Background
Polarized cells often act as barriers between the external environment and the underlying tissue. Due to their asymmetric plasma membranes, these cells contain distinct apical or basolateral membranes and can impose an obstacle for virus infection and spread. Viruses subvert this in a variety of ways, including disruption of the tight junctional barrier or transcytosis to gain access to the basal tissue [
1‐
5].
The outbreak of Ebola virus disease (EVD) that occurred from 2013 to 2016 in the West African countries of Guinea, Liberia, and Sierra Leone constituted a major humanitarian disaster. The outbreak numbered over 28,500 cases and 11,000 deaths [
6]. Two more outbreaks have since occurred in the Democratic Republic of Congo in 2017, and 2018. As of August 25 2018, the latest outbreak has caused 72 deaths with a total 111 cases [
7]. This highlights the fact that EBOV will remain a health threat in the near future, and development of therapeutics is urgently needed to effectively combat the virus.
Ebola virus infects a variety of polarized cells in vivo, and has been isolated from a number of tissues including the liver and gastrointestinal tract, both of which comprise of polarized cells [
8]. Gastrointestinal symptoms are among the earliest, most common, and life-threatening clinical manifestations of EVD in humans [
9]. In the 2014 outbreak in Western Africa, results of a study found that among patients admitted to the hospital with confirmed EVD, the most common clinical syndrome was one of gastrointestinal illness, intravascular volume depletion, and related complications [
10]. Owing to the difficulties in handling EBOV, knowledge of virus pathogenesis in polarized cells remains to be elucidated.
Differential availability of proteins on the cell surface can be a limiting step during the virus replication cycle. Indeed, a number of viruses induce downregulation of receptors to prevent superinfection [
11,
12]. In polarized cells, proteins can be selectively expressed on the apical or basolateral surface through specialized sorting mechanisms [
13]. Ebola virus entry is a complex and multifactorial process, and restriction of important entry factor(s) because of selective protein localization can potentially impact the efficiency of virus entry. The present study investigates the impact of polarity on EBOV infection using the colorectal adenocarcinoma (Caco-2) cell polarized model.
Methods
Cells and virus
Caco-2 cells (human epithelial adenocarcinoma cells, ATCC) were maintained in minimal essential medium (MEM; Invitrogen) supplemented with 2 or 10% fetal bovine serum (FBS) (Invitrogen). Only low passage Caco-2 cells (between passage 3 and 30) were used for seeding on transwells, and a single cell suspension was made each time to encourage formation of a monolayer. All experiments used EBOV isolate Kikwit (Ebola virus H.sapiens-rec/COD/1995/Kikwit), a widely used strain of EBOV, and were carried out at the biosafety level-4 facilities at Texas Biomedical Research Institute, San Antonio, TX or the Integrated Research Facility (IRF), National Institute of Allergy and Infectious Diseases (NIAID)/National Institutes of Health, Fort Detrick, MD.
RNA extraction and qPCR
TRIzol or TRIzol LS was added to the cell monolayer or supernatant samples in the appropriate amount and homogenized. RNA was extracted as per the manufacturer’s protocol. Primers targeting EBOV nucleoprotein (NP; F 5′- CATGCGTACCAGGGAGATTAC-3′, R 5′- ACTCCATCACGCTTCTTGAC -3′; amplicon length 80) were used to quantify EBOV vRNA in the infected cells using Verso™ 1 step RT PCR (Thermo Fisher Scientific Inc.) GAPDH was used as a reference (F 5′- CAACTCACCTCTTGGGATGAAG-3′, R 5′- CCTGGTTCAGTTTGGAGTCTATG-3′; amplicon length 90). The fold change values were calculated as described previously [
14].
SDS-PAGE and western blotting
Infected cells were harvested in RIPA lysis buffer supplemented with LDS buffer (Invitrogen) and boiled in reducing sample buffer for 10 min at 100 °C. The samples were subjected to reducing Novex 4–12% Bis-Tris gel electrophoresis. Separated proteins were electroblotted to PVDF membranes using the NOVEX Xcell Blot II module and probed using Rabbit anti-EBOV NP antibody (IBT Bioservices, Inc).
Transepithelial electrical resistance (TEER) assay
Caco-2 cells (4 × 104 cells/ well) were seeded onto 6.5-mm diameter, 1-mm pore size polycarbonate membrane transwells (Costar), and fresh medium was added at 2-day intervals. Resistance measurements were taken every other day and expressed in ohm (Ω). At day 6 post-seeding, the cells were verified to have around 100 (± 10%) Ω resistance before being used for infection. The EBOV suspension (50 μl) at a concentration of 3 pfu/cell was added either apically or basolaterally, incubated for 1 h at 37 °C, then washed three times with phosphate-buffered saline (PBS). MEM with 2% FBS medium was added, and cells were incubated at 37 °C for the required time. For infection studies, TEER measurements were taken 24 and 48 hpi.
Polarized infection
Caco-2 cells were seeded onto transwells (Costar), and fresh medium was added at 2-day intervals. At day 6 post-seeding, the cells were verified to have around 100 (± 10%) Ω resistance before being used for infection. Cell monolayers which did not have the required resistance were discarded and were not used for infection studies. EBOV suspension (50 μl) at a concentration of 3 pfu/cell was added either apically or basolaterally, incubated for 1 h at 37 °C, following which were washed three times with PBS. MEM supplemented with 2% FBS medium was added, and cells were incubated at 37 °C. Cells were harvested in TRIzol reagent and radioimmunoprecipitation assay (RIPA) buffer for RNA and protein analysis, respectively, at indicated time points, and EBOV NP vRNA was detected by quantitative reverse transcriptase (qPCR), or by western blot analysis.
Indirect immunofluorescence
Caco-2 cells were seeded into transwell inserts and infected with EBOV After infection, cells were fixed with 10% buffered formalin and processed for immunofluorescence as described with some modifications (
http://www.zonapse.net/protocols/id6.html). Cells fixed overnight were washed with PBS and incubated with immunofluorescence buffer (20 mM of HEPES, pH 7.5, 0.1% Triton-X-100, 150 mM of sodium chloride, 5 mM of EDTA, and 0.02% sodium azide as a preservative) for 5 min at room temperature (RT) and all further washes were performed with immunofluorescence buffer. Cells were then incubated with either Rabbit anti-E-cadherin antibody (Cell Signaling Technology, Inc) to visualize adherens junctions, or Mouse anti-EBOV GP antibody (IBT Bioservices, Inc) for visualizing EBOV infection overnight at 4 °C. For visualization tight junctions, the cells were fixed in methanol, and processed similarly as above. The cell monolayers were incubated with Rabbit anti-ZO-1 antibody (Cell Signaling Technology, Inc). Alexa fluor-conjugated secondary antibodies were added for 1 h at RT. Membranes were cut out using a scalpel blade, mounted on glass slides with Prolong anti-fade mounting reagent and stained with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). The glass slides were covered with cover-slips and left to dry overnight in the dark at RT. The membranes were visualized using an Eclipse Ti confocal microscope (Nikon) and NIS Elements Imaging Software.
Differential polarity assay
Caco-2 cells (4 × 104) were seeded onto 6.5-mm diameter, 1-mm pore size polycarbonate membrane transwells (Costar), and fresh medium was added at 2-day intervals. At day 4 (average resistance 36.63 Ω), day 6 (average resistance 107.32 Ω), and day 8 (average resistance 223.7 Ω) post-seeding, the cells were infected with EBOV (3 pfu/cell) either apically or basolaterally, incubated for 1 h at 37 °C, and washed three times with PBS. Then 2% FBS medium was added, and cells were incubated at 37 °C. Cells were harvested 6 hpi in TRIzol reagent for qPCR analysis.
Monolayer scratch assay
Monolayers of Caco-2 cells were gently scratched once on the apical side with a 10-μl pipette tip, followed immediately by apical addition of EBOV supernatant. Following an incubation of 1 h, the supernatant was removed, replaced with 2% FBS medium, and further incubated at 37 °C for 48 hpi. The cells were then fixed with 10% buffered formalin and analyzed using immunofluorescence assay [
15].
Ι-carrageenan assay
For the carrageenan assay, EBOV virus was pretreated with ι-carrageenan diluted in MEM without FBS supplementation for 30 min at 4 °C. Following incubation, cells were infected either apically or basolaterally with EBOV-carrageenan solution (50 μl) at a final virus concentration of 3 pfu/cell and further incubated at 37 °C for 1 h. The cells were then washed, the inoculum was replaced with MEM with 2% FBS medium, and cells were further incubated at 37 °C. At 24 hpi, the cells were harvested in TRIzol reagent. Quantification of the infection was measured by qPCR. For the binding assay, following addition of the ι-carrageenan pretreated virus, the cells were incubated for a further 30 min at 4 °C to allow attachment but not infection. Following incubation, the cells were washed with ice-cold PBS, and the cells were harvested immediately in TRIzol reagent for qPCR analysis as described earlier.
Heparin lyase assay
A stock solution of (1.0 U/μl) of HL Blend from Flavobacterium heparinum (Sigma) was prepared in sterile PBS. One hour before infection, 50 μl of 0.5 U/ well of HL in MEM without FBS was added to the culture medium (MEM with 2% FBS) and incubated at room temperature. Following treatment, cells were infected apically or basolaterally with EBOV (50 μl) at a concentration of 3 pfu/cell and incubated at 37 °C for 1 h. The cells were then washed, the inoculum was replaced with MEM with 2% FBS medium, and cells were further incubated at 37 °C. At 24 hpi, the cells were harvested in TRIzol reagent. Quantification of the infection was measured by qPCR. For the binding assay, following HL pre-treatment of Caco-2 cells, was added and incubated for 30 min at 4 °C. Following incubation, the cells were washed with ice-cold PBS and harvested in TRIzol reagent for analysis.
Statistical analysis
GraphPad Prism (version 5.0, GraphPad) software was used for statistical analysis. All data are shown as mean ± SD calculated from three independent experiments. Statistical significance was calculated using one-way ANOVA and significance was set at p < 0.05.
Discussion
Significant advances have been made in understanding EBOV infection in recent years, though studies in polarized epithelial cells have been lacking. Polarized epithelial cells establish an apical-basolateral axis with proteins localizing specifically to either the apical or basolateral membranes. We sought to determine the effect of cell polarity on EBOV infection.
The Caco-2 cell model used here has been used extensively in studies investigating virus pathogenesis as well as cellular permeability and absorption. Initially, we verified that the polarized monolayer is susceptible to EBOV infection. Further, it was found that EBOV infection efficiency is asymmetrical, with infection occurring more efficiently through the basolateral membrane. By breaking the tight junction barrier, apical infection was enhanced along the margins of the breach, indicating access to the basolateral membrane is a limiting factor during infection. Since the basolateral preference occurred as early as 6 hpi, the basolateral selection occurs early in the virus replication cycle, probably during the attachment or entry stages.
Other studies have investigated EBOV entry and attachment in the context of glycosaminoglycans (GAG). A recent report has shown that filoviruses utilize heparan sulfate proteoglycans, which are comprised of HS chains anchored to a protein core, for their attachment to host cells [
21,
26]. Further, expression of EXT1, a glycosyltransferase that is involved in the biosynthesis of heparan sulfate (HS), is required for efficient entry of the filoviruses [
27,
28]. Additionally, a competitive inhibitor of another GAG, heparin, suramin efficiently inhibited Ebola envelope-mediated gene transfer while vesicular stomatitis virus G protein pseudotyped vectors were only marginally affected [
29]. Thus, we sought to elucidate the involvement of heparan sulfate in EBOV infection of polarized Caco-2 cells. A competition assay using ι-carrageenan showed that the preferential basolateral infection in Caco-2 cells was dependent on HS and ι-carrageenan treatment selectively reduced the basolateral infection efficiency. However, though infection was reduced comparable to apical levels, it was not abrogated entirely, indicating that HS is not the sole factor influencing infection. Similarly, cells treated with HL prior to infection showed a reduction of only basolateral infection.
Aspects of HS distribution and glycosylation during Caco-2 cell polarization have been reported previously. Glypican, a heparan sulfate proteoglycan, was found to be mostly expressed at the basolateral surface, an unexpected finding for a glypiated protein. Interestingly, removal of the heparan sulfate glycanation sites from the glypican core protein resulted in the nearly exclusive apical targeting of glypican, indicating that heparan sulfate glycanation may be a determinant of the subcellular expression of glypican [
30]. Reports show that for Human cytomegalovirus, membrane-associated HS proteoglycan mediates both viral attachment and subsequent infection of Caco-2 cells. Further, the redistribution of HS is implicated in the basolateral entry of HCMV into differentiated Caco-2 cells [
31]. These results support our finding that differential distribution of HS can influence virus entry in polarized cells.
As HS is a key factor during polarized cell infection, the molecule may be a potential target for antiviral therapy. Chemical mimics can be used to competitively inhibit the initial virus attachment to the cell surface [
32]. Several strategies for prophylaxis that target HS are already being tested in other viruses including against human papillomavirus, herpes simplex virus, and influenza A virus, and a similar strategy can be explored for EBOV [
33‐
35]. Developing a topical prophylactic agent that can cover micro-abrasions on the skin may be especially useful in outbreak situations. This agent could provide an additional line of protection for healthcare workers during outbreak situations. Interestingly, GAGs are already being used to treat EVD, a report of two EVD patients exhibiting hypercoagulability were treated with heparin, a GAG analogue of HS [
36]. Though there was a possibility of heparin resistance in EVD patients, heparin administration may be of some therapeutic value as a competitive inhibitor of HS. However, hypercoagulaopathy occurs in later stages of infection, so the therapeutic window for HS-based inhibition to be effective may have already passed. More investigations are needed to see whether heparin administration at an earlier point of the disease may lead to better patient outcomes.
On a broader note, understanding the routes of infection of a virus through polarized surfaces can increase understanding of virus transmission and dissemination. In general, viruses that are transmitted through aerosols or surface contact with body fluids are generally thought to enter the epithelial barrier from the apical side, whereas virus infections due to injuries or transmission from animal bites and scratches enter epithelial cell monolayers from the basolateral side [
37]. Basolateral virus budding is thought to cause systemic infections, whereas local infections are a result of viruses that are released predominantly from the apical side.
Based on the presented data, we propose the following model for EBOV infection in the host. Since factors important for EBOV infection are segregated to the basolateral membrane in epithelial cells, the virus must first traverse the epithelial linings before it can interact with the entry factor(s). EBOV can enter through abrasions of the skin or through the mucous membrane, which have been hypothesized as the routes of transmission for EBOV [
38,
39]. The virus first infects monocytes or other early targets of EBOV infection, and systemic spread can occur through the extravasation of the infected cells into tissues. This extravasation of monocytes will give EBOV easy access to the basal membrane of cells, making them more susceptible to infection.
Though HS is ubiquitously expressed in mammalian tissues, their compositions may be tissue specific to carry out highly diverse yet specialized roles in mammalian physiology [
40,
41]. These HS mediated interactions are generally electrostatic in nature, and generally show a considerable specificity with regard to the HS structure involved [
42]. Varying distribution of HS can potentially have an impact on the cell susceptibility to the virus. Thus, different polarized cells may have a slightly different susceptibility and bias depending upon the HS distribution and thus have different outcomes of infection. Further studies are thus needed to elucidate the specificity of EBOV-HS interactions regards to glycosylation as well as structure and localization. Nevertheless, this study provides a good foundation to explore EBOV pathogenesis in polarized cells.