Gene expression profiling of the response to thermal injury in human cells

HONG-KHANH B. DINH1,2, BAITENG ZHAO1,3, STEVEN T. SCHUSCHEREBA2, GERALD MERRILL3 and PHILLIP D. BOWMAN3

1 Division of Pharmaceutics, College of Pharmacy, University of Texas at Austin, Austin 78712
2 United States Army Medical Research Detachment, Brooks Air Force Base 78235
3 United States Army Institute of Surgical Research, Fort Sam Houston, Texas 78234


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The genetic response of human cells to sublethal thermal injury was assessed by gene expression profiling, using macroarrays containing 588 complementary known genes. At 1, 4, 8, and 24 h following thermal injury, RNA was isolated, and a cDNA copy was generated incorporating 33P and hybridized to Atlas arrays. About one-fifth of the genes on the membrane exhibited a significant elevation or depression in expression (>=2-fold) by 4 h posttreatment. Genes for heat shock proteins (HSPs) were upregulated as well as genes for transcription factors, growth regulation, and DNA repair. Cluster analysis was performed to assess temporal relationships between expression of genes. Translation of mRNA for some expressed genes, including HSP70 and HSP40, was corroborated by Western blotting. Gene expression profiling can be used to determine information about gene responses to thermal injury by retinal pigment epithelium cells following sublethal injury. The induction of gene expression following thermal injury involves a number of genes not previously identified as related to the stress response.

cDNA arrays; retinal cell culture; ARPE-19 cells; heat shock protein; thermal injury


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
BECAUSE OF THE INCREASED use of lasers in medicine, research, and military applications (e.g., laser sighting devices), instances of laser injury, particularly to the eye, have considerably risen in the past four decades (5, 40). The damage may be categorized into three general types, photothermal, photomechanical, and photochemical, of which the thermal damage appears to be most important (28, 30, 31). We evaluated the thermal aspect of the injury, because the resulting pathophysiology following laser damage closely parallels the response following burn injuries. The eye is especially susceptible to laser injury, in particular the retinal pigment epithelium (RPE) layer. As the laser light passes through the retina to the pigmented epithelium, the eye’s own focusing mechanism can concentrate the radiant energy as much as 10,000 times (44). This injury is predominantly a small thermal injury produced by a short pulse of relatively high energy, focused by the lens to produce irreversible damage at the site of energy absorption. In some of the clinical treatment protocols, temperatures are elevated from 10 to 60°C or more (28, 29). A secondary injury then results from a variety of factors including the inflammatory response initiated by thermally injured cells (42).

Although larger external burns are one of the most common forms of accidental injury, and heat is often used in conjunction with radiation for control of cancerous tissue (16), very little is known about the gene response of cells to thermal injury (23, 38). In the case of thermal treatment of cancer, heat is often used in conjunction with radiation therapy for control of cancerous tissue; however, the effectiveness of secondary heat applications is limited by the induction of thermotolerance from the initial treatment (17). The induction of thermotolerance, evident by the heat shock response, has been shown to provide temporary protection against ensuing episodes of injurious stimuli (8). Analysis of an Arrhenius plot for the inactivation of keratinocytes from short exposure (seconds) to high temperatures (50–60°C) indicated that the absorbing species or the injury mechanism may be somewhat different from injury induced in the most studied range (40–50°C over minutes) (6). Above a certain threshold of heating all cells die from exposure to short-duration thermal injury, yet they express large amounts of heat shock protein (HSP) and cytokines before doing so. Human keratinocytes exposed for 1 s at 58°C synthesized the cytokines interleukin-8 (IL-8) as well as mRNA for tumor necrosis factor-{alpha} and IL-1{alpha} and exhibited an intense heat shock response. Depending on the intensity of the heating, the initial form of cell death was apoptotic but quickly became accidental cell death as the intensity of heating increased (33, 47). Understanding how cells respond to a sublethal, survivable thermal injury may provide a means by which to enhance protective cellular responses and increase survival. Although the heat shock response is the best studied response to thermal injury and can perhaps be explained as an adaptive, protective response to heat and stress (51), other adaptive responses may be involved. The progression of the heat shock response is usually detected by radiolabeling or by antibody detection of proteins in polyacrylamide gels after electrophoresis. Recent developments, particularly in gene expression profiling with macro- and microarrays, have greatly increased the efficiency of gene expression analysis. Arrays allow for the parallel analysis of many genes from the same sample, revealing the complex interplay of genes invoked at any given time. Here we have examined the genetic response of 588 known genes to thermal injury. In addition to the well-described HSP response, genes involved in a variety of other metabolic pathways were altered. Although most genes exhibited no alteration in expression by 1 h after thermal injury, a few were mildly suppressed. By 4 h posttreatment a greater than twofold elevation was observed for some genes, including growth arrest/repair, DNA binding, and heat shock and other stress response proteins. Of particular interest were the heat shock proteins HSP40, HSP70, and HSP86, the 45-kDa and 153-kDa growth arrest and DNA-damage-inducible genes, and genes for DNA-binding and repair proteins. The translation of protein product from the induced mRNA expression was confirmed for HSP70 and HSP40.

By observing the gene response to thermal stress in an in vitro model, we may better understand the benefits and targets of drug treatment for secondary damage associated with thermal injury. Gene expression profiling may be applied to both drug discovery and development to define the response of specific cell or tissue systems to a drug, or to evaluate a drug’s mechanism of action. As we gather more knowledge of how and why genes function, we can use this knowledge to systematically find targets and complementary drugs for enhancing treatment.


    METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

Cell culture.
ARPE-19 cells, derived from human RPE, were a gift from Dr. Larry Hjelmeland, Department of Ophthalmology, University of California at Davis. Stock cultures were grown in T75 flasks in Dulbecco’s modified Eagle’s medium (DMEM, GIBCO-BRL; Life Technologies, Grand Island, NY) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 100 µg/ml penicillin/streptomycin, and 5 µg/ml Fungizone (GIBCO-BRL) at 37°C, 5% CO2, and 95% relative humidity. These cells exhibit a normal human diploid karyotype and an epithelial-like morphology (11).

Thermal injury experiments.
For survival experiments the cells were seeded on 13-mm Thermanox coverslips in 24-well plates, and for gene expression profiling they were cultured in 100-mm culture dishes. Thermal injury experiments were carried out after the cells reached confluence by dipping the cultured cells in HEPES-buffered saline heated to 55°C for the appropriate time. To determine the optimal time of exposure for inflicting a sublethal thermal injury, preliminary experiments using increasing times of heating (0–9 s) with a constant temperature of 55°C were carried out with cells grown on coverslips. The percentage of viable cells at 24 h postheating was determined for each experiment, using mitochondrial redox activity as an indicator. The optimal time and temperature for producing a 25% decrease in redox activity was chosen. Although this level of injury did not lead to morphological changes in the cells, it produced both a decline in metabolic activity and induced a genetic response that triggers early response and stress response genes.

For gene expression profiling, cells seeded in 100-mm petri dishes were dipped for 3 s, then returned to medium and the incubator for the appropriate recovery period. RNA and protein were isolated from cells 1, 4, 8, and 24 h after heating. Control cells were treated in a manner identical to that for heated cells, excluding the heating step, and were isolated concurrently with the 1-h and 24-h thermally injured cells. In preliminary experiments (data not shown), control cells were isolated at every time point until the reproducibility of control samples between time points was verified. The gene expression experiment was repeated five times.

Cell viability.
Cell viability after thermal injury was assessed using Alamar Blue (Biosource International, Camarillo, CA), which is converted to a fluorescing compound in amounts proportional to the number of viable cells (24). The viability measurements were taken at 24 h to determine that a recoverable sublethal injury was inflicted. Under these conditions, good quality RNA as judged by gel electrophoresis, a necessity for performing gene expression analysis, could be obtained. The cells were incubated for 2 h at 37°C with culture medium containing 10% Alamar Blue according to the manufacturer’s instructions. After incubation, fluorescence was measured at 530-nm excitation and 590-nm emission (Cytofluor 2350; Millipore, Bedford, MA).

RNA isolation.
RNA isolation was performed following the protocol of the Molecular Research Center (Cincinnati, OH) for TRI Reagent. RNA yield was determined spectrophotometrically, and the quantity and quality of the RNA were estimated from absorption at 260 nm and 280 nm and gel electrophoresis on a 0.9% agarose gel containing 1:10,000 SYBR Gold nucleic acid stain (Molecular Probes, Eugene, OR). Only undegraded RNA free of genomic DNA contamination was used. Isolated RNA was solubilized in deionized H2O and stored at -20°C until 33P labeling.

cDNA production.
Two micrograms of total RNA was heated to 70°C for 5 min, then immediately frozen in liquid nitrogen and lyophilized. The lyophilized RNA was combined with a master mix containing 2 µg oligo dT (1 µg/µl; Research Genetics, Huntsville, AL), 6 µl of 5x first-strand buffer, 3 µl 0.1 M DTT (GIBCO-BRL, Grand Island, NY), 3 µl of 5 mM dNTPs mixture, 1 µl RNase inhibitor (5Prime->3Prime, Inc.), 1.5 µl SuperScript II RT (200 U/µl, GIBCO-BRL), and 1 µl [33P]dCTP (10 mCi/mmol; Amersham Pharmacia, Arlington Heights, IL) and then incubated at 39°C. The reaction was stopped by the addition of 4 µl EDTA (0.5 M, Ambion). Hydrolysis of the RNA after reverse transcription was carried out at 68°C with 4 µl 1 M NaOH for 20 min. Unbound nucleotides were separated from the labeled probe by passage through a Centri-Spin 20 column (Princeton Separations, Adelphi, NJ).

Hybridization and phosphor imaging.
Blocking of nonspecific binding was performed in 5 ml MicroHyb (Research Genetics) with 1.0 µg/ml poly-dA (Research Genetics) and 1.0 µg/ml COT-1 DNA (GIBCO-BRL) at 42°C for 2 h. The purified probe was hybridized to an Atlas human cDNA expression array (no. 7740-1; Clontech, Palo Alto, CA) membrane at 42°C for 18–24 h in a 35 x 150-mm roller bottle (Roller Oven, Hybaid). Washes were performed twice at 50°C in 20 ml of 2x SSC and 1% SDS for 20 min and once at 25°C in 100 ml of 0.5x SSC and 1% SDS for 15 min. The membranes were then placed on moistened Whatman filter paper and wrapped with plastic wrap before exposing to super resolution screens (Packard Instrument, Meridian, CT). Following exposure to the phosphor-imaging screens (1–4 days), the distribution and intensity of radioactivity were determined by scanning the screens with a phosphor imager (Cyclone, Packard Instrument). The images were digitally acquired using OptiQuant software (v. 3.0, Packard Instrument), and the autoradiographs were analyzed for changes in gene expression using AtlasImage software (v. 1.5, Clontech).

Data analysis.
For analysis of gene expression, the ratios of the integrated optical density (IOD) for each cDNA target were compared between heated and control cells. Cluster analysis was then performed with software originally from Stanford University, courtesy of Dr. Michael Eisen (http://rana.lbl.gov/EisenSoftware.htm).

To determine significance, two separate array membranes were hybridized with the same control sample. Then, the average and standard deviation of IOD ratios between the two were used to indicate the level of difference necessary to claim a significant difference. A scatter plot of the two control IOD values illustrates the distribution along a line with the slope of the average intensity (Fig. 1). With an average and standard deviation of 1.14 ± 0.32, an IOD ratio of at least 2.10 represents a significant difference (P = 0.01).



View larger version (11K):
[in this window]
[in a new window]
 
Fig. 1. Scatter plot of integrated optical density (IOD) intensities for two control samples illustrating the distribution of points.

 
Protein isolation.
Protein was extracted from the phenol phase remaining after extraction of RNA from the aqueous phase of cultured cells following the protocol of the Molecular Research Center (Cincinnati, OH) for TRI Reagent. The protein samples were solubilized in 500 µl of lysis buffer (NOVEX, San Diego, CA) containing 1% ß-mercaptoethanol (Sigma). When necessary, the samples were incubated at 50°C to effect solubilization.

Gel electrophoresis, total protein determination and western blotting.
Ten microliters, corresponding to ~10 µg of protein from each treatment, was run on NuPage 4–12% Bis-Tris Gels (NOVEX) for 1 h at 125 mA. Additionally, 5 µl of MultiMark Multicolored Standard (NOVEX) was run on each gel as the molecular weight marker. The protein was then transferred to a polyvinylidene difluoride membrane (PVDF, NOVEX) for 5 h at 95 mA (Semi-Dry Blotter II; Kem En Tec, Flander, NJ). After transferring the protein, 2-methoxy-2,4-diphenyl-3(2H)-furanone (MDPF) fluorescent stain was utilized for total protein quantification, following the procedure outlined by Alba et al. (2). MDPF fluorescently tagged all protein, and a relative total protein ratio was calculated by dividing the fluorescence value for each sample by that of the control. The ratio was then used to normalize for any variability in the amount of protein from sample to sample. The blots were hydrated in a 50% methanol/water solution for 10 min and then washed twice in 10 mM sodium borate, pH 9.5, for 5 min each time. After the second wash, the blots were incubated for 10 min in 10 ml borate buffer containing 50 µl (0.50%) of 35 mM MDPF in DMSO. The blot was then rinsed briefly in borate buffer before visualizing under an ultraviolet light source. The total fluorescence for each sample was calculated using Gel-Pro Analyzer 2.0 software (Media Cybernetics, Silver Spring, MD).

Following total protein quantification, the membranes were hydrated in a 50% methanol/water solution for 10 min. The membranes were then incubated three times, for 15 min each, in a blocking solution consisting of 0.2% I-Block (Tropix, Bedford, MA), 0.1% Tween-20 (Sigma, St. Louis, MO), and 0.1% thimerosal (Sigma) in PBS (Sigma). The blots were then incubated with a primary antibody for 2 h. Rabbit anti-human HSP40 (1:10,000 dilution) and mouse anti-human HSP70 (1:1,000 dilution) were purchased from StressGen Biotechnologies (Victoria, British Columbia, Canada). After the initial hybridization in primary antibody, the blots were washed another three times in blocking solution for 15 min each. A 1:4,000 dilution of alkaline phosphatase-labeled secondary antibody (goat anti-rabbit from Zymed, San Francisco, CA; and goat anti-mouse from Tropix) was then added, and the blots were incubated for another 2 h. Finally, the blots were washed three times in PBS, 15 min each, before staining. The blots were stained in a 10% solution of nitroblue tetrazolium 5-bromo-4-chloro-3-indolyl phosphate (NBT-BCIP, Zymed), until the bands developed. Quantitative analysis was performed using Gel Pro Analyzer 2.0 software on blots scanned into the computer.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

Determination of heating regimen.
To determine a time and temperature for production of a reproducible sublethal injury, cell survival was assessed at a temperature of 55°C (Fig. 2). At this temperature, increasing time of exposure (0–9 s) to heated saline caused a linear decrease in mitochondrial redox capacity, as measured by the Alamar Blue assay at 24 h. By 48 h postheating, preliminary data (not shown) indicated that cells heated to 55°C for 3 s had recovered to viability comparable to controls. Visual inspection of the cells at 24 h posttreatment, by phase-contrast microscopy, showed that heated cells exhibited the same morphological characteristics as control cells until they have been heated for at least 6 s. At 6 s and beyond, the cells began rounding up and detaching from the plate or coverslip surface. At the chosen exposure duration of 3 s, there was no morphologically observable injury and mRNA was not degraded; however, there was approximately a 25% reduction in Alamar Blue conversion, and alterations in gene expression were demonstrable. The further reduction in Alamar Blue with longer heating times demonstrated the relationship between mitochondrial activity and cell viability (24). Since we used this assay as a marker of the overall damage caused by heating, we wanted to account for both decreases in cell number and from reduced redox capacity from injured cells (1, 4, 46).



View larger version (9K):
[in this window]
[in a new window]
 
Fig. 2. ARPE-19 cell viability by the Alamar Blue fluorescence assay 24 h after heating at 55°C for 0 to 9 s (n = 3). Cell viability is expressed in terms of percent of heat-treated cells surviving compared with unheated control cells.

 
Gene array profiling.
The digitally acquired autoradiographs for the Atlas array membranes were analyzed for alterations in gene expression using AtlasImage software (Clontech). The Atlas Human cDNA Expression Arrays (588 known genes) were used for this initial study, since only limited information exists for expressed sequence tag (EST) data at this time. Deciphering the relationships and interactions of these several hundred genes presents a challenging task. Additionally, the data set presented here only identifies the most significantly altered genes of those arrayed on this particular membrane and should not be interpreted as the only significantly altered genes following thermal injury. Of 588 known human genes, the unheated 1-h control retinal pigment epithelial (RPE) cells exhibited some expression of all genes (Fig. 3). The mRNA for these genes must be transcribed constitutively to be present in unheated cells and to exemplify the cell type’s normal genotype. By 1 h following heating at 55°C for 3 s, 46 underwent significant changes in expression compared with control cells, and they were all due to suppression. The most highly suppressed genes include the bcl-1 oncogene, fra-1, CLK-1, the Duffy blood group antigen, RACH1, and LIM kinase (Table 1).



View larger version (57K):
[in this window]
[in a new window]
 
Fig. 3. Representative autoradiograph of control membrane showing distribution and intensity of radioactivity visualized from Atlas Human cDNA Expression Array membranes. The autoradiographs were analyzed for changes in density between ARPE-19 cells heated at 55°C for 3 s and unheated control samples.

 

View this table:
[in this window]
[in a new window]
 
Table 1. Fold elevation or suppression of most highly altered genes 1, 4, 8, and 24 h after thermal injury

 
By 4 h, 53 genes exhibited a significant fold elevation of at least 2.10. These include genes that function in growth arrest/repair, DNA binding, heat shock, and other stress responses, such as HSP70, HSP40, GADD45, GADD153, IL-12, neurotrophin-4, and placenta growth factor. The only gene significantly suppressed at 4 h was the c-myc proto-oncogene. At 4 h, the most highly upregulated genes encode for heat shock proteins HSP70, HSP40, and HSP86, as well as growth arrest and DNA damage-inducible proteins GADD45 and GADD153 (Table 1). Other significantly upregulated genes at 4 h encode for IL-12, retinoic acid-binding protein II, placenta growth factor, MCP-1RA, neurotrophin-4, c-src-kinase, and HSP60.

By 8 h, the extent of upregulation for several HSPs, as well as GADD genes, had begun to decrease, whereas upregulation of c-src-kinase and chaperonin increased. The message for glutathione S-transferase pi, which had not been significantly altered before, was significantly suppressed at 8 h. DNA-repair proteins and the HSPs at 70 kDa, 86 kDa, and 60 kDa were still elevated. By 24 h, however, most of the early upregulated genes had either fallen back to background levels, or were minimally elevated. An elevation of proliferation-associated gene, melanoma differentiation associated protein 6 (mda-6), transcription and RNA polymerase elongation factor subunits, and DNA-binding protein TAX became apparent. On the other hand, the message for glia maturation factor-ß was ~20 times lower at 24 h. Other genes significantly downregulated include glutathione S-transferase pi, IL-1ß, TDGF1, VEGF receptor 1 precursor, and IL-5{alpha} receptor.

The temporal pattern of expression for these genes is more easily recognized through clustering (Fig. 4). Analysis of the most highly regulated subset of genes would suggest which cellular functions are most important for repair of injured cells. The clustered genes were grouped into dendrograms containing sets of genes that possess similar functions. Groupings within each dendrogram show which genes are regulated in a similar manner over the 24-h time course. For each dendrogram, the average fold change in regulation for all genes in the set was also calculated at each time point. The plots of the average fold change over time indicate the trends in regulation for each set of functionally related genes.




View larger version (86K):
[in this window]
[in a new window]
 
Fig. 4. Gene clustering dendrograms and average fold change graphs. Genes undergoing >=2-fold change in regulation are cluster analyzed. Each dendrogram contains genes, which have similar functionality and are grouped by temporal expression. Clustering within each dendrogram shows which genes are regulated in a similar manner over the 24-h time course. The average fold change for all genes in a dendrogram is plotted over 24 h to show the trend in regulation in functional groups. A: cell cycle control proteins, oncogenes, and tumor suppressors genes. B: ion channel, transport proteins, and stress response proteins. C: Apoptosis-related proteins and DNA synthesis, repair, and recombination proteins.  D: DNA binding/transcription/transcription factors. E: cell receptors (growth factor and chemokine receptors, interleukin/interferon receptors, hormone receptors, neurotransmitter receptors) and cell surface antigens and adhesion molecules. F: extracellular cell signaling and communication proteins (growth factors, cytokines, chemokines, interleukins, interferons, hormones).

 
Western blotting for protein product.
Since the induction of HSPs was expected, verification of the production of HSPs served, in part, as a positive control to validate the gene expression experiment. To determine whether protein was produced, Western blotting was performed on protein isolated from the phenol phase remaining after isolation of total RNA. It was expected that protein production would follow mRNA synthesis, and that was confirmed here for two genes that were highly upregulated, HSP70 and HSP40. Western blotting over a 24-h period indicated that protein for both was increased at 8 h and possibly remained elevated beyond 24 h (Fig. 5A). Normalized HSP70 and HSP40 values against glyceraldehyde-3-phosphate dehydrogenase on cDNA arrays showed that both were transcribed constitutively and that transcription was elevated by 4 h and decreased by 20 h (Fig. 5B). Also, normalization of protein values for both HSP70 and HSP40 showed a 33-fold and 7-fold elevation, respectively (Fig. 5C). Although the elevation of message peaked at 4 h and steadily fell over the next 24 h for both HSPs, the production of protein was elevated at 4 h and continually increased up to and possibly beyond 24 h (Fig. 6).



View larger version (30K):
[in this window]
[in a new window]
 
Fig. 5. Production of protein as determined by Western blotting. Translation of mRNA into protein was verified for two genes that were highly upregulated according to the arrays (HSP70 and HSP40). A: protein regulation over 24 h by Western blotting. In duplicate, HSP40 first becomes elevated at 4 h and continually increase up to and possibly beyond 24 h. Although protein for HSP70 is not produced constitutively, it is also initially elevated at 4 h and increases up to and possibly beyond 24 h. B: Atlas array cDNA regulation over 24 h. Duplicate hybridized cDNAs for HSP70 and HSP40 were normalized with glyceraldehyde-3-phosphate dehydrogenase (G3DPH). Both genes are transcribed constitutively, and transcription is elevated at 4 h, but slowly decreases over the next 20 h. C: normalized values for protein elevation or suppression (in boldface) for HSP70 and HSP40. Values were calculated from relative total protein ratios and IOD ratios. MDPF, 2-methoxy-2,4-diphenyl-3(2H)-furanone; CTL, control.

 


View larger version (14K):
[in this window]
[in a new window]
 
Fig. 6. Comparison in trends of transcriptional and translational regulation of the heat shock proteins (HSPs). A: the fold upregulation (positive values) or downregulation (negative values) of HSP70 mRNA (open bar) and protein (closed bar) from 1 to 24 h after thermal injury at 55°C for 3 s compared with unheated control. B: the fold upregulation (positive values) or downregulation (negative values) of HSP40 mRNA (open bars) and protein (solid bars) from 1 to 24 h after thermal injury at 55°C for 3 s compared with unheated control.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The breadth of information available from gene array profiling far exceeds any other biochemical or molecular biological technique currently available. The parallel and global nature of gene expression profiling with cDNA array profiling allows for a much more systematic evaluation of biological processes. Until approximately 5 years ago, the regulation of gene expression was performed one gene at a time. Although Northern blotting and RT-PCR can produce information similar to that of gene arrays, these techniques are only effective for a few genes at a time. Gene expression profiling allows for screening of multiple genes as a function of the array organization. Gene microarrays comprise hundreds to thousands of gene sequences physically tethered to a glass slide or nylon membrane in an array pattern. Each spot is very specific to complementary sequences of that particular gene. The Atlas array membranes also have each gene spotted in duplicate to verify reproducible binding specificity.

We have observed that for gene expression profiling following a potentially lethal insult, care must be taken to ensure that the RNA is representative of injured cells, but not dead cells. An overly severe injury results in RNA that is too degraded to be useful in gene expression profiling. The quality of RNA from cells heated up to 7 s at 55°C remained undegraded, as shown by agarose gel electrophoresis.

Another major advantage to array profiling, unachievable by other techniques, is the ability to assess expression of genes not yet known to be involved in a biological process (20). Gene clustering applications can then be utilized to group genes by their patterns of expression over time or after specific treatments (12). By grouping genes together in this manner, genes that are interdependent and function together can be visualized. Emerging techniques such as clustering allow analysis of large amounts of data generated from an experiment.

The set of genes presented in this report is from a representative experiment. We have performed this experiment five times with similar results, but considerable variability was observed between experiments, resulting in variability in ratios and number of genes up- or downregulated. This may have resulted from the nature of the injury, using different lots of arrays, or other unknown variables. However, many of the same genes were consistently altered and, more importantly, show similar patterns of gene expression in response to thermal injury. It is imperative that the "fold change" in gene expression is interpreted in relative terms as an indication of trends in expression, not as concrete numbers. Again, the ratios of fold change are not quantitative, but qualitative, and therefore can change with each experiment. Much of the variability may be due to the nature of the injury itself and the difficulty in inflicting a reproducible thermal injury. Furthermore, it is only logical to expect many different modes of response to any single type of insult. Therefore, although the set of genes induced by thermal injury may fluctuate between experiments, they must all be involved in the repair process, with the most essential genes induced more consistently and presumably to a greater extent.

Of 588 known human genes analyzed in this experiment, 120 exhibited a greater than twofold alteration in expression after sublethal thermal injury of ARPE-19 cells heated for 3 s at 55°C. As expected, the HSPs were highly elevated in response to thermal injury. Messages for the 70-kDa and 40-kDa HSPs were the two most highly elevated following injury. These proteins are also elevated in response to oxidative stress and may indicate one pathway by which thermal injury causes secondary damage (22, 26). The heat shock response, first described by Rotissa (39), is one of the best understood forms of induction of gene expression. Trimerization of the heat shock transcription factor (HSTF) yields a configuration that binds favorably to the heat shock response element and enhances transcription of RNA for several HSPs (35, 45). Constitutively expressed HSPs are postulated to act as molecular chaperones involved in protein folding, assembly, and transport. During stress, induced HSPs probably participate in an attempt by the cell to restore denatured proteins to their active configuration and therefore prevent the accumulation of aggregated, misfolded, and damaged proteins (15, 18, 34).

The presence of elevated message does not always ensure translation into product. One example of this is the disparity between the presence of mRNA for both HSP40 and HSP70 in control cells and the translation of protein product for HSP40 only. Message for HSP70 was present in control cells and peaked at 4 h, but protein was not detected until 4 h after heating, with a 33-fold elevation by 24 h (Fig. 5). On the other hand, HSP40 protein was present in the control but was elevated only sevenfold by 24 h. This indicated that HSP70 was translationally activated, whereas activation of constitutively expressed HSP40 may be regulated by a different mechanism (36).

The protective characteristics of HSPs are also being studied for therapeutic purposes. Aside from the ability to establish a state of thermotolerance following sublethal thermal injury, the induction of HSPs has been implicated in protection against ischemic damage (7, 26). The ability of HSPs to protect against stresses other than heat implies that the mechanism of induction is not specific for thermal insult or that there is more than one mechanism. In addition to heat damage, a variety of other insults such as trauma, ischemia, heavy metal, and alcohol can produce similar responses in cells. Other artificial means of induction, such as the utilization of prostaglandin A1 (3, 13), herbimycin A (18, 36), and sodium arsenite (27), are being explored for possible pharmacological application. By observing the specific pathways of induction and the resulting downstream changes in cellular function and expression, the mechanism of protection may be elucidated.

The DNA-binding protein TAX and the 153-kDa GADD gene were also highly elevated. The GADD stress response genes have generally been reported to be induced by ionizing and ultraviolet radiation or alkylating agents and are proposed to be regulated by distinct kinase-mediated pathways (9). Genes previously described as affecting growth and differentiation, including the retinoic acid binding protein (43), placenta growth factor, mda-6 gene (21), and the protein tyrosine kinase c-src (14) were highly elevated. Both the proliferation-associated gene (pag) (37) and neurotrophin-4 (49, 50) additionally have protective roles in reaction to oxidative or ischemic damage. The elevation of IL-12, which leads to the production of interferon-{gamma} and which is largely induced by bacterial infection, supports the manifestation of a general immune response resulting from thermal injury (10, 48).

One aspect of visible laser light injuries to the retina is that in most instances, the laser beam produces a Gaussian distribution of energy and does not produce a homogeneous population of sublethally injured cells. Lethal thermal injury occurs immediately to the RPE cell layer in the center of injury zones, whereas sublethal injury occurs to a subset of cells on the periphery of the injury (19, 32). Evaluation of the response of those small number of cells at the periphery is therefore difficult, but they are the relevant cells for staging the healing response and preventing additional expansion of the injury site. The dipping method of heating was therefore employed in this study to inflict a reproducible high-intensity, short-duration thermal injury as a model for the damage on the burn perimeter. Furthermore, during the early stages of the injury and prior to the onset of wound healing, sublethally injured cells may either undergo accidental cell death or apoptosis (47). However, factors regulating these processes are unclear. Thermal damage induced by laser irradiation will have similar effects on other tissue types, and this type of tissue response is necessary to assess as new applications of lasers are found in medicine.

The cellular response to sublethal thermal insult has been highly conserved to promote both cellular and organism survival. All cells live within relatively narrow tolerances for heat and when subjected to elevated temperature adaptively respond to protect themselves. The cells may survive up to some point, but then may either die or recover. In multicellular organisms, the evolution of thermal homeostasis was an important advance for survival. Because the brain precisely regulates temperature, optimizing function over a very narrow range, a form of thermal insult occurs naturally in the form of fever, with prostaglandins elevating temperature as a physiological response to invasion by foreign substances (13, 41). The RPE cell may also have adapted to tolerate somewhat higher temperature levels by the fact that photon energy is converted to heat at intracellular sites of melanin accumulation.

Interestingly, many of the genes upregulated by heat shock have also recently been reported to be upregulated during normal aging, indicating the accumulation of denatured protein with age and its acceleration by heating (25). Although the entire significance of discovered alterations in gene expression is not yet clear, by observing the changes in metabolic processes due to thermal stress, we may better understand the benefits and targets of drug treatment for secondary damage associated with thermal injury. Based on gene expression profile data, potential therapeutics will be chosen depending on their ability to restore the genetic responses to a more normal pattern of expression or to reduce deleterious genetic responses to thermal injury.


    ACKNOWLEDGMENTS
 
The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: P. D. Bowman, US Army Institute of Surgical Research, 3400 Rawley E. Chambers Ave., Bldg. 3611, Fort Sam Houston, TX 78234-6315 (E-mail: phillip.bowman{at}amedd.army.mil).


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Ahmed SA, Gogal RM Jr, and Walsh JE. A new rapid and simple nonradioactive assay to monitor and determine the proliferation of lymphocytes: an alternative to [3H]thymidine incorporation assay. J Immunol Methods 170: 211–224, 1994.[ISI][Medline]
  2. Alba F and Daban J. Rapid fluorescent monitoring of total protein patterns on sodium dodecyl sulfate-polyacrylamide gels and western blots before immunodetection and sequencing. Electrophoresis 19: 2407–2411, 1998.[ISI][Medline]
  3. Amici C, Palamara AT, and Santoro MG. Induction of thermotolerance by prostaglandin A in human cells. Exp Cell Res 207: 230–234, 1993.[ISI][Medline]
  4. Back SA, Khan R, Gan X, Rosenberg PA, and Volpe JJ. A new Alamar Blue viability assay to rapidly quantify oligodendrocyte death. J Neurosci Methods 91: 47–54, 1999.[ISI][Medline]
  5. Barbanel CS, Ducatman AM, Garston MJ, and Fuller T. Laser hazards in research laboratories. J Occup Med 35: 369–374, 1993.[ISI][Medline]
  6. Bowman PD, Schuschereba ST, Lawlor DF, Gilligan GR, Mata JR, and DeBaere DR. Survival of human epidermal keratinocytes after short-duration high temperature: synthesis of HSP70 and IL-8. Am J Physiol Cell Physiol 272: C1988–C1994, 1997.[Abstract/Free Full Text]
  7. Bush KT, Goldberg AL, and Nigam SK. Proteasome inhibition leads to a heat-shock response, induction of endoplasmic reticulum chaperones, and thermotolerance. J Biol Chem 272: 9086–9092, 1997.[Abstract/Free Full Text]
  8. Carper SW, Duffy JJ, and Gerner EW. Heat shock proteins in thermotolerance and other cellular processes. Cancer Res 47: 5249–5255, 1987.[Abstract]
  9. Carrier F, Zhan Q, Alamo I, Hanaoka F, and Fornace AJ Jr. Evidence for distinct kinase-mediated pathways in gadd gene responses. Biochem Pharmacol 55: 853–861, 1998.[ISI][Medline]
  10. Chan SH, Kobayashi M, Santoli D, Perussia B, and Trinchieri G. Mechanisms of IFN-gamma induction by natural killer cell stimulatory factor (NKSF/IL-12). Role of transcription and mRNA stability in the synergistic interaction between NKSF and IL-2. J Immunol 148: 92–98, 1992.[Abstract/Free Full Text]
  11. Dunn KC, Aotaki-Keen AE, Putkey FR, and Hjelmeland LM. ARPE-19, a human retinal pigment epithelial cell line with differentiated properties. Exp Eye Res 62: 155–169, 1996.[ISI][Medline]
  12. Eisen MB, Spellman PT, Brown PO, and Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95: 14863–14868, 1998.[Abstract/Free Full Text]
  13. Elia G, Amici C, Rossi A, and Santoro MG. Modulation of prostaglandin A1-induced thermotolerance by quercetin in human leukemic cells: role of heat shock protein 70. Cancer Res 56: 210–217, 1996.[Abstract]
  14. Erpel T and Courtneidge SA. Src family protein tyrosine kinases and cellular signal transduction pathways. Curr Opin Cell Biol 7: 176–182, 1995.[ISI][Medline]
  15. Georgopoulos C and Welch WJ. Role of the major heat shock proteins as molecular chaperones. Annu Rev Cell Biol 9: 601–634, 1993.[ISI]
  16. Gerner EW and Manning MR. In: Hyperthermia in Modern Radiation Oncology, edited by Gilbert HA. Philadelphia, PA: Harper and Row, 1984, p. 571–580.
  17. Gerweck LE. Hyperthermia in cancer therapy: the biological basis and unresolved questions. Cancer Res 45: 3408–3414, 1985.[ISI][Medline]
  18. Hegde RS, Zuo J, Voellmy R, and Welch WJ. Short circuiting stress protein expression via a tyrosine kinase inhibitor, herbimycin A. J Cell Physiol 165: 186–200, 1995.[ISI][Medline]
  19. Humphrey MF, Parker C, Chu Y, and Constable IJ. Transient preservation of photoreceptors on the flanks of argon laser lesions in the RCS rat. Curr Eye Res 12: 367–372, 1993.[ISI][Medline]
  20. Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JCF, Trent JM, Staudt LM, Hudson J Jr, Boguski MS, Lashkari D, Shalon D, Botstein D, and Brown PO. The transcriptional program in the response of human fibroblasts to serum. Science 283: 83–87, 1999.[Abstract/Free Full Text]
  21. Jiang H, Lin J, Su ZZ, Herlyn M, Kerbel RS, Weissman BE, Welch DR, and Fisher PB. The melanoma differentiation-associated gene mda-6, which encodes the cyclin-dependent kinase inhibitor p21, is differentially expressed during growth, differentiation and progression in human melanoma cells. Oncogene 10: 1855–1864, 1995.[ISI][Medline]
  22. Keyse SM and Emslie EA. Oxidative stress and heat shock induce a human gene encoding a protein-tyrosine phosphatase. Nature 359: 644–647, 1992.[ISI][Medline]
  23. Langer R and Vacanti JP. Tissue engineering. Science 260: 920–926, 1993.[ISI][Medline]
  24. Larson EM, Doughman DJ, Gregerson DS, and Obritsch WF. A new, simple, nonradioactive, nontoxic in vitro assay to monitor corneal endothelial cell viability. Invest Ophthalmol Vis Sci 38: 1929–1933, 1997.[Abstract]
  25. Lee CK, Klopp RG, Weindruch R, and Prolla TA. Gene expression profile of aging and its retardation by caloric restriction. Science 285: 1390–1393, 1999.[Abstract/Free Full Text]
  26. Lewden O, Garcher C, Assem M, Morales C, Rochette L, and Bron AM. Changes of the inducible heat shock protein 70 mRNA level in rat retina after ischemia and reperfusion. Ophthalmic Res 30: 291–294, 1998.[ISI][Medline]
  27. Li GC. Induction of thermotolerance and enhanced heat shock protein synthesis in Chinese hamster fibroblasts by sodium arsenite and by ethanol. J Cell Physiol 115: 116–122, 1983.[ISI][Medline]
  28. Mainster MA. Decreasing retinal photocoagulation damage: principles and techniques. Semin Ophthalmol 14: 200–209, 1999.[Medline]
  29. Mainster MA, White TJ, Tips JH, and Wilson PW. Retinal-temperature increases produced by intense light sources. J Opt Soc Am A 60: 264–270, 1970.
  30. Marshall J. Structural aspects of laser-induced damage and their functional implications. Health Phys 56: 617–624, 1989.[ISI][Medline]
  31. Marshall J. Thermal and mechanical mechanisms in laser damage to the retina. Invest Ophthalmol 9: 97–115, 1970.[ISI][Medline]
  32. Marshall J and Mellerio J. Pathological development of retinal laser photocoagulations. Exp Eye Res 6: 303–308, 1967.[ISI][Medline]
  33. Matylevitch NP, Schuschereba ST, Mata JR, Gilligan GR, Lawlor DF, Goodwin CW, and Bowman PD. Apoptosis and accidental cell death in cultured human keratinocytes after thermal injury. Am J Pathol 153: 567–577, 1998.[Abstract/Free Full Text]
  34. Minowada G and Welch WJ. Clinical implications of the stress response. J Clin Invest 95: 3–12, 1995.[ISI][Medline]
  35. Morimoto RI. Cells in stress: transcriptional activation of heat shock genes. Science 259: 1409–1410, 1993.[ISI][Medline]
  36. Morris SD, Cumming DV, Latchman DS, and Yellon DM. Specific induction of the 70-kD heat stress proteins by the tyrosine kinase inhibitor herbimycin-A protects rat neonatal cardiomyocytes. A new pharmacological route to stress protein expression? J Clin Invest 97: 706–712, 1996.[Abstract/Free Full Text]
  37. Prosperi MT, Ferbus D, Rouillard D, and Goubin G. The pag gene product, a physiological inhibitor of c-abl tyrosine kinase, is overexpressed in cells entering S phase and by contact with agents inducing oxidative stress. FEBS Lett 423: 39–44, 1998.[ISI][Medline]
  38. Pruitt B, Goodwin CW, and Pruitt SK. Burns: Including Cold, Chemical, and Electrical Injuries. Philadelphia, PA: Saunders, 1991.
  39. Ritossa F. A new puffing pattern induced by temperature shock and DNP in Drosophila. Experentia 18: 571–573, 1962.
  40. Rockwell JR. Laser accidents: reviewing thirty years of incidents: what are the concerns: old and new? J Laser Applic 6: 203–211, 1994.[ISI]
  41. Rossi A and Santoro MG. Induction by prostaglandin A1 of haem oxygenase in myoblastic cells: an effect independent of expression of the 70 kDa heat shock protein. Biochem J 308: 455–463, 1995.[ISI][Medline]
  42. Schuschereba S. Current therapy for laser-induced retinal injury: overview of clinical and experimental approaches. In: Laser and Noncoherent Ocular Effects: Epidemiology, Prevention, and Treatment, edited by Stuck BE and Belkin M. Proceedings of SPIE, 1997, vol. 2974, p. 171–188.
  43. Si SP, Tsou HC, Lee X, and Peacocke M. Effect of cellular senescence and retinoic acid on the expression of cellular retinoic acid binding proteins in skin fibroblasts. Exp Cell Res 219: 243–248, 1995.[ISI][Medline]
  44. Smith R. Lasers and vision: high-powered revision. Occup Health Saf 63: 88–91, 1994.[Medline]
  45. Sorger PK. Heat shock factor and the heat shock response. Cell 65: 363–366, 1991.[ISI][Medline]
  46. Springer JE, Azbill RD, and Carlson SL. A rapid and sensitive assay for measuring mitochondrial metabolic activity in isolated neural tissue. Brain Res Brain Res Protoc 2: 259–263, 1998.[ISI][Medline]
  47. Trump BF, Berezesky IK, Chang SH, and Phelps PC. The pathways of cell death: oncosis, apoptosis, and necrosis. Toxicol Pathol 25: 82–88, 1997.[ISI][Medline]
  48. Valiante NM, Rengaraju M, and Trinchieri G. Role of the production of natural killer cell stimulatory factor (NKSF/IL-12) in the ability of B cell lines to stimulate T and NK cell proliferation. Cell Immunol 145: 187–198, 1992.[ISI][Medline]
  49. Vecino E, Caminos E, Ugarte M, Martin-Zanca D, and Osborne NN. Immunohistochemical distribution of neurotrophins and their receptors in the rat retina and the effects of ischemia and reperfusion. Gen Pharmacol 30: 305–314, 1998.[Medline]
  50. von Bartheld CS. Neurotrophins in the developing and regenerating visual system. Histol Histopathol 13: 437–459, 1998.[ISI][Medline]
  51. Welch WJ. Mammalian stress response: cell physiology, structure/function of stress proteins, and implications for medicine and disease. Physiol Rev 72: 1063–1081, 1992.[Free Full Text]