Micro-computed tomography (micro-CT) is commonly used to assess bone quality and to evaluate the outcome of experimental therapies in animal models of bone diseases. Generating large datasets is however challenging and data are rarely made publicly available through shared repositories. Here we describe a dataset of micro-CT reconstructed scans of the proximal part of 21 tibiae from wild-type mice, osteogenesis imperfecta mice (homozygous oim/oim) and oim/oim mice transplanted with human amniotic fluid stem cells. The dataset contains, for each sample, 991 8-bit Bitmap reconstructed images and a 3D reconstruction of the bone in the PLY format, available at the online repository Figshare. In line with the increasing effort to make scientific datasets open-access, our data can be downloaded and used by other researchers to compare their observations with ours and to directly test scientific questions on osteogenesis imperfecta bones without the need to generate complete datasets.
Machine-accessible metadata file describing the reported data (ISA-tab format)
Background & Summary
The availability of rodent models of bone diseases is essential for the fast progression of biomedical research and discovery of novel therapeutics. These models are used to mimic a variety of bone conditions, including osteoarthritis, osteoporosis, age-related bone loss and osteogenesis imperfecta1,
Since its introduction in the 1980s (ref. 9), micro-CT has become an indispensable tool across the skeletal research community for morphological analyses of both cortical and trabecular bone, leading to the wide-spread commercialisation of scanners. Micro-CT is a reliable and reproducible technique10,
We have made the micro-CT dataset used for the analysis of bone microarchitecture of WT, oim mice and oim mice transplanted with human amniotic fluid mesenchymal stem cells (hAFSCs) available by Figshare, an open access repository. These methods are expanded versions of descriptions in our related work (Ranzoni et al.18). All experimental protocols complied with UK Home Office guidelines (PPL 70/6857). Briefly, heterozygous male and female mice (B6C3Fe a/a-Col1a2oim/Col1a2oim, Jackson Laboratory) were housed at 21° C with a 12:12-hour light/dark cycle. All mice were handled throughout the study by the same operator. It was the first pregnancy for all females. Pregnant females were housed individually. Offsprings were genotyped by sequencing the oim fragment and housed at 21° C with a 12:12-hour light/dark cycle with their mother. All mice used were experimentally naïve prior to this study. At 3-4 days of age, homozygous oim neonates from the same litter were randomly assigned to the transplanted or non-transplanted group and we ensured that mice from at least six different litters were included in all groups (wild-type, oim non-transplanted and oim transplanted). Homozygous oim neonates were individually intraperitoneally injected (between 10:00 and 12:00) with hAFSCs (106 cells/mouse resuspended in 20μl cold PBS) under direct vision using a 33-G Hamilton Microlitre syringe (Bonaduz, Switzerland) without anaesthesia and replaced in their home cage with their mother immediately after injection. No blinding was implemented with regards to genotype. At the age of 30 days, mice were individually housed in filter cages (sawdust bedding) with a 12:12-hour light-dark cycle (21 °C), with water and chow (Purina, St Louis, MO) ad libitum. All experimental mice were culled at 8 weeks of age for analysis. The data set includes a total of 21 micro-CT mouse tibia axial scan sets including 6 wild-type mice, 6 oim and 9 oim transplanted with hAFSCs. Details on age and sex of the mice can be found in Table 1. One weakness of the study is that the oim non-transplanted group only includes males, whilst the two other groups (oim transplanted with stem cells and wild-type) contained both males and females; sex affects skeletal size and body weight. However, we have previously compared the effect of sex on the data reported by microCt, in particular bone mineral density, trabecular architecture and organisation, and found that sex had no significant impact on the microstructural properties assessed in the study. We have made this dataset accessible for re-use to the scientific community to encourage researchers to compare their datasets to ours and to therefore facilitate the investigation of new hypotheses.
Tibiae were isolated from immune competent 8-week old wild-type mice (B6C3Fe a/a-Col1a2+/+) (n=6), homozygous oim mice (B6C3Fe a/a-Col1a2oim/oim) (n=6) and homozygous oim mice transplanted with hAFSCs (n=9) and fixed in 10% neutral buffered formalin for 24 h. Samples were then washed in phosphate buffered saline and stored in 70% ethanol until scanning.
The bones were scanned with a Skyscan 1172 micro-CT scanner (Bruker, Coventry, UK), in small plastic tubes containing 70% ethanol. Double blinding was implemented during scanning measurement. The bones of wild type, oim non-transplanted and oim transplanted bones were randomly scanned to minimise any potential order effect. Scans were performed using a beam energy of 49KV and a flux 200μA, a 0.5mm aluminum filter and an isotropic pixel size of 5.06μm. The region scanned extended for 5 mm, starting from the proximal tibial epiphysis. We have compared the images and microstructure of oim bones scanned with frame averaging of 3 and frame averaging of 16 and found that the data related to microstructure were not significantly improved (data not shown). As there was no improvement in image quality, we have scanned all the bones of the study with a frame averaging of 3. Consequently, each image was acquired with an exposure time of 590ms, frame averaging of 3 and a 0.4 degree rotation step with a total rotation of 180 degrees and using medium camera resolution (2000x1048px). Random movement was set to ON and flat field correction was applied. Each scan includes a dataset of 991 stacked cross-sectional images, created for each scanned sample using NRecon segmentation software (Bruker) and stored as 8 bit Bitmap. The following settings were used for the reconstruction in NRecon: ring artifacts reduction was enabled and set to 1, smoothing was enabled and set to 1 to reduce background noise and beam hardening correction was enabled and set to 25%. Moreover, 3D reconstructions are available in the PLY format. Image stacks can be processed via CTAn software and 3D segmentation softwares CTVol and CTVox (Bruker) and via BoneJ, an online tool which can be downloaded for free19. Examples of reconstructed BMP images and 3D renderings obtained from CTVox are shown in Fig. 1.
The microCT dataset (Data Citation 1: Figshare https://dx.doi.org/10.6084/m9.figshare.c.3795019) comprises 21 sets of micro-CT scans of the proximal part of the tibia from 21 mice, including 6 from wild-type mice, 6 from non-transplanted oim mice and 9 hAFSCs-transplanted oim mice. Each scan set contains 991 stacked cross-sectional BMP images (approximate size of each folder 900MB). Additionally, 3D reconstructions are provided in the PLY format (Polygonal File Format).
The micro-CT scanner Skyscan 1172 is regularly serviced by Bruker UK engineers to verify the stability of the x-ray source. Density calibration was performed applying the Hounsfield unit method, before density measurement.
How to cite this article: Ranzoni, A. M. et al. Micro-computed tomography reconstructions of tibiae of stem cell transplanted osteogenesis imperfecta mice. Sci. Data 5:180100 doi: 10.1038/sdata.2018.100 (2018).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Ranzoni, A.M., Corcelli, M., Arnett, T.R., & Guillot, P.V. Figshare https://dx.doi.org/10.6084/m9.figshare.c.3795019 (2018)
This work was supported by grants from Action Medical Research, Rosetrees Trust and Newlife Foundation, and by the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital for Children NHS Foundation Trust and University College London.