Radiation induced osteosarcoma is initiated by disruption of the genetic programme controlling bone cell development. In conjunction with other participants in the programme the molecular and cellular basis of osteosarcomagenesis at low doses or radiation is being investigated.
Long term in vivo analysis of dose dependency.
A series of experiments have been concluded that were designed to determine the oncogenic risk following incorporation of a short lived bone seeking radionuclide (thorium-227) in the conjunction with a background exposure to low levels of a long lived isotope.
Low level long term exposure was modelled by injecting 3 month old female NMRI mice with 1.85 kBq/kg of actinium-227 (mother nuclide of thorium-227), giving a mean skeletal dose of 100 cGy per 600 days. At the age of 4 or 12 months cohorts received an additional treatment with a relatively low activity of the short lived alpha emitter thorium-227 (mean skeletal alpha dose burden of 50 cGy and 200 cGy at doses of 9.25 and 37 kBq/kg thorium-227 respectively).
The rates of osteosarcoma development under the different experimental conditions were determined. In the lower dose range (thorium-227 9.25 kBq/kg or actinium-227 plus thorium-227 9.25 kBq/kg) a similar (ie not significantly different) osteosarcoma rate was observed in animals treated at either 4 or 12 months of age. The older age of the animals presumably compensates for the later incorporation of the thorium-227. At the higher dose (thorium-227 37 kBq/kg or actinium-227 plus thorium-227 37 kBq/kg) the osteosarcoma rate in the older age groups was significantly lower than in those groups receiving thorium-227 at a young age. Despite this fact, low level background irradiation by actinium-227 seems to unmask the promoting efficiency of the older age. An increase of the osteosarcoma risk by incorporation of the additional activity of thorium-227 in animals with a low activity background burden of actinium-227 was significant only f or the groups older at the time of thorium-227 incorporation and is seen at both levels of thorium-227 activity.
Analysis of the early molecular events in osteosarcomagenesis.
Whilst the entire genome is subjected to damage from irradiation, alteration in the function of particular sets of genes has far reaching consequences for carcinogenesis. These include the protooncogene and tumour suppressor gene groupings, where a series of unrelated genes encode proteins regulating cell proliferation and differentiation. Disruption of the normal function of these genes by external noxae, including irradiation, can predispose the affected cell to tumorigenicity. We have chosen initially to analyze the 2 best characterized tumour suppressor loci, encoding the p53 and RB1 (retinoblastoma) proteins. These genes are subjected to loss of function mutation in many, if not all, induced and inherited tumours, and thus are prime candiadates for early molecular targets of oesteosarcoma inducing radiation damage.
To analyze these alterations in tumour suppressor genes in osteoarcoma a number of cell lines were established from radiation induced mouse osteosarcomas in collaboration with MZP-Neuherberg. Deoxyribonucleic acid and ribonucleic acid (RNA) were extracted, and the structural integrity of the p53 tumour suppressor gene locus analyzed by genomic Southern blotting. DNA was digested with several different restriction enzymes and the restriction fragments analyzed by hybridization with radiolabelled p53 complementary DNA (cDNA). Digestion with the restriction enzymes EcoRI, BamHI and TaqI revealed that in 5 of the 9 cell lines gross rearrangements of the p53 gene could be demonstrated. In each of these cases the normal allele was missing, such a loss of heterozygosity removes the unmutated allele and results in a characteristic total loss of function at the tumour suppressor locus.
The sensitivity of the Southern blotting analysis is restricted to detection of major rearrangem ents in the p53 locus. To determine if in other tumours there have been minor changes to the DNA we have chosen to analyze directly the p53 cDNA present in each tumour.
Tumour cell messanger RNA (mRNA) was converted into cDNA by reverse transcription, creating a DNA template for polymerase chain reaction (PCR) directed amplification. This is necessary as p53 sequences normally represent only a very small portion of cellular RNA. The complete translated region of the cDNA (comprising exon 2-11), as well as smaller regions containing selected exons were amplified, and the length of the product compared to wild type p53 amplificates using high resolution polyacrylamide gel electrophoresis developed in participation with the project of Technische Universitaet Munich. This analysis allows fine mapping of alterations in the DNA sequence and has a resolving power sufficient for the detection of length alterations as small as 1 base. If the target exon is missing from the cDNA there will be no product when that particular exon is amplified, and the product of the entire coding region will be smaller. This analysis indicated that, in addition to gross alterations in the geonomic organization, there have been a series of small deletions within the p53 gene in radiation induced osteosarcoma. These can be summarized as follows: in the cell line OS 46 a mutation has deleted parts of exon 8 and 9; in OS 47 exon 7 has been lost; and in OS 51 exon 5 is missing.
Whilst the 2 techniques described above can identify insertion or deletion of bases in the p53 gene, mutations that do not affect the length of the cDNA are undetectable. These include point mutations changing the sequence of the cDNA. It has been established that detectable levels of p53 protein in the cell are frequently associated with point mutation in the coding sequence. The alteration of a single amino acid residue results in accumulation of p53 protein to a detectable level. The expression of p53 has been analyzed immunohistochemical methods. Cells were stained with an anti-p53 specific monoclonal antibody (PAb 421) and an fluorescence isothiocyanate bound second antibody and visualized by fluorescence microscopy. p53 immunostaining was detected only in the cell lines OS 43 and OS 56. These 2 lines did not shown genomic alterations by either Southern blotting or PCR analysis. These are potential candidates for loss of function point mutations, and the entire coding region will be sequenced to determine the location of any point mutations in these tumours.
Osteosarcoma induction in man following irradiation has generated dose response and risk assessment data for radiation induced carcinogenesis. Using a murine model of alpha emitter induced oseteosarcoma in the mouse our collaboration with Pathology Neuherberg has implicated alterations in the tumour suppressor locus encoding p53. Our contribution to the programme is to compare the data obtained using the murine model with that in human osteosarcoma. As human osteosarcoma tissue is very scarce methods to establish the molecular biological analysis were developed using nonosteosarcoma tissue.
Loss of function mutations to the reinoblastoma gene (RB1) and p53 tumour suppressor gene has been implicated already in the development of osteosarcoma. Southern blot analysis is effective in detecting large delections or rearrangements to these genes. However, the pathogenetic mechanism of tumour suppressor gene inactivation dose not require large deletions in the genes, inactivating point mutations may also result in loss of the locus. In the p53 gene we have already established that polymerase chain reaction (PCR) amplication is adequate for the detection and characterizationof small gene alterations. In the case of the RB1 locus tremendous difficulties arise with this methodology because of the size of the gene (some 200 kilobases (kb) and 27 exons). However, be amplifying and sequencing numerous small regions of the gene it may be possible to detect point mutations with a high degree of sensitivity. The simultaneous amplification of different exons is necessary to contain costs for the programme, and we have developed the methodology for this by using formalin fixed, paraffin embedded human retinoblastoma tissue as a test target. This choice of material was made as the frequency of RB1 mutation will approach 100%. To date reaction conditions for analysis of 26% of the coding region of the RB1 gene have been established. Using this technique 3 different mutations have bee n identified and sequenced from the 24 cases tested so far.
Human osteosarcoma material from 11 cases has been collected. From 5 of these ribonucleic acid and deoxyribonucleic acid has been extracted for use at a later date in the PCR-based analytical process.
This study concerns:
characterization of cooperating effects of alpha-irradiation and retrovirus infection on cell proliferation in mandibular condyles;
pathogenicity of treated tissues and cell lines established therefrom;
expression of transcription factors and growth control genes in treated tissues.
Endogenous retroviruses are activated both in the early latent period of radiation induced osteosarcomas and in radiation induced tumours. In vitro and in vivo studies have shown that activated endogenous retroviruses induce significant effects on skeletal cells of the mouse, including induction of osteogenic differentiation and new bone formation, as well as induction of benign and malignant bone tumours in infected newborn mice.
In order to study a cooperative effect between alpha-irradiation (radium-224 [Ra-224]) and retroviral infection on neoplastic transformation of skeletal cells and tissue, mandibular condyles of newborn mice were irradiated with doses ranging between 0.007 kBq/ml and 7.4 kBq/ml Ra-224 for 3.5 7, 10.5 14, and 21 days in the presence or absence of infectious retinofibroblast murine leukaemia virus (RFB MLV) retrovirus. It has previously been shown that continuous irradiation of mice in this dose range, over a period of 36 weeks, induces osteosarcomas in 10 to 95% of animals. Control tissues were either infected with RFB MLV or untreated. After the end of the treatment period the tissues were transplanted into syngeneic mice (1 to 4 condyles/mouse. Within an observation period 3 months after transplantation no tumour development was observed, suggesting that under these conditions either the dose or the irradiation period are not sufficient to induce malignant transformation or mandibular condyles.
To study the effect of irradiation and retrovirus infection on cell growth, the cell number in manidibular condyles was determined after various treatment: irradation (7 days) was followed by RFB MLV infection (14 days) and vice versa; controls were either irradiated (21 days) or infected with RFB MLV. The highest cell number was observed in tissues which were irradiated for 7 days and infected thereafter with RFB MLV for 14 days. These data suggest that irradiation increases the susceptibility of skeletal cells to the proliferation inducing effects of RFB MLV.
Irradiated and/or retrovirus infected mandibular condyles were dissociated by collagenase and single cell suspensions were used to establish monolayer cell lines. The highest cell proliferation was found in cell lines established from tissue which had been irradiated and infected, indicating a cooperative effect of alpha irradiation and RFB MLV infection on growth of skeletal cells.
The c-fos protooncogene plays a significant role in osteogenic differentiation, in transdifferentiation of cartilage cells to bone cells, and in terminal differentiation processes of bone cells in osteosarcomas. The v-fos oncogene, on the other hand, is a poetent inducer of osteosarcomas. Further, in vivo studies with c-fos transgenic mice have indicated that endogenous retroviruses can cooperate with cellular oncogenes in bone tumour development. Northern blot analysis of irradiated mandibular condyles showed a high transient expression of the c-fos protooncogene between 30 min and 2 hours after start of the culture and down regulation of c-fos after 6 hours, similar to that observed in nonirradiated control condyles. In contrast to published reports on fibrobalst cell lines using toxic doses of gamma irradiation, increased expression of c-jun orc-myc was not found in Ra-224 irradiated condyles. These data suggest that the observed stimulation of cell proliferation is not a result of enhanced transcriptional activity of the growth control genes c-fos, c-jun, or c-myc.
Irradiated condyles, however, showed an enhanced ribonucleic acid (RNA) level of the T1 gene after treatment with Ra-223. T1 is a gene stimulated by the oncogene Ha-ras and was isolated f rom Ha-ras transformed NIH3T3 cells. T1 expression was analyzed by reverse transcription of total cellular RNA from irradiated condyles and sequence-specific amplification by polymerase chain reaction (PCR) of a 400 base pair sequence. In contrast to nonirradiated tissues, irradiated tissues showed expression of T1 and 2 hours after start of the culture, suggesting that T1 is activated by Ra-223. Further in situ hybridization analysis will be carried out to study the cell specific expression of genes, which are sensitive to Ra-223 in irradiated mandibular condyles.
During the reporting period Ra-224 became no longer commercially available. As an alternative, Ra-223 was obtained from the GSF-Institute of Pathology, Neuherberg, and subsequent experiments were carried out with Ra-223.
Endogenous retroviruses are frequently activated in association with radiation induced osteosarcomagenesis in laboratory strains of mice. This work addresses the following questions of interest for understanding of immediate and late effects of ionizing radiation:
How are endogenous proviruses activated?
Are the activated viruses pathogenic?
What features in a retroviral genome may affect its pathogenicity for bone tissues?
What molecular mechanisms may contribute to the effect of a retrovirus on bone cells?
All viruses included in the studies are available as molecular clones and all are closely related to the prototype endogenous ecotropic murine leukemia virus, Akv. The viruses induce lymphomas, osteopetrosis, and osteomas in NMRI mice, but differ in potency of disease induction. In CBA ice, on the other hand, the virus effects are frequent osteomas and only rare lymphomas. Interesting differences in organization of the tandem repeat sequences in U3 of the long terminal repeat (LTR) are found among the viruses. Studies in a number of cell lines show that the LTR differences affect the potency and specificity of virus directedtranscription. These expression assays reveal that the most potent bone pathogenic isolate in our collection retinofibroblast murine leukaemia virus (RFB MLV), may represent a novel transcriptional phenotype. In a study of the T-lymphoma inducing murine leukemia virus SL3-3, results revealed single nucleotide positions in the U3 region of the viral genome that are crucial both for expression in cultured T-lymphoma cells and for lymphomagenicity in the animals, thus directly linking in vivo and in vitro results.
The protooncogene c-fos may be activated as part of the stress response that is an immediate consequence of irradiation c-fos also plays a role in normal osteogenic differentiation, and potential interactions between this protooncogene and a virus are therefore of major interest for understanding the role of activated viruses in the effects of radiation on bone tissues. Transient expression studies were performed which indicated that the c-fos product can specifically trans-regulate expression from virus LTRs in an osteogenic cell line. The effect is stimulatory at lower levels of c-fos expression and inhibitory at higher levels.
The studies of proviral integration sites in DNA from virus induced and radiation induced tumours use both Southern hybridization techniques and polymerase chain reaction techniques, however no clear pattern of immediate biological significance has emerged during the reporting period.
For radiation protection purposes it is important to identify cells at risk for tumour developent after incorporation of low doses of bone seeking radionuclides. This may help to estimate osteosarcomogenic doses and allows tracing radation induced events which may be related to subsequent development of bone cancers.
This work focused on the identification of precursor cells for bone cells among bone marrow and assessment of their radiosensitivity; and the development of an efficient model for osteosarcomas in Balb/c mice. In these animals osteogenic cell populations will be studied during the latency period for bone tumour development.
In the conventional osteogenic culture assay for bone marrow cells, bone marrow was flushed as an organ fragment from the femur diaphysis of adult mice and incubated in vitro on a collagen sponge. After a lag time of 13 days mineralization could be followed quantitatively via measurement of strontium-85 uptake which had been added to the tissue culture medium (strontium-85 is used as a tracer for calcium).
To select further which cell populations among the heterogeneous marrow cells are involved in osteogenic differentiation, stromal marrow cells were selectively cultured via the long term bone marrow culture (LTBC) technique where the stromal cells form an adherent layer. Subsequently these LTBC were irradiated with 10 Gy of X-rays to kill haemopoietic stem cells residing in the adherent layer.
Involvement of proliferating cells in osteogenic differentiation was determined in marrow fragment cultures after in vitro incorporation of tritiated thymidine deoxy ribose (H-3 Tdr) with high specific activity (ie killing of proliferating cells). The osteogenic capacity of these various cultures was assessed by measuring the mineralization rate via uptake of strontium-85.
Similar to the intact marrow fragments, the stromal cells cultured from bone marrow according to the long term bone marrow culture technique, displayed osteogenic di fferentiation characteristics in vitro. After the stromal cells reached confluency, the cells were scraped from the bottom of the tissue culture flasks and loaded on a collagen matrix. The prerequisites to get calcification were a high cell density and 3-dimensional configuaration. The stromal cells were still an heterogeneous cell population of mainly adherent fibroblast-like cells, macrophages and haemopoietic stem cells which homed in the adherent layer. Killing the haemopoietic stem cells by irradiation of the confluent LTBC with 10 Gy of X-rays did not impair osteogenic differentiation after the irradiated cells were brought in 3-dimensional configuration. Haemopoietic stem cells were not further needed for the osteogenic differentiation.
Next to differentiation, proliferation of cells was also found to be needed for bone differentiation of marrow cells. This has implications for the sensitivity to irradiation and suggests a possible regulatory role of cell growth factors. Incorporation of H-3 Tdr in deoxyribonucleic acid (DNA) synthesizing cells of marrow plug cultures showed that the osteogenic capacity was more reduced if H-3 Tdr was added between day 6 and 9 of in vitro incubation than between day 3 and day 6. Cell proliferation related to osteogenic differentiation was lowest during the first 3 days of culture.
Bone marrow was irradiated by X-rays (1, 2, 4, 8, 16 Gy), either in vivo or in vitro prior to onset of the cultures. Femoral marrow was then incubated in vitro as an organ fragment and strontium-85 uptake was measured to evaluate the osteogenic capacity.
Male Balb/c mice were injected intravenously with americium-241. The doses ranged between 40 to 500 Bq/g mouse, noncontaminated mice were included as controls. Cultures were initiated 1 day, 1 week and several weeks (from 4 to 14) after americium-241 injection. Marrow cells were cultured either as intact fragments (plugs) or the stroma was cultured as an adherent layer and subsequently, when l oaded on a sponge, tested for its osteogenic capacity (strontium-85 uptake).
In vivo and in vitro irradiation of adult femoral bone marrow was found to damage the bone formation capacity in vitro. After doses higher than 4 Gy the mineralization capacity of marrow plugs was significantly reduced as shown by the measurement of strontium-85 uptake.
4 weeks after americium-241 injection of adult mice with doses higher than 150 Bq/mouse, the marrow showed a reduced mineralization capacity when cultured as a fragment. This contamination corresponds with a skeletal dose rate of 25 mGy/day and an accumulated skeletal dose of 0.75 Gy.
If stromal cells were cultured from the bone marrow and subsequently brought in 3-dimensional conditions allowing osteogenic differentiation, a reduced mineralization capacity in vitro was noticed 4 weeks and 13 weeks after americium-241 injection and this from the lowest dose tested (40 Bq/g mouse, skeletal dose rate: 7 mGy/day, accumulated dose after 1 month: 0.2 Gy).
The osteogenic capacity of marrow in vitro showed a high radioresistance for acute X-irradiation. In contrast, relatively low doses of alpha-irradiation from bone seeking radionuclides had effects on the osteogenic differentiation of marrow cells in vitro.
More than 2 years ago, male and female mice were injected, at 10 weeks of age, with americium-241 doses ranging between 45 and 250 Bq/g. All mice have now died and were autopsied, samples were taken for histopathology and carcasses were radiographed for diagnoses of bone tumours (preliminary results have been obtained).
43% of female Balb/c mice injected with 124 Bq/g developed bone tumours, while only 23% of male mice developed bone tumours after injection of 103 Bq/g. At lower doses of 45 and 90 Bq/g respectively, 71% and 64% of female mice developed bone tumours. These diagnoses are obtained from X-ray pictures and need further histopathological confirmation, but it is obvious that in female Balb/c mice a single inj ection with a relatively low dose of americium-241 induces a substantial proportion of bone tumours.
The oncogenic transforming principle in irradiated BALB/c mouse deoxyribonucleic acid (DNA) has proven to be homologous to the transregulator (TR) gene of the mouse mammary tumour virus (MMTV). Thus a baculovirus expression vector was constructed to allow the large scale production of the TR gene product.
A baculovirus expression vector containing the MMTV TR gene was constructed as follows:
Firstly, a plasmid was created placing the retroviral transactivator gene under the control of the baculovirus polyhedrin promotor. This plasmid was electroporated into host Sf9 insect cells together with DNA containing the wild type baculovirus Autographa californica multiple nuclear polyhydrosis virus. By means of homologous recombination the MMTV TR gene was introduced into the baculovirus vector within the polyhedrin gene. By means of repeated limiting dilution the recombinant virus was purified to homogeneity. Upon infection of large scale Sf9 cultures with the recombinant virus a 36 kD protein was seen to be present in the nuclei of the infected cells. This material is presumed to represent the TR protein produced by the vector. In NIH/3T3 cells transformed by retroviral transactivator genes (from the MMTV the bovine leukemia virus or the human immunodeficiency virus) multiple copies of these genes proved to be integrated into chromosomal DNA. Expression of these sequences was not detectable, however, at either the ribonucleic acid (RNA) or protein level. Prokaryotic genes, which were present in the plasmid vectros carrying the retroviral TR genes, also proved to be integrated along with the TR genes, but were also not expressed. This lack of active expression is possibly due to hypermethylation of the integrated sequences. Presumably, oncogenic transformation by these TR genes is caused by a hit and step aside phenomenon.
The human radiumAinduced tumours, resulting from former industrial occupation as dial painters with 226Ra and medical treatment with 224Ra, continue to provide basic information on the effects of internal irradiation and radiocarcinogenesis in man. Animal experimentation and in vitro systems help to elucidate the problems of radiation hazards, and in addition increase our understanding of the mechanisms of radiation carcinogenesis. Bone tumours and tumours of the haematopoietic system have been induced experimentally by internal administration of many boneAseeking radioactive isotopes. For several short lived radionuclides we have the basic knowledge of dose dependence and protraction effects in radiationAinduced osteosarcomogenesis and leukaemogenesis. The additional induction of leukaemia, particularly at lower dose levels, raises the question of the target cell in alphaAradiation oncogenesis. This is a further indication of the importance of extending studies of early events in haematopoietic and stroma boneAmarrow cells and osteogenic cells. Data from the epidemiological study on 224Ra patients also indicate the need for systematic investigations of the leukaemia risk after the incorporation of boneAseeking radionuclides. In addition experimental studies using 224Ra offer the possibility of estimating the plutonium risk in humans. In studies of different age periods the role of possible sensitive periods during the whole lifespan should be explored. In this proposal considerable emphasis is placed on molecular biological approaches to radiation carcinogenesis, both in appropriate animal models, in vitro systems and human tumours.
Studies with oncogenic retroviruses have revealed new aspects of the molecular basis of carcinogenesis. Several groups of genes have been detected which are involved in the development of tumours. These genes are normal constituents of the genetic makeup of the cell, but in carcinogenesis they seem to be overAexpressed, mutated or combined with other genes. A more detailed molecular analysis of target genes for radiation carcinogenesis is envisaged. The research projects in this proposal are concerned with the general problem of radiation risk, especially with respect to radiationAinduced cancer. The quantitative analysis of doseAeffect relationships with special emphasis on time and quality factors, and the development and analysis of appropriate animal experiments and in vitro systems with the possibility of studying mechanisms of oncogenesis, will have immediate implications for improving the establishment of dose limits.
Inclusion of the whole lifespan in the study of risk takes the age pattern of a real population into account. The determination of the genes and gene products responsible for radiation carcinogenesis would offer the possibility of defining the critical radiation doses which induce or activate events in the cellular genome subsequently leading to cell transformation. This would contribute to the detection of individuals at risk of developing cancer, and to the establishment of new strategies of radiation protection. This would include abolishing those early effects of radiation on cells which are the first steps in tumour development, or interacting with the mechanisms involved in promotion and progression of radiationAinduced malignancy. The work will be of great relevance to various groups within the CEC concerned with assessing the risk to human health following incorporation of boneAseeking radionuclides. The data will also be useful to the ICRP and various national bodies concerned with radionuclide to icity. Coordination within EULEP is planned.
Funding SchemeCSC - Cost-sharing contracts
8000 Aarhus C