OUP user menu

Transcytosis of Streptococcus iniae through skin epithelial barriers: an in vitro study

Marina Eyngor, Stefan Chilmonczyk, Amir Zlotkin, Elisabetta Manuali, Dan Lahav, Claudio Ghittino, Roni Shapira, Avshalom Hurvitz, Avi Eldar
DOI: http://dx.doi.org/10.1111/j.1574-6968.2007.00973.x 238-248 First published online: 1 December 2007


By constructing a biological model based on in vitro culture of polarized rainbow trout primary skin epithelial cell monolayers, the series of early events that precede Streptococcus iniae infection, particularly colonization and translocation through external barriers, were analyzed. Streptococcus iniae promptly invades skin epithelial cells, but the rapid decline of viable intracellular bacteria points out the limited capability of intracellular survival for this bacterium. Translocation assays, supported by electron microscopy microphotographs, demonstrate that following successful in vitro invasion of skin epithelial cell, the bacterium exists free in the cytoplasm after release from the endosome, and translocates through the skin barrier. Bacterial invasion and transcytosis is not accompanied by apparent cell-line damages or disruption of host cells' tight junctions. It is hypothesized that the phenomenon of epithelial invasion coupled to the rapid translocation through the barrier plays a crucial role in Streptococcus iniae infection.

  • Streptococcus iniae
  • skin
  • adhesion
  • invasion
  • translocation


Streptococcus iniae is one of the leading causes of morbidity and mortality of fish in numerous countries, including Israel and the United States (Eldar et al., 1995). The most common features of fish infected by virulent Streptococcus iniae strains (Weinstein et al., 1997; Fuller et al., 2001) are meningitis or meningoencephalitis (Eldar & Ghittino, 1999); systemic infection may also occur, primarily in outbreaks caused by invasive serotype II strains (Barnes et al., 2003; Lahav et al., 2004). Accidental injuries of humans following handling of fish can lead to infections, mainly of the skin. Blood isolations are less frequent (Weinstein et al., 1997).

Available data point out the central role of Streptococcus iniae capsular polysaccharide in pathogenesis (Bachrach et al., 2001; Barnes et al., 2003; Zlotkin et al., 2003; Lahav et al., 2004; Miller & Neely, 2005). The genetics of the capsule (21-gene) operon has been elucidated (Lowe et al., 2007), and further evidence for its main role was demonstrated through deletion mutants with decreased capsule expression, shown to be attenuated (Locke et al., 2007; Lowe et al., 2007). Phosphoglucomutase has also been proposed as a virulence factor (Buchanan et al., 2005). Additional mechanisms of Streptococcus iniae virulence, other than the presence of a streptolysin S-associated gene (Fuller et al., 2002), remain largely unknown. Additionally, as most previous models used to assess strain virulence and pathogenicity by intraperitoneal or intramuscular inoculation of bacteria (Eldar & Ghittino, 1999; Fuller et al., 2002; Zlotkin et al., 2003), data pertaining to initial sites of adhesion, colonization, and invasion were not made available.

While external surfaces provide a primary defense against invading organisms, several pathogenic bacteria possess the (in vitro) ability to translocate an epithelial or mucosal cell barrier without causing an apparent initial damages (Falkow et al., 1992; Kops et al., 1996; Bayles et al., 1998; Peters et al., 1999; Burns et al., 2001; Theodoropoulos et al., 2001; Spence et al., 2002). In vivo, this is an important virulence feature, as it allows the invader access to underlying tissues without causing initial inflammatory reaction and may permit the pathogen to disseminate throughout the host. By constructing a biologically relevant model, based on in vitro culture of rainbow trout (polarized) skin epithelial cell monolayers, the series of early events that precede the clinical stage of Streptococcus iniae infection, particularly colonization and translocation through external barriers, have now been investigated. Polarized cell monolayers are characterized by defined apical and basolateral cell surfaces separated by tight junctions (TJs) that are crucial for the development and maintenance of epithelial cell surface polarity (van Meer et al., 1986). TJs join individual epithelial cells and seal the intercellular space to create a primary barrier that limits the passage of solutes and pathogens through the paracellular spaces (Gumbiner, 1993; Madara, 1998), thereby providing an in vitro model to assess the ability of bacteria to translocate across an intact epithelial cell barrier. Monolayer integrity and the presence of TJs are monitored by transepithelial electrical resistance (TER) measurements; disruption of the intercellular TJs results in TER decrease.

In vitro cellular assays, supported by electron microscopy microphotographs, demonstrate that Streptococcus iniae successfully invades skin epithelial cells, exists free in the cytoplasm after release from the endosome, and translocates through rainbow trout skin barrier. Bacterial invasion and transcytosis is not accompanied by TER decrease, lactate dehydrogenase (LDH) release, or morphological changes of host cells.

Materials and methods

Bacterial strains and culture conditions

Streptococcus iniae Dan-15 (serotype I) is a clinical isolate recovered in 1992 from the brain of a diseased rainbow trout in a commercial a fish farm in the Upper Galilee, Israel; Streptococcus iniae KFP 404 (serotype II) was collected from the same site in 2000. Staphylococcus caseolyticus KFP 776 (control strain) is a commensal strain recovered from a healthy rainbow trout by striking a skin sample on a Baird–Parker agar base (Becton Dickinson, Sparks, MD) supplemented with 0.01% sodium azide. For the definitive identification of KFP 776 as Staphylococcus caseolyticus, the 16S rRNA gene was amplified [with primers U-1 and U-1500 (Escherichia coli 16S rRNA gene sequence, GenBank Accession no. J01859, positions 7–27 and 1516–1540, respectively) and the product was sequenced by the method of dye terminator cycle-sequencing using fluorescent-labeled dye terminators (PE Biosystems), and an ABI 377-automated DNA sequencer (PE Biosystems)]. The resulting nucleotide sequences were compared with the database deposited in GenBank. The highest similarity to known sequences in Genbank was 99% to Staphylococcus (Macrococcus) caseolyticus ATCC 13548 16S rRNA gene (Genbank Accession no. Y15711).

All strains were stored at −70 °C in brain heart infusion broth (Oxoid, Basingstoke, UK) with 15% glycerol. Gram-positive cultures were initially grown on Columbia blood agar (Oxoid) at 18 °C; for infection assays, bacteria were grown for 8 h in brain heart infusion broth. OD640 nm was measured with a spectrophotometer (Shimadzu Corporation, Kyoto, Japan), and viable CFU counts were determined. Mid-log-phase cultures (108 CFU) were found to correspond to an OD of 0.30–0.35. Bacterial suspensions were washed twice (by centrifugation at 6000 g for 15 min) in phosphate-buffered saline (PBS) [15 mM Na2HPO4, 145 mM NaCl (pH 7.20)] and concentrated so that, for experiments, c. 1.5 × 107 CFU in a 20-µL volume were added per well.

Streptococcus iniae enumeration from the skin

In order to quantify Streptococcus iniae, CFUs on the skin of fish living in an infected environment [a 100 m3 raceway where 50 000 rainbow trout (40 g each) are reared, and specific Streptococcus iniae daily mortality is 2–4%], 20 (apparently) healthy fish were sacrificed and 1 cm2 of skin from each fish was removed and milled until a complete and homogenous mixture was obtained. The tissue homogenate was suspended in PBS and briefly centrifuged to remove cellular debris; aliquots (in triplicate) were serially diluted and plated on phenylethyl alcohol agar (Becton Dickinson) supplemented with 5% defibrinated blood. Plates were incubated at room temperature. Definitive identification of the resultant colonies was achieved by PCR, as described elsewhere (Zlotkin et al., 1998).

The pathogenicity of Streptococcus iniae skin-isolates was assessed by cohabitating naïve fish with fish previously infected with these isolates by an intraperitoneal injection, as described elsewhere (Zlotkin et al., 2003). The mortality rate of fish infected by the (five randomly chosen) Streptococcus iniae skin isolates was similar to that of clinical isolates.

Preparation of rainbow trout skin epithelial primary cell cultures

As no rainbow trout skin epithelial cell lines are presently available, the development of appropriate culture methods was required in order to construct a biologically relevant model enabling to mimicate the series of events that take place during the initial phases of host–pathogen interactions.

Rainbow trout, weighting 100 g each, were obtained from an Streptococcus iniae-specific pathogen-free (SPF) facility. To reduce fungal contamination, fish were immersed in 1 µg mL−1. malachite green (Sigma) solution for 3 h. Fish were then sacrificed and scales were removed and cultured at 18 °C in 25 cm3 flasks (Costar Co., Cambridge, MA) in Modified Eagle's Medium (MEM; GIBCO Laboratories, Grand Island, NY) supplemented with 20% heat-inactivated fetal calf serum (Gibco), HEPES (1%), chloramphenicol (200 µg mL−1), ceftriaxone (200 µg mL−1), and amphotericin B (1 µg mL−1). The supernatant was removed daily, and cultures were washed in PBS before medium was replenished. After 6 days, when epithelial cells were grown to confluence of ≥95%, monolayers were trypsinized and scale debris removed. A major advantage of the present study is that cultures with mucous occurring throughout the outgrowth were obtained consistently (Fig. 3a). As mucus is an essential feature of the normal fish skin and contains peroxidase (Iger & Wendelaar Bonga, 1994; Brokken et al., 1998), lysozyme (Rainger & Rowley, 1993), immunoglobulins, complement, c-reactive protein, etc. (Shephard, 1994), the presence of mucus (and mucous cells) in the culture is likely to be important as it indicates that cultures closely approximate that of the teleost epidermis in vivo.

Figure 3

Transmission electron micrographs illustrating interactions between Streptococcus iniae and primary rainbow trout skin epithelial cells. Eukaryotic cells were grown on polycarbonate supports as described in ‘Materials and methods,’ and infected with Streptococcus iniae (MOI=100). All figures illustrate a 30-min coculture of eukaryotic cells with Streptococcus iniae. (a) Cross section of the apical limit of two adjacent epithelial cells (EC1, EC2). The components of the epithelial junctional complex (tight junction=thin arrow; desmosomes=large arrows) display their characteristic ultra structure, attesting to the integrity of the intercellular cohesion. A tenuous glycocalyx (arrow heads) overlays the cell surface. (b) First step of the translocation process: bacteria are located in deep plasmalemmal invagination of the apical surface of RBT epithelial cell and encircled by pseudopod-like structures. N, nucleus of the epithelial cell. (c) Endocytosis of Streptococcus iniae. c1, cross section of an epithelial cell overlaying a polycarbonate support (PS). Engulfed bacteria are enclosed within membrane-bound vacuoles. N, nucleus; m, mitochondria. c2, Vacuoles consisting of cytoplasmic membrane contortion associated with many of the infected cells. Each membrane-bound vacuole contains a single or more bacteria. Note the profile of a dividing bacterium (arrow). c3, endocytosis of Streptococcus iniae. Engulfed free bacterium located in direct contact with the cytosol of the epithelial cell.

Cells were subcultured for 1 day in antibiotics supplemented MEM–FCS medium; medium was then changed and cells were allowed to grow for an additional day in MEM–FCS with no antibiotic supplementation. These cells are referred to as rainbow trout skin epithelial primary cell cultures.

Adhesion assay

Adhesion of Streptococcus iniae to rainbow trout primary skin epithelial cell cultures was performed as described previously (Zlotkin et al., 2003). Briefly, cells were grown to confluence in 24-well tissue culture plates (Costar Co.) and used for quantitative adherence and invasion assays. The number of eukaryotic cells seeded per well and recovered after growth to confluence (1.5 × 105) was determined by quantitation in a counting chamber. Mid-log-phase bacterial suspensions (20 µL) were added to prewashed confluent monolayers at a multiplicity of infection (MOI) of 100 bacteria per eukaryotic cell. Owing to the high g required for Streptococcus iniae sedimentation, spinning was omitted; bacteria were therefore equally distributed in the 200 µL culture media overlaying the cell culture. Thus, only those bacteria suspended in the media directly spreading over the monolayer surface were offered to the cells. Based on these figures, the actual MOI was c. 25 (7.5 × 106 CFU in 50 µL — the volume that covers the 1.5 × 105 monolayer cells). After 30, 60, 90, and 120 min of incubation, nonadherent bacteria were removed by washing the cell cultures three times with PBS. For viable count determinations, infected monolayers were treated for 5 min with 0.05 mL of 0.25% trypsin and of 0.1% EDTA (Sigma) in Hanks balanced salt solution (Gibco), and streptococci were harvested by adding 0.1 mL of 0.025% Triton X-100 (U.S. Bio-chemical, Cleveland, OH) in sterile distilled water to each well. After 3 min, cell lysates were collected and serially diluted in PBS, and aliquots (in triplicate) were plated onto blood agar to assess bacterial CFU. All assays were independently repeated four times. Results are expressed as averages of all experiments.

Background activity due to nonspecific adherence of bacteria was determined by including an identically treated control plate that contained no cells. The level of nonspecific binding of bacteria to the plastic tissue culture wells was <1% of the specific adherence values. The results were considered as the total cell-associated (invading plus surface-adherent) bacteria. Adherent bacteria were quantified by subtracting the number of invasive bacteria (as described in the ‘invasion and survival assay’) from the total cell-associated bacteria.

Invasion and intracellular persistence assays

To quantify the invasion capacity and intracellular persistence of Streptococcus iniae, the ability of Streptococcus iniae to enter and survive within cells was next investigated. Primary skin epithelial cultures were infected as described above. Killing of extracellular bacteria was accomplished by adding (100 µg mL−1) ampicillin (Streptococcus iniae MIC >0.016 mg mL−1; Staphylococcus caseolyticus MIC >0.064 mg mL−1). It has been shown previously that following addition of ampicillin to a cell culture containing intracellular Streptococcus iniae, intracellular bacterial CFU counts continued to increase for several hours and were detectable for up to 2 days, while extracellular bacteria were killed rapidly (Zlotkin et al., 2003). All bacteria, independent of the strain tested, were killed in those control studies after 3 h in cell culture media containing 100 mg of ampicillin mL−1. All bacteria, independent of the strain, were also killed by incubating the bacteria with epithelial cells in medium containing ampicillin at 100 mg mL−1 from the beginning of the incubation period. For determination of invading bacteria (invasion assays), extracellular bacteria were removed after 30, 60, 90, and 120 min by three washes with PBS, and the original volume was reconstituted with MEM supplemented with 10% fetal calf serum and antibiotics. After the an addition of ampicillin, incubation was allowed to proceed for additional 3 h before cells were lysed and intracellular bacteria were plate counted. As ampicillin kills extracellular bacteria, a decrease of recovery over time (i.e. lower numbers of intracellular bacteria) would reflect failure of intracellular survival of the microorganism.

For the demonstration of intracellular persistence (survival assays), infection was allowed to proceed for 2 h, while incubation time in the presence of ampicillin was extended up to 5 days (4, 6, 24, 72, and 144 h) to give sufficient time for potential intracellular survival or replication to occur.

LDH release assay

A colorimetric assay, based on quantitation of LDH released from the cytosol of damaged cells, was used to assess sublethal damage to cells. Confluent cultures grown in 24-well tissue culture plates were infected with mid-log-phase Streptococcus iniae bacterial culture resuspended in 1 mL of MEM medium without fetal calf serum at a MOI of 100 bacteria per eukaryotic cell, as described previously. LDH from MEM alone and from Streptococcus iniae in MEM without a cell monolayer were assessed. A complete distilled H2O lysis of cell monolayer was used as a positive control. Plates were incubated for 30, 60, 90, and 120 min, at which time supernatants (in duplicates) from uninfected control and infected cells were collected and spun for 5 min at 280 g. Twenty-microliter portions of cleared supernatants were then transferred to a replica plate for quantification of LDH leakage into the medium using a miniaturized colorimetric commercial kit according to the manufacturer's instructions (Sigma, cat. no. 500-C). The A460 nm of each well was used to calculate the residual pyruvic acid activity, which is inversely proportional to LDH activity. All assays were repeated three times.

Polarized epithelial monolayers

Rainbow trout primary epithelial cells were washed three times in PBS, trypsinized, counted (by Trypan blue exclusion), and 7.5 × 104 cells were seeded on top of the apical side of a 6.5 mm-diameter Transwell-COL (collagen coated) polycarbonate membrane with a pore size of 3.0 µm (Costar Co.) in a volume of 200 µL of MEM antibiotic-free medium. The Transwell-COL inserts were placed into wells of a 24-well titer-plate (Corning Inc., Corning, NY), suspending the collagen-coated membrane in 1.5 mL of antibiotic-free MEM medium placed in the lower well.

Following overnight culture, eukaryotic epithelial monolayers were tested for integrity by excluding those that could not prevent the flow of media from the upper chamber into the lower chamber. Only those that could block the visible flow of media into the lower chambers for over 60 min were used for transcytosis experiments.

Electrical measurements

The integrity of the cell monolayer grown on polycarbonate filters and the presence of TJs was evaluated by measuring the TER (ohms cm−2; Ω cm−2) using a Millicell-ERS resistance system (Millipore). TER measurement (with SD) was calculated for each time point (30, 60, 120 and 240 min), and each measurement was repeated in triplicate. The TER was calculated from the following equation: (TERsample−TERblank) × surface area. The average TER measurement of polycarbonate filters in the absence of a cell monolayer was 90 Ω cm−2 (baseline). Monolayers of confluent primary cell sheets with a transepithelial resistance of ≥160 Ω cm−2 were considered confluent; inserts with TER <160 Ω cm−2 were discarded.

Paracellular permeability

As a second method for monitoring cell monolayer integrity and the presence of tight junctions, in a preliminary stage, and in two of the four independent repeats of the experiment, the transepithelial flux of FITC-labeled dextran (Sigma) was used as a marker of paracellular leakage (Madara, 1998). The mass of FITC-dextran, 70 kD, approximates that of human albumin, both of which have been used in similar endothelial paracellular permeability models (Sanders et al., 1995; Albelda et al., 1988; Madara, 1998) FITC-labeled dextran 70 kD (final concentration 5 mg mL−1) was poured on the surface of epithelial cells grown to confluence on the apical side of inserts (TER ≥160 Ω cm2). At 20-min intervals, 100-µL samples were removed from the lower chamber; sample volume was replaced with fresh media and the fluorescence intensity (excitation, 485 nm; emission, 530 nm) of each sample was measured. FITC-dextran concentrations were determined from standard curves generated by serial dilution of FITC-dextran. Paracellular flux was calculated by linear regression of sample fluorescence (Madara, 1998).

Infection and transcytosis of polarized epithelial monolayers

Transcytosis assays were carried out when electrical resistance was at its peak (160 Ω cm−2), indicating maximum formation of tight junctions. Inserts containing the cell monolayers were infected by adding 10 µL (7.5 × 106 CFU) suspensions of Streptococcus iniae Dan-15, Streptococcus iniae KFP 477, or Staphylococcus caseolyticus KFP 776 to the apical chamber (MOI=100), without spinning. At 30, 60, and 120 min postinfection, 100 µL aliquots of MEM medium from the lower chamber were sampled for enumeration of bacterial CFU (in triplicate) by plate counting on blood agar plates.

A control well, which did not contain epithelial cells, demonstrated that the majority FITC-labeled dextran passed through the membrane pores within the first hour. Similarly, Streptococcus iniae Dan-15, Streptococcus iniae KFP 477, and Staphylococcus caseolyticus KFP 776 immediately traversed a membrane-only control as well (data not shown).

Statistical analysis

Data are presented as means±SD from four independent experiments performed in triplicate. Results were analyzed by Student's t-test. The results of paracellular flux were analyzed by linear regression of sample fluorescence (sas software, version 5). The electron microscopy analysis was repeated three times, and data from a typical experiment are reported.

Transmission electron microscopy (TEM)

To confirm the intracellular localization of Streptococcus iniae, TEM was performed on infected monolayers using standard techniques. Briefly, after the initial 1-h infection the polycarbonate filters were washed seven times in PBS, fixed with 2% glutaraldehyde in 0.1 M cacodylate buffer (CB), pH 7.2, at 4 °C for 2 h, and washed three times in the same buffer. Specimens were postfixed with 2% osmium tetroxide in distilled water, dehydrated in a graded series of ethanol, and embedded in Epon 812. Ultrathin sections were poststained with uranyl acetate and lead citrate before viewing under a Philips EMC12 transmission electron microscope.


Purity and growth of primary skin epithelial cells

The purity of the primary cell cultures was determined by light microscopic observation, and typical cellular morphology was also confirmed for several experiments by TEM (data not shown). Preparations usually contained no more than 1% fibroblasts. Preparations that contained excessive fibroblasts were discarded.

Streptococcus iniae is present on the skin of fish and adheres to skin epithelial cells

Bacterial counting of skin samples revealed that 17/20 healthy fish were colonized by 16–80 Streptococcus iniae CFU 1 cm−2. Analysis of (in vitro) adherence to the epithelial cell layers, which is a precondition for colonization and for potential invasion of these cells, demonstrates that Streptococcus iniae readily adheres to monolayers of epithelial cells and that the number of attached bacteria increases through time (Fig. 1). For Streptococcus iniae Dan-15 (serotype I), an average of 1.9, 2.9, 5.0, and 11.9 CFU bounded to each primary epithelial cell after 30, 60, 90, and 120 min, respectively. For Streptococcus iniae KFP 404 (serotype II), the number of adherent bacteria to each primary epithelial at the same time intervals was found to be similar (1.0, 2.9, 3.5, and 11.3 CFU per cell). The apparent discrepancy between the (high) bacterial loads in the in vitro assay and the (relatively low) in vivo counts can be rationalized by the fact that the ‘infected environment’– rainbow trout farm raceways — is continuously flushed with pure effluent water (c. 5 complete water changes per hour). Thus, not only are the pathogens mechanically swiped from the skin (trout swim vigorously countercurrent), the environment itself is continually diluted. Indeed, by classical bacteriological tools the pathogen could not be isolated from the water (data not shown).

Figure 1

Adherence assays. Streptococcus iniae strains Dan-15 and KFP 404 adherence to rainbow trout primary epithelial cells. Bacteria were added to confluent monolayers of rainbow trout primary epithelial cells (MOI 100) and assessed for adherence as described in ‘Materials and methods.’ Adherence of Streptococcus iniae, expressed as CFU bound/cell (y-axis), was time dependent (x-axis). Data are means±SD from three experiments performed in quadruplicate. P<0.01 for Streptococcus iniae Dan-15 vs. Streptococcus iniae KFP 404 adherence to epithelial cells after 30 or 120 min; P>0.05 for Streptococcus iniae Dan-15 vs. Streptococcus iniae KFP 404 adherence to epithelial cells after 60 or 90 min (by Student's t-test).

Staphylococcus caseolyticus (control strain) adhered to primary epithelial cells (1.6, 2.8, 5.1, and 8.0 CFU per cell after 30, 60, 90, and 120 min of incubation) to a magnitude similar to that of Streptococcus iniae.

Streptococcus iniae (but not Staphylococcus caseolyticus) invades epithelial cells, but persists only for a limited time

The ampicillin protection assays (performed 30, 60, 90, and 120 min postinfection) revealed the presence of viable Streptococcus iniae in the intracellular compartment of the primary epithelial cells, with no significant variations according to the Streptococcus iniae strain tested. As shown in Fig. 2, intracellular localization was recorded as soon as 30 min postinfection, with maximal counts at 120 min (1.25 CFU per cell for Dan-15; 1.35 CFU per cell in the case of KFP 404). No CFU counts of intracellular Staphylococcus caseolyticus control strain were recorded, indicating that the control strain did not invade primary cells to any extent.

Figure 2

Invasion and intracellular persistence assays. Streptococcus iniae invaded primary rainbow trout epithelial but were able to survive intracellularly for only a limited period of time. Streptococcus iniae strains Dan-15 and KFP 404 were inoculated onto confluent epithelial cell monolayers; invasion and survival assays were performed as described in ‘Materials and methods.’ Invasion assay was performed for 0.5, 1.0, 1.5, and 2.0 h (x-axis) before extracellular bacteria were killed by ampicillin. For survival assays, epithelial cells had been infected for 2 h before ampicillin was added. Streptococcus iniae intracellular survival was determined at 4, 6, 24, 72, and 144 h postinfection. Intracellular bacteria were defined as the number of CFU recovered from each cell (y-axis). Data are means±SD from three experiments performed in quadruplicate. P<0.005 for Streptococcus iniae Dan-15 vs. Streptococcus iniae KFP 404 intracellular survival after 4 h; P>0.05 in all other cases (by Student's t-test).

After demonstrating the invasiveness of Streptococcus iniae for epithelial cells and determining conditions for optimal invasion, the persistence of both Streptococcus iniae strains within the epithelial cells was determined. Based on the results of the invasion assays, an initial 2-h period was chosen for invasion; invasion was then stopped by adding ampicillin to the media and the number of viable intracellular bacteria were determined at various times (up to 144 h) in order to demonstrate potential intracellular persistence (or intracellular replication). A rapid decline of intracellular viability was noticed, and after 4 or 6 h (Dan-15 and KFP 477, respectively), the number of intracellular bacteria was found to be below the maximal invasion efficiencies (0.3 CFU per cell; Fig. 2). Twenty-four hours after infection, the number of intracellular bacteria was reduced to 0.1–0.18. At day 3, the number of intracellular bacteria diminished to 0.05–0.08 CFU per cell. Intracellular bacterial counts continued to diminish and, at day 5, the intracellular presence of Streptococcus iniae Dan-15 (type I strain) was negligible (0.0001 CFU per cell). Instead, the intracellular presence of Streptococcus iniae KFP 404 (type II strain), although low (0.03 CFU per cell), was still appreciable. The results indicate that Streptococcus iniae could survive within the epithelial cells for a restricted time and that serotype II strain might also undergo a limited intracellular replication. Hence, Streptococcus iniae does not efficiently reside or survive within epithelial cells for long periods of time.

In vitro establishment of epithelial barrier

Monitoring of epithelial cultures demonstrated that overnight monolayer cultures were consistently characterized by TER values of ≥160 Ω cm−2 and that these monolayers possess barrier properties that excluded the flow of fluid from the upper to the lower chamber (the rate of FITC-labeled dextran passage across the monolayers was more than 1000-fold lower in comparison with control inserts). These data point out that, within a few hours under optimal conditions, rainbow trout primary epithelial cultures formed tight junctions, as described for several cell lines (Gonzalez-Mariscal et al., 1985, 1990; Haake & Lovett, 1994; Furuse et al., 2001; Barocchi et al., 2002).

TER is probably the result of a delicate balance between cell type to claudins/occludin levels and other regulatory components of TJs (Dombek et al., 1999; Furuse et al., 2001). Therefore, it is not surprising that TER values vary considerably according to the cell line, and even between lineages of the same line. For instance, the TER of confluent MDCK monolayers can range from 13 000 to 200 Ω cm−2 (Gonzalez-Mariscal et al., 1985). Yet, other authors found that monolayers of the same line with TER of 65 Ω cm−2 are still confluent (Abrami et al., 2003). Thus, relatively low TER values (i.e. 140–180 Ω cm−2) can be typical of certain confluent monolayers under those circumstances where paracellular permeability is intact. Because the primary skin epithelial monolayers used in the present study showed consistent TER values and barrier properties, the parameters for confluent monolayers with TJs and absence of gaps were met.

Rapid translocation of skin epithelial cell monolayers by Streptococcus iniae

The demonstration that Streptococcus iniae can invade primary skin epithelial cells led to the hypothesis that this bacterium might be able to traverse an epithelial cell monolayer without causing cell damage or apparent disruption of intracellular junctions, similar to what has been reported in group B streptococci (Winram et al., 1998). In vivo, such a mechanism would be advantageous for the pathogen, as it enables to translocate through intact skin and initiate infection.

After apical infection of epithelial cells, Streptococcus iniae was recovered from the lower chamber of the Transwell apparatus as early as 30 min postinfection. There was little variability in the time required for each of the two Streptococcus iniae strains to traverse the epithelial monolayers, and no significant variations in CFU counts of both serotypes were observed. Logically, CFU counts in the lower chamber increased as a function of the time that has elapsed between infections and sampling. At 30 min postinfection, a total of 200 or 315 CFU (Dan-KFP 404 and Dan-15, respectively) had crossed the epithelial barrier, and after 90 min the total number of bacteria recovered from the lower chamber increased to 360 for Dan-15 and 380 for KFP 404. At 120 min postinfection (last assay), CFU counts in the lower chamber were already 780 (Dan-15) and 920 (KFP 404). The possibility that the time-dependent increase of CFU counts in the lower chamber should be attributed to multiplication of bacteria that have crossed the cell membrane in former time points was discarded: cell cycle analysis demonstrated that dilution of mid-log cultures (as in the present case) resulted in a lag-phase of at least 90 min.

The minimal (1% or less) variations in TER that were detected at each time point throughout the experiment (starting from uninfected monolayers) indicated that the epithelial intracellular junctions remained intact. No LDH release beyond baseline (medium alone) was detected from monolayers (grown in 24-well) exposed for 2 h to 7.5 × 107 CFU of Streptococcus iniae types I or II strains, indicating a strong correlation between TER stability and the absence of cellular damages. Thus, Streptococcus iniae were able to transcytose rapidly through an intact cell monolayer.

The quick transcytosis through the epithelial layer and the (apparently) low invasion efficacy of the epithelial boundary are potentially controversial. This apparent conflict might be explained by the rapid exocytosis of Streptococcus iniae through the basal side of the cell membrane as follows: first, a detailed analysis of the figures points out that, as spinning was omitted, the number of bacteria that have been actually offered for the assay consisted only of the 9.375 × 105 CFU, which were dispersed in the 25 µL culture media directly overlaying the (7.5 × 104) cell monolayer; the effective MOI was therefore 12.5. Because at 30 min postinfection, 200 or 315 CFU (KFP 404 and Dan-15, respectively; P<0.005) were recovered from the lower chamber, data analysis reveals that 2.1% of Streptococcus iniae KFP 404 and 3.3% of Streptococcus iniae Dan-15 in actual contact with cells have translocated through the epithelial barrier. In a broader view, bacterial counts in the lower chamber consisted of 13.3% of the total invading Dan-15 and 4.2% of the total invading KFP 404 (0.2 and 0.1 internalized Dan-15 or KFP 404 per cell were recorded at 30 min postinfection). Thus, Streptococcus iniae is capable of translocating epithelial monolayers to a considerable extent, and CFU counts of the lower chamber are a logical outcome of intracellular CFU counts. Second, as survival assays indicate that Streptococcus iniae cells are rapidly eliminated from the intracellular compartment, significant numbers of internalized bacteria are not expected.

TEM examination of infected epithelial cells

Electron microscopy of infected monolayers demonstrated that intercellular junctions were not disrupted or involved in the translocation process (Fig. 3a). Invading bacteria were always observed within the cytoplasm of epithelial cells, comprising membrane-bound vacuoles. These derived from pseudopod-like structures of the plasma membrane involved in engulfing the bacteria, leading to the formation of membrane-bound endosomes (Fig. 3b). In some instances, several bacteria were located in a single vacuole, possibly indicating a limited intracellular replication. However, several vacuoles, most probably those internalized for longer periods of time, began to degrade and some bacteria were found entirely free in the cytoplasm (Fig. 3c).


The series of early events that precede clinical manifestations of infection, particularly colonization and translocation through external barriers, are a mainstay in Streptococcus iniae pathogenesis research. For this pathogen, which is transmitted by direct contact, adherence to the epithelial boundaries is a prerequisite for colonization. Subsequent to bacterial attachment, invasion and translocation through the epithelial cells could enable the pathogen to penetrate the first protective layer, evade the host innate defense mechanisms, and potentially establish an infection. By establishing a biologically relevant model, it is now apparent that (in vivo) skin epithelial cells provide a prompt colonizable surface for Streptococcus iniae and that (in vitro) this organism rapidly invades these cells. For Streptococcus iniae, as for other pathogenic bacteria, intracellular survival is ideally suited not only to overcome unfavorable host conditions but also to gain access to the invaded organism during the initial stages of the disease. In characterizing Streptococcus iniae pathogenesis further, it has been observed that many of the internalization events are consistent with the cellular events that are induced by other facultative intracellular pathogens (13, 14). Streptococcus iniae interacts or binds with the cell surface of the epithelial host cell and its uptake is probably facilitated by the formation of pseudopodia, which engulf the organism. A similar strategy has been described for group A (Dombek et al., 1999) streptococci, group B streptococci (Kallmanm & Kihlstrom, 1997; Valentin-Weigand et al., 1997) and staphilococci (Kahl et al., 2000). Although the internalized Streptococcus iniae cells were initially observed within membrane-bound vacuoles, bacterial cells eventually came to reside directly in the host cytoplasm; in this respect Streptococcus iniae mimics group B streptococci (Valentin-Weigand et al., 1997), and staphilococci (Bayles et al., 1998).

Although Streptococcus iniae has not been considered to be a classical facultative intracellular pathogen, the present data substantiate previous reports, revealing that, similar to several other streptococcal species, Streptococcus iniae is also capable of invading and surviving within eukaryotic cells (Zlotkin et al., 2003). Likewise, the finding that Streptococcus iniae intracellular survival and replication in epithelial cells is limited coincides with data regarding Gram-positive pathogenic cocci (Osterlund et al., 1997; Valentin-Weigand et al., 1997; Bayles et al., 1998; Benga et al., 2004). Exocytosis to the antibiotic-containing medium, leakage of the extracellular antibiotics, decreased bacterial access to nutrients, or intracellular killing by eukaryotic cells may account for the decrease in CFU recovered over time. This latter observation may be less critical because transcytosis of host epithelial cells begins within 30 min of apical contact.

Skin epithelial cells, a protective mechanical and immunochemical barrier, which represent the first line of innate defense mechanisms, are normally efficient in keeping the underlying tissues sterile. Interestingly, it is now demonstrated that Streptococcus iniae could translocate the epithelial cell layer without causing local damages or disruption of the cellular junctions that provide monolayer integrity. In vivo, complete cell translocation would provide the ultimate productive outcome of Streptococcus iniae invasion of skin epithelium, facilitating its rapid entry into the organism, into the bloodstream, and into target organs. Transcytosis of bacteria through intact skin is a rare event that has been associated mostly with members of the genus Leptospira. But even in this case the doubt whether penetration through the skin requires breaks in the skin is debatable (Barocchi et al., 2002). The likelihood that, under in vivo conditions, Streptococcus iniae is indeed capable of crossing intact skin is strengthened by the particular anatomic structure of fish skin, which is not shielded by a keratinized epidermis. Instead, the epidermis of teleosts consists of a thin living layer composed of two to 10 layers of epithelium whose outermost layer (Malpighian cell layer) retains the capacity to divide (Hawkes, 1983; Yonkos et al., 2000). The epidermis covers the flexible scales that arise in the dermis (stratum spongiosum) and project at an angle into the epidermis; the outer limit of epithelial surface is coated by an amorphous substance (glycocalyx) composed of mucus, mucopolysaccharides, immunoglobins, and free fatty acids (Hawkes, 1974). Scales are layered in overlapping rows like roof tiles; the gap between scales (scale pocket) allows cells from the epidermal basal layer (stratum germinativum) to migrate toward the surface. Hence, fish skin may well be regarded as a multilayered mucosal surface; mucosal surfaces are rapidly translocated by numerous pathogens as part of their invasion and dissemination strategy (Falkow et al., 1992; Kops et al., 1996; Meier et al., 1996; Finlay & Falkow, 1997; Winram et al., 1998; Peters et al., 1999; Burns et al., 2001; Barocchi et al., 2002). In vivo, the epidermis of the fish skin contains numbers of leukocytes, patrolling agents of the specific and nonspecific immune systems (van Muiswinkel, 1995; Sin et al., 1996; Dalmo et al., 1997) that are involved in disease resistance and maintaining healthy tissues (Secombes, 1994; van Muiswinkel, 1995). The fact that Streptococcus iniae is capable of residing within macrophages, exploiting these host cells for dissemination (Zlotkin et al., 2003), might also play a significant role in the events following primary transepithelial translocation. This hypothesis is further substantiated by data demonstrating that another external epithelial surface, the bilayered gill epithelium, is actively involved in engulfment of bacteria (Yersinia ruckeri), virus (Viral Hemorrhagic Septicemia virus), and particles (latex) (Chilmonczyk & Monge, 1980; Zapata et al., 1987).

For Streptococcus iniae, which — in an infected trout farm — cannot be isolated by conventional methods from the water (three to five complete changes of water are performed every hour) or the skin of healthy fish (M. Eyngor, unpublished data), the ability to adhere to cells may be as important as the capability to penetrate them. Successful colonization with subsequent rapid internalization and translocation may initiate dissemination, compensating for the low environmental presence and avoiding killing by local defense mechanisms. Streptococcus iniae might therefore be a more versatile and adaptive pathogen than believed previously. The present data might help in explaining the exceptional story of Streptococcus iniae, which, originally isolated from a freshwater Amazonian dolphin (Pier & Madin, 1976), has rapidly spread throughout continents and in <10 years turned out to be a major pathogen of fish.


  • Editor: Mark Enright


View Abstract