometric beads array assay. The amounts from infected samples were subtracted by the amounts from mock-infected samples and were expressed as pg/mL. Data is the mean and SD of 3 replicates for each cytokine and chemokine and is indicated for three individual lung donors (#248, #252, #253). Significance was calculated with Tukey’s multiple comparisons test in the GraphPad Prism program. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (F) Inhibition of SARS-CoV-2-caused histopathological abnormality in hPCLS by HHT. hPCLS were mock-infected or infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice and treated with HHT at 1 h p.i. at 10 µM. On day 5 p.i., slices were fixed with formalin and processed for H&E staining. Areas with cytopathic effects are marked with dashed lines. Data are representative of the experiments from three individual lungs. Scale bar, 50 µm.

We then determined the impact of HHT treatment on the induction of proinflammatory cytokines and chemokines mediated by SARS-CoV-2 infection. Results show that induction of 3 chemokines (IP-10, MIG, MCP-1) was completely blocked following the treatment of hPCLS with HHT at 1 µM (Fig. 7E). Interestingly, TNF-α was not affected in donor #248, increased in donor #252, but decreased in donor #253 by treatment with HHT, indicating a significant variability among different donor lungs. In contrast, IL-1β was induced for all donors at both time points by HHT treatment (Fig. 7E). The reason for the enhancing effect on IL-1β is unknown. The effect of HHT on the expression of IL-8 and RANTES was inconclusive (data not shown). Taken together, these results demonstrate that HHT has a general anti-inflammatory effect but that its effect on specific cytokines can vary between the individual donors.

We further evaluated the impact of HHT treatment on histopathological changes of the lungs. Five hPCLS were infected with SARS-CoV-2 and treated with HHT at 10 µM for 5 days. hPCLS infected with SARS-CoV-2 only were used as positive controls while mock-infected and HHT-treated hPCLS were used as negative controls. On day 5 p.i., hPCLS were fixed with formalin and processed for staining with H&E. As shown in Fig. 7F, focal cytopathic effects and cell accumulation in the alveolar space were observed in SARS-CoV-2-infected hPCLS (marked areas). In contrast, no cytopathic effect or cell accumulation was observed in SARS-CoV-2-infected and HHT-treated hPCLS, as in mock-infected and HHT-treated hPCLS (Fig. 7F). These results indicate that treatment with HHT reversed the lung histopathology caused by SARS-CoV-2 infection, and that the induction of IL-1β by HHT likely did not contribute to the histopathological abnormality. Thus, hPCLS provides a standardized platform for evaluating not only the antiviral efficacy but also how the antivirals act in the lungs and modulate host immune response and pathology.

DISCUSSION

In the present study, we establish an ex vivo hPCLS platform for evaluating the pathogenesis of SARS-CoV-2 infection in human lungs. Although SARS-CoV-2, especially the omicron variant, is highly transmissible, a large proportion of SARS-CoV-2 infections result in mild or no clinical symptoms. The most severe and fatal cases of COVID-19 are primarily caused by lung diseases; yet how SARS-CoV-2 replicates in the lungs is not well understood. Answers to this question are critical for understanding COVID-19 pathogenesis. Although airway epithelial cells are the initial targets of SARS-CoV-2 infection, many cell lines derived from airway epithelium, e.g., A549 (type II alveolar cells) and BEAS-2 (bronchial epithelial cells), are not or are minimally susceptible to SARS-CoV-2 infection. It has been shown that cell susceptibility to SARS-CoV-2 infection is directly correlated with the expression level of ACE2 (angiotensin-converting enzyme 2) receptor, as exogenous expression of ACE2 in the A549 cell line drastically increases its susceptibility to SARS-CoV-2 infection (Fig. 2C). Several studies reported controversial results about the expression levels of ACE2 in the normal human airway epithelium (18, 28, 29). Immunohistochemical staining and single-cell RNAseq analysis show that ACE2 is mainly expressed on type I and type II alveolar cells, weakly on bronchial epithelial cells, and minimally in the upper respiratory tract, such as oral and nasal mucosa and nasopharynx (28, 29). However, these results are inconsistent with the frequent detection of abundant SARS-CoV-2 viral RNAs and proteins in nasopharyngeal swabs during clinical diagnosis. In contrast, recent RNA in-situ hybridization experiments with tissues isolated from different regions of the respiratory tract reveal an opposing profile, i.e., ACE2 expression levels appear in a gradient, decreasing fashion from upper to lower respiratory tract, which also correlates with SARS-CoV-2 infectivity in the primary cell types isolated from these specific regions (18). This finding supports the clinical and epidemiological data that SARS-CoV-2 replicates efficiently in the nasopharynx and that a large proportion of the infected subjects do not suffer from pneumonia. It also suggests that SARS-CoV-2 replication in the lungs is likely different from that in the upper respiratory tract. Thus, most cell culture models using cell lines derived from airway epithelium or primary airway epithelial cells may not effectively recapitulate the characteristics of viral replication in the lungs (30). We show that SARS-CoV-2 infectivity in hPCLS as measured by virus titer (Fig. 1) differs from that seen in isolated primary cells, i.e., lower than in nasal or large airway epithelial cells but higher than in types I and II alveolar cells (18), indicating that the hPCLS platform represents mixed populations of different cell types in the native lungs.

Cell heterogeneity is a salient feature of the hPCLS platform. The hPCLS platform reflects not only the actual lung niche, preserving ciliary beat frequency and host cell responsiveness, but also cellular heterogeneity, including alveolar epithelial cells, endothelial cells, smooth muscle cells, and leukocytes, among other native populations. As leukocytes play critical roles in defense against respiratory pathogens as well as in mediating lung inflammation, the presence of leukocytes in hPCLS provides an excellent ex vivo infection platform more closely relevant to infection of actual human lungs. Indeed, a third of the cell populations in a typical hPCLS that contain alveoli and minor airways with at least one major airway used in this study are CD45 +leukocytes/lymphocytes, with alveolar macrophages, dendritic cells, classical monocytes, and interstitial macrophages representing the bulk of the leukocyte populations (31). Furthermore, the hPCLS platform preserves the actual 3D architecture of the lungs, including the blood vessels and interstitial spaces (Fig. 1A). Because interactions and communications among different cell types and structural components are essential for maintaining lung homeostasis and in responses to pathogen invasion, the hPCLS platform offers a unique perspective for evaluating the pathogenesis of COVID-19 over other recently reported in vitro models such as alveolospheres and lung organoids that lack the relevant cell heterogeneity and other structural components (13, 15). Supplementing additional peripheral blood mononuclear cells would make the hPCLS platform even more closely resembling the lungs in vivo.

Clinical signs following SARS-CoV-2 infection vary widely from asymptomatic to death, although the majority of deaths have occurred in the elderly and those with pre-existing medical conditions. What factors contribute to the varying degrees of clinical manifestations in individuals remains an important but unanswered question. We have attempted to assess the susceptibility of individuals to SARS-CoV-2 infection in hPCLS derived from healthy donors in the age group of 30 to 50 years. We found that while the susceptibility of individuals to SARS-CoV-2 was similar (Fig. 1F and G), their innate immune responses varied widely (Fig. 3), suggesting that different innate immune responses likely contribute to the different clinical presentation of the disease, even within the same age groups. Thus, hPCLS can be used to model variation in infectivity and innate immune responsiveness in human populations. Further experiments by employing increased numbers of hPCLS derived from different age groups, especially from the elderly and those with pre-existing medical conditions, will expand the utility of the hPCLS platform for addressing these critical questions.

Using the hPCLS platform, we observed that the kinetics of infectious SARS-CoV-2 production (as determined by virus titer) does not correlate with that of viral RNA (Fig. 2). For example, while SARS-CoV-2 RNA continues to replicate in the lungs during the period of 4 days p.i., infectious virus production is decreased from 48 h p.i. onward. This result is striking, because in general virus titer correlates with viral RNA accumulation during acute infection in a given host. Whether this phenomenon is unique to SARS-CoV-2 or to the hPCLS platform remains to be seen. This finding raises several interesting questions. For example, do viral RNAs continue to replicate in the lungs without apparent production of infectious virus after 4 days? If so, how long does the viral RNA persist in the lungs and what are the consequences to the host? SARS-CoV-2 RNAs have been detected in patients long after recovery from acute infection (32–34) and in multiple postmortem organs including the brain as late as 230 days after onset of symptoms (35). Viral RNA persistence in the absence of infectious virus has also been described in oligodendrocytes and in mouse brains infected with murine coronavirus (36–38). The observation that cytopathic effects are found in hPCLS on day 5 p.i., when infectious virus could no longer be detected, suggests that continuous viral gene expression in the absence of infectious virus production may directly result in the cytopathic effects seen in histopathological changes of the lungs (Fig. 4). An alternative explanation is that the observed focal cytopathic effect may partially result indirectly from migration of resident lung CD45 +leukocytes within a given lung slice to the local infected sites. Irrespective of the potential mechanisms, this finding may have important clinical implications in post COVID-19 sequelae and warrant further investigation.

While the precise mechanism by which SARS-CoV-2 causes lung diseases is unknown, our results from the multifaceted analysis of this experimental infection platform point an emerging picture of the SARS-CoV-2 pathogenic process that can be divided into three phases (Fig. 8). During the initial phase of infection (first 48 h), SARS-CoV-2 replicates rapidly (Fig. 2) and triggers immediate innate immune responses among others, as evidenced by rapid induction of pro-inflammatory cytokines and chemokines and high levels of secreted IL-8 and IP-10 (Fig. 3 and 8A). Rapid induction of these local inflammatory cytokines likely plays an important role in the development of lung diseases, as supported by clinical evidence that shows that high levels of IL-8 and IP-10 in patients sera correlate with severity of COVID-19 (19–22). Thus, both local (lung) and peripheral IL-8 and IP-10 can be considered reliable predictors for progressive and severe COVID-19. Changes in hPCLS cellular gene expression in response to infection appear in favor of promoting cell growth, survival, and trafficking in the lung environment (Fig. 5, 6, and 8B), which is also characteristic of an early stage of inflammation. At the second phase (48–96 h), while infectious virus production rapidly decreases (Fig. 2A), viral RNA continues to accumulate (Fig. 2B). This may suggest that even though infectious virus is cleared, continuous viral gene expression may have lingering effects on the host. Indeed, clusters of most enriched genes are involved in cellular pathways regulating cell death, such as p53 signaling pathway, mitophagy, and ferroptosis (Fig. 5F), which may play a role in the development of histopathological abnormality at the late stage (third phase) of infection (5 days p.i.) (Fig. 4 and 8), as observed in post-mortem COVID-19 patients (23).

Fig 8.

Schematic presentation of the pathogenic progression following SARS-CoV-2 infection in hPCLS. (A) The pathogenic progression is divided into three phases. Phase one is the initiation phase, represented by rapid virus replication and induction of proinflammatory cytokines and chemokines. Phase two highlights the inverse correlation between viral RNA replication and infectious virus production with no obvious macrostructural changes in the lungs. Phase three illustrates the consequence of SARS-CoV-2 infection in the lungs, as characterized by the development of histopathogical abnormality. Dashed line for viral RNA in this phase is the prediction. (B) Alteration of the host transcriptional landscape by SARS-CoV-2 infection. The number of genes that are up- and down-regulated by SARS-CoV-2 infection at indicated timepoints as in (A) are shown at the top (orange) and bottom (blue), respectively. The overall transcriptional landscape appears switching from promoting trafficking, survival and inflammation at the initial phase to cell death, disfunction, and disease in the second phase, which may drive the disease progression to the final phase.

Type I interferons (IFN-I) are key antiviral cytokines that can be induced upon infection by diverse viruses and play an important role both in host defense against invading viruses and in mediating inflammation. Hence fine tuning of the IFN-I responses often dictates the outcome of the infection or disease, and dysregulation of the IFN-I signaling pathways often contributes to disease pathogenesis. In our transcriptomic profiling, we found a complete absence of IFN-I response to SARS-CoV-2 infection in hPCLS (Fig. 5G; Supplementary Data set S1 to S4), suggesting a protective role of IFN-I in COVID-19 pathogenesis. This interpretation is supported by the identification of impaired IFN-I responses in severe COVID-19 patients that preceded clinical worsening (39), and genetic mutations in Toll-like receptor (TLR)−3-dependent and IRF7-dependent IFN-I immunity (40), TLR7 deficiency (41), or autoantibodies against IFNα, IFNβ, and IFNω (42), as major risk factors for the development of severe COVID-19 associated with pneumonia (43, 44). Defective activation and regulation of IFN-I immunity have also been linked to increased COVID-19 severity in patients as evidenced in postmortem lung tissues from lethal cases of COVID-19 (45) and in peripheral blood of patients with severe or critical COVID-19 (39, 46). Furthermore, upregulated IFN-I responses in asymptomatic COVID-19 infection are associated with improved clinical outcome (47). However, robust IFN-I responses have also been reported in peripheral blood mononuclear cells (PBMCs) from patients with severe COVID-19 (48–50), in bronchoalveolar lavage fluid from COVID-19 patients (51), and in human lung stem cell-based alveolospheres following SARS-CoV-2 infection (13). The apparent contradictory roles of IFN-I responses during SARS-CoV-2 infection as reported might be explained by a number of variables, such as the type of cells and tissues being analyzed, the methods being used, the timing of the sample collection, and specific subsets of IFN-I or interferon-stimulated genes (ISGs). Our transcriptomic profiling revealed a striking similarity in IFN-I signaling in hPCLS after SARS-CoV-2 infection and in postmortem lung tissues from lethal COVID-19 patients (45). Therefore, the hPCLS may provide an ideal platform for further delineating the roles and mechanisms of IFN-I responses in COVID-19 pathogenesis. It is worth noting that although IFN-α2 and IFN-L1 are predicted by bioinformatics analysis to be upstream molecules (Fig. 6A), their expression levels in hPCLS have not been altered by SARS-CoV-2 infection in both proteomics and transcriptomics analyses (Supplementary Data set S1 to S4).

Using the hPCLS platform we confirmed the antiviral activity of HHT against SARS-CoV-2 and identified new parameters that can help evaluate its therapeutic effect on the clinical outcome of COVID-19, as evidenced by the inhibition of inflammatory cytokine and chemokine expression and the elimination of histopathological abnormalities caused by SARS-CoV-2 infection (Fig. 7). Although it is not known at the present time whether HHT inhibits the induction of these cytokines and chemokines directly or through inhibition of SARS-CoV-2 replication indirectly, or both, the increase of IL-1β following virus infection and HHT treatment compared to virus infection alone suggests that HHT may have a direct effect on cytokine induction (Fig. 7E). This assumption is further supported by the evidence that HHT reduced the level of IP-10 in mock-infected hPCLS or reduced IP-10 in SARS-CoV-2-infected hPCLS to a level even lower than in mock-infected and mock-treated hPCLS (Fig. 7E). The varying degrees of innate immune responses among different donor hPCLS to SARS-CoV-2 infection and HHT treatment appear to reflect human population heterogeneity. Thus, the hPCLS platform has many advantages over traditional cell culture systems for preclinical testing of antiviral drugs, in that the hPCLS platform can evaluate not only cytotoxicity and antiviral efficacy, but also host factors involved in pathogenesis of respiratory viral pathogens and potential side-effects of a given drug. It is worth noting that hPCLS revived from cryopreserved lungs are susceptible to SARS-CoV-2 infection (data not shown), demonstrating the feasibility of using cryopreserved tissues. This allows the establishment of a “library” of donated lungs for continuous antiviral drug testing, which would resemble a clinical trial at the bench.

In summary, we have established the utility of hPCLS as an infection platform for studying the initial phase of SARS-CoV-2 pathogenesis and for evaluating the efficacy of antiviral drugs. We showed that during the initial stage of infection SARS-CoV-2 replicates rapidly in hPCLS, concomitant with a rapid induction of multiple pro-inflammatory cytokines and chemokines, which is consistent with the observations from COVID-19 patients. At the late stage, infectious viruses decreased rapidly while viral RNAs persisted and histopathological changes ensued. Transcriptomic and proteomic analyses identified molecular signatures and cellular pathways that are largely consistent with the disease progression. Furthermore, we have demonstrated that HHT is an effective antiviral that limits SARS-CoV-2 replication, may modulate host inflammatory responses to the advantage of the host and ameliorates histopathological abnormality caused by SARS-CoV-2 infection.

MATERIALS AND METHODS

Virus, cell lines, and reagents

SARS-CoV-2 strain USA-WA1/2020 was obtained through BEI Resources, NIAID, NIH, and was propagated in Vero cells. Vero cells were obtained from ATCC and A549-hACE2 cells were provided by Ralph Baric (University of North Carolina at Chapel Hill). All cells were cultured in Dulbecco modified Eagle medium (DMEM) (Gibco) containing 10% fetal bovine serum (FBS) and 1% penicillin and streptomycin (PS) at 37°C in 5% CO₂. Homoharringtonine (HHT), a natural plant alkaloid derived from Cephalotoxus fortunei, was purchased from Sigma (cat# SML-1091–10MG). A stock solution of 10 mM was prepared in DMSO and stored at −80°C for further use.

Preparation of human precision-cut lung slices (hPCLS)

Lungs were obtained from anonymous donors through transplant teams of the Arkansas Regional Organ Recovery Agency (ARORA) and by the National Disease Research Interchange (NDRI). The lung vasculature was perfused with phosphate-buffered saline (PBS) to wash out residual blood and clots. The lobes were surgically dissected, and the major bronchi were cannulated. Individual lobes were inflated with sterile 1.8% low-gelling-temperature agarose in PBS at 37°C. After inflation, the bronchi were clamped, and incubated at 4 to 7°C for 2 to 3 hours to allow the agarose to solidify. The hardened lungs were cut into ~12-mm-thick sections, and cross-sectioned airways were identified and collected with an 8.5-mm-diameter coring tool under a dissecting microscope. The cores (80–100 per lobe) were further cut into 600-µm-thick slices. Slices were then cultured in 48-well plates in DMEM/Ham’s nutrient mixture F-12 medium (DMEM-F12; 1:1) supplemented with 10% FBS, antibiotic-antimycotic, and antibiotic formulation (Primocin) at 37°C in 5% CO₂ with continuous agitation in a humidified incubator for 10–14 days prior to virus infection.

Infection of hPCLS and cells, and determination of virus titer

For virus infection, hPCLS were washed once with 1 × PBS and infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ (50% tissue culture infectious dose) per slice in a 48-well tissue culture plate at 37°C. At 1 h p.i., the virus inoculums were removed and hPCLS were rinsed with 1 × PBS. Following addition of DMEM-F12, hPCLS were incubated at 37°C in 5% CO₂ for various periods of time as indicated. Virus titers in the supernatants were determined by the standard TCID₅₀ assay on Vero cells, and were expressed as TCID₅₀/mL. For determining virus replication kinetics in cell cultures, Vero and A549/ACE2 cells were infected with SARS-CoV-2 at a MOI (multiplicity of infection) of 1, and the supernatants were harvested at various time points p.i. as indicated for determination of virus titers by TCID₅₀ assay.

Immunofluorescence assay and hematoxylin-eosin staining of hPCLS

SARS-CoV-2-infected and mock-infected hPCLS were fixed with 10% formalin for 30 min at room temperature, and then processed, embedded with paraffin, sectioned, and stained with hematoxylin-eosin (H&E) at the Experimental Pathology Core facility at UAMS. Unstained slides were used for detection of viral proteins by immunofluorescence assay. Briefly, to remove paraffin, slides were treated sequentially in the fume hood with: (i) xylene 3 times, 5 min each, (ii) xylene/100% ethanol, 15–20 dips, (iii) 100% ethanol, 15–20 dips, iv) 95% ethanol, 15–20 dips, v) 70% ethanol, 15–20 dips. The samples were fixed with 75% acetone/25% ethanol for 10 min at −20°C and rehydrated by washing slides twice with 1 × PBS for 2 min each. The slides were placed in a humidified chamber, and blocked with 2% normal horse serum (NHS) in 1 × PBS for 10 min. The slides were stained with the primary monoclonal antibody against SARS-CoV-2 nucleocapsid (N) protein (1:50 dilution) (BEI Resources, NR-619) at 4°C overnight followed by washing twice with 1 × PBS, and the secondary anti-mouse IgG antibody conjugated with FITC (1:100 dilution, Sigma) followed by washing twice with 1 × PBS. The cell nuclei were then stained with DAPI for 2 min at room temperature followed by washing twice with 1 × PBS. The slides were observed under a fluorescence microscope (Olympus IX-70), and images were captured with the attached digital camera (Zeiss).

Cytokine measurements

Virus-infected or mock-infected hPCLS were cultured in DMEM-F12 for various periods of time as specified in each experiment and culture supernatants were collected and stored at −80°C until use. To inactivate SARS-CoV-2 prior to removing the samples from the BSL-3 laboratory for cytokine assay, supernatants were placed at a distance of 14 cm from the UV light in a UV cross-linker (Fisher Scientific) and exposed to UV light at an energy level of 1,200 µW/ms for 15 min. Virus inactivation was confirmed by the absence of cytopathic effect on Vero cells following infection with the inactivated samples and the absence of viral nucleoprotein by immunofluorescence assay. Cytokines were measured by flow cytometry using cytometric bead array kits (BD Biosciences) following the manufacturer’s instruction. The human inflammatory cytokine kit (IL-8, IL-1β, IL-6, IL-10, TNF-α, IL-12p70) (Cat.# 551811) and human chemokine kit (CXCL8/IL-8, CXCL9/MIG, CXCL10/IP-10, CCL2/MCP-1, CCL5/RANTES) (Cat.# 552990) were used. The amount of each cytokine/chemokine in virus-infected hPCLS was normalized to that in mock-infected hPCLS and was expressed as pg/mL.

RNA isolation and quantitative reverse transcription-PCR (qRT-PCR)

For RNA isolation, 3 to 6 slices of the hPCLS were pooled and treated with 1 mL of TRIzol reagent (Invitrogen) at room temperature for 30 min. Total RNAs were then isolated according to the manufacturer’s instruction. Isolated RNA samples were treated with DNase at room temperature for 10 min to remove DNA contaminants, purified with RNeasy mini spin column kit (Qiagen, Cat#74104), and quantified with NanoDrop 2000c spectrophotometer (Thermo Scientific). qRT-PCR was carried out according to the manufacturer’s instruction (BioRad). Briefly, for cDNA synthesis, 1 µg of each RNA sample was used for RT with iScript RT Supermix (BioRad cDNA kit, cart#1708841) or iScript NO-RT control Supermix for negative control. The RT reaction was carried out in a thermal cycler (BioRad) at 25°C for 5 min and at 46°C for 20 min, and was terminated at 95°C for 1 min. PCR was carried out with iTaq Universal SYBY green Supermix kit (BioRad cat# 1725121) in the MicroAmp optical 96-well plate (Applied Biosystems, cat#N8010560) in a thermal cycler (QuantStudio 6 Flex, Applied BioSystems by Thermo Fisher Scientific). The primer pair specific to SARS-CoV-2 N gene (forward primer: 5′-ATG CTG CAA TCG TGC TAC AA-3′; reverse primer: 5′-GAC TGC CGC CTC TGC TC-3′) or to cellular housekeeping gene GAPDH (forward primer: 5' TGA TGA CAT CAA GAA GGT GGT GAA G −3'; reverse primer: 5'TCC TTG GAG GCC ATG TGG −3') were used for amplifying viral and cellular RNA, respectively. The amount of viral RNA in each sample was normalized to that of GAPDH and expressed as fold change relative to mock-infected sample.

Gene expression profiling by RNAseq

For RNAseq analysis, a group of 6 hPCLS each were mock-infected or infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀/slice. At various times (24, 48, 72, and 96 h) p.i., RNAs were extracted from the hPCLS with TRIzol reagent (Invitrogen) and quantified with NanoDrop (ThermoFisher). The RNA samples were checked for quality using Bioanalyzer (Agilent 2100) prior to RNAseq analysis per Novogene protocol (www.novogene.com). The samples were sequenced and analyzed by Novogene. The reads were mapped using STAR (v2.5) (52)to the reference genome and HTSeq (v0.6.1) (53) as used to count the reads mapped to each gene. FPKM of each gene was calculated and differential expression analysis was performed using DESeq2 (v2_1.6.3) (54). The resulting P-values were adjusted using the Benjamini and Hochberg’s approach for controlling the False Discovery Rate (FDR). Genes with an adjusted P-value < 0.05 were considered differentially expressed.

Gene ontology (GO) enrichment analysis of differentially expressed genes was performed using clusterProfiler R package to test the statistical enrichment of differential expression genes in KEGG pathways (55, 56). Volcano plots were created using VolcaNoseR (57).

Proteomic analysis

For proteomic analysis, 5 hPCLS each were mock-infected or infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀/slice. At 48 h p.i., each slice was lysed with 100 µL of the radioimmunoprecipitation assay (RIPA) buffer (10 mM Tris-Cl, pH 8.0, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl) (Thermo Scientific, Cat.# 89901) containing a cocktail of protease inhibitors (Sigma, Cat.# P8340-5ML) and phosphatase inhibitors (Fisher Scientific, Cat.# PIA32957) at room temperature for 30 min followed by repeated pipetting. Lysates were clarified of debris by centrifugation and proteins were analyzed by using CME bHPLC phosphoTMT Methods – Orbitrap Eclipse in the IDeA National Resource for Quantitative Proteomics facility on UAMS campus. Specifically, total protein from each sample was reduced, alkylated, and purified by chloroform/methanol extraction prior to digestion with MS-grade porcine trypsin/LysC (Promega). Resulting peptides were labeled using a tandem mass tag 11-plex isobaric label reagent set (Thermo), combined into four TMT multiplex groups with a common pooled reference sample, and enriched using High-Select TiO2 and Fe-NTA phosphopeptide enrichment kits in succession (Thermo) following the manufacturer’s instructions. Both enriched and un-enriched labeled peptides were separated into 46 fractions on a 100 × 1.0 mm Acquity BEH C18 column (Waters) using an UltiMate 3000 UHPLC system (Thermo) with a 40 min gradient from 99:1 to 60:40 buffer A (0.1% formic acid, 0.5% acetonitrile):B (0.1% formic acid, 99.9% acetonitrile) ratio under basic pH conditions, and then consolidated into 18 super-fractions. Each super-fraction was then further separated by reverse phase XSelect CSH C18 2.5 um resin (Waters) on an in-line 150 × 0.075 mm column using an UltiMate 3000 RSLCnano system (Thermo). Peptides were eluted using a 75 min gradient from 98:2 to 60:40 buffer A:B ratio. Eluted peptides were ionized by electrospray (2.4 kV) followed by mass spectrometric analysis on an Orbitrap Eclipse Tribrid mass spectrometer (Thermo) using multi-notch MS3 parameters. MS data were acquired using the FTMS analyzer in top-speed profile mode at a resolution of 120,000 over a range of 375 to 1500 m/z. Following CID activation with normalized collision energy of 31.0, MS/MS data were acquired using the ion trap analyzer in centroid mode and normal mass range. Using synchronous precursor selection, up to 10 MS/MS precursors were selected for HCD activation with normalized collision energy of 55.0, followed by acquisition of MS3 reporter ion data using the FTMS analyzer in profile mode at a resolution of 50,000 over a range of 100–500 m/z.

Data analysis—ProteoViz (phosphoTMT)

Proteins were identified and reporter ions quantified by searching the UniprotKB database using MaxQuant (Max Planck Institute) with a parent ion tolerance of 3 ppm, a fragment ion tolerance of 0.5 Da, a reporter ion tolerance of 0.001 Da, trypsin/P enzyme with two missed cleavages, variable modifications including oxidation on M, Acetyl on Protein N-term, and phosphorylation on STY, and fixed modification of Carbamidomethyl on C. Protein identifications were accepted if they could be established with less than 1.0% false discovery. Proteins identified only by modified peptides were removed. Protein probabilities were assigned by the Protein Prophet algorithm (58). TMT MS3 reporter ion intensity values are analyzed for changes in total protein using the unenriched lysate sample. Phospho(STY) modifications were identified using the samples enriched for phosphorylated peptides. The enriched and un-enriched samples are multiplexed using two TMT10-plex batches, one for the enriched and one for the un-enriched samples. Following data acquisition and database search, the MS3 reporter ion intensities were normalized using ProteiNorm (59). The data were normalized using Cyclic Loess (60)and analyzed using ProteoViz to perform statistical analysis using Linear Models for Microarray Data (limma) with empirical Bayes (eBayes) smoothing to the standard errors (61). A similar approach is used for differential analysis of the phosphopeptides, with the addition of a few steps. The phosphosites were filtered to retain only peptides with a localization probability >75%, filter peptides with zero values, and log2 transformed. Limma was also used for differential analysis. Proteins and phosphopeptides with an FDR-adjusted P-value < 0.05 and an absolute fold change >2 were considered significant.

Antiviral drug testing

hPCLS were infected with SARS-CoV-2 at 2 × 10⁵ TCID₅₀ per slice and treated with homoharringtonine (HHT) at 1 h p.i. at various concentrations as indicated. hPCLS treated with vehicle (medium containing 0.1% or 0.01% DMSO) were used as a negative control. At 48 h p.i., culture supernatants were collected for determination of virus titer by TCID₅₀ assay and for measuring the cytokines and chemokines using the human inflammatory cytokine and chemokine CBA kits as described above.

Cell viability assay

Cells grown in 96-well plates were incubated for 48 h with HHT at various concentrations as indicated, and then cell viability was determined using the XTT assay kit TOX2-1KT according to the manufacturer’s instruction (Sigma-Aldrich). Medium containing DMSO at 0.1% or less was used as vehicle control.

Lactate dehydrogenase (LDH) assay

For assessing the cytotoxity of HHT in hPCLS, hPCLS were treated with various concentrations of HHT for 48 h, or with 0.1% DMSO as a vehicle control, or untreated as a negative control for background level of LDH. Culture media were then collected for determining LDH activity using CyQUANT LDH cytotoxicity assay kit (Cat.# C20300) according to the manufacturer’s instruction (Invitrogen). For positive control, hPCLS were treated with 1% triton X-100 at 37°C for 45 min, and the supernatants were then assayed for LDH activity.

Statistical analysis

Statistical analyses on cytokine and chemokine data were performed using the Tukey’s multiple comparisons test in the GraphPad Prism 9 program (v9.5.0). Other statistical analyses were carried out with unpaired t test or one way ANOVA in the same Prism 9 program. Results with P values of >0.05, <0.05, <0.01, <0.001, and <0.0001 are indicated in the figures and legends.

DATA AVAILABILITY

All data supporting the findings of this study are found within the paper and its supplemental material and are available from the corresponding author upon request. RNAseq data have been deposited in the Gene Expression Omnibus (GEO) database under accession GSE226702. The proteomics data have been deposited in the MassIVE repository with accession link ftp://massive.ucsd.edu/MSV000091383/.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00794-24.

Dataset S1. jvi.00794-24-s0001.xlsx.

List of differentially expressed genes in hPCLS following SARS-CoV-2 infection at 24 h p.i. (COV24H) as compared to mock infection (COVM).

Dataset S2. jvi.00794-24-s0002.xlsx.

List of differentially expressed genes in hPCLS following SARS-CoV-2 infection at 48 h p.i. (COV48H) as compared to mock infection (COVM).

Dataset S3. jvi.00794-24-s0003.xlsx.

List of differentially expressed genes in hPCLS following SARS-CoV-2 infection at 72 h p.i. (COV72H) as compared to mock infection (COVM).

Dataset S4. jvi.00794-24-s0004.xlsx.

List of differentially expressed genes in hPCLS following SARS-CoV-2 infection at 96 h p.i. (COV96H) as compared to mock infection (COVM).

Supplemental figures. jvi.00794-24-s0005.pdf.

Fig. S1 and S2.

Dataset S1 legend. jvi.00794-24-s0006.pdf.

Legend for Dataset S1.

Dataset S2 legend. jvi.00794-24-s0007.pdf.

Legend for Dataset S2.

Dataset S3 legend. jvi.00794-24-s0008.pdf.

Legend for Dataset S3.

Dataset S4 legend. jvi.00794-24-s0009.pdf.

Legend for Dataset S4.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.