Use of B-mode ultrasonography for sex determination and maturation monitoring in Prochilodus brevis (Steindachner 1875)

ABSTRACT The Brazilian bocachico is a rheophilic fish from the northeast region of Brazil that holds great economic and scientific interest. The absence of evident sexual dimorphism and the invasive character of sexing methods and evaluation of gonadal maturation are obstacles to its artificial reproduction. In this scenario, ultrasonography emerges as a non-invasive method for sexing. The objectives of this study were to establish an ultrasound sexing method for Brazilian bocachico and select equations to estimate the gonadosomatic index (GSI), gonad weight (GW), and gonad volume (GV) of live bocachico. A total of 32 bocachico fish (18 females and 14 males) were analyzed ultrasonographically and morphologically. The association of morphometric and ultrasonographic parameters was investigated using Pearson Correlations. The following prediction equations were selected using multiple regression analysis: GSI = 16.22 × transvarea; GW = 30.3 × transvarea; and GV = 0.001260036 × pixels. A model capable of predicting the sex of the animals based on the ultrasonographic parameters was obtained using partial least-squares discriminant analysis with principal component analysis. The significance level adopted was P < 0.05. In conclusion, ultrasonography is a useful tool to perform sexing and evaluate the gonads during the breeding season, contributing to artificial reproduction in captivity.


INTRODUCTION
The Brazilian bochachico is a reophilic fi sh of the order Characiformes and the family Prochilodontidae.It is endemic to the semiarid region of northeast Brazil and highly appreciated by the local population, mainly due to the consumption of its spawn, which holds notorious scientifi c and economic interest (BOMFIM et al., 2015).This species is in decline due to predatory fi shing and the construction of dams to perpetuate intermittent rivers, where it naturally occurs, as these constructions prevent their reproductive migration, known as 'piracema' (NASCIMENTO et al., 2012).Currently, some wild fi sh populations are challenged by the negative impacts of generalized environmental variations caused by humans, which can lead to phenotypic changes with signifi cant ecological and evolutionary consequences that ultimately put the survival of these animals at risk (JENKINS et al., 2021).
Due to this threat, research is being carried out to enhance the biotechnologies used in the reproduction of this species (ALMEIDA-MONTEIRO et al., 2020;TORRES et al., 2022).According to Bomfim et al. (2015), males and females of Brazilian bochachico reach all stages of maturity (immature, maturing, mature, and in regression) even in captivity, with mature individuals possibly being found after 247 days after hatching (DAH).
However, some obstacles are encountered in the captive breeding of this species, e.g. the sexing of breeding stock and the evaluation of the gonadal maturation stage of the available specimens (CASTAÑEDA-CORTÉS; FERNANDINO, 2021;NASCIMENTO et al., 2012), which makes it diffi cult to separate breeding stock in lots of the same sex.Moreover, because it is a reophilic species, it is not able to spontaneously spawn in these conditions due to the absence of environmental stimuli, requiring the application of pituitary extract.
Existing methods for sexing and for the evaluation of gonadal maturation, such as the calculation of the gonadosomatic Index (GSI), ovarian biopsy, and hormonal measurements, among others, are considered invasive, requiring the sacrifice of part of the animals in the school or harming their welfare and health (COLOMBIER et al., 2015).Thus, adequate sexing and effective monitoring of gonadal maturation in males and, mainly, females, will enable the development of aquaculture and species preservation (CASTAÑEDA-CORTÉS; FERNANDINO, 2021;ZOHAR, 2021).
In this context, ultrasound appears as an alternative non-invasive method for assessing fi sh gonads.This method has the advantage of using portable equipment, having a shorter execution time, and the optional use of anesthesia in some species (NOVELO; TIERSCH, 2012).In addition to sex identifi cation, ultrasonography can also be used to distinguish immature from mature specimens.The evaluation of gonadal structure is important for estimating gamete quality and gonadal morphology.
In view of this and the existing variations in hormonal induction protocols for reproduction regarding the dose of hormone to be used for each sex (PIRES et al., 2018), as well as the proportion between males and females to be allocated in each tank (SHIMODA et al., 2015), an eff ective and non-invasive method must be adopted for sexing and for evaluating gonadal maturation.Despite the use of ultrasonography to identify both sex and maturation stage in other fi sh species, the evaluation of gonadal morphology in Prochilodus brevis by ultrasound has yet to be reported in the literature.
Therefore, the objectives of this study were: (1) to establish an ultrasound sexing method for Brazilian bocachico fi sh; and (2) to select equations to estimate GSI, gonad weight (GW), and gonad volume (GV) in non-anesthetized live Brazilian bocachico.

MATERIAL AND METHODS
This project was approved by the Ethics Committee on Animal Experimentation at the State University of Ceará (approval no.6699198/2017).Its execution was divided into two stages, considering the breeding season of the species.The fi rst was in February 2018 and the second in August of the same year, during the breeding and non-breeding seasons, respectively.
A total of 32 Brazilian bocachico fi sh were selected at random from a school of 55 animals collected from a reservoir in the municipality of Canindé-CE.These were identifi ed individually through the intramuscular implantation of a microchip and kept for at least two months in a permanent-recirculation tank of approximately 7,000 L at the State University of Ceará, located in Fortaleza.
The fi sh were fed a commercial maintenance feed.The temperature remained at approximately 28 °C degrees, ranging from 26.5 to 29.5.The pH remained stable, at approximately 7.0.Ammonia and nitrite were systematically monitored and water was exchanged so as not to compromise the welfare of the animals, always keeping the levels of these substances below 0.25 ppm.
In the two experimental stages, a fi shing net was used to remove the Brazilian bocachico from the tanks.Then, the animals were handled carefully so as to cause the least stress and subjected to ultrasound examination.
Use of B-mode ultrasonography for sex determination and maturation monitoring in Prochilodus brevis (Steindachner 1875) In February, during the breeding season, 11 females and fi ve males were analyzed ultrasonographically.The fi sh were manually contained in the supine position, completely submerged in water, without sedation.They were evaluated along the abdominal surface, in the transverse plane, from the cranial to the caudal region.Two longitudinal and one cross-sectional image were captured, totaling three images per animal.
In August, in the non-breeding season, seven females and nine males were analyzed ultrasonographically. Due to the difficulty in obtaining the images, it was necessary to perform the ultrasound analysis outside the water.Therefore, the animals were sedated with 1 mL Eugenol (Shanghai Medical Instrument Co., Ltd) for each 10 mL of absolute alcohol for each 10 L of water.During this analysis, an appropriate gel based on carboxymethylcellulose was used and the examination sequence was similar to the first step.The longitudinal and cross-sectional images were captured digitally and stored in JPEG format.The total ultrasound length of the gonad was measured by viewing its extremities, using the pelvic fins as a reference (Figures 1A, B, and C), with the first image captured with the caudal region of the transducer positioned on the ventral fin and the second image with the cranial region of the transducer on the same fin, thus allowing full coverage of the gonad.
For both sexes, cross-sectional images were taken at half the length of the gonad, following the methodology used by Colombier et al. (2015).After ultrasound examinations, the fi sh were euthanized with an overdose of Eugenol (Shanghai Medical Instrument Co., Ltd) to confi rm sex and morphometry and thus validate the use of ultrasound to calculate GSI, GW, and GV.
In the fi rst stage, ultrasound analyses were performed using portable equipment (Logic E, General Eletric); and in the second stage, fi xed equipment (Mylab 40, Esaote).Both were coupled to linear 12-MHz transducers.The signal amplifi cation (gain) was adjusted to obtain quality images and the B-mode two-dimensional ultrasound was performed in the resolution mode.
The images were evaluated using ImageJ ® software, and the following gonadal ultrasound parameters were measured: length (uslength), width (uswidth), height (usheight), cross-sectional area (transvarea), and number of pixels in the cross-sectional area (pixels).
Pearson's correlation was used to investigate the association between morphometric and ultrasound parameters and multiple regression analysis was performed to obtain prediction equations for IGS, GW, and GV, depending on the ultrasound parameters evaluated.The equations were selected by the values of the coeffi cient of determination (R 2 ) and the corrected Akaike criterion (AICc), weighting the predictive capacity and the parsimony of the models.
Discriminant analysis was performed by the partial least squares (PLS-DA) and principal component analysis (PCA) methods to develop a model capable of predicting the sex of animals using ultrasound parameters.Statistical analyses were carried out using R software (R CORE TEAM, 2018).The hypothesis tests were considered significant for P < 0.05.

RESULTS AND DISCUSSION
Based on the ultrasound analyses, the following images were obtained as representative of the patterns found in the breeding and non-breeding seasons, in longitudinal section (Figures 2A, B, E, and F) and in cross-section (Figures 2 C, D, G, and H).The application of ultrasound in the study of animal reproduction makes it possible to acquire knowledge of reproductive biology in different species (KAUFFOLD et al., 2019;MANTZIARAS;LUVONI, 2020;MEDAN;EL-ATY, 2010;NAEVE et al., 2019).
The best images were obtained using the maximum frequency of 12 MHz, corroborating the study by Colombier et al. (2015), which used frequencies between 11 and 15 MHz.In other studies (CREPALDI  , 2006;GUITREAU et al., 2012;KUCHARCZYK et al, 2016;NAEVE et al., 2018), diff erent frequencies were used, probably due to the diff erent body dimensions of the species.Brazilian bocachico are known to be smaller than the other fi sh species studied, which showed better results at lower frequencies.Based on the images, it was observed that, in the breeding season, the ovaries and testes showed greater measurements when compared with the non-breeding season, with greater changes observed in the size of the ovaries.These results are in accordance with the reproductive biology of the bocachico species, corroborating Bomfi m et al.In the breeding season, the ovaries were subjectively more echogenic than the testes, and it is possible to observe the level of association between the ultrasound parameters and the morphometric parameters analyzed (Table 1).The ultrasound images obtained are compatible with the ultrasound fi ndings described in other species, such as in salmon (Salmo salar), by Naeve et al. (2018), who observed that ultrasound parameters are highly correlated with GSI and ovarian weight (R = 0.99, P < 0.001), and can be a non-invasive alternative for estimating GSI, detecting females in advanced maturation stage even before detection through hormonal dosages, eliminating sacrifi ce, reducing stress, and improving the welfare of breeding stock.
The diffi culty in sexing varies between fi sh species and between maturation stages; however, it is usually based on the echotexture, grayscale, and visibility of ovaries and testes in ultrasound images.Monitoring fi sh maturation generally depends on the estimation of gonad diameter or volume, but there are still few studies involving the ultrasound estimate of gonad weight for calculating GSI.The total gonad length and/or cross-sectional area at diff erent locations is estimated and then used for prediction, either using linear models or associating these measurements with an early representation of the gonad shape (COLOMBIER et al., 2015).
In this study, it was only possible to subjectively sex fish in the breeding season.Thus, a predictive model was created based on PLS regression, which classifies males and females based on two principal statistical components (Figure 3).To obtain a prediction equation for GSI, GW, and GV (Figures 4 A, B, and C) based on longitudinal and cross-sectional gonad ultrasound during the breeding season, all morphometric and ultrasound parameters were analyzed statistically and then a model with maximum RSQ and minimum AICc, based on a single variable: GSI = 16.22 × transvarea, GW = 30.3× transvarea, GV =11 0.001260036 × pixels.The histological slides of ovaries and testes, obtained by conventional processing and hematoxylin-eosin staining (Figures 5 A, B, C, and D), revealed the stage of maturation at which the animals were on each occasion.
The ovaries showed a great histological diff erence in the obtained images (Figures 6 A, B, C, and D).In the breeding season, type-V (larger) and type-II (smaller) oocytes were observed, demonstrating that the females were in the mature stage.In the non-breeding season, type-II, -III, and -IV oocytes were found, indicating that the animals were in the maturing stage.
In the testes, a large amount of sperm was present in the seminiferous tubules, with higher values found in the breeding season.These results are in agreement with the study by Bomfi m et al. (2015), who performed a histological evaluation of the gonads of this species.
Reproductive ultrasound proved to be an effective, noninvasive tool for determining the sex of Neosho madtoms and performing repeated fish examinations without causing injury or appreciable  , 2005), and its use has already been reported in some catfish species (GUITREAU et al., 2012).However, there are still no reports for the species P. brevis or even for the order Characiformes.
Ultrasonography showed to be an eff ective technique, as it allows for sexing and accurate gonad evaluation and is a non-invasive procedure, since it causes little stress and can be performed with the animal submerged in water.Additionally, it has a low cost and the equipment is easy to acquire, given the availability of ultrasound devices used in cattle farming, a very common activity throughout Brazil, and whose use in fi sh breeding started over 30 years ago (MARTIN et al., 1983).
Therefore, further studies are clearly warranted, since ultrasound can improve the selection of breeding stock and the monitoring of maturation, making it possible to maintain animals with superior growth characteristics alive in the herd for scientifi c, conservational, or commercial purposes.In addition, it may also be used in future works to determine the exact time of hormonal induction in the species (CREPALDI et al., 2006).

F i g u r e 1 -
Transducer placement during the ultrasound examination of Prochilodus brevis in the supine position using the ventral fi ns as an anatomical reference.(A) longitudinal cranial, (B) longitudinal caudal, and (C) transverse et al.

Figure 2 -
Figure 2 -Ultrasonographic aspect of longitudinal and cross-sectional scans of Prochilodus brevis ovary in the breeding season (A and E) and in the non-breeding season (B and F).Ultrasonographic aspect of longitudinal and transverse scans of Prochilodus brevis testes in the breeding season (C and G) and in the non-breeding season (D and H).Images obtained using linear 12-MHz transducers

Figure 3 -
Figure 3 -Partial Least Squares (PLS) analysis used to develop a statistical model for sexing Prochilodus brevis

Figure 5 -
Figure 5 -Histological sections of Prochilodus brevis ovary in the breeding season (A) and in the non-breeding season (C).Histological sections of testes of P. brevis in the breeding season (B) and in the non-breeding season (D).Hematoxylin-Eosin.40x magnifi cation

Figure 6 -
Figure 6 -Histological sections of Prochilodus brevis ovary at diff erent stages, showing type-II oocytes (A), type-III oocytes (B), type-IV oocytes (C), and yolk granules in type-V oocytes (D).Hematoxylin-Eosin.400x magnifi cation.The mature stage contains types II and V.The maturing stage contains types II, III, and IV ( ) ( )

Table 1 -
Correlations between the analyzed parameters slength: standard length; width: animal width; depth: animal depth, abdcirc: abdominal circumference.Above the diagonal (R 2 = 1) are the values of R 2 and below it, the values of P