COVID-19 Diarrhea Is Inflammatory, Caused by Direct Viral Effects Plus Major Role of Virus-induced Cytokines

Abstract

Background & aims: Diarrhea occurs in up to 50% of cases of COVID-19. Nonetheless, the pathophysiologic mechanism(s) have not been determined.

Methods: This was examined using normal human enteroid monolayers exposed apically to live SARS-CoV-2 or non-replicating virus-like particles (VLPs) bearing the 4 SARS-CoV-2 structural proteins or irradiated virus, all of which bound and entered enterocytes.

Results: Live virus and VLPs incrieased secretion of multiple cytokines and reduced mRNAs of ACE2, NHE3, and DRA. Interleukin (IL)-6 plus IL-8 alone reduced NHE3 mRNA and protein and DRA mRNA and protein. Neither VLPs nor IL-6 plus IL-8 alone altered Cl^- secretion, but together they caused Cl^- secretion, which was Ca²⁺-dependent, CFTR-independent, blocked partially by a specific TMEM16A inhibitor, and entirely by a general TMEM16 family inhibitor. VLPs and irradiated virus, but not IL-6 plus IL-8, produced Ca²⁺ waves that began within minutes of VLP exposure, lasted for at least 60 minutes, and were prevented by pretreatment with apyrase, a P2Y1 receptor antagonist, and general TMEM16 family inhibitor but not by the specific TMEM16A inhibitor.

Conclusions: The pathophysiology of COVID-19 diarrhea appears to be a unique example of a calcium-dependent inflammatory diarrhea that is caused by direct viral effects plus the virus-induced intestinal epithelial cytokine secretion.

Keywords: CaCC; Cytokines; DRA; Diarrhea; NHE3; SARS-CoV-2; Virus-like Particles.

Graphical abstract

Figure 1

Live SARS-CoV-2 enters and replicates in human enteroid monolayers. (A) The apical membrane of differentiated human small intestinal and proximal colonic enteroid monolayers was exposed to live SARS-CoV-2 10⁶ PFUs/ml for 90 minutes, then washed to remove virus, and studied for up to 5 days after viral exposure. Apical and BL media from differentiated duodenal or colonoid monolayers and lysates of enterocytes were studied. Total RNA was prepared and expression of viral nucleocapsid was measured by qRT-PCR. The concentration of VPs corresponding to RNA is shown. n = 3. (B) Differentiated human colonoid monolayers were treated as in (A), and 48 hours later, viral entry was confirmed in colonocytes, using double stranded RNA (dsRNA) antibody (red) and viral Spike protein (green). This did not occur in goblet cells. (C) Spike protein and dsRNA were not detected in uninfected monolayers as a negative control for (B). (D) Infectious virus was detected in colonoid monolayers as well as in apical media by Vero cell plaque forming assay under conditions in (A). Y axis, PFU/ml is for live virus in apical media; PFU/layer is for the cells in monolayer. n = 3. (E) Viral infection of differentiated colonoid monolayers did not affect epithelial permeability of fluorescent 10 or 70 kDa dextran when studied 2 days after initial viral exposure. Dextran was added apically, and subsequent BL collection is shown with results normalized to the amount of dextran in untreated conditions set as 100% in each experiment. n = 3. All results are means ± SEM.

Figure 2

SARS-CoV-2 stimulates epithelial cytokine/ chemokine secretion in human duodenal enteroid monolayers. Differentiated human enteroid monolayers were exposed on the apical surface to SARS-CoV-2 (10⁶ PFU/ml) for 90 minutes; the virus was then washed out and evaluated 48 hours after initial exposure. Some studies, as shown, were performed in the presence of IL-6 and IL-8 (50 ng/ml each) on the basolateral surfaces for 48 hours. BL media were collected and analyzed for secreted cytokines/chemokines by multiplex enzyme-linked immunosorbent assay (ELISA). Live virus significantly increased IL-6, MCP-1, and IL-1b. Two separate experiments with triplicate assays in each. Results are means ± SEM.

Figure 3

Live SARS-CoV-2 reduces NHE3 and DRA but not CFTR mRNAs and NHE3and DRAbut not CFTR protein expression. (A) Effect of SARS-CoV-2 infection on the mRNA expression of ACE2, NHE3, DRA, and CFTR human colonoid monolayers. Live SARS-Cov-2 virus was apically exposed for 90 minutes (10⁶ PFU/ml) to proximal colonoid monolayers from 3 different donors and studied at 48 hours. There was reduced mRNAs for ACE2, NHE3, and DRA, but not CFTR. Results are means ± SEM; n = 3–5. Normalization was to S18 ribosomal protein mRNA. (B) Effect of SARS-CoV-2 infection on the protein expression of NHE3, DRA, and CFTR in human ileal enteroid monolayers studied as in (A) above. Results are means ± SEM. In (B), data shown as ratio of paired virus/control monolayers due to wide variation among NHE3, DRA, and CFTR protein/GAPDH ratios. n = 3–5 monolayers. P values are comparison with untreated control enteroids; paired t-tests.

Figure 4

SARS-CoV-2 VLPs directly bind and stimulate cytokine/chemokine secretion in human ileal enteroid monolayers and reduce mRNAs for ACE2, NHE3, DRA, and CFTR. (A) IF confocal microscopy (XZ plane shown) demonstrating binding of 10⁶ particles/ml VLPs were added at times 0 and 24 hours, and monolayers fixed at 48 hours. Exo = empty exosomes; green = anti-rabbit polyclonal antibodies to SARS-COV-2 spike protein; white = phalloidin (actin); blue = Hoechst 33342 (nuclei). Scale bar, 10 um. Similar results of experiments repeated twice. (B) Differentiated human ileal enteroid monolayers were apically exposed to empty exosomes (Exo) or 10⁶ particles/ml SARS-CoV-2 VLPs, added as in (A) at times 0 and 24 hours and sampled at 48 hoours. BL media were collected and analyzed for secreted cytokines/chemokines by multiplex enzyme-linked immunosorbent assay (ELISA). (C) VLPs expressing the 4 SARS-CoV-2 structural proteins, including Spike D614G, were exposed apically to ileal enteroid monolayers as in (B) and studied at 48 hours, with IL-6 and IL-8 (50 ng/ml each) on the BL surface. Effects on mRNAs demonstrated that, compared with exosomes similarly exposed, there was significant reduction in ACE2, NHE3, DRA, and CFTR. Data shown are normalized to effect of empty exosome controls set as 1.0 for each experiment (horizontal line). Results are means ± SEM. n = 3.

Figure 5

VLPs plus IL-6 and IL-8 cause active anion secretion in human ileal enteroids. Ussing chamber/voltage clamp studies of active electrogenic ion transport were performed in the presence of IL-6 plus IL-8 (both 50 ng/ml added basolaterally ∼18 before study). Arrows show time of VLP addition. Three × 10⁸ VLPs/IL-6, IL-8 stimulated an increase in Isc which began 4 to 20 minutes after VLP addition. The increase in Isc occurred in 21 of 23 monolayers exposed to the VLPs. The increase in Isc (A) was prevented by pretreatment by BAPTA-AM (35 uM, 30-minute pretreatment); (B) Reversed by BL exposure to the K channel inhibitor clotrimazole (30 uM); (C and D) Not affected by the CFTR inhibitor BPO-27 (1 μM) but reversed by the CaCC inhibitor- Ao1 (25 μM); and (E) Partially reversed by the specific TMEM16A inhibitor- Ao1 (20 uM). All studies shown are of single monolayers. Figure 6 shows these studies as means ± SEM of a series of experiments. (F) Representative trace of 5 similar experiments showing lack of effect on Isc of VLP 3 × 10⁶ added as in (A–E) but without pretreatment with IL-6 plus IL-8.

Figure 6

VLPs plus IL-6 and IL-8 cause active anion secretion in human ileal enteroids. Effects of apical VLPs/IL-6, IL-8 on ileal Isc: (A) BAPTA-AM; (B) Clotrimazole; (C) BPO-27; (C and D) CaCC inh-Ao1; (E) Comparison of Isc increase caused by VLP/IL-6, IL-8 in normal ileal enteroids (name of ileal enteroid is 46i) and human homozygous FD508 ileal enteroids (CF ileal enteroid CF-U3); (F) Effect of TMEM16A_inh-Ao1 and general CaCC_inh- Ao1 on increase in Isc caused by VLP/IL-6, IL-8 on normal ileal enteroids and human homozygous FD508 ileal enteroids. Results are means ± SEM; n = 3–7.

Figure 7

Irradiated virus is taken up in human colonic enteroid monolayers but does not alter enterocyte cytokine secretion or mRNAs of NHE3, DRA, CFTR, TMEM16F, and TMEM16, whereas IL-6 plus IL-8 reduce mRNAs of NHE3 and DRA. (A) 10⁶ viral particles/ml of irradiated virus were exposed apically to differentiated human enteroid monolayers for various times and intracellular virus identified by IB using anti-Spike antibodies. Representative study, repeated 4 times, is shown. Uptaken Spike is present at 2 and 4 but not 24 hours after exposure. (B) Irradiated virus exposure for 48 hours did not alter cytokines/chemokine secretion from differentiated ileal enteroid monolyes. 10⁶ PFU/ml irradiated virus was added apically at times 0 and 24 hours to ileal enteroid monolayers and BL media sampled at 48 hours for cytokines and chemokine assays. In parallel, enteroids that were not treated were used as controls. Data shown set the untreated controls as 1 for each experiment. Two separate ileal enteroid lines were studied. Results are means ± SEM; n = 6. (C) IL-6/IL-8 but not irradiated virus decreased mRNAs of NHE3 and DRA but not CFTR, TMEM16A, or TMEM16F. Differentiated ileal enteroid monolayers from 3 separate ileal enteroid cultures were exposed on the apical surface to 10⁶ PFU/ml irradiated virus, the combination of irradiated virus with IL-6/IL-8 (50 nl/ml each) on the BL surface, or the IL-6/IL-8 alone. Irradiated virus was added at times 0 and 24 hours and IL-6/IL-8 at time 0 and all sampled at 48 hours for mRNA determinations. Results are means ± SEM; n = 4–9. ANOVA used to calculate P values shown above bars. Bracketed P values compare IL-6/IL-8 effect to combination of irradiated virus plus IL-6/IL-8. (D) Irradiated virus exposure for 48 hours did not alter NHE3 protein expression (left), but IL-6 plus IL-8 for this time reduced NHE3 (right). Three × 10⁶/ml irradiated virus was exposed apically at times 0 and 24 hours to ileal enteroid monolayers to IL-6/IL-8 (50 ng/ml each) added at time 0 to the BL surface, and 48 after initial exposure, the entroids were lysed for IB or fixed for IF study of NHE3. Irradiated virus alone did not alter NHE3 protein expression by IB (left). Right: representative confocal XY images at the level of the the BB (Ei, Eiii) and XZ sections (Eii, Eiv). Ei, Eii, untreated controls; Eiii, Eiv, irradiated virus. NHE3, green; F-actin, red; nuclei, blue. IL-6/IL-8 greatly reduced total NHE3 protein expression. Scale bar, 20 um. Experiment on left was done once, and those on right were repeated 3 times.

Figure 8

Irradiated virus but not IL-6 plus IL-8 induce Ca²⁺waves in ileal enteroids. (A) IF image of undifferentiated ileal enteroids stably expressing GCamP-6s. This image at excitation 488 nm and emission 500–580 nm was after exposure to 1 uM ionomycin. Scale bar, 5 um. (B) Irradiated virus stimulation of ileal Ca²⁺ waves (total Ca²⁺ wave signal). The total pixel counts in these waves is shown. Studies were also performed with pretreatment with IL-6 plus IL-8 (both 50 ng/ml) 18 hours before study, after pretreatment with apyrase, BPTU, the general CaCC inhibitor-Ao1, or the specific TMEM16A inhibitor-Ao1. ANOVA was used to calculate P values comparing total Ca²⁺ wave signal to results with irradiated virus. n = 2–10. (C) Irradiated virus stimulation of ileal Ca²⁺ waves (total number of waves). Studies were performed with pretreatment with IL-6 plus IL-8 18 hours before study, after pretreatment with apyrase, BPTU, the general CaCC inhibitor-Ao1, or the specific TMEM16A inhibitor-Ao1. ANOVA was used to calculate P values comparing total Ca²⁺ wave signal to results with irradiated virus. n = 2–10. (D) Ca²⁺ waves in ileal enteroid monolayers illustrated as changes in F-Fo/Fo over 60 minutes comparing effect of irradiated virus with and without pretreatment 18 hours earlier with IL-6 plus IL-8 (both 50 ng/nl). Single experiments from those shown in Figure 8 (B and C).

Figure 9

VLP stimulation of ileal Ca²⁺waves. (A) Ca²⁺ waves (total signal in pixels) induced by VLPs and effects on the signal by pretreatment with apyrase, BPTU, the general CaCC inhibitor-Ao1, or the specific TMEM16A inhibitor-Ao1 with exosome controls. ANOVA used for statistical evaluation. n = 3–9. (B) Ca²⁺ waves (total number of waves) induced by VLPs and effects on the signal by pretreatment with BPTU, CaCC inhibitor-Ao1, or TMEM16A inhibitor-Ao1 with exosome controls. ANOVA used for statistical evaluation. n = 3–9. (C) Pseudocolored images of single field of human ileal enteroid expressing GCamP-6s showing areas involved in Ca²⁺ waves over 60 minutes. Different colors were used for each Ca²⁺ wave. If multiple Ca²⁺ waves occurred in the same area, only a single color is shown. (i) Irradiated virus. (ii) Exosomes. (iii) VLPs. (iv) VLPs after TMEM16inh-Ao1. Results from a single monolayer are shown using data from Figure 9 and (C). To generate the pseudocolored images, all pixels from within the areas of interest, representing the extent of Ca²⁺ waves, were binarized (value 1), then the MetaMorph Stack Topographic Surface Map command was used to generate the figure, which was colorized using MetaMorph Rainbow lookup tables (blue, low; green, middle; red, high). Each of the 1800 time points had a grayscale value (0–255) assigned, then the 8-bit color image was converted to 24-bit color for illustration.