An assessment of responses to egg production and liver health of Japanese quails subjected to different levels of metabolizable energy

Objective Current quail production is configured as an economic activity in scale. Advancements in quail nutrition have been limited to areas such as breeding and, automation of facilities and ambience. The objective of this study was to evaluate the performance responses, liver and oviduct morphometry, and liver histology of Japanese laying quails subjected to different levels of nitrogen-corrected apparent metabolizable energy (MEn). Methods A completely random design was used that consisted of nine levels of MEn, six replicates, and five hens per cage with a total of 270 quails. The experimental period lasted for 10 weeks. The variables of performance were subjected to analysis of variance and then regression analysis using the broken-line model. The morphometric and histological variables were subjected to multivariate exploratory techniques. Results The MEn levels influenced the responses to zootechnical performance. The broken-line model estimated the maximum responses for feed intake, egg production, egg weight, and egg mass as 3,040, 2,820, 1,802, and 2,960 kcal of MEn per kg of diet, respectively. Multivariate analysis revealed that the occurrence of hepatic steatosis and increased levels of Kupffer cells were not related to MEn levels. Conclusion The level of 2,960 kcal/kg of MEn meets performance variable requirements without compromising hepatic physiology.


INTRODUCTION
Advances in genetics have allowed for the segregation of the multiplication and produc tion sectors, and this is indicative of the degree of professionalization that the quail egg industry has achieved in recent decades. The facilities were adapted for Japanese quails, and currently, sheds possessing temperature control automation, dung removal, and egg collection are present at the main commercial eggproducing locations [13].
Attention should be focused on the concentration of nitrogencorrected apparent me tabolizable energy (MEn) in the diet. Therefore, the concentration of MEn in the diet should be established based on certain criteria. However, publications during the last two decades do not show a consensus regarding the amount of MEn required. Here, it is understood that a reduction in the difference between the values recommended in the literature is necessary. We located 11 studies examining MEn for quail hens, and the recommenda tions ranged from 2,600 to 3,100 kcal/kg and egg production (EP) ranged from 77% to 94% using Japanese and European quails [414]. When ana lyzing these publications, a common theme was that the mean concentration of MEn in the diet was 2,876 kcal/kg with an amplitude ranging from 87% to 110%. The lowest MEn tested was 2,500 kcal/kg [11]. The association of narrow amplitude with the ad libitum feed intake (FI) option may have been an attenuating factor in the imposed deficiency and, consequently, in the absence of an effect of the MEn levels tested in most of the studies reviewed here.
The possibility of obtaining a greater amplitude of the MEn values in the experimental diets may involve the use of the dilution technique; however, this is a common practice in amino acid studies [15]. For MEn, only one study using commercial laying hens utilized this technique [15], and with quails, the use of this technique is nonexistent. Another missing description in the literature is the impact of the concentration of MEn on hepatic physiology and oviduct morphometry of current Japanese quail lines improved for commercial posture. For other nutrients such as protein, the possible effects have already been described [16]. Thus, in this study we investigated the performance responses, liver and oviduct morphometry, and hepatic histology of laying Japanese quails subjected to different levels of MEn.

MATERIALS AND METHODS
The Animal Ethics and Welfare Committee of Universidade Estadual Paulista approved all experimental procedures used in this study (protocol number 6.725/15).

Birds and husbandry housing
Two hundred and seventy female Japanese quail (VICAMI strain at 16 weeks of age) were housed in galvanized wire cages measuring 1.0 m×0.5 m×0.15 m. The cages were equipped with feeders and nipple drinkers in a climatic chamber at room temperature. The birds were selected and uniformly distributed in the experimental units based on their weight (185±7 g) and EP (78%±6%). The light program that was used was a 16 L:8 D. The minimum, average, and maximum temperatures were 18°C, 22°C, and 26°C, respec tively. The minimum, average, and maximum humidity values were 50%, 60%, and 68%, respectively. The experiment lasted ten weeks. The first four weeks involved adaptation, and the last six weeks were for data collection. All birds were fed 27 g of the diet per day and were supplied twice each day (morning and afternoon).

Experimental diets
The treatments were distributed in a completely randomized design that consisted of nine increasing levels of MEn (1,609, 1,740, 1,895, 2,260, 2,394, 2,643, 2,892, 3,045, and 3,058 kcal/kg, respectively) with six replicates of five quails per experimental unit. Two diets were formulated with high and low MEn contents (Table 1) according to previously published recom mendations [17] ( Table 2).
The high MEn diet was formulated to contain 114% greater than the MEn recommendation (2,850 kcal/kg), and the low MEn diet was formulated to contain approximately 52% less than the MEn recommendation. The remaining samples were obtained by dilution (Table 3). Dietary protein, other nutrients, amino acids, vitamins, and minerals were main tained at constant levels so as not to be limiting.

Measurements of responses
Egg production (EP, %) was recorded daily, and egg weight (EW, g) was measured on three consecutive days each week. Egg mass (EM, g/bird·d) was determined using EP and EW. Feed leftovers were weighed weekly to quantify the weekly FI (g/bird·d). The feed conversion ratio (FCR, g/g) was cal culated by dividing FI by EM and was corrected for mortality. Body weight (BW, g/bird) was measured at the first and tenth week of the assay. The crude fiber (CF) content of the feed or its indigestible portion was used as a measure of physical ca pacity intake or scaled feed intake (SFI, g/kg metabolic BW 0.67 ).
At the end of the tenth week, excreta were collected using the total collection method to determine the MEn of the ex perimental diets. Ferric oxide (10 g/kg) was added to each treatment diet with proper mixing as a marker to define the beginning and end of the excreta collection period. The pe riod of adaptation to diet was 72 h. The excreta were collected in adapted trays under the cages twice each day for three days, packed in plastic bags daily, and stored in a freezer (-20°C) until the end of the collection period when the samples were then homogenized per experimental unit, predried, and ground in a ball mill. The diets and excreta samples were sent to the Laboratory of Poultry Sciences for analysis of dry matter (AOAC Official Method 930.15), total nitrogen (AOAC Official Method 2001.11), and crude energy (Model Parr 6200 oxygen bomb calorimeter), and MEn values were cal culated. The CF of the diet was analyzed on a dosifiber machine using the AOAC Official Method 920.39.

Tissue sampling and euthanasia
At the end of the experiment, 54 birds were selected (one bird per replicate) to analyze the oviduct morphometry and liver histology. The birds were individually identified and euthanized after an eighthour fasting period. Carbon diox ide was used in a transparent chamber to allow individual birds to be observed. The flow rate was maintained at approxi mately 20% of the chamber volume per min. The gas flow was maintained for at least 3 min after apparent clinical death. The chamber was sanitized after each euthanasia procedure.
The livers and oviducts were collected, measured using an electronic digital device, and weighed on a digital scale (0.001 g).
The right and left lobes of the livers were separately weighed. The number of folds in each magnum and isthmus was quantified.

Physiological responses
For analysis of the histological parameters of the liver, sam ples from the median region of the left lobe (approximately 1 cm 2 ) were collected and fixed immediately in Bouin's solu tion for 24 h. They were then washed in alcohol (70%) to remove the fixative, dehydrated in a series of alcohols, di aphanized in xylol, and embedded in paraffin. Four semi serial histological sections (7 μm thickness) were obtained from each bird and placed on the histological slides, and they were subsequently stained using the hematoxylineosin (HE) technique. Five fields were photographed randomly from each cut with the aid of a digital camera (Leica DFC 295) at tached to a microscope (LeicaDM 2500), and the images were analyzed using LAS V.3.8 (Leica, Wetzlar, Germany).
Three photomicrographs per slide based on a randomiza tion table were randomly chosen to determine variables. The 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 ME, metabolizable energy. 1) Vitamin premix provided the following (per kg of diet): Vit. A 1,750.000 U. I; Vit. D 3 500,000 U. I; Vit. E 2,000 U. I; Vit. K 3 500 mg; Vit. B 1 250 mg; Vit. B 2 875 mg; Vit. B 6 500 mg; Vit. B 12 1,250 mcg/kg; niacin 6,250 mg; choline 65 g; pentatonic acid 2,500 mg; copper 2,000 mg/kg; ferro 12,500 g; manganese 17,500 g; zinc 12,500 g; iodine 300 mg; selenium 50 mg. 2) Butylhydroxytoluene. areas of Kupffer cells and steatosis were measured [18]. To count Kupffer cells and steatosis, a grid with five frames (50×50 μm each) composed of dotted and continuous lines was super imposed onto each of the photos to randomize the counting area. Cell counting was performed on all frames (including cells on the dotted lines) while discarding those located on the continuous lines [18].

Statistical analysis
Univariate statistics: The data from the current experiment were analyzed using the PROC NLIN statement of SAS 9.4 (Statistical Analysis for Windows, SAS Institute Inc., Cary, NC, USA). The linear plateau models (one slope and two slopes) were adjusted according to previously described pro cedures [19]. The responses (ξ) were considered as dependent variables, and the mean levels were analyzed (X) as independent vari ables. A brokenline model was used with a slope (Eq. 1), quadratic (Eqs. 2), and two slopes or two lines (Eq. 3).
where (τ-X) is defined as zero when X>τ for the singleslope Eq. 1 and quadratic Eq. 2. The twoslope brokenline model expressed in Eq. 3, (τ-X) is defined as zero at X>τ, and (X-τ) is defined as zero when X<τ. We used the parameters for the breakpoint X value (τ), an asymptote for the first segment (ψ), and slopes for the 2 line segments (α, β). Parameters were estimated using the procedure described previously [19].
Multivariate statistics: The morphometric and histological variables were subjected to factorial analysis that is a multi variate exploratory technique. Thirteen variables were used for this analysis, including BW, liver weight, liver length, right liver lobe weight, left liver lobe weight, oviduct weight, oviduct length, number of isthmus folds, number of ovary folds of magnum, number of Kupffer cells, number of ste atoses, area of Kupffer cells, and area of steatosis. First, the adequacy of the sample space was analyzed through the cor relation matrix, KaiserMeyerOlkin test, Bartlett sphericity test, antiimage matrix, and commonalities [20]. The results led to the exclusion of the weight of the bird, number of folds of the isthmus, and number of Kupffer cells from the factorial analysis. Then, we used principal components as a method to extract the factors while considering that the value 1.00 is the minimum for an eigenvalue to be significant and for the load of a variable to be considered significant. The minimum value was 0.70 within one factor.
Factor analysis was used to create latent variables associ ated with the originally measured variables to thus allow analysis of variance considering several variables simultane ously to verify the statistical significance of the relationship between the latent variables classified within each factor and its relationship with the levels of MEn. Considering the means of extracting latent variables, we here may refer to multivariate analysis of variance.

RESULTS
In general, the various MEn levels modified the bird responses (p<0.010) for all variables of zootechnical performance (Table 4). Considering FI, none of the MEn levels were the result of birds consuming the full 27 g/d of feed offered. The level of MEn at 1,609 kcal/kg exhibited the highest FI of 25.7 g/d compared to the level of 3,250 kcal/kg for 23.3 g/d, and this represented a difference of 2.4 g. Comparing the results of FI and metabolizable energy intake (MEI) it was observed that at levels of 1,609, 1,740, and 1,895 kcal/kg, the birds consumed more feed and consumed less energy. For diets possessing a greater amount of fiber and as MEn levels increased, the FI decreased and the MEI increased.

Physical capacity intake
According to Figure 1, the brokenline model with two slopes allowed for the description of the SFI that was closer to the mean contour of the treatment responses. The level possess ing a higher CF content exhibited a reduction in the SFI. This reduction influenced the response of asymptote shaping when the brokenline model was adjusted with a slope according to the equation: For X<τ and when X>τ, [τ-X] = 0. The reduction in FI as By definition, [186-X] = 0 when X>τ and [X-186] = 0 when X≤τ. The rate of reduction in the SF after 186 g/kg of CF in diet is 4.5 (0.1614/0.7276)fold greater than is the rate that describes increasing SFI.

Zootechnical performance
According to the model fitted for FI, the asymptote (ψ) in the FI occurred at 3,040 kcal/kg. According to our model, FI is expected to increase by 0.38% for every 100 kcal reduction in the diet ( Table 5).
The lowest EP was 2.3% for the 1,609 kcal/kg level, and the highest EP was 77.8% for the 3,058 kcal/kg level. The as ymptote (ψ) for EP was 74%, and this occurred from 2,829 kcal/kg according to the model presented in Table 5. Accord ing to this model, a reduction of 100 kcal/kg in the diet resulted in a 7.9% reduction in the EP.
There was a significant difference for the EW (Table 4),  3 Figure 1. The mean relationship between the scaled feed intake (SFI) of Japanese quails at the laying stage (g feed/kg metabolic body weight/d) and crude fiber content in the feed (g/kg).
We need the file of Figure 1.
Please prepare new pictures and send them (through this submission system). The preferred file type for figures is TIFF. If figures are to appear in color in the print journal, the files must be submitted in CMYK color (not RGB).
The minimum resolution is 300 dpi for color and grayscale figures, and 600 dpi for line art. Photomicrographs must have their unmagnified size designated either in the caption or with a scale bar on the figure. where the lowest EW was 10.2 grams (2,260 kcal/kg) and the highest EW was 11.7 g (3,058 kcal/kg). There was a linear increase in EW from 1,802 kcal/kg ( Table 4). The twoslope model fitted to this variable is presented in Table 5.
As the birds increased their MEI, there was also an increase in EM. The smallest EM was 0.2 grams (1,609 kcal/kg), and the largest was 9.1 g (3,058 kcal/kg). The response asymptote (ψ) to EM stabilized at 8.44 g/bird·d, and this corresponded to the 2,965 kcal/kg level. According to the model (Table 5), a reduction of 100 kcal/kg in the diet causes the EM to be come reduced by 7.0%.
Overall, there was an improvement in FCR as the levels of MEn in the diet increased ( Table 4). The highest FCR (98.1± 38.6) was observed for the first level of MEn, and the lowest FCR was observed for the highest level of 3,058 kcal/kg for MEn. The model with a quadratic ascending fit to the vari able FCR is presented in Table 5. According to this model, stabilization of the response (ψ) at 3.12 g/g occurred at a concentration of 2,611 kcal/kg.
The BW of the birds at the start of the trial varied from 184 to 188 g (a difference of 3.24 g), and these data are not provided in the tables. At the end of the assay, the weight of the birds varied from 153 to 194 (a 41 g difference) as pre sented in Table 4. The first four levels of MEn resulted in weight loss in the birds, and the largest losses were observed for the first two levels of MEn (average of 31.76 g). Weight loss de creased the extent to which the MEI was increased. Table 4 indicates a linear increase in the ΔBW of birds that occurred with an increase in the level of MEn in the diet until it reached the point where there was no further response (the response asymptote [ψ]). According to the model (Table 5), the point where the break (τ) occurred was 2,892 kcal/kg, and at this level the ΔBW close to zero (0.007 g).

Morphometry and histology
The MEn levels did not significantly affect the morphometric variables of the liver and oviduct (Table 6) or the morpho logical variables of the liver (Table 7). In general, there was a numerical difference between the level of 1,609 kcal/kg and the low average values compared to the 3,059 kcal/kg level. The datasets presented in Tables 6 and 7 were subjected to an analysis of factors to evaluate the interrelationship be tween the variables (Table 8). Table 8 presents the four factors obtained in the factorial  analysis that was performed for the extracted parameters and the variables comprising each factor. There was no sig nificant effect (p>0.050) for energy level on the parameters that were evaluated by variance analysis ( Table 8). The analy sis of variables that correlate with each of the factors revealed that factor 1, factor 2, and factor 3 exhibited a positive corre lation among a group of variables. Factor 1 indicated that when the weight of the liver was increased, the right lobe weight, left lobe weight, and the number of steatoses increased. Factor 2 indicated that when the weight was increased, the oviduct became increased in length. Factor 3 indicated that when the number of pleats in the magnum was increased, the length of the liver became increased, and factor 4 revealed a negative correlation between Kupffer cells and steatosis cells. This indicated that when the area of the Kupffer cells was increased, the area of the steatosis cells was decreased.

DISCUSSION
To obtain range of responses using the doseresponse method, the evaluated nutrient must be limited. To achieve this, we applied the concept of "theoretical" dilution technique [15]. According to the results obtained for EP and EM, MEn was the most limiting nutritional resource for the validation of the study. This can be verified by the theory of FI regulation that was proposed previously [21] and that states that if the diet is not properly balanced, the bird will increase its FI in an attempt to compensate for the most limiting nutritional resource. We have observed this in our current study as evi denced by the ratio of 0.38% for each reduction of 100 kcal in the diet. This rate is expected to continue to increase. Al though it has not been measured, it is speculated that the physicochemical characteristics of the ingredients used in  1) 0.148 0.500 0.148 0.937 F1-4, factor. * Significance variables within each factor. 1) The factor coefficients in bold were used for interpretation. Probability F multivariate analysis of variance for nitrogen-corrected apparent metabolizable energy (MEn) levels.