The MEK1/2-inhibitor ATR-002 efficiently blocks SARS-CoV-2 propagation and alleviates pro-inflammatory cytokine/chemokine responses

Coronavirus disease 2019 (COVID-19), the illness caused by a novel coronavirus now called severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has led to more than 260 million confirmed infections and 5 million deaths to date. While vaccination is a powerful tool to control pandemic spread, medication to relieve COVID-19-associated symptoms and alleviate disease progression especially in high-risk patients is still lacking. In this study, we explore the suitability of the rapid accelerated fibrosarcoma/mitogen-activated protein kinase/extracellular signal-regulated kinase (Raf/MEK/ERK) pathway as a druggable target in the treatment of SARS-CoV-2 infections. We find that SARS-CoV-2 transiently activates Raf/MEK/ERK signaling in the very early infection phase and that ERK1/2 knockdown limits virus replication in cell culture models. We demonstrate that ATR-002, a specific inhibitor of the upstream MEK1/2 kinases which is currently evaluated in clinical trials as an anti-influenza drug, displays strong anti-SARS-CoV-2 activity in cell lines as well as in primary air–liquid-interphase epithelial cell (ALI) cultures, with a safe and selective treatment window. We also observe that ATR-002 treatment impairs the SARS-CoV-2-induced expression of pro-inflammatory cytokines, and thus might prevent COVID-19-associated hyperinflammation, a key player in COVID-19 progression. Thus, our data suggest that the Raf/MEK/ERK signaling cascade may represent a target for therapeutic intervention strategies against SARS-CoV-2 infections and that ATR-002 is a promising candidate for further drug evaluation. Supplementary Information The online version contains supplementary material available at 10.1007/s00018-021-04085-1.

Moreover, DTA pose the risk of emerging resistant virus strains, as with the M2 ion channel inhibitor Amantadine that rapidly forced influenza virus to develop a fully resistant phenotype [15]. In this regard, host targeting antivirals (HTA) directed against cellular factors that are required for viral life cycle might be advantageous, due to the inability of the virus to circumvent the affected cellular function. In the last decade, replication of DNA [16][17][18] and RNA viruses [15,[19][20][21][22][23][24][25] has been found to rely on the Raf/MEK/ERK signaling pathway which is involved in a vast variety of cellular processes including cell proliferation, differentiation, and survival [26], suggesting that this signaling axis is a promising antiviral target. Importantly, SARS-CoV spike (S) protein affects calcium-dependent activation of PKCα and promotes COX-2 protein synthesis via Raf/MEK/ERK activation [27]. Moreover, genomic and subgenomic RNA synthesis of the murine betacoronavirus mouse hepatitis virus (MHV) can be blocked through inhibition of MEK1/2 [28]. Encouraged by these findings, we thus explored the antiviral capacity of ATR-002 (Zapnometinib), an inhibitor of the two kinase isoforms MEK1 and MEK2, in the SARS-CoV-2 infection scenario. ATR-002 and its parental drug CI-1040 inhibit the replication of influenza viruses in vitro and in vivo [21,29]. A recent phase 1 clinical trial in healthy individuals has already confirmed that ATR-002 is safe and well tolerated (ClinicalTrials.gov Identifier: NCT04385420), paving the way for the planned phase 2 trial to treat influenza in hospitalized patients. In addition to the direct effect on viral replication, both ATR-002 and CI-1040 are immunomodulatory, decrease the expression of pro-inflammatory cytokines/chemokines [20,30], and thus might prevent the life-threatening hyperinflammation observed in patients suffering from severe COVID-19. In this study, we report the impact of the Raf/MEK/ERK signaling on SARS-CoV-2 replication and evaluate the anti-SARS-CoV-2 activity of the MEK1/2-inhibitor ATR-002 to counterbalance virusinduced pro-inflammatory cytokine/chemokine expression.
Human alveolar lung epithelial cells (A549) from ATCC, African green monkey kidney epithelial cells (VeroE6) and human embryonic kidney cells (HEK293T) from the cell line collection of the IVM were grown in Dulbecco's modified Eagle medium (DMEM) (Sigma-Life Science) supplemented with 10% (v/v) FBS.
Human primary nasal AECs (Table S2) were harvested from healthy adult volunteer donors by nasal brushing of the inferior turbinate and processed for cultivation at the air-liquid interface (ALI) as described previously [31]. Briefly, immediately after collection, the sampling brush was agitated to detach cells from the brush tip into 5 ml of serum-free isolation medium (Minimum Essential Medium, MEM) (Thermo Fisher Scientific) supplemented with 100 U/ml P/S (Lonza), 0.25 µg/ml amphotericin B (Sigma-Life Science), 50 µg/ml gentamicin (Sigma-Life Science), and 100 U/ml nystatin (Thermo Fisher Scientific). Subsequently, AECs were seeded at a concentration of 5 × 10 4 /ml in epithelial cell serum-free growth medium (BEGM; LHC basal medium) (Thermo Fisher Scientific) plus additives [31] supplemented with 1% P/S and 50 µg/ml gentamicin (from day 6 on without gentamicin) into culture flasks precoated with collagen solution from human fibroblasts (Sigma-Aldrich). Media were changed 24 h after seeding and thereafter every 3 days. After 7-10 days of culture, after reaching about 70-90% confluence, cells were detached with trypsin-EDTA 0.05% (Gibco) for 5 min and seeded apically on Transwell permeable supports (6.5 mm diameter; 0.4 μm pore size) (Corning) in ALI medium (LHC basal medium and DMEM at 1:1 plus additives [31] supplemented with 1% P/S) at a density of 50,000 cells in 200 µl ALI medium per Transwell. About 500-600 µl ALI medium was placed at the basolateral side and changed every other day. Once the cells had reached full confluency (after 5-7 days), they were induced to differentiate at ALI by removing medium at the apical side of the Transwell. After 21 days of culture at ALI, welldifferentiated AECs were used for virus infection.
All cells were incubated at 37 °C and 5% CO 2 atmosphere.  (Table S1) were diluted in cell culture medium according to the desired MOI. Cells were washed twice with PBS (Sigma-Aldrich) and incubated with the viral dilutions for 1 h at 37 °C. Cells were washed once with PBS and incubated in cell culture medium. Primary AECs were washed twice with HBSS (Thermo Fisher Scientific) and incubated apically with the viral dilutions for 2 h at 37 °C. Cells were washed twice with HBSS and incubated at ALI.

SARS-CoV-2 viruses
After the indicated total incubation times, supernatants were collected and cells were subjected to the respective analysis method.
Inhibitor treatment was performed as described in the respective experiments. siRNA knockdown ERK1/2 was knocked down in Calu3 cells using Lipofectamine ® 2000 (Invitrogen) according to the manufacturer's protocol.

Cell cytotoxicity assay
Calu3 cells were grown in 96-well flat-bottom tissue plates. 24 h after seeding cells were incubated with different inhibitor concentrations (Table S3)

Quantitative real-time PCR
Total RNA was isolated using the RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. cDNA was synthesized mixing 1 µg RNA, Revert AID H Minus Reverse Transcriptase (Thermo Fisher Scientific), and oligo (dT) primers (Eurofins MWG Operon). The qRT-PCR reaction mix Brilliant III SYBR Green QPCR Master Mix (Agilent Technologies), specific primers (Table S1), and the LightCycler ® 480 II (Roche) in combination with the program LightCycler ® 480 SW V1.5.1.62 were used for the qRT-PCR analysis. GAPDH was used as reference.

HTRF-based quantification of MAPK/ERK1/2 phosphorylation levels
To quantify the levels of pERK1/2 (Thr202/Tyr204) in mock or SARS-CoV-2-infected Calu3 cells, a plate-based assay (Cisbio) was used according to the manufacturer's protocol. Briefly, cells grown on 96-well plates (50,000 cells/well) were non-infected or infected with different SARS-CoV-2 isolates for 1 h at 37 °C. Lysates were then transferred to a 384-well low-volume plate, incubated with the antibodies for 4 h at room temperature, and luminescence was recorded with a CLARIOstar reader (BMG Labtech) (200 flashes/well, integration start 60 µsec, integration time 400 µsec, and settling time 100 µsec). HTRF ratios were normalized to maximum system output obtained through stimulation with 100 nM phorbol 12-myristate 13-acetate (PMA) (AppliChem).

VSV-pseudotyped system
SARS-CoV-2 S protein receptor-mediated entry was analyzed with the SARS-CoV-2 S protein-bearing vesicular stomatitis virus (VSV) pseudotyped system, containing a G gene substitution by a green fluorescence protein gene (VSV ΔG/GFP-Luc + S Δ21). The pseudotyped virus was generated according to Berger Rentsch and Zimmer, 2011 [32]. 24 h after transfection of the S coding plasmid into HEK293T cells, the inoculation with VSV ΔG + G-transcomplemented virus particles was initiated (provided by G. Zimmer). After 1 h, incubation cells were washed with PBS and treated with culture medium supplemented with anti-VSV-G antibody produced by I1-hybridoma cells to inactivate residual VSV-G particles. After 18 h incubation, the supernatant was collected and centrifuged through a 100 kDa spin column to concentrate the produced virus. Afterward, virus titration was performed in VeroE6 cells using the virus carrying GFP reporter to obtain FFU/ml. Cells were infected with the pseudotyped VSV ΔG/ GFP-Luc + S Δ21 using an MOI of 0.003 for 1 h. ATR-002 (100 µM) treatment was initiated simultaneously with the infection (0 h) or 1 h post-infection. The solvent DMSO served as control. 16 h.p.i. GFP-positive cells were quantified by light microscopy using the Axiovert 200 M microscope (Zeiss).

Quantification and statistical analysis
Western blot signal intensities were measured using the LiCor Image Studio™ version 5.2.5 (LiCor). Immunofluorescence pictures were formatted using the AxioVersion V4.8.2.0 software (Zeiss) and quantified with ImageJ 1.48v (NIH Image). LEGENDplex data were analyzed using the LEGENDplex™ Data Analysis Software version 8.0.
GraphPad PRISM version 8.4.3 (GraphPad Software) was used to create the shown graphs and perform statistical analysis. Information about sample size and statistical tests are shown in the respective figure legends.

SARS-CoV-2 infection triggers an early monophasic activation of the Raf/MEK/ERK signaling pathway
We first assessed whether SARS-CoV-2 infection activated the Raf/MEK/ERK kinase cascade. Western blot analysis of SARS-CoV-2 (D614G-FI)-infected Calu3 cells revealed a prominent ERK1/2 activation in the early phase of the infection, with a peak of ERK1/2 phosphorylation 1 h post-infection (h.p.i.), followed by a decrease over time ( Fig. 1a-c). To explore the correlation between viral titers   Fig. S1a, b. a, d Shown are means ± SD of three independent experiments. Data passed an one-way ANOVA followed by Dunnett's multiple comparison test (*p ≤ 0.0332; **p ≤ 0.0021; ***p ≤ 0.0002; ****p ≤ 0.0001) and ERK1/2 activation, we infected Calu3 cells with various multiplicities of infection (MOI: 0.5, 1.0, 2.0) and analyzed the changes in ERK1/2 phosphorylation in a homogeneous time-resolved fluorescence (HTRF)-based assay. A significant virus concentration-dependent increase in ERK1/2 phosphorylation was seen 1 h.p.i. (Fig. 1d; Fig. S1a, b). We included a Wuhan-like wild type isolate (WT), additional D614G isolates (LP, NK), and variants of concern (VOC) (α-B1.1.7, β-B1.351) to exclude a strain-or variant specific effect (Fig. S1a, b). While the activation of the Raf/MEK/ ERK signaling cascade seemed to be crucial during the very early phase of infection, a second activation phase of the pathway could not be detected at later stages. At 4 h.p.i., the time point at which viral protein expression (S, N) was detected, the ERK1/2 phosphorylation intensity dropped below the level observed in mock-infected cells (Fig. 1a-c).
The antiviral interferon (IFN) response was induced 1 h.p.i. represented by an increase in the phosphorylation of the interferon regulatory factor 3 (IRF3) and the inflammatory response was activated 2 h.p.i. indicated by an induction in the NFκB-p65 phosphorylation (Fig. 1a-c) Between 5 and 6 h.p.i., first virus particles were released and signal transducer and activator of transcription (STAT1) phosphorylation occurred, a prerequisite for the onset of the antiviral IFN response (Fig. 1a-c; Fig. S1c).
The resulting very favorable selectivity indices (SI) between 24.05 and 56.64 (Table 1) confirmed that ATR-002 activity was selective toward the viruses and not the host.
Notably, 72 h.p.i. viral titers (D614G-FI) were reduced by a half log unit at 50 µM ATR-002 up to 6.5 log units at 150 µM ATR-002, demonstrating a sustained inhibitory impact on the viral life cycle at non-cytotoxic concentrations (Fig. S2c, d). A cytopathic effect (CPE) caused by viral infection was first detected 48 h.p.i. which was not visible upon treatment with 50 µM ATR-002. Higher ATR-002 concentrations of 100 µM and 150 µM prevented CPE even 72 h.p.i. (Fig. S2e). Together, these results strongly support the suitability of the MEK1/2inhibitor ATR-002 as a host targeting antiviral.

The Raf/MEK/ERK signaling pathway plays a role in the early phase of the SARS-CoV-2 life cycle
We next addressed the question whether the antiviral activity of ATR-002 was limited to the early phase of the SARS-CoV-2 infection cycle. Therefore, cells were treated with ATR-002 either 1 h pre-infection or 2 h/4 h postinfection and the expression of viral proteins (S, N) and progeny viral titers were determined 8 h.p.i. (Fig. 3a-c;  Fig. S3). While the ERK1/2 phosphorylation was reduced by approximately 50% for the 10 µM ATR-002 concentration, higher amounts (50-150 µM) resulted in an almost complete inhibition of the phosphorylation signal. Interestingly, the expression of N protein remained unchanged, and S protein expression was only marginally reduced (18% ± 0.13%) when cells were treated with 10 µM of ATR-002 1 h pre-infection ( Fig. 3b; Fig. S3b). In addition, no changes in viral titers were found, indicating that the remaining ERK1/2 activity was sufficient to maintain its role in the viral life cycle ( Fig. 3c; Fig. S3c). A 10 µM ATR-002 treatment 2 h.p.i. and 4 h.p.i. did also not affect the protein expression and viral particle release. A significant reduction of viral protein expression and viral titers could be found at ATR-002 concentrations ranging from 50 to 150 µM, when treatments were started 1 h pre-infection (Fig. 3a-c; Fig. S3). In contrast, the inhibitory effect was less pronounced in the post-infection scenarios. Viral protein expressions started to occur for 100 µM 2 h.p.i. and for 150 µM 4 h.p.i. (Fig. 3a, b; Fig. S3a, b). In addition, viral titers increased with delayed ATR-002 treatment  Fig. 3c; Fig. S3c). Immunofluorescence analysis of the SARS-CoV-2 N protein expression confirmed these results (Fig. 3d, e). Comparable to the data shown in Fig. 3a-c, a strong inhibitory effect on the N expression was found for the 1 h pre-infection treatment but not for the 2 h.p.i. treatment. These data indicate that a very early activation of the Raf/MEK/ERK signaling is an important requirement for optimal viral replication.

ATR-002 restricts SARS-CoV-2 replication in different cell lines albeit with variable efficiency
To investigate a role of the early ERK1/2 activation during the viral internalization, we used the SARS-CoV-2 S protein-bearing vesicular stomatitis virus (VSV) pseudotype system, containing a G gene substitution by the green fluorescent protein gene (VSV Δ/GFP-Luc + S Δ21) [32,34]. With this S-pseudotyped VSV, specifically designed to analyze the SARS-CoV-2 S protein receptor-mediated entry, we observed a significant reduction of GFP-positive Calu3 cells of 36.81% ± 6,9% if the inhibitor treatment was initiated together with viral infection (0 h), whereas a not significant reduction of 9.9% ± 5,5% was found if the inhibitor treatment was initiated 1 h.p.i. (Fig. 4a). Comparable results were found for the colorectal cancer cell line CaCo2, that serves as a model for an epithelial barrier (Fig. S4q). Both cell lines are described to endogenously express transmembrane serine protease 2 (TMPRSS2) [35] (Fig. S4a). A strongly reduced inhibitory effect on the SARS-CoV-2 S protein receptor-mediated entry was found in the Vero cell lines VeroE6 and Vero76-TMPRSS2, which do express high amounts of Cathepsin L (Fig. S4a, q). These results indicate that virus-induced ERK1/2 activation is important for the early TMPRSS2-mediated internalization process.
To assess whether SARS-CoV-2-mediated activation of Raf/MEK/ERK signaling was cell type specific, we next used A549 cells genetically modified to render them permissive to SARS-CoV-2 [35]. In the A549 cell line, expressions of both the main SARS-CoV-2 entry receptor Angiotensinconverting enzyme 2 (ACE2) and TMPRSS2 responsible for the priming of the spike protein [36] are only expressed at low levels (Fig. S4a). Hence, A549 cells are only marginally permissive for SARS-CoV-2 [37]. As expected, we could not detect progeny viral titers in the A549 cell lines lacking robust levels of endogenous ACE2 (parental A549 cells and A549-TMPRSS2 cells stably expressing TMPRSS2), whereas ACE2 overexpression allowed for a successful infection at MOI 0.01. At 48 h.p.i. and 72 h.p.i., viral titers were 3 log units and 4 log units, respectively, higher in ACE2/TMPRSS2-expressing A549 cells compared to A549-ACE2 cells. Highest viral titers were obtained in the naturally permissive Calu3 cells (Fig. 4b). For successful infection at lower MOIs of 0.001, the combined expression of ACE2 and TMPRSS2 was required (Fig. 4b). Notably, infection of A549-ACE2, A549-ACE2/TMPRSS2 and Calu3 cells using an MOI 2 led to a significant phosphorylation of ERK1/2 1 h.p.i. The activation intensity increased in A549-ACE2 cells by the factor 1.6 ± 0.18, in A549-ACE2/ TMPRSS2 cells by the factor 2.26 ± 0.36, and in Calu3 cells by the factor 2.7 ± 0.07 compared to mock-infected cells. In line with their poor permissiveness, no increase in ERK1/2 phosphorylation was found for parental A549 and A549-TMPRSS2 cells (Fig. 4c, d). We additionally analyzed the ERK1/2 activation in VeroE6, Vero76-TMPRSS2, and CaCo2 cell lines. No ERK1/2 activation was found in the Cathepsin L expressing Vero cell lines, whereas in TMPRSS2-expressing CaCo2 cells, the ERK1/2 phosphorylation was increased by the factor 2.47 ± 1.46 (Fig. S4a, r, s).
The efficacy of ATR-002 to inhibit SARS-CoV-2 replication varied between the different cell lines, depending on the chosen MOI. Viral titers were significantly decreased in Calu3, A549-ACE2/TMPRSS2, and CaCo2 cells for all analyzed time points using an MOI of 0.001 ( Fig. 4e; Fig. S4b-d, h-j, t). Comparable effects were found for Calu3, A549-ACE2, and CaCo2 cells using an MOI of 0.01 ( Fig. 4e; Fig. S4e-g, n-p, u). No inhibitory effect was found for the Cathepsin L expressing Vero cell lines (Fig. S4t, u). Surprisingly, viral titers in A549-ACE2/TMPRSS2 cells were also not significantly decreased when the cells were infected with a MOI of 0.01 ( Fig. 4e; Fig. S4k-m).
These results confirm that the antiviral activity of ATR-002 is not restricted to the Calu3 cell line and that activation of ERK1/2 is linked to efficient SARS-CoV-2 infection of cells via ACE2 and TMPRSS2. Nevertheless, the extent of antiviral activity seems to vary among cell lines with respect to the viral dose used for inoculation.
It is already known that lower amounts of ATR-002 are sufficient to achieve high inhibitory effects on the viral replication of Influenza A viruses in primary cells due to their lower basal ERK1/2 activity compared to immortalized cell lines [21]. To confirm these findings and investigate the role of ERK1/2 activation on SARS-CoV-2 infection in a more physiological scenario we next assessed the antiviral activity of ATR-002 in air-liquid-interface (ALI) cultures of primary nasal airway epithelial cells (AEC) which represent one of the most accurate models for infection of the upper respiratory tract. In this infection model, treatment with 1 µM ATR-002 already reduced viral titers by 20-40%. A similar reduction was only achieved in Calu3 cells upon treatment with 50 µM ATR-002, a concentration that fully inhibited virus production in the AEC cultures ( Fig. 5; Fig. S5; Fig. S2a).

ATR-002 dampens the induction of pro-inflammatory cytokine expression
A hallmark of severe progression of COVID-19 is excessive lung damage that is mainly caused by acute inflammation  Fig. S4b, e, h, k, n. c Depicted cell lines were infected with SARS-CoV-2 (D614G-FI) (MOI 2) ( +). Immunoblots were prepared 1 h.p.i. and probed with anti-pERK1/2, anti-ERK1/2, and anti-Tubulin antibodies. Shown are results of one out of three independent experiments. Mock (-) infected cells served as negative control. d Quantification of (c). Intensity ratios of pERK1/2 were normalized to the total protein amount. Mock was arbitrarily set to 1.0. Data passed a paired two-tailed t test (*p ≤ 0.0332; ***p ≤ 0.0002) for each cell line separately. e Titer reduction of SARS-CoV-2 (D614G-FI) in different cell lines after MEK1/2-inhibition. 1 h.p.i. cells were treated with ATR-002 (100 µM). Untreated (SARS-CoV-2) and DMSO (0.1%) treated cells served as negative controls. Data passed an one-way ANOVA followed by Dunnett's multiple comparison test (**p ≤ 0.0021; ***p ≤ 0.0002; ****p ≤ 0.0001) for each cell line separately. DMSO was used as reference and arbitrarily set to 100%. See also Fig. S4c, d, f, g, i, j, l, m, o, p. a, b, d, e Data represent means ± SD of three independent experiments, each performed in triplicates (b, d, e) or in quadruplicates (a) ◂ and hypoxemia [38]. The resulting acute respiratory distress syndrome (ARDS) and extrapulmonary multi-organ dysfunction are provoked by the excessive production of pro-inflammatory cytokines (IL-6, CXCL8) and chemokines (CXCL10, CCL5, CCL2), a phenomenon also referred as cytokine storm [39]. Such exaggerated dysregulated cytokine response strongly increases the mortality rate in COVID-19 patients. An early and controlled alleviation of the cytokine storm can increase the likelihood of a mild course of the disease [39]. It could already be shown that MEK1/2-inhibitors have an inhibitory effect on the expression of pro-inflammatory cytokines/chemokines without interfering with the interferon-induced antiviral response [20,30]. Therefore, the question arose if ATR-002 has a similar beneficial immunomodulatory effect during the course of a SARS-CoV-2 infection. Cytokine and chemokine expression analysis of the infection kinetic shown in Fig.  S2d revealed a strong reduction of pro-inflammatory cytokines/chemokines, such as IL-6, CXCL8, CXCL10, CCL2, and CCL5, for ATR-002 concentrations of 100 µM and 150 µM ( Fig. 6a; Fig. S6a). Because mRNA levels of the antiviral acting proteins IFNβ and MxA were also decreased in this assay, we hypothesized that a potential direct effect of ATR-002 on cytokine/chemokine expression was dominated here by the virus-inhibiting effect of ATR-002, leading to reduced stimulus. Thus, in a next step, we used the A549-ACE2/TMPRSS2 cells, which showed no inhibitory effect on viral titers after ATR-002 treatment when infected with a MOI of 0.01 ( Fig. 4e; Fig. S4k-m), to exclude any effects caused by reduced viral growth. The IFN-mediated antiviral response was not altered in these cells. Comparable levels of IFNβ and MxA mRNAs were found for the controls (SARS-CoV-2, DMSO) and the ATR-002 treatment (100 µM). Despite the unchanged antiviral IFN response, pro-inflammatory cytokine and chemokine expression was significantly decreased. Strongest effects were found for CXCL8, with a reduction of 89.3% ± 0.86% compared to the DMSO control and weakest effects for IL-6 with a reduction of 49.5% ± 6.2% in mRNA expression (Fig. 6b). Comparable to the results in Fig. 4e, ATR-002 did not significantly reduce progeny viral titers (Fig. S6b). We furthermore aimed to analyze the direct inhibitory effect of ATR-002 on the expression of pro-inflammatory cytokines/chemokines in a virusfree system by transfecting and stimulating Calu3 cells with a synthetic analog of double-stranded RNA (polyI:C) mimicking viral RNA. This allows us to discriminate whether ATR-002 directly interferes with gene induction by a pathogen-associated molecular pattern (PAMP), because indirect effects due to reduced virus replication caused by ATR-002 in an infection scenario are avoided. No change in the cellular IFN response to the polyI:C stimulation was observed for the ATR-002 treatment using increasing amounts of the inhibitor (10-150 µM) (Fig. 6c). In contrast and in line with the results obtained in the A549-ACE2/TMPRSS2 cell line (Fig. 6b), expression of pro-inflammatory cytokines/ chemokines was significantly decreased. The strongest effect was found for the expression of CXCL8 with a reduction of 78.2% ± 12.9% in cells treated with 100 µM of ATR-002; the weakest effects were observed for the expression of CCL5 and IL-6 upon treatment with 100 µM of ATR-002 with reductions of 42.5% ± 7.1% and 48.4% ± 4.4%, respectively (Fig. 6c). In an additional attempt, we aimed to analyze the effect of the ATR-002 treatment on the secretion of proinflammatory cytokines/chemokines. Therefore, we stimulated Calu3 or A549-ACE2/TMPRSS2 cells with polyI:C in a combinational treatment with ATR-002 (10-150 µM). Comparable to the mRNA expression levels in Fig. 6b, c, the MEK1/2 inhibition via ATR-002 reduced the secretion of the pro-inflammatory cytokines/chemokines IL-6, CXCL8,  Fig. S6c, d. a-e Data passed an one-way ANOVA followed by Dunnett's multiple comparison test (*p ≤ 0.0332; **p ≤ 0.0021; ***p ≤ 0.0002; ****p ≤ 0.0001). a Oneway ANOVA was conducted for each time point separately. b-e DMSO was used as reference and arbitrarily set to 100% CXCL10, GM-CSF, and TNFα in a concentration-dependent manner (Fig. 6d, e; Fig. S6c, d).
These results strongly implied that the MEK1/2-inhibitor ATR-002 reduces the risk of a COVID-19-associated cytokine storm by attenuating the production of pro-inflammatory cytokine/chemokine expression.

Discussion
A multitude of antiviral interventions against SARS-CoV-2/COVID-19 are under evaluation. However, the development of novel therapeutics including interferons [40], convalescent plasma treatment [41], new small-molecule drugs [42], and monoclonal antibodies [14] might require years. An alternative approach is the repurposing of already licensed drugs to target either viral components or cellular mechanisms involved in SARS-CoV-2 replication. In this study, we explored the involvement of the cellular Raf/MEK/ERK pathway in the SARS-CoV-2 infection, and assessed its antiviral potential and the suitability of the MEK1/2-inhibitor ATR-002 as a potential drug repurposing candidate. Our observation of a monophasic activation of ERK1/2 in the early phase of SARS-CoV-2 infection ( Fig. 1; Fig. S1) adds to previous studies that already linked Raf/MEK/ERK signaling to the infection process of various viruses [15-23, 43, 44], including coronaviruses [27,28,[45][46][47]. The significantly reduced virus production of all tested SARS-CoV-2 variants in cells in which this signaling pathway was blocked by either ERK1/2-knockdown or MEK1/2-inhibition (Figs. 2, 3, 4, 5; Table 1; Fig. S2-S5) confirms the potential as a druggable target, with efficient inhibitory effects of ATR-002, reflected by the low SI values. It is worth to mention that these values, which do reflect the ratio between the drug compatibility (CC 50 ) and the inhibitory effect (IC 50 ), are steady within the different virus isolates, indicating that the mutational changes of the SARS-CoV-2 variants did not influence the inhibitory ability of ATR-002. Inhibition of MEK1/2 signaling prior to SARS-CoV-2 infection was most effective, strongly arguing for an important role of the Raf/MEK/ERK pathway in the very early stage of the viral life cycle (Fig. 3; Fig. S3). This assumption is affirmed by the finding that entry of SARS-CoV-2 S-pseudotyped VSV was significantly reduced in case of ATR-002 treatment simultaneous with infection, whereas a 1 h lagged treatment did not show a significant entry reduction ( Fig. 4a; Fig. S4q). As coronaviruses enter host cells either through a serine protease (e.g., TMPRSS2) dependent cell surface entry pathway or through an endocytic pathway using lysosomal cysteine proteases (Cathepsins), depending on the availability of serine proteases [48], the missing inhibitory effect of ATR-002 in A549-ACE2/TMPRSS2 cells when infected with higher viral titers ( Fig. 4; Fig. S4) might suggest that Raf/MEK/ERK activation is connected to SARS-CoV-2 endocytosis, resembling the IAV-induced early ERK1/2 activation that stimulates V-ATPase-dependent endosomal acidification required for HA-mediated fusion of the viral envelope with the endosomal membrane [49]. However, we found only moderate inhibitory effects of ATR-002 on the SARS-CoV-2 S-pseudotyped VSV in Cathepsin L expressing VeroE6 and Vero76-TMPRSS2 cells, compared to cell lines for which we could not detect high amounts of Cathepsin L (Calu3, Caco2) ( Fig. 4a; Fig.  S4a, q). These findings are supported by the absence of the ERK1/2 activation in the Vero cell lines (Fig. S4r, s).
Another result pointing toward the involvement of TMPRSS2 is the high inhibitory ability of ATR-002 in primary human airway epithelial cells (AEC) (Fig. 5;  Fig. S5), as it is already described that SARS-CoV-2 S-driven entry in AECs can efficiently be blocked with the TMPRSS2-inhibitor camostat mesylate [36]. If the Raf/MEK/ERK pathway would be related to the cathepsin pathway, strong inhibitory effects could be expected in the Cathepsin L expressing Vero cell lines, which we did not find in our experiments. Despite the evidence of the involvement of active ERK1/2 in the TMPRSS2-mediated cell surface entry pathway, we could not decipher the exact contribution of ERK1/2 in this process. Our data suggest that it might act as a cofactor or a posttranslational modifier for the entry process. Different studies have identified several cofactors that seem to play a role in the SARS-CoV-2 entry process. Besides TMPRSS2 and Cathepsin L [36], a contributing role of NRP-1 [50], all known to be modified by phosphorylation, as well as HSPGs [51] are described. Of note, except Cathepsin L, these cofactors are localized at the cell surface, where they directly interact with the spike protein. As intracellular kinase, a direct interaction of ERK1/2 during the binding process can be excluded. Nevertheless, active ERK1/2 or a downstream kinase may phosphorylate cofactors directed to plasma membrane, allowing for an indirect contribution to the internalization process.
Since hyperinflammation and multiple organ failure (lung, heart, kidney, liver, and brain) [5,38,39,52] drive severe COVID-19 in late stages, pharmacological suppression of the pro-inflammatory cytokine/chemokine induction rather than interference with viral replication might be a promising treatment option. Most cytokines/ chemokines are involved in immune cell activation and recruitment to the side of infection and inflammation; however, hyperactivation can have fatal consequences for the patients. MEK1/2-inhibition can reduce but not fully block the production of pro-inflammatory cytokines/ chemokines ( Fig. 6; Fig. S6) [20]. However, a drug should not completely suppress all cytokines/chemokines, since, e.g., induction of the antiviral interferon response is vital to mount an antiviral defense. As shown here and elsewhere, MEK1/2-inhibitors did not affect IFNβ-and interferon stimulated gene (ISG) mRNA expression such as, MxA. The data using a synthetic analog of vRNA (polyI:C) (Fig. 6c) and comparable results in A549-ACE2/TMPRSS2 cells (Fig. 6b) show that this is a direct effect on the induction mechanism of pro-inflammatory cytokines/chemokines and not indirectly caused via viral titer reduction.
In conclusion, the low efficiency of DTAs against acute, self-limiting and hyperinflammatory viral diseases, such as severe influenza and COVID-19, observed so far might call for a rethinking of antiviral concepts in general. A DTA attempt might be simply too late to rescue severely diseased patients in late stages of the disease. Here, host targeting antivirals (HTA) that at the same time have immunomodulatory effects, such as ATR-002, might offer a suitable toolbox for antiviral therapies of different viral origin (different viruses misuse similar cellular factors) without posing the risk of emerging resistances. Besides promising data on bioavailability [21] and safety of the compound (ClinicalTrials. gov Identifier: NCT04385420), a phase 2 clinical trial was initiated to verify the antiviral properties of ATR-002 (Zapnometinib) (RESPIRE: EudraCT:2020-004206-59) in intermediate stages of hospitalized COVID-19.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.