Journal of Fish Diseases 2015 doi:10.1111/jfd.12352
Effects of Loma morhua (Microsporidia) infection on the cardiorespiratory performance of Atlantic cod Gadus morhua (L).
M D Powell1 and A K Gamperl2
1 Norwegian Institute for Water Research, Bergen, Norway 2 Department of Ocean Sciences, Memorial University of Newfoundland, St. John’s, NF, Canada
Abstract Introduction The microsporidian Loma morhua infects Atlantic Microsporidial diseases pose significant challenges cod (Gadus morhua) in the wild and in culture and to the development of marine fish aquaculture, spe- results in the formation of xenomas within the gill cifically in the North Atlantic (Murchelano, filaments, heart and spleen. Given the importance Despres-Patanjo & Ziskowski 1986; Bricknell, of the two former organs to metabolic capacity and Bron & Bowden 2006; Kahn 2009), the North thermal tolerance, the cardiorespiratory perfor- Pacific (Brown, Kent & Adamson 2010) and more mance of cod with a naturally acquired infection of recently the Red Sea (Abdel-Ghaffar et al. 2011). Loma was measured during an acute temperature Of particular significance is the infection of gadoid increase (2 °Ch 1)from10°C to the fish’s criti- fishes [e.g. Atlantic cod (Gadus morhua)] with the cal thermal maximum (CTMax). In addition, oxy- microsporidian Loma morhua; this species is gen consumption and swimming performance were recently identified separately from Loma branchialis measured during two successive critical swimming (Brown et al. 2010). Microsporidian xenomas are ° speed (Ucrit)testsat10 C. While Loma infection characterized by their distinct morphology, with had a negative impact on cod cardiac function at those produced by Loma sp. having a well-defined warm temperatures, and on metabolic capacity in and characteristic thick granular amorphous wall both the CTMax and Ucrit tests (i.e. a reduction of and various developmental stages of the parasite 30–40%), it appears that the Atlantic cod can lar- (Lom & Dykov a 2005). It is believed from data on gely compensate for these Loma-induced cardiore- related species (Loma salmonae) that the na€ıve host spiratory limitations. For example, (i) CTMax ingests spores that enter the host through the gut, ° ~ 1 (21.0 0.3 C) and Ucrit ( 1.75 BL s ) were that the sporoplasm is injected into a host cell that very comparable to those reported in previous stud- migrates to the heart and enters a merogony-like ies using uninfected fish from the same founder phase and finally that the parasite utilizes macro- population; and (ii) our data suggest that tissue phage-mediated transport to the gill where endo- oxygen extraction, and potentially the capacity for thelial and pillar cells hypertrophy to form anaerobic metabolism, is enhanced in fish infected xenomas (Sanchez, Speare & Markham 2000; San- with this microsporidian. chez et al. 2001a; Rodriguez-Tovar et al. 2003). However, L. morhua xenomas regularly occur in Keywords: cardiac performance, critical thermal the heart as well as in the gills and other organs maximum, gill pathology, heart pathology, metab- (Murchelano et al. 1986). Infection with Loma olism, oxygen consumption. leads to a reduced body condition, a decrease in energy stores (liver somatic index), and mild anae- Correspondence M D Powell, Norwegian Institute for Water mia and leukaemia (Khan 2005). In salmonids, Research, Thormøhlens gate, 53D Bergen 5006 Norway such as rainbow (Oncorhynchus mykiss) and brook (e-mail: [email protected])
Ó 2015 John Wiley & Sons Ltd 1 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
trout (Salvelinus fontinalis) that are infected with and the Norwegian Animal Welfare Authority Loma salmonae, decreases in specific growth rate are (FDU). The Atlantic cod G. morhua L. used in correlated with increases in routine and maximum this study (mean body mass SEM, metabolic rate, and reductions in routine but not 509.1 12.3; total length 37.1 0.2 cm; maximum metabolic rate, respectively (Powell et al. N = 42) were raised and maintained at the Dr. 2005): the metabolic cost of the disease is largely Joe Brown Aquatic Research Building (JBARB; attributed to changes in branchial O2 permeability Ocean Sciences Centre, Memorial University of (Powell, Speare & Becker 2006). Newfoundland) in a 3000-L tank supplied with The thermal biology of Atlantic cod has aerated sea water at 10–11 °C for at least received considerable attention over the past dec- 6 months prior to experimentation. The fish were ade (e.g. see Drinkwater 2005; Gollock et al. fed daily with a commercial cod diet (EWOS) 2006; Perez-Casanova et al. 2008a,b; Hori et al. under a photoperiod of 8 hours light: 16 hours 2012). Although cod prefer cooler waters (8– dark. 15 °C; Pettersen & Steffansen 2003), local tem- peratures may fluctuate significantly on a seasonal Critical thermal maximum experiment and diurnal basis, and cod in the wild and in cul- ture may not be able to escape acutely elevated Surgical procedures. Fish were individually netted water temperatures as high as 20 °C (Gollock and anaesthetized in sea water containing et al. 2006; Righton et al. 2010). Given that it is 100 mg L 1 MS-222 (tricaine methane sulpho- widely recognized that limitations in metabolic nate) until ventilatory movements ceased. The fish scope and cardiac output are a key determinant of were then weighed and measured before being the upper thermal limits of fishes (Farrell 2002; transferred to a surgical table where their gills Farrell et al. 2009; Gamperl, Swafford & Rodnick were constantly irrigated with oxygenated, and 2011; Keen & Gamperl 2012; Sandblom et al. chilled, sea water containing MS-222 2014), that Loma infections impact fish metabo- (0.5 mg L 1). To allow for the direct measure- lism (Powell et al. 2005) and that this parasite ment of cardiac function (cardiac output Q; heart forms xenomas in the cod heart and gills, it might rate fH; and stroke volume Hsv), a 2PS Tran- be expected that infection with this microsporidi- sonicTM blood flow probe (Transonic Systems) was an constrains the upper temperature that infected implanted around the ventral aorta using a cod can tolerate and/or reduces the capacity of method modified from Thorarensen, Gallaugher individuals to perform other important physiologi- & Farrell (1996) and Gollock et al. (2006). In cally (and metabolically) demanding functions brief, the left operculum and underlying gills were such as swimming. Thus, we measured the follow- elevated and secured in this position with umbili- ing: (i) the cardiorespiratory physiology (oxygen cal tape, before a small 5–7 mm incision was consumption and cardiac performance) of adult, made at the base of the junction between the sec- 10 °C acclimated, cod with varying severities of ond and third gill arches. The ventral aorta was L. morhua infection when acutely exposed to then located and cleared of the surrounding con- increasing temperatures up to their critical thermal nective tissue by careful dissection, and the flow maximum (CTmax); and (ii) oxygen consumption probe was placed around the ventral aorta anterior in this same population of cod when they were to the intact pericardium. Finally, the cable from subjected to two consecutive critical swimming the flow probe was secured to the skin using 1-0 speed (Ucrit) tests. silk at positions immediately posterior to the oper- culum, just above the lateral line behind the left pectoral fin, and in between the first and second Materials and methods dorsal fin. A PE50 (Clay Adams Inc.) cannula, with a 45-degree bend approximately 2 cm Experimental animals from its tip, was also sutured at the upper All experiments were carried out in accordance edge of the opercular cavity for measurements with the guidelines of the Canadian Council on of ventilation frequency. This cannula was then Animal Care and approved by the Institutional secured to the skin with 1-0 silk sutures using Animal Care committee of Memorial University the same attachment points as for the flow of Newfoundland (protocol number 11–25 KG) probe cable.
Ó 2015 John Wiley & Sons Ltd 2 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
Following surgery, the fish were transferred to transducer. This transducer was calibrated daily 6.5-L tubular clear acrylic respirometers supplied against a static water column. Signals from the with flush and recirculating pumps (EHEIM Transonic flow meter and pressure transducers GmbH & Co. KG) and left to recover overnight were amplified and filtered using a data acquisi- (typically 20–24 h) with the pumps in the ‘flush’ tion system (MP100A-CE; BIOPAC Systems, mode. These respirometers were submerged in a Inc.) and universal interface module (UIM100C; ‘water table’ supplied with temperature-controlled BIOPAC Systems Inc.) connected to a laptop â (10 °C), and aerated, sea water from a large computer running AcqKnowledge software (ver- (~300 L) reservoir; the temperature in the reser- sion 3.8.2; BIOPAC Systems, Inc.). As Transonic voir controlled using a custom-designed heater/ flow probes are temperature sensitive, and the chiller (Technical Services, Memorial University characteristic of this relationship varies between of Newfoundland). probes, the temperature–flow relationship for each probe was measured at different flow rates (5– Experimental protocol. To begin the experiment, 25 mL min 1) over the temperature range used cardiac output Q, ventilation frequency (Vf) and in the experiment using Transonic calibration oxygen consumption (MO2) were measured at tubing, a high precision peristaltic pump, a foam- 10.0 °C. The fish were then challenged with an lined chamber, and saline with a haematocrit of increase in temperature of 2 °Ch 1 until loss of 15–20%. These relationships were then used to equilibrium [i.e. they reached their critical thermal correct in vivo flow rates (Q) at various tempera- maximum (CTMax)], with cardiorespiratory tures. Measurements of MO2, Q, fH and Vf were ° parameters and MO2 measured at every 2 C made for 10 min, just prior to the temperature increase in temperature. This protocol was chosen being increased. to be consistent with previous studies that have From the above data, the following parameters assessed temperature-dependent cardiovascular were calculated: limits and the relative contributions of stroke vol- Cardiac stroke volume ðHSV; in mLÞ¼Q=fH ume (HSV) and heart rate (fH) to increases in Q
with temperature in various fishes (Gollock et al. Arterio-Venous oxygen difference ðCaO2 2006; Mendonca & Gamperl 2010; Gamperl 1 CvO2; in mg O2 mL bloodÞ et al. 2011; Keen & Gamperl 2012). After the ¼ MO2=Q fish lost equilibrium, water temperature was rap- idly decreased to 10 °C, and the fish was removed Perfusion conductance ratio (in mL blood from the respirometer and immediately killed in 1 1 1Þ¼ = 10 °C sea water containing 0.3 g L 1 MS-222. min kg mg O2 Q MO2
Data collection. Oxygen consumption measure- 1 1 Swimming performance ments (in mg O2 kg h ) were made using a â laptop computer running AutoResp software To examine the effect of Loma sp. infection upon (v2.1.0; Loligo Systems) that was interfaced with a swimming performance, infected cod were initially fibre-optic oxygen meter (model OXY-4 mini) placed in a custom-built (Technical Services, and associated with precalibrated dipping probe Memorial University of Newfoundland) 81-L (PreSens) and Loligo’s DAQ 4 and TEMP-4 Blazka-type swim tunnel filled with aerated sea ° modules. To make MO2 measurements, the Aut- water maintained at 10 C, and with water veloc- â oResp software switched between the ‘flushing’ ity set to approximately 0.5 BL s 1. This speed and ‘recirculating’ EHIEM pumps for 8 min allowed the fish to maintain their orientation and (making the respirometer a closed circuit), and position, without having to swim actively. At the began recording water oxygen levels after a 2-min end of a 24-h acclimation period, the swim tunnel wait period. was closed and oxygen consumption was measured Cardiac output (Q, in volts) was recorded by over a 20-min period using a dipping probe that connecting the flow probe leads to a Transonic was inserted into the swim tunnel. This dipping 1 T206 flow meter, whereas Vf (in min ) was probe was connected to a Fibox 3 oxygen meter measured by attaching the opercular cannula to a (PreSens, Precision sensing GmbH) interfaced with Gould Statham (Model P23-10) pressure a portable computer, and the output recorded using
Ó 2015 John Wiley & Sons Ltd 3 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
â Oxyview software (LCDPST3-v2.01; PreSens). then quickly dissected, and the whole heart was The water velocity was then increased in steps carefully removed to prevent damage to the equivalent to 0.2 BL s 1, with each step comprised atrium or bulbus arteriosus. After rinsing the heart of equivalent (10 min) flush and measurement in saline, it was blotted dry and weighed, the periods. The water velocity was increased until the atrium and ventricle were weighed separately, and fish was unable to move away from the grid at the the ventricle measured for length (VL), width back of the swim chamber, at which point the (VW) and height (VH) as described by Powell, velocity was reduced to 0.5 BL s 1 and the fish to Nowak & Adams (2002) and Powell, Burke & allowed to recover. Oxygen consumption measure- Dahle (2011, 2012). The chambers of the heart ments were also made after 30 and 60 min of were then fixed in 10% neutral buffered formalin recovery at ~0.5 BL s 1, and the fish was subse- for histological examination and xenoma quently given a 2nd swim test (without measuring quantification. oxygen consumption) prior to being killed in The filaments on the second gill arch on the 0.3 g L 1 of MS-222. right side were subsequently separated from the From these data, the following parameters were arch, and 40–80 individual filaments were then calculated: digitally photographed (at 109 magnification ; ð Þ including an internal scale) and the number of Critical swimming velocity Ucrit visible xenomas recorded. The fixed ventricles ¼ U þ½U ðt =t Þ f i f i were also digitally photographed, and the number
where Uf is the last velocity at which the fish of visible xenomas on the ventricle surface completed the entire period (20 min), Ui is the recorded. 1 water velocity increment (0.2 BL s ), tf is the Formalin-fixed gills and hearts were then – time to fatigue and ti is the duration of each embedded in paraffin wax and sectioned at 3 velocity step. 5 µm, stained with haematoxylin and eosin and examined using an Olympus BX51 microscope ð ; 1 1Þ Gross cost of transport GCOT in mg kg m with an Olympus DP71 camera (Olympus Life ¼ð Þ=ðð Þ Þ MO2i Ui L 36 and Material Sciences, Europa Gmbh) using Cell^B image software (Olympus Soft Imaging where MO2i is the oxygen consumption at a given Solutions Gmbh). velocity (Ui) and L is the length of the fish. Standard oxygen consumption (MO2standard) was derived from a semi-log plot of swimming Cardiac measurements and morphometrics speed vs. log MO2, and using the derived linear 1 From measurements of the heart and its cham- regression to extrapolate back to 0 BL s . Rou- bers, the following parameters were calculated: tine oxygen consumption was that measured in fish that were resting quietly in the swim tunnel Cardiac, Ventricular and Atrial somatic indices; at 0.5 BL s 1. Maximum oxygen consumption CSI, VSI and ASI, respectively. (MO2max) was the highest oxygen consumption CSI = heart mass/fish mass; VSI = ventricular that each individual fish achieved, and metabolic mass/fish mass; ASI = atrial mass/fish mass. scope was calculated by subtracting standard oxy- % contribution to heart mass (%V, %A): % gen consumption from MO2max. V = ventricular mass/heart mass; %A = atrial mass/heart mass. Ventricular shape ratios (L:W, L:H, H:W): L: Determination of the level of Loma morhua W=VL/VW;L:H=VL/VH; H:W = VH/VW. infection
After the fish were killed, the second gill arch on Statistical analyses the left side was removed and fixed in 10% neu- ƒ tral buffered formalin for histological examination The data for MO2, Q, H, HSV, arterial-venous and confirmation of Loma sp. infection and the (AV) oxygen difference, perfusion conductance, second gill arch on the right side was removed Vƒ, and GCOT were all analysed using a one-way and placed in phosphate-buffered saline (PBS) for repeated-measures analysis of variance (RM ANO- branchial xenoma quantification. The fish was VA). Where the data were not normally
Ó 2015 John Wiley & Sons Ltd 4 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
(a) (b)
(c) (d)
(e) (f) Figure 1 Images of gill filaments from Atlantic cod that were lightly (a) and heavily (b) infected with Loma morhua. The white ‘spots’ are xenomas. Histological sections of Loma morhua xenomas in the gills of lightly infected (c) and heavily infected (d) Atlantic cod (bar = 500 µm) showing limited (e) and extensive (f) filamental epithelial hyperplasia and (g) (h) mononuclear infiltration (*) (bar = 100 µm). Xenomas were also located in the gill arch at the base of the gill filaments (g, bar = 500 µm), where extensive inflammatory infiltration (*) and spongious hyperplasia (S) were associated with localized haemorrhage and xenoma disruption (h) (bar = 100 µm). Arrows indicate xenomas.
distributed, Freidman’s RM ANOVAs on ranks surface of the heart. Comparisons between data were conducted. Individual differences were iden- collected in this study and those reported in Gol- tified using either Student–Newman–Keuls or Tu- lock et al. (2006) and Petersen & Gamperl key’s post hoc analyses. The routine metabolic (2010) were performed using unpaired t-tests. All rate of cod scales allometrically with body mass analyses were performed using SigmaPlot 10.0 with a slope of 0.8–0.85 (Post & Lee 1996; Killen and SigmaStat 3.5 (Systat Software Inc.) graphical et al. 2007). However, we report isometrically and statistical software. All values in the text, scaled metabolic rate data for comparison with tables and figures are means + 1 standard error previous studies on cardiorespiratory function in (SE). cod (Gollock et al. 2006; and Petersen & Gam- perl 2010). Furthermore, the range of fish masses Results was small, and thus, the error in using isometric scaling was minimal. Pearson product–moment All of the fish used in this study were infected correlations were used to examine the relationship with L. morhua as determined by the presence of between the measured and derived parameters and xenomas on the gills (Fig. 1a,b) and heart the number of xenomas per filament or on the (Fig. 2a,b). However, the number, size and
Ó 2015 John Wiley & Sons Ltd 5 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
(a) (b)
(c) (d)
(e) (f)
Figure 2 Image of lightly (a) and heavily infected (b) Atlantic cod ventricles showing white xenomas. Histological sections (g) (h) showing Loma morhua xenomas in the atrium (c), ventricle (d) and bulbus arteriosus (e) of the heart, with an accompanied fibrocytic and granulomatous inflammatory reaction (bar = 200 µm). Endocardial (f) and epicardial (g) and bulbus arteriosus endothelial (h) xenomas with limited inflammatory reaction (bar = 50 µm). Arrows indicate xenomas.
severity of the infection varied considerably small spores, and were often associated with a between individuals, with some fish having few fibrocytic or granulomatous tissue reaction pathological signs of infection (Fig. 1c,e), whereas (Fig. 1g). In the heart, xenomas were observed in others showed extensive xenoma formation and the atrium, ventricle and bulbus arteriosus branchial hyperplasia; the latter often associated (Fig. 2c–e). Some of these were surrounded by an with the filament tips and with fusion of gill amorphous eosinophilic coat and a notable granu- lamellae (Fig. 1d,f,h). The level of infection in the lomatous reaction (Fig. 2c–e), whereas those asso- gill ranged from 0.6 to 26.8 xenomas per filament ciated with the endo- and epicardium generally (7.6 1.1, median 5.5), whereas in the heart the showed no apparent inflammatory reaction number of xenomas ranged from 0 to 35 (Fig. 2f–h). As a population, the density of bran- (5.9 1.30, median 3.00) and they had an aver- chial xenomas was not predictive of cardiac xeno- age density of 14.8 3.3 xenomas g 1 ventricle ma density (Pearson correlation (median 7.0; range of 0– 91.9) as determined by coefficient = 0.130, P value = 0.417). The only external examination. The xenomas in the fila- significant correlation between morphological ments, both in the lamellae and at the base of the parameters and the number of xenomas or xeno- filaments on the gill arch, showed a characteristic ma density was a negative correlation between amorphous external layer, contained numerous condition factor (K) and cardiac xenoma density
Ó 2015 John Wiley & Sons Ltd 6 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
Table 1 Pearson correlation coefficients and P values for the 20 °C), and the perfusion convection requirement relationships between heart-somatic and morphological parame- decreased significantly with temperature (by 30%; ters, and (1) the mean number of branchial xenomas per fila- from 0.30 0.02 at 10 °C to 0.21 0.02 at ment and (2) the total number of visible cardiac xenomas per ° g ventricle mass in the Atlantic cod population used in the 20 C; Fig. 3g) (Freidman RM ANOVA on ranks 2 = < study HSV; X 5 24.4, P 0.001: MO2/Q; 2 = < X 5 31.3, P value 0.001). Finally, ventilation Gill xenomas Cardiac xenomas filament 1 g 1 ventricle frequency (Vf) increased significantly with increas- ing temperature (by 1.6-fold; from 40.2 1.4 Parameter Pearson P value Pearson P value at 10 °C to 63.0 1.7 ventilations min 1 at ° = K 0.081 0.615 0.522 <0.001 20 C; Fig. 3e) (RM ANOVA F5,119 83.4, P THM 0.137 0.393 0.242 0.122 value <0.001). TVM 0.023 0.888 0.196 0.214 TAM 0.107 0.504 0.259 0.098 There was no significant correlation between CSI 0.101 0.530 0.085 0.591 the number of branchial xenomas and oxygen VSI 0.062 0.698 0.171 0.278 consumption at most temperatures. However, ASI 0.090 0.575 0.098 0.538 %V 0.097 0.548 0.231 0.141 there was a strong negative correlation between %A 0.053 0.741 0.167 0.290 these two parameters at 18 °C (Pearson correla- VL 0.105 0.512 0.0205 0.897 tion coefficient 0.58, P = 0.002) (Fig. 4a). VW 0.088 0.584 0.018 0.908 Cardiac output was not significantly correlated VH 0.026 0.871 0.0643 0.686 L:W 0.112 0.487 0.026 0.870 with either branchial xenomas per filament or L:H 0.119 0.459 0.057 0.719 cardiac xenoma density at any of the tempera- H:W 0.069 0.670 0.193 0.222 tures (data not shown), and there were no sig- K represents condition index; THM, TVM and TAM are total heart, nificant correlations between heart rate and ventricle and atrial masses; CSI, VSI and ASI are cardio-, ventricular- branchial xenomas per filament. Cardiac stroke and atrial-somatic indices; %V and %A are the relative proportion of cardiac mass contributed by ventricle and atrium, respectively; V , V volume and Qmax were also both negatively, but L W > < and VH are ventricular length, width and height, respectively; and L:W, marginally (i.e. 0.01 P 0.05), correlated L:H, H:W are ratios between the various measurements. with cardiac xenoma density (Hsv; Pearson corre- lation coefficient = 0.36, P value = 0.093: Q; (Pearson correlation coefficient = 0.522, P value Pearson correlation coefficient = 0.37, P <0.001) (Table 1). value = 0.093) (Fig. 4b,d). However, heart rate was positively correlated with cardiac xenoma density at all temperatures from 10 to 18 °C CT experiment max (Fig. 4c). Oxygen consumption increased as temperature The mean critical thermal maximum (CTMax) rose and reached a maximum value at 18 °C that for this group of cod was 21.0 0.3 °C. Upon was approximately twofold higher than measured reaching this upper temperature limit, all cardiore- ° 1 1 at 10 C (73.5 3.5 mg O2 kg h ) (Freid- spiratory variables with the exception of HSV (i.e. 2 = man RM ANOVA on ranks X 5 102.8, P value Q, fH, and Vf) were substantially lower as com- <0.001) (Fig. 3a). Similarly, there were significant pared with values recorded at 20 °C (Fig. 3b–e). temperature-induced increases in cardiac output There was a large degree of variation in the values ° (Q, by 1.45-fold; from 21.5 1.1 at 10 Cto of these parameters at CTmax, and thus, there 31.1 2.4 mL kg 1 min 1 at 20 °C; Fig. 3b), were no statistically significant correlations heart rate (fH, by 1.7-fold; from 46.8 0.9 at between CTmax and either branchial xenomas 10 °C to 80.1 2.3 beats min 1 at 20 °C; (Pearson correlation coefficient = 0.017, P Fig. 3c) and the AV difference in blood oxygen value = 0.94) or cardiac xenoma density at this content (by 1.45-fold; from 3.7 0.30 at 10 °C temperature (Pearson correlation coeffi- to 5.4 0.5 at 20 °C; Fig. 3f) [Freidman RM cient = 0.141, P value = 0.492) (Fig. 5). 2 = ANOVA on ranks (Q; X 5 38.8, P value Compared with the data presented by Gollock < 2 = < 0.001: fH; X 5 113.7, P 0.001: CaO2-CvO2; et al. (2006) for uninfected fish of the same stock 2 = < ° X 5 33.1, P value 0.001)] (Fig. 3f). In con- origin, CTmax was only approximately 1 C lower ° = =< trast, cardiac stroke volume (HSV; Fig. 3d) (21 vs. 22.2 C; t33 6.11, P value 0.001). decreased slightly with temperature (by 14%; from This result was despite maximal MO2 and Q ° = =< 0.46 0.03 at 10 C to 0.40 0.03 mL at (MO2; t33 6.900, P value 0.001: Q;
Ó 2015 John Wiley & Sons Ltd 7 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection ) –1 h
160 ) –1 36
(a) c c –1 (b) kg b 2 34 kg
140 –1 bc 32 b
(mg O 30 2 b 120 b 28 a 26 a 100 ab 24 a 80 22 a 20
60 Cardiac output, Q (mL min 18 8 10121416182022 8 10121416182022 Oxygen consumption, MO Temperature oC Temperature oC
90 0.52 (c) (d) ) cd 0.50 a a –1 80 d (mL) 0.48 sv 0.46 c ab 70 0.44 ab
(beats min bc 0.42 ab H b 60 0.40 ab 0.38 50 a 0.36 Heart rate, f 0.34 Cardiac stroke volume H 40 0.32 8 10121416182022 8 10121416182022 Temperature oC Temperature oC )
–1 70 6.5 (e) (f) b e 65 ) 6.0 b –1 60 d 5.5 b
55 c 5.0 b ) (mg mL 2 (Ventilations min f O 50 c v 4.5 ab - C 45 2 4.0 a a O a a
(C 3.5 40 Arterial-venous difference
35 3.0 Ventilaton rate, V 8 10121416182022 8 10121416182022 Temperature oC Temperature oC
0.34 ) a
–1 (g) 2 0.32 ab
mg O 0.30 –1 0.28 kg b –1 0.26
0.24 bc bc bc (mL min
2 0.22
0.20 Q/MO 0.18 8 10121416182022 Temperature oC
Figure 3 Oxygen consumption (a), cardiac output (b), heart rate (c), cardiac stroke volume (d), ventilation rate (e), arterial-venous
blood O2 content difference (f) and perfusion convection requirement (Q/MO2) (g) of Loma-infected Atlantic cod acutely exposed to increasing temperature at 2 °Ch 1. Points (temperatures) without a letter in common are significantly different at P < 0.05. The open symbol represents values observed just after the fish reached their CTMax. Values are means 1 SE.
Ó 2015 John Wiley & Sons Ltd 8 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
(a) Pearson correlation = –0.362 P value = 0.090 (b) Pearson correlation = 0.141 P value = 0.492 0.8 24
22 0.6 C)
o 20 0.4 18 CTMax ( 0.2 16 Cardiac stroke volume (mL)
0.0 14 0 10203040 0 10203040 –1 Cardiac xenoma density (xenomas g ventricle) Cardiac xenoma density (xenomas g–1 ventricle)
60 (c) Pearson correlation = –0.371 P value = 0.093 (d) Pearson correlation = –0.0165 P value = 0.936 ) 24 –1 50 kg
–1 22
40 C)
o 20
30 18 CTMax (
20 16 Cardiac output (mL min
10 14 0 10203040 0 5 10 15 20 25 30 Cardiac xenoma density (xenomas g–1 ventricle) Branchial xenomas filament–1
Figure 4 The relationship between (a) routine oxygen consumption rate at 18 °C and the mean number of branchial xenomas per gill filament; and between (b) stroke volume, (c) cardiac frequency at temperatures ranging from 10 to 18 °C and (d) cardiac output
and cardiac xenoma density in Atlantic cod infected with Loma morhua and used in the CTMax experiment.
= =< = = t33 4.86, P 0.001) being 30 and 40% (2006) (t33 1.827, P value 0.077) and the lower, respectively, in this study. Resting resting and maximal values for AV difference were = = (t33 2.49, P value 0.018) and maximum only significantly different for Loma-infected fish = < = = HSV (t33 6.11, P value 0.001) were also (t50 3.403, P value 0.001) (Table 2). lower in the infected cod, and this was compen- sated for by an increase in resting heart rate (by = < Swimming performance 60%; t33 5.75, P value 0.001) but not maxi- = = mum fH (t33 1.123, P value 0.270). Neither Oxygen consumption increased significantly with = = resting (t33 1.123, P value 0.270) nor maxi- water velocity and reached a maximum value = = mal (t33 1.123, P value 0.270) perfusion con- (approximately twofold resting values) at 1 = ductance ratios (Q/MO2) were statistically 1.5 BL s (RM ANOVA F5,89 13.82, P value significant from those calculated from Gollock <0.001). This swimming speed was approximately 1 et al. (2006). However, Q/MO2 was different 0.2 BL s below the cod’s critical swimming between rest and maximal values for Loma- velocity (1.71 BL s 1; Fig. 6). After a 1-h recov- = = 1 infected fish (t50 3.182, P value 0.003) ery interval at 0.5 BL s , oxygen consumption (Table 2). Similarly, while there were no signifi- had, on average, only decreased to 77.1% of max- cant differences between this study and Gollock imum. Nevertheless, the Ucrit determined in the 1 et al. (2006) with regard to AV difference at rest second Ucrit trial was 1.76 BL s . Gross cost of = = (t33 0.261, P value 0.796), the maximal transport decreased with swimming velocity up to value for AV difference for Loma-infected fish was the critical swimming speed (Ucrit) (Freidman RM 2 = < higher than that derived from Gollock et al. ANOVA on ranks X 6 68.981, P value 0.001)
Ó 2015 John Wiley & Sons Ltd 9 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection
(a) Pearson correlation = 0.141 P value = 0.492 24
22 C) o 20
CTMax ( 18
16
14 010203040 Cardiac xenoma density (xenomas g–1 ventricle)
P (b) Pearson correlation = –0.0165 value = 0.936 24
22
20 C) o
18 CTMax (
16 Figure 5 The relationship between CTMax and cardiac xenoma density (a) and 14 number of branchial xenomas per filament 0 5 10 15 20 25 30 (b) in Atlantic cod infected with Loma –1 Branchial xenomas filament morhua.
(Fig. 6). There were no significant correlations Discussion between, standard, resting (i.e. at 0.5 BL s 1)or maximum oxygen consumption, Ucrit1 or Ucrit2, The Atlantic cod population used in this study and either gill xenomas per filament or cardiac xe- was naturally infected with the microsporidian noma density. L. morhua. Nonetheless, the fish exhibited variable In comparison with data presented in Petersen & levels of infection, with all fish having branchial Gamperl (2010) for uninfected cod from the same xenomas and cardiac xenomas as revealed through founder population, maximal MO2 histological examination, but not external ventric- (183.24 10.31 as compared to 234.60 21.20; ular xenomas. The degree of host tissue inflamma- = = t22 2.444, P value 0.023) and metabolic tory responses also varied greatly between scope (95.51 11.97 as compared to individuals with some fish exhibiting xenomas = = 152.10 20.70; t19 2.463, P value 0.023) with apparently little or no inflammatory reaction, were significantly lower. In contrast, there were no as compared to others where extensive branchial = significance differences between Ucrit (t22 0.227, hyperplasia or fibrocytic and granulomatous reac- = = P value 0.822), standard MO2 (t19 1.993, P tions were observed. Extensive hyperplastic and = = value 0.061) or routine MO2 (t22 0.542, P granulomatous reactions, and the associated infil- value = 0.593) between the two studies (Table 3). tration of mononuclear cells, are characteristic of
Ó 2015 John Wiley & Sons Ltd 10 Journal of Fish Diseases 2015 M D Powell & A K Gamperl Effects of Loma morhua infection