key: cord-0958038-mkk3530e
authors: Yang, J.-K.; Zhao, M.-M.; Yang, W.-L.; Yang, F.-Y.; Zhang, L.; Huang, W.; Fan, C.; Hou, W.; Jin, R.; Feng, Y.; Wang, Y.
title: Cathepsin L plays a key role in SARS-CoV-2 infection in humans and humanized mice and is a promising target for new drug development
date: 2020-10-27
journal: nan
DOI: 10.1101/2020.10.25.20218990
sha: faeb2e66738776ac11227fbd00f7b00bb066838c
doc_id: 958038
cord_uid: mkk3530e

To discover new drugs to combat COVID-19, an understanding of the molecular basis of SARS-CoV-2 infection is urgently needed. Here, for the first time, we report the crucial role of cathepsin L (CTSL) in patients with COVID-19. The circulating level of CTSL was elevated after SARS-CoV-2 infection and was positively correlated with disease course and severity. Correspondingly, SARS-CoV-2 pseudovirus infection increased CTSL expression in human cells in vitro and human ACE2 transgenic mice in vivo, while CTSL overexpression, in turn, enhanced pseudovirus infection in human cells. CTSL functionally cleaved the SARS-CoV-2 spike protein and enhanced virus entry, as evidenced by CTSL overexpression and knockdown in vitro and application of CTSL inhibitor drugs in vivo. Furthermore, amantadine, a licensed anti-influenza drug, significantly inhibited CTSL activity after SARS-CoV-2 pseudovirus infection and prevented infection both in vitro and in vivo. Therefore, CTSL is a promising target for new anti-COVID-19 drug development.

1 lower in patients with severe disease than in patients with nonsevere disease, while plasma levels 1 of ACE2 were the same between the two groups (Table S3 ). Interestingly, the plasma level of 2 CTSL was markedly higher, while that of CTSB was slightly lower, in patients with severe 3 disease than in patients with nonsevere disease ( Figure 1A and Table S2 ). 4 To further confirm the correlation between changes in cathepsin levels and COVID-19, plasma 5 CTSL and CTSB levels were measured in the 125 healthy volunteers, and the reference ranges 6 for each parameter were established as the mean values ± 2 SD in the healthy participants, 7 indicated by the green boxes in the figures ( Figure 1A and 1B). The CTSL level was markedly 8 higher in patients with COVID-19 than in healthy volunteers, while the CTSB level was 9 unchanged in the patients ( Figure 1B) .

The circulating level of CTSL changes with the course of COVID-19 11 Notably, a strong correlation was found between the CTSL level and the number of days from 12 symptom onset to blood collection before therapy ( Figure 1C) . We further performed a follow-13 up study to determine the correlation between the CTSL level and COVID-19. Patients with 14 COVID-19 experienced a mean hospitalization time of 14 days (Day 14) and were followed up 15 on the 14 th day (Day 28) and 28 th day (Day 42) after hospital discharge ( Figure 1D ). In the follow-16 up study, the elevated levels of CTSL were dramatically decreased on Day 28 and remained at a 17 stable level on Day 42 ( Figure 1E ). However, the difference between the severe and nonsevere 18 groups persisted for 42 days after admission to the hospital, although in most patients, the CTSL 19 level returned to the normal range after discharge ( Figure 1F ). These results indicated that CTSL 20 was significantly associated with SARS-CoV-2 infection.

CTSL is an independent factor for severity in patients with COVID-19 22 The correlations between disease severity and clinical parameters, including CTSL, CTSB, 23 ACE2, Ang(1-7), age, sex, and coexistence of diabetes or hypertension, were estimated using 24 Spearman's rho correlation coefficient. Severity was positively correlated with CTSL and age but 25 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 was negatively correlated with CTSB. In addition to being correlated with severity, CTSL was 1 also positively correlated with age and history of hypertension but was negatively correlated with 2 CTSB (Table S4 ). In univariable logistic regression analysis, the odds ratio (OR) of experiencing 3 critical condition was significantly higher in patients with a higher CTSL level (OR, 1.53 per 4 ng/ml; 95% CI, 1.19-1.96; P = 0.001) and older age (OR, 1.94 per 10 years; 95% CI, 1.31-2.88; 5 P = 0.001), while CTSB, ACE2, Ang(1-7), sex, and coexistence of diabetes or hypertension did 6 not significantly contribute to the odds of experiencing critical condition ( Figure 1G ). 7 Multivariable logistic regression indicated that CTSL was an independent factor for severe 8 disease status after adjustment for hypertension, diabetes, sex, age, Ang(1-7), ACE2 and CTSB 9 ( Figure 1H ). Taken together, these findings led us to conclude that CTSL was highly correlated 10 with SARS-CoV-2 infection and associated with the severity of the disease.

The CTSL level is elevated in SARS-CoV-2 pseudovirus-infected cells in vitro 12 As the CTSL level is significantly elevated in the plasma of patients with COVID-19, we 13 speculated that CTSL may be an important biomarker and therapeutic target for COVID-19. To 14 test this hypothesis, we conducted a series of in vitro and in vivo experiments using a SARS- 15 CoV-2 pseudovirus system. This pseudovirus is composed of replication-defective vesicular 16 stomatitis virus (VSV) particles bearing SARS-CoV-2 S proteins (SARS-2-S), faithfully reflects 17 key aspects of SARS-CoV-2 cell entry (Hoffmann et al., 2020b) as we previously reported (Nie 18 et al., 2020) and can be used safely in biosafety level 2 (BSL-2) laboratories ( Figure S1 ). 19 First, to determine the cell line with the highest susceptibility to SARS-2-S-driven entry, we is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 AST in COVID-19 patients compared with healthy volunteers observed in our study (Table S1 1 and S2) and the hepatocellular injury observed in patients with COVID-19 by others suggested 2 that SARS-CoV-2 attacks hepatocytes as target cells (Gupta et al., 2020) . Therefore, Huh7 cells 3 were selected for the subsequent experiments in this study.

Next, to verify whether CTSL expression increases after SARS-CoV-2 infection in vitro, the 5 protein and mRNA levels of CTSL and CTSB were measured in Huh7 cells infected with SARS-6 CoV-2 pseudovirus ( Figure 2A ). Huh7 cells were infected with different doses of SARS-CoV-2 7 pseudovirus, as indicated by the luciferase activities ( Figure 2B ) and VSV phosphoprotein (VSV-8 P) mRNA levels ( Figure 2C ). Consistent with our clinical data, the mRNA ( Figure 2D ) and 9 protein ( Figure 2E ) levels of CTSL increased in a dose-dependent manner after SARS-CoV-2 10 pseudovirus infection. These results confirmed our findings in patients with indicated that SARS-CoV-2 infection caused CTSL upregulation.

To investigate whether CTSL is required for cell entry of SARS-CoV-2, we used siRNAs against 14 human CTSL (si-CTSL) and plasmids encoding human CTSL (pCTSL) to knockdown and 15 overexpress the CTSL gene in Huh7 cells, respectively, as shown in Figure 3A . si-CTSL treatment 16 dose-dependently downregulated CTSL without affecting CTSB expression ( Figure 3B ).

Knockdown of CTSL led to a significant dose-dependent reduction in pseudovirus cell entry, as 18 evidenced by the luciferase activity and VSV-P mRNA level ( Figure 3C -3E). In contrast, 19 overexpression of CTSL markedly increased pseudovirus cell entry in a dose-dependent manner 20 without affecting CTSB expression ( Figure 3F -3I). These results suggested that CTSL but not 21 CTSB was critical for SARS-CoV-2 infection.

CTSL cleaves the S protein, and this cleavage promotes cell-cell fusion 23 CTSL cleaves the SARS-CoV-1 S protein into S1 and S2 subunits and proteolytically activates 24 cell-cell fusion (Simmons et al., 2005) . Unlike the SARS-CoV-1 S protein, the SARS-CoV-2 S 25 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 protein is precleaved by the proprotein convertase furin at the S1/S2 cleavage site ( Figure 4A ).

Hence, the effects of CTSL in SARS-CoV-2 seem to be replaced by those of furin. However, no 2 report has examined the effects of CTSL on SARS-CoV-2 S protein cleavage and the function of 3 CTSL in cell-cell fusion. Here, we directly detected the cleavage of purified SARS-CoV-2 S 4 protein by CTSL. Treatment with CTSL resulted in successful cleavage of purified SARS-CoV-5 1 S protein, suggesting that the experimental system was feasible. Notably, CTSL also efficiently 6 cleaved purified SARS-CoV-2 S protein in a dose-dependent manner ( Figure 4B ). To further 7 confirm the specificity, CTSL inhibitors were employed. Given that no currently available drug 8 specifically inhibits CTSL (Dana and Pathak, 2020) , two compounds that have been 9 demonstrated to have inhibitory activity against CTSL (E64d, a broad-spectrum cathepsin 10 inhibitor, and SID 26681509, a relatively selective CTSL inhibitor) were used. The cleavage 11 activity of CTSL was blocked by E64d and SID 26681509 ( Figure 4B ). These results indicated 12 that CTSL efficiently cleaved the SARS-CoV-2 S protein into smaller fragments after its initial 13 cleavage by furin.

14 To investigate whether CTSL functionally cleaves the SARS-CoV-2 S protein, we performed 15 a cell-cell fusion assay by recording SARS-2 S protein-driven formation of multinucleated giant 16 cells (syncytium). No syncytia were observed without S protein expression, while SARS-2-S 17 expression alone (treated with PBS, pH=7.4) resulted in syncytium formation. Trypsin treatment

(2 µg/ml in PBS, pH=7.4), which has been shown to induce S-protein-mediated cell-cell fusion 19 (Bosch et al., 2008; Ou et al., 2020) , markedly increased syncytium formation nearly two-fold.

In addition, compared to mock treatment (PBS, pH=5.8), CTSL (in PBS, pH=5.8) dose-21 dependently induced an increase in syncytium formation of up to ~70%. These data indicated 22 that CTSL activity, not acidic conditions, was responsible for the increase in syncytium formation 23 ( Figure 4C and 4D). Therefore, these results led us to conclude that CTSL efficiently cleaved the 24 SARS-CoV-2 S protein and that this cleavage promoted S-protein-mediated cell-cell fusion.

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 1 To further confirm the role of CTSL in SARS-CoV-2 infection, Huh7 cells were treated with 2 CTSL inhibitors, as shown in Figure 5A . Both SID 26681509 and E64d significantly inhibited 3 SARS-CoV-2 pseudovirus infection. As E64d exhibited less cytotoxicity than SID 26681509, it 4 was selected for the subsequent experiments ( Figure 5B and 5C).

Furthermore, we were interested to find that amantadine, a prophylactic agent approved by the 6 US FDA in 1968 for influenza and later for Parkinson's disease, has been reported to suppress 7 the gene transcription of CTSL (Smieszek et al., 2020) . We next examined the impact of 8 amantadine on SARS-CoV-2 infection and found that it significantly inhibited pseudovirus 9 infection with little cytotoxicity ( Figure 5D ). Moreover, both E64d and amantadine also 10 significantly prevented SARS-CoV-1 S protein-driven but not VSV G protein-driven or Rift is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 technology, as we previously reported (Sun et al., 2020) -were employed. The hACE2 1 humanized mice were randomly divided into four groups and treated with either vehicle or 2 different drugs as indicated. Bioluminescence was measured and visualized in pseudocolor as an 3 indicator of SARS-CoV-2 pseudovirus infection severity. Pseudovirus-infected humanized mice 4 showed a significantly higher luminescence signal than healthy control mice, indicating that the 5 mice were successfully infected ( Figure 6A and 6B). Compared to the vehicle treatment, E64d 6 significantly prevented SARS-CoV-2 pseudovirus infection. Amantadine also showed 7 suppressive effects on pseudovirus infection, but the differences were not statistically significant 8 (P = 0.058) ( Figure 6B ). The hepatic VSV-P mRNA level was markedly increased in humanized 9 mice after SARS-CoV-2 pseudovirus infection, but this increase was significantly suppressed by 10 pretreatment with either E64d or amantadine ( Figure 6C ), indicating that both drugs indeed 11 prevented SARS-CoV-2 pseudovirus infection.

Notably, the protein level of CTSL in the liver was significantly increased in SARS-CoV-2-13 infected mice; this increase was reversed by treatment with either E64d or amantadine ( Figure   14 6D), while the CTSB level was not significantly affected ( Figure 6E ). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020.

infection-related genes in SARS-CoV-2 infection requires further exploration. However, CTSL 1 would be a promising therapeutic target for inhibitors that could not only inhibit entry of the virus 2 but also block the vicious circle ( Figure 6F ). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint vitro and inhibition data in vivo. Therefore, it is reasonable to conclude that CTSL, TMPRSS2 1 and furin are all required for SARS-CoV-2 infection. Broadening the range of therapeutic targets 2 for COVID-19 is important, and the effects of CTSL should not be underestimated.

The COVID-19 pandemic has motivated the most immense efforts to date to identify drugs 4 that can safely, quickly, and effectively reduce morbidity and mortality. Focusing on repurposing 5 of a licensed drug for COVID-19 may be more efficient than starting with a preclinical drug. 6 Thus, repurposing of many FDA-approved drugs for COVID-19 has been suggested, e.g., as was 7 accomplished with the antimalarial drugs chloroquine (CQ) and hydroxychloroquine (HCQ) for 8 rheumatoid arthritis and with the anti-influenza drug amantadine for Parkinson's disease (Ballout 9 et al., 2020; Gautret et al., 2020; Mitja and Clotet, 2020) .

Amantadine is a preventive agent first used for influenza and later for Parkinson's disease.

Amantadine has been hypothesized to have therapeutic utility for COVID-19 partially because In conclusion, we report that SARS-CoV-2 infection promoted CTSL expression and enzyme 20 activity, which, in turn, enhanced viral infection. CTSL functionally cleaved the SARS-CoV-2 S 21 protein and enhanced viral entry. Hence, CTSL is likely an important therapeutic target for 22 COVID-19. Furthermore, we showed that amantadine, a licensed anti-influenza drug, is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint for COVID-19 treatment would be a worthwhile endeavor. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint Tables S1 to S4 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.25.20218990 doi: medRxiv preprint (A) Plasma CTSL and CTSB levels patients with severe (n=20) and nonsevere COVID-19 (n=67) 1 upon hospital admission (Day 0). Statistical significance was assessed by the Mann-Whitney U 2 test (two-sided).

Plasma CTSL and CTSB levels in COVID-19 patients (n=87) and age-/sex-matched healthy 4 volunteers (n=125). The green lines in panels A and B indicate the reference ranges for each 5 parameter, established as the mean values ± 2 SD in the healthy participants. Statistical 6 significance was assessed by the Mann-Whitney U test (two-sided). 7 (C) Correlation between CTSL in plasma from COVID-19 patients (n=87) and the number of 8 days from symptom onset to blood collection before therapy. Statistical significance was assessed 9 by Spearman correlation analysis (two-sided). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint (H) values, and VSV-P mRNA levels (I). n=5. Statistical significance was assessed by the 1 Kruskal-Wallis test with Dunn's post hoc test.

The data are expressed as the mean ± s.e.m. values. *P<0.05, **P < 0.01, ***P<0.001, 3 ****P<0.0001. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. protein was incubated in the presence or absence (assay buffer, pH=5.5) of CTSL (2 or 10 µg/ml 8 in assay buffer, pH=5.5) at 37℃ for 1 h. The reaction system of 2 µg/ml CTSL was further 9 supplemented with CTSL inhibitors (20 µM E64d or 20 µM SID 26681509) as indicated.

Proteins were subjected to SDS-PAGE and detected by silver staining. Representative data from 11 three independent experiments are shown. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. The data are expressed as the mean ± s.e.m. values. *P<0.05, **P < 0.01, ***P<0.001. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ;

(A) Schematic of the CTSL inhibitor assay setup. Huh7 cells were pretreated with different drugs is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint Human ACE2 transgenic mice were randomly divided into four groups and pretreated with 1 vehicle or different drugs (E64d or amantadine) as indicated 2 days prior to virus inoculation via 2 tail vein injection (1.5×10 6 TCID50 per mouse). Mice without pseudovirus inoculation were used 3 as the healthy control group. Bioluminescence was measured one day post infection and 4 visualized in pseudocolor.

The relative intensities of emitted light are presented as the photon flux values in 6 photon/(sec/cm 2 /sr) and displayed as pseudocolor images, with colors ranging from blue (lowest 7 intensity) to red (highest intensity). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. Hofmann, H., Li, X., Zhang, X., Liu, W., Kühl, A., Kaup, F., Soldan, S.S., González-Scarano, F., Weber, F., 32 He, Y., et al. (2013) . Severe fever with thrombocytopenia virus glycoproteins are targeted by neutralizing 33

antibodies and can use DC-SIGN as a receptor for pH-dependent entry into human and animal cell lines. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020 .10.25.20218990 doi: medRxiv preprint Mitja, O., and Clotet, B. (2020 . Use of antiviral drugs to reduce COVID-19 transmission. Lancet Glob Health 1 8, e639-e640. 2 Nie, J., Li, Q., Wu, J., Zhao, C., Hao, H., Liu, H., Zhang, L., Nie, L., Qin, H., Wang, M., et al. (2020) . 3

Establishment and validation of a pseudovirus neutralization assay for SARS-CoV-2. Emerg Microbes Infect 4 9, 680-686. 5

Ou, X., Liu, Y., Lei, X., Li, P., Mi, D., Ren, L., Guo, L., Guo, R., Chen, T., Hu, J., et al. (2020) . Characterization 6 of spike glycoprotein of SARS-CoV-2 on virus entry and its immune cross-reactivity with SARS-CoV. Nat 7

Commun 11, 1620. Zhou, N., Pan, T., Zhang, J., Li, Q., Zhang, X., Bai, C., Huang, F., Peng, T., Zhang, J., Liu, C., et al. (2016) . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint al. (2020) . A pneumonia outbreak associated with a new coronavirus of probable bat origin. Nature 579, 270-1 273. 2 Ziegler, C.G.K., Allon, S.J., Nyquist, S.K., Mbano, I.M., Miao, V.N., Tzouanas, C.N., Cao, Y., Yousif, A.S., 3

Bals, J., Hauser, B.M., et al. (2020) . SARS-CoV-2 Receptor ACE2 Is an Interferon-Stimulated Gene in Human 4

Airway Epithelial Cells and Is Detected in Specific Cell Subsets across Tissues. Cell 181, 1016-1035 e1019. 5 6 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 /2020.10.25.20218990 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 27, 2020. ; https://doi.org/10.1101 https://doi.org/10. /2020 1 / 6 

The data that support the findings of this study are available from the Leading Contact, Dr. Jin-Kui Yang (jkyang@ccmu.edu.cn).

All the unique reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.

This study did not generate any unique datasets or code. bicarbonate, 0.1 mM non-essential amino acids and 1.0 mM sodium pyruvate and 10% FBS. All the cells were maintained at 37°C in a humidified atmosphere containing 95% air and 5% CO2.

The study used 4-5-week-old human ACE2 transgenic mice (weight 13-17g), a mouse model expressing human ACE2 (hACE2) generated by using CRISPR/Cas9 knockin technology as previous reported (Sun et al., 2020) . All animal protocols were approved by the Ethical Review

Committee at the Institute of Zoology, Capital Medical University, China.

Plasma samples of patients with COVID-19 collected at admission (d0), 14th day after discharge (d28) and 28th day after discharge (d42) were collected and stored at -80°C within 2h. Angiotensin

(1-7), ACE2 , CTSL and CTSB were analyzed using commercially available enzyme-linked immunosorbent assays (ELISA) following the manufacturer's instructions. All samples were detected without virus inactivation to retain the original results in a P2+ biosafety laboratory.

The SARS-CoV-2, SARS, RVF and VSV pseudovirus were generated with the incorporation of SARS-CoV-2 S protein, SARS-CoV-1 S protein, RVF G protein, and VSV G protein into VSVbased pseudovirus system, respectively. For this VSV-based pseudovirus system, the backbone was provided by VSV G pseudotyped virus (G*ΔG-VSV) that packages expression cassettes for firefly luciferase instead of VSV-G in the VSV genome (Nie et al., 2020) . Therefore, the luciferase activity and the mRNA level of VSV phosphoprotein (VSV-P) were used for indicators of pseudovirus infection.

Luciferase assay

The activities of firefly luciferases were measured on cell lysates using luciferase substrate following the manufacturer's instructions. Briefly, for 96-well plates, the culture supernatant was aspirated gently to leave 100μl in each well; then, 100μl of luciferase substrate was added to each well. Two minutes after incubation at 37°C, 150μl of lysate was aspirated to a clean 1.5 ml sterile EP tube to measuring the firefly luciferase activity for each well rapidly using a luminometer (Turner BioSystems) as described previously (Yang et al., 2017) .

Huh7, 293T, A549 and Calu-3 cells were plated in 48-well plates respectively and infected with different dose of SARS-CoV-2 pseudovirus (starting from 0 to 1.3×10 4 TCID50/ml). The cells were cultured for another 24 hours before luciferase activity analysis. Cells without the addition of pseudovirus as the cell control. The most susceptible cell line was selected for subsequent experiments. To verify the clinical data, Huh7 cells were plated in 48-well plates, and allowed to adhere until the cells are about 70% confluent, followed by infecting with different dose of SARS-CoV-2 pseudovirus (starting from 0 to 1.3x10 4 TCID50/ml). After 24 hours incubation, the cells were lysed for analysis the firefly luciferase activity, VSV-P mRNA and the detection of CTSL and CTSB by ELISA assays.

Purified ectodomain of SARS-CoV-2 S protein (NCBI reference sequence YP_009724390.1, residues 16-1213) and ectodomain of SARS-CoV (isolated: Tor2) S protein (NCBI reference sequence NP_828851.1, residues 1-1195) were expressed in baculovirus-insect cells. One microgram of each protein was incubated with 2 or 10 µg/ml CTSL in assay buffer (400 mM sodium acetate, pH 5.5, with 4 mM EDTA and 8 mM DTT) for 1h at 37℃. CTSL was preactivated 7 / 10 in 30℃ for one minute before use. Where indicated, 20 µM E64d or 20 µM SID26681509 was added in the reaction system with 0.5 µg SARS-CoV-2 S protein and 2 µg/ml CTSL. The proteins were then subjected to sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by silver stain.

Huh7 cells were seeded in 24-well plates and transfected with SARS-CoV-2 S protein expression plasmids (NCBI reference sequence YP_009724390.1, QHD43416.1) (0.65 µg/well) using

Lipofectamine 3000 reagent. The transfection solutions were changed to standard culture medium 6 h post-transfection and cells incubated for additional 12 h. Next, cells were treated in the absence (PBS, pH=7.4) or presence of 2 µg/ml trypsin (Sigma-Aldrich) (in PBS, pH=7.4), or in the absence (PBS, pH=5.8) or presence of 2 or 4 µg/ml CTSL (Sigma-Aldrich) (in PBS, pH=5.8) for 20 min at 37℃. Then, the solutions were changed to standard culture medium and the cells further incubated for 16 h. The pictures were captured under bright-field microscopy (Olympus) and

analyzed the formation of syncytia by counting the nuclei in syncytia in five random microscopic fields.

For CTSL knockdown, Huh7 cells were plated in 48-well plates, and transfected with 50 nM or 100 nM siRNAs against homo CTSL mRNA (si-CTSL) or 50 nM negative control siRNA (scramble) using Lipofectamine 3000 reagent. For CTSL overexpression, Huh7 cells were plated in 48-well plates, and transfected with 0.2 μg or 0.4 μg human CTSL expression plasmid (pENTER-CTSL, pCTSL) or 0.2 μg control plasmid (pENTER-vector, Con). 24 hours post transfection, the cells were lysed for analysis the CTSL and CTSB mRNA level to evaluate the efficiency of si-CTSL and pCTSL. To evaluate the effect of CTSL on SARS-CoV-2 entry, Huh7 8 / 10 cells were plated in 48-well plates, and transfected with si-CTSL or pCTSL under the same conditions stated above. 24 hours post transfection, the medium was replaced with fresh medium.

Then the cells were infected with SARS-CoV-2 pseudovirus (1.3x10 4 TCID50/ml) and cultured for another 24 hours before firefly luciferase activity and VSV-P mRNA analysis.

The anti-SARS-CoV-2 activity of SID26681509, E64d and amantadine were performed in 96-well plates by quantification of the firefly luciferase activity. Huh7 cells were pretreated with different concentrations of drug or the equivalent amount of solvent for 1 hour and then infected with SARS-CoV-2 pseudovirus (1.3x10 4 TCID50/ml) in a 5% CO2 environment at 37°C for 24 hours before firefly luciferase activity analysis. In detail, the concentrations of different drugs as follow:

SID26681509(0.2μM, 2μM, 4μM, 20μM, 40μM and 100μM), E64d (0.14μM, 0.42μM, 1.23μM, 3.7μM, 11.1μM and 33.3μM), and amantadine (1.56μM, 6.25μM, 25μM,100μM, 400μM and 1600μM). The effects of SID26681509, E64d and amantadine on cell viability were measured by MTT assay. Huh7 cells were seeded into a 96-well plate at a cell density of 0.5×10 4 per well and allowed to adhere until the cells are about 70% confluent, followed by treatment with different concentrations of drugs or the equivalent amount of solvent for 24 hours. The concentrations of different drugs were detailed above. Cells without any treatments as the blank control. After treatments, MTT was added into the culture medium to the final concentration of 0.5mg/ml, and then the cells were incubated for 4 hours at 37°C in an incubator. After removing the culture medium, the cells were lysed by gently rotating in 200µl DMSO for 10 minutes in darkness at room temperature. The absorbance at 570nm was measured using an automatic plate reader. The average absorbance reflected cell viability with the data normalized to the blank control group.

Experiments were done in quintuplicates and repeated at least three times.

The anti-SARS-CoV-2 activity of E64d and amantadine were performed in human ACE2 transgenic mice by bioluminescent imaging (BLI) assay as before (Zhang et al., 2017) . In brief, 4-5-week-old human ACE2 transgenic mice were treated with E64d (12.5mg/kg body weight) or amantadine (50mg/kg body weight) or the equivalent amount of solvent once a day via the intraperitoneal (IP) route 2 days prior to virus inoculation. Then mice were injected with 1.5x10 6 TCID50 SARS-CoV-2 pseudovirus per mouse via tail vein injection. Mice pre-treated with drug solvent but without pseudovirus inoculation served as the healthy control group. Bioluminescence was measured one day post-infection and visualized in pseudocolor. Finally, mice were sacrificed for experimental analysis immediately after bioluminescence measurement.

Total RNA was extracted from cultured cells or mouse livers and the reverse transcription was performed. The real-time qPCR was then performed on the LightCycler® 96 Real-Time PCR System (Roche) using SYBR Green I Master Mix reagent with the primers and using GAPDH as the house-keeping gene.

Clinical data were expressed as median (interquartile range (IQR)) or percentage, as appropriate.

Comparison of continuous data between groups were determined using Mann-Whitney U test. Chisquare (χ 2 ) test or Fisher's exact tests were used for categorical variables as appropriate. To explore the risk factors associated with severity, univariate and multivariate logistic regression models were used. Spearman's rho test (2-tailed) were used to analyze nonparametric correlations of parameters correlated with SARS-CoV-2 infection and severity of the disease. SPSS for Windows

Establishment and validation of a pseudovirus neutralization assay for SARS-CoV-2

A Mouse Model of SARS-CoV-2 Infection and Pathogenesis

NFE2 Induces miR-423-5p to Promote Gluconeogenesis and Hyperglycemia by Repressing the Hepatic FAM3A-ATP-Akt Pathway

A bioluminescent imaging mouse model for Marburg virus based on a pseudovirus system