HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by MARKO, N. F. Articles by HU, V. W. Search for Related Content PubMed PubMed Citation Articles by MARKO, N. F. Articles by HU, V. W.

Does metabolic radiolabeling stimulate the stress response? Gene expression profiling reveals differential cellular responses to internal beta vs. external gamma radiation

NICHOLAS F. MARKO^*,1, PAUL B. DIEFFENBACH^*,1, GAI YAN^*, SUSAN CERYAK, ROGER W. HOWELL, TIMOTHY A. MCCAFFREY^* and VALERIE W. HU^*,2

Departments of
^* Biochemistry and Molecular Biology and
Pharmacology The George Washington University School of Medicine, Washington, DC, USA; and
Department of Radiology, UMDNJ-New Jersey Medical School, Newark, New Jersey, USA

²Correspondence: Department of Biochemistry and Molecular Biology, The George Washington University Medical Center, 2300 Eye St., N.W., Washington, DC 20037, USA. E-mail: bcmvwh{at}gwumc.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
DNA microarray analyses were used to investigate the effect of cell-incorporated ³⁵S-methionine on human colorectal carcinoma cells. This ß-radiation-induced gene expression profile was compared with that induced by external -radiation. The extent of DNA fragmentation was used as a biomarker to determine the external dose that was bioequivalent to that received by cells incubated in medium containing ³⁵S-methionine. Studies showed that ³⁵S-methionine at 100 µCi/mL induced a much more robust transcriptional response than -radiation (2000 cGy) when evaluated 2 h after the labeling or irradiation period. The cellular response to internal ß-radiation was greater not only with respect to the number of genes induced, but also with respect to the level of gene induction. Not surprisingly, the induced genes overlapped with the set of -responsive genes. However, a distinct ß-gene induction profile that included a large number of cell adhesion proteins was also observed. Taken together, these studies demonstrate that metabolic incorporation of a low energy ß-emitter, such as ³⁵S-methionine, can globally influence a diverse set of cellular activities that can, in turn, affect the outcome of many experiments by altering the cell cycle, metabolic, signaling, or redox status (set point) of the cell. Additional studies of the mechanism of ß-induced proliferation arrest and cell death and of the significance of its differential gene induction/repression profile in comparison to pulsed -irradiation may lead to new insights into the ways in which ionizing radiation can interact with cells.—Marko, N. F., Dieffenbach, P. B., Yan, G., Ceryak, S., Howell, R. W., McCaffrey, T. A., Hu, V. W. Does metabolic radiolabeling stimulate the stress response? Gene expression profiling reveals differential cellular responses to internal beta vs. external gamma radiation.

Key Words: DNA microarray • DNA damage • ionizing radiation • radioisotopes • sulfur-35 methionine

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
STUDIES OF CELLULAR responses to ionizing radiation have typically used acute external beams of photons (i.e., X-rays or -rays) because of their widespread use in diagnostic and therapeutic radiology and their straightforward dosimetry and safety in processing experimental samples postirradiation (1 2 3 4 5 6 7 8 9 10) . On the other hand, although the long-term cytotoxicity of cell-incorporated low energy ß-emitters has long been established by radiobiological studies (11 , 12) , the immediate cellular effects and mechanism of action of intracellular ionizing radiation from such compounds have not been well characterized, in part because of the complexity in determining actual cellular and nuclear absorbed doses in cultured cells. Yet cells are often exposed to such internal radiation in a wide range of experimental protocols for tracking a variety of biological processes. Metabolic labeling with ß-emitting radioisotopes such as ³⁵S-methionine has been widely used as a standard technique for studying the biosynthesis, trafficking, and degradation of proteins in vitro and in vivo, whereas ³H-thymidine has been used to follow DNA synthesis and cell proliferation (13) . For these applications, it has often been assumed that the low level of ionizing radiation emitted by ³⁵S- and ³H-labeled compounds affords the sensitivity of radioactive tracers without producing cellular damage. This assumption is based on observations that cells apparently remain viable in culture (e.g., by trypan blue dye exclusion assays) after pulse-chase labeling protocols.

The biological impact of tritium and other radionuclide labeled organic compounds localized within the cell has been a subject of interest for more than 40 years. Early studies focused primarily on biological end points such as survival, mutation, chromosomal aberrations, and transformation. The very early studies were summarized by Halpern and Stöcklin (14) and the National Council on Radiation Protection (12) . Interest in the effects of incorporated ß-emitters continued over the ensuing years (15 , 16) . Technological advances in addition to increased understanding of cellular stress and apoptosis at the molecular level have recently sparked an interest in elucidating the explicit molecular mechanisms that lead to cellular damage from organic compounds radiolabeled with ß-particle emitters. In the early to mid-1990s, Solary et al. (17) observed DNA fragmentation after radiolabeling with ³H-thymidine, and two groups independently described up-regulation of p53 and growth arrest in ³⁵S-methionine- and ³H-thymidine-labeled cells (18 , 19) . Similarly, ³²P-orthophosphate and ³²P-phosphorothioate-modified oligonucleotide respectively were found to cause cell cycle arrest of human fibroblasts and reduced proliferation and migration of vascular wall cells (20 , 21) . More recently, our laboratory reported that metabolic incorporation of ³⁵S-methionine induces cell cycle arrest, prolonged inhibition of proliferation, and apoptosis in rabbit smooth muscle cells (22) , whereas Yanokura et al. reported cell cycle arrest and apoptosis after ³H-thymidine incorporation into hematopoietic cells (23) . We later demonstrated that the cell-incorporated ³⁵S-methionine induced p53-independent DNA fragmentation even under hypoxic conditions and that reactive oxygen species (ROS) were also induced in radiolabeled cells, but in a p53-dependent manner (24) . These findings led us to suggest that multiple signaling pathways might be affected in radiolabeled cells, either in direct response to DNA damage or possibly as a result of ROS-induced stress signaling pathways.

Given these results, we hypothesized that multiple stress response pathways are induced by cell-incorporated low energy ß-emitters. The goals of this study were to more thoroughly characterize the molecular and cellular response to internal ß-radiation from incorporated ³⁵S-methionine, to compare this response with that induced within the same cells by a standard radiobiological reference radiation (-rays) that produced comparable initial DNA damage, and to illuminate stress response pathways that are affected by cell-incorporated radioactivity.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
Chemicals and reagents
³⁵S-Trans label (70% ³⁵S-Met and 30% ³⁵S-Cys; spec. act. >1000 Ci/mmol) and its nonradioactive carrier buffer were obtained from ICN Radiochemicals (Irvine, CA, USA). All media, cell culture reagents and materials, and buffers were purchased from Fisher Scientific (Pittsburgh, PA, USA) and FBS was from Hyclone (Atlanta, GA, USA). Unlabeled L-methionine and L-cysteine were from Sigma (St. Louis, MO, USA). Antibodies to human p21, p53, cdc2, FasL, cIAP1, and JNK2 were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA) and antibodies to human caspase 8, Bax, Bcl-X_L were obtained from Trevigen, Inc. (Gaithersburg, MD, USA). Antibody to cJun-P-Ser63 was from Cell Signaling Technology (Beverly, MA, USA) and antibody to 14-3-3 was the generous gift of Dr. Bert Vogelstein (Johns Hopkins University School of Medicine, Baltimore, MD, USA). HRP-conjugated secondary antibodies to rabbit and mouse IgG were obtained from Zymed Laboratories, Inc. (South San Francisco, CA, USA). Atlas Human 1.2 DNA Microarrays and hybridization solutions were obtained from Clontech (Palo Alto, CA, USA). For RNA isolation and purification, RNAzol B was purchased from Tel-Test, Inc. (Friendswood, TX, USA) and a DNA-free^TM Kit (containing DNaseI and appropriate buffers), TE (10 mM Tris-HCL, 1 mM EDTA, pH 8.0), and TAE (40 mM Tris-acetate, 1 mM EDTA, pH 8.0) buffers were purchased from Ambion, Inc. (Austin, TX, USA). The CometAssay^TM Kit was obtained from Trevigen, Inc. and the BCA Protein Assay Kit from Pierce Biotechnology, Inc. (Rockford, IL, USA).

Cell model and culture
Human colorectal carcinoma cells (HCT-116) were kindly provided by Dr. Bert Vogelstein. The cells, which were homozygous for wild-type p53, were cultured in McCoy’s 5A medium supplemented with 10% FBS, 2 mM L-glutamine, and antibiotic/antimycotic solution (final concentration of 100 U/mL penicillin, 100 µg/mL streptomycin, and 0.25 µg/mL amphotericin B). The cells were grown in a monolayer and maintained at 37°C with 5% CO₂.

Radiolabeling
Cells were preincubated for 5 min at 37°C in methionine/cystine-free Dulbecco’s minimum essential medium supplemented with 10% dialyzed FBS, 2 mM L-glutamine, and 1 µg/mL L-cysteine (referred to here as deficient medium) before addition of radiolabel. ³⁵S-methionine at the appropriate dose was added to the medium and the cells were incubated for 2 h at 37°C. The media was then removed and the cells were washed twice in sterile PBS. Cells were either processed immediately for Western and Comet analyses or for RNA isolation or "chased" for an additional 2 h (or more for Western and Comet analyses) in complete medium at 37°C. Controls for all experiments were run in parallel with the experimental samples and followed identical protocols except that the radiolabel was replaced by an equivalent volume of carrier buffer obtained from ICN and contained 1 µg/mL unlabeled L-methionine.

Determination of cellular and nuclear radioactivity
Cellular uptake of radiolabel was determined after harvesting of cells in trypsin/EDTA solution containing 1 µg/mL unlabeled methionine to inhibit efflux of radiolabeled methionine from cells. The cells were pelleted, resuspended in complete medium, and counted for both cell number and dpm per 0.1 mL. For determination of cell-associated radioactivity, 0.1 mL of a cell suspension was placed in 10 mL of Ecolite (+) (ICN Radiochemicals) and counted in a Beckman LS3801 scintillation system operating at 94% counting efficiency for ³⁵S. In the experiments for which absorbed dose is determined, the activity/cell was found to be 27 dpm/cell. Half of the resuspended cells were further processed to isolate nuclei using a procedure described by Osborn et al. (25) . This procedure involved washing the cell pellet 1x with PBS, 2x with 10 mM HEPES, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT (buffer A), and incubation for 10 min on ice in buffer A with 0.1% NP-40 detergent. The nuclei were collected after vortexing for 10 s and pelleting in a microfuge at 10,000 rpm for 10 min at 4°C. A clean nuclear fraction was verified by resuspension of the resulting pellet in 0.1% sodium citrate solution containing propidium iodide at 50 µg/mL and examination with an Olympus BH2-RFL microscope with a reflected light fluorescence attachment. An aliquot of the nuclear suspension was counted in a liquid scintillation counter to determine nucleus-associated radioactivity. The percent nuclear uptake of cell-incorporated radioactivity was calculated by dividing the dpm associated with the nuclear fraction by the dpm associated with the equivalent number of cells. This value was determined to be 4% after 1 and 2 h of incubation in radiolabeling medium as well as after a 2 h chase period in complete medium. Although seemingly low, this level of nuclear incorporation of ³⁵S-methionine is consistent with that observed (5%) with V79 Chinese hamster lung fibroblasts (15) . During the 2 h chase, 10% of the intracellular activity was cleared from the cells.

Determination of cell and nuclear dimensions
Laser scanning cytometry using a Meridian ACAS 570 Interactive Cytometer (Meridian Instruments, Okemos, MI, USA) was used to determine the dimensions of the HCT 116 cells and nuclei. Cell size was obtained by imaging live cells stained with 5(6)-chloromethyl-2',7'-dichlorofluorescein whereas nuclear size was estimated from propidium iodide-stained fixed cells. Staining procedures for cells and nuclei have been described previously (24 , 22) . Meridian’s image analysis software was used to determine the average radial dimensions of the stained cells and nuclei, R_C and R_N, which were 7 and 5 µm, respectively. These data were corroborated by confocal fluorescence analyses using a Bio-Rad MRC1024 Confocal Fluorescence Microscope housed in the Center for Microscopy and Image Analysis at GWU. In addition, in cell monolayers that were 60–70% confluent, each roughly spherical cell was surrounded by an average of six neighboring cells. These data were used to calculate the mean absorbed dose to the cell nucleus, as shown in the Appendix.

Gamma irradiation
To determine the amount of DNA fragmentation as a function of dose of -radiation, samples were placed in deficient medium supplemented with unlabeled methionine and irradiated at room temperature with a ¹³⁷Cs source (Shepherd Mark I Model 25 Gamma Irradiator) at 964 cGy/min to deliver stated doses. Cells were harvested and prepared for Comet analyses immediately after irradiation. The goal of these analyses was to establish initial DNA damage as a biological dosimeter of early radiation damage. The biological dosimeter was calibrated with the standard reference radiation, ¹³⁷Cs -rays. A similar approach (based on survival of granulocyte-macrophage colony-forming cells) has been used for bone marrow toxicity caused by incorporated yttrium-90 (26) . To monitor the rate of DNA repair, Comet analyses were performed at various times after -irradiation at 2000 cGy.

For the microarray analyses, cells were first incubated in deficient medium with unlabeled methionine and cysteine each at 1 µg/mL for 2 h before irradiation to establish culture conditions identical to those used for radiolabeled samples. This protocol was designed to inflict an equivalent amount of DNA damage in ß-exposed and -irradiated samples at matched time points before harvesting RNA for the microarray analyses. Parallel cultures were set up for Western analyses for which the cells were harvested either immediately after the radiation exposure or after 2, 4, and 22 h of chase in complete medium. Controls for all experiments were run in parallel with the experimental samples and followed identical protocols except that they were mock-irradiated (i.e., the ¹³⁷Cs irradiator was not on).

Quantitative Comet analysis for DNA fragmentation
The Comet analysis is a sensitive assay for DNA fragmentation that involves single-cell electrophoresis in agarose (27) . Cells for the Comet assay were irradiated per protocol for ³⁵S-methionine radiolabeling, -irradiation, or control at stated doses. At appropriate times, cells were harvested and analyzed for double-stranded DNA damage using the CometAssay^TM kit as described previously (24) . The resulting fluorescently stained DNA Comets were scanned using a Meridian ACAS570 Interactive Laser Cytometer for quantitative determination of the extent of DNA fragmentation. The head (cell body) and tail (fragmented DNA) portions of individual DNA Comets were circumscribed using Meridian’s image analysis software and the integrated fluorescence values of each defined area were recorded for a minimum of 25 individual Comets. The ratio of tail/head fluorescence was used as a relative measure of DNA fragmentation and is referred to here as the DNA fragmentation index (DFI). Normalized DFI values were obtained by dividing the DFI values for irradiated samples by that of the unirradiated controls.

Western analysis
For Western blot analyses, radiolabeled, -irradiated, and control samples were lysed directly in 2x SDS gel loading buffer (100 mM Tris-HCl, 200 mM DTT, 4% SDS, and 20% glycerol, pH 6.8) before electrophoretic separation of 25 µg protein per sample on 12% SDS-polyacrylamide gels according to standard protocols (28) . BCA protein assays were performed in order to standardize the amount of cellular protein for SDS-PAGE. The gel-separated proteins were then transferred to nitrocellulose membranes using a Bio-Rad Trans-Blot SD semi-dry electrophoresis transfer cell. After blocking the membrane in a 10% solution of nonfat dry milk (NFDM) in 50 mM Tris, 150 mM NaCl, 0.05% Tween-20, pH 7.4–7.5 (TTBS), the blot was incubated with the respective primary antibody in a 3% solution of NFDM in TTBS before washing and exposure to the appropriate secondary antibody-HRP conjugate. Finally, the blot was stained with SuperSignal West Dura Extended Duration Substrate (Pierce) and exposed to X-ray film (Hyperfilm, Amersham, Piscataway, NJ, USA), which was then developed with a Kodak M35 X-OMAT X-Ray Film Processor.

RNA isolation and purification
RNA for cDNA microarray analysis was isolated from control cells as well as from cells irradiated per the respective ³⁵S-methionine radiolabeling or -irradiation protocols. RNA isolation was accomplished using RNAzol B following the manufacturer’s protocol with extensive homogenization procedures added to improve yield and purity. Isolated RNA was stored at -70°C until it was purified by DNase treatment using Ambion’s DNA-free^TM Kit.

Verification of RNA purity and integrity
RNA purity was assessed by diluting a 10 µL aliquot of isolated RNA to 200 µL in TE, pH 8.0, and measuring the OD_260/280 at 22°C. All samples for cDNA microarray analyses had a 260/280 ratio 2.0 and were verified for purity by gel electrophoresis in 0.5x TAE/1% agarose using a 1x TAE buffer. The purified RNA was then stored at -70°C until it was used for cDNA microarray analysis.

Preparation of cDNA
Poly-A⁺ RNA enrichment was conducted using biotinylated oligo(dT) and streptavidin magnetic beads per Clontech’s protocol to isolate mRNA for microarray analysis. cDNA probes were constructed from the mRNA according to the Atlas Pure Total RNA Labeling System^TM protocol for Clontech Nylon Arrays, which was optimized for RNA purity and yield by using the more active SuperScript Reverse Transcriptase (Invitrogen, Gaithersburg, MD, USA) and an excess of enzyme. After purification of the cDNA through NucleoSpin^TM columns, scintillation counts were obtained to determine efficiency of the labeling procedure. Only samples with total radioactivity above 5 x 10⁶ cpm were included in the microarray analysis.

cDNA hybridization
Radiolabeled (³²P) cDNA was hybridized to Clontech Atlas^TM Human 1.2 Gene Arrays (containing cDNAs to 1176 named human genes) overnight at 68°C in ExpressHyb solution following a 30 min prehybridization step with 0.5 mg of heat-denatured sheared salmon testes DNA. The membranes were washed 4x with Wash Solution 1 (2x SSC, 1% SDS) and 1x with Wash Solution 2 (0.1x SSC, 0.5% SDS), both at 68°C, 1x with 2x SCC at RT, then immediately wrapped in plastic wrap before exposure to a PhosphorImager screen (Molecular Dynamics) for 96 h at room temperature. After development of the PhosphorImager screens, the membranes were reexposed to HyperFilm X-ray film (Amersham) for 7 days at -70°C to detect additional low level hybridization.

RNase protection assay
RNase protection assays were conducted with RNA isolated as described from cells irradiated per the respective ³⁵S-methionine radiolabeling, -irradiation, or control protocols. mRNA expression was quantitated with a ³²P-labeled multitranscript probe containing gene sequences for anti-apoptotic proteins bcl-W and bcl-X_L; the cdk inhibitors, p21^waf1/cip and p15^INK4B; and GADD45 and cyclin A, following the manufacturer’s instructions (BD PharMingen, San Diego, CA, USA). Protected ³²P-labeled probes were resolved on a 5% acrylamide sequencing gel and the dried gel was exposed to a PhosphorImager screen (Amersham, Piscataway, NJ, USA).

Densitometry, microarray image acquisition, and analysis
Exposed X-ray film from Western analyses was analyzed using a Molecular Dynamics [PDSI] Densitometer and the data was quantified using ImageQuant image analysis software (Molecular Dynamics). Data from the Western analyses were normalized according to the amount of protein in each sample and relative changes in protein expression are given in relation to the level of the respective protein level in control samples. PhosphorImager screens from the cDNA microarrays and RNase protection assays were analyzed using the STORM 860 PhosphorImager and quantified using ImageQuant software on the densitometer. Relative expression in the RPA analyses was determined by normalization to the respective expression of two housekeeping genes (L32 and GAPDH) and the normalized results were expressed as percent of the respective control. Data from the densitometric scans of the microarrays were passed directly as raw data into GeneSpring 4.1.1 Expression Analysis Software (Silicon Genetics) for further analysis.

Statistical analysis
Normalization and statistical analysis of cDNA microarray data were conducted entirely with the GeneSpring Version 4.1.1 expression analysis software. All raw data was first corrected for background intensity by subtraction of negative controls (averaged data from four blank regions of each respective membrane). Then, "per chip" normalization was conducted by normalizing all data on a given array to the average value of 3 positive controls, which included cytoplasmic ß-actin, 60S ribosomal protein L13A, and 40S ribosomal protein S9. These genes were selected as positive controls because of the magnitude of their expression in all arrays, the uniformity of their expression across arrays, and their general "housekeeping" functions. In addition to the positive and negative normalization procedures for each array, data for each gene in the irradiated samples were normalized to the respective gene in time-matched, unirradiated controls. This "per gene" normalization to time-matched controls was done to correct for experimental artifacts that may result solely as a consequence of the incubation conditions or the incubation time. Hierarchical clustering of samples was performed using GeneSpring’s "experimental tree" analysis software. This analysis, which shows the relatedness between samples based on similarities in expression patterns, was conducted on the six irradiated samples using a standard correlation model with a separation ratio of 0.5 and a minimum distance of 0.001, both of which are default values that determine the branching behavior of the tree. Expression restrictions were applied such that only genes with raw densitometric values 900 (background defined on the basis of negative controls) and with normalized values 0.1 with respect to positive controls in all eight arrays were included in the analyses in order to determine changes in only highly expressed genes.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
Empirical determination of gamma ray dose that causes DNA damage equivalent to incorporated ³⁵S-radionuclide
To compare the gene expression profiles of ³⁵S-radiolabeled cells with that of cells exposed to the standard reference -radiation, we needed to define a dose of -radiation that would give a comparable biological end point. We selected a short-term biomarker of radiation damage for the empirical determination of a biologically equivalent dose of acute -radiation that would cause the same effect as internal ß-radiation resulting from the incorporation of ³⁵S-methionine after a 2 h incubation with the radionuclide at a concentration of 100 µCi/mL, which is a standard recommended dose for metabolic labeling protocols (13) . Because DNA is considered to be a primary target of ionizing radiation, the extent of DNA fragmentation immediately after radiation exposure was used as an early biomarker of radiation damage. To construct a biological dosimeter for internal ß-radiation based on DNA damage, quantitative Comet analyses, as developed in our laboratory using laser scanning cytometry (24) , were used to measure the amount of fragmented DNA immediately after irradiation with various doses of -radiation from a ¹³⁷Cs source and after incubation of cells for 2 h with different amounts of ³⁵S-methionine. Figure 1 A, B shows that both cell-incorporated ß-radiation and external -radiation respectively induced dose-dependent fragmentation of DNA in HCT116 cells as revealed by quantitative Comet assays. To facilitate comparison between the extent of DNA damage mediated by different doses of -radiation and that induced by ³⁵S-methionine at different levels of cell incorporation (which was found to be linearly dependent on the amount in the incubation medium), normalized DFI for ß- and -irradiated samples were least squares fitted and the gamma ray dose D() required to achieve the same level of DNA damage (i.e., DFI) induced by a given concentration of the radionuclide in the culture medium C(ß) was calculated to be D() = (47 ± 1.3)C(ß), where the units are cGy and µCi/mL, respectively. Figure 1C shows this direct relationship between the concentration of radionuclide in the incubation medium and the dose of external -radiation in terms of inducing a specific level of initial DNA damage. Based on this ß- calibration curve, radiolabeling with ³⁵S-methionine for 2 h at 100 µCi/mL causes DNA damage that is equivalent to that caused by 4700 cGy of gamma rays. However, during the course of these studies, we noticed that cell exposure to 5000 cGy of acute -radiation led to early and substantial cell damage as indicated by morphological changes (e.g., blebbing). Thus, in subsequent microarray studies, we limited our high dose of acute -irradiation to 2000 cGy, which produced a level of DNA fragmentation that was 86% of that produced by ³⁵S-methionine at 100 µL/mL according to the fitted curve derived by nonlinear regression analysis (Fig. 1D ), without inducing obvious morphological changes within the experimental time period. This dose also facilitated comparisons of our microarray data with that obtained by others who have studied cellular responses to 2000 cGy of -radiation (7) .

View larger version (15K): [in this window] [in a new window]   Figure 1. Comparison of DNA fragmentation after A) 2 h incubation with 35S-methionine at increasing doses of incorporated radioactivity and B) external -radiation at different doses. Quantitative Comet analyses were used to determine the extent of DNA fragmentation (or DFI) reflected in the DNA Comet’s "tail/head" fluorescence ratio as described in Materials and Methods and ref 24 . A, B) Dashed line represents the DFI value of the unlabeled or unirradiated controls. C) A ß- calibration curve, derived from least squares fitting of normalized data from panels A, B was constructed in panel C by plotting the dose of -radiation that would yield the same DNA fragmentation index (DFI) that would be expected after a 2 h incubation in a given amount of ß-emitter (35S-methionine). D) DFI normalized by unirradiated control values for -irradiated samples. The data were fitted to a second order polynomial by nonlinear regression analysis (GraphPad PRISM version 3.00). Dashed line indicates the normalized DFI corresponding to ß exposure at 100 µCi/mL.

The rate of delivery (dose rate) of internal ß-radiation (50 cGy/hr, Appendix 1) is substantially lower than that of the external pulsed -radiation used to deliver 2000 cGy (2.07 min at 964 cGy/min). Accordingly, the persistence of DNA damage in ß- and -irradiated cells is different as shown in Fig. 2 . The apparent DNA repair rate, as inferred from the time-dependent decrease in the amount of fragmented DNA detected by Comet assay (DFI), is much lower in ß-irradiated cells (t_1/2 >24 h) than in -irradiated cells (t_1/21 h), and reflects the chronic nature of irradiation by the intracellularly incorporated ß-emitter, which increases the cumulative absorbed dose over time beyond the initial 2 h radiolabeling period, partially offsetting cellular attempts at DNA repair. Thus, it is predicted that, despite nearly equivalent initial DNA damage (Fig. 2) , molecular and cellular responses to internal ß-emitters would be different from that induced by pulsed -radiation, and therefore worthy of independent study.

View larger version (16K): [in this window] [in a new window]   Figure 2. Time-dependent decrease in level of fragmented DNA (DFI) in cells after incubation for 2 h in 35S-methionine at 100 µCi/mL vs. cells irradiated with 2000 cGy of -radiation. Comet analyses were performed at various times after washout of the 35S-methionine and return to complete unlabeled medium or after -irradiation.

Kinetics of activation of p53, p21, and cdc2 by ß- and -radiation
Given that p53 plays a major role in the transcriptional response to -radiation (1 2 3 4) , we monitored the activation of p53 and p21, one of its downstream targets, to define the temporal changes in these cell cycle regulators to internal ß-radiation as well as external -radiation. As shown in Fig. 3 A, p53 is substantially activated in radiolabeled cells within 2 h of the labeling period, as indicated by increases in the amount of p53 protein present as well as its level of phosphorylation, inferred from the upper band of the doublet reacting with p53-specific antibody on the Western blot (Fig. 3B ). In all Western analyses, equal amounts of total protein per sample were applied to the gel and the levels in the ß- or -exposed samples were normalized by the amount of the respective protein present in unlabeled, unirradiated control samples. Although p53 activation apparently peaks by 4 h after initial exposure to ³⁵S-methionine, the level of p21 protein increases steadily from 2–6 h after label addition, and at an apparently slower rate thereafter up to 24 h (Fig. 4 ). Similar temporal changes in p53 and p21 were observed in cells irradiated with 2000 cGy of -radiation, but they were attenuated and biphasic in cells irradiated with low dose -irradiation at 20 cGy. The biphasic nature of the p21 time-course at 20 cGy is consistent with that observed for -irradiation of ML-1 myeloid leukemia cells in the low-dose range of 2–50 cGy (29) . Because ³⁵S-methionine-labeled cells were previously observed to be arrested in G2 phase (22) , changes in cdc2 were also investigated. cdc2 phosphorylation levels were significantly elevated at the 2 h time point and followed a similar biphasic time course in all 3 samples (ß at 100 µCi/mL and at 20 and 2000 cGy), whereas cdc2 protein levels followed a different time course for the ß- and -irradiated samples (Fig. 5 A, B). Based on these studies demonstrating a protein product of p53-mediated transcriptional activation (p21) by 4 h with post-translational changes in cdc2 occurring as early as 2 h after introduction of the radiolabel, we decided to investigate gene expression profiles of ß- and -irradiated cells at two time points: 1) immediately after radiation exposure, and 2) after a 2 h chase period in complete medium without ³⁵S-methionine.

View larger version (15K): [in this window] [in a new window]   Figure 3. Relative changes in p53 A) or phospho-p53 B) expression as a function of time after 2 h pulse label in 35S-methionine at 100 µCi/mL or after -irradiation at 20 or 2000 cGy. The relative amounts of p53 or phospho-p53 (inferred from slower migrating p53 staining band) were revealed by Western analyses and quantitation of Western blots by densitometry. In all Western analyses, the amount of protein in radiolabeled or irradiated samples was normalized by the amount of protein in unlabeled, unirradiated controls.

View larger version (15K): [in this window] [in a new window]   Figure 4. Relative changes in p21 protein expression as a function of time after 2 h pulse label in 35S-methionine at 100 µCi/mL or after -irradiation at 20 or 2000 cGy. The relative amounts of p21 were revealed by Western analyses and quantitation of Western blots by densitometry.

View larger version (14K): [in this window] [in a new window]   Figure 5. Relative changes in cdc2 A) or phospho-cdc2 B) expression as a function of time after 2 h pulse label in 35S-methionine at 100 µCi/mL or after -irradiation at 20 or 2000 cGy. The relative amounts of cdc2 or phospho-cdc2 were revealed by Western analyses and quantitation of Western blots by densitometry.

cDNA expression profiling of ß-exposed and -irradiated samples
cDNA microarray analyses were used to determine the gene expression profiles of cells radiolabeled with a standard dose of ³⁵S-methionine (100 µCi/mL) after a 2 h labeling period and following an additional 2 h "chase" period in unlabeled complete medium. For simplicity, these samples are identified as ß2 and ß4, respectively. Such conditions are typical of many experimental protocols that use radioactive methionine as a tracer to follow protein modification, maturation, and degradation in a wide variety of cellular processes (13) . For the purpose of comparison with cells exposed to internal ß-radiation, cells were irradiated with external -radiation at low and high doses: 20 cGy, which resulted in transient induction/activation of p53 and p21 (Figs. 3 and 4) , and 2000 cGy, which produced comparable DNA damage to that induced by ³⁵S-methionine at 100 µCi/mL (Figs. 1 and 2) . The -irradiated samples were incubated under identical conditions as the radiolabeled samples and analyzed at analogous time points relative to the start of the incubation in deficient medium. -irradiation took place shortly prior to the end of the 2 h incubation in deficient medium in order to induce the maximum amount of DNA damage at a matched time point (with respect to the radiolabeled sample) during the experiment. That is, RNA was harvested for the first time point (labeled "2 h") within minutes of the respective exposures when the amounts of DNA fragmentation induced by the ß-emitter or 2000 cGy -radiation are equivalent and maximal, as shown by Comet analyses (Fig. 2) . The "4 h" time point includes a 2 h "chase" in normal medium following the irradiation procedures to allow further development of the response and to model a typical pulse-chase experimental protocol using ³⁵S-methionine. The -irradiated samples are respectively identified as 20-2, 20-4, 2000-2, and 2000-4, with the former number denoting the dose in cGy and the latter number denoting the time point relative to the start of incubation in deficient medium.

Figure 6 summarizes the results of the DNA microarray analyses of radiolabeled and -irradiated HCT116 cells at the selected time points in terms of number of genes expressed at different-fold levels of induction. Mock-irradiated control samples incubated under identical conditions as the irradiated samples were included for each of the two time points to control for any experimental artifacts resulting solely from the conditions of the irradiation protocols. Of the six samples exposed to radiation, the ß4 sample exhibits the strongest response to radiation both in terms of number of genes up-regulated and the magnitude of the increases. In contrast, immediately after the radiolabeling period, only nine genes are up-regulated > twofold in the ß2 sample. Eight of these nine genes are also up-regulated in the ß4 sample and the 20-2 (but not 20-4) sample (Table 1 ). Less than half (2 3) of these genes are induced > twofold in either of the -2000 samples. These results suggest that these nine genes, which include redox-active, stress, and DNA repair proteins, are representative of early cellular response to low levels of ionizing radiation. A time-dependent enhancement of overall gene induction is also observed with the 2000 samples (Fig. 6) . In contrast, the 20 samples show a moderate (2–3x) up-regulation of a substantial number of genes at 2 h but a marked reduction in this number at the 4 h time point. This biphasic response at low radiation dose may be indicative of the activation of transient and rapid repair responses to limited DNA damage (see Fig. 1 ).

View larger version (29K): [in this window] [in a new window]   Figure 6. Summary of microarray data showing the total number of genes expressed above specific threshold levels in each of the 6 irradiated samples relative to respective time-matched unirradiated, unlabeled controls.

Of the 162 genes induced twofold over the time-matched control by incorporated ß-radiation at the 4 h time point, roughly 40% (or 63) are also induced by -radiation at either 20 or 2000 cGy. These gene inductions most likely represent a general response to ionizing radiation. A Venn diagram comparing genes expressed twofold (relative to the 4 h control) in the ß4 and 2000-4 samples reveals a set of 38 common genes that includes transcription, signaling, DNA repair, cell cycle, apoptosis, and redox regulator proteins as depicted in Fig. 7 . These genes and their respective expression levels for these 2 samples are listed in Table 2 . Although the Venn diagram shows that there are a total of 124 genes up-regulated at least twofold only in ß4, of special note are the 35 genes that are differentially expressed by fivefold relative to the 2000-4 sample (Fig. 7 /Table 3 ). These include a proportionally large number of proteins involved in transcription, signaling, and cell adhesion. The majority of these genes (28/35) are also highly expressed ( fourfold) relative to the time-matched control. By contrast, the set of genes selectively up-regulated at least twofold in the 2000-4 sample includes a large number of apoptosis, inflammatory, and degradative proteins, which may be reflective of the acute nature of the high dose radiation (Fig. 7 /Table 4 ). These differences in gene expression between the ß4 and 2000-4 samples are consistent with a GeneSpring tree clustering analysis demonstrating early divergence between these 2 samples relative to the other irradiated samples as described below.

View larger version (48K): [in this window] [in a new window]   Figure 7. Overlapping and distinct classes of genes are both induced by internal ß- and external -radiation. A Venn diagram constructed from: all genes in ß4 expressed 2-fold greater than control (left circle), all genes in 2000-4 expressed 2-fold greater than control (right circle), and all genes in ß4 differentially expressed 5-fold relative to the respective genes in 2000-4 (bottom circle). Pie charts show the functional classification of genes up-regulated in both ß4 and 2000-4 (chart A); differentially expressed in ß4 in comparison to 2000-4 (chart B), and genes selectively elevated in 2000-4 but not in ß4 (chart C).

View this table: [in this window] [in a new window]   Table 2. Genes up-regulated by at least 2-fold in both ß4 and 2000-4

View this table: [in this window] [in a new window]   Table 3. Genes >2-fold up-regulated in ß4 that are also >5-fold differentially expressed from the same genes in 2000-4

View this table: [in this window] [in a new window]   Table 4. Genes up-regulated at least 2-fold in 2000-4 but not ß4

Statistical analyses of gene expression profiles induced by ß- and -radiation
The differences in cellular response to ß- and -irradiation were further analyzed using the statistical clustering techniques in the GeneSpring microarray analysis software package. The experimental tree-clustering model was used to examine the expression profile of each sample as a whole and to organize the samples in a hierarchical "experimental tree" depending on their relatedness to one another. Those samples that divide at the most proximal branches of the tree represent samples with the greatest overall difference in expression profiles in the experiment, whereas those that divide only at the most terminal branches share the greatest similarities in overall expression profiles. As shown in Fig. 8 , the "experimental tree" had three primary branches, representing the ß4 sample, the 2000-4 sample and, collectively, the remaining four irradiated samples. This statistically derived, hierarchical organization showing a proximal split between the ß and 2000 samples at the 4 h time point supports the argument that the expression profiles of these 2 samples are fundamentally different from each other despite sharing an overlapping set of response genes. On the other hand, the two most distal branches of the experimental tree were the ß-exposed and the 20 samples at the 2 h time point, supporting the argument that the early response to ionizing ß-radiation from cell-incorporated ß-emitters is most similar to the early response to a low dose of external -radiation. Both of these samples are most closely related to the unirradiated controls, which have a normalized value of 1 for every gene in each of the graphs.

View larger version (42K): [in this window] [in a new window]   Figure 8. An experimental "tree" diagram demonstrating the relationships between the irradiated samples in terms of gene expression profile for genes meeting the criteria of having a raw value of at least 900 and having a normalized value of at least 0.1 relative to the respective genes in time-matched controls. For this statistical analysis, a separation ratio of 0.5 and a minimum distance of 0.001 was used. The tree analyses shows that the ß4 and 2000-4 samples are distinctly different from one another, thus splitting off at the lowest branches, whereas the ß2 sample is most similar to the 20-2 sample. Prominently overexpressed or underexpressed genes (relative to time-matched unirradiated controls) are noted for each sample as follows: a, Microsomal stress 70 protein ATPase core precursor; b, selenium binding protein; c, ZO-1; d, HINT; e, GSTT1; f, GRB2; g, neurotrophin-4; h, SWI2; i, diaphanous 1; j, p53; k, IRF2; l, telomeric repeat binding factor 1; m, cadherin 11; n, APAF1; o, cIAP1; p, M-cadherin; q, GADD45 ß; r, EGR1; s, caspase 8; t, guanine nucleotide binding protein ß subunit 1; u, BARD1; v, MAPKAPK-2; x, transferrin receptor; y, MKK2; z, Na/K ATPase ß 3 subunit; 1, Ku 70; 2, AP4; 3, microsomal GSTII; 4, glutathione reductase; 5, A-raf; 6, cationic amino acid transporter 3; 7, integrin ß-8 precursor; 8, ZFM1, alternatively spliced; 9, retinoblastoma binding protein 2; 10, ZFM1; 11, VCAM1; 12, telomeric repeat binding factor 2; 13, integrin ß-6; 14, MLK3; 15, I-TRAF; 16, cdc2; 17, SP1; 18, translin; 19, MAPK3; 20, cyclin H.

Gene expression profiling also reveals distinct patterns of gene repression
The gene expression line graphs in Fig. 8 also illustrate that in addition to gene induction, both forms of radiation result in down-regulation of the mRNA of select genes. As for the gene induction profiles, the gene repression profiles are characterized by both overlapping and differential expression. Similarly down-regulated genes are those involved in mitogenic and proliferative responses (Table 5 ). Some of the genes most prominently affected positively as well as negatively in each sample have been identified in the legend to Fig. 8 .

View this table: [in this window] [in a new window]   Table 5. Repressed genes (<0.5 relative to control) in both ß4 and 2000-4

Partial confirmation of microarray data
RNase protection assays (RPA) and Western analyses were used to examine changes in gene expression that were indicated by the microarray data. Figure 9 shows the relative changes in mRNA expression for select genes as indicated by microarray and RPA analyses of the ß4 and 2000-4 cells. Table 6 and Table 7 summarize data from Western analyses that show protein changes partially consistent with the changes in mRNA level for select proteins. Collectively, these data verify gene expression changes or trends for several of the proteins involved in cell cycle arrest, cellular stress, and apoptosis. Table 7 , in particular, shows that relative changes in protein levels are time-dependent and show a lag period between changes in message level and corresponding changes at the protein level. Thus, analysis at the protein level is especially important when studying cellular stress response, because a significant number of the proteins involved are regulated by post-transcriptional as well as post-translational modifications, such as altered protein stability or phosphorylation state, respectively. p53 is one such protein whose activity is altered by both mechanisms (30 , 31) , as confirmed in this study (Fig. 3) . The RPA and Western analyses together corroborate the microarray data and further illustrate the importance of conducting research at the genomic and proteomic levels.

View larger version (35K): [in this window] [in a new window]   Figure 9. Relative changes in expression of select stress response genes in ß4 and 2000-4 samples as revealed by RNase protection assay. The filled and open bars show RPA data for ß4 and 2000-4, respectively; dotted and hatched bars show microarray data for these respective samples.

View this table: [in this window] [in a new window]   Table 6. Changes at the protein level for select proteins in ß4 and 2000-4 samples

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
The primary objectives of this project were to evaluate the effect of cell-incorporated ³⁵S-methionine on the gene expression profile of the radiolabeled cells and to compare this profile with that induced by a "biologically comparable" dose of acute external -radiation that was determined on the basis of roughly equivalent DNA damage as measured by the Comet assay. Results from DNA microarray analyses show that cellular incorporation of ³⁵S-methionine at 100 µCi/mL for 2 h induces substantial changes in gene expression within 4 h of label addition and that the resulting gene expression profile is statistically distinguishable from that induced by 2000 cGy of external -radiation. We also present a new empirical approach to quantifying the biologically effective dose of a given amount of internal ß-radiation under specific labeling conditions based on identifying the dose that induces similar molecular outcomes. This approach, unlike standard cellular dosimetric calculations, is independent of the size and shape of the cells and nuclei as well as independent of the availability of tabulated S values for the specific isotope. However, unlike cellular dosimetry approaches, this method cannot differentiate between the many reasons that may be involved with differences in observed responses, such as possible intrinsic differences in the relative biological effectiveness (RBE) of low energy ß particles and acute external gamma rays, and dependence of RBE on subcellular distribution of the radiochemical. Nor does it take into consideration the differences in dose rates of the two types of radiation delivery.

Empirical determination of gamma ray dose that causes DNA damage equivalent to incorporated ³⁵S-radionuclide
Initial post-irradiation DNA fragmentation was used as an early molecular marker to establish bioequivalence between a given concentration of radiolabel in the culture medium over a specified labeling period and a specific dose of acute -radiation. Analysis of the dose dependence of DNA fragmentation as revealed by quantitative Comet assays (Fig. 1) indicates that exposure of HCT116 cells to 100 µCi/mL ³⁵S-methionine for 2 h is roughly equivalent to irradiating cells with 4700 cGy of -radiation. However, Fig. 1D shows that the DFI relative to control samples approaches a limiting value at 2000 cGy. In fact, the normalized DFI induced by 2000 cGy is 86% of that induced by ³⁵S-methionine at 100 µCi/mL. A noteworthy comparison of the empirically derived limiting dose (2000 cGy) with the calculated absorbed dose to the nucleus based on cellular uptake of ³⁵S (see Appendix) reveals that the dose is roughly an order of magnitude higher than that derived for ³⁵S from cellular dosimetry calculations (120 cGy after a 2 h incubation, and 300 cGy after the 2 h chase period). It should be noted that both of these calculated doses would be expected to yield normalized DFI values substantially lower than that resulting from incubation with ³⁵S-methionine at 100 µCi/mL (Fig. 1D ). This result suggests that under our experimental conditions, internal ionizing radiation from cell-incorporated ³⁵S-methionine may be significantly more effective in causing radiation damage to cellular DNA than a comparable dose of external pulsed -radiation. This relatively high efficiency of radiation damage needs to be taken into account when assessing biological risk factors of incorporated radionuclides.

Relative rates of DNA repair in cells exposed to internal ß-radiation or external -radiation
A major difference between the two irradiation protocols is that the cell-incorporated radionuclide accumulates over 2 h in the labeling medium and is a chronic source of radiation even after the removal of extracellular label whereas the external -radiation is administered as a brief pulse. Consequently, it is not surprising that the apparent rate of DNA repair, as inferred from the time-dependent decrease in fragmented DNA, is different in the two samples after the respective exposure periods (Fig. 2) . With the pulsed -irradiation, the t_1/2 for DNA repair is 1 h, which is consistent with published studies (32 , 33) , whereas the effective t_1/2 for repair (as indicated by reduction in DFI) in the radiolabeled samples is >24 h. This difference, due in part to the acute vs. chronic nature of the radiation, may be reflected in the unique characteristics of the cellular responses to the respective forms of radiation as revealed in part by the genomic analysis. Although it may be possible to mimic chronic internal ß-radiation more accurately with a chronic -irradiator (34) , this experimental setup is neither commonly available nor typically used to study the effects of ionizing radiation on cells.

Global changes in gene expression are induced by cell-incorporated ³⁵S-methionine
Most investigators who use low energy ß-emitters as tracers to study biological processes assume there are little or no untoward effects induced by the radiolabels, despite early radiobiological studies demonstrating the cytotoxicity of such molecules (11 , 12) . More recently, we and others have documented that cell-incorporated radionuclides (³H, ³⁵S, and ³²P) can induce significant molecular and cellular aberrations ranging from DNA damage to cell cycle arrest to apoptosis (17 18 19 20 21 22 23 24 , 35) . Still, these findings may be easy to ignore if one assumes that these changes have no impact on short-term experiments or are limited to cellular activities necessary for mitosis or continued cell proliferation. Here, we demonstrate for the first time that metabolic labeling with a commonly used radioactive tracer molecule (³⁵S-methionine) induces global changes in gene expression that activate a number of stress response signaling pathways. In effect, we are stressing cells by radiolabeling them even for brief periods at commonly used concentrations of ³⁵S-methionine. Among the stress signaling pathways transcriptionally activated are 1) the p53 DNA damage-inducible pathway leading to activation of cell cycle- and apoptosis-regulatory genes (see Fig. 10 for changes in relevant genes); 2) oxidant stress pathways; 3) the DNA damage and repair pathways; 4) the Fas signaling network; and 5) a multitude of transcription factors and signaling molecules, including an abundance of cell adhesion proteins, some of which have recently been implicated in signal transduction across the plasma membrane (36) . On the other hand, genes for mitogenic factors (e.g., cdks and early cyclins) and signaling molecules (e.g., ERKs) are simultaneously down-regulated (Table 5) .

View larger version (22K): [in this window] [in a new window]   Figure 10. A schematic diagram illustrating some of the stress signaling pathways that are transcriptionally activated by cell-incorporated 35S-methionine. The majority of the proteins shown as up-regulated were confirmed by RNase protection assays or Western analyses. LEBE, low energy beta emitters.

Although we have not yet corroborated all of the gene expression changes with corresponding changes in protein expression, the data in Fig. 9 and Tables 6 and 7 confirm changes in a number of key proteins in some of the pathways mentioned above. Furthermore, a comparison of the temporal changes in mRNA and protein levels for p53, p21, and cdc2 (Table 7) shows a time lag of 2 h between changes at the mRNA level and corresponding changes at the protein level. Variable lag times in protein expression may account for the relatively modest increases in protein levels detected in comparison to the level of the respective mRNA expressed at the 4 h time point (Fig. 9) .

Differential effects of chronic internal ß-radiation and pulsed external -radiation
Although similarities in response to internal ß-radiation and external -radiation are not unpredictable, it is interesting to note some of the highly differential responses to the two forms of radiation, which are presented in Fig. 7 and Table 3 . The extent to which these differences are attributable to the chronic vs. acute nature of the radiation delivery remain to be determined. It is nevertheless intriguing to speculate that further research on these differences and their functional significance may enhance our understanding of the effects of ionizing radiation on cells.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
In summary, DNA microarray analyses were used to analyze the gene expression profile of cells with incorporated ³⁵S-methionine and to compare this profile with that induced by a biologically comparable dose of acute, external -radiation. Our results show not only a significant up-regulation of stress genes expressed in the radiolabeled cells, but also distinct differences between the genes induced (or repressed, data not shown) by the two forms of ionizing radiation. Although these differences may be solely the result of the chronic nature of the internal ß-radiation in contrast to the acute -radiation, the possibility also exists that the intracellular location of the ß-emitters may influence cellular responses, as has been implicated for other incorporated radionuclides (37 , 38) . These results underscore the potential for experimental artifacts resulting from the use of standard metabolic radiolabeling techniques and demonstrate the need for additional research to further characterize the effects of cell-incorporated radioactivity and the significance of differential cellular responses to this source of ionizing radiation.

	APPENDIX

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS APPENDIX REFERENCES

 
Cellular Dosimetry
There are three sources that are responsible for the mean absorbed dose to the cell nucleus from decays of ³⁵S. These include the absorbed dose from extracellular decays in the culture medium D_medium, the self-dose from intracellular decays in the target cell D_self, and the cross-dose from intracellular decays in neighboring cells D_cross. The total mean absorbed dose to the cell nucleus D_total is therefore

(A1)

Medium dose D_medium
The mean absorbed dose to cell nuclei from extracellular ³⁵S decays in the culture medium can be estimated by the mean absorbed dose to the culture medium. According to the formalism of the Medical Internal Radiation Dose (MIRD) Committee (39) and the International Commission on Radiation Units and Measurements (40) , this is given by D_medium = _medium/m_medium, where _medium is the cumulated activity in the culture medium, is the mean energy emitted per nuclear transition, and m_medium is the mass of the culture medium. The mean absorbed dose per unit cumulated activity for ³⁵S is = 7.82 x 10^-15 Gy-kg/Bq-s (41) . The quantity _medium/m_medium = ], where A_medium(t) is the activity in the culture medium as a function of time. Assuming that the 100 µCi/mL (3700 kBq/mL) of ³⁵S undergoes negligible decay during the 2 h labeling period, and assuming that the culture medium is unit density, then _medium/m_medium = 3700 kBq/g (2 hr) = 2.66 x 10¹³ Bq-s/kg. Therefore, D_medium = 7.82 x 10^-15 Gy-kg/Bq-s (2.66 x 10¹³ Bq-s/kg) = 0.21 Gy.

Cumulated Activity in the cell
Using the notation outlined in the MIRD Cellular S Values monograph (42)