Controlling the cannibalism of African catfish juvenile by 17β‒estradiol hormone administration and the stocking density determination

An effort to increase the production of juvenile catfish is limited by the high mortality rate, especially caused by cannibalism. The hormonal treatment has been conducted as an effort to control cannibalism. This study used completely randomized factorial design, consisted of six treatments and three replications. There were two factors examined in this study, the first factor was different doses of 17β‒estradiol hormone (0, 30, and 60 mg estradiol‒17β/kg) incorporated in the feed, and the second factor was stocking density (150 and 300 fish/m2). Juvenile catfish with the size of 4.0 ± 0.1 cm were reared for 30 days in the 84 L aquarium. The results showed that hormone treatment could reduce cannibalism rate, type-I and II cannibalism compared to control (P<0.05). The lowest of mortality was obtained in the treatment B (17β‒estradiol administration of 30 mg/kg, at stocking density of 150 fish/m2; P<0.05). The lowest cortisol level was found in treatment F, and the highest estradiol level was found in treatment F (17β‒estradiol administration of 60 mg/kg, at stocking density of 300 fish/m2; P<0.05). The lowest of blood glucose level was found in treatment B (P<0.05). The highest specific weight growth was found in the stocking density of 150 fish/m2 (P<0.05). The results of this study indicated that administration of 17β‒estradiol in feed could reduce the level of cannibalism in African catfish juvenile.


INTRODUCTION
African catfish Clarias gariepinus is one of the most popular freshwater species in Indonesia. Domestic production of African catfish showed a promising level with an increase of 26.43% per year (KKP, 2014). However, it cannot be said that African catfish culture was free of challenge. Cannibalism is one of the major challenges faced by the farmers because it impacted the survival. Król and Zakes (2016) supported the later statement that cannibalism contributes more than 50% of mortality during the rearing period. It is believed that intraspecific aggressive behavior is found in most catfish cannibalism cases (Onwuteaka & Prince, 2015). Particularly, it distracted most farmers because of its impact on production and undoubtedly profit. Cannibalism is often stated as killing and consuming amongst similar species, partially or entirely of the prey (Liu et al., 2017). Cannibalism frequently happens on catfish started from larvae stadium (Mukai et al., 2013). According to Obirikorang et al. (2014), cannibalism on catfish commonly occurs on fingerling amongst 700-1900 hours or 29-79 day old. At this particular time, cannibalism contributes 42.5% of mortality. Size sorting is required to be conducted regularly and low stocking density will help to reduce cannibalism. Nevertheless, the previously mentioned methods are high labor and need a wider area. It is not costeffective considering that catfish can be cultured in high stocking density.
Catfish broodstock has a high testosterone content in the gonad maturation phase and right before ovulation (Zairin et al. 1992). A maternal steroid hormone, testosterone, for example, can be directly transferred by the broodstock to the egg and it will be detected until it hatched (Paitz et al., 2015). High testosterone content particularly affected aggressive behavior. Kim and Park (2012) mentioned that the estrogen increase causes feedback mechanism that will decrease testosterone content. Several former studies found that estrogenic compound and their waste (estradiol-17β, E2) and their analogue (17α-ethinylestradiol, EE2) located in an aquatic environment have negative effects and influence aquatic reproduction organism. The 17α-ethinylestradiol (Saaristo et al., 2010) and 17β-estradiol (Clotfelter & Rodriguez, 2006) degrade the aggression behavior in male broodstock. Estrogen has other functions, i.e. antidepression and antistress (Calmarza-Font et al., 2012). Andrade et al. (2005) stated that estrogen deficiency is often related to stress, anxiety, and depression. Afterward, Prasad et al. (2015) presented that serotonin is essential in inhibiting aggressive behavior by boosting brain serotonergic. It has been proved by Kumar et al. (2017) that serotonin can reduce cannibalism on snapper. A similar result is assumed to be applied on African catfish by using 17β-estradiol. In brief, this study aimed to analyze and determine the 17β-estradiol dosage and stocking density to reduce cannibalism behavior on African catfish.

Experimental fish
The tested fish was Sangkuriang catfish fingerling with average length 4.0 ± 0.1 cm from the Center of Freshwater Aquaculture, Sukabumi. It was stocked according to the stocking density treatment (150 individuals/m 2 and 300 individuals/m 2 ).

Experimental design
This study used the factorial randomized complete design. There were two experimental factors, i.e., 17β-estradiol dosage (0 mg/kg, 30 mg/kg, and 60 mg/kg) and stocking density (150 individuals/m 2 and 300 individuals/m 2 ) ( Table  1). Each treatment was repeated three times. The 17β-estradiol hormone was delivered orally using the commercial feed with 39-41% of the protein content for 30 days. Feeding was done three times a day as much as 10% of the total biomass.

Experimental feed preparation
Each hormone treatment was weighed according to the treatment. After that, the feed was sprayed using alcohol 96% (300 mL of alcohol for 1 kg feed). Furthermore, it was sprayed to the feed. The feed was dried at room temperature for 12 hours to release the alcohol smell. Thereafter, it was stored at -20°C freezer.

Fingerling sample preparation
This analysis was particularly aimed to measure the hormone and glucose content of the tested fingerling. Ten fingerlings were used for each treatment. The fingerling sample was fainted using a stabilizer with 1 mL/L dosage for 5 minutes. The sample was crushed and thoroughly mixed in PBS solution (phosphate buffer saline). The ratio was 1:4 and 1:2 for hormone and glucose measurement, respectively. Afterward, the samples were centrifuged at 5000 rpm and 4°C for 10 minutes. Supernatant then was separated into another microtube and stored at -20°C.

Experimental parameters
The cannibalism was calculated using the formula by Obirikorang et al. (2014). The mortal fingerling with wounds was categorized as cannibalism type I, while the swallowed fingerling was classified as cannibalism type II. The mortal fingerling without wounds grouped in the mortality caused by other factors and the fingerlings with 2-3 times bigger than the others were categorized as a potential cannibal. Masingmasing parameter tersebut dihitung berdasarkan rumus yang dikemukakan oleh Król dan Zakes (2016). Production performance was calculated using a formula by Effendie (1997). Hormone analysis at the end of the study was conducted using ELISA (enzyme-linked immunosorbent assay) method. Estradiol, cortisol, and blood glucose were analyzed using Kit Estradiol-17ß (DRG EIA 2693), Kit Cortisol (EIA 1887), and Kit GLUCOSE, respectively. Water quality analysis was managed at the beginning, in the middle, and at the end of the study. Water temperature was measured using a thermometer, pH using a pH meter. Dissolved oxygen and total ammonia nitrogen were measured using a DO meter and spectrophotometer, respectively.

Cannibalism
The cannibalism rate was measured using the formula according to Obirikorang et al. (2014).

Cannibalism category
Deceased and wounded fish were categorized into cannibalism type I. Meanwhile, completely lost or swallowed fish was categorized to cannibalism type II (Król & Zakes, 2016).

Mortality caused by other factors
Mortal fish without bite wounds was categorized to mortality by other factors (Król & Zakes, 2016).

Potential of cannibal
The bigger fish, two or three times bigger, from the average weight was categorized as a potential cannibal (Król & Zakes, 2016).

Mortality
Total mortality was calculated at the end of the study.

Absolute growth
Absolute growth was calculated based on final length and weight minus the initial length and weight.

Note: PM
: Absolute growth (cm) Pt : Final length (cm) Po : Initial length (cm) Specific growth rate Specific growth rate was calculated at the end of the study. Note: SGR : Specific growth rate (%/day) WT : Final average weight (g) W0 : Initial average weight (g) T : Rearing period (day)

Coefficiant of length variance
Both length and weight variance were calculated at the initial and the end of the study. version 2014. A post-hoc test was also conducted at 95% of confidence level. Particularly, water quality parameters were explained descriptively.

Mortality and cannibalism rate
The result showed that there were interactions amongst the hormonal treatment and stocking density towards the mortality, cannibalism type, and mortality by other factors (Table 2; P<0.05). Moreover, the lowest mortality was discovered in treatment B. It had 2.94 times of mortality and 3.84 times of cannibalism lower than the non 17β-estradiol hormonal treatment (Table 2; P<0.05). The application of 17β-estradiol also reduced the cannibalism, both type I and type II ( Table 2). The discovered cannibalism type I at the 17β-estradiol hormonal treatment (B, F, C, and E) was lower that without 17β-estradiol hormonal treatment (A and D). Additionally, the lowest mortality by other factors was also disclosed at treatment B (Table 2; P<0.05). The potential of cannibal declined along with the 17β-estradiol hormonal treatment (Table 2).
According to Figure 1A, mortality by other factors occurred from day 1 to day 14. However, the highest mortality of all treatments was found on day 7. Treatment E and F had the highest mortality, respectively. Moreover, mortality caused by cannibalism type I showed various results, the mortality started on day 1 to day 14 and the highest mortality was found on day 7

Feed conversion ratio
Feed conversion ratio is defined as a certain amount of feed required to produce 1 kg of biomass. It was calculated using the following formula: Note : FCR : Feed conversion ratio P : Feed amount (g) Wt : Final biomass (g) Wm : Mortal fish biomass (g) W0 : Initial biomass (g)

Data analysis
Data were tabulated using Microsoft Excel 2010. Cannibalism, production performance, hormone content, and blood glucose were analyzed using two-way ANOVA at P=0.05 of significant level assisted by XLSTAT software  ( Figure 1B). Meanwhile, treatment D showed consistent mortality until day 21.

Hormone content and blood glucose
The result indicated interactions amongst 17βestradiol hormonal treatment and stocking density towards estradiol, cortisol, and blood glucose of catfish fingerling (Table 3; P<0.05). Treatment F presented the highest estradiol content, but the lowest cortisol (Table 3; P<0.05). Meantime, blood glucose was reported declining along with the 17β-estradiol hormonal treatment (Table 3; P<0.05).

Production performance
Stocking density had no significant difference towards length specific growth rate (SGR length), size variation, and feed conversion. In contrast, it decreased SGR weight (Table 4; P<0.05). Furthermore, the 17β-estradiol hormonal treatment did not affect growth, size variance, and feed conversion (Table 4; P>0.05).

Water quality
Water quality results are shown in Table 5. The results indicated in the tolerable range to support the growth and survival of African catfish fingerling (Table 5).

Discussion
According to Naumowicz et al. (2017) cannibalism is aggressive behavior, caused by stress, variance, and environmental factors occurred in most live phases. Fundamentally, cannibalism is indicated by mortality caused by hurting, preying, and or killing. Different stocking density did not increase mortality and cannibalism (Table 2). However, the 17β-estradiol hormonal treatment could diminish mortality and it was in line with the cannibalism decrease, both type I and II ( Table 2). The decrease in mortality and cannibalism was presumably caused by the low level of aggressivity. Supported by Colman et al. (2009), who reported that the 17β-estradiol hormonal treatment   potentially reduced aggressive behavior on male broodstock of zebra danio Danio rerio, fathead minnows Pimephales promelas (Salierno et al., 2009), and Betta splendens (Clotfelter & Rodriguez, 2006). Conversely, testosterone and aromatase inhibitors conceivably boost aggressivity (Forsatkar et al., 2013). A similar result was also reported by Zairin et al. (2016) who discovered an aggressively boost caused by 17α-methyltestosterone. Rajakumar et al. (2012) reported in their study that testosterone and 11-ketotestosterone were detected in a 6 cm catfish fingerling. In this study, we used 4 cm in length catfish fingerlings so that we assumed that all of the experimental fingerlings already contented both two hormones mentioned before. Androgen hormone natural existence can be inhibited by increasing estrogen hormone through the feedback mechanism (Dinsdale & Ward, 2010). They also stated that 17β-estradiol could elevate the estradiol content in the catfish fingerling (Table 3). The feedback mechanism and estradiol would eliminate aggressivity, decrease mortality, and cannibalism in African catfish fingerling (Table 2).
According to Solomon and Udoji (2011), cannibalism type II is related to the mouth size along with the growth of catfish. It is also dependent on the size variation and the potential cannibal individuals. The potential cannibal individual undoubtedly will increase the cannibalism probability. In this study, the size variance (CV) at the end of the study did not differ significantly amongst treatments (Table  4). It was presumably caused by the mortality of the smaller individuals prey by the larger and more aggressive individuals. Hseu and Huang (2014) stated that sorting can be done to prevent cannibalism on orange-spotted grouper larvae Ephinephelus coioides. The supported previous statement, Ribeiro and Qin (2013) also declared that sorting was the simplest and most practical solution in preventing cannibalism on directly ontogeny species, e.g. barramundi Lates calcarifer has the potential to be a cannibal when their size was 50% bigger than the overall population. Szczepkowski et al. (2011) added that sorting contributed a beneficial effect in boosting survival through inhibiting cannibal behavior on pikeperch. Biu et al. (2015) confirmed the latter theory on African catfish Clarias gariepinus.
Furthermore, the high potential cannibal in this study was found in treatment A (0 mg/kg of hormone dosage and 150 individuals/m 2 of stocking density; P<0.05). Without 17β-estradiol hormonal treatment decreased cannibalism type II. In contrast, it increased cannibalism type I ( Table 2). The decrease of cannibalism type II was impacted by the stocking density so that each individual has no space and opportunity to prey its kind. They tended to keep moving, forming groups, and unquestionably lessened the aggressivity (Nieuwegiessen et al., 2008).
The lowest mortality was noticed in treatment B (30 mg/kg of hormone dosage and 150 individuals/m 2 of stocking density) (Table 2). However, there was no significant difference in cannibalism between the hormonal treatment and both high and low stocking density. As mentioned in Table 2, the highest mortality caused by other factors was observed in treatment E. Thereby, lower mortality at treatment B was presumably caused by other factors. Table 2 also presented that mortality caused by other factors elevated along with stocking density. Li et al. (2010) described that the increase of blood glucose was the secondary effect of stress, which is mediated by the corticosteroid and catecholamine release. Manuel et al. (2014) stated that the plasm cortisol content in catfish fingerling was below 10 ng/mL and increased when exposed by handling, sorting, and transport up to 50 ng/mL. The normal range of blood glucose content is 40-90 mg/dL and will reach two to three times higher than normal when exposed to stress (Patriche, 2009). Hastuti and Subandiyono (2015) stated a slightly different range of normal blood glucose, i.e. 70-100 mg/ dL. Table 3 presented that the experimental catfish fingerling experienced stress during the study, which was described by the cortisol and blood glucose content. The 17β-estradiol hormonal treatment can presumably inhibit stress proved by the decline of cortisol and blood glucose content along with the 17β-estradiol hormone dosage (Table 3). Estradiol is known as an antidepressant hormone, functions as an anxiety inhibitor of several animal species (Walf & Frye, 2009). The definition of estrogen is frequently related to stress, anxiety, and depression because of its ability to modulate neurotransmitter changes, included boosting serotonin and noradrenaline content. It also regulates the serotonin receptor and controls serotonergic neuron activity (Andrade et al., 2005). Gore (2008) added that estrogen was involved in various aspects of the brain neuroendocrine system which affects behavior. The serotonergic activity rise in the fish brain decreases the aggressivity and cannibalism on pikeperch larvae Sander lucioperca (Król & Zakes, 2016).
According to SNI 01-6484.2-2000, the rearing of catfish fingerling (3-5 cm) in a plastic/concrete tank which applies 25 fish/m 2 of stocking density, shows an excellent survival, precisely >80%. Statistically, the 300 fish/m 2 treatment presented higher mortality than the 150 fish/m 2 treatment (P<0.05). The water quality in this study was considered as similar to each other and proper for fish culture (Table 5). Thereby, the difference 1.14 ± 0.19 a Different superscript in the same column indicates significant difference (P<0.05). Note: A (0 mg/kg of hormone dosage and 150 individuals/m 2 of stocking density); C (60 mg/kg of hormone dosage and 150 individuals/m 2 of stocking density); D (0 mg/kg of hormone dosage and 300 individuals/m 2 of stocking density); E (30 mg/kg of hormone dosage and 300 individuals/m 2 of stocking density); F (60 mg/kg of hormone dosage and 300 individuals/ m 2 of stocking density). ALG (absolute length growth), AWG (absolute weight growth), SGR (specific growth rate), CV (coefficient of variance), FCR (feed conversion ratio). amongst mortality results (Table 2) was not caused by environmental factors. The high mortality caused by other factors in this study might be caused by the inability to quickly adapt in a new environment. It was supported by the high mortality on day 1 to day 7 in all treatments.
The hormonal treatment and stocking density did not affect the growth performance, included SGR length, but the increased stocking density diminished the SGR weight (Table 4). It was considerably caused by the low energy consumption in the wider area compared to the narrower area, where they tended to consume more energy (Slavĭk et al., 2014). In a high stocking density, each individual tends to move actively and less rest so it requires more energy (Boerreigter et al., 2015). It happened because, in high stocking density, they lack the space to move around and less chance on food as well. Nevertheless, the production performance on fingerling rearing is determined by the length and survival of the fingerling.

CONCLUSION
The 17β-estradiol hormone treatments could decrease the mortality and cannibalism on African catfish fingerling in high-density culture. The best treatment was reasonably the 30 mg/kg of 17β-estradiol and 150 individuals/m 2 .