Nile tilapia served as experimental fish because they could be held at warm water temperatures closer to the normal human body temperature of 37°C than other available fish species. Seven groups of 10 tilapia (mean ± SE weight 28.20 ± 0.51 g) each were housed in 57-L aquariums at the Aquatic Animal Health Research Laboratory (Chestertown, MD, USA). All tanks were supplied with flow-through dechlorinated tap water and 2 submersible heaters and air stones to maintain desired water temperature and dissolved oxygen (DO) levels. Water quality (mean ± SE temperature, DO, and ammonia concentration) was measured daily by using a YSI 85 meter (Yellow Springs Instrument Co., Yellow Springs, OH, USA) and a Fresh Water Aquaculture Kit (Model AG-2; LaMotte Company, Chestertown, MD, USA). Temperature was 32.1 ± 0.09°C, DO was 4.2 ± 0.14 mg/L, and ammonia concentration was 0.74 ± 0.08 mg/L. Fish were fed daily (4% of bodyweight) with Aquamax Grower 400 fish feed (Purina, Brentwood, MO, USA) and maintained and handled according to Institutional Animal Care and Use Committee–approved guidelines.

The serotype Ia ST-7 human GBS isolate (ID# 510012) was obtained from a patient in Japan who had neonatal meningitis. The isolate was cultured overnight on 5% sheep blood agar (SBA; Remel, Lenexa, KS, USA) at 32°C. Before the study, the isolate was passed through 5 Nile tilapia (weight 18.4 ± 0.48 g) 1 time each by intraperitoneal injection of 10⁷ CFU of GBS/fish. Specimens from a fish that died 3 days postchallenge were cultured on SBA, and GBS was recovered from nostrils, intestines, posterior kidney, and brain. One GBS colony isolated from brain was cultured on SBA, typed as GBS by Lancefield grouping (6), and used for experimental infection.

Serial dilutions of the GBS isolate were prepared in tryptic soy broth (TSB; Remel), and 10 fish (weight 28.2 ± 0.51 g) were each injected intraperitoneally with 0.1 mL of inoculum at 10⁷, 10⁶, 10⁵, 10⁴, 10³, or 10² CFU/fish. Ten control fish were injected with 0.1 mL of TSB only. Fish were placed in separate 57-L aquariums at 32°C according to dose and monitored daily for signs of disease and death for 14 days postchallenge. Moribund fish were humanely euthanized by an overdose of tricaine methanesulfonate (MS-222; Argent Chemical Laboratories, Redmond, WA, USA).

Bacterial samples were obtained from nostrils and brains of all moribund or dead fish and cultured on SBA at 32°C for 24 h for GBS. Identification of GBS was performed by using methods of Evans et al. (6) and the BIOLOG MicroLog3 Microbial Identification System (BIOLOG, Hayward, CA, USA) according to the manufacturer’s instructions. BIOLOG results were compared with a Microlog database (www.biolog.com/mID_product.html); a similarity index >0.50 and high probability (>90%) were considered a strong confirmation for GBS.

Figure

Figure. Mortality rates for 60 Nile tilapia at all doses (black line) and 10 tilapia each challenged with a human Streptococcus agalactiae isolate (#510012): 10² (gray line), 10³ (green line), 10⁶ (red...

Within 7 days postchallenge, the 10², 10³, and 10⁷ CFU/fish groups had a mean cumulative mortality rate of 11.7% (7/60) (Figure). Overall, deaths occurred on day 2 and the mortality rate reached 20% after 14 days of observation. Deaths after day 10 occurred in fish that received 10⁶ CFU. Sampled organs were negative for GBS and deaths were attributed to tank mate aggression to weakened fish. A linear dose response was not seen. Deaths occurred at low (10²–10³ CFU/fish) and high (10⁷ CFU/fish) doses but not at median doses (Table). Disease signs in tilapia exposed to human GBS were lethargy, anorexia, dark coloration, opaque eyes, and remaining stationary at the bottom of the tank.

All sampled organs from fish dying within 7 days of infection contained β-hemolytic, gram-positive, oxidase-negative, catalase-negative bacteria. BIOLOG analysis confirmed identification as GBS (similarity index 0.79, probability 100%). None of the TBS-injected control fish showed signs of disease or died.

Deaths among experimentally infected Nile tilapia indicate that an Ia ST-7 human GBS isolate can be pathogenic to fish. Such isolates have been associated with human (4,12) and fish disease as well as with a marine mammal sampled during a GBS fish kill (15). The human isolate was virulent in tilapia at 10², 10³, 10⁶, and 10⁷ CFU/fish, and GBS was reisolated from diseased fish. These experimental findings suggest GBS transmission between mammals and fish and that GBS-induced fish epidemics can originate from mammalian sources. Although we studied only 1 human isolate, other isolates or isolates repeatedly passed through fish may be more virulent. Susceptibility to GBS may also be enhanced by suboptimal environmental conditions, such as low DO, high ammonia levels, euthrophication, harmful algae, and changing or extreme water temperatures (5,6). Future histopathologic investigations may characterize the distribution and nature of the host response to human GBS.