Rapalogs downmodulate intrinsic immunity and promote cell entry of SARS-CoV-2

Abstract

SARS-CoV-2 infection in immunocompromised individuals is associated with prolonged virus shedding and evolution of viral variants. Rapamycin and its analogs (rapalogs, including everolimus, temsirolimus, and ridaforolimus) are FDA-approved as mTOR inhibitors for the treatment of human diseases, including cancer and autoimmunity. Rapalog use is commonly associated with increased susceptibility to infection, which has been traditionally explained by impaired adaptive immunity. Here, we show that exposure to rapalogs increases susceptibility to SARS-CoV-2 infection in tissue culture and in immunologically naive rodents by antagonizing the cell-intrinsic immune response. By identifying one rapalog (ridaforolimus) that is less potent in this regard, we demonstrate that rapalogs promote Spike-mediated entry into cells by triggering the degradation of antiviral proteins IFITM2 and IFITM3 via an endolysosomal remodeling program called microautophagy. Rapalogs that increase virus entry inhibit the mTOR-mediated phosphorylation of the transcription factor TFEB, which facilitates its nuclear translocation and triggers microautophagy. In rodent models of infection, injection of rapamycin prior to and after virus exposure resulted in elevated SARS-CoV-2 replication and exacerbated viral disease, while ridaforolimus had milder effects. Overall, our findings indicate that preexisting use of certain rapalogs may elevate host susceptibility to SARS-CoV-2 infection and disease by activating lysosome-mediated suppression of intrinsic immunity.

Keywords: COVID-19; IFITM; SARS-CoV-2; TFEB; coronavirus; interferon; mTOR inhibitor; membrane fusion; microautophagy; rapalog; rapamycin.

Figure 1:. Rapamycin and its analogs share a macrolide structure but differ by the functional group present at carbon-40.

Violet and green bubbles indicate the FKBP- and mTOR-binding sites, respectively.

Figure 2:. Rapalogs promote SARS-CoV-2 infection in lung epithelial cells to different extents by counteracting the intrinsic antiviral state.

(A) A549-ACE2 were treated with or without type I interferon (250 U/mL) for 18 hours and then treated with 20 μM rapamycin (Rap), everolimus (Eve), temsirolimus (Tem), ridaforolimus (Rid), or an equivalent volume of DMSO (D) for 4 hours. HIV-CoV-2 (100 ng p24 equivalent) was added to cells and infection was measured by luciferase activity at 48 hours post-infection. Luciferase units were normalized to 100 in the DMSO condition in the absence of interferon. (B) A549-ACE2 cells from (A) were subjected to SDS-PAGE and Western blot analysis. Immunoblotting was performed with anti-IFITM2/3, anti-ACE2, and anti-actin (in that order) on the same nitrocellulose membrane. Numbers and tick marks indicate size (kilodaltons) and position of protein standards in ladder. (C) Primary HSAEC were treated with 20 μM Rap, Eve, Tem, Rid, or an equivalent volume of DMSO for 4 hours. VSV-CoV-2 (50 μL) was added to cells and infection was measured by GFP expression at 24 hours post-infection using flow cytometry. (D) A549-ACE2 were treated with varying concentrations of Eve or DMSO (equivalent to 30 μM of Eve) for 4 hours. SARS-CoV-2 (nCoV-WA1–2020; MN985325.1) was added to cells at an MOI of 0.1 and infectious titers were measured in VeroE6 cells by calculating the TCID₅₀ per mL of supernatants recovered at 24 hours post-infection. TCID₅₀ (pfu per mL) values are shown. Means and standard error were calculated from 3–4 experiments. Statistical analysis was performed with one-way ANOVA and asterisks indicate significant difference from DMSO. *, p < 0.05; **, p < 0.01. Rel.; relative. pfu; plaque forming units.

Figure 3:. Rapalogs promote SARS-CoV-2 infection in HeLa-ACE2 cells.

(A) HeLa-ACE2 were treated with varying concentrations of Eve or DMSO for 4 hours. SARS-CoV-2 (nCoV-WA1–2020; MN985325.1) was added to cells at MOI 0.1 and infectious titers were measured in VeroE6 cells by calculating the TCID₅₀ of supernatants recovered at 24 hours post-infection. TCID₅₀ (pfu per mL) values are shown. (B) HeLa-ACE2 were treated with 20 μM Rap, Eve, Tem, Rid, or an equivalent volume of DMSO for 4 hours. SARS-CoV-2 (nCoV-WA1–2020; MN985325.1) was added to cells at MOI 0.1 and infectious titers were measured in VeroE6 cells by calculating the TCID₅₀ per mL of supernatants recovered at 24 hours post-infection. TCID₅₀ per mL values were normalized to 100 in the DMSO condition. (C) HeLa-ACE2 were treated with 20 μM Rap, Eve, Tem, Rid, or an equivalent volume of DMSO for 4 hours. HIV-CoV-2 (100 ng p24 equivalent) was added to cells and infection was measured by luciferase activity at 48 hours post-infection. Luciferase units were normalized to 100 in the DMSO condition. (D) HeLa-ACE2 cells from (C) were subjected to SDS-PAGE and Western blot analysis. Immunoblotting was performed with anti-IFITM2, anti-IFITM1, anti-IFITM3, anti-ACE2, and anti-actin (in that order) on the same nitrocellulose membrane. (E) IFITM3 levels from (D) were normalized to actin levels and summarized from 5 independent experiments. (F) HeLa-ACE2 were treated with 20 μM Rap, Eve, Tem, Rid, or an equivalent volume of DMSO for 4 hours and cells were fixed, stained with DAPI and anti-IFITM2/3, and imaged by confocal immunofluorescence microscopy. Images represent stacks of 5 Z-slices and one representative image is shown per condition. Means and standard error were calculated from 3–6 experiments. Statistical analysis was performed with one-way ANOVA and asterisks indicate significant difference from DMSO. *, p < 0.05; **, p < 0.01. Rel.; relative. pfu; plaque forming units.

Figure 4:. Rapalogs promote cell entry mediated by diverse viral fusion proteins.

(A) HeLa-ACE2 were treated with 20 μM Rap, Eve, Tem, Rid, or an equivalent volume of DMSO for 4 hours. HIV-CoV-2 S pseudovirus incorporating BlaM-Vpr (HIV-BlaM-CoV-2) was added to cells for 2 hours and washed. Cells were incubated with CCF2-AM for an additional 2 hours and fixed. Cleaved CCF2 was measured by flow cytometry. Dot plots visualized as density plots from one representative experiment are shown on the left and the percentage of CCF2+ cells which exhibit CCF2 cleavage is indicated. Summary data representing the average of four experiments is shown on the right. (B) HIV-CoV-1, (C) HIV-MERS-CoV, (D) HIV-IAV HA, or (E) HIV-VSV G were added to HeLa-ACE2 or HeLa-DPP4 cells as in (A) and infection was measured by luciferase activity at 48 hours post-infection. Luciferase units were normalized to 100 in the DMSO condition. Means and standard error were calculated from 3–4 experiments. Statistical analysis was performed with one-way ANOVA and asterisks indicate significant difference from DMSO. *, p < 0.05; **, p < 0.01. Rel.; relative.

Figure 5:. Select rapalogs enhance Spike-mediated infection in HeLa-ACE2 by inhibiting IFITM2 and IFITM3.

(A) HeLa WT and HeLa IFITM1–3 KO cells were transiently transfected with 0.150 μg pcDNA3.1-hACE2 for 24 hours. Whole cell lysates were subjected to SDS-PAGE and Western blot analysis. Immunoblotting was performed with anti-IFITM2, anti-IFITM3, anti-IFITM1, anti-ACE2, and anti-actin (in that order) on the same nitrocellulose membrane. (B) HeLa IFITM1–3 KO were transfected with IFITM2 or IFITM2 and IFITM3 and SDS-PAGE and Western blot analysis was performed. (C) HIV-CoV-2 was added to transfected cells from (B) and infection was measured by luciferase activity at 48 hours post-infection. Luciferase units were normalized to 100 in HeLa WT cells treated with DMSO. (D) HeLa WT were transiently transfected with 0.150 μg pcDNA3.1-hACE2 for 24 hours. HIV-CoV-2 decorated with ancestral Spike (WA1) or Omicron Spike (BA.1) was added and infection was measured by luciferase activity at 48 hours post-infection. Luciferase units were normalized to 100 in cells treated with DMSO for both pseudoviruses. Means and standard error were calculated from 3 experiments. Statistical analysis was performed with one-way ANOVA and asterisks indicate significant difference from nearest DMSO condition. *, p < 0.05; **, p < 0.01. ns; not significant. Rel.; relative.

Figure 6:. Nuclear TFEB triggers IFITM2/3 turnover, promotes Spike-mediated infection, and is required for enhancement of infection by rapalogs.

(A) A549-ACE2 were treated with 20 μM Rap, Eve, Tem, Rid, tacrolimus (Tac), or DMSO for 4 hours and whole cell lysates were subjected to SDS-PAGE and Western blot analysis with anti-TFEB and anti-pTFEB (S211). (B) pTFEB (S211) levels were divided by total TFEB levels and summarized as an average of 3 experiments. (C) HeLa-ACE2 were transfected with TFEB-GFP for 24 hours, treated with Rap, Eve, Tem, Rid or Tac for 4 hours, stained with DAPI and CellMask (not shown), and imaged by high-content microscopy. Representative images are shown. (D) Ratio of nuclear to cytoplasmic TFEB-GFP was calculated in individual cells and average ratios derived from 9 separate fields of view (each containing 20–40 cells) are shown. (E) HeLa-ACE2 were transfected with 0.5 μg TFEBΔ30-GFP for 24 hours, fixed, stained with anti-IFITM2/3, and imaged by high-content microscopy (representative field on left). Average intensity of IFITM2/3 levels in 150 GFP-negative and 150 GFP-positive cells were grouped from two transfections (right). (F) HeLa-ACE2 were transfected (or not) with 0.5 μg TFEBΔ30-GFP, for 24 hours and HIV-CoV-2 (100 ng p24 equivalent) was added. Infection was measured by luciferase at 48 hours post-infection. Luciferase units were normalized to 100 in the non-transfected condition. (G) HeLa WT or TFEB KO were transfected with 0.3 μg pcDNA3.1-hACE2 for 24 hours and treated with 20 μM rapalogs/DMSO for 4 hours. HIV-CoV-2 (100 ng p24 equivalent) was added and luciferase activity measured at 48 hours post-infection. Luciferase units were normalized to 100 in the non-transfected condition. Means and standard error were calculated from 3 (A), 5 (F), and 3 (G) experiments. Statistical analysis was performed with one-way ANOVA or student’s T test (E and F) and asterisks indicate significant difference from DMSO or non-transfected conditions. *, p < 0.05; **, p < 0.01. Rel.; relative. A.u.; arbitrary units.

Figure 7:. Rapamycin increases susceptibility of primary human nasal epithelial cells to SARS-CoV-2 infection while limiting pro-inflammatory cytokine induction.

Primary human nasal epithelial cells (hNAEC) pooled from 12 donors were cultured at the liquid-air interface for 30–60 days were infected with 5^10⁵ plaque forming units (pfu) SARS-CoV-2 (WA1). At 24 hours and 48 hours post-infection, Trizol was added to cells and total RNA extraction was performed. RT-qPCR was performed using primers and probes specific to viral ORF1a (A), cellular IL6 (B), and cellular IFNB1 (C). Means and standard error were calculated from 2 experiments (infection of pooled cells from 12 human donors was performed in duplicate). Relative RNA levels are presented Comparative CT method with beta actin (ACTB) serving as an endogenous control. RNA levels present in the DMSO condition at 24 hours were normalized to 1. ORF1a was not detected in non-infected cells. Statistical analysis was performed using one way ANOVA. *, p < 0.05; **, p < 0.01. ns; not significant. rel.; relative.

Figure 8:. Rapamycin injection into hamsters intensifies viral disease during SARS-CoV-2 infection.

(A) Golden Syrian hamsters were injected intraperitoneally with 3 mg/kg Rap, Rid, or equivalent amounts of DMSO (4 animals per group). Four hours later, hamsters were infected intranasally with 6 × 10³ plaque forming units of SARS-CoV-2. At 2 days post-infection, half of the animals received a second injection of Rap, Rid, or DMSO. Oral swabs were taken and used for measurement of oral viral RNA load by qPCR. At 10 days post-infection (or earlier, if more than 20% weight loss or agonal breathing was detected), hamsters were euthanized, and lungs were harvested for determination of infectious virus titer by TCID₅₀ assay and IL-6 ELISA. (B) Individual body weight trajectories for each treatment group are plotted by day post-infection. Red lines indicate animals that required euthanasia for humane endpoints (more than 20% weight loss or agonal breathing). (C) Kaplan-Meier survival curves were generated according to the dates of euthanasia (or in one case, when an animal was found dead). (D) Infectious virus titers in lungs were determined by TCID₅₀ in Vero-TMPRSS2 cells. Data is depicted as floating bars (minimum, maximum, and mean shown). (E) Viral RNA copy number was determined by qPCR from oral swab at 2 days post-infection. Data is depicted as box and whiskers plots. (F) IL-6 protein levels in lungs were determined using a hamster IL-6 ELISA kit. Statistical analysis in (C) was performed by comparing survival curves between Rap and DMSO or Rid and DMSO using the Log-rank (Mantel-Cox) test. Statistical analysis in (D) was performed by comparing all individuals (survivors and euthanized) in the Rap and Rid groups using the Mann-Whitney test. Statistical analysis in (E) and (F) was performed by one way ANOVA. Illustration created with BioRender.com. *, p < 0.05; **, p < 0.01. ns; not significant.

Figure 9:. Rapamycin injection into mice downmodulates IFITM3 in lungs and boosts MA SARS-CoV-2 titers.

(A) C57BL/6 mice were injected with 3 mg/kg of Rap or an equivalent amount of DMSO (6 or 7 mice per group, respectively). The following day, mice were infected intranasally with 6 × 10⁴ TCID₅₀ mouse-adapted (MA) SARS-CoV-2. Mice received second and third injections of Rap or DMSO on the day of infection and on day 1 post-infection. (B) Lungs were harvested from infected mice upon euthanasia at day 2 post-infection and infectious viral loads were determined by TCID₅₀ (B) and IL-6 protein was measured by a mouse IL-6 ELISA kit (C). Geometric mean TCID₅₀ per gram was calculated per treatment group. Statistical analysis was performed with Mann-Whitney test and asterisks indicate significant difference from DMSO. *, p < 0.05; **, p < 0.01. (D) Lung homogenates (3 μg) from mice injected with Rap or DMSO were subjected to SDS-PAGE and Western blot analysis. Immunoblotting was performed with anti-Fragilis/IFITM3 (ab15592) and anti-actin. Illustration created with BioRender.com.

Figure 10:

Model for rapalog-mediated enhancement of SARS-CoV-2 infection. Rapamycin, everolimus, and temsirolimus potently inhibit the phosphorylation of TFEB by mTOR, while ridaforolimus is a less potent inhibitor. As a result, TFEB translocates into the nucleus and induces genes functioning in lysosomal activities, including autophagy-related pathways. Nuclear TFEB triggers a microautophagy pathway that results in accelerated degradation of membrane proteins IFITM2 and IFITM3. Loss of IFITM2/3 promotes SARS-CoV-2 entry into cells by facilitating fusion between viral membranes and cellular membranes. Illustration created with BioRender.com.