Distinct Severe Acute Respiratory Syndrome Coronavirus-Induced Acute Lung Injury Pathways in Two Different Nonhuman Primate Species

Approval for animal experiments was obtained from the Institutional Animal Welfare Committee. Four cynomolgus macaques (Macaca fascicularis) and four African green monkeys (Chlorocebus aethiops), 3 to 5 years old with active temperature transponders in the peritoneal cavity, were inoculated with 5 × 10⁶ 50% tissue culture infective doses (TCID₅₀) of SARS-CoV strain HKU39849 suspended in 5 ml phosphate-buffered saline (PBS). Four milliliters was applied intratracheally, 0.5 ml intranasally, and 0.25 ml on each conjunctiva, as described previously (7, 12, 15, 27, 46). The four SARS-CoV-infected young adult cynomolgus macaques have been described previously (46). For reverse transcription (RT)-PCR and immunohistochemistry, data from previously SARS-CoV-infected aged cynomolgus macaques were used (46). Two young adult macaques and African green monkeys (age range, 3 to 9 years old) were mock (PBS) infected. The animals were checked daily for clinical signs. Just before infection and at days 2 and 4 postinfection, animals were anesthetized with ketamine, and oral, nasal, and rectal swabs were taken and placed in 1 ml Dulbecco's modified Eagle's medium supplemented with 100 IU penicillin/ml and 100 μg of streptomycin/ml (virus transport medium). The swabs were frozen at −70°C until RT-PCR analysis was performed. The animals were euthanized 4 days after infection by exsanguination under ketamine anesthesia.

Necropsies were performed according to a standard protocol. For semiquantitative assessment of gross pathology, the percentage of affected lung tissue from each lung lobe was determined at necropsy and recorded on a schematic diagram of the lung, and the area of affected lung tissue was subsequently calculated (gross pathology score). One lung from each monkey was inflated with 10% neutral-buffered formalin by intrabronchial intubation and suspended in 10% neutral-buffered formalin overnight. Samples were collected in a standard manner (from the cranial, medial, and caudal parts of the lung), embedded in paraffin, cut at 3 μm, and used for titration (see below) or immunohistochemistry (see below) or stained with hematoxylin and eosin (H&E). The lung, liver, spleen, kidney, intestine, trachea, and tracheobronchial lymph node H&E sections were examined by light microscopy.

For semiquantitative assessment of SARS CoV infection-associated inflammation in the lung, two H&E-stained slides per animal (cranial and caudal lobes) were examined for inflammatory foci by light microscopy using a 10× objective. Each focus was scored for size (1, smaller than or equal to the area of the 10× objective; 2, larger than the area of the 10× objective and smaller than or equal to the area of the 2.5× objective; 3, larger than the area of the 2.5× objective) and degree of inflammation (1, low density of inflammatory cells; 2, intermediate density of inflammatory cells; 3, high density of inflammatory cells). The cumulative scores for the inflammatory foci provided the total score per animal. Sections were examined without knowledge of the identities of the animals.

Three lung tissue samples (one from the cranial part of the lung, one from the medial part, and one from the caudal part) were homogenized in 2 ml virus transport medium, using Polytron PT2100 tissue grinders (Kinematica). After low-speed centrifugation, the homogenates were frozen at −70°C until they were inoculated on Vero E6 cell cultures in 10-fold serial dilutions. The identity of the isolated virus was confirmed as SARS-CoV by RT-PCR of the supernatant (15).

Serial 3-μm lung sections were stained using mouse anti-SARS nucleocapsid IgG2a (clone Ncap4; Imgenex; 1:1,600), mouse anti-osteopontin (clone AKm2A1; Santa Cruz), rabbit anti-ACE2 (angiotensin 1-converting enzyme 2) (ab15348; Abcam), or isotype control antibodies (clones 11711 and 20102; R&D) according to standard protocols (15, 27). Quantitative assessment of SARS-CoV antigen expression in the lungs was performed as described previously (15). For phenotyping (to confirm cell types), we used a destaining-restaining technique (8). Briefly, the red precipitate (aminoethylcarbazole [AEC]) used for visualization of osteopontin antigen staining was dissolved in graded 100% to 70% alcohols. To detach the primary antibody and red immunohistochemistry signals, slides were soaked in eluting buffer (5 ml 0.1 M HCl, 95 ml 0.1 M NaCl containing 1 M glycine) for 2 h. The sections were treated for two 5-min intervals by heating them in citric acid buffer, pH 6.0, to denature any undetached primary antibody. The slides were then incubated with mouse anti-human CD68 (clone KP1; Dako) 1/200 in PBS-0.1% bovine serum albumin (BSA) for 1 h at room temperature. After being washed, the sections were incubated with horseradish peroxidase-labeled goat-anti-mouse IgG1 (Southern Biotech) 1/100 in PBS-0.1% BSA for 1 h at room temperature. Peroxidase activity was revealed by incubating the slides in 3,3′-diaminobenzidine-tetrachlorhydrate (DAB) (Sigma) for 3 to 5 min, resulting in a brown precipitate, followed by counterstaining with hematoxylin.

RNA from 200 μl of swabs was isolated with the Magnapure LC total nucleic acid isolation kit (Roche) external lysis protocol and eluted in 100 μl. SARS-CoV RNA was quantified on the ABI prism 7700, with use of the TaqMan Reverse Transcription Reagents and TaqMan PCR Core Reagent kit (Applied Biosystems), using 20 μl isolated RNA, 1× TaqMan buffer, 5.5 mM MgCl₂, 1.2 mM deoxynucleoside triphosphates (dNTPs), 0.25 U AmpliTaq gold DNA polymerase, 0.25 U Multiscribe reverse transcriptase, 0.4 U RNase inhibitor with 200 nM primers and 100 nM probe specific for the SARS-CoV nucleocapsid protein gene (7, 46). The amplification parameters were 30 min at 48°C, 10 min at 95°C, and 40 cycles of 15 s at 95°C and 1 min at 60°C. RNA dilutions isolated from a SARS-CoV stock were used as a standard. Average results (± standard errors of the mean [SEM]) for macaque (n = 4) and African green monkey (n = 4) groups were expressed as SARS-CoV equivalents per ml swab medium.

Lung tissue samples (0.3 to 0.5 g) were taken for RT-PCR and microarray analysis in RNA-later (Ambion, Inc.). RNA was isolated from homogenized postmortem tissue samples using Trizol reagent (Invitrogen) and the RNeasy minikit (Qiagen). cDNA synthesis was performed with ∼1 μg total RNA and Superscript III reverse transcriptase (Invitrogen) with oligo(dT), according to the manufacturer's instructions. Semiquantitative RT-PCR was performed as described previously to detect SARS-CoV mRNA and to validate cellular gene expression changes as detected with microarrays of SPP1, CCL3, and CCL20 (Applied Biosystems; Hs00167093_m1, Rh02788104_gH, and Rh02788116_m1) and IL-8, IL-6, and IFN-β (7). Differences in gene expression are represented as the fold change in gene expression relative to a calibrator and normalized to a reference, using the 2^−ΔΔCt method (29). GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as an endogenous control to normalize quantification of the target gene. The samples from the mock-infected macaques or African green monkeys were used as a calibrator. Average results (±SEM) for groups were expressed as the fold change compared to PBS-infected animals (29).

Data (RT-PCR, gross pathology, and histology scores for SARS-CoV-infected animals) were compared using Student's t test. Differences were considered significant at a P value of <0.05.

Pooled total RNA (2.4 μg) from one to three separate lung pieces per animal, with substantial SARS-CoV replication (>10⁵-fold change), was labeled using the One-Cycle Target Labeling Assay (Affymetrix) and hybridized onto Affymetrix GeneChip Rhesus Macaque Genome Arrays (Affymetrix) according to the manufacturer's recommendations. Image analysis was performed using Gene Chip Operating Software (Affymetrix). Microarray Suite version 5.0 software (Affymetrix) was used to generate .dat and .cel files for each experiment. The data were normalized using a variance stabilization algorithm (VSN) (20). Transformed probe values were summarized into one value per probe set by the median polish method (53). Primary data are available at http://www.virgo.nl in accordance with proposed minimum information about a microarray experiment (MIAME) standards.

Principal-component analysis (PCA) was performed on a single normalized data set, including all African green monkey and cynomolgus macaque samples. For PCA, 1,147 probe sets were included that deviated ≥1.5-fold from the geometrical mean (absolute values) for at least one of the samples. Differentially expressed mRNA transcripts were identified using LIMMA (version 2.12.0) (47). Correction for multiple testing was achieved by requiring a false-discovery rate (FDR) of 0.05, calculated with the Benjamini-Hochberg procedure (1). To understand the gene functions and the biological processes represented in the data and to obtain differentially expressed molecular and cellular functions, Ingenuity Pathways Knowledge Base (Ingenuity Systems) was used. Heat maps of proinflammatory pathways were generated using complete linkage and Euclidian distance in Spotfire DecisionSite for Functional Genomics version 9.1 (Spotfire; Tibco) and Ingenuity Pathways Knowledge Base (Ingenuity Systems), using log (base 2)-transformed expression values, with minimum and maximum values of the color range of −4 and 4, respectively.

Cynomolgus macaques develop pathology upon SARS-CoV infection, and the severity is associated with aging (46). In this study, we compared SARS-CoV infection in two related nonhuman primate host species, young adult cynomolgus macaques and African green monkeys, focusing on the latter. Four young adult African green monkeys were infected with SARS-CoV HKU39849, and as a control, two were PBS infected. During the 4-day experiment, all animals remained free of clinical symptoms. Their body temperatures, measured by transponders in the peritoneal cavity, demonstrated a rhythmic pattern, with temperatures fluctuating between 35°C at night and 39°C during the day, prior to infection (Fig. 1A). Between days 1 and 2 after SARS-CoV infection, an increase in body temperature was recorded during the night in all animals. In addition, some African green monkeys showed an increase in body temperature up to 40°C during the first 2 days after SARS-CoV infection (Fig. 1A). Even at days 3 and 4 postinfection, elevated temperatures were still observed in some African green monkeys. At gross necropsy, 4 days postinfection, the lungs of African green monkeys showed (multi)focal pulmonary consolidation (Fig. 1B and C). The consolidated lung tissue was gray-red, firm, level, and less buoyant than normal. Strikingly, one animal had severe lesions, with up to 30% of total lung tissue affected (Fig. 1B). Four young adult cynomolgus macaques infected with SARS-CoV HKU39849 and two PBS-infected cynomolgus macaques were used for comparative analyses (46). In contrast to infected African green monkeys, no increase in body temperature of up to 40°C during the first 2 days after SARS-CoV infection was observed in macaques (46). Between days 1 and 2 after SARS-CoV infection, an increase in body temperature was recorded during the night in three out of four animals (46). Temperatures returned to normal from day 3 postinfection onward (46). The lungs of young adult macaques showed small, patchy, macroscopic lesions (Fig. 1C) (46).

The lesions in the lungs of African green monkeys consisted of acute exudative diffuse alveolar damage characterized by thickened alveolar septa with infiltration of moderate numbers of neutrophils and macrophages, lymphocytes, and plasma cells, and necrosis with karyolysis, karyorrhexis, and loss of cellular detail (Fig. 2A). Multifocally, moderate hypertrophy and hyperplasia of type II pneumocytes was observed (Fig. 2A and B). A variable amount of eosinophilic proteinaceous fluid (edema and exudate) with fibrin and eosinophilic hyaline membrane formation (Fig. 2A and B), few syncytia, and moderate numbers of foamy alveolar macrophages, neutrophils, and erythrocytes (hemorrhage) was seen in alveolar lumina. Mild necrosis, hypertrophy, and hyperplasia of bronchiolar epithelium were encountered (Fig. 2A). Peribronchiolar, peribronchial, and perivascular infiltrations of moderate numbers of lymphocytes, plasma cells, macrophages, and neutrophils with mild to moderate edema were seen (Fig. 2A). There was a mild to moderate proliferation of the bronchus-associated lymphoid tissue (BALT), with focal nodular lymphocytic proliferation and formation of follicles with additional exocytosis of neutrophils in the tracheal epithelium (Fig. 2A). A mild multifocal lymphoplasmacytic tracheobronchoadenitis, characterized by a multifocal mild infiltration of lymphocytes, macrophages, plasma cells, and neutrophils in and surrounding the bronchial seromucous glands with mild necrosis of the epithelium, was observed in all African green monkeys (Fig. 2A). Two animals showed evidence of multifocal mild lymphoplasmacytic pleuritis consisting of a mild multifocal distention of the pleura with infiltration of a few lymphocytes, plasma cells, macrophages, and neutrophils and mild edema. The lymphoid organs were activated, with proliferation of the tracheobronchial lymph node and spleen, characterized by increased numbers of neutrophils and histiocytes and distinct germinal centers in the lymphoid follicles. No significant lesions were seen in other tissues examined or in the tissues of negative-control animals. The histopathological lesions in the respiratory tracts of SARS-CoV-infected cynomolgus macaques were largely similar in character to those described for the African green monkeys (46). However, hyaline membrane formation was not observed, and the lesions in macaques were significantly less severe and extensive than those in African green monkeys (Fig. 3A), which corroborated the gross pathology scores (Fig. 3B).

SARS-CoV mRNA levels were detected in throat (Fig. 4A) and nasal (Fig. 4B) swabs from African green monkeys on days 2 and 4 postinfection and in rectal swabs on day 4 postinfection (data not shown). In addition, virus replication in the lungs was demonstrated by RT-PCR and titration, and immunohistochemistry revealed SARS-CoV antigen expression in the lungs of African green monkeys within or directly adjacent to areas with lesions (Fig. 4C and D). In alveoli of African green monkeys, cells that resembled type I and type II pneumocytes and rarely alveolar macrophages expressed SARS-CoV antigen (Fig. 5A). SARS-CoV antigen was not detected in bronchioles or bronchi but was observed in ciliated epithelial cells in the tracheas of two African green monkeys (Fig. 5A). The differences in the percentages of SARS-CoV-infected cells in the lungs or virus titers in the throat, noses, and lungs of African green monkeys and macaques were not statistically significant (Fig. 4). SARS-CoV, however, was not detected in ciliated epithelial cells in the trachea on day 4 postinfection in young adult macaques (Fig. 5) or in rectal swabs. In both PBS-infected African green monkeys and macaques, we could show expression of ACE2, which is the receptor for SARS-CoV, in cells morphologically resembling type II epithelial cells (Fig. 5B). Significant differences between ACE2 expression levels in African green monkeys and macaques could not be observed. Overall, African green monkeys displayed more severe pathology than macaques upon infection with SARS-CoV, which correlated with a broader range of cell types infected at day 4 postinfection, but not with the extent of virus replication.

To understand why SARS-CoV infection in African green monkeys results in more severe pathology than in young adult macaques, we generated genome-wide mRNA expression profiles from lung tissue with substantial SARS-CoV replication (Fig. 6A). Principal-component analysis was performed on a single normalized data set of African green monkey samples, together with macaque samples. This unsupervised method clearly separated SARS-CoV-infected lung mRNA expression profiles from PBS control expression profiles for both species. In addition, the infected and uninfected expression profiles of the two species also separated well. Species- and infection-related expression profile differences coincided with principal components 1 and 2, respectively (Fig. 6B). These results show that species-associated mRNA expression profile differences are maintained upon SARS-CoV infection, suggesting that African green monkeys and macaques respond differently to SARS-CoV infection.

To analyze the host response to infection in African green monkeys and macaques in more detail, mRNA transcripts differentially expressed between infected and uninfected lung samples were identified using the analysis of variance (ANOVA)-like method LIMMA (47) for both species separately. To obtain a maximum view of infection-induced host response differences between African green monkeys and macaques, the microarray data for the two species were normalized separately. Infected African green monkeys expressed 556 gene transcripts differently from PBS-infected animals. These differently expressed genes were analyzed within the context of molecular pathways, using a functional-analysis approach with biologically related genes (Ingenuity). The most significantly regulated molecular/cellular functions in African green monkeys compared to PBS-infected animals were associated with a proinflammatory response and included cellular growth and proliferation, cell death, cell movement, and cell-to-cell signaling (Fig. 7A). Upon analysis of the gene transcripts within the context of biological pathways using Ingenuity Knowledge Base, the top gene interaction network in African green monkeys showed a strong antiviral response with differentially expressed type I interferon-stimulated genes and genes associated with the induction of type I interferons (Fig. 7B), including ISG15, ISG20, DDX58, DHX58, IFI/IFITs, MX1, MX2, IRF7, IRF9, OAS1, and OAS2. In addition, the network contained NF-κB (Fig. 7B), a transcription factor implicated in proinflammatory host responses and development of ALI/ARDS (10, 17, 21). Differential expression of genes associated with acute lung injury, inflammation, and/or hypoxia signaling in African green monkeys included C3AR1, CEBPD, CCL2, CCL3, IL1RN, TIMP1, CDKN1A, SPARC, VEGFA, SPP1, CSF1, CSF3R, CD86, KIT, CCL8, CXCL10, PML, SP100, and SOD2 genes. Some of these genes are also upregulated in SARS patient sera or in patients with ARDS (51, 55). Young adult SARS-CoV-infected macaques expressed 475 gene transcripts differently from PBS-infected animals. These genes were associated with the same molecular/cellular functions observed in African green monkeys, with similar numbers of differentially expressed genes per function (Fig. 7A). Analysis of the gene transcripts in the context of biological pathways revealed differential expression of genes in both antiviral and proinflammatory responses (Fig. 7C). These data suggest that African green monkeys and macaques display similar types of responses upon infection, with strong induction of antiviral and proinflammatory pathways, which has been described for SARS-CoV-infected cynomolgus macaques previously but based on regulation of a different array of individual genes within the antiviral and proinflammatory pathways (7, 46).

As the principal-component analysis suggested that the host response to SARS-CoV infection was different in African green monkeys and macaques, the gene expression profiles of SARS-CoV-infected African green monkeys and macaques (n = 4) were directly compared using LIMMA (8 samples normalized in a single set). A total of 1,607 gene transcripts were differentially expressed. Upon analysis of these 1,607 gene transcripts within the context of biological processes and pathways using Ingenuity Pathways Knowledge Base, this subset of genes showed indications of an innate host response to viral infection. Among the top significantly differentially regulated (P < 0.005) functional categories were inflammatory disease, cell movement, and cell-to-cell signaling and interaction, which included genes like those for CCL20, CXCL1, CXCL2, DEFB1, IL1RL1, IL1RN, MMP7, MMP9, IL-8, SERPINE1, SPP1, TFPI2, and VCAM1, which were either up- or downregulated, depending on the gene (Fig. 8A to C), indicating that they are upregulated in either African green monkeys or macaques. These data suggest that both host species and SARS-CoV infection are factors involved in determining the transcription of cellular genes and that significant differences exist in the proinflammatory host response to SARS-CoV infection, corresponding to the host species.

In order to understand the host responses in the context of host species, we directly compared lung samples from PBS-infected African green monkeys (n = 2) and macaques (n = 2). LIMMA analysis revealed that 2,198 gene transcripts were differentially expressed (change, ≥2-fold; P < 0.05), with categories such as inflammatory disease, cell death, cell movement, and cellular growth and proliferation among the top significantly differentially regulated functions (P < 0.005). Of the 2,198 differentially expressed genes, 917 were also differentially expressed in the direct comparison of SARS-CoV-infected African green monkeys and macaques, but they did not include genes for cytokines/chemokines, such as CCL20, CCL3, and SPP1. Our data indicate that significant differences exist in the basal gene expression levels of African green monkeys and macaques, which may partly explain why differences in pathology were observed after SARS-CoV infection.

To obtain more insight into the differences in the host responses to SARS-CoV infection of African green monkeys and macaques, mRNA expression profiles for specific molecular pathways of infected animals of both species were compared to those of their PBS controls, but also directly against each other. Heat maps were generated for differentially regulated genes involved in ARDS (Fig. 8A), NF-κB signaling (Fig. 8B), and cytokine/chemokine signaling (Fig. 8C), which indicated that the host responses to infection with respect to these pathways in African green monkeys and macaques are different. Surprisingly, inflammatory cytokines that play a major role in mediating and amplifying ALI/ARDS or have neutrophil chemoattractant activity, such as IL-6, IL-8, CXCL1, and CXCL2 (10, 25, 55), were differentially expressed in macaques, but not in African green monkeys (Fig. 8C). TaqMan analysis confirmed that IL-6 and IL-8, key players in ALI/ARDS, were differentially expressed at significantly lower levels in African green monkeys than in young adult macaques (Fig. 9A and B), in contrast to the IFN-β levels, which were not significantly different (Fig. 9C). Moreover, in the direct comparison of the microarray data on SARS-CoV-infected African green monkeys and macaques, IL-8, CXCL1, and CXCL2 were also differentially expressed (Fig. 8A to C). These observations suggest that classical pathways known to be involved in development of ALI/ARDS are differently regulated in these two nonhuman primate species. From previous studies, we know that in aged cynomolgus macaques, IL-8 gene expression is further upregulated (Fig. 9A) (46) and correlated with significantly more severe pathology. Because African green monkeys showed even more pathology than young adult macaques, other pathogenic pathways besides the induction of genes such as those for IL-6 and IL-8 are most likely involved in the observed pathological changes (21, 46).

Interestingly, CCL3, SPP1, and CCL20 were induced strongly in African green monkeys, but not in young adult macaques (Fig. 8A to C and 9D to F). This observation was corroborated by the direct comparison of SARS-CoV-infected African green monkeys with macaques, which showed that CCL20 and SPP1 were differentially expressed (Fig. 8B and C). It is notable that both these genes were also statistically significantly upregulated in aged macaques compared to young adult macaques (Fig. 9D to F) (46). CCL3 is involved in the acute inflammatory state in the recruitment and activation of polymorphonuclear leukocytes, whereas CCL20 is strongly chemotactic for lymphocytes but weakly attracts neutrophils. Osteopontin, encoded by SPP1, has been implicated in a broad range of disease processes, including chronic obstructive pulmonary disease, asthma, multiple sclerosis, rheumatoid arthritis, and cardiovascular disease (34, 54). Osteopontin has both matrix protein-like and cytokine-like properties and is expressed by different cell types of the immune system (34, 54). To determine which cell types are responsible for the osteopontin expression, immunohistochemical detection of the protein was performed on lung sections of SARS-CoV-infected African green monkeys (Fig. 10). SARS-CoV-infected African green monkeys showed marked staining of osteopontin in the alveoli in areas with lesions. Strong expression of osteopontin was observed in infiltrating cells, and staining with an antibody directed against CD68 indicated that the infiltrating cells that expressed osteopontin were mainly macrophages (Fig. 9). Similar staining was performed on the lungs of young adult and aged macaques, also showing that mainly macrophages were stained for osteopontin expression (data not shown). The immunohistochemistry data, however, do not allow quantification of the amounts of osteopontin expression in the different species.