Pituitary Gland

Figure 1 visually shows the improvement of the robust T1w image; however, we know that this comes at the expense of an introduced intensity bias across the images. Therefore two experienced radiologists (PH/PM) familiar with the MP2RAGE contrast were asked to qualitatively assess, from a clinical perspective, the image quality regionally over the brain. The anonymized and randomized images were rated for: i) general image quality, in line with the quality rating procedure employed in ADNI [14] (link), ii) 7T specific artifacts including residual inversion, susceptibility and intensity inhomogeneity, and; iii) the definition of local structures including the hippocampus, thalamus, striatum, pituitary gland, and the temporal and cerebellar lobes. The scoring system was based on the ADNI rating system: 0 = severe, 1 = moderate, 2 = mild and 3 = none [15] (link). A positive sided Wilcoxon signed rank test was performed on the average score difference (robust – uniform) across categories for each observer assuming that the observations over subjects and repetitions can be considered independent.
To demonstrate the utility of the robust T1w image for morphometry packages, the image volumes were segmented with SPM8 and MorphoBox. The segmentation quality between the uniform and robust images was visually assessed. The reproducibility of MorphoBox's volumetric estimates of the brain's structures between repeat scans was used to compare the segmentation quality of the uniform verse robust T1w image. The average relative volumetric difference and the worst-case errors are reported. The worst-case error was defined for each subject as the maximum across structures of relative volumetric differences between repeats in absolute value. A positive Wilcoxon signed rank test was performed on the difference of worst-case reproducibility error between the uniform and robust segmentations. The same was done between the pre-processed and robust segmentations. In addition the proposed pre-processing step for MP2RAGE images proposed in [16] was also applied, Figure 3, using the software package [17] . Here the background noise is removed by creating a mask based on the GRE_TI2 magnitude image, followed by a region growing algorithm and level set smoothing to finally yield a skull-stripped image volume.

Each tissue type was homogenized in three different solvents: (1) 100% methanol (MeOH), (2) 10 mM phosphate buffer pH 7.5 (PB), or (3) an 85/15 (v/v) ethanol/10 mM phosphate buffer pH 7.5 mixture (EtOH/PB). Further, for each tissue type, two defined tissue to solvent ratios were tested [1 mg tissue to X µL solvent, denoted by 1:X (w/v)]. For most tissues, 40–100 mg frozen tissue were mixed in a ratio of 1 mg tissue with either 3 or 6 µL solvent per mg tissue [ratio 1:3 or 1:6 (w/v), respectively]. For bone, 6 or 9 µL solvent per mg tissue [ratio 1:6 or 1:9 (w/v), respectively] were used, because smaller solvent volumes often resulted in damaged tubes during homogenization. Tissues with low input material (e.g., adrenal and pituitary glands, of which only 1–2 mg were available) were homogenized with 12 or 18 µL extraction solvent per mg tissue [ratio 1:12 or 1:18 (w/v), respectively], since at least 10 µL homogenate were needed for the subsequent metabolite quantification. Table 1 provides an overview of all used tissue to solvent ratios.
Table 1

Tissue to solvent ratios used for the extraction of metabolites

Tissue type	Tissue to solvent ratio (w/v)
Liver	1:3 and 1:6
Kidney	1:3 and 1:6
Skeletal muscle (M. quadriceps femoris)	1:3 and 1:6
Fat (visceral)	1:3 and 1:6
Brain (cerebrum)	1:3 and 1:6
Pituitary gland	1:12 and 1:18
Lung	1:3 and 1:6
Bone	1:6 and 1:9
Adrenal gland	1:12 and 1:18
Testis	1:3 and 1:6
Ovary	1:3 and 1:6

Tissue to solvent ratios are denoted as 1:X, indicating 1 mg of tissue was homogenized with X µL solvent

For homogenization, the frozen tissue pieces were weighed and placed into pre-cooled (dry ice) homogenization tubes containing ceramic beads with a diameter of 1.4 mm (Precellys Homogenization Kit, CK14, PeqLab Biotechnology, Erlangen, Germany). Ice-cold extraction solvent was added to each tube and the tissues were then homogenized in a Precellys24 homogenizer equipped with an integrated cooling unit (PeqLab Biotechnology, Erlangen, Germany) for three times over 20 s at 5500 rpm with 30 s pause intervals to ensure constant temperatures during homogenization. Samples containing organic extraction solvents (MeOH, EtOH/PB) were homogenized at − 4 °C, and samples containing aqueous extraction solvent (PB) at 4 °C. After homogenization, the samples were centrifuged at 4 °C and 2300×g for 5 min, and the supernatants (“tissue extracts”) were used for metabolite quantification.
For reproducibility experiments, several portions of the same tissues were homogenized. In case of in-homogenous organs (e.g., kidney, liver, brain), regions of comparable consistency were taken as replicates (e.g., cortex region for kidney). In case the available amount of organ material was very small (pituitary and adrenal glands), organs of littermate mice were used for replicates.

Bovine ovaries from randomly cycling cattle and other endocrine tissues (pituitary glands, testes and adrenal glands) were obtained from an abattoir. Ovaries were dissected to obtain antral follicles ranging in diameter from 3 mm to 18 mm, and these were further processed to isolate the GC layer, TC layer, and follicular fluid, as described previously (Glister et al. 2010 (link)). Briefly, follicles were sorted into five different size classes: 3–4 mm (n = 8), 5–6 mm (n = 8), 7–8 mm (n = 9), 9–10 mm (n = 6) and 11–18 mm (n = 12). Each follicle was hemisected, and GC and TC layers were recovered for RNA extraction, while follicular fluid was recovered for steroid hormone analysis. Large follicles (11–18 mm) were reclassified according to their oestrogen to progesterone ratio (E:P ratio) in follicular fluid as either large oestrogen-active (LEA; E2:P4 ratio >1) or large oestrogen-inactive (E2:P4 ratio <1) follicles. Corpora lutea (CL) at growing (n = 4), mid-luteal (n = 5) and regressing (n = 4) stages were also harvested. All tissue samples were homogenized in Trizol reagent for total RNA extraction, as described previously (Glister et al. 2010 (link)).
Follicles (4–8 mm diameter) were also retrieved for isolation of GC and TC to be used for primary cell culture experiments, as described in detail elsewhere (Glister et al. 2001 (link), 2005 (link)). GC and TC were seeded into 96-well plates (Nunclon, Life technologies Ltd.) at a density of 75,000 cells/250μL/well for serum-free culture (non-luteinized cells) or 10,000 cells/250 μL/well for serum-supplemented culture (luteinized cells). Cells were cultured for 6 days at 38.5°C with saturating humidity in 5% CO₂ in air. The culture medium consisted of McCoy’s 5A medium (Sigma), supplemented with antibiotic/antimycotic solution (1% v/v; Sigma), apo-transferrin (5 μg/mL; Sigma), sodium selenite (5 ng/mL; Sigma), bovine insulin (10 ng/mL; Sigma), HEPES (20 mM; Sigma) and bovine serum albumin (0.1% w/v; Sigma). Medium used for serum-free GC culture was also supplemented with 10^–7 M androstenedione as aromatase substrate. For GC and TC cultured under conditions that promote luteinization, 2% fetal calf serum (FCS) was also included as a supplement. In all four culture models, media were changed after 48 h and replaced with fresh media containing treatments as specified below. This was repeated after a further 48-h incubation period. Media were retained after the final 48-h period (i.e. 96–144 h) for subsequent analysis of steroid hormone secretion. Viable cell number at the end of culture was determined by neutral red uptake, as described elsewhere (Glister et al. 2001 (link)).
It should be noted that culturing TC and GC using defined serum-free medium preserves a non-luteinized phenotype reflected by LH-induced androstenedione (A4) secretion by TC and follicle-stimulating hormone (FSH)-induced oestradiol (E2) secretion by GC. Henceforth, these cells will be referred to as non-luteinized TC (NLTC) and non-luteinized GC (NLGC). In contrast, culturing TCs and GCs under serum-supplemented conditions promotes spontaneous luteinization, as indicated by reduced A4/E2 secretion and greatly increased secretion of P4 (Glister et al. 2001 (link), 2005 (link), Kayani et al. 2009 (link)). Henceforth, these cells will be referred to as LTC and LGC.

We selected from GTEx in the following way. We required all samples to have a RIN score of greater than 6. For performance evaluation we chose to evaluate a comparison between cerebellum and skeletal muscle. We randomly selected 150 samples from both tissues, excluding the same donor from being selected in both tissues (Supplementary Data 2). For the brain subregions analysis, we selected all samples in GTEx v8 associated with brain tissue (not including pituitary gland). Samples were downloaded as FASTQ or as BAM and converted to FASTQ depending on when they were released. Samples that were part of v7 are available on SRA, so they were downloaded using SRA Tools (v2.9.6) as FASTQ files. New samples from the v8 release were only available as BAMs on the cloud, so they were downloaded using gsutil (v4.46) and converted to FASTQ using samtools (v1.9).

All patients were simulated in a head-first, supine position, immobilised with five-clamp thermoplastic covering the brain to chest level. If the patient was non-cooperative and no sedition was used, an additional two-clamp thermoplastic was used to immobilise the abdomen pelvis region^{7 (link)}. Anesthesia was used if needed. All patients were simulated in a Brilliance Big Bore CT scanner (Philips, Eindhoven, The Netherlands) with 3-mm uniform slice thickness from brain to mid-thigh, with the first marker in the brain and second marker at the abdomen level to keep the patient straight during the simulation. CT Images were transferred to the SomaVision (Varian Medical Systems, Palo Alto, CA) contouring station and co-registered with three-dimensional (3D) T1-contrast, T2-flair magnetic resonance images (MRI).
The gross tumour volumes (GTV) of the brain and the spine were delineated as follows: the cranial contouring included the whole brain and up to the junction of the cervical vertebrae C5 and C6. The superior end of the spinal cord starts from the end of brain GTV and goes up to the inferior end of the thecal sac, as seen on the sagittal view of the MRI. The planning target volume (PTV) for the brain was generated by applying a 3 mm margin on the GTV. For the spinal cord, the PTV was generated using a 7 mm margin over GTV^{7 (link)}. The brain and spinal PTVs were summed to generate a single PTV for the plan optimisation. To standardise the contouring of organs at risk for all patients, a predefined structure template consisting of bladder, bowel, brain stem, chiasm, cochlea (bilateral), duodenum, esophagus, eyes (bilateral), thyroid gland, heart, humerus head (bilateral), kidneys (bilateral), lacrimal gland (bilateral), larynx, lens (bilateral), lung (bilateral), mandible, optic nerve (bilateral), oral cavity, ovary (bilateral for female patients), parotid (bilateral), pituitary gland, rectum, stomach, and submandibular glands (bilateral) was used.