Evaluation of Reproductive Characteristics of 21 Highly Inbred Lines of White Leghorns Divergently Selected for or Segregating in Tumor Resistance

Reproduction performance of 21 inbred experimental lines of White Leghorns was evaluated based on samples of reproduction records over a period of eight consecutive years. Two lines (63 and 72) have been extensively used in studies, especially in research seeking for genetic and epigenetic factors underlying resistance to avian tumor virus-induced diseases in chickens. The other 19 lines are recombinant congenic strains (RCS), which were generated by crossing lines 63 and 72 followed by two consecutive backcrosses to the line 63 and then full-sib mating. In theory, each RCS processes 7/8 of progenitor background line 63 genome and a random sample (1/8) of the progenitor donor line 72 genome. All 21 inbred lines share a common major histocompatibility complex haplotype, B*2. The estimated average fertility of the 21 inbred lines ranged from 72.9% (RCS-J) up to 96.8% (RCS-P). Both progenitor lines 63 and 72 were observed with lower average fertility (82.4% and 81.6%, respectively) in comparison with the RCS except the RCS-J, suggesting a substantial polygenic component underlying the fertility phenotype. The average embryo mortality rate ranged from 14.5% (RCS-P) up to 47.0% (RCS-M). The background line 63 fell at about the middle of the range (28.3%) significantly higher than the donor line 72 (15.7%), which was among the group with the lowest embryo mortality. By definition, hatchability of fertile eggs is reversely correlated with embryo mortality. The average hatchability ranged from 26.5% (RCS-M) up to 66.8% (line 72) while the background line 63 remained (46.6%) at about the middle of the range. The variability of the average embryo mortality and hatchability observed among the 21 inbred lines indicated the two correlated traits also follow polygenic models of inheritance. Findings from this study paves the way for further investigation on genetic and environmental influence over reproductive performance of inbred lines of chickens, and particularly in understanding and improving the reproduction fitness of invaluable genetic resources like these inbred lines. Corresponding author. G. Kulkarni, H. Zhang


Introduction
Reproduction characteristics are of profound importance for livestock and poultry both in evolutionary and economic perspectives.Fertility and hatchability are major parameters of reproductive performance for egg-layer type of chickens.The heritability estimates for fertility, embryo mortality, and hatchability of fertile eggs in chickens reportedly range from 0.05 to 0.13, which indicate these important traits bear heavy influence from management measures, nutrient supplies, and surrounding environmental conditions [1]- [5].
Since the foundations for genetic improvement were laid by Robert Bakewell in the 18 th century [6], a variety of ways to employ inbreeding in combination with intensive selection to achieve breeding goals has been attempted.The adverse effects of inbreeding, commonly referred to as inbreeding depression, are also richly documented.In poultry, inbreeding depression not only adversely affect production performance, but also reproductive characteristics, including fertility, embryo mortality, and hatchability [7]- [13].
The adverse consequence resulted from inbreeding, particularly severe in closed small populations of domesticated poultry, leads to partial or complete loss of poultry lines or flocks from time to time [12].For instance, five out of a series of 24 recombinant congenic strains (RCS) of White Leghorns, initiated and maintained at the USDA, Agriculture Research Service (USDA-ARS) laboratory at East Lansing, Michigan, completely failed to reproduce by the sixth generation of full-sib mating [14].Inbreeding escalates homozygosity and speeds up fixation, by chance, of deleterious recessive alleles simultaneously at multiple loci.Due to natural or artificial selection, such recessive alleles generally remain with low or very low frequencies at heterozygous state in outbred populations, and their harmful effects are hardly detectable phenotypically.However, when inbreeding takes place, the chance for homozygosity of such deleterious alleles is drastically increased.Especially when inbreeding occurs consecutively over generations, homozygosity of such alleles at multiple loci generally show up at high frequencies, which would have significant negative impact on production, reproduction, and fitness characteristics of the populations [6] [8] [11] [12] [15] [16].
Despite of the hurdles and barriers along the way in developing inbred animals, inbred, recombinant inbred, and/or recombinant congenic inbred strains have been generated in many organisms including mice, cattle, and chicken [10] [14] [17].Inbred lines, especially characterized inbred lines of animals are invaluable resources for genetic studies [17]- [21].In particular, many unique genetic lines of chickens were developed and thirty five of which have been maintained at the USDA-ARS, Avian Disease and Oncology Laboratory (ADOL) at East Lansing, Michigan.Two-thirds of the genetic lines are highly inbred lines or RCS.Chickens from all of the lines have been serving as critical resources for successful completion of important projects in the past and remains being critically vital for new projects that are conducted in the genomic and post-genomic era at ADOL.The ADOL inbred lines as well as outbred lines of chickens also have been one of the important sources for research and diagnostic tests at other research centers and laboratories in the United States and around the world.The National Animal Germplasm Program (NAGP; http://www.ars.usda.gov/Research/docs.htm?docid=22314) approved and collected all chicken lines developed at ADOL, including the thirty five lines that are currently maintained at ADOL.This study was aimed to provide estimates on fertility, embryo mortality, and hatchability of fertile eggs for 21 out of the 35 lines, which include two highly inbred lines of chickens and a series of 19 RCS using sampled reproduction records during a period of eight years.

Lines of Chickens
A total of 21 inbred experimental lines of White Leghorn was included in this study to examine reproduction characteristics.Two of which, the line 6 3 and line 7 2 , are highly inbred and diversely selected for tumor resistance.Line 6 3 is resistant to Marek's disease (MD) and line 7 2 is highly susceptible to MD.Both lines 6 3 and 7 2 share the same major histocompatibility complex (MHC) haplotype, B*2.The other 19 lines are a series of RCS derived from the lines 6 3 and 7 2 by crossing the two lines followed with two consecutive backcrosses to the background line 6 3 .The details in development of all the 21 lines of chickens were reviewed by Bacon et al. [14].

Reproduction Data
Records collected from Year 2005 to Year 2012 were sampled to evaluate the reproduction characteristics of the inbred genetic lines of chickens.The line reproduction each year during the 8 years began with artificial insemination (AI) during January to March.AI and incubation handlings were all accomplished by a team of same four people during the period of time.A total of 84 hens and 12 males from each of lines 6 3 and 7 2 were used for line reproduction.The 12 males were selected from two families, six from each, within each of the lines 6 3 and 7 2 .Each male was mated to 7 random females.For the RCS, a total of 21 hens and 2 full-sib males for each line were used for reproduction.One male was mated to 10 females, and the other, to 11 females.The reproduction performance was measured by collecting the eggs for incubation from 1 to 2 weeks at around 35 weeks of age for each hatch for a total of three hatches per laying hen for each line.Fertility was calculated as the ratio of number of fertile eggs divided by the total number of eggs set for incubation multiplied by 100; the embryo mortality (EM) was defined as the number of dead embryos determined during candling around 8 -9 days of incubation divided by the total number of embryos multiplied by 100; hatchability refers to the percentage of fertile eggs that hatched, which was calculated as the ratio of the number of hatched chicks divided by the total number of embryos (dead or alive determined at candling) multiplied by 100.The numbers of hens and records included in this study for each line and year are listed in Table 1.

Statistical Analyses
The records of the dataset were filtered first to remove any record of hens that had fewer than 3 eggs at the time set for incubation.The percentage data for fertility, embryo mortality, and hatchability of fertile eggs were subjected to square root transformation to normalize the residuals prior to a reduced general linear model, Y ij = Year i + Line j + e, was fitted to examine the year and chicken line effects on the Y variables, where Y refers to the fertility, embryo mortality, hatchability of fertile eggs, and the e refers to the residual error term.The statistical significance of the year and chicken line effects was examined by F test under the linear model.The pairwise comparisons among the lines of chickens within each of, as well as across, the years for each of the three variables, fertility, embryo mortality, and hatchability of fertile eggs, were also tested under the linear model for statistical significance using the Duncan's multiple-range test.For direct and easy interpretation, the averages of fertility, embryo mortality, and hatchability of fertile eggs for each of the lines in each of the years were tabulated and plotted with the data prior to the square root transformation (Tables 2-4).All statistical analyses were performed using the JMP ® 11 SAS package [22].

Fertility
Fertility measures the capacity of poultry to initiate its reproduction process.The variability of fertility from year to year during the 8 years for all the inbred lines was statistically significant (P < 0.0001).The average fertility of all the lines ranged from 81.7% ± 0.6% in one year (2012) to 91.6% ± 0.5% in another (2010).Statistically significant difference was also detected for average fertility over the eight years between the inbred experimental lines (P < 0.0001) with an even larger range between 72.9% ± 1.3% for RCS-J and 96.8% ± 0.4% for RCS-P.Examining the average fertility by line over the eight years, the average fertilities of the progenitor lines 6 3 and 7 2 were ranked on the lower side with averages of 82.4% ± 0.6% and 81.6% ± 0.8%, respectively, and were not significantly different from each other (P > 0.05; Table 2).A hierarchical clustering tree was constructed to graphically illustrate the relative average fertility ranks among the 21 inbred lines over the eight year period (Figure 1).Table 2 also lists the average fertility estimates for by line and year for each of the 21 inbred experimental lines of chickens in Year 2005 to Year 2012.The average fertilities for RCS-P and RCS-B were between 91% to 98% and were consistently ranked on the higher side among the 21 inbred lines (P < 0.05), and were signifi-

Embryo Mortality
Embryo mortality measures the percentage of embryos dead-in-shell during early period of incubation, also known as early embryonic mortality, which directly leads to reduction in hatchability of fertile eggs in poultry.
Embryo mortality of all the inbred lines varied significantly during the years (P < 0.0001) with average embryo mortality of all the inbred lines ranged from 20.7% ± 0.6% in the Year 2012 to 35.2% ± 0.7% in Year 2011.The average embryo mortality of the eight years varied significantly among the inbred lines (P < 0.0001) with a range from 14.5% ± 0.7% for RCS-P up to 47.0% ± 1.2% for RCS-M.The average embryo mortality of the RCS-X 82.1 ± 3.1 cd 81.5 ± 2.9 defg 60.9 ± 5.6 f 81.5 ± 3.5 cdefg 81.4 ± 2.9 de 96.3 ± 1.1 ab 94.9 ± 1.3 ab 89.4 ± 2.3 abc 84.8 ± 1.1 ef progenitor background line 6 3 was 28.3% ± 0.6%, which was significantly different from the 15.7% ± 0.6% average embryo mortality of the progenitor donor line 7 2 (P < 0.01).Figure 2 graphically illustrates the relative rank of average embryo mortalities of the 21 inbred lines during the period of the eight years.
The estimates of the average embryo mortality for the 21 inbred experimental lines of chickens in each year are given in Table 3.Among the 21 inbred lines, RCS-M was ranked with the highest average embryo mortality in year 2005-2007 and year 2010-2012, followed by RCS-V with relatively high embryo mortality consistent throughout the 8 years (P < 0.05) in comparison to the other lines.In contrast, the RCS-K and RCS-P were consistently on the lowest or lower side of the rank in average embryo mortality from Year 2005 to Year 2012 (P < 0.05).The progenitor donor line 7 2 was observed with consistently lower embryo mortality percentages (12.1% -20.6%) compared to the background line 6 3 (19.2%-43.6%), and statistically differed in 4 (2006, 2008, 2009, and 2011) out the 8 years (P < 0.05, Table 3).

Hatchability of Fertile Eggs
Hatchability of fertile eggs measures the percentage of viable chicks hatched from fertile eggs within given numbers of eggs set in incubation.The average hatchability of fertile eggs of all the 21 inbred lines varied significantly during the eight years (P < 0.0001) with a range from 36.1% ± 0.8% in Year 2011 to 54.8.2% ± 0.8% in Year 2012.The average hatchability of fertile eggs during the eight year period also varied significantly among the inbred lines (P < 0.0001) with a range between 26.5% ± 1.1% for RCS-M and 66.8% ± 0.8% for the Figure 1.A hierarchical clustering tree depicts differences and similarities between the progenitor lines 6 3 and 7 2 and the recombinant congenic strains in fertility realized during a consecutive period of eight years (2005-2012).The progenitor lines 6 3 and 7 2 were similar to each other and were on the lower fertility side of the whole group (82.4% ± 0.60% and 81.6% ± 0.77%, respectively), similar to RCS-T, RCS-V, and RCS-W, and relatively distant to the rest of RCS in fertility.As shown, RCS-J had the lowest (72.9% ± 1.29%) and RCS-P had the highest (96.8% ± 0.41%) average fertility among the group during the years.A hierarchical clustering tree depicts differences and similarities among the 21 lines in embryo mortality during the eight years.The progenitor donor line 7 2 had much lower average embryo mortality (15.7% ± 0.63%) than the background progenitor line 6 3 (28.3%± 0.64%), similar with the RCS-P and RCS-K and relatively distant from the rest of RCS.RCS-P had the lowest and the RCS-M had the highest average embryo mortality (14.5% ± 0.71% and 47.0% ± 1.24%, respectively) during the years.RCS-X 24.1 ± 3.4 cdef 41.3 ± 5.1 bcd 33.6 ± 5.0 cde 14.2 ± 3.4 hi 32.9 ± 3.9 cde 24.1 ± 2.8 def 29.1 ± 2.9 ef 14.5 ± 2.3 fg 25.6 ± 1.3 fg donor line 7 2 .The average hatchability of the progenitor background line 6 3 was 46.6% ± 0.8%, which was significantly lower than the progenitor donor line 7 2 (P < 0.01, Table 4).A hierarchical clustering tree was also constructed to graphically depict the relative average hatchability of the 21 inbred lines for the eight year period (Figure 3).The estimates for hatchability of fertile eggs for each of the 21 inbred experimental lines of chickens from Year 2005 to Year 2012 are given in Table 4.The progenitor donor line 7 2 was ranked with the highest hatchability of fertile eggs in 6 (2005-2009, 2012) out of the 8 years, and was also among the highest hatchability group of lines in the other two years (2010, 2011).In contrast, the RCS-M was ranked as the line with the lowest hatchability in four years (2005, 2006, 2010, and 2012) and among the relatively low hatchability groups in the other four years (2007)(2008)(2009)2011).Some of the lines, including RCS-C, RCS-A and RCS-D, varied greatly in hatchability from year to year (Table 4).

Discussion
Inbred lines are invaluable resources for biological research, which provide experimental animals of high constancy within each of individual lines, and of distinctly genetic diversity between lines.The animals with such characteristics enable high detection power of biological experiments for difference in treatment and/or genetic effects on biological traits of interest over time and places with relatively high repeatability [10] [14] [18] [20] [23].RCS-K 53.9 ± 5.9 bcd 70.9 ± 3.6 a 45.4 ± 5.9 bcd 50.6 ± 5.0 bcde 41.5 ± 5.2 bcd 69.3 ± 2.9 a 56.7 ± 3.0 ab 71.4 ± 3.2 a 57.2 ± 1.7 b RCS-L 51.9 ± 4.4 cde 39.8 ± 4.0 cd 37.8 ± 4.6 cde 34.2 ± 4.3 efg 36.9 ± 3.7 cde 40.6 ± 3.4 def 34.4 ± 3.5 efgh 51.1 ± 3.6 cd 41.0 ± 1.4 efg RCS-M 27.4 ± 3.4 h 21.5 ± 3.5 f 18.9 ± 3.3 fg 32.9 ± 2.9 fg 22.4 ± 3.1 hi 25.8 ± 2.7 g 28.0 ± 3.4 gh 28.7 ± 2.9 f 26.5 ± 1.  ADOL at East Lansing, Michigan, currently maintains 35 special experimental lines of White Leghorns.Twothirds of these lines are highly inbred.These lines are tested annually for purity by blood-typing tests using over 40 antisera.The breeders of all the lines are unique in that they are maintained in a quarantined state and, on the basis of frequent serologic tests for 11 pathogens, are considered free of infection from common poultry pathogens.Many of the lines each year produce either embryos or day-old chickens that are shipped out to meet the research or diagnosis needs of institutes or laboratories in the United States and around the world upon requests, in addition to fulfill the needs for research conducted on the site.Line 6 3 and line 7 2 were initiated for establishment as inbred lines in 1939 at ADOL with hatching eggs from similar White Leghorn strains.Line 6 3 was one of the lines that were initiated to define genetic resistance to avian virus-induced tumors collectively termed as avian leucosis complex at the time.It was later shown by studies that the avian leucosis complex was consisted of tumors caused by two types of avian viruses commonly co-existed at the time in chicken houses, they are avian leucosis retroviruses (ALV) inducing principally lymphoid leukosis and Marek's disease herpesviruses (MDV) inducing MD in susceptible chickens [14] [20] [24] [25].Line 6 3 chickens were selected for susceptibility to both ALV and MDV infection but uniquely resistance to tumor progression.Line 7 2 chickens were selected to define genetic resistance to subgroup A and B exogenous and subgroup E endogenous ALV infection but for susceptibility to MD tumors [14] [20].
A total of 19 RCS was set up for establishment as described by Bacon et al. [14] and Chang et al. [26].Briefly, the series of RCS were developed by crossing the MD resistant line 6 3 and susceptible line 7 2 to produce the F 1 Figure 3.A hierarchical clustering tree depicts differences and similarities among the 21 lines in hatchability of the fertile eggs during the eight years.As illustrated, the progenitor line 7 2 achieved the highest average hatchability (66.8% ± 0.82%) during the years among all the lines and the line 6 3 failed close to the middle of the range (46.6% ± 0.76%).The lowest average hatchability was observed in the RCS-M (26.5% ± 1.14%) among all the lines over the years.
followed by two consecutive backcrosses to the background line 6 3 .The series of RCS were established from full-sib mating of the backcross-2 progeny [14].In theory, each of the RCS on average maintains 7/8 of the line 6 3 genome background and received a random sample of 1/8 of the donor line 7 2 genome.All lines, 6 3 , 7 2 , and RCS, are MHCB*2 haplotype.The series of 19 RCS are the only extensive set of RCS under development in chickens or any other livestock species.The 19 RCS are under development to identify non-MHC genes that influence traits differing between the progenitor lines 6 3 and 7 2 .The progenitor lines 6 3 and 7 2 differ in many ways, including resistance to avian virus-induced tumors [14] [26]- [28], protective efficacy in response to vaccination followed by Marek's disease virus challenge [26] [29] [30], behavior and behavior associated characteristics [31] [32], primary and secondary lymphoid organ sizes [33], and genetic and epigenetic differences [34]- [40].This study showed the two lines 6 3 and 7 2 significantly differ in embryo mortality and hatchability of fertile eggs (P > 0.01).When a RCS differs in phenotype of a trait from the rest of RCS, a limited number of responsible non-MHC genes may be identified for further analysis.Each of the RCS is now over 90% inbred, and inbreeding increases each generation.
Fertility in chickens is reported with a very low heritability of about 0.05 -0.10 [41] [42], suggesting the fertility of chicken eggs is heavily influenced by factors other than genetics.The fertility of layer-type of chickens in commercial populations is as high as 97% [43].The progenitor lines 6 3 and 7 2 were observed with relatively lower average fertilities, 82.4% ± 0.6% and 81.5% ± 1.1%, respectively, over the eight years in comparison with either the commercial populations or some of the RCS.This can be in part attributable to the high degree of inbreeding of the lines.The fertility within each of the inbred lines did fluctuate from year to year in the eight years and the both of the year and the line effects on fertility were statistically significant (P < 0.0001).Yet three of the RCS (RCS-B, RCS-M, and RCS-P) maintained fertility consistently over 90% each year in the eight years and had 93.6% ± 0.5% 93.0% ± 0.7%, and 96.8% ± 0.4% average fertility over the year period, respectively (Table 2).Furthermore, although the progenitor lines 6 3 and 7 2 did not differ in fertility during the eight years evaluated, ten of the 19 RCS were observed with average fertility significantly higher than the progenitor lines (P < 0.05) during the eight year period, which strongly suggested the existence of a substantial polygenic component underlying the fertility phenotype despite the known fact that fertility is a trait of low heritability.
Embryo mortality is regarded as a direct fitness trait that could significantly reduce the efficiency of reproduction and increase the costs of production performance in poultry [44].Like fertility, embryo mortality is also characterized with relatively low heritability, reportedly ranging from 0.06 to 0.12 [5].The two progenitor lines 6 3 and 7 2 significantly differed (P < 0.01) in embryo mortality with an eight year average of 28.3% ± 0.6% and 15.7% ± 0.6% embryo mortality, respectively.The donor line 7 2 was among the group of the lowest average embryo mortality and was significantly lower in average embryo mortality than 17 out of 19 RCS while the background line 6 3 was only significantly lower than 3 RCS (P < 0.05) over the eight year period (Table 3, Figure 2).The sizable range (14.5% ± 0.7% -47.0%± 1.2%) of the average embryo mortality of the inbred lines during the eight year period indicated that the embryo mortality ratios of the inbred lines of chickens followed a polygenic model of inheritance.
Hatchability of fertile eggs is a major component of reproductive fitness.Reported estimates of heritability for hatchability of fertile eggs fall into the low to moderate category [5] [41].Industry performance target for hatchability is set for greater than 83% [45].The average hatchability of fertile eggs in the eight year period for the 21 inbred experimental lines ranged from 26.5% ± 1.1% (RCS-M) to 66.8% ± 0.8% (line 7 2 ), significantly lower than the industrial target, which was highly likely, in a large part, resulted from high inbreeding, since inbreeding is known to have a seriously negative impact on reproductive characteristics including hatchability in chickens [8].The progenitor donor line 7 2 was one of the lines with the observed lowest average embryo mortality but was the line with the observed highest average hatchability of fertile eggs over the eight year period and significantly different from the background line 6 3 (Table 3 & Table 4, P < 0.01).The large range of hatchability of fertile eggs among the RCS indicated this trait also follow a model of polygenic inheritance.
Earlier studies showed genetic background due to MHC haplotype differences does exert significant effects on production and reproduction characteristics in White Leghorns [18] [46].The fact that the year to year effect on fertility, embryonic mortality, and hatchability of fertile eggs was very statistically significant (P < 0.0001) suggesting that environmental effects and random variations highly likely predominate.However, since all of the 21 inbred lines of chickens in this study were housed and managed under standardized control conditions on a specific pathogen free farm and all of the lines share the same MHC B*2 haplotype, it should be within reason to assume that the differences in these traits within each year attributable to environmental influence were moderate to minimal.Thus, the statistically significant differences in fertility, embryo mortality, and hatchability of fertile eggs among the inbred lines strongly suggested that polygenic effects play a substantial role influencing the phenotypes of the reproduction characteristics of the inbred lines and the polygenic effects were imposed by non-MHC genes.Furthermore, both of the progenitor lines 6 3 and 7 2 averaged with relatively low fertility in contrast to over half of the RCS, which suggested multiple loci of genes positively affecting fertility are reciprocally fixed in the two lines, respectively.This speculated model for fertility, to some extent, would also be applicable for interpreting the progenitor lines' ranking positions with the RCS for embryonic mortality.On the contrary, the progenitor line 7 2 was ranked on the very top for hatchability of fertile eggs, and the line 6 3 , at the middle, which indicated both of the progenitor lines 6 3 and 7 2 are likely oppositely fixed at multiple loci of genes negatively impacting on hatchability of fertile eggs, in addition to fixations at loci that positively affect hatchability, especially in the line 7 2 .Findings from this study paves the way for further investigation on genetic and environmental influence over reproductive performance of inbred lines of chickens, and particularly in understanding and improving the reproduction fitness of invaluable genetic resources like these inbred lines.

Figure 2 .
Figure 2.A hierarchical clustering tree depicts differences and similarities among the 21 lines in embryo mortality during the eight years.The progenitor donor line 7 2 had much lower average embryo mortality (15.7% ± 0.63%) than the background progenitor line 6 3 (28.3%± 0.64%), similar with the RCS-P and RCS-K and relatively distant from the rest of RCS.RCS-P had the lowest and the RCS-M had the highest average embryo mortality (14.5% ± 0.71% and 47.0% ± 1.24%, respectively) during the years.

Table 1 .
Numbers of hens and set eggs sampled from ADOL farm records for the retrospective evaluation.higherthan that of the progenitor lines 6 3 and 7 2 in Year 2005 to Year 2008 (P < 0.01) but comparable from Year 2006 to Year 2012 (P > 0.05).In contrast, the fertility of RCS-J was estimated with the lowest average fertility in 3 out of the 8 years (2005, 2010, and 2012), on the lower side in 3 of the 8 years(2007, 2009, and  2011), and at about the middle of the range in the other two years(2006 and 2008).The average fertilities of the progenitor lines 6 3 and 7 2 varied from year to year and were not significantly different from each other (P > 0.05) within any of the 8 years.
Note: RCS stands for recombinant congenic strains.RCS-A… RCS-X refer to a series of 19 RCS derived from the inbred lines 63 and 72.cantly

Table 3 .
Embryo mortality of 21 inbred experimental lines of White Leghorns.