SARS-CoV-2 infection in hiPSC-derived neurons is cathepsin-dependent and causes differential accumulation of HIF1ɑ and phosphorylated tau

Abstract

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been shown to infect areas of the human brain and a subset of neurons in vitro. We have previously demonstrated that the virus enters human induced pluripotent stem cell (hiPSC)-derived neurons via an endosomal-lysosomal pathway. Here, we show that neuronal infection with both SARS-CoV-2 Wuhan and Omicron XBB.1.5 variants is dependent on cathepsins and can be blocked by an inhibitor of cathepsin B. The result was reproducible in non-transgenic hiPSC-derived cortical organoids. We further show that SARS-CoV-2 can replicate in neuron cultures, but the infectivity of the newly produced virions declined at 24 h post-infection despite a further increase in released viral RNA at later time points. The number of infected neurons decreased within five days, suggesting virus-induced neuronal cell death. The infection also caused the accumulation of the hypoxia-inducible stress factor HIF1-α in infected neurons under normoxia. Finally, expanding previous findings, in SARS-CoV-2 infected neurons, the microtubule-associated protein tau was hyperphosphorylated at multiple loci, including S202/T205, and mislocalized to the soma of infected 2D-neuronal cultures, but not in 3D-organotypic models. Hence, the neurodegenerative potential of SARS-CoV-2 infection should be carefully considered in different infection models.

Keywords: CA-004 ME; HIF-1α; MT: Clinical Applications; SARS-COV-2; antivirals; cathepsins; microglia; neurodegeneration; tau; virus entry.

Graphical abstract

Figure 1

Neuronal infection by SARS-CoV-2 is productive (A) Representative image of a neuron-astrocyte co-culture. Scale bar, 100 μm. MAP2 (microtubule-associated protein 2). GFAP (Glial fibrillary acidic protein). (B) ACE2 mRNA expression in wildtype and ACE2-transduced hiPSCs, 35-day-old neuron-astrocyte co-cultures, astrocytes and negative and positive control cell lines (A549 wildtype and A549-ACE2-TMPRSS2 (AT)), respectively. Two hiPSC-lines derived from two individuals were used for brain cell differentiation. (C) Comparison of the infection rate between co-cultures that contain control (wt neuron) or ACE2 overexpressing neurons (ACE2 neuron). Statistical significance was tested using Mann-Whitney U test. ∗∗, p ≤ 0.01. N = 6 replicate wells. (D) The effect of viral MOI and washing on infection rate. (E) A representative image of an infected neuron with intact neurites at 120 hpi. Scale bar, 300 μm. (F) Comparison of the number of infected neurons per 96-well at 24, 48, 72, and 120 hpi. Statistical significance tested using Kruskal-Wallis test with Dunn’s correction for multiple comparison. ∗, p ≤ 0.05. N = 3–14 replicate wells. (G) Viral RNA copy number per one μL of conditioned medium at 6, 12, 24, 48, 72, and 120 hpi. N = 4 replicate wells. (H) Production of viral RNA copies per hour at 6, 12, 24, 48, 72, and 120 hpi. (I) Infectivity of susceptible Vero E6 cells by conditioned medium collected from infected neuron-astrocyte co-cultures at 6, 12, 24, 48, 72 and 120 hpi. N = 4 replicate wells. (J) Representative images of Vero E6 cultures treated with conditioned medium from infected neuron-astrocyte co-cultures are 6, 24, 48, 72, and 120 hpi. Scale bar, 800 μm. Only significant results are denoted. In all the graphs, the mean is indicated and the error bars indicate the standard deviation.

Figure 2

Cathepsin inhibitors block SARS-CoV-2 infection in neurons (A) Effectiveness of 0.2 μM apilimod, 25 μM nafamostat, 3–30 μM CA-074-ME and 3–30 μM SB412515 in blocking neuronal infection by the Wuhan strain of SARS-CoV-2. For each drug treatment, values were normalized to the corresponding concentrations of vehicle (DMSO) controls. Analysis included 6 replicate wells in two independent batches. (B) Effectiveness of 0.2 μM apilimod, 25 μM nafamostat, 3–30 μM CA-074-ME and 3–30 μM SB412515 in blocking neuronal infection by the Omicron XBB.1.5 strain of SARS-CoV-2. N = 6 replicate wells. (C) Representative images of data shown at A. Scale bars 600 μm (main) and 100 μm (zoom-in). (D) mRNA expression of cathepsin B (CTSB) and L (CTSL) in hiPSC-derived neurons, astrocytes, microglia, and neuron-astrocyte co-cultures. Data shown as fold change to the housekeeping gene GAPDH. (E) Effectivity of 25 μM nafamostat, 0.5 μM apilimod, 3–22.5 μM SB41251, and 3–30 μM CA-074-ME in blocking SARS-CoV-2 infection in Vero E6 cells, (F) and A549-AT cells. The data are displayed as fold-change to DMSO (reference level). N = 6 replicate wells. (G) Cytotoxicity assay of 3–22.5 μM SB41251 and 3–30 μM CA-074-ME in Vero E6 and A549-AT cell lines. DMSO was used as negative and 9 μM UCN-01 as a positive control. N = 8 replicate wells. All statistical significances were assessed using Kruskal-Wallis test with Dunn’s correction for multiple comparison. All comparisons were made to the SARS-CoV-2 infected DMSO control. Only significant results are denoted. ∗, p ≤ 0.05; ∗∗, p ≤ 0.01; ∗∗∗, p ≤ 0.001; ∗∗∗∗, p ≤ 0.0001. In all the graphs, the mean is indicated and the error bars are the standard deviation.

Figure 3

Cathepsin-B inhibitor CA-074-ME protects choroid plaxus cortical organoids from SARS-CoV-2 infection (A) Representative images of untreated and SARS-CoV-2-infected cortical organoids with choroid plexus (arrowheads). Scale bars, 300 μm. (B) Quantification of the percentage of infected neurons at 24, 48, 72, and 120 hpi. (C) Close-up of SARS-CoV-2 infected neurons (MAP2+). An infected cell marked with arrowheads. (D) Quantification of the percentage of SARS-CoV-2 infected neurons in infected organoids with (10 and 30 μM CA-074-ME) and without (DMSO) drug-treatment at 72 hpi. (E) Representative images of SARS-CoV-2 positive cells in infected organoids with (10 and 30 μM CA-074-ME) and without (DMSO) drug-treatment at 72 hpi. Arrowheads mark SARS-CoV-2 infected cells. Scale bar, 500 μm. (F) Representative images of cleaved caspase-3 positive neurons with and without drug treatment. Arrowheads correspond to figure. Scale bar, 500 μm. D. (G) Quantification of the % of cleaved caspase-3 positive neurons with and without drug treatment. All statistical significances were assessed using Kruskal-Wallis test with Dunn’s correction for multiple comparison. N = 3 organoids. Only significant results are denoted. ∗, p ≤ 0.05; ∗∗, p ≤ 0.01. In all the graphs, the mean is indicated and the error bars are the standard deviation.

Figure 4

Stress and immune responses in 2D neuron-astrocytes co-cultures infected with SARS-CoV-2 (A, D, and E) Mean intensity of AT8 antibody signal in untreated, treated but noninfected, and infected neurons at 24, 48, and 72 hpi. N = 2–6 replicate wells. (B and C) Representative images of AT8 staining at 24 hpi. (F) Mean intensity of anti-S396 signal in untreated, treated but noninfected, and infected neurons at 24 hpi. (G) Representative images of anti-S396 staining at 24 hpi. N = 3 replicate wells. (H and I) Mean intensity of anti-HIF-1α signal in untreated, treated but noninfected, and infected neurons at 24 and 72 hpi. N = 3 replicate wells in one batch. (J) Representative images of anti-HIF-1α staining at 24 hpi. (K) Secreted cytokine profile of SARS-CoV-2 infected thick neuron-astrocyte co-cultures and untreated controls at 48 hpi. Fold-changes to control are shown next to the four main hits. N = one sample pooled from two replicate wells. (L) Annotated image of the membrane used to detect secreted cytokine profile of a SARS-CoV-2 infected neuron-astrocyte co-culture and positive control. All statistical significances were tested using Kruskal-Wallis test with Dunn’s correction for multiple testing. Only significant results are denoted. ∗ ≤0.05, ∗∗p ≤ 0.01. Scale bars: 100 μm. In all the graphs, the mean is indicated and the error bars are the standard deviation.

Figure 5

Microglia cells are not pruductivelly infected by SARS-CoV-2 in 2D co-cultures and do not prevent Tau phosphorulation in infected Neurons (A) Representative image of anti-Iba-1 (microglia) and anti-MAP2 (neuron) staining of an untreated neuron-astrocyte-microglia co-culture. (B) Same image but the Iba-1 staining has been thresholded with ImageJ into binary format to better display the microglial morphology. (C) Similar thresholded image of SARS-CoV-2 infected culture. (D) Representative image of a SARS-CoV-2 infected neuron (stained with anti-N protein, arrowhead) among Iba-1 positive microglia in triplecultures. (E) Mean intensity of AT8 fluorescent signal in untreated, treated but noninfected, and infected neurons at 48 hpi. N = 6 replicate wells. (F) and at 72 hpi. N = 3 replicate wells. Scale bars, 100 μm. All statistical significances were tested using Kruskal-Wallis test with Dunn’s correction for multiple testing. Only significant results are denoted. ∗, p ≤ 0.005.