Induction of Thermotolerance in EarlyPostimplantation Rat Embryos Is AssociatedWith Increased Resistance toHyperthermia-Induced ApoptosisPHILIP E. MIRKES,1,2* L.M. CORNEL,1 H.W. PARK,3 AND M.L. CUNNINGHAM1,2

1Birth Defects Research Laboratory, Division of Congenital Defects, Department of Pediatrics,University of Washington, Seattle, Washington 981952Birth Defects Research Laboratory, Division of Congenital Defects, Department of Biological Structure,University of Washington, Seattle, Washington 981953Department of Anatomy, Yonsei University College of Medicine, Seoul, Korea

ABSTRACT Previously we reported that hyper-thermia (437C) induces cell death in neurulation stagerat embryos as part of the pathogenesis culminating inabnormal growth and development. We now show thathyperthermia-induced cell death occurs by a processtermed apoptosis. DNA fragmentation, a hallmark ofapoptosis, was noted as early as 2.5 hr after embryoswere exposed to 437C. A smaller but significant in-crease in DNA fragmentation was also observed inembryos exposed to 427C, but only at the 5 hr timepoint. In control embryos, TUNEL-positive apoptoticbodies were consistently observed in the neuroepithe-lium at the point of neural tube closure and in the opticstalk. In embryos exposed to 437C, the number ofTUNEL-positive apoptotic bodies was significantly in-creased. Using both gel electrophoresis and TUNEL,we also show that the induction of thermotolerance isassociated with a significant reduction in DNA fragmen-tation. Together our results show that specific pro-grammed cell death and hyperthermia-induced celldeath correlate with internucleosomal DNA fragmenta-tion characteristic of apoptosis. Finally, we show thatthe induction of thermotolerance in rat embryos isassociated with a significant reduction in internucleo-somal DNA fragmentation and associated apoptosis.Teratology 56:210–219, 1997. r 1997 Wiley-Liss, Inc.

Programmed cell death (PCD) is an integral part ofthe normal development of most, if not all, embryonictissues and organs (Glucksman, ’62). The most well-studied example is the PCD that occurs during limbmorphogenesis, particularly digit formation (Saunderset al., ’62). Appropriate PCD is essential for normaldevelopment as evidenced by mutants such as hammer-toe in which altered PCD underlies the abnormaldevelopment of the limbs (Zakeri et al., ’94). In additionto the well-known role of PCD in normal development,it is also known that many developmental toxicantsinduce an episode of cell death in excess of PCD andthat this toxicant-induced cell death is part of the

pathogenesis that culminates in specific malformations(Scott, ’77). For example, hyperthermia is known toinduce an episode of cell death in postimplantationembryos exposed in vitro (Mirkes, ’85) and in vivo(Webster et al., ’85).

Recent work has also shown that PCD and cell deathinduced by a wide variety of physical and chemicalagents occur by a process of apoptosis (Arends andWyllie, ’91). Apoptosis is an active process controlled byspecific gene products (White, ’96). At the morphologi-cal level, apoptosis is characterized by chromatin con-densation/margination, cell shrinkage, membrane bleb-bing, and the formation of apoptotic bodies (Wyllie etal., ’80). Biochemically, one of the characteristic fea-tures of apoptosis is the cleavage of nuclear DNA intolarge fragments (Brown et al., ’93; Oberhammer et al.,’93) and/or multiples of internucleosomal DNA (185 bp)typically referred to as a DNA ladder (Compton, ’92).One of the goals of the present study was to determinewhether normal PCD and/or hyperthermia-induced celldeath in the rat embryo occurs by the process ofapoptosis, i.e., is characterized by internucleosomalcleavage of DNA.

Previously, we showed that a brief in vitro exposureto acute hyperthermia (43°C) induced cell death inearly postimplantation rat embryos (Mirkes, ’85). Hyper-thermia-induced cell death occurs preferentially in thedeveloping central nervous system and is most pro-nounced in the developing forebrain. In addition toinducing cell death, in vitro exposures to hyperthermiaalso induce growth retardation and abnormal develop-ment characterized by dysmorphogenesis of the fore-brain and edema of the hindbrain. We and others have

Contract grant sponsor: NIH; Contract grant numbers: HD22095,HD29137, ES07026, ES08744.

*Correspondence to: Philip E. Mirkes, Ph.D., Department of Pediat-rics, Box 356320, University of Washington, Seattle, WA 98195.E-mail: pemst@u.washington.edu

Received 21 May 1997; Accepted 23 July 1997

TERATOLOGY 56:210–219 (1997)

r 1997 WILEY-LISS, INC.

also shown that the acute effects of hyperthermia ongrowth and development can be partially amelioratedby exposing embryos to a non-embryotoxic exposure tohyperthermia (42°C) prior to the 43°C exposure, i.e., bythe induction of thermotolerance (Mirkes, ’87). Al-though not completely normal, thermotolerant embryosare characterized by growth and development param-eters more closely associated with control embryosgrown at 37°C than those exposed only to 43°C.

The mechanisms underlying this thermotoleranceresponse are incompletely understood; however, a sub-stantial body of information suggests that heat shockproteins (Hsps) play a critical role (Mirkes, ’96). Severallaboratories have reported that Hsp72, the heat-inducible member of the Hsp70 family, is associatedwith the induction of thermotolerance in mammalianpostimplantation embryos (Mirkes, ’87; Walsh et al.,’87; Kapron-Br’as and Hales, ’91). The relationshipbetween Hsp72 and improved growth and developmentinduced under thermotolerance conditions is unclear;however, the knowledge that hyperthermia-inducedabnormal development is associated with elevated lev-els of cell death suggests that thermotolerance may be,in part, related to resistance to hyperthermia-inducedcell death. Thus, the second goal of the present studieswas to determine whether the induction of thermotoler-ance is correlated with reduction in the level of celldeath induced by a 43°C exposure.

MATERIALS AND METHODS

Animals

On day 9 or 10 of gestation, overnight-mated, primi-gravida Sprague-Dawley rats were obtained from TylerLaboratories (Bellevue, WA) or from B & K UniversalLaboratories (Kent, WA). If a copulation plug was notedthe morning after mating, that morning was consideredday 0. The rats were maintained on a 14 hr light/10 hrdark cycle in an animal facility of the University ofWashington. The animals were fed Purina lab chow andgiven water ad libitum.

Whole embryo culture

On the afternoon of gestational day 10, pregnantanimals were anesthetized with halothane and eachuterus was removed and placed in Hank’s balanced saltsolution (HBSS; Gibco BRL, Gaithersburg, MD). Theembryos were then explanted and placed into a modi-fied whole embryo culture system as described by New(’78). Embryos were cultured overnight at 37°C inmedium containing 50% rat serum/50% HBSS. Culturemedium was gassed with 20% O2/5% CO2/75% N2 for 20min before adding embryos. On the morning of day 11,culture media were regassed with 95% O2/5% CO2.

In vitro exposure conditions

To assess the kinetics of hyperthermia-induced celldeath, day 11 embryos (22–25 somites) were assigned to1 of 3 groups. One group of embryos was exposed to43°C for 15 min by placing culture bottles in a water-

bath and then returned to culture at 37°C. A secondgroup was exposed to 42°C for 15 min (waterbath) andthen returned to culture at 37°C. The third group wascultured at 37°C throughout the experiment. At 1, 2.5,and 5 hr after the exposures to hyperthermia, groups of5–10 embryos were removed from culture. Groups of5–10 embryos were also removed at comparable time-points from the group of embryos cultured at 37°C.Embryos were dissected free of surrounding amnionand yolk sac, washed in HBSS, and then appropriatelyprocessed.

To assess the effects of thermotolerance on the level ofcell death, day 11 embryos were assigned to 1 of 4groups. One group was exposed to 42°C for 15 min andthen returned to culture at 37°C. A second group wasexposed to 43°C for 15 min and then returned to cultureat 37°C. A third group was exposed to 42°C for 15 minthen 1 hr later exposed to 43°C (15 min) and thenreturned to culture at 37°C. The final group wascultured at 37°C throughout the experiment. Afterexposures and culture at 37°C, 5–10 embryos wereremoved from each group at timepoints correspondingto 5 hr after the 43°C exposure in group three. Thistimepoint was chosen because available data indicatedthat hyperthermia-induced cell death was maximal 5hr after exposure to 43°C.

DNA fragmentation assay

Isolation of DNA. Groups of embryos were snap-frozen with liquid nitrogen and stored at 270°C. DNAwas isolated by a method adapted from Tilly and Hsueh(’93). After thawing embryos, homogenization buffer[0.1 M NaCl, 10 mM sodium ethylenediaminetetraace-tic acid (NaEDTA; pH 8), 0.3 M Tris HCl (pH 8), 0.2 Msucrose] was added: 100 µl of buffer per embryo wasadded to gestational day 11 embryos. The amount ofbuffer added was approximately 10 µl/µg of the ex-pected yield of DNA. The embryos were homogenizedwith 50 strokes in Dounce homogenizers. To eachhomogenate, 12.5 µl of 10% sodium dodecyl sulfate(SDS) per 200 µl homogenization buffer was added andthen incubated for 30 min at 65°C. To precipitateproteins, 35 µl of 8 M potassium acetate per 200 µlhomogenization buffer was added to each sample andplaced on ice for 60 min. Samples were then centrifuged(5,000g) 10 min at 4°C. The supernatants were ex-tracted with phenol:chloroform (1:1) and then chloro-form, precipitated with ethanol, and resuspended. Toaid precipitation, 10 mM magnesium chloride wasadded. RNA in each sample was digested with 1 µl (0.5mg/ml) DNase-free RNAase for 60 min at 60°C (10 mMTris, pH 7.5, 15 mM NaCl). Samples were again ex-tracted with phenol:chloroform (1:1) and chloroform,ethanol precipitated, and dried. Samples were resus-pended in 25–50 µl distilled/deionized water (ddH2O)and stored at 220°C. Concentration was determined byOD260 measurement.

Alternatively, genomic DNA was isolated using thetissue DNA isolation protocol from the Qiagen Genomic

RAT EMBRYO APOPTOSIS 211

DNA kit (Qiagen, Inc., Chatsworth, CA). Approxi-mately 21–30 mg (wet weight) of day 11 embryos wasused in minipreps (tip 20/G) according to the manufac-turer’s protocol. DNA was resuspended in 25–50 µl ofTE (10 mM Tris, 1 mM EDTA, pH 8.0) and theconcentration was determined by OD260 measurement.

End-labeling reaction. For each reaction, 0.5 µg ofisolated DNA was used. The total reaction volume was50 µl and included the following reagents: 1 µl (25 U)terminal deoxynucleotidyl transferase (TdT; Boehringer-Mannheim, Indianapolis, IN), 5 µl of 25 mM cobaltchloride, 10 µl of 53 reaction buffer (Boehringer-Mannheim), and 2.5 µl (25 µCi) of a-32P-dideoxy-ATP(Amersham, Arlington Heights, IL). The reactions werecarried out at 37°C for 60 min and stopped withaddition of 2.5 µl of 0.5 M EDTA. Labeled DNA wasprecipitated at 220°C with 3 volumes of absoluteethanol, 0.3 volumes of 7.5 M ammonium acetate, and 5µl yeast tRNA (10 mg/ml), and then centrifuged 20 minat 14,000g, 4°C. The dried pellets were resuspended in50 µl TE buffer. The samples were then reprecipitatedin ethanol as described above, dried, and resuspendedin 20–30 µl TE (pH 8.0).

Gel electrophoresis. Half of each sample (10 or 15µl) was loaded onto a 2% agarose, TAE (Tris-acetate/EDTA; 40 mM Tris, 1 mM EDTA) gel, with 0.1 volume103 DNA sample buffer (50% glycerol, 1 mM EDTA,0.4% bromophenol blue, 0.4% xylene cyanol). Sampleswere heated for 10 min at 55°C prior to loading.Samples were electrophoresed for 3.5–4 hr at 50 V. Thegel was dried on a slab gel dryer, and then exposed toKodak X-OMAT film or Reflection AutoradiographicFilm (DuPont NEN, Boston, MA).

Densitometric analysis. Films from 3 indepen-dent experiments were exposed for 1.5–3 hr.Acomputer-generated, scanned image of each film was created andsubsequently analyzed with the Millipore Bio Imagesystem. For each lane, integrated optical density (IOD)was determined for high molecular weight DNA as asum of the upper two bands (see Fig. 1, areas bracketedand labeled 1 and 2). The IOD of the rest of the lane wasused as a measure of low molecular weight DNA (areabracketed and labeled 3 in Fig. 1), and divided by thetotal lane IOD. Thus, the proportion of low molecularweight DNA was used to estimate relative amounts ofDNA fragmentation in each sample. Background levelsranged from 0.20 to 0.23 IOD units. Average valueswere determined for each treatment group and time-point. Statistical analyses were done using pairedt-tests (Statview 512).

Tunnel assay

The same 38 end-labeling reaction with TdT that wasused to label DNA for the DNA fragmentation assaywas utilized in situ when using the TUNEL method(TdT-mediated dUTP-biotin nick end labeling; Gavrielliet al., ’92). Embryos were exposed to hyperthermia andremoved from culture as described above and fixed 4–5hr in 4% phosphate-buffered saline (PBS) formaldehyde

at room temperature. Embryos were rinsed in 0.87%saline and then dehydrated through a graded series ofethanol. Three changes in Histoclear (National Diagnos-tics, Atlanta, GA) were done prior to infiltrating in 3changes of paraffin. Embryos were embedded in freshparaffin, sectioned at 5 µm through the area of theprosencephalon containing the optic cup, and mountedon silane-treated slides. Slides were deparaffinized andrehydrated through a graded series of ethanol. Thefollowing protocol for pretreatment and the in situreaction were adapted from Gavrielli et al. (’92). Theslides were treated with 10 µg/ml proteinase K for 30min at room temperature, washed 4 times, 2 min each,in ddH2O, and then treated with 2% H2O2 in ddH2O for5 min. Slides were again rinsed in ddH2O.

As a positive control, one slide from each group wastreated with DNase I (1 µg/ml in DNase buffer) for 10min at room temperature, after pretreating in DNasebuffer [30 mM Tris base (pH 7.2), 140 mM sodiumcacodylate, 4 mM magnesium chloride, and 0.1 mMdithiothreitol]. DNase-treated slides were then rinsedwell with ddH2O. All slides were pretreated in TdTbuffer [30 mM Tris base (pH 7.2), 140 mM sodiumcacodylate, and 1 mM cobalt chloride] for 5 min at roomtemperature. Except for negative control slides, theslides were incubated for 60 min, 37°C with TdT diluted1:250 in TdT buffer containing 5 µM biotinylated dUTP(Boehringer-Mannheim). The reactions were stoppedby placing in 300 mM sodium chloride/30 mM sodiumcitrate for 15 min at room temperature. Slides werethen rinsed in ddH2O.

The slides were blocked in 2% aqueous bovine serumalbumin for 10 min at room temperature, rinsed inddH2O, and then placed in PBS for 5 min. The sectionswere then covered with Vectastain Standard ABC re-agent (Vector, Burlingame, CA: PBS) for 30 min at37°C. The slides were washed 3 times in PBS, 5 mineach, and then developed with 3,38 diaminobenzidine(DAB/H2O2; Sigma, St. Louis, MO). The slides werecounterstained 13–15 min with methyl green, dehy-drated through a graded series of ethanols, and cover-slipped.

An alternative apoptotic detection system was alsoused. The TACS 2 TdT In Situ Apoptosis Detection Kitwas obtained from Trevigen, Inc. (Gaithersburg, MD).Instead of the streptavidin binding, DAB detection, andmethyl green counterstaining used in the Gavriellimethod, the TACS Blue Label Detection Protocol (Trevi-gen, Inc.) was utilized. Modifications to this protocolwere as follows: 1) proteinase K digestion at 10 µg/mlfor 30 min, 2) DNase I (1 µg/ml in 50 mM Tris, pH 7.4,and 5 mM MgCl2) treatment (positive controls) for 10min at room temperature prior to endogenous peroxi-dase removal, 3) slides were stained with Blue Label for2–3 min, and 4) counterstained with Red CounterstainC for 30 sec to 1 min. Slides were then dehydratedthrough a series of ethanol solutions and Histoclear

212 P.E. MIRKES ET AL.

prior to mounting with Histomount (National Diagnos-tics).

Quantitation of cell death after TUNEL labeling

To permit comparisons of cell death in tissue sectionsfrom embryos exposed to 37°C, 42°C, 43°C, or 42°C/43°C, we limited our analysis to those 5 µm sectionsthat encompass the prosencephalon containing theoptic cup. Typically, this area comprised 20–30 sections.TUNEL-positive apoptotic bodies in individual sectionswere counted manually using a microscope equippedwith a 340 objective and then the average number ofapoptotic bodies for all sections encompassing the opticcup was calculated. At least 2 different embryos wereexamined from each exposure group in 2 independentexperiments. In addition, assessments of the number ofapoptotic bodies in each of the 2 experiments wereperformed independently by 2 different investigators.Because cells produce multiple apoptotic bodies duringthe process of apoptosis, it is not possible to determinethe exact number of cells that are apoptotic at any one

timepoint. Thus, we have presented our data as thenumber of TUNEL-positive apoptotic bodies.

RESULTS

Previous work (Mirkes, ’85) has demonstrated thatan exposure to 43°C elicited an increase in embryoniccell death that could be detected in histologically stainedsections of rat embryos approximately 5 hr after expo-sure. To determine whether this hyperthermia-inducedcell death was associated with the internucleosomalDNA cleavage characteristic of apoptosis, we initiallycompared the state of DNA isolated from control em-bryos grown at 37°C (Fig. 1A) with DNA isolated fromembryos 5 hr after an exposure to 43°C (Fig. 1B). Figure1C shows that the majority of DNA from control em-bryos was high molecular weight; however, there wasclear evidence of some DNA cleavage as evidenced bythe DNA ladder composed of multiples of 185 bp piecesof DNA. The DNA from embryos exposed to 43°C (Fig.1D) showed a distinct increase in fragmented DNA.

Fig. 1. A: Day 11 rat embryo grown in vitro for 24 hr at a temperature of 37°C. B: Day 11 rat embryopreviously exposed on day 10 to 43°C for 15 min and then cultured for 24 hr at 37°C. C: Autoradiograph oflabeled DNA isolated from embryos cultured at 37°C and electrophoresed in 2% agarose. D: Similar to C,except DNA was isolated from embryos 5 hr after exposure to 43°C for 15 min. Brackets in C and Drepresent high molecular weight DNA (1, 2) and low molecular weight DNA (3).

RAT EMBRYO APOPTOSIS 213

Densitometric analysis of these data indicated thatapproximately 10% of the DNA from control embryosand 60% of the DNA from 43°C-exposed embryos pre-sent as fragmented DNA.

Previous histological analyses of the temporal pat-tern of induction of cell death after exposure to hyper-thermia indicated that cell death became histologicallyevident between 2.5 and 5 hr after embryos are exposedto a temperature of 43°C (unpublished data). To com-pare this histological timeframe of hyperthermia-induced cell death with the onset of DNA cleavagecharacteristic of apoptosis, we next undertook a kineticanalysis of hyperthermia-induced DNA cleavage. Fig-ure 2A depicts a representative autoradiograph show-ing the state of DNA (high molecular weight vs. inter-nucleosomal DNA fragments) at various times afterembryos were exposed to either 37°C, 42°C, or 43°C.The 42°C exposure had little, if any, effect on embryonicgrowth and development, whereas 43°C induced signifi-cant effects on embryonic growth and development.Visual inspection of this autoradiograph indicates thefollowing: 1) over a 5 hr period of culture, the level ofDNA fragmentation varies from 10 to 15% in embryoscultured at 37°C, 2) the level of DNA fragmentation inembryos cultured at 42°C is slightly increased abovecontrol levels at 2.5 and 5 hr after exposure, and 3) thelevel of DNA fragmentation is obviously increased atboth 2.5 and 5 hr in embryos exposed to 43°C comparedto control embryos. Figure 2B presents mean values forpercent fragmented DNA from 3 separate experiments.These data show that no significant differences in theamount of DNA damage exist at the 1 hr time periodbetween the 3 temperatures. At the 2.5 hr timepoint,there is a significant elevation in the level of DNAfragmentation in embryos exposed to 43°C compared to37°C. At the 5 hr timepoint, the levels of DNA damageinduced by exposures to 42°C and 43°C are significantlyelevated compared to 37°C (control). In addition, theamount of DNA fragmentation induced by exposure to43°C is significantly elevated compared to that inducedby 42°C.

The previous data were obtained using whole em-bryos and therefore provide no information concerninghyperthermia-induced DNA fragmentation/apoptosis indifferent embryonic tissues. To provide this informa-tion, we used the TUNEL method and focused on thecell death induced in the prosencephalic neuroepithe-lium in the area encompassing the optic cup. Figure3A–L shows representative TUNEL-stained sectionsfrom embryos exposed to 37°C (A–D), 42°C (E–H), and43°C (I–L) and examined at 5 hr after exposure. Incontrol embryos (37°C), TUNEL-positive nuclei areobserved at the point of neural tube closure (Fig. 3A,B)and in the optic stalk (Fig. 3D). Occasional apoptoticbodies were observed in other areas of the neuroepithe-lium and surrounding mesenchyme (Fig. 3C). An identi-cal distribution of TUNEL-positive nuclei was observedin comparable sections from stage-matched embryosthat were not cultured in vitro but were removed from

the uterus and immediately fixed prior to TUNELstaining (data not shown). Thus, programmed celldeath (TUNEL-positive) occurs by the process of apopto-sis. In embryos exposed to 42°C (Fig. 3E–H), no obviousincrease in TUNEL-positive nuclei could be detected;however, an obvious increase in TUNEL-positive nucleiwas observed in sections from embryos exposed to 43°C(Fig. 3I–L). A majority of TUNEL-positive nuclei werefound in the neuroepithelium and many TUNEL-positive nuclei had been sloughed into the neurallumen (Fig. 3L). These data, combined with the DNA

Fig. 2. Kinetics of DNA fragmentation induced by hyperthermia(42°C or 43°C, 15 min.) in day 11 rat embryos. A: Autoradiograph oflabeled DNA isolated from control (37°C), 42°C-exposed, and 43°C-exposed embryos at 1, 2.5, and 5 hr after exposure. B: Histogramdepicting percent fragmented DNA determined by densitometricanalysis of autoradiographs such as the one presented in A. Verticalbars represent standard errors. *P 5 0.05 (43°C compared to 37°C);**P , 0.05 (42°C compared to 37°C); ***P , 0.01 (43°C compared to37°C); ****P 5 0.01 (43°C compared to 42°C).

214 P.E. MIRKES ET AL.

fragmentation data, indicate that hyperthermia-induced cell death in postimplantation rat embryosoccurs by the process of apoptosis.

To assess the kinetics of hyperthermia-induced celldeath in the prosencephalon, TUNEL analysis wasperformed on sections encompassing the optic cupportion of the prosencephalon of embryos taken atvarious times after exposure to 37°C, 42°C, or 43°C. Ateach time and exposure, the total number of TUNEL-positive cells in the optic cup region of prosencephalonwas determined. Figure 4 presents data for 2 separateexperiments and shows that little, if any, increases inTUNEL-positive bodies are observed except at the 5 hrtimepoint in embryos exposed to 43°C.

We next turned to the relationship between thermo-tolerance and apoptosis. Previous studies showed thatembryos exposed to 42°C prior to 43°C exhibited bettergrowth and development compared to embryos exposedto only 43°C (Mirkes, ’87), i.e., the doubly exposedembryos were thermotolerant. We now asked whetherthis thermotolerance was associated with reduced lev-els of DNA fragmentation and therefore apoptosis.Figure 5A compares the amount of DNA fragmentationin embryos cultured under 1 of 4 conditions: 1) 37°Conly, 2) 42°C only, 3) 43°C only, and 4) 42°C/43°C doubleexposure. As noted previously, 42°C alone induces anincrease in DNA fragmentation even though this expo-sure does not produce observable effects on subsequentgrowth, development, and cell death. Likewise, 43°Calone, an exposure that alters subsequent growth anddevelopment, induces an increase in DNA fragmenta-tion above that observed in embryos exposed to 37°C or42°C. In comparison, the amount of DNA fragmenta-tion was significantly decreased in embryos exposedinitially to 42°C and subsequently to 43°C. Similarresults were obtained by quantitating the number ofTUNEL-positive nuclei in the optic cup region of theprosencephalon (Fig. 5B). Thus, the 42°C exposurepartially protects the embryo from the embryotoxiclevels of DNA fragmentation induced by a subsequentexposure to 43°C.

DISCUSSION

The associations of PCD with normal development(Glucksman, ’62) and teratogen-induced cell death withabnormal development (Scott, ’77) are well docu-mented. Until recently, both types of cell death werethought to be passive processes, as evidenced by theconsistent reference to teratogen-induced cell death asnecrosis. On the basis of the pioneering work of Wyllieet al. (’80), we now know that cell death induced in avariety of contexts, primarily non-embryonic, occurs bymeans of an orderly process termed apoptosis. Wyllie(’80) was the first to show that part of this orderlyprocess is the degradation of DNA. This DNA degrada-tion is mediated by a Ca/Mg-dependent endonucleasethat induces multiple nicks in internucleosomal DNA.This apoptosis-specific DNA degradation has typicallybeen visualized by gel electrophoresis of isolated DNA

as a ‘‘ladder’’ of DNA fragments that are multiples of185 base pairs (Tilly and Hsueh, ’93) or by an immuno-cytochemical approach that allows one to detect thisDNA fragmentation in situ, i.e., the TUNEL method(Gavrielli et al., ’92). Since this initial discovery, theformation of this DNA ladder has become a hallmarksignature of apoptosis (Compton, ’92).

One of the findings of the present study is that thePCDs associated with neural tube closure and with theeye (optic stalk) exhibit this hallmark of apoptosis. We(Little and Mirkes, ’95) and others (Zakeri et al., ’94)have previously shown that interdigital PCD in thedeveloping limb is associated with internucleosomalDNA fragmentation. Cell death associated with pri-mary palate closure (unpublished data) also exhibitsthis hallmark of apoptosis. Although we have notexamined every instance of PCD, available data sug-gest that cell death observed at different developmentalstages and in different tissues occurs via the process ofapoptosis. Thus, one aspect of development is theembryo’s ability to regulate cell death in specific popula-tions of cells. The observation that specific populationsof cells die in specific embryonic tissues at specific timesin development suggests that PCD is essential tonormal morphogenesis. Although limited, data can bemarshalled to support this thesis. For example, pa-tients with cryptophthalmos syndrome (Fraser’s syn-drome) exhibit failure of eye fissures to form, cutaneoussyndactyly, and atresia of the ear canals, anus, vagina,and alimentary canal (Fraser, ’62). The etiology of all ofthese defects is consistent with failed PCD. On thebasis of linkage analysis, Takayama et al. (’96) suggestthat Fraser’s syndrome may result from a mutation inBAG1. BAG1 encodes a protein that interacts withBcl-2 and suppresses cell death induced by staurospo-rine, anti-Fas antibody, and cytolytic T cells (Takayamaet al., ’95). Thus, the failed PCD associated withFraser’s syndrome may result from mutation in a keygene involved in regulating apoptosis. Another exampleinvolves the mouse mutant, hammertoe, in which web-bing occurs between digits 2 and 5 of all 4 limbs (Green,’64). This webbing has been shown to be directly relatedto the failure of cell death between digits 2 and 5 duringlimb development (Zakeri et al., ’94). Interestingly,normal patterns of cell death occur between the otherdigits.

Another finding of our study is that hyperthermia-induced cell death in the rat embryo is also associatedwith DNA fragmentation characteristic of apoptosis.This finding is in agreement with several reports(Barry et al., ’90; Harmon et al., ’90) showing thatcultured cells exposed to hyperthermia are induced todie by the process of apoptosis. The importance of ourfindings is that the cell death induced by hyperthermiais not a passive process (necrosis) but an active processnow known to involve the activities of an ever-growingnumber of cellular proteins (White, ’96). As such, celldeath induced by hyperthermia and potentially other

RAT EMBRYO APOPTOSIS 215

Fig. 3. Representative histological sections through prosencephaloncontaining the optic cup from day 11 rat embryos cultured at 37°C(A–D), 42°C, 15 min (E–H), and 43°C, 15 min (I–L) stained by theTUNEL method. Embryos were removed from culture 5 hr afterexposure to hyperthermia and at a comparable time for control

embryos. Lettered boxes in A, E, and I indicate areas that are shown inhigher magnification in B–D, F–H, and J–L, respectively. Arrowshighlight individual or groups of apoptotic bodies stained by theTUNEL method.

216 P.E. MIRKES ET AL.

teratogens is susceptible to manipulation by geneticand/or pharmacologic means.

Before this can be achieved, more details are requiredconcerning the essential events associated with theapoptotic process and how various events are tempo-rally related. Our results indicate that hyperthermia-induced DNA fragmentation is a relatively early eventin the apoptotic process. By gel electrophoresis, asignificant increase in DNA fragmentation is observed2.5 hr after a 43°C exposure, well before histologicalevidence of cell death is observed. Although the in-creased DNA fragmentation observed at 1 hr after a43°C exposure does not reach statistical significance,these data suggest that DNA fragmentation may beinitiated as early as 1 hr after exposure to 43°C. In turn,this would indicate that the Ca/Mg-dependent endo-nuclease(s) responsible for the initial cleavage of DNAinto high molecular weight fragments (Earnshaw, ’95)and subsequently into smaller fragments that aremultiples of 185 bp (Wyllie, ’80) is activated early in theprogram of hyperthermia-induced apoptosis. Little isknown about the sequence of events initiated by expo-sure to hyperthermia in rat embryos; however, prelimi-nary data show that the activation of the cysteineprotease CPP32 and the cleavage of poly ADP-ribosepolymerase (PARP) also occur in rat embryos exposed tohyperthermia (unpublished data). CPP32 is homolo-gous to the Ced-3 gene product in Caenorhabditiselegans and interleukin-1b converting enzyme (ICE)and is known to play a key role in carrying out the deathsentence once the decision to die is made (Yuan, ’95).PARP is a nuclear enzyme thought to play a role in DNA

repair (de Murcia and de Murcia, ’94). Cleavage ofPARP is associated with inactivation of this enzyme(Kaufmann et al., ’93), and thus the observed cleavageof PARP in rat embryos may indicate that a cell’sdecision to die is irreversible. Work in a variety ofnon-embryonic systems is identifying an ever-increas-ing number of proteins associated with the process ofapoptosis and one of the challenges is to identify thoseproteins playing a role in cell death induced by a varietyof developmental toxicants.

Finally, our results demonstrate that the level ofDNA fragmentation induced by an embryotoxic expo-sure of hyperthermia (43°C) is significantly reduced inembryos exposed to conditions that we (Mirkes, ’87;Thayer and Mirkes, ’97) have shown confer thermotoler-ance, i.e., 42°C exposure 1 hr before a 43°C exposure.Similar results have been reported for cultured humancells that were made thermotolerant to a cytotoxicexposure of hyperthermia (43°C for 2 hr) by pretreatingcells 10 hr earlier to an exposure of 43°C for 30 min, anexposure that by itself results in little DNA fragmenta-tion (Mosser and Martin, ’92). In a more recent study,Samali and Cotter (’96) report that human U937 cellsand murine Wehi-s cells are resistant to the inductionof apoptosis by actinomycin-D, campothecin, and etopo-side if they are pretreated with exposure to mildhyperthermia (42°C for 1 hr) 2 hr before exposure tothese cytotoxic agents. Thus, our data using postimplan-tation rat embryos and the data cited for several celllines clearly demonstrate that thermotolerance is asso-ciated with the induction of resistance to apoptosis.

Fig. 4. Histogram showing the kinetics of cell death (number of TUNEL-positive apoptotic bodies)induction in the prosencephalic region containing the optic cup in day 11 control embryos or embryosexposed for 15 min to hyperthermia (42°C or 43°C). The 2 histograms in each treatment group representresults from 2 independent experiments.

RAT EMBRYO APOPTOSIS 217

Although the mechanisms underlying this thermotol-erance-induced reduction in DNA fragmentation andapoptosis are not completely understood, recent datahave implicated Hsps, specifically Hsp70 and Hsp27. Liet al. (’96) have shown that overexpression of human

Hsp70 in rat fibroblasts significantly reduces the levelof heat-induced apoptosis. Similarly, Samali and Cotter(’96) showed that the overexpression of human Hsp70in murine Wehi-s cells also protects them from thecytotoxic effects of campothecin. The latter study also

Fig. 5. Effects of induced thermotolerance on the amount of DNA fragmentation (A) and TUNEL-positiveapoptotic bodies (B) in day 11 control embryos (37°C), embryos exposed to 42°C for 15 min, embryosexposed to 43°C, and embryos exposed to 42°C followed 1 hr later by exposure to 43°C. Embryos wereremoved from culture 5 hr after the 43°C exposure and at a comparable time for all other treatmentgroups. The 2 histograms in each treatment group represent results from 2 independent experiments.

218 P.E. MIRKES ET AL.

showed that overexpression of Hsp27 afforded the samelevel of protection as overexpression of Hsp70. Similaroverexpression studies have not been performed usingmammalian embryos; however, we have shown thatHsp70 is the dominant Hsp synthesized in response totemperatures that induce thermotolerance (Mirkes,’87; Thayer and Mirkes, ’96). In addition, we haveshown that the level of Hsp27 and its phosphorylationstate are increased when rat embryos are exposed tohyperthermia (Mirkes et al., ’96). When viewed in thelight of the overexpression data obtained from culturedcells, the data obtained from postimplantation mamma-lian embryos support the hypothesis that Hsp70 andHsp27 play a role in protecting embryonic cells from thecytotoxic effects of heat and developmental toxicants.The mechanisms by which these Hsps protect cells fromapoptosis are not known; however, the known role ofHsps as molecular chaperones (Georgopoulos and Welch,’93) suggests that Hsps may protect cells by interactingwith any one of a number of proteins known to regulateapoptosis (Hale et al., ’96).

ACKNOWLEDGMENTS

This work was supported by NIH grants HD22095,HD29137, ES07026, and ES08744 to P.E.M.

LITERATURE CITEDArends, M.J., and A.H. Wyllie (1991) Apoptosis: Mechanisms and roles

in pathology. Int. Rev. Exp. Pathol., 32:223–254.Barry, M.A., C.A. Behnke, and A. Eastman (1990) Activation of pro-

grammed cell death (apoptosis) by cisplatin, other anticancer drugs,toxins and hyperthermia. Biochem. Pharmacol., 40:2353–2362.

Brown, D.G., X.M. Sun, and G.M. Cohen (1993) Dexamethasone-induced apoptosis involves cleavage of DNA to large fragments priorto internucleosomal fragmentation. J. Biol. Chem., 268:3037–3039.

Compton, M.M. (1992) A biological hallmark of apoptosis: Internucleo-somal degradation of the genome. Cancer Metastas. Rev., 11:105–119.

de Murcia, G., and J.M. de Murcia (1994) Poly (ADP-ribose) polymer-ase: A molecular nick-sensor. Trends Biol. Sci., 19:172–176.

Earnshaw, W.C. (1995) Apoptosis: Lessons from in vitro systems.Trends Cell Biol., 5:217–220.

Fraser, G.R. (1962) Our genetical ‘‘load’’: A review of some aspects ofgenetical variation. Ann. Hum. Genet., 25:387–415.

Gavrielli, Y., Y. Sherman, and S.A. Ben-Sasson (1992) Identification ofprogrammed cell death in situ via specific labeling of nuclear DNAfragmentation. J. Cell Biol., 119:493–501.

Georgopoulos, C., and W.J. Welch (1993) Role of the major heat shockproteins as molecular chaperones. Annu. Rev. Cell Biol., 9:601–634.

Glucksman, A. (1962) Cell deaths in normal vertebrate ontogeny. Biol.Rev., 26:59–86.

Green, M.C. (1964) New Mutation. Mouse News Lett., 31:27.Hale, A.J., C.A. Smith, L.C. Sutherland, V.E.A. Stoneman, V.L.

Longthorne, A.C. Culhane, and G.T. Williams (1996) Apoptosis:Molecular regulation of cell death. Eur. J. Biochem., 236:1–26.

Harmon, B.V., A.M. Corder, R.J. Collins, G.C. Gobe, J. Allen, D.J.Allan, and J.F.R. Kerr (1990) Cell death induced in a murinemastocytoma by 42–47 C heating in vitro: Evidence that the form ofdeath changes from apoptosis to necrosis above a critical heat load.Int. J. Radiat. Biol., 58:845–858.

Kapron-Br’as, C.M., and B.F. Hales (1991) Heat shock induced toler-ance to the embryotoxic effects of hyperthermia and cadmium inmouse embryos in vitro. Teratology, 43:83–94.

Kaufmann, S.H., S. Desnoyers, Y. Ottaviano, N.E. Davidson, and G.G.Poirier (1993) Specific proteolytic cleavage of poly (ADP-ribose)

polymerase: An early marker of chemotherapy-induced apoptosis.Cancer Res., 53:3976–3985.

Li, W.X., C.H. Chen, C.C. Ling, and G.C. Li (1996) Apoptosis inheat-induced cell killing: The protective role of hsp-70 and thesensitization effect of the c-myc gene. Radiat. Res., 145:324–330.

Little, S.A., and P.E. Mirkes (1995) Clusterin expression duringprogrammed and teratogen-induced cell death in the postimplanta-tion rat embryo. Teratology, 52:41–54.

Mirkes, P.E. (1985) Effects of acute exposures to elevated tempera-tures on rat embryo growth and development in vitro. Teratology,32:259–266.

Mirkes, P.E. (1987) Hyperthermia-induced heat shock response andthermotolerance in postimplantation rat embryos. Dev. Biol., 119:115–122.

Mirkes, P.E. (1997) Cellular responses to stress. In: Handbook ofExperimental Pharmacology. R.J. Kavlock and G. Daston, eds.Springer-Verlag, New York, pp. 245–276.

Mirkes, P.E., S.A. Little, L. Cornel, M.J. Welsh, T.-N.N. Laney, andF.H. Wright (1996) Induction of heat shock protein 27 in rat embryosexposed to hyperthermia. Mol. Reprod. Dev., 45:276–284.

Mosser, D.D., and L.H. Martin (1992) Induced thermotolerance toapoptosis in a human T lymphocyte cell line. J. Cell. Physiol.,151:561–570.

New, D.A.T. (1978) Whole embryo culture and the study of mammalianembryos during organogenesis. Biol. Rev., 53:81–122.

Oberhammer, F., J.W. Wilson, C. Dive, et al. (1993) Apoptotic death inepithelial cells: Cleavage of DNA to 300 and/or 500 kb fragmentsprior to or in the absence of internucleosomal fragmentation. EMBOJ., 12:3679–3684.

Samali, A., and T.G. Cotter (1996) Heat shock proteins increaseresistance to apoptosis. Exp. Cell Res., 223:163–170.

Saunders, J.W., Jr., M.T. Gasseling, and L.C. Saunders (1962) Cellulardeath in morphogenesis of the avian wing. Dev. Biol., 5:147–178.

Scott, W.J. (1977) Cell death and reduced proliferative rate. In:Handbook of Teratology. J.G. Wilson and F.C. Fraser, eds. PlenumPress, New York, Vol. 2, pp. 81–98.

Takayama, S., T. Sato, S. Krajewski, K. Kochel, S. Irie, J.A. Millan,and J.C. Reed (1995) Cloning and functional analysis of BAG-1: Anovel Bcl-2 binding protein with anti-cell death activity. Cell,80:279–284.

Takayama, S., K. Kochel, S. Irie, J. Inazawa, T. Abe, T. Sato, T. Druck,K. Huebner, and J.C. Reed (1996) Cloning of cDNAs encoding thehuman BAG1 protein and localization of the human BAG1 gene tochromosome 9p12. Genomics, 35:494–498.

Thayer, J.M., and P.E. Mirkes (1997) Induction of Hsp72 and transientnuclear localization of Hsp73 and Hsp72 correlate with the acquisi-tion and loss of thermotolerance in postimplantation rat embryos.Dev. Dyn., 208:1–17.

Tilly, J.L., and A.J.W. Hsueh (1993) Microscale autoradiographicmethod for the qualitative and quantitative analysis of apoptoticDNA fragmentation. J. Cell. Physiol., 154:519–526.

Walsh, D.A., N.W. Klein, L.E. Hightower, and M.J. Edwards (1987)Heat shock and thermotolerance during early rat embryo develop-ment. Teratology, 36:181–191.

Webster, W.S., M.-A. Germain, and M.J. Edwards (1985) The inductionof microphthalmia, encephalocele, and other head defects followinghyperthermia during the gastrulation process in the rat. Teratology,31:73–82.

White, E. (1996) Life, death and the pursuit of apoptosis. Genes Dev.,10:1–15.

Wyllie, A.H. (1980) Glucocorticoid-induced thymocyte apoptosis isassociated with endogenous endonuclease activation. Nature, 284:555–556.

Wyllie, A.H., J.F.R. Kerr, and A.R. Currie (1980) Cell death: Thesignificance of apoptosis. Int. Rev. Cytol., 68:251–306.

Yuan, J. (1995) Molecular control of life and death. Curr. Opin. CellBiol., 7:211–214.

Zakeri, Z., D. Quagliano, and H.S. Ahuja (1994) Apoptotic cell death inthe mouse limb and its suppression in the hammertoe mutant. Dev.Biol., 165:294–297.

RAT EMBRYO APOPTOSIS 219