Hen selection and tissue collection
All animal procedures were approved by the Institutional Animal Care and Use Committees at USDA Beltsville Agricultural Research Center (BARC) and at the University of Maryland, College Park (reference numbers 16-002 and XR-16-09, respectively). A total of 400 turkey hens over two equal sized flocks from the same commercial line were provided by a poultry primary breeding company (Hendrix Genetics, Kitchener, Ontario) and housed at BARC individually in wire cages during two separate time periods six months apart (200 hens per flock). Hens were managed with artificial lighting (14L:10D) and were provided feed and water ad libitum to NRC standards (diet composition in Supplemental File 4). Daily egg records were kept from the onset of lay (around 28 weeks of age) until sampling occurred (at 37 weeks of age). Daily egg records were used to identify the bottom and top 15% of egg production for classification into two groups, LEPH and HEPH, as previously described [6]. Daily egg records were also used to predict the timing of the PS prior to ovulation, as described previously [15]. Average egg production and distribution for LEPH and HEPH groups did not differ between the two flocks used.
All hens were euthanized by cervical dislocation prior to tissue isolation. Additionally, all hens were sampled on the second day of the hen’s sequence. From the LEPH and HEPH groups, a total of 12 hens were sampled from the first flock and 6 hens were sampled from the second flock, based on the progression of the hen’s sequence and the timing of the preovulatory surge. Experimental replicate numbers per group were determined through a power analysis (α = 0.05, power = 0.8, |μ1 − μ2| = 0.5, σ2 = 0.2), with a recommended sample size of three replicates per group. The first flock (12 hens sampled total) was used to perform transcriptome analysis of the hypothalamus and pituitary, with six LEPH and six HEPH, half sampled outside of the PS and half sampled during the PS (n=3 per group). These samples were also used to confirm gene expression results obtained from RNA sequencing through RT-qPCR (n=3 per group). For this experiment, the hypothalamus and pituitary were isolated from each hen, snap frozen as whole tissues in liquid nitrogen, and stored at -80°C prior to assessment through RNA sequencing and confirmation of RNA sequencing results as described below. The second flock (6 hens sampled total) was used to perform follow up in vitro pituitary cultures, with three LEPH and three HEPH sampled exclusively outside of the PS (n=3 per group). For the follow-up experiment, the pituitary was isolated from each hen and placed in ice cold Dulbecco's Modified Eagle Medium (DMEM) until dispersion, cell culture, and RT-qPCR as described below. For both experiments, blood samples were taken from the wing vein immediately before sampling to measure plasma progesterone levels to confirm correct sampling outside or during the PS. Blood samples were collected and fractionated as previously described [7]. Plasma samples were stored at -20°C prior to assessment through radioimmunoassays as described below.
Radioimmunoassay
Plasma progesterone levels were measured using a coated tube radioimmunoassay (RIA) kit (MP Biomedicals, Solon, OH) to confirm that sampling occurred at the correct time during the ovulatory cycle, based on experimental group assigned. All protocols were performed as directed by the supplier and samples were assayed in duplicate in a single RIA. Ether extraction of plasma samples prior to progesterone assessment and standard curve assessment were performed as previously described [6]. Hypothalamus and pituitary samples taken from a hen with plasma progesterone levels less than 1 ng/dL were considered to be sampled outside of the PS, while hypothalamus and pituitary samples taken from a hen with plasma progesterone levels greater than 4 ng/dL were considered to be sampled during of the PS. The plasma progesterone cutoffs for outside and during the PS were based on previous studies determining average plasma progesterone levels during the ovulatory cycle [15].
RNA isolation, cDNA library construction, and sequencing
Total RNA was isolated from whole tissue hypothalamus and pituitary samples with RNeasy Mini kits (Qiagen, Valencia, CA), including on-column deoxyribonuclease digestion. Quantification of RNA was performed as previously described [15]. Amplified cDNA was generated using a SMART-Seq v4 Ultra Low Input RNA kit (Takara Bio Inc., Kusatsu, Japan) following the manufacturer’s procedure, starting with 10 ng of total RNA per sample. cDNA was amplified by long distance PCR (LD PCR) (8 cycles as per the manufacturer recommendation for 10 ng of starting RNA). Amplified cDNA was purified using Agencourt AMPure XP kit (Beckman Coulter, Indianapolis, IN). Amplified cDNA was quantified using an Agilent 2100 Bioanalyzer and High Sensitivity DNA Kit (Agilent, Santa Clara, CA).
Sequencing libraries were generated using a Nextera XT DNA Library Prep kit (Illumina, San Diego, CA) with input of 150 pg of amplified cDNA per library, following the manufacturer’s procedure. For each sample, two libraries were produced (from the same amplified cDNA), with a unique index pairing for library. Libraries were purified using Agencourt AMPure XP kit (Beckman Coulter, Indianapolis, IN) and were quantified using an Agilent 2100 Bioanalyzer and High Sensitivity DNA Kit. For sequencing 24 libraries (2 tissues) were pooled (10nM). Libraries were pooled so that set 1 for each tissue was sequenced in a different pool than set 2. Pools were submitted to NC State’s GSL facility for paired-end sequencing (75 bp reads) on an Illumina NextSeq 500.
Bioinformatic analysis of sequencing data
All FASTQ sequencing files have been deposited to the NIH Short Read Archive (accession numbers SAMN11624488-SAMN11624511). Processing and analysis of sequencing data was performed using the Galaxy (https://usegalaxy.org/) suite. Adapter sequences and low-quality sequences (Phred < 20) were removed from FASTQ files using the TrimGalore tool. Trimmed reads were mapped to the Meleagris gallopavo reference genome (Turkey_2.01 using ENSEMBL annotation release 98; https://uswest.ensembl.org/Meleagris_gallopavo/Info/Index). TopHat was used to analyze mRNA libraries. DEGs were determined using the Cuffdiff tool. Pairwise comparisons were made between LEPH and HEPH for each timepoint in the ovulatory cycle as well as between timepoints in the ovulatory for each egg production group. Due to poor annotation of the turkey genome, the protein sequences for DEGs that were unannotated in the turkey were subjected to orthologous comparisons in human, mouse, and chicken protein sequences using Ensembl Biomart (https://useast.ensembl.org/info/data/biomart/index.html). Unannotated DEGs were assumed orthologous if greater than 50% identity to the human, mouse, and chicken was seen at the protein level.
Ingenuity pathway analysis
Ingenuity pathway analysis (IPA) (Qiagen, Valencia, CA) was performed on the differential expression data. IPA was used to construct gene networks as well as to predict upstream biological regulators for each pairwise comparison. Only DEGs with RPKM>0.2 were used for IPA. The RPKM threshold was selected based on the distribution of log2 transformed RPKM values across all of the comparisons examined. The threshold of DEGs was set at P <0.05 and absolute fold change ≥1.5. Pathways and predicted upstream regulators with P-value < 0.05 (Fischer’s exact test) were considered to be statistically significant. For upstream regulators, published findings in the Ingenuity knowledge database were used to calculate the activation z-score to infer activation or inhibition of transcriptional regulators. Upstream regulators with a z-score greater than 2 or less than -2 and P<0.05 were considered to be significantly activated or inhibited.
Culture of pituitary cells
All cell isolation procedures were performed using Minimum Essential Medium, Spinner modification (SMEM) or DMEM as noted below. Media was supplemented with 0.1% bovine serum albumen, 100-U/mL penicillin G, and 100-μg/mL streptomycin sulfate (0.1% BSA and P/S). Dispersion pituitary cells was performed as previously described [7]. Briefly, isolated pituitaries were dispersed in SMEM (0.1% BSA and P/S) using trypsin and collagenase (1 mg/mL of each) and filtered through 70 μm nylon mesh.
Cells were diluted, plated, and allowed to attach for 2 hours before pretreatment, as previously described [7]. Pituitary cells were pretreated with either no pretreatment (NPT) (10 ml DMEM/F12 added), T3 (1.5 ng/mL of thyroid hormone), or E2 (1.5 ng/mL of estradiol) for 12 hours, followed by treatment with chicken GnRH-I (Phoenix Pharmaceuticals, Burlingame, CA) at 0, 10-9, or 10-8 M for 6 hours (n=3 per pretreatment, treatment, and egg production level combination). Cells were maintained in a 37.5°C, 5% CO2 atmosphere during incubation. Pituitary cells were harvested at the end of incubation by retrypsinization, immediately frozen in liquid nitrogen, and stored at -80 °C until RNA extraction.
RT-qPCR
Total RNA was isolated from pituitary cell cultures with RNeasy Mini kits (Qiagen, Valencia, CA), including on-column deoxyribonuclease digestion. Quantification of RNA, RT, and RT-qPCR were performed as previously described [15]. A pool of total RNA was made and the reaction conducted without reverse transcriptase as a control for genomic DNA contamination. PCR reactions (15 μL) were carried out as previously described using a CFX Connect Real-Time PCR System (Bio-Rad, Hercules, CA) [15]. Pituitary data were normalized to phosphoglycerate kinase 1 (PGK1) and were analyzed by the 2-ΔΔ Ct method. All PCR reactions for each gene in a given tissue were analyzed in a single 96-well plate, allowing accurate performance of relative quantification without the need to include a reference control sample in multiple plates. Primers (Integrated DNA Technologies, Skokie, IL) for turkey PGK1, LHB, FSHB, and CGA mRNA were designed and used with cycling parameters described previously [15]. Data are presented as fold increase over levels in LEPH basal cells for each pretreatment/treatment combination and time point.
Confirmation of gene expression as determined by RNA sequencing
RNA extracted and quantified from whole tissue hypothalamus and pituitary samples for RNAseq was reverse transcribed as described previously [15], with controls for genomic DNA contamination. PCR reactions were carried out as described above. Hypothalamic data were normalized to GAPDH, whereas pituitary data were normalized to PGK1. Data were analyzed as described above. For each tissue, mRNA levels for 12 genes total were determined. DEGs selected for RNAseq confirmation fit the following parameters: P<0.05, absolute fold change greater or equal to 1.5, annotated in the turkey genome, and encoded by a single transcript. DEGs fitting these parameters were selected with the following RNAseq expression profiles: 3 DEGs up-regulated in LEPH both outside and during the PS, 3 DEGs up-regulated in HEPH both outside and during the PS, 3 DEGs which showed up-regulation in one egg production group outside of the PS and up-regulation in the other egg production group during the PS, and 3 control genes which did not show expression changes between egg production levels or during the ovulatory cycle. Primers were designed as described above. Data are presented as fold increase over mRNA levels for LEPH outside of the PS for each gene.
Statistics
All data were analyzed using SAS software (SAS Institute, Cary, NC). Normalized RT-qPCR data were log2 transformed before statistical analysis. A three-way ANOVA using the mixed models procedure (PROC MIXED) was conducted to compare the log2 transformed gene expression between LEPH and HEPH pituitary cell culture samples subjected to different pretreatment and treatment combinations. A two-way ANOVA using the mixed models procedure (PROC MIXED) was used to compare the log2 transformed gene expression between LEPH and HEPH from hypothalamus and pituitary samples used for RNAseq confirmation. The least squares means for each group were compared using the test of least significant difference (PDIFF statement) when this indicated an overall significance level of P < 0.05.