| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Thermal and Mountain Medicine Division, US Army Research Institute of Environmental Medicine, Natick 01760; and 2 Division of Pulmonary and Critical Care Medicine and 3 Partners Gene Array Technology Center, Brigham and Women's Hospital/Harvard Medical School, Boston, Massachusetts 02115
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
We studied the effect of heat shock on gene expression by normal human cells. Peripheral blood mononuclear cells (PBMCs) were obtained from healthy adults. Paired samples from each subject were subjected to either 20 min of heat shock (43°C) or control (37°C) conditions and then returned to 37°C. RNA was isolated 160 min later, and five representative samples were analyzed on Affymetrix gene chip arrays containing ~12,600 probes. A biologically meaningful effect was defined as a statistically significant, twofold or greater difference in expression of sequences that were detected in all five experiments under control (downregulated sequences) or heat shock (upregulated sequences) conditions. Changes occurred in 395 sequences (227 increased by heat shock, 168 decreased), representing 353 Unigene numbers, in every functional category previously implicated in the heat shock response. By RT-PCR, we confirmed the findings for one upregulated sequence (Rad, a G protein) and one downregulated sequence (osteopontin, a cytokine). We conclude that heat shock causes extensive gene expression changes in PBMCs, affecting all functional categories of the heat shock response.
apoptosis; gene chip array technology; cell stress response
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
THE EFFECT OF ACUTE HEAT SHOCK on cells can range from survival and adaptation (usually with acquired thermotolerance) to apoptosis and, in the extreme, necrosis (5, 22, 41). These effects of heat on the cell can occur as a direct effect of heat itself on cellular constituents (for example, the denaturation of proteins) or as a secondary effect, that is, mediated by cellular adaptive mechanisms such as the expression of heat shock proteins (HSPs) and the activation of selected signaling pathways. These adaptive mechanisms can be further divided into those that involve a change in the activity of existing proteins (such as by phosphorylation/dephosphorylation or protein-protein interactions) and those that involve a change in protein expression.
The effects of moderate heat on cellular structure and function are well established (26, 27), although much of the knowledge on general cellular physiological effects of heat shock in eukaryotes has been derived from studies in yeast, Drosophila melanogaster, and tissue culture cell lines. These effects include 1) activation of the transcription factor heat shock factor-1 (HSF-1), which increases transcription of HSPs (32); 2) an otherwise generalized early inhibition of gene expression (27, 31) through effects on transcription (27), disruption of RNA processing (especially splicing) (3, 27, 51), and inhibition of translation (39); 3) changes in the intracellular distribution of certain proteins [such as the migration of HSF-1 from cytoplasm to nucleus (32)]; 4) protein degradation, especially through ubiquitin-proteasome (30, 40) and lysosomal pathways (49); 5) changes in the activity of a variety of kinases and phosphatases [especially in the mitogen-activated protein kinase (MAPK) pathways] (9, 16, 17, 36); 6) alterations in membrane function leading to increases in intracellular calcium, proton, and sodium concentrations (13, 25, 45, 48); 7) a decrease in cellular ATP concentration (48); and 8) alteration of components of the cytoskeleton (16). In response to heat, cells have evolved a stress response that includes a large increase in cellular levels of HSPs, whose chaperonin function promotes protein refolding and maintenance of native conformational state (15, 22); expression of molecules that are involved in redox homeostasis, such as heme oxygenase-1 (10, 11, 29, 38) and Cu,Zn superoxide dismutase (50); arrest of cell cycle progression (19, 26); and activation of antiapoptotic mechanisms (5, 41). The net physiological effect of these adaptations is generally to render the cell more tolerant to subsequent thermal challenges and cross-tolerant to other noxious stimuli, such as certain toxins (27, 33). In addition, certain effects of heat are cell-type specific, such as temperature-mediated effects on immune cell function (18, 33, 34).
The full extent to which proteins other than the HSPs undergo changes in expression as a result of heat shock has only recently become amenable to experimental investigation, due to the advent of gene chip array technology. This technology has made it possible to screen thousands of sequences in a single experiment to search for evidence of changes in gene expression. An early paper that used this technology screened ~1,000 genes before and after heat shock in human lymphocytes (44). The authors confirmed the effect of heat on expression of HSPs and demonstrated the potential of the technology to identify novel sequences that are affected by pathophysiological stimuli. The power and sophistication of gene chip arrays have since grown 10-fold, and it is now possible to analyze ~12,600 gene sequences on a single chip. More recently, a gene chip array containing 588 genes was used to identify the response of retinal pigment epithelial cells to a sublethal heat shock (8). Cells exposed to 55°C for 3 s subsequently showed changes in expression of twofold or greater in 120 sequences, many of which were DNA transcription factors and other regulatory proteins. In mouse testis (43), heat shock led to changes in expression of 176 genes, many of which are relevant to pathways known to be involved in the cell stress response.
The aforementioned studies suggest that the genomic response of mammalian cells to thermal stress is considerably more complex than previously recognized. However, a large-scale analysis of the effects of heat shock on normal human cells has not yet been reported, to our knowledge. The purpose of this study was to characterize the genomic response of normal human peripheral blood mononuclear cells (PBMCs) to a conventional acute heat shock stimulus (43°C for 20 min, followed by a return to 37°C) with the use of a late-generation human sequence gene chip array containing ~12,600 sequences. In addition to being easily obtained, PBMCs are part of the one tissue (blood) that is exposed to all body compartments during in vivo hyperthermia; thus we felt this to be a rational choice of tissue for this in vitro study. Our hypothesis was that gene chip array technology would identify genes not previously known to be part of the cellular response to acute heat shock.
| |
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Volunteers.
Adult subjects were recruited without regard to age, ethnic origin, or
gender. With the exception of the experiment illustrated in Fig.
1, which was performed separately on a
sample obtained from a 37-yr-old male Caucasian to illustrate the time
course of HSP expression (and was therefore not included in the
statistical analysis presented here), all 20 subjects reported in this
study donated samples between May 23, 2000 and June 29, 2000 in Natick, Massachusetts, well before the onset of summer temperatures capable of
inducing heat acclimatization.
|
PBMC preparation and heat shock. Blood samples were drawn from subjects in the morning, between 8:00 AM and 10:00 AM. Human mononuclear cells were obtained from 180 ml of EDTA anti-coagulated peripheral blood with the use of Lymphoprep (Nycomed, Oslo, Norway) density gradient centrifugation according to the instructions of the manufacturer. This procedure yielded a mean (±SD) of 1.4 ± 0.4 × 108 cells (n = 20), with a PBMC purity of 95% [83 ± 12% (SD) lymphocytes and 12 ± 5% monocytes] as determined by Coulter counter classification (Onyx analyzer, Coulter, Miami, FL). PBMCs were resuspended in 6 ml of RPMI 1640 and buffered with HEPES to pH 7.4 and osmolality of 285 mosmol/kgH2O (GIBCO) in 15-ml polypropylene tubes (Beckton-Dickenson Laboratories, Lincoln Park, NJ) at a concentration of 1 × 107 to 1.5 × 107 cells/ml. Cells in this suspension were equilibrated in a water bath at 37°C for 30 min and then subjected to 43°C for 20 min (heat shock) or maintained at 37°C (control). Cells were harvested by centrifugal pelleting at 3 and 4 h after the start of heat stress or control exposure. In six preliminary experiments in which volunteers of both genders were used, cell viability by trypan blue exclusion at 4 h after initiation of heat shock was 95 ± 0.8% (SE) in control cells and 93 ± 1.3% in heat-shocked cells (P = 0.18 by unpaired t-test).
RNA purification. RNA was extracted from cells at the 3-h time point by using the RNeasy method, according to the directions of the manufacturer (Qiagen, Valencia, CA). The quality of the extracted RNA was assessed as recommended by Farrell (12) by absorption spectroscopy in the range of 230-320 nm (to ensure that the peak of absorption did in fact occur at 260 nm and to identify evidence of contamination) and by examination of the 18S and 28S bands on gel electrophoresis (to exclude samples with band smearing indicative of degradation).
Measurement of intracellular HSP70. Intracellular HSP70 was measured at the 3- and 4-h time points by using a commercially available ELISA kit (StressGen, Vancouver, BC, Canada), following the directions of the manufacturer. The results were normalized to number of cells, as determined by Coulter counter enumeration.
Gene chip array. Samples were considered suitable for gene chip array analysis only 1) if heat-shocked cells showed expression of HSP70 protein at 3 or 4 h that was greater than or equal to twofold of the value measured in the control cells and 2) if the RNA was of suitable yield (a minimum of 25-35 µg) and quality, as judged by examination of the 18S and 28S bands on gel electrophoresis and by absorbance spectroscopy in the range of 240-320 nm. Of the 11 subjects from whom both RNA and protein were obtained, 9 met our criteria for inclusion by differences in HSP70 levels. Of these nine, two were excluded because of low-RNA yields and one was excluded because the quality of RNA was judged to be suboptimal for further analysis. Five of the six remaining samples were from male subjects, and these were subjected to gene chip array analysis.
Transcript profiling with Affymetrix gene chips (Affymetrix, Santa Clara, CA) was performed by using HG-U95A chips containing ~12,600 sequences (representing ~11,500 unique GenBank accession numbers), as detailed elsewhere (20, 28). Briefly, 8 µg of total RNA were reverse-transcribed with an oligo(dT) primer coupled to a T7 RNA polymerase binding site. Biotinylated complementary RNA (cRNA) was then synthesized from the resulting complementary DNA (cDNA) with the use of T7 polymerase. Twenty-five micrograms of biotinylated cRNA were then randomly sheared (to an approximate length of 50 nucleotides) and hybridized for 16 h to the Affymetrix gene chips. The hybridized gene chips were incubated with phycoerythrin-streptavidin and washed, and the signal was further amplified by incubating the chips with a polyclonal anti-streptavidin antibody coupled to phycoerythrin. Hybridization to the array was quantified with a Hewlett-Packard gene array laser scanner. First-pass analysis of the scanned data was performed with GeneChip software (version 3.3). In each experiment, external standards (cRNA transcribed from several cloned bacterial genes) were included to control for hybridization efficiency, test for sensitivity, and assist in the comparisons between data sets from different experiments. Furthermore, before hybridization on the HG-U95A chips, cRNAs were first hybridized to a Test II Chip (Affymetrix), to ensure the quality of the preparation.Confirmatory RT-PCR.
Each RNA sample used for gene chip array analysis was also subjected to
a separate poly(T)-primed RT-PCR by using a commercially available kit
(Retrocript first-strand synthesis kit, Ambion, Austin, TX). Each
resulting mixture of cDNA and unreacted primers was diluted to 50 ng/µl, and 100-ng aliquots were subjected to 30 cycles (Rad, HSP70B',
-actin) or 35 cycles (osteopontin) of PCR at an annealing
temperature of 60°C, with primers designed specifically to
recognize HSP70B', -actin, Rad, and osteopontin. The -actin
primers were obtained from a commercial source (Clonetech, Palo
Alto, CA), and the other primers were designed with PRIMER-3 software. The sequences of these primers were as follows: HSP70B', 5'-AGGAGATCTCGTCCATGGTG-3' (forward) and 5'-TTCCATGAAGTGGTTCACGA-3' (reverse), designed to yield a 380-bp amplicon; Rad,
5-GGGGATGCCTATGTCATTGT-3' (forward) and 5'-CTGTTACGAGCTACGATGCG-3'
(reverse), designed to yield a 386-bp amplicon; osteopontin,
5'-CATCACCTGTGCCATACCAG-3' (forward) and 5'-AACCACACTATCACCTCGGC-3'
(reverse), designed to yield a 415-bp amplicon.
Data analysis.
Data analysis was performed with Microsoft Excel, Microsoft Access,
SigmaStat 2.0 for Windows, and SPSS 10.0 for Windows. The multiples of
change (fold-change) in gene expression reported by the Affymetrix
software in the paired experiments were used to determine whether a
statistically significant change in expression had occurred. Because
there is a discontinuity in the possible values reported by the
software in the range 
1 to +1, the reported multiples of change were
first transformed by taking the natural logarithm of all values greater
than or equal to +1 [i.e., ln(x)] and the natural
logarithm of the absolute value of the reciprocal of all values less
than or equal to 1 [i.e., ln(1/|x|)]. This transformation produced a continuous distribution with a mode of zero,
with values greater than zero representing upregulation and values less
than zero representing downregulation. The 95% confidence intervals
were computed by adding (or subtracting) to the mean the product of
standard error times the two-tailed t-distribution factor
for P = 0.05 with four degrees of freedom. Means and
95% confidence intervals on the transformed data for each sequence
were computed, and a gene was considered to have undergone a
statistically significant change in expression if the 95% confidence
intervals of the mean excluded zero. Because of the data transform
used, the means reported in this manuscript therefore represent
geometric, not arithmetic, means.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Study population. Samples suitable for protein analysis were obtained from 20 subjects, and RNA was obtained from 11 subjects. Of these, 7 were women and 13 were men. The age range of the subjects was 19-58 yr [39 ± 10 (SD) yr, median of 39 yr]. All but two of the subjects were of Caucasian origin. Five of the six samples that met our criteria for RNA gene chip array analysis (i.e., a twofold or greater increase in HSP70 expression at any time after heat shock and RNA of sufficient quantity and quality) were from men (ages 29-46), and these were submitted for gene chip array analysis. Four of these subjects were active-duty Army personnel and were required to maintain a level of physical fitness sufficient to pass the Army Physical Fitness Test. The fifth subject (subject 4) in this cohort had retired from the military but reported exercising at least three times per week.
Expression of HSP70 protein. Figure 1 shows a representative experiment that illustrates the time course of HSP70 protein expression. After initial equilibration, cells were subjected to either a conventional heat shock (43°C × 20 min) or maintained at 37°C. The heat-shocked cells were then returned to 37°C, and both heat-shocked and control cells were assayed for intracellular levels of HSP70 by ELISA over the next 6 h. We found that maximum expression of intracellular HSP70 protein occurred between 4 and 6 h after the initiation of heat shock. We therefore chose to assess changes in RNA expression at 3 h after the initiation of heat shock.
The mean ± SE levels of HSP70 protein expression were 4.2 ± 0.6 ng/106 cells at 3 h (n = 20) and 4.4 ± 0.6 ng/106 cells at 4 h (n = 20) in the control cells. In the cells subjected to heat shock, these levels were 8.3 ± 1.0 ng/106 cells at 3 h and 10.6 ± 1.1 ng/106 cells at 4 h. The differences in HSP70 protein expression between control and heat-shocked cells were statistically significant by two-way, repeated-measures ANOVA (P < 0.001). The magnitude of the differences in HSP70 levels between control and heat-shocked cells was also affected by time. Two-way, repeated-measures ANOVA showed a statistically significant interaction between the time at which measurement was taken (3 vs. 4 h) and the difference between control and heat-shocked cells (P = 0.028). The mean ± SE change in HSP70 expression as a result of heat shock was 2.4 ± 0.3-fold at 3 h (n = 20) and 2.7 ± 0.2-fold (n = 20) at 4 h. In 15 of the 20 subjects tested, there was a twofold or greater expression of HSP70 protein in the cells subjected to heat shock at 3 h, 4 h, or both. Of the five subjects who showed a less than twofold increase in HSP70 expression as a result of heat shock, three had increases of between 1.9- and 2.0-fold, and the remaining two had increases between 1.5- and 1.7-fold. Figure 2 shows the changes in HSP70 protein expression that occurred in the five individual samples that were subjected to gene chip array analysis. At 3 h, the mean ± SE level of HSP70 expression in these cells was 3.4 ± 0.7 ng/106 cells in the control cells and 8.4 ± 1.6 ng/106 cells in the heat-shocked cells. At 4 h, these values were 3.1 ± 0.5 ng/106 cells and 10.4 ± 0.8 ng/106 cells, respectively. The mean ± SE change in expression was 2.6 ± 0.3-fold at 3 h and 3.6 ± 0.6-fold at 4 h. A two-way, repeated-measures ANOVA of these data [with heat shock vs. control as the within-subject factor and time (3 vs. 4 h) as the between-subject factor] revealed that there was a significant difference in expression between control cells and heat-shocked cells (P < 0.001). At both the 3- and 4-h time points, in both control and heat-shocked cells, the levels of HSP70 protein expression in the 5 samples submitted for gene chip array analysis were not significantly different from those in the 15 samples that were not (P > 0.05 by unpaired t-tests, equal variances not assumed).
|
Expression of HSP70B' RNA.
To ensure that our choice of 3 h after initiation of heat shock
was an adequate time point at which to assess cells for changes in RNA
expression, we examined RNA for expression of HSP70B' (HSPA6), a highly
inducible member of the HSP70 family. The results of RT-PCR analysis
are illustrated in Fig. 3 and demonstrate
that HSP70B' expression was barely detectable in the control cells but
was strongly expressed in cells that had been subjected to heat shock.
In contrast, -actin was strongly expressed in all five samples under
both control and heat shock conditions, and there was no consistent
effect of heat shock on the apparent level of expression of this gene
(Fig. 3, bottom).
|
Gene chip array results. Of the 12,626 sequences on the chip, 7,142 (56%) were identified as present or marginal by the chip-reading software in at least one experiment in the control cells and 7,112 (56%) in the heat-shocked cells. However, only about half of these ~7,100 sequences (3,711 in the control cells and 3,677 in the heat-shocked cells) were present or marginal in all five experiments.
A total of 395 sequences (representing 375 unique GenBank numbers and 353 UniGene numbers) met our criteria for a biologically meaningful change in expression. This number was derived as follows: 2,903 sequences (including sequences detected as absent in one or more experiments) showed a statistically significant difference in expression between the control cells and heat-shocked cells (1,311 were decreased as a result of heat shock and 1,592 were increased). Of these 2,903 sequences, 1,362 (547 decreased and 815 increased sequences) were excluded from further analysis because they were identified as absent in at least one of the five experiments either under control conditions (for the decreased sequences) or after heat shock (for the increased sequences). It is noteworthy that, of the sequences excluded because of absent calls, about three-fourths (394 in the control cells, 603 in the heat-shocked cells) were absent in more than half of the experiments. Finally, another 1,146 sequences (596 decreased and 550 increased sequences) were excluded because of changes in expression of less than twofold. Of the 395 sequences that met our criteria, 227 were increased by heat shock and 168 were decreased. Slightly more than half of these sequences fell into one of the following functional categories: HSPs/chaperonins/cochaperonins (45 sequences plus heme oxygenase-1, Table 1); immune function (54 sequences); signal transduction (45 sequences); cell growth, proliferation, and differentiation (31 sequences); metabolism (31 sequences); and transcription (22 sequences). Another 45 sequences of unknown identity were affected by heat shock (representing 11% of all sequences affected). Several of these were among the most highly affected genes in the experiment (Table 3).
|
Expression of HSPs and housekeeping genes. To validate the statistical and biological criteria that we applied to our gene chip array data, we specifically examined genes that were expected to show a change in expression as a result of heat shock (HSPs and chaperonins) and genes that were not expected to show a change in expression (housekeeping genes, sequences that are generally highly expressed in most cell types and that rarely show a change in expression under most experimental conditions).
Table 1 lists the sequences for HSPs and chaperonins that were affected by heat shock, according to our criteria. At least one sequence in each of the major HSP families was increased by heat shock. Of the sequences in this functional family included in our analysis, the smallest mean increase was 2.3-fold. Only four mRNAs in this class were decreased: oxygen-regulated protein 150, a member of the HSP70 family;
2-macroglobulin receptor-associated protein; and
tubulin-specific cofactors C and D.
By contrast, with one exception, we found no statistically significant
changes in the expression of several housekeeping gene sequences (Table
2). The sole exception was the 5' end of
GAPDH, which showed a statistically significant, 1.3-fold increase in expression as a result of heat shock; this was, however, below the
2.0-fold criterion we required for a change to be considered biologically meaningful. Furthermore, when we examined the 50 sequences
most highly expressed in the control cells (which included many
ribosomal proteins, -actin, laminin receptor-1, and
2-microglobulin), none showed any statistically
significant changes as a result of heat shock.
|
|
Effect of heat shock on the expression of genes other than HSPs/chaperonins. Tables 3 lists genes other than HSPs that were affected by heat shock. This table includes all sequences whose expression was increased by 4-fold or more, decreased by at least 0.3-fold, or otherwise judged to be of interest (the entire database can be obtained by contacting the corresponding author). The genes have been assigned to functional categories, on the basis of known or presumed biological functions, with the recognition that several could justifiably be classified into more than one of the categories listed. It is noteworthy that the genes identified as being affected by heat shock fell into every major functional category previously identified as playing a role in the cellular response to heat stress.
Genes involved in signal transduction and transcription accounted for one-seventh of all sequences affected by heat shock in this experiment. Among these were several known to be involved in or to modulate the activity of MAPK pathways, including dual-specificity phosphatase (DUSP) 1 (2.4-fold), DUSP6/MAPK phosphatase 3 (0.36-fold), DUSP8 (12-fold), and hematopoietic protein tyrosine phosphatases (HePTPase) (2.3-fold) (Table 3). Furthermore, transcription factor jun, which is known to be phosphorylated by stress-activated protein kinases (47), was also significantly increased (2.6-fold, Table 3). We saw no significant effect of heat shock on the expression of the transcription factors HSF-1-4 at the time point examined. Although we saw effects on transcription factors jun and fosB (a member of the fos family of transcription factors), we saw no effect on c-fos itself (mean change of 0.90-fold, 95% confidence interval of 0.38- to 2.1-fold).Expression of Rad and osteopontin.
To further validate our gene chip array findings, we chose to confirm
one increased and one decreased gene by RT-PCR. Our choices were
dictated by the limitations of RT-PCR, and we chose genes that on gene
chip array analysis had almost undetectable expression under one
condition but a high level of expression under the other. The results
are shown in Fig. 4 and show that, in
keeping with the gene chip array findings, expression of Rad was
increased by heat shock, whereas osteopontin was decreased.
|
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Heat shock appears to produce extensive changes in gene expression that, at the time point examined, are nevertheless confined to only a few functional categories. In the experiments reported, the number of sequences whose expression was altered (equal to 395) represents only 11% of the ~3,700 transcripts consistently expressed. Over half of these could be assigned to one of six major functional categories: 1) HSPs and chaperonins, 2) immune function, 3) signal transduction, 4) cell growth, proliferation, and differentiation, 5) metabolic enzymes, and 6) transcription.
The number of sequences that we report as being affected by heat shock probably represents a lower limit estimate of the total, given our very stringent criteria for accepting a change in gene expression as biologically meaningful, which we applied in an effort to minimize false-positive results. Statistically speaking, a highly stringent strategy to minimize false-positive reports will increase the number of false-negative results. It is therefore quite possible that our post hoc filtering process excluded a number of genes that are in fact affected by heat shock.
It is also noteworthy that there were a large number of sequences that
were identified by the chip-reading software as absent in one or more
experiments but that also showed statistically significant changes in
expression. This apparent incongruity may in part be explained by the
fact that the chip-reading software call is based on the intensity of
fluorescence in pixels containing perfectly matched sequences, compared
with pixels containing sequences with a mismatched base pair. Thus
statistically significant changes in absent genes might represent
changes in genes of sequence that are similar to, but not identical to,
those probed for in each pixel. Furthermore, some of these absent genes
may well be significantly altered by heat stress but expressed at very
low levels within the cell, below levels that yield a sufficiently
large signal-to-noise ratio for the current gene chip array technology
to identify them as present or marginal. Finally, some of the genes
excluded by our criteria probably are in fact affected by heat shock
but were excluded because of our use of presence calls in the post hoc filtering. In principle, a gene that showed a statistically significant increase in expression by heat shock but was detected in only four of
five heat-shocked cells would have been excluded. An example of this is
-tubulin 5 (GenBank accession no. X00734), which was significantly
increased (9.2-fold, 95% confidence interval of 3.1-27) but was
detected as absent in one of the five heat-shocked cell preparations
and was therefore excluded.
There was variation in the total number of genes expressed from one preparation to the next. This variation likely arises from several causes: 1) because PBMCs are a heterogeneous population of cells that contain monocytes, B cells, and various subsets of T cells in addition to any contaminating granulocytes; 2) because samples were obtained from five different individuals; 3) because there is experimental error associated with gene chip array technology (as with any technology); and 4) because the ability of any algorithm to accurately distinguish present from absent calls becomes increasingly difficult as levels of expression approach background noise.
Despite these limitations, we were able to validate our findings by applying both external standards (specifically, by examining genes that would be expected to respond to heat shock as well as genes that would not) as well as internal controls (by confirmatory RT-PCR). We found that every major family of HSPs showed changes in expression at the time point examined and, furthermore, that there were no biologically meaningful changes in expression of the housekeeping genes examined. We were also able to confirm the results of the gene chip array for two genes not previously known to be part of the heat shock response (Rad and osteopontin).
The major strengths of this study were the use of freshly isolated normal human cells, the use of an experimentally conventional in vitro heat shock that produced a well-defined response (i.e., a measurable increase in HSP70 protein), the large number of sequences present on each chip, and the fact that enough gene chip array experiments were performed to allow us to apply statistical methods to our analysis. The choice of human PBMCs was dictated by the ease of obtaining samples from volunteers; furthermore, blood is the one tissue that is exposed to all major body compartments (including the hottest ones) and is thus, in principle, a rational choice for studying cellular responses to hyperthermia.
Although our method of transforming and analyzing the gene chip array data represents a conventional statistical approach to the problem, it has not, to our knowledge, been reported elsewhere. It has the advantage of simplicity and ease of use, it produces a quantitative assessment of experimental variability (through the estimation of the 95% confidence intervals on the mean), and it takes advantage of the chip-reading software's present, marginal, and absent calls to perform post hoc filtering of the data. The principal disadvantage of our approach is that it relies strictly on the proprietary chip-reading software's algorithm to measure changes in expression and to make accurate presence calls. This appears not to have been a major disadvantage, as the final analysis was clearly able to distinguish between transcripts that were expected to show a change in expression as a result of the experimental conditions applied (for example, HSPs) and those that were not (such as housekeeping genes).
The most significant design limitations of our study are the lack of time course data and the fact that the experimental heat shock was delivered in vitro. Our experimental design was intended to obtain a representative sampling of the breadth of the cell stress response while minimizing the number of false positives detected. This required a large number of repeated experiments at a sufficiently representative time point. Because the time point chosen was 3 h after the initiation of heat shock, the reported results probably do not include a number of the genes involved in the initiation of the cell stress response. Further confirmatory work, specifically examining different time points, will be required to address this issue. Work involving experimentally induced hyperthermia in vivo will also be required to precisely identify which of the genes that we have detected in vitro are likely to play an important physiological role in the systemic response to hyperthermia. In addition, our experiments do not distinguish between changes in levels of RNA attributable to changes in transcription and those attributable to changes in mRNA stability. The observed changes in gene expression could have resulted from either or both. Finally, it is important to remember that our experiments only addressed changes in gene expression; these do not necessarily correlate quantitatively with changes in cellular protein expression.
The volunteers for our study were all adult men. This had the advantage of excluding variations in genomic responses that are attributable to gender. However, there are known physiological differences in how women and men respond to heat stress (46). Isolated cardiac myocytes from male rats show estrogen- and progesterone-induced activation of HSF-1 and increased expression of HSP70 (24), suggesting that there may be important gender differences in the response to heat also at the cellular level. The full extent of these differences remains to be elucidated.
Although changes in gene expression were observed in every functional class known to be affected physiologically by heat, more than half of the genomic effect could be accounted for by genes in only six functional categories: HSPs/chaperonins/cochaperonins, immune function, cell proliferation and differentiation, signal transduction, transcription, and metabolism. Thus, although far more extensive than previously realized, much of the genomic response to heat stress nevertheless appears to be confined to a select number of functional classes.
Among the affected transcription factors, we found increased levels of jun and of fosB, a member of the fos family of transcription factors. We did not see an effect on c-fos itself. Increased expression of jun after heat shock has been noted in HeLa cells (7). In these cells, fos expression also increases during recovery from heat shock (1, 7), at least in part by a posttranscriptional RNA stabilization mechanism (1). Others have found, in Chinese hamster ovary fibroblasts, that serum-induced upregulation of fos is inhibited by concurrent heat exposure (4). Because of its broad transcriptional effects, transcription factor AP-1 (activator protein-1, a dimer of jun and fos) is likely to be a mediator of a significant number of the cellular responses to heat stress. However, our data also show that transcription factors other than AP-1 are significantly affected by heat stress, suggesting that the transcriptional response to heat shock is complex.
Although activation of stress kinase signal transduction pathways by heat shock is well established (16, 17, 36), little is known of the effect of heat shock on gene expression of individual members of these regulatory pathways. Expression of two phosphatases involved in feedback inhibition of MAPK pathways has previously been shown to be affected by heat shock. Expression of DUSP1 (also known as CL100) has been shown to be increased by heat shock (21, 23), and, in our experiment, the mRNA for this molecule was significantly increased (2.4-fold, Table 3). DUSP5 expression has also been shown to be increased by heat shock (21) in human skin fibroblasts but did not change significantly in our experiment. We did find an increase in DUSP8 (12-fold) and HePTPase (2.3-fold) and a decrease in DUSP6 (0.35-fold). Furthermore, mRNA for transcription factor jun, which is a known phosphorylation target of MAPK cascades (47), was significantly increased (2.6-fold). Our data therefore suggest that modulation of MAPK pathways during the cellular response to heat also occurs at the level of gene expression. Because heat shock is known to activate stress kinase signaling pathways by means of phosphorylation of key components, it has been suggested (21) that increased expression of stress kinase phosphatases in response to heat serves as a physiological way to "reset" these this important signaling pathway to a maximally stress-sensitive state. This interesting hypothesis warrants further testing, in light of our observation that mRNAs for several stress kinase pathway phosphatases are increased in response to heat stress.
Using RT-PCR, we confirmed the increase in G protein Rad that was detected by the gene chip. This Ras-like regulatory protein was originally identified by subtractive hybridization in skeletal muscle of individuals with Type 2 diabetes (42). It is a GTPase (53) that is known to interact with the cytoskeleton (2, 52), and it may play a role in regulating glucose uptake (35) and metabolism (14). These properties, coupled with our observation that Rad expression is increased during the cellular response to heat, suggest that it may play an important regulatory role in the cell stress response to heat and provide a candidate functional role for this gene of previously unclear function.
Osteopontin is a multifunctional secreted protein that is expressed by activated macrophages, lymphocytes, and by many other cell types, including injured and healing bone (6, 37). It is a cytokine that appears to be involved in the regulation of tissue repair and inflammation and has been associated with processes that involve a Th-1 type immune response, such as cell-mediated inflammation and granuloma formation. It is a chemoattractant to lymphocytes and mononuclear cells; supports adhesion of macrophages, B cells, T cells, and platelets; induces expression of other chemokines; and may also play a role in mononuclear cell survival. Accordingly, the observation that this cytokine is downregulated by heat shock may have significant immunologic implications.
In conclusion, nonlethal heat shock affects PBMC gene expression at the level of RNA in every major functional class previously known to be involved in the cellular response to heat, as well as in functional classes of potential systemic pathophysiological importance, such as immune function and coagulation. The genomic response to heat stress would thus appear to be broader than previously recognized.
| |
ACKNOWLEDGEMENTS |
|---|
We thank Nicholas J. Messinese for performing the reverse transcriptions and Dr. Mark Kellogg and Janet Staab for excellent technical assistance.
| |
FOOTNOTES |
|---|
This work was supported in part by National Heart, Lung, and Blood Institute Grant 1RO1 HL-AI64104.
The views, opinions, and findings contained in this publication are those of the authors and should not be construed as an official US Department of the Army position, policy, or decision, unless so designated by other documentation. Approved for public release; distribution unlimited.
For the protection of human subjects, the investigators adhered to the policies of applicable Federal Law CFR 46.
Previously presented at the Experimental Biology Meeting, Orlando, FL, April 2001.
Address for reprint requests and other correspondence: L. A. Sonna, Thermal and Mountain Medicine Division, US Army Research Institute of Environmental Medicine, 42 Kansas St., Natick, MA 01760 (E-mail: larry.sonna{at}na.amedd.army.mil).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published January 11, 2002;10.1152/japplphysiol.01002.2001
Received 1 October 2001; accepted in final form 4 January 2002.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
1.
Andrews, GK,
Harding MA,
Calvet JP,
and
Adamson ED.
The heat shock response in HeLa cells is accompanied by elevated expression of the c-fos proto-oncogene.
Mol Cell Biol
7:
3452-3458,
1987.
2.
Bilan, PJ,
Moyers JS,
and
Kahn CR.
The ras-related protein rad associates with the cytoskeleton in a non-lipid-dependent manner.
Exp Cell Res
242:
391-400,
1998.
3.
Bond, U.
Heat shock but not other stress inducers leads to the disruption of a sub-set of snRNPs and inhibition of in vitro splicing in HeLa cells.
EMBO J
7:
3509-3518,
1988.
4.
Chen, C,
Xie Y,
Stevenson MA,
Auron PE,
and
Calderwood SK.
Heat shock factor 1 represses Ras-induced transcriptional activation of the c-fos gene.
J Biol Chem
272:
26803-26806,
1997.
5.
Creagh, EM,
Sheehan D,
and
Cotter TG.
Heat shock proteins
modulators of apoptosis in tumour cells.
Leukemia
14:
1161-1173,
2000.
6.
Denhardt, DT,
Noda M,
O'Regan AW,
Pavlin D,
and
Berman JS.
Osteopontin as a means to cope with environmental insults: regulation of inflammation, tissue remodeling, and cell survival.
J Clin Invest
107:
1055-1061,
2001.
7.
Diamond, DA,
Parsian A,
Hunt CR,
Lofgren S,
Spitz DR,
Goswami PC,
and
Gius D.
Redox factor-1 (Ref-1) mediates the activation of AP-1 in HeLa and NIH 3T3 cells in response to heat shock.
J Biol Chem
274:
16959-16964,
1999.
8.
Dinh, HK,
Zhao B,
Schuschereba ST,
Merrill G,
and
Bowman PD.
Gene expression profiling of the response to thermal injury in human cells.
Physiol Genomics
7:
3-13,
2001.
9.
Dubois, MF,
and
Bensaude O.
MAP kinase activation during heat shock in quiescent and exponentially growing mammalian cells.
FEBS Lett
324:
191-195,
1993.
10.
Ewing, JF,
and
Maines MD.
Rapid induction of heme oxygenase 1 mRNA and protein by hyperthermia in rat brain: heme oxygenase 2 is not a heat shock protein.
Proc Natl Acad Sci USA
88:
5364-5368,
1991.
11.
Ewing, JF,
Raju VS,
and
Maines MD.
Induction of heart heme oxygenase-1 (HSP32) by hyperthermia: possible role in stress-mediated elevation of cyclic 3':5'-guanosine monophosphate.
J Pharmacol Exp Ther
271:
408-414,
1994.
12.
Farrell, RE, Jr.
Determination of Nucleic Acid Concentration and Purity. RNA Methodologies. San Diego, CA: Academic, 1998, p. 94-103.
13.
Gaffin, SL,
Koratich M,
and
Hubbard RW.
The effect of hyperthermia on intracellular sodium concentrations of isolated human cells.
Ann NY Acad Sci
813:
637-639,
1997.
14.
Garvey, WT,
Maianu L,
Kennedy A,
Wallace P,
Ganaway E,
Hamacher LL,
Yarnall DP,
Lenhard JM,
and
Burns DK.
Muscle Rad expression and human metabolism: potential role of the novel Ras-related GTPase in energy expenditure and body composition.
Diabetes
46:
444-450,
1997.
15.
Georgopoulos, C,
and
Welch WJ.
Role of the major heat shock proteins as molecular chaperones.
Annu Rev Cell Biol
9:
601-634,
1993.
16.
Han, SI,
Ha KS,
Kang KI,
Kim HD,
and
Kang HS.
Heat shock-induced actin polymerization, SAPK/JNK activation, and heat-shock protein expression are mediated by genistein-sensitive tyrosine kinase(s) in K562 cells.
Cell Biol Int
24:
447-457,
2000.
17.
Han, SI,
Oh SY,
Woo SH,
Kim KH,
Kim JH,
Kim HD,
and
Kang HS.
Implication of a small GTPase Rac1 in the activation of c-Jun N-terminal kinase and heat shock factor in response to heat shock.
J Biol Chem
276:
1889-1895,
2001.
18.
Hasday, JD,
and
Singh IS.
Fever and the heat shock response: distinct, partially overlapping processes.
Cell Stress Chaperones
5:
471-480,
2000.
19.
Helmbrecht, K,
Zeise E,
and
Rensing L.
Chaperones in cell cycle regulation and mitogenic signal transduction: a review.
Cell Prolif
33:
341-365,
2000.
20.
Hsiao, LL,
Dangond F,
Yoshida T,
Hong R,
Jensen RV,
Misra J,
Dillon W,
Lee KF,
Haverty P,
Weng Z,
Mutter GL,
Frosch MP,
MacDonald ME,
Milford EL,
Crum CP,
Bueno R,
Mahadevappa M,
Warrington JA,
Stephanopoulos G,
Stephanopoulos G,
and
Gullans GR.
A compendium of gene expression in normal human tissues reveals tissue-selective genes and distinct expression patterns of housekeeping genes.
Physiol Genomics
7:
97-104,
2001.
21.
Ishibashi, T,
Bottaro DP,
Michieli P,
Kelley CA,
and
Aaronson SA.
A novel dual specificity phosphatase induced by serum stimulation and heat shock.
J Biol Chem
269:
29897-29902,
1994.
22.
Jaattela, M.
Heat shock proteins as cellular lifeguards.
Ann Med
31:
261-271,
1999.
23.
Keyse, SM,
and
Emslie EA.
Oxidative stress and heat shock induce a human gene encoding a protein-tyrosine phosphatase.
Nature
359:
644-647,
1992.
24.
Knowlton, AA,
and
Sun L.
Heat-shock factor-1, steroid hormones, and regulation of heat-shock protein expression in the heart.
Am J Physiol Heart Circ Physiol
280:
H455-H464,
2001.
25.
Koratich, M,
and
Gaffin SL.
Mechanisms of calcium transport in human endothelial cells subjected to hyperthermia.
J Therm Biol
24:
245-249,
1999.
26.
Kuhl, NM,
and
Rensing L.
Heat shock effects on cell cycle progression.
Cell Mol Life Sci
57:
450-463,
2000.
27.
Lindquist, S.
The heat-shock response.
Annu Rev Biochem
55:
1151-1191,
1986.
28.
Lockhart, DJ,
Dong H,
Byrne MC,
Follettie MT,
Gallo MV,
Chee MS,
Mittmann M,
Wang C,
Kobayashi M,
Horton H,
and
Brown EL.
Expression monitoring by hybridization to high-density oligonucleotide arrays.
Nat Biotechnol
14:
1675-1680,
1996.
29.
Maines, MD,
and
Ewing JF.
Stress response of the rat testis: in situ hydridization and immunohistochemical analysis of heme oxygenase-1 (HSP32) induction by hyperthermia.
Biol Reprod
54:
1070-1079,
1996.
30.
Mathew, A,
and
Morimoto RI.
Role of the heat-shock response in the life and death of proteins.
Ann NY Acad Sci
851:
99-111,
1998.
31.
Mok, EH,
Smith HS,
DiBartolomeis SM,
Kerrebrock AW,
Rothschild LJ,
Lange TS,
and
Gerbi SA.
Maintenance of the DNA puff expanded state is independent of active replication and transcription.
Chromosoma
110:
186-196,
2001.
32.
Morimoto, RI.
Regulation of the heat shock transcriptional response: cross talk between a family of heat shock factors, molecular chaperones, and negative regulators.
Genes Dev
12:
3788-3796,
1998.
33.
Moseley, PL.
Heat shock proteins and the inflammatory response.
Ann NY Acad Sci
856:
206-213,
1998.
34.
Moseley, PL.
Exercise, stress, and the immune conversation.
Exerc Sport Sci Rev
28:
128-132,
2000.
35.
Moyers, JS,
Bilan PJ,
Reynet C,
and
Kahn CR.
Overexpression of Rad inhibits glucose uptake in cultured muscle and fat cells.
J Biol Chem
271:
23111-23116,
1996.
36.
Obata, T,
Brown GE,
and
Yaffe MB.
MAP kinase pathways activated by stress: the p38 MAPK pathway.
Crit Care Med
28:
N67-N77,
2000.
37.
O'Regan, A,
and
Berman JS.
Osteopontin: a key cytokine in cell-mediated and granulomatous inflammation.
Int J Exp Pathol
81:
373-390,
2000.
38.
Otterbein, LE,
and
Choi AM.
Heme oxygenase: colors of defense against cellular stress.
Am J Physiol Lung Cell Mol Physiol
279:
L1029-L1037,
2000.
39.
Panniers, R.
Translational control during heat shock.
Biochimie
76:
737-747,
1994.
40.
Parag, HA,
Raboy B,
and
Kulka RG.
Effect of heat shock on protein degradation in mammalian cells: involvement of the ubiquitin system.
EMBO J
6:
55-61,
1987.
41.
Punyiczki, M,
and
Fesus L.
Heat shock and apoptosis. The two defense systems of the organism may have overlapping molecular elements.
Ann NY Acad Sci
851:
67-74,
1998.
42.
Reynet, C,
and
Kahn CR.
Rad: a member of the Ras family overexpressed in muscle of type II diabetic humans.
Science
262:
1441-1444,
1993.
43.
Rockett, JC,
Mapp FL,
Garges JB,
Luft JC,
Mori C,
and
Dix DJ.
Effects of hyperthermia on spermatogenesis, apoptosis, gene expression, and fertility in adult male mice.
Biol Reprod
65:
229-239,
2001.
44.
Schena, M,
Shalon D,
Heller R,
Chai A,
Brown PO,
and
Davis RW.
Parallel human genome analysis: microarray-based expression monitoring of 1000 genes.
Proc Natl Acad Sci USA
93:
10614-10619,
1996.
45.
Skrandies, S,
Bremer B,
Pilatus U,
Mayer A,
Neuhaus-Steinmetz U,
and
Rensing L.
Heat shock- and ethanol-induced ionic changes in C6 rat glioma cells determined by NMR and fluorescence spectroscopy.
Brain Res
746:
220-230,
1997.
46.
Stephenson, LA,
and
Kolka MA.
Effect of gender, circadian period and sleep loss on thermal responses during exercise.
In: Human Performance Physiology and Environmental Medicine at Terrestrial Extremes, edited by Pandolf KB,
Sawka MN,
and Gonzalez RR.. Carmel, IN: Cooper, 1986, p. 267-304.
47.
Tibbles, LA,
and
Woodgett JR.
The stress-activated protein kinase pathways.
Cell Mol Life Sci
55:
1230-1254,
1999.
48.
Weitzel, G,
Pilatus U,
and
Rensing L.
The cytoplasmic pH, ATP content and total protein synthesis rate during heat-shock protein inducing treatments in yeast.
Exp Cell Res
170:
64-79,
1987.
49.
Xu, C,
Fracella F,
Richter-Landsberg C,
and
Rensing L.
Stress response of lysosomal cysteine proteinases in rat C6 glioma cells.
Comp Biochem Physiol B Biochem Mol Biol
117:
169-178,
1997.
50.
Yoo, HY,
Chang MS,
and
Rho HM.
The activation of the rat copper/zinc superoxide dismutase gene by hydrogen peroxide through the hydrogen peroxide-responsive element and by paraquat and heat shock through the same heat shock element.
J Biol Chem
274:
23887-23892,
1999.
51.
Yost, HJ,
and
Lindquist S.
Heat shock proteins affect RNA processing during the heat shock response of Saccharomyces cerevisiae.
Mol Cell Biol
11:
1062-1068,
1991.
52.
Zhu, J,
Bilan PJ,
Moyers JS,
Antonetti DA,
and
Kahn CR.
Rad, a novel Ras-related GTPase, interacts with skeletal muscle beta-tropomyosin.
J Biol Chem
271:
768-773,
1996.
53.
Zhu, J,
Reynet C,
Caldwell JS,
and
Kahn CR.
Characterization of Rad, a new member of Ras/GTPase superfamily, and its regulation by a unique GTPase-activating protein (GAP)-like activity.
J Biol Chem
270:
4805-4812,
1995.
This article has been cited by other articles:
|
|
 |
|
  I. S. Singh, A. Gupta, A. Nagarsekar, Z. Cooper, C. Manka, L. Hester, I. J. Benjamin, J.-r. He, and J. D. Hasday Heat Shock Co-Activates Interleukin-8 Transcription Am. J. Respir. Cell Mol. Biol., August 1, 2008; 39(2): 235 - 242. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. P. McClung, J. D. Hasday, J.-r. He, S. J. Montain, S. N. Cheuvront, M. N. Sawka, and I. S. Singh Exercise-heat acclimation in humans alters baseline levels and ex vivo heat inducibility of HSP72 and HSP90 in peripheral blood mononuclear cells Am J Physiol Regulatory Integrative Comp Physiol, January 1, 2008; 294(1): R185 - R191. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. Espinoza, J. A. Goodrich, and J. F. Kugel Characterization of the structure, function, and mechanism of B2 RNA, an ncRNA repressor of RNA polymerase II transcription RNA, April 1, 2007; 13(4): 583 - 596. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. T. Selsby, S. Rother, S. Tsuda, O. Pracash, J. Quindry, and S. L. Dodd Intermittent hyperthermia enhances skeletal muscle regrowth and attenuates oxidative damage following reloading J Appl Physiol, April 1, 2007; 102(4): 1702 - 1707. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Sonna, M. M. Kuhlmeier, H. C. Carter, J. D. Hasday, C. M. Lilly, and K. D. Fairchild Effect of moderate hypothermia on gene expression by THP-1 cells: a DNA microarray study Physiol Genomics, September 14, 2006; 26(1): 91 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Moran, L. Eli-Berchoer, Y. Heled, L. Mendel, M. Schocina, and M. Horowitz Heat intolerance: does gene transcription contribute? J Appl Physiol, April 1, 2006; 100(4): 1370 - 1376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Ilany, P. J. Bilan, S. Kapur, J. S. Caldwell, M.-E. Patti, A. Marette, and C. R. Kahn Overexpression of Rad in muscle worsens diet-induced insulin resistance and glucose intolerance and lowers plasma triglyceride level PNAS, March 21, 2006; 103(12): 4481 - 4486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. T. Selsby and S. L. Dodd Heat treatment reduces oxidative stress and protects muscle mass during immobilization Am J Physiol Regulatory Integrative Comp Physiol, July 1, 2005; 289(1): R134 - R139. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Lilly, H. Tateno, T. Oguma, E. Israel, and L. A. Sonna Effects of Allergen Challenge on Airway Epithelial Cell Gene Expression Am. J. Respir. Crit. Care Med., March 15, 2005; 171(6): 579 - 586. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Horowitz, L. Eli-Berchoer, I. Wapinski, N. Friedman, and E. Kodesh Stress-related genomic responses during the course of heat acclimation and its association with ischemic-reperfusion cross-tolerance J Appl Physiol, October 1, 2004; 97(4): 1496 - 1507. [Abstract] [Full Text] [PDF]
|
|
