Calcification
Saos-2 osteosarcoma model of calcification/bone deposition
Time course ID: human_Saos-2_osteosarcoma
Sample provider: Kim Summers
Introduction
The formation of bone is a multi-stage process. It begins with mesenchymal cells committed to becoming cartilage. These cells condense into nodules and differentiate into chondrocytes which proliferate rapidly to form the template for the developing bone. They secrete a cartilege-specific extracellular matrix. blood vessels sinvade the cartilage structure. Mesenchymal precursors surrounding the cartilage cells then differentiate into osteoblasts which begin to form an extracelular matrix specific for bone. At the same time the chondrocytes within the osteoblast shell begin to die from apoptosis leaving a space that becomes the bone marrow. osteoblasts themselves become embedded into the bone matrix and differentiate further into osteocytes. Bone is constantly being remodelled by the action of osteoclasts, cells of the monocyte lineage that degrade the bone and osteoblasts on the surface which replace it with new bone. Thus the formation and maintenance of bone is a dynamic process.
During the formation of bone, osteoblasts lay down a matrix of hydroxyapatite, a mineral containing calciumand phosphorous which makes up to 50% of the weight of bone. this process of mineralisation can be simulated in vitro in cell cultures treated with chemicals to induce calcification. Many cell lines, primarily derived from osteosarcomas, can be induced to mineralise in this way, as can primary vascular smooth muscle cells. It is important to understand the process of mineralisation because formation of the bony skeleton is critical to the proper functioning of the vertebrate organism and because ectopic mineralisation can occur in genetic and environmental disease. For example, calcification is a key finding in forms of arterial disease including atherosclerosis.
Samples
In this study we used the human osteosarcoma cell line Saos-2. This line was first established from an osteosarcoma isolated from an 11 year old female patient in 1973 [1][2]. The line has been commonly used to study extracellular matrix and mechanical mechanisms in both expression and mineralisation studies [3] although the utility of Saos-2 as a human model system for normal osteoblast formation and function has been debated. Varying cellular morphology and proliferation has been observed [4], but Saos-2 demonstrates physiological levels of multiple osteoblastic markers including, osteocalcin (OC), bone sialoprotein (BSP) and decorin (DCN) [3]. In addition, active levels of alkaline phosphatase (ALPL) and the capacity to mineralize render Saos-2 a useful model system in the study of osteoblast function and ECM formation [5].
This experiment was designed to determine promoter specificity and gene expression patterns of key regulators early and late in the mineralization process using this human model system. To control for effects of serum cell proliferation and hyperconfluence we also submitted two control samples: mock-treated Saos-2 cells that had medium changes at the same time points but were not treated with BGP/ascorbic acid and MG63 osteosarcoma cells that were treated but failed ot mineralise. MG63 cells lack high levels of alkaline phosphatase and are not able to mineralise under these conditions, but they may respond to the signal by upregulating some required genes. [3][6].
The experiment was designed across 18 time points: 0, 15, 30, 45, 60, 80,100,120,150 and 180 minutes, 4, 8, and 24 hour as well as 4, 7, 14, 21, and 28 days following calcification induction. Cells were plated into 6 well plates (NUNC) or flasks (T25 or T75) at approximately 100,000 cells / well and mineralization was induced two days later with 50µg/ml ascorbic acid (Sigma) and 2.5mM ß-glycerophosphate (BGP) (Sigma) in medium with 10% serum. Medium was replaced with fresh medium including 10% serum and BGP/ascorbic acid every two to three days. All time point samples were processed for RNA at the same time relative to medium-change. RNA from the 18 time point samples were isolated through the following methods. The cells were lifted using 1x Trypsin-EDTA solution (Sigma) and RNA was isolated using the RNA Bee protocol (Ambisco). The sample RNA was quantified using a Nanodrop spectrophotometer (Nanodrop, USA). Three biological replicates were performed.
Quality control
Mineralization was verified using alizarin red staining of calcium at time point 0, 7, 14, 21 and 28 days. The cells were fixed with 4% paraformaldehyde (Sigma) for 10 minutes at room temperature, and rinsed with 1% PBS solution. The cells were stained using 1ml of 2% Alizarin Red (pH 4.2) for 5 minutes at room temperature. The wells were washed 3 times with dH20. The wells were extracted with 10% cetypyridium chloride and the extract used for quantification. The optical density was recorded using Multiskan ascent (Thermo) plate reader at a wavelength of 570nm.
The image shows Saos-2 cells at 0, 1, 2, 3 and 4 weeks after initiation of calcification. Red staining shows calcium deposition.
The graph shows quantification of calcium deposition in Saos-2 and MG63 cells treated with BGP and ascorbic acid and in untreated Saos-2 cells. Under treatment, Saos-2 cells deposit calcium in the matrix which MG63 do not and resemble the untreated Saos-2 cells. Pictures show the Saos-2 culture at 0 and 3 weeks where dark patches are caused by the deposition of calcium.
Marker gene expression
Key upregulated marker genes that indicate success of the mineralization:
COL1A2 [7], PHOSPHO1 [8], IFITM5 [9],SOST [10], DMP1 [10]
All these are markers of mineralisation in osteoblasts or the later stage osteocytes. All markers increased from Day1 or Day 4 and in general remained high for the remainder of the time course, indicating that calcification of the Saos-2 cultures involved induction of genes associated with miineralisation in vivo.
References
[1] Rodan S.B. et al., Characterizatin of a human osteosarcoma cell line (Saos-2) with osteoblastic properties. Cancer Res, 1987. 46(18): p. 4961-6.
[2] Fogh, J., J.M. Fogh, and T. Orfeo, One hundred and twenty-seven cultured human tumor cell lines producing tumors in nude mice. J Natl Cancer Inst, 1977. 59(1): p. 221-6.
[3] Pautke, C., et al., Characterization of osteosarcoma cell lines MG-63, Saos-2 and U-2 OS in comparison to human osteoblasts. Anticancer Res, 2004. 24(6): p. 3743-8.
[4] Boskey, A.L. and R. Roy, Cell culture systems for studies of bone and tooth mineralization. Chem Rev, 2008. 108(11): p. 4716-33.
[5] McQuillan, D.J., et al., Matrix Deposition by a Calcifying Human Osteogenic Sarcoma Cell Line (SAOS-2).Bone, April 1995. 16(4):p.415-26
[6] Billiau, A., et al., Human interferon: mass production in a newly established cell line, MG-63. Antimicrob Agents Chemother, 1977. 12(1): p. 11-5.
[7] Rodan, G.A., et al., Diversity of osteoblastic phenotype. Ciba Found Symp, 1988. 136: p. 78-91.
[8] Houston, B., et al., PHOSPHO1 - A novel phosphatase specifically expressed at sites of mineralisation in bone and cartilage. Bone, 2004. 34(4): p. 629-37.
[9] Moffatt P., et al., Bril. A novel bone-specific modulator of mineralization, J Bone Miner Res, 2008. 23:1497-508
[10] Bonewald, L.F., The amazing osteocyte. J Bone Miner Res, 2011. 26(2): p. 229-38.
Beginning of non-public section
hCAGE sequencing
Quality control of the hCAGE data was performed at RIKEN by Erik Arner. In short, data was RLE normalised and the subjected to hierarchial clustering and principal component analysis to check that sample clustered according to differentiation time points. K-means clustering (k=10) was also performed to detect any "abnormal" clusters. A few replicates were filtered away after this before continueing further with analysis. Some osteogenic marker genes were also checked to confirm an expected expression pattern during differentiation.
Related samples
Osteosarcoma cell lines 143B/TK^(-)neo^(R) and HS-Os-1 from RIKEN BRC - contact Al Forrest
Additional untreated timecourse is stored
Additional timecourse using MG63 is stored
Mesenchymal stem cells differentiating into adipocytes or osteocytes time course
Retinal pigment cell EMT time course
Results
Zenbu configuration and status
Saos-2 osteosarcoma treated with ascorbic acid and BGP
Saos-2 osteosarcoma timecourse human, October 26th 2012
Expression profiles
Gene and CAGE cluster expression for the Saos-2 series
https://fantom5-collaboration.gsc.riken.jp/webdav/home/arner/timecourse/time_course_main_paper_freeze_feb2013/qc_release_130226/human_Saos-2_osteosarcoma/expression_tables/
MARA based network results
Self organising maps
Switch Engine
TSS dynamics plots can be found here:Saos-2 osteosarcoma_bgp_ascorbic_acid.plots.pdf (it is recommended to download these files instead of viewing in browser as they are large in size)
These figures only show those genes for which Switch Engine has detected TSS switching to occur. For now, only the first and last time points are being compared to define a switch.
One gene per page, RefSeq and Gene Symbol identifiers on top. Each panel is a separate TSS associated with the gene. TPM expression on y-axis and time on x-axis. Colored points correspond to specific expression, with colors referring to different replicates (key on top). Replicates with a * next to the name are those that have missing data from the currently available release of the DPI clustered and normalized TPM matrix. Such missing data were replaced by imputed values. Black lines are the mean trajectories across replicates. Purple vertical bars correspond to the time points being compared for switch definition. Magnitude of switch is expressed in each panel under TSS identifier with approximate 95% confidence interval in brackets. Note: some confidence intervals may not be accurate or may be missing due to the small sample size. TSS identifiers colored GREEN show an increasing TSS, those colored RED show decreasing TSS, WHITE show no statistically significant change.
TSS switch defined as the simultaneous presence of one or more increasing, together with one or more decreasing TSSs per gene. At least one increase and at least one decrease has to be statistically significant (alpha approximately 5%). At least one of the two time points being compared has to be greater than 5 tpm. At least 2 replicates have to agree to call an increasing/decreasing trajectory. TSSs have to be at least 250 base pairs away from one another to call a switch.
A summary file can be found here: Saos2osteosarcoma_bgp_ascorbic_acid.summary.xls
The file contains the following columns: RefSeq ID, Gene Symbol, TSS ID, X{value}: mean tpm expression at {value} time point, Delta: change in tpm expression from first to last time point, FC: fold change in tpm expression from first to last time point, DeltaSum: net change in tpm expression for this gene (summed across all TSS in this gene), Balance: ratio of increasing to decreasing tpm expression TSSs per gene (e.g. a balance value of 1 means that the total increase in tpm expression in increasing TSSs is equal to the total decrease in tpm expression in decreasing TSSs), Chromosome, Strand, Start position of TSS cluster, Stop position of TSS cluster, DBTSS: distance between TSSs relative to NA value.
A list of genes in which TSS switching occurs can be found here: Saos2osteosarcoma_bgp_ascorbic_acid.genelist.refseq.doc and Saos2osteosarcoma_bgp_ascorbic_acid.genelist.symbol.doc
BED file with regions corresponding to TSS clusters involved in TSS switching can be found here: Saos2osteosarcoma_bgp_ascorbic_acid.switchinglist.bed.xls (please remove .xls extension after downloading)
Short RNA
ISMARAanalysis results
All samples: http://ismara.unibas.ch/timecourses/saos/ismara_report/index.html
Replicate averaged: http://ismara.unibas.ch/timecourses/Saos2-avgd/averaged_report/index.html
For more information, see ISMARA.
Paper outline
Tentative title:
Transcriptional regulation during mineralisation of Saos-2 cells, a model for human bone mineralisation
Margaret R Davis., RIKEN OSC collaborators, bioinformatics collaborators, Kim M Summers
Introduction
Bone formation, mineralisation
Osteosarcoma model for bone formation
What’s known so far about gene expression changes
Guilt by association – finding new candidate for bone diseases
Why do early points (refer to main umbrella paper)
CAGE etc (refer to phase 1 main paper)
Methods
Culture of Osteosarcoma cell lines
RNA extraction and CAGE (phase 1 paper)
qPCR of control sets
Bioinformatics
Results
Validation of the calcification
• Figure – alizarin red levels, phase contract images, alizarin red quantification
Detection of differentially expressed genes
• Figure 1a – Biolayout graph of all genes image
Genes associated with early responses (elaboration of umbrella paper)
• Figure 1b – images for key early genes, highlighting how we can identify dimerisation partners and also the upregulation of some at the late time points
Genes associated with late responses, discussion of how these are mainly bone specific
• Figure 1c – images for late genes
Genes that go down early and don’t come up again (or do very late on), discussion of how these tend to be (?) general connective tissue genes
• Figure 1d – image for down regulated genes
Genes that peak after early genes (4 hrs to 4 days)
• Figure 1e – images for later peaking genes
Expression in the control data sets
• Figure or table – qPCR for selected time points
Transcription factors/motifs that change during the time course (MARA/Michael analysis)
Changes in enhancers (Albin analysis)
Promoter switching within time course and compared with other celltypes – overall and some examples
• Figure – relevant genes that switch in the time course
Short RNAs – what do they add to the story?
• Figure – relevant short RNAs
Comparison between SAOS-2 and osteoblasts (mouse/human) in FANTOM5
• Figure – Biolayout of time courses with Phase 1 samples
• Figure - TF profiles of SAOS and osteoblasts (Phase 1 and wiki)
• Figure – enhancer profiles of SAOS and relevant phase 1 samples – maybe all time course then mature osteoblasts (do we have them?) in same image
• Figure – relevant genes that switch between bone and others (including Phase 1 data)
Discussion
1. transcriptional changes during mineralisation; use of the control data sets
2. regulation of these changes
3. SAOS-2 as a model for bone formation in vivo
4. new candidates for bone formation and diseases
Analysis required
Biolayout analysis of expression patterns (Kim will do this)
TF/novel motif analysis for time course and relevant Phase 1 samples - have spoken to Michael Rehli about this
Small RNA analysis for time course and relevant Phase 1 samples - need to work out how/who to do this
Enhancer analysis for time course and relevant Phase 1 samples - have spoken to Albin about this
Promoter switching analysis for time course and relevant Phase 1 samples - have spoken to Emmanuel about this
