Growth of Piscirickettsia salmonis to High Titers in Insect Tissue Culture Cells

Salmonid rickettsial septicemia (SRS), or piscirickettsiosis, was first recognized in 1989 in Chile (1, 3), where it remains the most significant disease affecting salmonids in aquaculture (4, 11). The causative organism was identified as a highly fastidious obligate intracellular bacterium that replicates within membrane-bound cytoplasmic vacuoles in tissue culture cells or those of the host fish (4, 5). Despite its rickettsia-like properties, the 16S DNA sequence of the organism shows that it is a member of the γ-proteobacteria, unlike the Rickettsiales, which are α-proteobacteria (6). Therefore, it was named Piscirickettsia salmonis, as a new genus and species within the γ-proteobacteria (6). Subsequently, SRS has been recorded in Canada (2), Norway (12), Ireland (15), and Scotland (7), but both the incidence and losses have been much lower than those in Chile. Rickettsiae are usually transmitted to hosts via an invertebrate vector from a reservoir (14), but with no vector or reservoir established for P. salmonis, transmission is generally considered to be via water (4). However, a Scottish isolate of P. salmonis (SCO-95A), although highly virulent by injection, is very poorly transmitted by cohabitation or bath challenge (T. H. Birkbeck and S. Wadsworth, unpublished data), leading us to investigate possible invertebrate reservoirs or vectors of the organism. We show here that P. salmonis SCO-95A replicates to much higher titers in an insect cell line than in the Chinook salmon embryo (CHSE-214) cells normally used to culture the organism and that P. salmonis retains virulence for Atlantic salmon after repeated culture in insect cells.

P. salmonis isolate SCO-95A was obtained from an outbreak of SRS in Atlantic salmon in Scotland in 1995 (7) and was cultured in CHSE-214 cells (ECACC 91041114) in Eagle's minimal essential medium (MEM) with 4 mM glutamine and 10% fetal bovine serum in the absence of antibiotics. Rainbow trout gonad cells (RTG-2; ECACC 90102529) and the Xenopus laevis cell line XTC-2 (13) were cultured similarly. Sf21 cells (Spodoptera frugiperda; ECACC 89070101) were cultured in TC100 medium plus 10% fetal bovine serum. The Sf21 insect cell line, derived from the fall armyworm (16), was chosen because many rickettsiae are capable of growth in insects; the cell line is readily available and grows well in culture. All media, sera, and supplements were obtained from Gibco (Invitrogen).

The concentration of infectious units (IU) of P. salmonis in cell cultures was determined by plaque assay (3) in which serial dilutions of homogenates of tissue culture cells in Eagle's MEM were added to duplicate wells of monolayer cultures of CHSE-214 cells in Costar 24-well culture plates. After culture for 10 days at 18°C, medium was removed, cells were Giemsa stained, and plaques resulting from P. salmonis replication were counted. Each plaque was considered to represent 1 IU of P. salmonis in the original culture fluid.

The yield of P. salmonis from growth in CHSE-214 and Sf21 cells in four separate experiments is shown in Table 1. This was consistently greater in Sf21 than in CHSE-214 cells, with a mean recovery of 5.5 × 10⁷ IU ml of culture fluid⁻¹ after culture for 7 to 14 days, approximately 100 times the yield in CHSE-214 cells over 14 to 21 days. P. salmonis type strain LF-89 (ATCC VR-1361) (6) produced similar cytopathic effects in Sf21 cells, although the yield of piscirickettsiae was not quantified.

The ceiling temperature for growth of P. salmonis was established in Sf21 cells as they can grow at >30°C. P. salmonis was able to grow at 24°C; but at 25 to 26°C, the bacteria retarded Sf21 cell growth slightly and no cytopathic effects were seen. The optimum temperature range for growth was 20 to 23°C.

Virulence of P. salmonis passaged at 7- to 10-day intervals for 4 months in Sf21 cells was compared with organisms cultured in CHSE-214 cells. Cells and culture fluid containing bacteria and cell debris were harvested from six 175-cm² tissue culture flasks of CHSE-214 and Sf21 cells in which all cells had become detached from the surface of the flasks. After centrifugation at 10,000 × g for 10 min, the pellet was resuspended, gently homogenized, and resuspended in a one-tenth volume of Eagle's MEM (without serum). Atlantic salmon postsmolts, mean weight 70 g, were marked by dye injection and adipose fin clipping to identify 10 groups. Eight groups of five fish were injected with 0.1 ml of four 10-fold serial dilutions of the infective material in Eagle's MEM (CHSE-214-grown P. salmonis) or TC100 medium (Sf21-grown P. salmonis). The IU of P. salmonis injected per fish was determined retrospectively as described above. The 40 fish, along with 20 cohabitants, injected with Eagle's MEM or TC100 medium were held in running seawater at 14 ± 0.5°C in a 1.5-m diameter tank at the Marine Harvest Trials Unit, Lochailort, Scotland, for 28 days, after which time the temperature was raised to 17.5 ± 0.5°C until the experiment was terminated on day 73. Dead or moribund fish were removed, and kidney homogenate samples were applied to CHSE-214 cell monolayers to confirm the presence of P. salmonis. The number of viable IU of P. salmonis injected into fish could not be determined beforehand, as the titration took at least 10 days, and this led to differing quantities of P. salmonis being used from the two cell lines.

For piscirickettsiae grown in Sf21 cells, all fish injected with the highest dose (2 × 10⁸ IU) died 10 or 11 days postinfection and from 14 to 17 days postinfection with the lowest injected dose (2 × 10⁵ IU) (Table 2). For CHSE-214-cultured P. salmonis, fish administered the two highest doses (2 × 10⁶ and 2 × 10⁵ IU) died within 15 to 20 and 13 to 21 days, respectively. As the incidence of SRS is temperature dependent, the water temperature was raised to 17.5°C on day 28 in an attempt to provoke mortalities in cohabitant fish.

With the lowest injected dose (2 × 10³ IU), four of the five fish died by day 43 and one fish survived to day 73 when the trial was terminated. Only 1 of 20 cohabitants succumbed to infection (on day 51). The IU of P. salmonis required to kill half the fish was <2 × 10³, with no discernible difference in virulence between P. salmonis cultured in the two tissue culture cell lines.

The ability to culture P. salmonis in both insect and frog tissue culture cells suggests that the bacterium may persist in invertebrates and nonfish poikilotherms. A search for possible reservoirs of infection or vectors within marine invertebrates is justified, but our initial studies using specific PCR primers described by Marshall et al. (9) failed to reveal evidence of P. salmonis in sea lice (Lepeophtheirus salmonis and Caligus elongatus) taken from heavily infected Atlantic salmon or from filter-feeding bivalves (Mytilus edulis, Venerupis spp., and Tellina tenuis) from the vicinity of infected farms. However, as P. salmonis has been detected by PCR in a bacterioplankton fraction from seawater off the Oregon coast (10), further studies in the environment close to aquaculture sites regularly affected by SRS would be worthwhile.

For much work on P. salmonis, Sf21 cells offer considerable advantages over CHSE-214 cells, with a much higher yield of piscirickettsiae and the absence of fish-derived antigens. In addition, since Sf21 cells can be cultured in suspension, much larger scale production of P. salmonis is possible for killed bacterin vaccines until effective recombinant vaccines (e.g., reference 8) are available commercially.

• Copyright © 2004 American Society for Microbiology