RESEARCH Open Access Establishment of feline intestinal epithelial cell cultures for the propagation and study of feline enteric coronaviruses Lowiese MB Desmarets Sebastiaan Theuns Dominique AJ Olyslaegers Annelike Dedeurwaerder Ben L Vermeulen Inge DM Roukaerts and Hans J Nauwynck Abstract Feline infectious peritonitis FIP is the most feared infectious cause of death in cats induced by feline infectious peritonitis virus FIPV This coronavirus is a virulent mutant of the harmless ubiquitous feline enteric coronavirus FECV To date feline coronavirus FCoV research has been hampered by the lack of susceptible cell lines for the propagation of serotype I FCoVs In this study long term feline intestinal epithelial cell cultures were established from primary ileocytes and colonocytes by simian virus 40 SV40 T antigen and human Telomerase Reverse Transcriptase hTERT induced immortalization Subsequently these cultures were evaluated for their usability in FCoV research Firstly the replication capacity of the serotype II strains WSU 79 1683 and WSU 79 1146 was studied in the continuous cultures as was done for the primary cultures In accordance with the results obtained in primary cultures FCoV WSU 79 1683 still replicated significantly more efficient compared to FCoV WSU 79 1146 in both continuous cultures In addition the cultures were inoculated with faecal suspensions from healthy cats and with faecal or tissue suspensions from FIP cats The cultures were susceptible to infection with different serotype I enteric strains and two of these strains were further propagated No infection was seen in cultures inoculated with FIPV tissue homogenates In conclusion a new reliable model for FCoV investigation and growth of enteric field strains was established In contrast to FIPV strains FECVs showed a clear tropism for intestinal epithelial cells giving an explanation for the observation that FECV is the main pathotype circulating among cats Introduction Feline coronaviruses FCoVs are associated with both en teric and systemic diseases in domestic and wild Felidae The feline enteric coronavirus FECV is an ubiquitous enteropathogenic virus replicating in epithelial cells of both small and large intestine after oral uptake 1 5 The mild enteritis caused by this replication is usually un apparent or is manifested by a transient diarrhoea in young kittens 3 Around 13 of all infected cats are not able to clear the virus 6 In these cats the virus persists for several months or even years in the epithelium of the large intestine 2 5 7 Since FECVs are easily trans mitted from cat to cat by faecal oral route they are enzootic among most cat populations 3 8 Although FECV infections manifest subclinically they may be the start of a lethal outcome During replication mutations can occur in the viral genome providing the virus with tools to productively replicate in monocytes macrophages 9 12 This mutational variant designated feline infec tious peritonitis virus FIPV causes a chronic and highly fatal systemic disease FIP characterized by a dif fuse pyogranulomatous peri phlebitis and serositis in presence wet form or absence dry form of fibrinous exudate in the affected body cavities 13 15 In contrast to FECV which is highly infectious but seldom causes disease FIPV shows a low infectivity but high mortality 95 100 16 Losses from FIP are typically unpredict able and occur in only a restricted fraction 90 in the colon cultures was still of epithelial origin For the ileum cultures the vimentin positive mes enchymal cells had expanded in between the epithelial clusters occupying around 50 of the wells Remarkably Figure 1 Morphological features and immunocytochemical characterization of the primary ileum A C and colon D F cultures A Epithelial cells were isolated in cell clusters B D Polygonal cells started to spread from these clusters giving rise to several foci of cells E Sub confluent layers were reached 3 4 days after seeding due to a restricted proliferation of the cells C F Double immunostainings against cytokeratin red and vimentin green filaments 4 days after isolation confirming the epithelial nature of the polygonal cells Figure 2 Kinetics of FCoV replication in primary ileum and colon cultures from 3 conventional cats Cells were inoculated with FCoV WSU 79 1683 or FCoV WSU 79 1146 at a moi 1 At different time points post inoculation cytoplasmically expressed viral proteins were visualized and the percentage of infected epithelial cells was determined Desmarets et al Veterinary Research 2013 44 71 Page 6 of 13 http www veterinaryresearch org content 44 1 71 some of the ileum epithelial cells did also express vimentin resembling dedifferentiated epithelial cells typ ically found after injury or in tumours Figure 1C Expression kinetics of viral antigens in FCoV WSU 79 1683 and WSU 79 1146 infected primary ileocytes and colonocytes Primary ileocytes and colonocytes were susceptible to in fection with both serotype II FCoV strains However the antigen expression kinetics differed greatly between the avirulent FCoV WSU 79 1683 and the virulent FCoV WSU 79 1146 Figure 2 For both strains the first antigen positive cells appeared at 6 h pi and increased further over time However the avirulent enterotropic WSU 79 1683 strain infected the cells significantly more effi cient P 0 05 for both ileum and colon compared to WSU 79 1146 Morphological features and characterization of the established continuous ileocyte and colonocyte cultures By introducing a combinational expression of SV40 large T antigen and hTERT a successfully transformed cell line was generated for both ileocytes and colonocytes Figures 3 and 4 Indeed a various number of the trans duced cells started to proliferate from 1 week after trans duction onwards forming layers of cobblestone like cells with a cell diameter of 20 25 mand30 35 mfor ileocytes and colonocytes respectively Both SV40 large T antigen as hTERT expression was detected in these cultures confirming the success of transduction These cell lines could be further expanded and passaged for over 30 passages now which is in sharp contrast to the primary cultures Besides its typical cobblestone like appearance the epithelial character was confirmed by the expression of cytokeratin and dome formation in the cultures The latter is indicative for the polarization of cells in monolayers Remarkably most of the cells in both cultures co expressed both cytokeratin and vimentin in the freshly formed monolayers suggesting a more dedifferentiated state of the cells For further charac terization APN expression in the cultures was investi gated since APN is an intestinal brush border associated hydrolase and moreover an important receptor for serotype Figure 3 Morphological and immunocytochemical characterization of the continuous ileocyte cultures A Proliferating isles B Cobblestone morphology of the monolayer C Dome formation D Double immunostaining against cytokeratin red and vimentin green filaments Desmarets et al Veterinary Research 2013 44 71 Page 7 of 13 http www veterinaryresearch org content 44 1 71 II FCoVs All cells expressed APN at their surface However the expression levels varied great from cell to cell in both cultures most probably due to different differentiation levels of the cells in culture Antigen expression kinetics of FCoV WSU 79 1683 and WSU 79 1146 in continuous ileocyte and colonocyte cultures Since the continuous cultures seemed to be less differen tiated compared to the primary cultures the reliability of the established cell lines as model for intestinal epi thelial cells was further investigated Therefore antigen expression kinetics were assessed in both continuous ileocyte and colonocyte cultures as was done for the pri mary cells Figure 5 In accordance with the results obtained for the primary cultures FCoV WSU 79 1683 significantly infected both ileocytes as colonocytes more efficiently than WSU 79 1146 At 24 h pi FCoV WSU 79 1683 had infected 19 46 4 37 and 18 47 4 61 of the colonocytes and ileocytes respectively whereas only 0 03 0 02 of the colonocytes and 0 22 0 18 of the ileocytes were infected by FCoV WSU 79 1146 at that time point Titration of field strains in faecal and tissue suspensions A major restriction in FCoV research is the lack of cell lines supporting the growth of serotype I enteric strains Therefore the newly established cell lines were further validated by investigating their susceptibility for different field strains All those strains were serotype I viruses as confirmed by PCR Table 1 gives the results obtained by titration of different faecal and tissue suspensions on colonocyte cultures Comparable results were obtained by titration on ileocyte cultures with FECV UCD Hence titration of other field strains was not repeated on this cell line All but two of the samples collected from healthy cats were positive for coronavirus with qPCR titres ranging from 10 4 18 to 10 9 06 viral copies g faeces Infectious virus was detected by IPMA in 50 of all posi tive samples 8 16 with 57 of positivity in samples with qPCR titres above 10 5 This number increased to 64 7 11 and 80 4 5 when the cutoff was made at qPCR Figure 4 Morphological and immunocytochemical characterization of the continuous colonocyte cultures A Proliferating isles B Cobblestone morphology of the monolayer C Dome formation D Double immunostaining against cytokeratin red and vimentin green filaments Desmarets et al Veterinary Research 2013 44 71 Page 8 of 13 http www veterinaryresearch org content 44 1 71 titres above 10 6 and 10 7 viral copies g faeces respect ively In the one sample UG FH9 with a qPCR above 10 7 that was negative on IPMA enterotropic virus was detected by immunofluorescence staining All but one of the samples collected from FIP cats were positive for coronavirus on qPCR with the number of viral copies g ranging from 10 3 98 to 10 9 16 As determined by both IPMA and immunofluorescence staining none of those samples except for one contained enterotropic virus However 3 tissue samples UG TF5 UG TF9 and UG TF17 did contain infectious virus as determined on monocyte derived macrophages Despite its high viral load no infectious virus neither on enterocytes nor on monocytes macrophages was found in faecal suspensions of FIP cat 1 UG FF1 Faeces of FIP cat 2 UG FF2 did contain enterotropic virus that was not infectious for macrophages Propagation and titration of FECV UCD and UG FH8 To date no serotype I enteric field strains have been prop agated in vitro and availability of such FECVstrains would be valuable in feline coronavirus research Therefore two faecal strains FECV UCD and UG FH8 were further propagated in colonocyte cultures Table 2 After 3 passages both strains were raised in titre with around 3log 10 TCID 50 mL In addition ORF3 and ORF7 from each of the third passage strains were sequenced to check for signs of cell culture adaption Both strains still carried intact accessory genes that were 100 identical to the original strain Typically a lot more CPE was noticed in UG FH8 infected wells compared to FECV UCD Figure 6 After 3 passages both strains still showed a specific enterotropism since no infection was seen after inoculation of other feline cell lines fcwf and CrFK cells Discussion In this study immortalized cultures of both small ileum and large colon intestinal epithelial cells were established and validated for their use in feline coronavirus research Intestinal epithelial cells are important target cells in FCoV pathogenesis but to date such cell lines are not available The establishment of primary intestinal epithe lial cell cultures has been proven to be difficult because of the induction of programmed cell death after disrup tion from the extracellular matrix the uncontrolled contamination with stromal cells and the still unknown homeostatic components needed for the maintenance of these cultures 38 To avoid induction of apoptotic signals by disrupting cell matrix adhesions a combin ation of collagenase and dispase was used in this study to digest the mucosa allowing the isolation of epithelial cell clusters These were subsequently separated as much as possible from the contaminating single stromal cells by D sorbitol density centrifugation The primary colon cultures showed a relative high purity of epithelial cells whereas primary ileum cultures were much more con taminated with stromal cells The contamination with mesenchymal cells is intrinsic to the isolation method used and therefore inevitable Yet the epithelial cells Figure 5 Kinetics of FCoV replication in continuous ileocyte and colonocyte cultures Cells were inoculated with FCoV WSU 79 1683 or FCoV WSU 79 1146 at a moi 1 At different time points post inoculation the percentage of infected cells was determined Data are expressed as the means standard deviation of the results of 3 separate experiments Desmarets et al Veterinary Research 2013 44 71 Page 9 of 13 http www veterinaryresearch org content 44 1 71 could be cultured for a week without overgrowth by these cells making both primary cultures ideal models for studying interactions with enterotropic infectious agents Remarkably some primary cells co expressed cytokeratin and vimentin filaments which is often found in injured epithelial cells tumours and in primary cultures due to the detachment of the cells from their natural en vironment during isolation In these cells the epithelial differentiation is turned back to a more embryonic state amongst others characterized by de novo expression of vimentin filaments 39 Only a minority of the cells did express vimentin suggesting that most cells were able to restore their differentiation with the used culture conditions Although the doubtful origin and clear signs of cell culture adaptation 17 40 FCoV WSU 79 1683 and FCoV WSU 79 1146 were the only available strains representing an avirulent and related virulent strain at that time of the study Hence those strains were initially used for investigating the susceptibility of primary enterocytes to both virulent and avirulent FCoVs Replication of both strains have been studied in CrFK cells fcwf cells peri toneal macrophages bone marrow derived macrophages and peripheral blood monocytes 9 10 12 29 In contrast to the available continuous cultures CrFK and fcwf cells the difference in virulence between both strains was reflected in vitro when using primary FIPV target cells monocytes macrophages The highly efficient and mostly sustained infection of FIPV in macrophages and monocytes from susceptible cats in contrast to an inef ficient and not sustained infection of the avirulent WSU 79 1683 in those cells may explain why FIPV behaves as a harmful invasive virus causing this progressive Table 1 QPCR and infectious titre of different faecal and tissue suspensions from healthy and FIP cats Sample Source QPCR titre Log 10 copies g Infectious titre Log 10 TCID 50 g UG FH1 Faeces healthy cats 6 03 UG FH2 Faeces healthy cats 6 64 2 67 UG FH3 Faeces healthy cats 5 51 UG FH4 Faeces healthy cats 5 41 2 36 UG FH5 Faeces healthy cats 7 22 2 50 UG FH6 Faeces healthy cat 6 88 UG FH7 Faeces healthy cat UG FH8 Faeces healthy cat 6 30 3 33 UG FH9 Faeces healthy cats 7 69 UG FH10 Faeces healthy cat 7 89 2 50 UG FH11 Faeces healthy cats 8 44 2 67 UG FH12 Faeces healthy cats 4 66 UG FH13 Faeces healthy cats UG FH14 Faeces healthy cats 6 27 UG FH15 Faeces healthy cat 6 62 2 50 UG FH16 Faeces healthy cats 4 18 FECV UCD Faeces healthy cat 6d pi 9 06 5 00 UG FF1 Faeces FIP cat 1 7 57 UG FF2 Faeces FIP cat 2 9 16 3 50 UG FF3 Faeces FIP cat 3 UG FF4 Faeces FIP cat 4 3 98 UG TF2 Kidney FIP cat 1 6 79 UG TF5 Omentum FIP cat 2 6 87 UG TF9 Spleen FIP cat 3 5 83 UG TF17 Omentum FIP cat 4 8 00 Table 2 Infectious titre and status of ORF3 and ORF7 in cell culture propagated viruses Strain Infectious titre Log 10 TCID 50 mL Status ORF3 at P 3 Status ORF7 at P 3 P 0 P 3 FECV UCD 3 97 6 30 Intact Intact UG FH8 2 63 5 97 Intact Intact Desmarets et al Veterinary Research 2013 44 71 Page 10 of 13 http www veterinaryresearch org content 44 1 71 systemic disease 9 10 12 As was previously shown for monocytes macrophages the present study confirms that both strains exhibit clear differences in cell tropism In contrast to FCoV WSU 79 1146 the avirulent WSU 79 1683 efficiently infected and replicated in intestinal epithelial cells resulting in exactly opposing kinetics as were found for macrophages 9 Primary cultures are ideal tools to reliably investigate virus host interactions Nevertheless isolation of primary epithelial cells is labour intensive the cultures are often contaminated with a various amount of mesenchymal cells and the yield is variable and rather low To activate re search with those cells long term cultures were derived from both primary ileocytes and colonocytes by SV40 T antigen and hTERT induced immortalization resulting in the generation of two feline intestinal epithelial cell cultures The epithelial nature of both cell lines was con firmed by their cobblestone morphology dome formation and cytokeratin expression These newly established cell lines could be valuable tools for virus research How ever immortalized cell lines are often phenotypically transformed making reliable research with these cells questionable In the present study it was shown that in contrast to the primary cultures the majority of the cells co expressed cytokeratin and vimentin filaments suggesting that the cultures were less differentiated compared to their primary counterparts Therefore the reliability of the established cell lines for their use in feline corona virus research was further investigated and confirmed Antigen expression kinetics of FCoV WSU 79 1683 and FCoV WSU 79 1146 were comparable with the results obtained with the primary cultures showing a signifi cant difference in cell tropism between both strains As mentioned before comparative studies in the available continuous feline cell lines CrFK and Fcwf cells showed no replicative differences between both serotype II strains 9 10 29 However both cultures are hardly sensitive to serotype I FCoVs To date cultivation of serotype I FECVs has never been achieved and only few serotype I FIPV strains could be adapted to grow in continuous cell cul tures In addition most of these strains seem to have lost their pathogenicity through cell culture adaptation 17 31 In the present study the newly established intestinal epithelial cell cultures were further evaluated for their sus ceptibility to serotype I field strains Infectious enterotropic virus was found in 57 8 14 of all FCoV positive faecal samples originating from healthy cats in 3 geographically distinct multi cat environments One of those samples was detected only by immunofluorescence staining This higher sensitivity can be explained by the use of more inoculum in that test In the majority of the posi tive samples infectious titres were always between 10 3 05 to 10 5 77 times lower compared to the total virus titre This difference can be attributed to the presence of defective particles but infectious titres in such faecal samples can possibly be underestimated due to faecal toxicity to the cells and the presence of neutralizing IgA antibodies as well In infection experiments with FECV UCD the amount of infectious particles was typically 3 4log 10 times lower compared to the total amount of particles in the first week pi but this further increased thereafter most probably due to the generation of neu tralizing antibodies data not shown It is impossible to estimate when cats in multi cat environments became infected and the presence of neutralizing antibodies can explain why infectious virus in some of the faecal samples with a quite high viral load was not detectable Corona virus was detected in 3 4 of the tested faecal samples from FIP cats Previously it has been shown that faecal viruses from FIP cats did not cause enteric infect