Positron emission tomography and functional characterization of a complete PBR/TSPO knockout

The evolutionarily conserved peripheral benzodiazepine receptor (PBR), or 18-kDa translocator protein (TSPO), is thought to be essential for cholesterol transport and steroidogenesis, and thus life. TSPO has been proposed as a biomarker of neuroinflammation and a new drug target in neurological diseases ranging from Alzheimer’s disease to anxiety. Here we show that global C57BL/6-Tspotm1GuWu(GuwiyangWurra)-knockout mice are viable with normal growth, lifespan, cholesterol transport, blood pregnenolone concentration, protoporphyrin IX metabolism, fertility and behaviour. However, while the activation of microglia after neuronal injury appears to be unimpaired, microglia from GuwiyangWurraTSPO knockouts produce significantly less ATP, suggesting reduced metabolic activity. Using the isoquinoline PK11195, the ligand originally used for the pharmacological and structural characterization of the PBR/TSPO, and the imidazopyridines CLINDE and PBR111, we demonstrate the utility of GuwiyangWurraTSPO knockouts to provide robust data on drug specificity and selectivity, both in vitro and in vivo, as well as the mechanism of action of putative TSPO-targeting drugs.

A ccording to current concepts, the most important function of the mitochondrial membrane protein peripheral benzodiazepine receptor (PBR)/18-kDa translocator protein (TSPO) is the regulation of steroid hormone production by helping to translocate cholesterol, the precursor to pregnenolone, across the aqueous mitochondrial intermembrane space, a view emphasized by the recent renaming of the PBR as translocator protein (TSPO) 1,2 . A large body of literature 2-4 on the regulatory influence of the PBR/TSPO on steroid-mediated homeostasis, including the reported embryonic lethal effects of a Tspo knockout 5 , has established the PBR/TSPO as essential for life. Its critical role as an endocrine regulator is the prevailing explanation for the observed actions of PBR/TSPO binding drugs across an exceptionally broad therapeutic spectrum from the antiinflammatory treatment of Alzheimer's disease 6 to anxiolysis without direct effects on the central GABA A receptor protein complex 4,7 . However, recent observations that a conditional knockout in testicular Leydig cells appeared not to affect hormone production 8 have controversially been interpreted as evidence that the PBR/TSPO, unlike the steroidogenic acute regulatory protein (StAR) 9,10 , is not an essential requirement for steroid hormone biosynthesis 11,12 . No further data indicating other potential impairments have been reported.
An important observation that has underpinned the growing interest related to the PBR/TSPO is the regularly seen increase of the PBR/TSPO in areas of brain injury and during 'neuroinflammation', most prominently in activated microglia 1,13,14 .
Our study provides a first extensive reference description of the constitutive phenotype of a global Tspo knockout animal model in vitro and in vivo. It addresses some of the fundamental questions, in regard to the long-term viability of animals without any potentially compensating PBR/TSPO-expressing cells, normalized to Gapdh and Actb) (d) and measurement of TSPO protein by western blot of lysates from the kidney, spleen and testis (e) confirmed the complete absence of any Tspo gene product in Tspo À / À mice.
whether they have altered steroid levels or any other loss of function and whether specificity and selectivity can be assumed for some of the most widely used PBR/TSPO-ligands.

Removal of exons 2 and 3 of the Tspo gene resulted in viable
Tspo þ / À and Tspo À / À animals. Following the removal of exons 2 and 3 only exons 1 and 4 remain, both of which do not contain any start codons in the TSPO reading frame. Therefore, no TSPO protein, or truncated TSPO protein can be produced (Fig. 1a).
A more detailed illustration of how the loss of exons 2 and 3 and subsequent merger of exon 1 and exon 4 cannot give rise to any functional fragment of the PBR/TSPO but possibly only an unrelated protein with no sequence similarity is shown in Supplementary Fig. 1. The targeted deletion of Tspo and complete loss of TSPO protein was confirmed by Southern blot, PCR, RT-PCR, RT-qPCR, Western blot, (Fig. 1b-e and Supplementary Fig. 1), specific antibody staining against amino acids 156-169 at the C-terminus of the PBR/TSPO in tissues and macrophages from Tspo þ / þ , Tspo þ / À and Tspo À / À mice (Fig. 2), in vivo tracer kinetic PET/CT studies using the PBR/TSPO ligand [ 18  (a) Anti-TSPO antibody staining showed the presence of TSPO (here shown in the kidney and testis) in the wild-type and absence in the knockout Tspo À / À mice. The slides were the same as previously used for autoradiography with the selective TSPO-binding ligand [ 3 H]PK11195 (Fig. 4). (b) Antibody staining of Tspo þ / þ and Tspo þ / À macrophages validates mitochondria as the primary site of the TSPO, which is entirely absent in Tspo À / À mice. No obvious difference in intracellular density or distribution of the mitochondria was detected in the Tspo À / À mice (green ¼ TSPO; red ¼ mitochondria; yellow ¼ merged image; blue ¼ nucleus; scale bars: (a) 500 mm and (b) 20 mm). In addition, PBR/TSPO receptor membrane-binding data as well as extensive receptor autoradiographic validation for all major organs and the whole body of neonatal mice in all three genotypes confirm the absence of the PBR/TSPO protein in the Tspo À / À mice and the high selectivity of [ 3 H]PK11195 in tissues where the PBR/TSPO is present (Fig. 4a,c Fig. 6a-e). Our data also indicate that the early stage of perineuronal microglial activation, with its typical change in microglial morphology, is not noticeably influenced by the loss of the PBR/TSPO and that the neuro-glial signaling mechanism remains intact ( Fig. 6f,g).
We further demonstrate the background-free detection of syngeneic PBR/TSPO-expressing glioma cells growing in the brains of Tspo À / À animals in vitro and in vivo using the selective PBR/TSPO ligands, [ 3 H]PK11195 and [ 18 F]PBR111, as well as antibody staining against the PBR/TSPO. This approach tests simultaneously for the absence (respectively presence) of several recognition or binding domains that make up the full PBR/TSPO, whereby the PBR/TSPO-expressing tumour serves as an internal positive control within the same animal. As predicted from the readable sequences remaining after the deletion of exons 2 and 3, the tissue of Tspo À / À animals cannot express any functional domains of the PBR/TSPO or similar proteins, whereas the Tspo þ / þ tumours express the protein as expected ( Fig. 7 and Supplementary Fig. 3).
having the same litter size, length of 1-day-old pups, sex ratio, and subsequent growth rate and weight increase from 3 to 50 weeks (weekly measurement) as the colony control wild-type Tspo þ / þ animals ( Fig. 8a,b). Follow-up monthly measurements from 50 to 80 weeks failed to reveal any significant genotypedependent weight differences in older animals.
Haematological analysis and blood biochemistry. Analysis of blood, including cell sorting, demonstrated that the blood phenotype in both sexes remained largely unaffected by the heterozygous or homozygous loss of the Tspo gene, the only exception being a statistically significant trend increase in natural killer (NK) cells in female Tspo À / À mice compared to female Tspo þ / þ (Supplementary Table 3). Thus, the Tspo À / À mice did not reveal the changes in cellular blood composition seen in zebrafish tspo gene silencing experiments 15 .
The concentrations of protoporphyrin IX (PPIX), thought to be an endogenous PBR/TSPO ligand, in the blood of Tspo þ / þ , Tspo þ / À and Tspo À / À mice were indistinguishable and increased similarly after treatment with 5-aminolevulinic acid (ALA), the first intermediate in haem biosynthesis and a precursor of PPIX (Fig. 8d), thus indicating normal haem metabolism.
Absence of compensatory transcriptomic regulation. Likewise, quantitative RT-PCR showed that the gene expression profiles for StAR, P450scc and Tspo2 across the major organs known to express the Tspo were similar across genotypes, indicating that the global loss of Tspo had not led to any compensatory transcriptomic regulation of those genes thought to be most closely linked to PBR/TSPO as a regulator of steroidogenesis ( Fig. 8e).

Mitochondrial respiration and ATP production in microglia.
Measurements of mitochondrial respiration and ATP production in primary cultures of mitochondria-rich microglia from Tspo þ / þ and Tspo À / À mice using the Seahorse XF96 Cell Analyzer and selective inhibitors of the electron transport chain revealed that Tspo À / À mitochondria, compared to Tspo þ / þ , have a consistently and significantly reduced ATP production as determined by inhibition of complex V (ATP synthase) with oligomycin ( Fig. 9).

Discussion
Orthologues of the PBR, or TSPO, are found widely among eubacteria, archaea and eukaryotes 2,3 . Most members of the PBR/TSPO protein family contain a specific binding site for the isoquinoline carboxamide PK11195, which has been used to pharmacologically define the PBR/TSPO and distinguish it from other benzodiazepine-binding receptors, such as the GABA A receptor protein complex 1,3,16 . An evolutionarily younger C-terminal cholesterol recognition amino-acid consensus (CRAC) domain is primarily found in the animal phylum. A paralogue of the PBR/TSPO, the TSPO2, exists in birds and mammals as the result of a gene duplication, which has retained the CRAC domain but lost the isoquinoline-binding site 17 .
In light of the evolutionary pressure to conserve the gene, its documented influence on steroidogenesis and the previously observed embryonic lethality caused by the knockout of either the Tspo 5 or the endogenous PBR/TSPO ligand, Acyl-CoA binding protein 18 , the survival and overtly normal phenotype of a global Tspo knockout animal might seem unexpected. However, the evolutionary preservation of an ancient gene does not necessarily imply that it is essential, that is, embryonic lethal, nor that an ancient and non-essential gene is less likely to be relevant for the emergence of disease. On the contrary, phylostratigraphic analysis of the human genome suggests that non-essential ancient genes predominate among the disease-associated genes 19 . PBR/TSPO may be considered a non-essential disease gene that lies at the functional periphery of the interactome with more limited direct interactions than an in utero essential hub gene 20 . Consequently, the functional impact on the Tspo knockout would be predicted to be discreet.
Our observation that microglia isolated from Tspo knockouts have altered oxygen consumption and ATP synthesis rates points to a potential latent phenotype that may be elicited under those disease conditions where a significant regulation of the PBR/TSPO is known to occur 13,14 . This raises the intriguing possibility that the disease course notably in inflammatory pathology of the brain, may be different in the absence of the normally observed, concomitant upregulation of microglial PBR/TSPO. Immune-modulatory functions of the PBR/TSPO have previously been suspected based on pharmacological studies using PBR/TSPO binding compounds that appeared to abrogate inflammatory responses [21][22][23] .
Earlier work in bacteria 24 suggests other functions than cholesterol transport for the PBR/TSPO, such as oxygen sensing and thus co-regulation of mitochondrial ATP production 25 , which may become relevant under a range of disease conditions. Possible functional links of the TSPO to the F o sub-unit of the ATP synthase through the VDAC on the outer and ANT on the inner mitochondrial membrane have been described 25,26 .
Perceivably, PBR/TSPO-mediated changes in ATP production might explain indirect regulatory effects on the energy-dependent steroid biogenesis 1,3,10 , particularly under stress challenges. Tspo Anti-CD11b Tspo -/-  Many studies use compounds with purported selectivity for the PBR/TSPO for diagnostic or therapeutic purposes, such as those modulating complex behaviour in the apparent absence of binding to neural receptor systems 7 . Our study paradigmatically demonstrated for a long-established as well as a recently developed PBR/TSPO compound, that is, PK11195 and CLINDE/PBR111, respectively, that the Guwiyang Wurra Tspo À / À strain is a powerful tool to discern selective from non-selective off-target binding and thus validate the purported drug actions in live animals under naturalistic conditions. We also anticipate innovative lipid membrane studies into the biophysical mechanisms by which ligands for the PBR/TSPO may either modify or introduce additional, potentially nonreceptor, actions in the presence or absence of the protein 1,27,28 . Lastly, the apparent irrelevance of the PBR/TSPO for the normal activation of microglia after neuronal injury indicates that neuro-glial interactions and inflammatory tissue responses are separate processes, though in vivo neuroimaging studies on both may regularly show high levels of PBR/TSPO expression 13,14 . The current concepts of 'neuroinflammation' do not adequately distinguish these and, therefore, too, merit revisiting.
The extensively characterized Guwiyang Wurra Tspo À / À strain provides the broader research community with a means of dissecting the network of pathways through which the PBR/TSPO exerts its regulatory influence at the intersection of cell metabolism, innate immune response and endocrine regulation of behaviour in health and disease.

Methods
Generation of Tspo À / À mice. Tspo knockout mice were generated using a targeting construct that contained loxP sites flanking exons 2 and 3, and a neomycin cassette inserted between exons 3 and 4. The targeting construct was electroporated into C57BL/6 Bruce4 embryonic stem (ES) cells and cells correctly targeted by homologous recombination were injected into albino C57BL/6 blastocysts. Male chimeras were mated to female albino C57BL/6 mice and the resulting offspring with a black coat were screened for the presence of a targeted Tspo allele.
Mice positive for the presence of the targeted allele were crossed with C57BL/6 Cre deleter mice to remove the neomycin cassette and exons 2 and 3 to create heterozygous global Tspo deficient mice. To remove the Cre transgene, animals were bred to wild-type C57BL/6 mice.
To produce animals for experiments, heterozygous mice were crossed to generate wild-type littermate controls. All animal procedures were approved by the University of Sydney Animal Ethics Committee and the ANSTO Animal Care and Ethics Committee. The Tspo À / À mouse strain has been given the additional designation Guwiyang Wurra ('fire mouse' in the local Dharawal language). Future naming will thus be C57BL/6-Tspo tm1GuMu(GuwiyangWurra) .
Mice were genotyped by Southern blot analysis using genomic DNA isolated from tail biopsies. For routine genotyping, genomic DNA was isolated from tail biopsies, ear biopsies or stool samples and amplified by PCR to generate a 489-bp product for the wild-type allele and a 246-bp product for the knockout allele. See Supplementary Table 4 for primer sequences. The PCR consisted of an initial incubation at 95°C for 2 min, followed by 4 cycles at 95°C for 30 s, 68°C for 30 s and 72°C for 2 min, then 4 cycles at 95°C for 30 s, 65°C for 30 s and 72°C for 2 min, then 30 cycles at 95°C for 30 s, 62°C for 30 s and 72°C for 2 min, and a final step at 72°C for 5 min.
RT-PCR and quantitative real-time PCR. Tissue samples were homogenised in TRIzol (Life Technologies) and total RNA isolated using the PureLink RNA mini kit (Life Technologies) with on-column DNase treatment following the manufacturer's instructions. Purified RNA (1 mg) was reverse-transcribed to cDNA using the SuperScript III First-Strand Synthesis kit (Life Technologies). For RT-PCR, cDNA from the kidneys of Tspo þ / þ and Tspo À / À mice was amplified using and subsequent immunohistochemical staining with anti-TSPO antibody on the same brain sections (c) confirm that PBR/TSPO ligand binding is strictly limited to the brain tumour region and no PBR/TSPO binding or recognition sites are present in Tspo À / À tissue. Scale bars, 1 mm. The PET images show the time frame from 5 to 10 min after injection and scaling is 0-5.3% ID per cm 3 for (a). Protoporphyrin IX. Blood from euthanized mice was placed into heparin tubes on ice, plasma removed, cells washed twice in ice-cold phosphate-buffered saline (PBS) and then resuspended in PBS to the original volume. Samples were evenly split (300 ml each), one receiving 1 mM 5-ALA and the other sterile H 2 O, and incubated in a shaking incubator at 37°C for 4 h. Ethyl acetate/acetic acid (4:1) was added to each sample, then mixed and centrifuged before transferring the supernatant (1 ml) to a new tube with 1 ml of 1.5 M HCl. An aliquot of the lower HCl phase, after mixing and centrifuging the sample, was measured in a spectrofluorimeter (excitation wavelength 407 nm, slit width 10 nm, emission spectrum 450-800 in 1 nm increments). Glass tubes and containers were used throughout and samples were protected from light.
A small incision (0.5-0.7 cm) in the skin 0.5 cm dorsal and lateral to the ear was made and the facial nerve transected under an operating microscope. The incision was closed with Vetbond Tissue Adhesive (3M, St Paul, MN, USA). The success of the operation was confirmed by the loss of ipsilateral whisker movement. Mice were monitored carefully and euthanized with CO 2 3 days after the operation. Brains were dissected, placed in OCT, transferred to liquid nitrogen and stored at À 80°C until cryo-sectioning.
Stereotactic implantation of GL261 mouse glioblastoma cells. The Tspo þ / þ (n ¼ 6) and Tspo À / À mice (n ¼ 4) were operated under isofluorane anaesthesia and aseptic conditions using a Kopf stereotactic apparatus and precision dental drill (Kopf Instruments, Tujunga, CA, USA). A small burr hole was made 3 mm anterior of the bregma and 2 mm lateral over the right frontal lobe (3 mm ventral). GL261 glioblastoma cells (NCI, Bethesda, MD, USA) at 3 Â 10 3 per ml in a volume of 10 ml were slowly injected over 2 min using a 10-ml Hamilton syringe connected to a pump by means of a 32-gauge needle. The animals were closely monitored daily for good recovery, including body weight and absence of neurological signs. Three and four weeks after implantation of the GL261 cells, the animals were imaged by means of microPET/CT and brains were dissected for histological studies and autoradiography.
PET and CT imaging. Mice, anaesthetized (5% (v/v) with isoflurane and maintained at 1-2%, were scanned using a small-animal Inveon PET/CT scanner (Siemens, Knoxville, TN, USA) according to the methods described previously 31,32 . Body temperature was maintained with a feedback regulated heating pad and . Independent from genotypes, females weigh significantly less than males, as is known to be the case for this mouse strain (b). (c-e) No significant differences between Tspo þ / þ and Tspo À / À mice are found in blood pregnenolone concentrations (n ¼ 4 for all groups; Student's t-test) (c); blood protoporphyrin IX (PPIX; left: three typical closely overlapping spectra from the different genotypes after addition of 5-aminolevulinic acid (ALA), right: no significant differences in levels of PPIX (PPIX per ml of blood in arbitrary units (A.U.); n ¼ 3-5 per genotype; Student's t-test)) (d); and mRNA expression of steroidogenic acute regulatory (StAR) protein, P450scc and Tspo2 in nine organs (normalized to reference genes Actb and Gapdh, n ¼ 3 per genotype) (e). Error bars denote standard deviation. respiration monitored (BioVet; m2m Imaging Corp, Cleveland, OH, USA). Scans started with the tail vein injection of [ 18 F]PBR111 (8-18 MBq per 100 ml, 0.2 nM). After 40 min of imaging, PBR111 (1 mg per kg in 2% acetic acid-saline) was injected to determine non-specific accumulation in organs and imaged for 10 min. Then, mice underwent a 10-min CT scan for anatomical information. All PET data were corrected, normalized and reconstructed with an OSEM3D-MAP algorithm 31 to produce PET volumes of activity concentration (kBq per ml).
Autoradiography. Tissue sections from Tspo þ / þ , Tspo þ / À and Tspo À / À mice were incubated at room temperature for 20 min in 170 mM Tris-HCl pH 7. Immunohisto-and cytochemistry. Immunohistochemistry was performed as previously described 13,33,34 . Tissue sections (used for autoradiography) were fixed with 3.7% formaldehyde in PBS for 5 min, then permeabilized with ice-cold acetone. Non-specific antibody binding was blocked with 10% horse serum and 2% bovine serum albumin (BSA) in PBS. Sections were incubated with TSPO monoclonal antibody (#109497, Abcam) at 4°C overnight and secondary HRP-conjugated goat anti-rabbit antibody at RT for 1 h (#A0545, Sigma-Aldrich). The activity of HRP was detected with the peroxidase 3,3 0 -diaminobenzidine tetrahydrochloride liquid substrate system (Sigma-Aldrich). Sections were dehydrated with ethanol, incubated with xylene, and slides were mounted in DPX mounting media with a coverslip. Sections were visualized under an inverted Olympus BX51 microscope (Olympus, Tokyo, Japan), and captured with a Q-imaging camera and ImagePro 5.1 program. A Zeiss LSM 710 confocal microscope with a 5 Â EC Plan-Neofluar NA 0.16 objective was used for the fluorescence microscopy of the facial nucleus. Alexa Fluor 568 was excited with a 561-nm laser, using a 488/561/633 dichroic mirror, and emission captured from 565 to 640 nm.
For immunocytochemistry, peritoneal macrophages were isolated according to Zhang et al. 35 . After 2 days in complete cultural medium the purity of macrophages was 497% as confirmed with the macrophage-specific marker IB4 conjugated with FITC (Sigma-Aldrich). Cells on coverslips were fixed and incubated using the same method as for tissue sections, with the addition of a mouse anti-human/mouse mitochondrial electron transport chain complex IV antibody (#AB14705, Abcam), then incubated with the secondary antibodies Alexa Fluor (AF) 594-conjugated goat anti-rabbit antibody (Life Technologies) and AF488-conjugated goat antimouse antibody (Life Technologies). Cells on coverslips were mounted on microscope slides with ProLong Gold antifade reagent containing DAPI (Life Technologies) and viewed under a BX61WI Olympus microscope. Images were acquired with a digital camera (CoolSNAP, Photometrics, Tucson, AZ, USA) and the Image InVivo program (Photometrics). Deconvolution of images was performed with the AutoDeblur program (Photometrics) and further processed with ImageJ (NIH, Baltimore, MD, USA).
Further extensive hematological and behavioural data are presented in Supplementary Tables 2 and 3.
Microglia were purified by shaking at 350 revolutions per minute (r.p.m.) for 50 min at 37°C, pelleted by centrifugation, resuspended in supplemented DMEM/ F12 and placed (4 Â 10 4 cells per well) into a 96 well Seahorse XF cell culture plate (Seahorse Bioscience, North Billerica, MA, USA) pre-coated with FBS for 15 min at room temperature. After 15 min the wells were washed twice to remove unattached cells. The purity of microglial cultures was 498% as confirmed by staining with the microglial marker Alexa Fluor-conjugated isolectin GS-IB4 (Life Technologies). The 5 Â 10 4 microglia from wild-type and knockout mice were plated into each well of the XF96 plate and were grown for 2-3 days before evaluating mitochondrial functions with the Seahorse XF 96 Cell Analyzer (Seahorse Bioscience). To rule out systematic differences in cell density and viability between Tspo þ / þ and Tspo À / À microglia, cell density was monitored and assured throughout the experimental measurements, while viability, including apoptosis, was confirmed using caspase 3/7 (ref. 36).
Measurements of mitochondrial respiration. The oxygen consumption rate (OCR) and extracellular acidification rate in primary microglial cultures were determined using the Seahorse XF 96 Cell Analyzer. OCR is an indicator of mitochondrial respiration 37,38 . The day before the experiment 200 ml of the Seahorse XF Calibrant (pH 7.4) was added to each well of the XF 96-well utility plates. The sensor cartridge was placed on top of the plate and hydrated for 16 h at 37°C in a non-CO 2 incubator. The protocol for basal measurements of the cells at 37°C (start, calibrate probes, equilibrate, mix 2 min, 1 min delay, measure 3 min) was repeated twice. Mitochondrial stress compounds (oligomycin, rotenone combined with antimycin A) were then injected, followed by 2 min mixing, 1 min delay, measure 3 min (repeated 3 times).
On the day of the experiment, microglia (plating density 5 Â 10 5 per well, purity 498% and viability 492% for both wild-type and knockout microglia) were washed 3 times with un-buffered DMEM plus 20 mM glucose (no buffers) and incubated at 37°C (no CO 2 ) for 1 h. Oligomycin and rotenone combined with antimycin A were used to determine mitochondrial ATP production and basal mitochondrial OCR 37,38 . Oligomycin and the combination of rotenone and antimycin A were injected at concentrations of 0.3, 1 and 3 mM (optimal for oligomycin) and 10 mM (optimal for rotenone and antimycin A).
The data for each reported parameter reflect 11-14 independent Seahorse XF measurement series from a total of seven independently repeated microglial cell culture preparations.
Radioligand binding. Organs were homogenized using a T25 digital Ultra-Turrax homogenizer (Ika, Wilmington, NC, USA) at speed setting 5 and 20,000 r.p.m. in B45 ml of ice-cold TRIS buffer (pH 7.4) and washed twice in ice-cold 50 mM Tris-HCl (pH 7.4) by centrifugation at 48,000 g, resuspended in 50 volumes of ice-cold 50 mM Tris-HCl, and protein concentration was determined using a BCA protein assay kit (Thermo Scientific). Bmax and Kd were determined with saturation binding by incubating membrane protein (60 ml) with [ 3 H]PK11195 (0.56-20 nM) on ice for 90 min before harvesting by rapid filtration through Whatman GF/C filters (GE Healthcare) presoaked in a 0.5% polyethylenimine solution. Non-specific binding was determined with 5 mM PK11195. Filters were placed in scintillation cocktail (Perkin Elmer) for 12 h and radioactivity was determined using a Tri-Carb 2100 TR Liquid Scintillation Counter (Perkin Elmer). Bmax and Kd values were obtained by non-linear regression using GraphPad Prism 5.04 (GraphPad Software, La Jolla, CA, USA) by fitting total and nonspecific binding.
Cholesterol transport and biosynthesis. P450scc metabolism was measured in mouse testis according to Tuckey et al. 39 and Slominski et al. 40 . Briefly, an enriched mitochondrial fraction was prepared by centrifugation in 250 mM sucrose, 50 mM Tris (pH 7.4), and the enzymatic conversion of cholesterol to pregnenolone was conducted under the following conditions out to a maximum 2-h time-point: 50 mM HEPES pH 7.4, 250 mM sucrose, 20 mM KCl, 5 mM MgSO 4 , 0.2 mM EDTA, 1 mg per ml BSA, 8 mM Trilostane, 0.05 mCi 3 H-cholesterol, 500 mM isocitrate and 5 mM NaNADP. Reactions were terminated with the addition of 4°C dichloromethane. The organic phase was retained and the aqueous phase was re-extracted twice more into dichloromethane. Extracts were combined, dried under nitrogen and dissolved in 100 ml of ethyl acetate. Pregnenolone and cholesterol were separated by thin-layer chromatography, with hexane:diethylether:acetic acid (15:15:1) mobile phase, and the amount of radioactivity in each spot quantified by liquid scintillation counting.
Detection of serum pregnenolone. Serum pregnenolone of male and female Tspo þ / þ and Tspo þ / þ mice was determined by ELISA performed according to the manufacturer's instructions (Abnova Cooperation, Taiwan). Briefly, blood was collected, allowed to clot, and the serum collected. Standards, controls and samples were incubated in duplicate in a sealed plate with or without conjugate and antibody for 1 h at room temperature on a plate shaker (KS4000ic, IKA, Selangor, Malaysia) at 200 r.p.m. After the incubation, wells were emptied, washed five times, and the plate tapped on lint free paper towel to remove any remaining buffer. Wells were incubated with HRP conjugate on the plate shaker at 200 r.p.m. for 30 min at room temperature, washed five times, and incubated with TMB substrate for 10 min at room temperature. Stop solution was added to each well and after 20 min the optical density at 450 nm was measured. Data were normalized against blood volume and analysed using a four-parameter logistic curve fitting program MasterPlex ReaderFit: Curve-Fitting Software for ELISA Analysis (Hitachi Solution America, San Francisco, CA, USA).
Blood phenotyping. Thirty mice (10 Tspo þ / þ , 10 Tspo þ / À , 10 Tspo À / À , 5 months old, males/females evenly distributed) used in behavioural experiments were euthanized with CO 2 . Blood was collected from the pulmonary cavity by cutting the inferior vena cava and placed in microtubes with heparin (haematology and lymphocyte analysis) or gel clot activator (biochemical analysis). Spleens were collected in 10 ml of ice-cold sterile PBS. Blood (room temperature) and spleens (on ice) were delivered (within 3-7 h of collection) to the Australian Phenomics Facility for flow cytometric and biochemical analysis.
For general haematology, blood was diluted 1:2 and analysed with fluorescenceactivated cell sorting using the Advia 2120 Haematology System (Siemens, Munich, Germany). Blood cell subsets, including white blood cells and platelets, were assessed. For lymphocyte analysis, blood was analysed for the presence or absence of abnormalities in T cells, B cells, NK cells and monocytes. The T-cell subsets included CD4 þ and CD8 þ subsets, while the B-cell subsets included immature and mature B cells and IgE þ monocytes. Spleens were analysed by flow cytometry and stained for NK cells, for B cells, including mature and immature subsets, and for T cells, including CD4 þ and CD8 þ cells and their activated subsets. See Supplementary Table 3 for a list of cell types examined.
For biochemistry, blood was centrifuged, the serum collected and run through the Olympus AU400 Chemistry Analyzer (Olympus), which detects levels of cholesterol, triglycerides, glucose, high-density lipoprotein (HDL), albumin, creatine kinase activity and alanine aminotransferase activity. See Supplementary Table 3 for a list of biochemical parameters examined.
Mouse sperm analysis. Sperm were released from the cauda epididymides of Tspo þ / þ and Tspo À / À mice by cutting each cauda with a scalpel and incubating in 150 ml of pre-warmed Biggers-Whitten-Whittingham (BWW) medium for 5 min at 37°C. After the incubation, an additional 250 ml of BWW medium was added, the tissue removed and the sperm assessed using a computer-assisted sperm analysis system (Hamilton Thorne, Beverly, MA, USA). Parameters assessed were total motile, progressive motile, VAP, VSL and VCL.
Assessment of anxiety-related behaviours. Behavioural tests were performed on mice aged between 3-4 months, handled regularly to reduce the impact of handling stress during testing. Anxiety-related behaviours were examined with the following: open-field test, emergence test, light/dark preference test and elevated plus maze. In each test the mice were allowed free exploration in the apparatus for 15 min. The duration and frequency of entering each compartment, total distance travelled, time spent active and the number of stool droppings were assessed. In the elevated plus maze risk assessment was defined as when an animal's head was facing towards an open arm with more than 15% of its body in an open arm while maintaining its centre of mass within the closed arm or central platform. The genotype of the animals was blind to the experimenter during testing. Tests were carried out at least 7 days apart to reduce the effects of training history 41 . All experiments were recorded with an overhead camera on to DVD and later analysed using Motman Tracker 4.5 software (Motion Mensura, Cooks Hill, NSW, Australia).
Data analysis. Statistical significance was calculated using an unpaired two-tailed Student's t-test for two-group comparisons. For comparisons of more than two groups univariate analysis of variance (ANOVA) with a Bonferroni's post-hoc test was performed and considered significant at Po0.05. Data are presented as mean ± standard deviation, except where indicated otherwise.