Oocyte distribution within and between ovary lobes is largely homogeneous in Lachnolaimus maximus ( Perciformes : Labridae )

In fish reproduction, previous information of ovary oocyte distribution is necessary, when oocytes quantitative estimates are required to estimate batch or annual fecundity. Heterogeneous oocyte distribution requires a standardized sampling protocol to prevent bias in estimates, whereas homogeneous distribution, allows sampling of any portion of gonads with no risk of bias. We studied gonad oocyte distribution patterns in the hogfish Lachnolaimus maximus population from Southern Gulf of Mexico. For this, 23 mature females were selected from a total of 47 individuals exhibiting visible oocytes in the ovaries. These females were classified by reproductive phase and sub-phase (early developing-ED, developing-D, spawning capable-SC and actively spawning-AS). Six histological sections were taken from the anterior, middle and posterior regions of the left and right ovary lobes of each individual. Digital image processing (AxioVision and Image ProPlus programs) was used to estimate oocyte density per unit area, and for different development stages. Contingency tables were used to analyze oocyte distribution frequencies between the regions of each lobe, and between the lobes of each ovary. This was supported with a Pearson’s χ2 test for goodness-of-fit and a replicated G test to confirm distribution heterogeneity. Oocyte stage distribution was homogeneous in almost all 23 females regardless of the ovary lobe. In the left ovary lobe, oocyte distribution was uniform in all three regions sampled regardless the female phase or sub-phase. In the right ovary lobe, oocyte frequencies were similarly uniform for the ED, D and AS phase and sub-phases; nevertheless, during SC phase, some heterogeneity was observed in tertiary vitellogenesis-Vtg3 oocytes, especially in samples from middle and posterior regions of this lobe. Females in AS sub-phase are normally used to estimate batch fecundity in fish such as L. maximus, which has shown to have asynchronous oocyte development and batch spawning. Given the homogeneous oocyte distribution pattern within and between the ovary lobes in females in AS sub-phase, no systematization is required of the gonad histological sampling protocol to estimate species batch fecundity. Nevertheless, due to the heterogeneous Vtg3 oocytes distribution in the right ovary lobe of females in the SC phase, it is best to systematically take sections of any region in the left ovary lobe when conducting a study encompassing all of a species’ reproductive aspects. Rev. Biol. Trop. 65 (1): 293-303. Epub 2017 March 01.

Reproduction has direct implications in the exploitation, management and cultivation of fishes (Wootton & Smith, 2015).Fish reproduction studies involving microscopic observation of gonads require preliminary analyses of oocyte distribution homogeneity in females at different ovary development stages.During maturation, fish ovaries undergo voluminous development.For practical reasons, histological observations of large ovaries are normally done using randomly sampled areas within a section of one of the ovary's two lobes.Quantitative analysis of the structure of any biological tissue using histological sections requires basic data on cell type distribution in the studied organ.If distribution of the different oocytes stages is homogeneous within the gonad during a fish species' reproductive season, samples can be collected from any region of either lobes, and will be representative of the female´s degree of maturity (e.g.Ma, Kjesbu, & Jirgensen, 1998;Witthames et al., 2009;McElroy, Wuenschel, Press, Towle, & McBride, 2013).On the other hand, if distribution among the different regions of a lobe or between the two lobes is heterogeneous, a systematic protocol is required for collecting samples for histological analysis (Witthames & Greer Walker, 1995;Nichol & Acuna, 2001;Kennedy, Witthames, & Nash, 2007;Hainfellner, de Souza, Nascimento, Freitas, & Batlouni, 2011).Generally, a lack of a standardized ovary sample collection method, may create significant bias in data for fecundity estimations (Murua et al., 2003).
Gonadal development studies have focused in standard histological methods, and very few studies have used digital image analysis, to calculate average oocyte diameter and estimate oocyte density (Nichol & Acuna, 2001;McElroy et al., 2013); although this type of analysis has been in use in the biological sciences for over a decade (Klibansky & Juanes, 2008).The application of digital image processing techniques may generate data on structure from images normally intended to improve image quality and in the search of data (Pertusa, 2003).They can also be used to count and measure cells or other small particles in an image, and to make precise measurements; besides, the use of these techniques has reduced the processing time, health risks and costs of standard histological methods.
Hogfish Lachnolaimus maximus (Walbaum, 1792) is a labrid fish species distributed from the mid-Atlantic seaboard of the United States (North Carolina) to the North coast of South America (Venezuela and Guyana), including the Bermudas, the Gulf of Mexico and the Caribbean Sea (Carpenter, 2002).It is important in commercial and recreational fisheries throughout its range, and presently, the IUCN (2016) has considered the species to be at high risk of extinction in the wild (Vulnerable A2bd).Growth and recruitment stage overfishing has been reported in a L. maximus population in South Florida (US) (McBride & Murphy, 2003;McBride & Richardson, 2007;McBride, Thurman, & Bullock, 2008).The L. maximus population in Southern Gulf of Mexico, on the continental shelf of the Yucatan Peninsula (known as Campeche Bank), is considered a high value alternative target fishing resource.To date, no studies have been done on the biology, exploitation level and conservation status of L. maximus population in Campeche Bank.
The reproductive biology of this species' has been studied in populations of Puerto Rico (Colin, 1982), Cuba (Claro, García-Cagide, & Fernández de Alaiza, 1989), North Carolina (Parker, 2000), the Northeast Gulf of Mexico, and South Florida (Davis, 1976;McBride, Johnson, Bullock, & Stengard, 2001;Robinson & Prince, 2003;McBride & Johnson, 2007;McBride et al., 2008;Muñoz, Burton, Brennan, & Parker, 2010;Collins & McBride, 2015).Of these studies, six have included histological analysis of gonads to identify species sexuality (protogynous hermaphrodite) and reproductive cycle, and to estimate fecundity.Only two (Davis, 1976;McBride & Johnson, 2007) have addressed spatial development of oocytes in the ovaries.Neither of these reported studies have evidence of heterogeneous oocyte development between lobes or within sections of the same lobe.But Davis (1976)  The objective of the present study was to characterize oocyte distribution patterns in different developmental stages and different sections of the ovary of L. maximus, by the use digital images taken from histological sections of female gonads, in different development stages.The resulting oocyte distribution patterns will determine the sampling protocol for histological sampling from gonads of this species, thus providing more accuracy to analyses of main reproductive parameters (i.e.maturity, reproductive cycle and fecundity) of L. maximus in Campeche Bank.

Female selection and ovary histology:
From the total sample of females, a preselection was made of those exhibiting visible oocytes in the ovaries (N = 47;, taking care to identify mature adult individuals with secondary growth oocytes.Six samples were extracted from each preselected ovary for histological analysis: one from the anterior, middle and posterior sections of the left and right ovarian lobes.These were fixed in Bouin's solution for four to five days, embedded in paraplast, sectioned at a 6 µm thickness (4 to five sections per sampled region), and stained with Gabe and Martoja's one-step trichrome stain (Gabe, 1968).Based on the histological analyses, the preselected females were classified by their gonad development phase into one of the following reproductive phases or sub-phases (Brown-Peterson, Wyansky, Saborido-Rey, Macewicz, & Lowerre-Barbieri, 2011): early developing (ED), developing (D), spawning capable (SC) and actively spawning (AS).
The histological sections of only 23 females (20.3-34.0cm FL; 191-814 g WW) were of sufficient quality (i.e.sections with no broken or incomplete oocyte and adequate color contrast) to allow digital image processing.
Image analysis: Digital images were taken from three microscopic fields (each field = 3.8 mm 2 ) of each ovary region section using a camera (Axiocam MRc) attached to a microscope (Axioscop; 25X).Each image was analyzed with the Image-Pro Plus 6.0 program (Media Cybernetics, Inc.).
Before beginning the oocyte count, all oocytes (that did not appear complete in the right and upper portions of each field) were excluded to prevent recounting of the same oocyte.Each image was processed with spectral analysis and a binary mask generated within which the selected oocytes were counted.Oocyte stages were identified based on the criteria developed by Wallace and Selman (1981), Brown-Peterson, Thomas and Arnold (1988) and Lowerre-Barbieri et al. ( 2009): primary growth (PG) oocyte; cortical alveolar (CA) oocyte; primary, secondary and tertiary vitellogenic (Vtg1, 2 and 3, respectively) oocytes; and oocyte maturation (OM).The oocyte maturation stage included: germinal vesicle migration, germinal vesicle breakdown, yolk coalescence, and hydration events.
Contingency tables (rows x columns) were applied to each group of females which had been classified into the four reproductive phases or sub-phases to analyze the frequency distributions of oocyte stage (columns) at two levels (rows): 1) anterior, middle and posterior regions of each ovary lobe, and 2) right and left lobes of the ovary.A Pearson's Chisquared (c 2 ) was used to determine if oocyte frequencies were uniform between rows and columns.When distributions were not uniform, a replicated goodness-of-fit test (heterogeneity G-test) (Sokal & Rohlf, 1995) was applied to identify which rows differed in terms of oocyte frequency between columns.All statistical analyses (a = 0.05) were run with the InfoStat program (Di Rienzo et al., 2013).

RESULTS
Following the degree of ovary development, the 23 selected females analyzed were classified in four reproductive phases or subphases: eight were in the ED sub-phase; four in the D phase; five in the SC phase; and six in the AS sub-phase (Fig. 1).The latter did not exhibit newly collapsed post-ovulatory follicles (POFs) in the ovary.The different oocyte stages characteristic of each female reproductive phases and sub-phases were observed in all the examined ovary sections and lobes (Table 1).In addition, the oocyte density typical of each phase and sub-phase was generally very similar between ovary regions and lobes; the one exception being Vtg3 oocytes in the right ovary of SC phase females.
The oocyte frequency analysis showed that the oocyte distribution was uniform in all three regions of the left lobe (c 2 , 0.1423 ≤ P ≤ 0.3858), no matter the phase or sub-phase (Table 2).In the right lobe, oocyte frequencies were also uniform in females at the ED subphase (c 2 , P = 0.2683), the D phase (c 2 , P = 0.4006) and the AS sub-phase (c 2 , P = 0.6852).However, females in the SC phase exhibited significant differences in oocyte frequencies in this lobe (c 2 , P = 0.0146).This discrepancy in frequencies in the SC phase females was confirmed by the G test (P = 0.0086; Table 3).Significant differences were present between any combination of the three regions in the right lobe of these females, the highest being between the middle and posterior lobes.This difference was probably caused by variability in Vtg3 oocyte frequency between the three regions of the right lobe (Table 1).The SC phase females analyzed here were likely just entering that phase.Nonetheless, if this oocyte category is excluded from the statistical analysis, distribution of all the other oocyte categories (PG, CA, Vtg1 and 2 oocytes) was homogeneous throughout the three regions  n = number of females analyzed; a = reproductive sub-phases; b = three 3.8 mm 2 microscopic fields x n; A = anterior, M = middle; P = posterior; PG = primary growth oocyte; CA = cortical alveolar oocyte; Vtg1 = primary vitellogenic oocyte; Vtg2 = secondary vitellogenic oocyte; Vtg3 = tertiary vitellogenic oocyte; OM = oocyte maturation.a = reproductive sub-phases; n = number of females analyzed; c 2 = chi-square goodness-of-fit statistic; d.f.= degrees of freedom (r-1)(c-1), with r=anterior, middle and posterior regions of each right and left lobes of ovaries, and c=oocyte stages (see Table1), for a chi-square calculated from contingency table data; p = probability value; b = three 3.8 mm 2 microscopic fields x three ovary regions x n. * = significant statistical difference (P < 0.05).

DISCUSSION
Most of females analyzed in the present study exhibited a high degree of homogeneity in oocyte distribution within and between the ovary lobes.The only observed difference in distribution was between different regions of the right lobe in SC phase females.These females had low Vtg3 oocyte densities in both lobes, as well as notable variation in the frequencies of this oocyte stage among the regions of the right lobe.This variation was apparently one of the possible factors causing the heterogeneity observed in the SC phase females.During the spawning season of L. maximus from the Eastern Gulf of Mexico, Collins and McBride (2015) observed a diel pattern of oocyte maturation in females: germinal vesicle migration of the vitellogenic oocytes occurs in the late morning-early afternoon preceding spawning day.As a consequence, few vitellogenic oocytes without any evidence of maturation were present in the ovaries at mid-day.Notwithstanding, SC phase females analyzed in the present study presented neither oocytes undergoing early nor late germinal vesicle migration.These females were not undergoing OM and probably were barely entering the SC phase and therefore displaying very few oocytes progressing towards the final vitellogenesis stage.
In this latter case, then females classified as being in the SC phase, may not necessarily be representative of this stage, and data from them must be used with caution.Larger sample sizes of SC phase females collected throughout this species' reproductive season would help to eliminate any uncertainty about Vtg3 oocyte distribution during this phase.However, even if Vtg3 oocyte distribution is heterogeneous in the right lobe during this phase, it did not affect the protocol for estimating fecundity.Female L. maximus have asynchronous oocyte development and engage in successive batch spawning (Claro et al., 1989;McBride & Johnson, 2007).This means that only females in the AS subphase, not exhibiting newly collapsed POFs in the ovary, should be selected to estimate batch fecundity (Murua et al., 2003).In the present data, AS sub-phase females exhibited a homogeneous distribution of all oocyte stages, including OM, which are those considered when estimating batch fecundity.
Two previous studies have been done on oocyte development homogeneity in L. maximus, both with a Florida population (Davis, 1976;McBride & Johnson, 2007).a = reproductive sub-phases; n =number of females analyzed; c 2 = chi-square goodness-of-fit statistic; d.f.= degrees of freedom (r-1)(c-1), with r=left or right lobes of the ovaries, and c=oocytes stages (see Table 1), p = probability value; b = three 3.8 mm 2 microscopic fields x three ovary regions x two lobes x n.
Given the homogeneous oocyte distribution pattern within and between the ovary lobes in L. maximus females in the AS sub-phase, no systematization is required of the gonad histological sampling protocol to estimate species batch fecundity.Stereological methods in tandem with digital image processing can therefore be used to make accurate measurements of oocyte size, shape, count, area and volume in the ovaries of L. maximus, and generate unbiased fecundity estimates in this species.However, ovary sampling for characterizing the reproductive cycle in this species, based on analysis of the seasonal development of reproductive phase and sub-phase frequency, needs to be done with caution due to the potential for heterogeneity among Vtg3 oocytes in the right ovary lobe of females in the SC phase.If this is a possible risk in a given sample, the present results suggest it is best to systematically take sections of any region in the left ovary when conducting a study encompassing all of a species' reproductive aspects.

ACKNOWLEDGMENTS
Collections were authorized by fishing licenses: PPF/DGOPA-037/14 and PPF/ DGOPA-080/15, from Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación/Comisión Nacional de Acuacultura y Pesca.The authors thank the professional fishers Leonardo Pech, Santos Efraín Sosa Pech, José Nicolás Flores Aceves, Luis Emilio Aceves Nadal and Cesar Alexander Tun Pacheco for their assistance with sample collection.We also thank the fishing cooperatives Nohoch Cuch S.C.L., S.C.P.P., and Pescadores de Río Lagartos S.C.D.R.L, S.C.P.P.The Consejo Nacional de Ciencia y Tecnología (CONACYT) financed a doctoral research stay in La Paz, Baja California Sur through the 2013 PNPC RM program.Technical assistance with gonad histological processing was provided by Teresa Colás-Marrufo, and the digital images were taken by Ximena Renán.We thank R. S. McBride, one anonymous reviewer and the editor for providing insightful comments that improved the quality of the manuscript.

TABLE 1
Oocyte density (number per unit area) frequencies observed in anterior, middle and posterior regions of left and right ovary lobes of Lachnolaimus maximus females in different reproductive phases and sub-phases

TABLE 2
Chi-square goodness-of-fit test results for oocyte density frequencies in anterior, middle and posterior regions of left and right ovary lobes of Lachnolaimus maximus females in four reproductive phases and sub-phases

TABLE 3
Replicated goodness-of-fit test (heterogeneity G-test) results for oocyte density frequencies in anterior, middle and posterior regions of the right ovary lobe of five Lachnolaimus maximus females in spawning capable (SC) reproductive phase.Total area analyzed = 57.0mm 2 Chi-square goodness-of-fit test results for oocyte density frequencies in left and right ovary lobes of Lachnolaimus maximus females in four reproductive phases and sub-phases