Hole‐in‐the‐head disease in discus fish, Symphysodon (Heckel, 1840): Is it a consequence of a dietary Ca/P imbalance?

Abstract Hole‐in‐the‐head (HITH) disease‐affected fish develop characteristic lesions in the skin above sensory pores of the head and the trunk. This study investigated whether an unfavourable Ca/P ratio in the diet could provoke lesions consistent with HITH disease in discus fish Symphysodon (Heckel, 1840) as a comparable condition to secondary hyperparathyroidism of tetrapod species. Two groups of five fish were fed a plain beef heart diet (Ca/P of 0.03), whereas two other groups were kept on commercial discus feed (Ca/P of 2.73). Each feeding group was submitted to two different water hardness regimes (35.66–71.39 mg/L CaCO3 and 124.94–196.33 mg/L CaCO3, respectively). All fish were observed for the development of the characteristic lesions for 16 weeks. At the end of the study, histological, bacteriological and parasitological examinations were conducted and plasma Ca, P and Mg values were determined. Diplomonad flagellates were detected in two fish. Isolated bacteria of all groups mostly belonged to Aeromonadales and Pseudomonadales. No significant difference of plasma mineral values between the groups was observed. Compared to the results of other authors, Ca stayed mainly in the range and P exceeded the reference values. Histological examinations did not indicate HITH disease, and no fish developed signs of the disease during the study. Clinical trial registration number GZ 68.205/0135‐WF/V/36/2014.

on the subject are scarce. Most of the authors mentioned below agree that unfavourable and stressful environmental conditions may play a role in the development of the disease. Among the proposed specific causes, Paull and Matthews (2001)
A feeding regime exemplary for such mineral imbalance is the use of raw beef heart as a source of protein, which was suggested by Chong, Hashim, Chow-Yang, and Ali (2002) for commercial discus farming. The authors found good digestibility values for dry matter and protein of the raw beef heart, resulting in accelerated growth and intense coloration of the fish. Thus, feeding beef heart to discus has become very popular among breeders and hobbyists, often with no or insufficient supplementation (Untergasser, 1991).
Nevertheless, Ca contents of beef heart are far below suggested for the diet of the comparable species Cichlasoma urophthalmus (0.6 g/ kg vs. 1.8 g/kg), while P is exceeding recommendations (20.6 g/kg vs. 1.5 g/kg), leading to a Ca/P of 0.03, while ratios of 1.3 are recommended (Chavez-Sanchez, Martinez-Palacios, Martinez-Perez, & Ross, 2000). Such low-Ca and high-P diets are known to cause alimentary secondary hyperparathyroidism in non-aquatic species like horses (Lacitignola et al., 2018), rabbits (Bas et al., 2005), dogs (Kawaguchi, Braga, Takahashi, Ochiai, & Itakura, 1993), reptiles (Mans & Braun, 2014), cats (Tomsa et al., 1999) and birds (Wallach & Flieg, 1969). This condition is caused by calcium deficiency or an imbalance of Ca/P in the diet and leads to hypocalcaemia, which causes increased release of parathyroid hormone, resulting in chronic bone resorption.
In teleost fish, the role of parathyroid hormone-related protein (PTHrP) for Ca and P metabolism was demonstrated for some fish species (Abbink, 2004;Abbink & Flik, 2007;Guerreiro, Renfro, Power, & Canario, 2007). Abbink and Flik (2007) suggest a role of PTHrP for Ca mobilization from bones and scales while Witten and Huysseune (2009) attribute bone resorption rather to a lack of P than to Ca deficiency. These authors also state that the teleost endoskeleton is not used as a source of minerals when in demand.
On the other hand, Takagi and Yamada (1993) showed that in tilapia (Oreochromis niloticus) Ca depletion led to bone resorption and decreased bone Ca and P contents. While the endocrine mechanisms regulating the Ca/P balance differ substantially from terrestrial animals, decreasing Ca/P in the diet of carp (Cyprinus carpio) was demonstrated to decrease the Ca absorption rate significantly (Nakamura & Yamada, 1980). Until now, evidence of bone resorption or involvement of an impaired Ca and P metabolism in HITH disease has not been described.
This study was conducted to determine whether feeding a plain beef heart diet would cause HITH disease in discus fish as a result of an altered osseous metabolism comparable to secondary hyperparathyroidism of tetrapods. Furthermore, a possible correlation between the selected dietary regimes, water parameters and plasma values of the concerned minerals should be evaluated. Twenty-four discus fish (captive bred, mixed gender, mean length 11.7 ± 4.5 cm, mean weight 57.45 ± 52.8 g) from a commercial breeding facility were assigned at random to four groups (A-D)

| MATERIAL S AND ME THODS
of six animals each. One fish of each group was killed and sampled upon arrival as described below to serve as a reference (Ref), while the other fish were acclimated for one week. During acclimatization, the groups were maintained in 250-L recirculation tanks in filtered and UV-treated well water with pH 7.0 and 124.9-196.3 mg L CaCO 3.
Each tank was filtered individually, and temperature was set to 30°C.
During that time, all fish were fed a commercial discus diet (JBL Grana Discus®, JBL). After one week of acclimatization, two groups were designated as experimental groups (A and C) and two as controls (B and D). Water CaCO 3 content in one experimental (A) and one control group (B) was changed to 35.7-71.4 mg L CaCO 3 . Diet was switched to beef heart in the experimental groups (A and C), whereas the commercial feed was retained in the control groups (B and D) ( Table 1). All groups were fed twice a day at 2% of their body weight. The raw beef heart diet was prepared by removing all tendons and fat from the heart, grinding the meat, and stored at −20°C.
Diet contents of Ca, P and Mg were analysed by the Austrian Agency for Health and Food Safety, Ltd. (AGES).
Carbonate hardness, as well as temperature and pH, was controlled daily. Temperature and pH were determined using a multiparameter portable meter (WTW multi 3420, Xylem Analytics). Water CaCO 3 content was calculated daily from the results of complexometric hardness titration with hydrochloric acid. Based on these results, demineralized water was added as needed to the tanks to maintain the desired water CaCO 3 content during the experiment.
Determination of the needed amounts of demineralized water was accomplished by the use of a graph, which was established before the experiment started. Other divalent ions contributing to total water hardness were negligible.
Lighting was provided for nine hours daily at true light quality (2× 15-W T8 Sylvania®, Feilo Sylvania). The duration of the experiment was 16 weeks.

| Examination of the fish
Upon arrival, all fish were measured and weighed. Photographs of all fish were taken from both sides of the head, with emphasis of the supra-and infraorbital region and the lateral line region with a digital camera (NIKON D300s digital camera, Nikon Corp.).
Reference fish (one of each group) were killed and sampled in the same way as the other fish at the end of the experiment. All other fish were acclimated and assigned to the different groups as described above. Photodocumentation was repeated after four, eight and twelve weeks and at the end of the study (16 weeks).
During the experiment, all fish were examined daily for the devel-

| Statistical analysis
The hypothesis of this study was that a severe imbalance of dietary Ca and P may lead to HITH disease. A plain beef heart diet meets these criteria and was therefore fed to half of the fish in the experiment (n = 10). The other half served as a control group and was fed a commercial maintenance diet. Should the hypothesis prove true, all fish of one group should be affected. Thus, the number of used animals was kept low. Besides the mere observation of signs of HITH disease, various quantitative parameters were collected, which allowed a cautious comparing of the groups.
To reveal possible statistical correlations between water parameters, feeding regime and plasma levels of Ca, P and Mg, the following tests were used in IBM® SPSS version 24.0: descriptive statistic to provide the mean and standard deviation for each evaluated parameter, a test of between-subjects effects to show whether either of the two independent variables (feeding regime F I G U R E 1 Exemplary discus from beef heart group at the beginning (a and b) and at the end (c and d) of the study.

| RE SULTS
For the duration of the study, no mortalities occurred and all fish fed well. Length gain rates (LGR) were 5% for groups A and C and 5.96% for groups B and D, while weight gain rates (WGR) were 10.16% and 7.76% respectively (Table 1). Macroscopic observations and evaluation of the photographs revealed no lesion development consistent with HITH disease and no obvious changes of sensory pores in any of the fish (Figure 1).
Results of the food analysis are given in Table 2 Pseudomonas sp. were found in all groups except for one fish of group B, where Ochrobactrum anthropi and Acinetobacter baumannii were detected. In fish of group B, also Proteus vulgaris was identified (Table S1).
Plasma values for Ca, P and Mg were determined (Table 3). The results of plasma Ca and P of reference fish were lower compared to the results of fish from groups A-D, and this also applies to Ca values of the reference fish compared to other studies (Table 4)

| Statistical results
The results of the descriptive statistics demonstrated the highest mean plasma value of Ca and P in group B (commercial discus food, low water Ca), whereas the highest mean value of the plasma levels TA B L E 2 Results of diet analysis and dietary requirements regarding Ca, P and Mg (g 1000g -1 wet weight) Beef heart (g 1000 g −1 ) Commercial diet (g 1000 g −1 )

Dietary requirements (g 1000 g −1 )
Ca 0.06 20. of Mg was detected in group A (fed beef heart, low water Ca). The test of between-subjects effects revealed no statistically significant interaction between feeding regime and blood parameters (Table   S2). There was a significant effect of water CaCO 3 on plasma Ca levels (p = 0.010) in any group. Accordingly, the one-way ANOVA showed a statistically significant difference of plasma Ca levels between the different levels of CaCO 3 in the water (p = 0.012). Finally, Fisher's LSD post hoc test illustrated a significant difference of mean plasma Ca levels between groups B and C (p = 0.012) and groups B and D (p = 0.006) and of Mg between groups A and D (p = 0.020) and groups C and D (p = 0.039) (Table S3). Still, these results are biased by the high SD (>1) in group B for Ca and the small number of specimens use. No significance was revealed for P. In summary, the feeding regime did not cause statistically significant differences of Ca, P and Mg plasma levels between the groups. Water CaCO 3 did cause a significant difference in plasma Ca.

| D ISCUSS I ON
The study was conducted to investigate whether an imbalanced beef heart diet could cause HITH disease in discus fish. In this context, several additional parameters were evaluated to improve assessment of the condition. During the experiment, no signs of HITH disease were evident in any fish despite the unfavourable Ca/P ratio of both diets. The growth rate of the fish in all groups was rather low when compared with other studies (Tibile et al., 2016), despite good food intake. While the growth rate did not differ substantially between the groups, weight gain was higher in the beef heart groups. In a study of Wen, Chen, Qu, and Gao (2018), discus on a plain beef heart diet showed significant higher WGR (90.33%) and LGR (60.84%) than in our study. These differences can be explained by different feed quantities and the age of the fish, assumingly plus other environmental factors. Moreover, the small number of fishes per group did not allow a statistical sound statement regarding growth and weight gains.
In context with a publication by Peyghan, Boloki, and Ghorbanpour (2010) who described HITH disease associated with the isolation of Aeromonas hydrophila, bacteria were cultured from the sensory pores and the surrounding skin of each fish in our experiment. All isolated bacteria represented only a few orders (Table S1) and can be found in waters and biofilms as opportunistic microbes (Drzewiecka, 2016;Walczak, Puk, & Guz, 2017). Despite their ability to cause diseases and mortality of fish under favourable conditions (Austin & Austin, 2016), this did not apply to any fish in this experiment. However, it is notable that the bacteria isolated from fish from group B differed considerably in their species composition compared to the other groups. It must be assumed that fish of all groups showed a similar composition of bacteria upon arrival. This assumption is supported by isolates from the reference fish (Table   S1). Thus, either a shift of the bacterial population or the introduction of bacteria during the experiment might have happened, which we consider of no significant importance for the experiment.  Parasitic examinations were included in the study because parasitic infestation may serve as an indicator for a compromised health status in general (Williams et al., 2013) but also may cause disease itself, thus being an important unspecific factor for the anticipated development of HITH disease. On the other hand, protozoa belonging to the genus Spironucleus have been suggested as the causative agent of HITH disease (Bassleer, 1983;Herkner, 1969Herkner, , 1970Paull & Matthews, 2001). Except for one fish of group C (Dactylogyrus sp. in the gills), no external parasites were detected in fish from this study. bone structures were identified, and no evidence of HITH disease as described by Morrison, O'Neil, and Wright (2007) was observed.
The macroscopic and histological comparison of healthy and HITH disease-affected discus fish is shown in Figure 2. To demonstrate a possible connection between selected plasma values and the development of HITH disease, the analysis of Ca, P and Mg values was included in the study. Haemolysis, which may result in increased levels of Ca, P and Mg (Mirghaed, Ghelichpour, Hoseini, & Amini, 2017), was prevented by appropriate handling of the samples. MS 222 anaesthesia should not impact Ca, P and Mg levels (Velíšek, Stejskal, Kouřil, & Svobodová, 2009). Among environmental factors, the water Ca contents and the availability of dietary P are mostly responsible for meeting the requirements of the two minerals in fish (Nose & Arai, 1979).
Discus are endemic to the Amazon River Basin and inhabit black water streams as well as white and clear water regions. Water chemistry, including Ca contents of these waters, differs considerably, and it has been shown that the Ca content of tail vertebrae from discus differed accordingly to their habitat (Geisler & Schneider, 1976 to soundly confirm the differences, a higher number of samples would have been required. Nevertheless, comparing the results of all measured plasma levels, they show a quite homogenous picture with similar Ca and high P levels when compared to the references as given in Table 4. The high P results from our study are only in accordance with the results of Mauel, Miller, & Merrill, 2007. These authors contribute the higher values of some of the parameters in their study to high stocking density and differences between species/strains. Our discus were stocked at low densities, and the high plasma P levels most likely result from the excessive amounts of P in the diets, which contrasts with the much lower natural intake of the mineral with insects as the main source of protein in the wild (Bleher, 2010). We also compared our values with former results from four healthy discus, which were fed a mixed diet and found comparable results for Ca and Mg, whereas the value of P was much lower in this group (Table 4), which gives an additional indication that the high P contents of the diets resulted in high plasma values of the mineral. Otherwise, the high plasma P values were comparable to the archived results from a HITH diseased discus, whereas this fish also had a high plasma Ca content. In conclusion, our study indicates that an unfavourable Ca/P in the diet of discus may result in altered P plasma levels and justifies further research with larger numbers of animals to verify these results. On the other hand, an unfavourable Ca/P in the diet did not provoke HITH disease or evidence of bone resorption or remodelling during the observation period under the selected environmental conditions. Thus, HITH disease was not observed in discus fed a plain beef heart diet.

ACK N OWLED G EM ENTS
We would like to thank Diskuszucht Stendker for providing the fish and the team of the Central Laboratory of our University for processing the blood samples. We are also thankful to Ass. Prof. Dr. M.
Reifinger and Mag. P. Wernsdorf for technical assistance in histology and Christopher Pichler, MSc., for animal care.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.