Clinical Infectious Diseases

MAJOR ARTICLE

Comparative Replication and Immune Activation Profiles of

SARS-CoV-2 and SARS-CoV in Human Lungs: An Ex Vivo
Study With Implications for the Pathogenesis of COVID-19
Hin Chu,1,a Jasper Fuk-Woo Chan,1,2,3,4,a Yixin Wang,1,a Terrence Tsz-Tai Yuen,1,a Yue Chai,1 Yuxin Hou,1 Huiping Shuai,1 Dong Yang,1 Bingjie Hu,1
Xiner Huang,1 Xi Zhang,1 Jian-Piao Cai,1 Jie Zhou,1 Shuofeng Yuan,1 Kin-Hang Kok,1 Kelvin Kai-Wang To,1,2,3 Ivy Hau-Yee Chan,5 Anna Jinxia Zhang,1
Ko-Yung Sit,5 Wing-Kuk Au,5 and Kwok-Yung Yuen1,2,3,4
1
State Key Laboratory of Emerging Infectious Diseases, Carol Yu Centre for Infection, Department of Microbiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Pokfulam, Hong
Kong Special Administrative Region, China, 2Department of Microbiology, Queen Mary Hospital, Pokfulam, Hong Kong Special Administrative Region, China, 3Department of Clinical Microbiology

Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

and Infection Control, The University of Hong Kong–Shenzhen Hospital, Shenzhen, China, 4Hainan Medical University–The University of Hong Kong Joint Laboratory of Tropical Infectious Diseases,
Hainan Medical University, Haikou, Hainan, and The University of Hong Kong, Pokfulam, Hong Kong Special Administrative Region, China, and 5Department of Surgery, The University of Hong
Kong, Queen Mary Hospital, Pokfulam, Hong Kong Special Administrative Region, China

(See the Editorial Commentary by O’Brien et al on pages 1410–2.)

Background.  Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an emerging coronavirus that has resulted in
more than 2 000 000 laboratory-confirmed cases including over 145 000 deaths. Although SARS-CoV-2 and SARS-CoV share a
number of common clinical manifestations, SARS-CoV-2 appears to be highly efficient in person-to-person transmission and fre-
quently causes asymptomatic or presymptomatic infections. However, the underlying mechanisms that confer these viral character-
istics of high transmissibility and asymptomatic infection remain incompletely understood.
Methods.  We comprehensively investigated the replication, cell tropism, and immune activation profile of SARS-CoV-2 infec-
tion in human lung tissues with SARS-CoV included as a comparison.
Results.  SARS-CoV-2 infected and replicated in human lung tissues more efficiently than SARS-CoV. Within the 48-hour in-
terval, SARS-CoV-2 generated 3.20-fold more infectious virus particles than did SARS-CoV from the infected lung tissues (P < .024).
SARS-CoV-2 and SARS-CoV were similar in cell tropism, with both targeting types I and II pneumocytes and alveolar macrophages.
Importantly, despite the more efficient virus replication, SARS-CoV-2 did not significantly induce types I, II, or III interferons in the
infected human lung tissues. In addition, while SARS-CoV infection upregulated the expression of 11 out of 13 (84.62%) represen-
tative proinflammatory cytokines/chemokines, SARS-CoV-2 infection only upregulated 5 of these 13 (38.46%) key inflammatory
mediators despite replicating more efficiently.
Conclusions.  Our study provides the first quantitative data on the comparative replication capacity and immune activation pro-
file of SARS-CoV-2 and SARS-CoV infection in human lung tissues. Our results provide important insights into the pathogenesis,
high transmissibility, and asymptomatic infection of SARS-CoV-2.
Keywords.  coronavirus; COVID-19; ex vivo; interferon; SARS-CoV-2.

The novel coronavirus disease 2019 (COVID-19) pandemic with 774 deaths [1, 2]. The genome of SARS-CoV-2 is most
caused by severe acute respiratory syndrome coronavirus 2 homologous to the Chinese horseshoe bat SARS-related cor-
(SARS-CoV-2) has affected more than 2 000 000 patients with onaviruses and SARS-CoV-2 was speculated to have been trans-
over 145 000 deaths within 4 months, whereas the 2002–2003 se- mitted from bats to a yet unknown wild animal being traded in
vere acute respiratory syndrome (SARS) caused by SARS-CoV a Wuhan wet market in China where this pandemic agent was
was controlled within 6 months and only affected 8096 patients first detected in the environmental samples [2, 3]. Apparently
more efficient person-to-person transmission than 2003 SARS
was soon noticed in the settings of families, churches, commu-
 

Received 4 April 2020; editorial decision 6 April 2020; accepted 8 April 2020; published online
nities, cruise ships, nursing homes, and hospitals [4–8]. Despite
April 9, 2020. the much lower crude fatality rate of COVID-19 (0.25–5%) than
a
Co–first authors.
that of SARS (~10%), the high number of patients affected in this
Correspondence: K.-Y. Yuen, Department of Microbiology, Li Ka Shing Faculty of Medicine,
State Key Laboratory of Emerging Infectious Diseases, Carol Yu Centre for Infection, Queen ongoing pandemic due to a complete lack of population herd im-
Mary Hospital, The University of Hong Kong, Room T19-026, Pokfulam, Hong Kong Special munity has resulted in a staggering number of deaths [9]. The
Administrative Region, China (kyyuen@hku.hk).
clinical manifestations of COVID-19 can vary from asympto-
Clinical Infectious Diseases®  2020;71(6):1400–9
© The Author(s) 2020. Published by Oxford University Press for the Infectious Diseases Society matic virus shedding among household contacts, mild upper
of America. All rights reserved. For permissions, e-mail: journals.permissions@oup.com. respiratory tract infection, and acute asymptomatic walking
DOI: 10.1093/cid/ciaa410

1400 • cid 2020:71 (15 September) • Chu et al

pneumonia, to symptomatic pneumonia with bilateral multifocal West Cluster (UW13-364). The freshly obtained lung tissues were
ground-glass opacities on lung imaging studies and severe pneu- processed into small rectangular pieces and were rinsed with
monia with acute respiratory distress syndrome and multiorgan advanced Dulbecco’s Modified Eagle’s Medium (DMEM)/F12
failure [10]. Although animal models of nonhuman primates, medium (Gibco, Thermo Fisher Scientific) supplemented with
human angiotensin-converting enzyme 2 (ACE2) transgenic 2  mM of 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
mice, and golden Syrian hamsters successfully challenged by (HEPES) (Gibco), 1× GlutaMAX (Gibco), 100 U/mL penicillin,
SARS-CoV-2 have recently been reported, studies on the patho- and 100 μg/mL streptomycin. The specimens were infected with
genesis of COVID-19 in human subjects are still largely lacking SARS-CoV-2 or SARS-CoV with an inoculum of 1 × 106 plaque-
[11–13]. While bronchoalveolar lavage was frequently performed forming units (PFU)/mL at 500 µL per well. After 2 hours, the
in patients with severe COVID-19 to make a virologic diagnosis, inoculum was removed and the specimens were washed 3 times
transbronchial or open-lung biopsy for detecting the histopath- with phosphate-buffered saline (PBS). The infected human lung
ological changes and host response in the human lung induced tissues were then cultured in 1.2  mL of advanced DMEM/F12
by SARS-CoV-2 was almost never performed to understand the medium with 2  mM HEPES (Gibco), 1× GlutaMAX (Gibco),

Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

pathogenesis in these patients with severe disease and respiratory 100 U/mL penicillin, 100  μg/mL streptomycin, 20  μg/mL van-
failure. Postmortem examinations were also rarely performed comycin, 20  μg/mL ciprofloxacin, 50  μg/mL amikacin, and
until multiorgan failure developed after weeks of intensive care 50 μg/mL nystatin. Supernatants were collected at 2, 24, and 48
by intubation and ventilation [14]. In this study, we used ex vivo hours postinoculation (hpi) for plaque assays. The lung tissues
human lung tissue challenged by SARS-CoV-2 with SARS-CoV were harvested at 2, 24, and 48 hpi in RL buffer (Qiagen) with
as a control to quantitatively characterize the viral kinetics, cell dithiothreitol (DTT) (Qiagen) for quantitative reverse transcrip-
type tropism, virus protein expression, and host cytokine/che- tion–polymerase chain reaction (qRT-PCR) analysis or fixed in
mokine responses. It is notable that the ex vivo lung tissue from 10% formalin for immunostaining and immunohistochemistry
each individual donor patient can be divided into 2 parts for a studies.
parallel challenge with SARS-CoV-2 and the control SARS-CoV.
Thus, this ex vivo lung tissue organ culture model system pro- Immunofluorescence Staining and Confocal Microscopy
vides a unique opportunity for side-by-side comparison of the Immunofluorescence staining and confocal microscopy were
virologic and host response characteristics that clinical studies on performed as previously described with slight modifications
patients or animal models cannot achieve. The findings are im- [15, 18, 19]. Briefly, infected human lung tissue samples were
portant in understanding the relatively more efficient transmissi- harvested and fixed overnight using 10% formalin, and then
bility of COVID-19 and the innate immune mechanism behind processed with a TP1020 Leica semi-enclosed benchtop tissue
this phenomenon. processor with serially increasing concentrations of ethanol,
xylene, and wax for 16 hours, before being embedded in wax.
METHODS Formalin-fixed, paraffin-embedded human lung tissues were
Viruses and Biosafety then sectioned at 4  μm with a Thermo Fisher Scientific HM
SARS-CoV-2 HKU-001a (GenBank accession number 355S rotary microtome. Tissue sections were retrieved and
MT230904​) was isolated from the nasopharyngeal aspirate of dried to fix on Thermo Fisher Scientific Superfrost Plus slides
a patient with laboratory-confirmed COVID-19 in Hong Kong, at 37oC overnight before deparaffinizing and immunofluores-
as previously described [15]. SARS-CoV GZ50 (GenBank ac- cence staining. Serially increasing concentrations of xylene and
cession number AY304495​) was an archived clinical isolate ethanol were applied for dewaxing. In order to expose the viral
at the Department of Microbiology, The University of Hong nucleocapsid (N) antigens, the slides were heated with antigen
Kong (HKU). Both SARS-CoV-2 and SARS-CoV were propa- unmasking solution (Vector Laboratories) in a pressure cooker
gated and titered in VeroE6 cells with plaque assays, as previ- for 90 seconds. After blocking with serum-free protein block
ously described [15, 16]. All experiments involving infectious and Sudan black B, the slides were stained with an in-house
SARS-CoV-2 and SARS-CoV followed the approved standard mouse anti–SARS-CoV-2-N immune serum or an in-house
operating procedures of our Biosafety Level 3 facility at the mouse anti–SARS-CoV-N immune serum. The secondary anti-
Department of Microbiology, HKU. body goat anti-mouse immunoglobulin G (IgG) heavy and light
chains (H&L) (Alexa Fluor 488)  was obtained from Thermo
Human Ex Vivo Lung Tissues Fisher Scientific. Mounting was performed with the antifade
Human lung tissues for ex vivo studies were obtained from pa- mounting medium with 4′,6-diamidino-2-phenylindole
tients undergoing surgical operations at Queen Mary Hospital, (DAPI) (Vector Laboratories). Images were acquired with con-
Hong Kong, as we previously described [17]. All donors gave focal microscopy using a Zeiss LSM710 system with the 20× ob-
written consent as approved by the Institutional Review Board jective as previously described [20]. Image fields were selected
of the University of Hong Kong/Hospital Authority Hong Kong and processed with the ZEN software blue edition (Zeiss) at a

SARS-CoV-2 Infection in an Ex Vivo Model  •  cid 2020:71 (15 September) • 1401

resolution of 1024 × 1024 pixels. To quantify the relative fluo- RNase-free water, 0.2 μL of QuantiNova SYBR Green RT-Mix,
rescence intensity, multiple images from infected human lung 1.6 μL each of 10 μM gene-specific forward and reverse primer,
tissues of 2 representative donors were split into separate chan- and 4 μL of extracted RNA as template. Reactions were incu-
nels and the intensity of channel 488 nm was quantified with bated at 45°C for 10 minutes for reverse transcription, 95°C for
ImageJ software (National Institutes of Health). 5 minutes for denaturation, followed by 45 cycles of 95°C for 5
seconds and 55°C for 30 seconds. Signal detection and meas-
Histology and Immunohistochemical Staining urements were taken in each cycle after the annealing step. The
Histology and immunohistochemical staining was performed cycling profile ended with a cooling step at 40°C for 30 seconds.
as previously described with slight modifications [13]. Briefly, The primer sequences are available upon request.
fixed human tissues were processed, embedded, and cut to pre-
pare 4-μm tissue sections on glass slides. Slides were dewaxed Statistical Analysis
with xylene and serially decreased concentrations of ethanol Statistical analyses for the quantification of viral N antigen ex-
before staining. To detect SARS-CoV-2-N and SARS-CoV-N pression from SARS-CoV-2– or SARS-CoV–infected human

Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

antigens, tissue sections underwent the same antigen retrieval lung tissues were performed with 1-way ANOVA using
procedures as those of immunofluorescence staining, followed GraphPad Prism 6 (GraphPad Software, Inc). Statistical ana-
by blocking with 0.3% hydrogen peroxide for 30 minutes to lyses for the area under the curve (AUC) comparison between
abolish the activity of any potential endogenous peroxidase. SARS-CoV-2– and SARS-CoV–generated infectious virus par-
The slides were subsequently incubated with an in-house mouse ticles from the infected human lung tissues were performed with
anti–SARS-CoV-2-N immune serum or an in-house mouse Student’s t test using GraphPad Prism 6 as previously described
anti–SARS-CoV-N immune serum, and then incubated with [15]. Statistical analyses for SARS-CoV-2– or SARS-CoV–in-
Mouse-on-Mouse Polymer (Abcam). The signal was developed duced cytokine/chemokine expressions were performed with
with the DAB (3,3′-diaminobenzidine) substrate kit (Vector 2-way ANOVA using GraphPad Prism 6. Differences were con-
Laboratories). The nuclei were developed with Gill’s hematox- sidered statistically significant when P < .05.
ylin before mounting the slides with VectaMount permanent
mounting medium (Vector Laboratories). Images were acquired RESULTS
with the Olympus BX53 light microscope using a 40× objective. Comparative Infection and Replication Capacity of SARS-CoV-2 and SARS-
CoV in Ex Vivo Human Lung Tissue Explants
Plaque Assays A total of 6 patient donors were included in this study. These in-
VeroE6 cells were maintained in DMEM with 10% FBS, 1% cluded 3 females and 3 males with a mean age of 53 years (range,
penicillin, and 1% streptomycin. The cells were seeded 1  day 47–64 years) who underwent wedge resection or lobectomy for
before the experiment at 2 × 105 cells per well in 12-well plates lung tumor. To evaluate their differential capacities to infect
and were allowed to grow at 37oC overnight. For the plaque and replicate in human lung tissue, we inoculated human lung
assay, serially diluted samples were inoculated on the cells for 1 tissue explants with SARS-CoV-2 or SARS-CoV with an inoc-
hour. Afterwards, 3% agarose/PBS was mixed with DMEM with ulum of 1 × 106 PFU/mL at 500 µL per well. The infected sam-
1.5% fetal bovine serum at a 1:2 ratio and applied to the cells. ples were harvested at 24 hpi and examined for viral N antigen
The cells were incubated for 96 hours for plaque formation. The expression. With immunostaining and confocal microscopy, we
plaques were visualized by staining the plates with 1% crystal demonstrated that both SARS-CoV-2 and SARS-CoV could in-
violet in 20% ethanol/distilled water for 15 minutes. fect human lung tissues as evidenced by viral N antigen expres-
sion in representative images of 2 donors (Figure 1). Intriguingly,
RNA Extraction and Quantitative Reverse Transcription–Polymerase Chain SARS-CoV-2-N antigens were consistently detected in higher
Reaction abundance and in broader areas of the lung tissues of all donors
Cell lysate samples from the infected lung tissues were har- than SARS-CoV-N antigens (Figure  1). Quantitative analysis
vested at 2, 24, and 48 hpi for qRT-PCR detection of mRNA confirmed that the relative fluorescent intensity of SARS-CoV-
levels of interferon (IFN) and cytokine/chemokine markers. 2-N antigen was significantly (P < .0001 to .041) higher (2.30- to
Briefly, lung tissues were homogenized and lysed with RL 2.87-fold) than SARS-CoV-N antigen in the donors’ lung tissues
buffer for RNA extraction using the RNeasy Mini kit (Qiagen). (Supplementary Figure 1). To more thoroughly quantify the dif-
Quantitative reverse transcription–polymerase chain reac- ferent infectiveness and replication of the 2 viruses in human
tion was then performed using the QuantiNova SYBR Green lung tissues, we inoculated additional lung tissue samples and
RT-PCR kit (Qiagen) with the LightCycler 480 Real-Time PCR measured the infectious virus titers generated at 2, 24, and 48 hpi
System (Roche) to quantify the expression level of different with plaque assays. Among the 4 lung tissues evaluated, the max-
host markers. Each 20-μL reaction mixture contained 10 μL of imum SARS-CoV-2 titer increased between 2 hpi and 48 hpi and
2× QuantiNova SYBR Green RT PCR Master Mix, 2.6  μL of ranged from 1.20-log to 2.04-log. With the same experimental

1402 • cid 2020:71 (15 September) • Chu et al

 Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

Figure 1.  Confocal microscopy examination of SARS-CoV-2 and SARS-CoV replication in human lung tissues. Human lung tissues were challenged with SARS-CoV-2 or
SARS-CoV with an inoculum of 1 × 106 PFU/mL. At 24 hpi, the infected lung tissues were harvested, fixed in 10% formalin, and immunolabeled for viral antigen. SARS-CoV-
2-nucleocapsid (N) antigens were identified with an in-house mouse anti–SARS-CoV-2-N immune serum, and SARS-CoV-N antigens were identified with an in-house mouse
anti–SARS-CoV-N immune serum. Representative images from 2 lung donors are shown. The experiment was repeated with another 4 donors with consistent findings. Bars
represent 50 μm. Dotted square, areas to be enlarged and shown in the insets. Abbreviations: hpi, hours postinoculation; PFU, plaque-forming units; SARS-CoV, severe acute
respiratory syndrome coronavirus; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

SARS-CoV-2 Infection in an Ex Vivo Model  •  cid 2020:71 (15 September) • 1403

setting, the maximum SARS-CoV titer increase between 2 hpi generated 3.20-fold more infectious virus particles than did
and 48 hpi ranged from 0.61-log to 1.15-log only (Figure 2A). To SARS-CoV from the infected lung tissues over a period of 48
determine the relative replication capacity of the 2 viruses in lung hours (P < .024) (Figure 2C). Taken together, our results illustrate
tissues, we set the virus titer at 2 hpi as the baseline, which repre- that SARS-CoV-2 was capable of infecting and replicating more
sented the residual inoculum after inoculation and wash, and cal- robustly than SARS-CoV in human lung tissues.
culated the virus titer increase from the baseline at 24 and 48 hpi.
As shown in Figure 2B, SARS-CoV-2 consistently demonstrated Cell Tropism of SARS-CoV-2 and SARS-CoV in Human Lung Tissues
higher virus titer increases than did SARS-CoV in lung tissues To identify the cell types targeted by SARS-CoV-2 and SARS-CoV
from all 4 donors. Our AUC analysis revealed that SARS-CoV-2 in the human lung, we performed immunohistochemical staining

Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

Figure 2.  Comparative replication capacity of SARS-CoV-2 and SARS-CoV in human lung tissues. Human lung tissues from 4 independent donors were challenged with
SARS-CoV-2 or SARS-CoV. Supernatant samples from the infected lung tissues were harvested at 2, 24, and 48 hpi. The virus titers were determined with plaque assays. A,
The virus titers of the 4 individual lung tissues. B, The amount of infectious virus particles increased over the 48-hour period upon SARS-CoV-2 or SARS-CoV inoculation for
each lung donor. C, AUC analysis of SARS-CoV-2– and SARS-CoV–infected human lung tissues, demonstrating the total amount of live infectious virus particles generated
over the 48-hour period. Results in panel C represent the means and SDs from 4 independent donors from 4 independent experiments. Statistical significance in panel C
was determined with Student t test. *P < .05. Abbreviations: AUC, area under the curve; hpi, hours postinfection; SARS-CoV, severe acute respiratory syndrome coronavirus;
SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

1404 • cid 2020:71 (15 September) • Chu et al

 Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020
Figure 3.  Cell tropism of SARS-CoV-2 and SARS-CoV in human lung tissues. Human lung tissues were challenged with SARS-CoV-2 or SARS-CoV. At 24 hpi, the infected
lung tissues were harvested, fixed in 10% formalin, and processed for immunohistochemical examination of viral antigen. A–H, SARS-CoV-2-nucleocapsid (N) antigens
could be identified from type I pneumocytes (B, C), type II pneumocytes (C, F, H), and alveolar macrophages (D). I–L, SARS-CoV-N antigens could be identified from type
I pneumocytes and type II pneumocytes (J), as well as alveolar macrophages (L). M and N, Mock-infected human lung tissues processed and immunolabeled with the same
set of antibodies did not show any specific signal. SARS-CoV-2-N antigens were identified with an in-house mouse anti–SARS-CoV-2-N immune serum, and SARS-CoV-N
antigens was identified with an in-house mouse anti–SARS-CoV-N immune serum. Blue arrows represent type I pneumocytes, orange arrows represent type II pneumocytes,
and green arrows represent alveolar macrophages. The experiment was repeated with 3 donors with consistent findings. Bars represent 50 μm. Dotted square, areas to be
enlarged and shown in the insets. Abbreviations: hpi, hours postinoculation; SARS-CoV, severe acute respiratory syndrome coronavirus; SARS-CoV-2, severe acute respiratory
syndrome coronavirus 2.

of the virus-infected lung tissues (Figure 3). Our results clearly Expression Profiles of Interferons and Proinflammatory Mediators in
SARS-CoV-2– and SARS-CoV–Infected Human Lung Tissues
demonstrated that SARS-CoV-2 could infect type I pneumocytes
(Figure  3A–C), type II pneumocytes (Figure  3E–H), as well as Next, to assess the innate immune responses induced by SARS-CoV-2
alveolar macrophages (Figure 3D). In this regard, the cell tropism and SARS-CoV in human lung tissues, we investigated the expres-
of SARS-CoV-2 in the human lung was similar to that of SARS- sion of a panel of representative IFNs and proinflammatory cyto-
CoV, which also targeted types I and II pneumocytes (Figure 3I, J) kines/chemokines upon virus challenge. Our data demonstrated that
and alveolar macrophages (Figure 3K, L). Viral N antigen signal the 2003 SARS-CoV infection resulted in significant upregulation of
was not detected from mock-infected samples (Figure 3M, N). types I (IFNβ), II (IFNγ), and III (IFNλ1, IFNλ2, and IFNλ3) IFNs

SARS-CoV-2 Infection in an Ex Vivo Model  •  cid 2020:71 (15 September) • 1405

 Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020
Figure 4.  Interferon response of SARS-CoV-2– and SARS-CoV–infected human lung tissues. Human lung tissues were challenged with SARS-CoV-2 or SARS-CoV. At 2,
24, and 48 hpi, the infected lung tissues were harvested for qRT-PCR analysis of type I (IFNα and IFNβ), type II (IFNγ), and type III (IFNλ1, IFNλ2, and IFNλ3) interferons. The
results represent values from 4 independent lung donors in 4 independent experiments, with 3–5 lung samples per donor. Statistical significance was determined with 2-way
ANOVA. *P < .05, **P < .01, ***P < .001. Abbreviations: ANOVA, analysis of variance; hpi, hours postinfection; qRT-PCR, quantitative reverse transcription–polymerase chain
reaction; SARS-CoV, severe acute respiratory syndrome coronavirus; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

in human lung tissues. In contrast, SARS-CoV-2 infection did not type I and type II pneumocytes, and alveolar macrophages as target
significantly trigger the expression of any IFN at all evaluated time cells, the amount of viral N antigen expression in the virus-challenged
points (Figure 4). In addition to types I, II, and III IFNs, we evalu- lung tissues was significantly higher and more intensive in the tissues
ated the expression of 4 key proinflammatory cytokines (Figure 5A) infected with SARS-CoV-2 than in those infected with SARS-CoV.
and 9 key proinflammatory chemokines (Figure 5B) that have critical Moreover, SARS-CoV-2 produced 3.20-fold higher amounts of infec-
roles in immune cell recruitment and activation. Our data show that tious virus particles than did SARS-CoV within 48 hpi. These find-
SARS-CoV infection resulted in significant activation of 11 out of 13 ings might explain the high viral load in the respiratory secretions of
(84.62%) evaluated proinflammatory factors at either 24 hpi or 48 hpi. patients with COVID-19 during the early days after presentation or
In contrast, SARS-CoV-2 infection only significantly upregulated 5 even during incubation, and thus its high person-to-person trans-
of these 13 (38.46%) inflammatory mediators including interleukin-6 missibility [4, 21]. Importantly, we demonstrated that SARS-CoV-2
(IL6), monocyte chemoattractant protein-1 (MCP-1), chemokine triggered lower levels of IFNs and proinflammatory cytokines/
(C-X-C motif) ligand 1 (CXCL1), chemokine (C-X-C motif) ligand 5 chemokines despite being capable of infecting and producing signifi-
(CXCL5), and chemokine (C-X-C motif) ligand 10 (CXCL10/IP10). cantly higher amount of virus in human lung tissues. While the innate
In addition, among all assessed IFNs and cytokines/chemokines, the immune response in infected cells is our first line of defense against
expression of 12 out of 19 (63.16%) genes were significantly lower acute viral infection, SARS-CoV-2 infection did not significantly
in SARS-CoV-2–infected human lung tissues in comparison to that trigger any IFN response at all, and only significantly activated 5 of
of the SARS-CoV–infected samples (Figures  4 and 5). Only IP10 13 proinflammatory mediators. Thus, unlike patients with SARS who
was significantly more induced by SARS-CoV-2 than SARS-CoV. generally presented with high fever followed by rapidly progressive
Collectively, our results illustrate that SARS-CoV-2 generally trig- pneumonia and then respiratory failure with a mortality rate of ap-
gered significantly lower levels of IFNs and proinflammatory cyto- proximately 10%, only approximately 16% of patients with COVID-
kines/chemokines despite being capable of infecting and replicating 19 are severely ill and less than 5% of them died [2, 10]. The low degree
in the human lung at a significantly higher efficiency. of innate immune activation could also account for the mild or even
lack of symptoms in many patients with COVID-19 who were not
DISCUSSION even tested and unknowingly spreading the virus in both community
Using ex vivo human lung tissue explants respectively challenged and hospital settings, making the control of the pandemic much more
by the same inoculum of SARS-CoV-2 and SARS-CoV, we showed difficult than SARS.
that SARS-CoV-2 was more capable than SARS-CoV of infecting While both 2019 SARS-CoV-2 and 2003 SARS-CoV can at-
and replicating in human lung tissues. While both viruses infected tach and enter host cells via ACE2, the mechanism of how

1406 • cid 2020:71 (15 September) • Chu et al

 Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020
Figure 5.  Expression profile of proinflammatory cytokines and chemokines from SARS-CoV-2– and SARS-CoV–infected human lung tissues. Human lung tissues were
challenged with SARS-CoV-2 or SARS-CoV. At 2, 24, and 48 hpi, the infected lung tissues were harvested for qRT-PCR analysis of representative proinflammatory cytokines
(A) and chemokines (B). The results represent values from 4 independent lung donors in 4 independent experiments, with 3–5 lung samples per donor. Statistical significance
was determined with 2-way ANOVA. *P < .05, **P < .01, ***P < .001, ****P < .0001. Abbreviations: ANOVA, analysis of variance; CXCL1, chemokine (C-X-C motif) ligand 1;
CXCL10, chemokine (C-X-C motif) ligand 10; CXCL2, chemokine (C-X-C motif) ligand 2; CXCL5, chemokine (C-X-C motif) ligand 5; CXCL9, chemokine (C-X-C motif) ligand 9; hpi,
hours postinoculation; IL, interleukin; IL12, interleukin 12; IL1β, interleukin 1 beta; IL6, interleukin 6; IL8, interleukin 8; MCP1, monocyte chemoattractant protein 1; MIP1α,
macrophage inflammatory protein 1 alpha; qRT-PCR, quantitative reverse transcription–polymerase chain reaction; RANTES, regulated on activation, normal T cell expressed
and secreted; SARS-CoV, severe acute respiratory syndrome coronavirus; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; TNFα, tumor necrosis factor α.

SARS-CoV-2 overcomes the innate immune response and sup- response in a dose-dependent manner [22]. Moreover, 2 other
presses the IFN and proinflammatory cytokines and chemokines SARS-CoV proteins, namely open reading frame (Orf) 3b and
to achieve a higher degree of viral replication is still elusive. Orf6, also antagonize IFN synthesis and signaling [23]. While
Although the N protein of both of these coronaviruses share 94% SARS-CoV-2 is also postulated to possess IFN-antagonizing pro-
of amino acid similarity [3], significantly higher N expression teins, the exact identities of these proteins remain undetermined
in terms of both the extent and intensity was readily observed at this stage. The differential degrees of IFN inhibition by these
in SARS-CoV-2–infected ex vivo lung tissues. Studies of SARS- viruses’ IFN-antagonizing proteins may explain the discrepant
CoV in cell culture suggested that its N protein antagonized IFNβ viral replications and inflammatory responses in human lungs

SARS-CoV-2 Infection in an Ex Vivo Model  •  cid 2020:71 (15 September) • 1407

observed in the present study. Moreover, most proinflammatory adaptive immune response. Second, the supply of human lung
cytokines and chemokines except for IP10, a chemoattractant tissues is limited and thus it was not possible for us to investi-
for monocytes/macrophages, T cells, and natural killer (NK) gate the characteristics of different SARS-CoV-2 strains in this
cells, were more significantly induced by SARS-CoV than SARS- ex vivo model. More studies should be performed using both
CoV-2. This finding may explain the more severe disease and ex vivo lung and intestinal organoids to elucidate additional
higher mortality of SARS than of COVID-19. details on the pathogenesis of SARS-CoV-2 during pulmonary
In contrast to patients with SARS who had increasing viral and extrapulmonary involvement.
load in nasopharyngeal secretions that peaked at approximately
day 10, the viral load in the respiratory secretions of patients Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online.
with COVID-19 peaked much earlier at the time of symptom
Consisting of data provided by the authors to benefit the reader, the posted
onset [21, 24]. The viral kinetics and innate immune response materials are not copyedited and are the sole responsibility of the authors, so
profiles in ex vivo human lung tissue characterized in this study questions or comments should be addressed to the corresponding author.
might help to  explain this observation. The suboptimally ac-
Notes

Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

tivated innate immune response would allow SARS-CoV-2 to
Author contributions. H.  C., J.  F.-W. C., and K.-Y. Y.  had roles in the
replicate to high levels in the respiratory tract early on and may study design, data collection, data analysis, data interpretation, and writing
contribute to its efficient person-to-person transmission via of the manuscript. Y. W., T. T.-T. Y, Y. C., Y. H., H. S., D. Y., B. H., X. H.,
droplets or contact with contaminated respiratory secretions X. Z., J.-P. C., J. Z., S. Y., K.-H. K., K. K.-W. T., I. H.-Y. C., A. J. Z., K.-Y. S.,
and W.-K. A. had roles in the experiments, data collection, data analysis,
containing high viral loads [4, 21, 25]. Notably, the higher de- and data interpretation. All authors reviewed and approved the final version
gree of viral infection and replication of SARS-CoV-2 in human of the manuscript.
lung tissues corroborated our recent finding that SARS-CoV-2 Disclaimer. The funding sources had no role in the study design, data
collection, analysis, interpretation, or writing of the report
exhibited more efficient replication in Calu3 (human lung ade-
Financial support. This study was partly supported by the donations of May
nocarcinoma) cells than SARS-CoV [15]. Tam Mak Mei Yin, Richard Yu, and Carol Yu, the Shaw Foundation Hong Kong;
Our findings have important implications for the infection- Michael Seak-Kan Tong, Respiratory Viral Research Foundation Limited; Hui
control and treatment strategies for COVID-19 as patients Ming, Hui Hoy and Chow Sin Lan Charity Fund Limited; Chan Yin Chuen
Memorial Charitable Foundation; Marina Man-Wai Lee, the Hong Kong Hainan
with SARS-CoV-2 may be infectious during the incubation Commercial Association South China Microbiology Research Fund; the Jessie &
period before onset of clinical symptoms due to its efficient George Ho Charitable Foundation, and Perfect Shape Medical Limited and received
modulation of host innate immune and proinflammatory re- funding from the Consultancy Service for Enhancing Laboratory Surveillance of
Emerging Infectious Diseases and Research Capability on Antimicrobial Resistance
sponse, as evidenced by our findings [26]. Thus, in addition for Department of Health of the Hong Kong Special Administrative Region
to hand washing and social distancing, universal masking was Government; the Theme-Based Research Scheme (grant number T11/707/15) of
recommended in East Asia as a key epidemiological control the Research Grants Council; Hong Kong Special Administrative Region; Sanming
Project of Medicine in Shenzhen, China (grant number SZSM201911014); and the
measure to prevent virus shedding from subclinical patient
High Level-Hospital Program, Health Commission of Guangdong Province, China.
sources and to prevent infection of susceptible individuals in Potential conflicts of interest. The authors: No reported conflicts of
the community [27]. Similar to influenza, which is the most interest. All authors have submitted the ICMJE Form for Disclosure of
well-studied acute respiratory virus infection, the viral load Potential Conflicts of Interest.

in the respiratory secretions of patients with COVID-19 also References

peaks at the time of symptom onset, and antiviral treatment 1. Cheng  VC, Lau  SK, Woo  PC, Yuen  KY. Severe acute respiratory syndrome
for COVID-19 may be less effective if given later than 48 hours Coronavirus as an agent of emerging and reemerging infection. Clin Microbiol
Rev 2007; 20:660–94.
after symptom onset. Thus, unlike SARS and Middle East res- 2. World Health Organization. Coronavirus disease (COVID-19) situation re-
piratory syndrome which have viral loads peaking at day 7 to port—74. 3 April 2020. Available at: https://www.who.int/docs/default-
source/coronaviruse/situation-reports/20200403-sitrep-74-covid-19-mp.
day 10 and therefore sufficient time for antivirals to act and pdf?sfvrsn=4e043d03_4. Accessed 4 April 2020.
reduce the peak viral loads, early initiation of antiviral therapy 3. Chan  JF, Kok  KH, Zhu  Z, et  al. Genomic characterization of the 2019 novel
human-pathogenic coronavirus isolated from a patient with atypical pneumonia
would be even more important to improve the clinical out- after visiting Wuhan. Emerg Microbes Infect 2020; 9:221–36.
come of COVID-19 [24, 28]. Moreover, the use of high-dose 4. Chan JF, Yuan S, Kok KH, et al. A familial cluster of pneumonia associated with
the 2019 novel coronavirus indicating person-to-person transmission: a study of
corticosteroids and antagonists against major inflammatory
a family cluster. Lancet 2020; 395:514–523. doi:10.1016/S0140-6736(20)30154-9
mediators such as IL6 should only be used together with ef- 5. Cheng VCC, Wong SC, Chen JHK, et al. Escalating infection control response
fective antivirals to avoid oversuppression of the innate im- to the rapidly evolving epidemiology of the coronavirus disease 2019 (COVID-
19) due to SARS-CoV-2 in Hong Kong. Infect Control Hosp Epidemiol 2020;
mune response in patients with COVID-19 as SARS-CoV-2 1–6. Epub ahead of print. doi:10.1017/ice.2020.58
is already modulating the host innate immune response at the 6. Rocklov  J, Sjodin  H, Wilder-Smith  A. COVID-19 outbreak on the Diamond
Princess cruise ship: estimating the epidemic potential and effectiveness of public
beginning of the infection [29]. health countermeasures. J Travel Med 2020; pii:taaa030. Epub ahead of print. doi:
There are limitations to this study. First, the ex vivo human 10.1093/jtm/taaa030
7. McMichael TM, Currie DW, Clark S, et al. Epidemiology of Covid-19 in a long-
lung tissue explant culture is short-lasting and cannot repre- term care facility in King County, Washington. N Engl J Med 2020. Epub ahead of
sent the effect of host systemic inflammatory response and the print. doi:10.1056/NEJMoa2005412

1408 • cid 2020:71 (15 September) • Chu et al

 8. Pung  R, Chiew  CJ, Young  BE, et  al; Singapore 2019 Novel Coronavirus 18. Chu H, Chan CM, Zhang X, et al. Middle East respiratory syndrome coronavirus
Outbreak Research Team. Investigation of three clusters of COVID-19 in and bat coronavirus HKU9 both can utilize GRP78 for attachment onto host cells.
Singapore: implications for surveillance and response measures. Lancet J Biol Chem 2018; 293:11709–26.
2020; 395:1039–46. 19. Chan  CM, Chu  H, Wang  Y, et  al. Carcinoembryonic antigen-related cell adhe-
9. Wilson  N, Kvalsvig  A, Barnard  LT, Baker  MG. Case-fatality risk estimates for sion molecule 5 is an important surface attachment factor that facilitates entry of
COVID-19 calculated by using a lag time for fatality. Emerg Infect Dis 2020; 26. Middle East respiratory syndrome coronavirus. J Virol 2016; 90:9114–27.
Epub ahead of print. doi:10.3201/eid2606.200320 20. Chu H, Zhou J, Wong BH, et al. Middle East respiratory syndrome coronavirus
10. Guan WJ, Ni ZY, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 efficiently infects human primary T lymphocytes and activates the extrinsic and
in China. N Engl J Med 2020. Epub ahead of print. doi:10.1056/NEJMoa2002032 intrinsic apoptosis pathways. J Infect Dis 2016; 213:904–14.
11. Bao L, Deng W, Gao H, et al. Reinfection could not occur in SARS-CoV-2 infected 21. To KK, Tsang OT, Leung WS, et al. Temporal profiles of viral load in posterior oro-
rhesus macaques. 2020. Available at: https://www.biorxiv.org/content/10.1101/20 pharyngeal saliva samples and serum antibody responses during infection by SARS-
20.03.13.990226v1. Accessed 3 April 2020. CoV-2: an observational cohort study. Lancet Infect Dis 2020.
12. Bao  L, Deng  W, Huang  B, et  al. The pathogenicity of SARS-CoV-2 in hACE2 22. Hu  Y, Li  W, Gao  T, et  al. The severe acute respiratory syndrome coronavirus
transgenic mice. 2020. Available at: https://www.biorxiv.org/content/10.1101/20 nucleocapsid inhibits type I interferon production by interfering with TRIM25-
20.02.07.939389v3. Accessed 3 April 2020. mediated RIG-I ubiquitination. J Virol 2017; 91.
13. Chan JF, Zhang AJ, Yuan S, et al. Simulation of the clinical and pathological mani- 23. Kopecky-Bromberg  SA, Martínez-Sobrido  L, Frieman  M, Baric  RA, Palese  P.
festations of coronavirus disease 2019 (COVID-19) in golden Syrian hamster model: Severe acute respiratory syndrome coronavirus open reading frame (ORF) 3b,
implications for disease pathogenesis and transmissibility. Clin Infect Dis 2020. ORF 6, and nucleocapsid proteins function as interferon antagonists. J Virol 2007;
pii:ciaa325. Epub ahead of print. doi:10.1093/cid/ciaa325 81:548–57.

Downloaded from https://academic.oup.com/cid/article/71/6/1400/5818134 by guest on 10 November 2020

14. Tian S, Hu W, Niu L, Liu H, Xu H, Xiao SY. Pulmonary pathology of early-phase 24. Peiris JS, Chu CM, Cheng VC, et al; HKU/UCH SARS Study Group. Clinical pro-
2019 novel coronavirus (COVID-19) pneumonia in two patients with lung gression and viral load in a community outbreak of coronavirus-associated SARS
cancer. J Thorac Oncol 2020. pii:S1556-0864(20)30132-5. Epub ahead of print. pneumonia: a prospective study. Lancet 2003; 361:1767–72.
doi:10.1016/j.jtho.2020.02.010 25. To KK, Tsang OT, Chik-Yan Yip C, et al. Consistent detection of 2019 novel coro-
15. Chu H, Chan JF, Yuen TT, et al. Comparative tropism, replication kinetics, and navirus in saliva. Clin Infect Dis 2020.
cell damage profiling of SARS-CoV-2 and SARS-CoV with implications for clin- 26. Wei WE, Li Z, Chiew CJ, et al. Presymptomatic transmission of SARS-CoV-2—
ical manifestations, transmissibility, and laboratory studies of COVID-19: an ob- Singapore, January 23–March 16, 2020. MMWR Morb Mortal Wkly Rep 2020
servational study. Lancet Microbe 2020. doi:10.1016/S2666-5247(20)30004-5 [Epub ahead of print] doi: 10.15585/mmwr.mm6914e1.
16. Chan JF, Yip CC, To KK, et al. Improved molecular diagnosis of COVID-19 by 27. Chan  KH, Yuen  KY. COVID-19 epidemic: disentangling the re-emerging con-
the novel, highly sensitive and specific COVID-19-RdRp/Hel real-time reverse troversy about medical facemasks from an epidemiological perspective. Int J
transcription-polymerase chain reaction assay validated in vitro and with clinical Epidemiol 2020; pii:dyaa044.
specimens. J Clin Microbiol 2020. pii:JCM.00310-20. Epub ahead of print. doi: 28. Chan JF, Lau SK, To KK, Cheng VC, Woo PC, Yuen KY. Middle East respiratory
10.1128/JCM.00310-20 syndrome coronavirus: another zoonotic betacoronavirus causing SARS-like di-
17. Zhou J, Chu H, Li C, et al. Active replication of Middle East respiratory syndrome sease. Clin Microbiol Rev 2015; 28:465–522.
coronavirus and aberrant induction of inflammatory cytokines and chemokines 29. Xu  X, Han  M, Li  T, et  al. Effective treatment of severe COVID-19 patients
in human macrophages: implications for pathogenesis. J Infect Dis 2014; with Tocilizumab. Available at: http://chinaxiv.org/home.htm ChinaXiv 2020.
209:1331–42. Accessed 3 April 2020.

SARS-CoV-2 Infection in an Ex Vivo Model  •  cid 2020:71 (15 September) • 1409