Lenalidomide downregulates ACE2 protein abundance to alleviate infection by SARS-CoV-2 spike protein conditioned pseudoviruses

Dear Editor, The recent health emergency caused by SARS-CoV-2 created a global pandemic. Similar to other CoVs, SARS-CoV-2 utilizes its Spike (S) protein to specifically recognize the human angiotensin converting enzyme 2 (ACE2) membrane receptor for infection. Thus, blocking SARS-CoV-2-S/ACE2 interactions has been investigated as a therapeutic direction in treating SARS-CoV-2, including recombinant hACE2 proteins, ACE2-derived peptides, neutralizing antibodies, engineered ACE2 traps, various heparins, TMPRSS2 inhibitors, and others. Deep mutagenesis assays identified engineered ACE2 mutants with enhanced binding affinity with SARS-CoV-2-S proteins, facilitating development of approaches targeting SARS-CoV-2 S protein binding to ACE2. ACE2 is a plasma membrane attached enzyme converting angiotensin II into angiotensin (1–7) that lowers blood pressure. ACE2 is broadly expressed but detectable expression is enriched in alveolar cells, enterocytes, arterial, and venous endothelial cells. The E3 ligase MDM2 was reported to target ACE2 for degradation in an AMPK phosphorylation-dependent manner. Intriguingly, increased ACE2 levels were observed in lungs from patients with comorbidities associated with severe COVID-19. Modulating ACE2 protein homeostasis is an attractive direction to prevent/control SARSCoV-2 infection; however, regulatory mechanisms of ACE2 protein control remain elusive. We profiled ACE2 protein expression in a panel of commonly used cell lines, and only observed a detectable ACE2 signal (~130 KD) in UMRC2 cells using an antibody targeting ACE2 C-terminus (sc-390851) (Supplementary Fig. 1a–c). This signal was also detected by an ACE2 N-terminus targeting antibody (CST#4355), with an additional band at ~75 KD, which may correspond to a shorter isoform of human-ACE2 (uniport Q9BYF1). We further examined additional kidney cancer cells and observed full-length ACE2 in UMRC2, UMRC6, and RCC4 cells, with the lower 75 KD band (Fig. 1a). Depleting endogenous ACE2 by shRNAs in UMRC2, UMRC6, or A498 cells led to reduced intensity of the upper 130 KD band with minimal effects on the lower 75KD band (Supplementary Fig. 1d–f), suggesting only the ~130 KD ACE2 band is ACE2. This was further confirmed by PCR analyses of cDNAs obtained from these kidney lines that full-length ACE2 genes were observed in UMRC2, UMRC6, and RCC4 cells (Supplementary Fig. 1g). Thus, we use these three cell lines to study ACE2 protein homeostasis, and the ACE2-C antibody to monitor full-length ACE2 proteins. Depleting ACE2 in either UMRC2 or UMRC6 cells didn’t significantly affect cell growth in vitro (Supplementary Fig. 1h–k). ACE2 protein sequence includes a putative degron for the E3 ligase SPOP, defined as 1-MSSSS-5, located at the extreme Nterminus of ACE2 signal peptide (amino acid 1–17) (Fig. 1b and Supplementary Fig. 2a). SPOP depletion reduced ACE2 protein abundance in UMRC2, UMRC6 (Fig. 1c) and RCC4 cells (Supplementary Fig. 2b), and didn’t affect UMRC6 cell growth in vitro (Supplementary Fig. 2c), nor constantly downregulating ACE2 mRNAs (Supplementary Fig. 2d, e). Ectopic SPOP expression reduced ubiquitination of either C-terminal HA-tagged ACE2 (Fig. 1d) or N-terminal GST-tagged ACE2 (Supplementary Fig. 3a), suggesting SPOP may interfere with ACE2 ubiquitination. Mutating SPOP degron 1-MSSSS-5 into 1-MSAAA-5 (3A), or deleting 3-SSS-5 reduced ACE2 binding to SPOP (Fig. 1e and Supplementary Fig. 3b), supporting 1-MSSSS-5 as a major motif for SPOP recognition. Given casein kinase(s) phosphorylates SPOP degrons to modulate SPOP binding, depletion of endogenous CK1α reduced ACE2 protein abundance in UMRC6, RCC4, and UMRC2 cells (Fig. 1f). CK1α depletion-induced ACE2 downregulation was partially rescued by MG132 (Supplementary Fig. 3c). Inhibiting CK1 by D4476 reduced ACE2 binding to SPOP and increased ACE2 ubiquitination (Fig. 1g, h and Supplementary Fig. 3d, e). D4476 treatment-induced ACE2 downregulation was observed in WT but not 3A-ACE2 expressing cells (Supplementary Fig. 3f), supporting 3-SSS-5 as major CK1 phosphorylation sites. CK1α expression reduced ACE2 ubiquitination (Fig. 1i). Intriguingly, lenalidomide, an FDA approved drug in treating multiple myeloma, follicular and marginal zone lymphoma, as a CK1α PROTAC, efficiently induced degradation of WT but not 3A-ACE2 (Fig. 1j) and attenuated ACE2 binding to SPOP (Fig. 1k), resulting in enhanced ACE2 ubiquitination (Fig. 1l and Supplementary Fig. 3g). These data support CK1-mediated ACE2-Ser3/Ser4/Ser5 phosphorylation triggering ACE2 recognition by SPOP. Next, we examined if inactivating CK1 affects endogenous ACE2 abundance. Inhibiting CK1 by D4476 or epiblastin A reduced ACE2 proteins in UMRC2 (Fig. 1m, n), UMRC6 (Supplementary Fig. 4a, b) and RCC4 (Supplementary Fig. 4c, d) cells. Lenalidomide as a CK1α PROTAC led to a canonical biphasic response of ACE2 expression in UMRC2, UMRC6, RCC4, and Calu-3 (Fig. 1o–q, and Supplementary Fig. 4e, f) cells. Lenalidomide-like compounds including CC122, pomalidomide, and thalidomide didn’t reduce CK1 expression nor ACE2 protein abundance (Supplementary Fig. 4g–i). These data suggest CK1 inactivation reduces ACE2 protein abundance (Fig. 1r). Given NEDD4 and MDM2 have been reported to be involved in ACE2 homeostasis control, we found ectopic expression of NEDD4 or MDM2 reduced ACE2 expression (Supplementary Fig. 5a, b). However, depleting neither NEDD4 nor MDM2 by sgRNAs upregulated ACE2 proteins in UMRC2 and UMRC6 cells (Supplementary Fig. 5c–h), suggesting NEDD4 and MDM2 may not be physiological ACE2 degrading E3 ligases in this setting. Thus, if SPOP/CK1 protects ACE2 by competing with ACE2 degrading E3 ligase(s) remains to be determined. Given the SPOP degron presenting in ACE2 N-terminal signal peptide is essential for ACE2

Pictures were taken using a Keyence BZ-X700 microscope.
SARS-CoV-2 S conditioned pseudovirus infection assays ~15,000 UMRC2 cells in 100 µL culture media were plated into each well with triplicates in 96well plates. Indicated concentrations of lenalidomide were added to cell culture 8-12 hrs post cell seeding, followed by adding 50 µL or 25 µL SARS-CoV-2 S protein conditioned pseudoviruses (Montana Molecular C1110G) as indicated in figures. The viral titer is 2 x 10 10 viral genes (VG) per milliliter (mL) from manufacturer's instructions. 24 hrs post-infection, cells were washed with sterile DPBS twice and stored in 100 µL DPBS for reading GFP signals using the BioTek Cytation 5 Cell Imaging reader.
Packaging of home-made SARS-CoV-2 S protein condition pseudoviruses and infection assays SARS-CoV-2 spike-pseudotyped HIV was generated via co-transfection of HEK293T cells with HIV clone pNl4.3-luciferase (4.5 µg) and pCAG-nCoV-S-FLAG (0.5 µg) by Lipofectamine 3000 as per manufacturer's instructions. Spike-pseudotyped HIV viral containing culture supernatant was harvested 3-days post-transfection and stored in 1 mL aliquots at -80°C. UMRC2 cells (1 x 10 5 cells/well) in 1 mL culture media were plated in 24-well plates and incubated overnight. Target cells were then pre-treated with 40 µM lenalidomide for 7 hrs. Cell medium was removed prior to spininoculation and each well received 500 µL viral containing culture supernatant and proceeded to spininoculation at 1200 x g for 2 hr at 25°C. Indicated target cell groups were treated with 40 µM lenalidomide immediately after centrifugal inoculation, 24 hr and 48 hr post centrifugal inoculation. Cells were lysed 72 hr post spininoculation with Reporter Lysis Buffer (Promega, WI, USA) and firefly luciferase activity was detected by luciferase assays (Promega, WI, USA).

Statistics Differences between control and experimental conditions were evaluated by Student's t test or
One-way ANOVA. These analyses were performed using the SPSS 11.5 Statistical Software and p < 0.05 was considered statistically significant.
DATA AND SOFTWARE AVAILABILITY All data supporting the findings in this study are available from the corresponding author upon reasonable request.  and quantification (i) of colony formation assays using control and ACE2 depleted UMRC2 cells. 600 indicated cells were plated in 6-well plates with triplicates. (j-k) Representative images (j) and quantification (k) of colony formation assays using control and ACE2 depleted UMRC6 cells. 600 indicated cells were plated in 6-well plates with triplicates.