DNA damage and p21WAF1/CIP1/SDI1 in experimental injury of the rat adrenal cortex and trauma-associated damage of the human adrenal cortex

J. Pathol. 189: 119–126 (1999)

DNA DAMAGE AND p21WAF1/CIP1/SDI1 INEXPERIMENTAL INJURY OF THE RAT ADRENAL

CORTEX AND TRAUMA-ASSOCIATED DAMAGE OFTHE HUMAN ADRENAL CORTEX

. , †, . *

Huffington Center on Aging and Department of Cell Biology, Baylor College of Medicine, Houston, TX 77030, U.S.A.

SUMMARY

In vivo models are needed to study the reactions of tissues to DNA damage, such as the induction of the cyclin-dependent kinaseinhibitor p21, indicating potential repair of the damage, versus apoptosis, indicating the elimination of the damaged cells. Damage toDNA occurs in tissues during shock, sepsis, and other critical medical conditions. Previous studies have found evidence of damage to thecortex of adrenal glands from organ donors who had undergone severe trauma prior to death. The present experiment studied rats underexperimental interventions of clinical relevance to patients with conditions that put them at risk for damage to the adrenal glands. Theseinterventions comprised ischaemia and reperfusion injury, sepsis following caecal ligation and puncture, acute pancreatitis, andadministration of chemical agents (zymosan and acrylonitrile). All the interventions caused an increase in p21 mRNA as assessed bynorthern blotting and in situ hybridization. Increased nuclear p21 protein was shown by immunohistochemistry. All the interventionscaused damage to DNA, as shown by labelling of available 3* termini of single-strand breaks with terminal transferase. The number ofcells undergoing apoptosis, visualized by ligation of a hairpin oligonucleotide probe to double-strand breaks in DNA, was much lower.In rat adrenal glands, apoptotic cells were infrequent under all the conditions studied. They were more abundant in human organ donoradrenal glands that were previously shown to have extensive DNA damage accompanied by induction of p21. The similarity of the effectsof a wide variety of surgical interventions and chemical agents suggest a common pathophysiological mechanism which is not specific tothe initiating injury. Experimental injury of the rat adrenal cortex provides a model for investigating the role of organ DNA damage andof mediators of the response to DNA damage, such as p21. Copyright ? 1999 John Wiley & Sons, Ltd.

KEY WORDS—DNA damage; p21WAF1/CIP1/SDI1; ischaemia/reperfusion injury; sepsis; adrenal cortex; apoptosis

*Correspondence to: Peter J. Hornsby, PhD, Huffington Center onAging, Baylor College of Medicine, 1 Baylor Plaza M320, Houston,TX 77030, U.S.A. E-mail: [email protected]

†Current address: Department of Surgery, Lund University,S-221 85 Lund, Sweden.

Contract/grant sponsor: National Institute on Aging; Contract/grant numbers: AG 12287, AG 13633.

INTRODUCTION

Protection against DNA damage comprises repair,if damage is moderate, or commitment of the damagedcell to apoptosis, if damage exceeds a threshold level.1,2

Both apoptosis and cell cycle arrest, to permit repairof damage, may occur within a cell population follow-ing DNA damage.3,4 DNA damage causes arrest viathe p53-dependent transcriptional activation of thecyclin-dependent kinase inhibitor p21WAF1/CIP1/SDI1.5–9

p21 was originally characterized as a target gene for p53,as well as an inhibitor of cyclin-dependent kinases andas a cell cycle inhibitor that is expressed in senescentcells.10–12 Most studies of DNA damage, p21 induction,and apoptosis have been performed in cells in culture;model systems in which the relationships among theseprocesses can be studied in vivo have been much lessfrequently described. Examples of in vivo activation ofp21 and apoptosis by DNA damage are radiation withultraviolet light7 or X-rays,13,14 and ureteric obstruc-tion.15,16 The paucity of in vivo systems for studying p21has contributed to the current poor understanding of the

CCC 0022–3417/99/100119–08$17.50Copyright ? 1999 John Wiley & Sons, Ltd.

role of p21 in vivo. The elucidation of multiple functionsfor p21 in cultured cells17 has been difficult to reconcilewith the observation that p21 null mice have relativelyfew detectable abnormalities.13,14

The possible involvement of p21 versus apoptosis asreactions to DNA damage in tissues during shock,sepsis, and other critical medical conditions has not beenaddressed. We show here that experimental injury of therat adrenal cortex provides a model for investigating therole of DNA damage in organ injury and for investigat-ing the role of mediators of the response to DNAdamage, such as p21. The adrenal cortex may sustaindamage under a wide variety of pathophysiologi-cal conditions in humans18–21 and in experimentalanimals.22–25 The extent to which the adrenal cortex maybe affected in critical medical conditions has probablybeen underrecognized.18,19 In humans, the most seriousconsequence of this injury is massive haemorrhage intothe adrenal glands, leading to adrenal insufficiency andpossibly death.19,21 A specific cause of adrenal haemor-rhage is severe meningococcal sepsis (Waterhouse–Friderichsen syndrome26), but haemorrhage has beenobserved following many other initiating events.19,21

Less severe damage to the adrenal cortex appears tooccur in hypotension27 and in patients dying fromtrauma.28

In previous studies, we found evidence of damage tothe adrenal cortex in glands from organ donors whohad undergone severe trauma prior to death.29 Cells

Received 17 September 1998Revised 17 December 1998

Accepted 20 April 1999

120 V. V. DIDENKO ET AL.

of the adrenal cortex had extensive DNA damageaccompanied by increased p53 and p21WAF1/CIP1/SDI1,detected by immunocytochemistry. Some cells showedevidence of apoptosis. We hypothesized that damage tothe adrenal glands in human trauma victims wouldinvolve periods of low blood perfusion alternating withperiods of higher blood flow, as observed clinically.30 Totest the effects of fluctuating blood flow, we subjectedrat adrenal glands to ischaemia and reperfusion, whichinduced DNA damage, p53, and p21 in the adrenalcortex.29

In the present experiments, we performed a moreextensive set of studies in rats involving interventions ofclinical relevance to patients with conditions that putthem at risk for damage to the adrenal glands, such asshock and sepsis. A specific aim was to examine whetherthe intervention used previously (ischaemia/reperfusion)uniquely damaged the adrenal cortex, or whether, assuggested by the clinical literature, common end-pointsmight result from damage by a variety of surgical andchemical agents. Additionally, we reinvestigated theextent of apoptosis in adrenal damage, both in ratadrenal glands and in human adrenal glands previouslyobserved to have high levels of DNA damage and p21expression. Previously, we detected apoptosis by nuclearmorphology using light and electron microscopy.29 Wenoted that the number of apoptotic cells detected bynuclear morphology was much smaller than the numberof cells with DNA strand breaks detected by an in situend-labelling technique using terminal transferase.31 Weinterpreted the terminal transferase labelling as indicat-ing the presence of single-strand breaks in DNA, butcells with this form of DNA damage are not necessarilycommitted to undergoing apoptosis. In the presentexperiments, we used an in situ labelling technique thatdetects double-strand breaks in DNA by ligation of adouble-stranded DNA probe.32,33 Nuclear morphologymust still be relied on to confirm that labelled cells areundergoing apoptosis, because cells in the early phasesof necrosis also have double-strand breaks.34 However,in situ labelling using ligation does not stain the exten-sive single-strand breaks that may occur in necrotictissue, or in cells undergoing DNA damage from avariety of sources, all of which are labelled by terminaltransferase, and thus the method can be more reliably

32,33
used as a histological stain for apoptosis.
MATERIALS AND METHODS

Surgical interventions

Ischaemia/reperfusion—Ischaemic damage to theadrenal gland was induced as described previously.29

Sprague–Dawley rats (male, 150 g body weight) wereanaesthetized with a ketamine/xylazine/acepromazinecombination. The peritoneal cavity was opened and theleft renal artery was ligated adjacent to the aorta, using4-0 surgical silk tied around the artery and a short lengthof 1 mm polyester tubing. Ischaemia was monitored byloss and gain of colour of the affected organs (adrenalgland and kidney). After 30 min, the ligation wasremoved. This time was selected on the basis of a

Copyright ? 1999 John Wiley & Sons, Ltd.

previous report that 30 min of ischaemia followed byreperfusion produced small (0·1–0·5 mm) infarcts in theinner rat adrenal cortex.35 The skin incision was closedwith wound clips. The animals were allowed to recoverfrom the anaesthetic and were killed after 4 or 8 h. Theadrenal glands were bisected. Part of each gland wasprocessed for histochemistry and part was used forRNA preparation, as described below.

Hypovolaemia—Rats were anaesthetized as above.The femoral artery was cannulated to enable removal ofblood and monitoring of blood pressure (model 1290Cpressure transducer, Hewlett Packard Inc., Andover,MA, U.S.A.): 5 ml of blood was withdrawn over5–8 min and following this, 1 ml of blood was with-drawn every 10 min for 30 min, keeping the arterialpressure at 30–40 mm Hg. The artery was then ligatedand the skin incision was closed. The animals were killedafter 4 or 8 h.

Caecal ligation and puncture—Under anaesthesia, theperitoneal cavity was opened and the distal two-thirds ofthe caecum was ligated and punctured once with an18-gauge syringe needle.36 The animals were killed after4 or 8 h.

Acute haemorrhagic pancreatitis—The proximal endof the common bile duct was catheterized by puncturethrough the duodenal wall.37 0·2 ml of 5 per cent sodiumdeoxycholate (Sigma Chemical Co., St. Louis, MO,U.S.A.) was injected. The animals were killed after 27 h.

Acrylonitrile—(Aldrich Chemical Co., St. Louis, MO,U.S.A.) was administered at 0·05–0·15 mg/g body weightin 0·2 per cent Tween 20 via the femoral vein.38 Animalswere killed after 8 or 16 h.

Zymosan—(Sigma Chemical Co.) was administeredat 0·75 or 1·0 mg/g body weight intraperitoneally.39

Animals were killed after 8 or 16 h.

Human adrenal tissue

The human adrenal tissue investigated here wasfrom an 8-year-old male organ donor as describedpreviously.29 The adrenal gland was removed duringrecovery of the kidneys for potential transplantation. Ithad been perfused with cold University of Wisconsinsolution40 prior to organ recovery and was transportedon ice in the same solution. Adrenal cortex fragmentswere fixed in paraformaldehyde without allowing anincrease in temperature of the tissue.

Probes and northern blotting

A mouse p21 cDNA41 was used to probe rat RNA,as in earlier experiments.29 RNA was prepared fromadrenal glands by immediate immersion of dissectedtissue into cold RNAzol (Tel-Test Inc., Houston, TX,U.S.A.) followed by freezing in liquid nitrogen forstorage. On thawing, the tissue was homogenized in

J. Pathol. 189: 119–126 (1999)

121DNA DAMAGE/p21 IN ADRENOCORTICAL DAMAGE

RNAzol using a Teflon/glass motor-driven homogen-izer, followed by RNA preparation in accordancewith the manufacturer’s directions. Northern blottingand probe labelling were performed using previouslyestablished techniques.42 To check loading, blots werereprobed, after washing in 85)C water, with an oligonu-cleotide complementary to human ribosomal 28SRNA (Oncogene Science Inc., Cambridge, MA, U.S.A.)labelled with 32P using polynucleotide kinase.

In situ hybridization

Tissues were fixed in 4 per cent paraformaldehyde,dehydrated, and embedded in paraffin. Sections (6 ìm)were deparaffinized and rehydrated using graded alcoholconcentrations. In situ hybridization for p21 mRNA wasperformed as previously described.43 An antisense probeagainst rat p21 was synthesized from a DNA templatecorresponding to part of exon 2 of the rat p21 gene.Primers against this fragment of the exon were designedusing the human sequence as described previously;44

both the rat and the human sequences are amplified bythese primers. The 3* primer was modified by adding onits 5* end a T7 polymerase consensus promoter site.45

Polymerase chain reaction (PCR) using rat genomicDNA as template was performed as previouslydescribed.44 Using the PCR product as template,fluorescein-labelled RNA was synthesized using T7polymerase (New England Biolabs, Beverly, MA,U.S.A.) in the buffer recommended by the manufacturerwith 50 ì fluorescein-UTP (Molecular Probes, Eugene,OR, U.S.A.) together with unlabelled ATP, CTP,and GTP at 200 ì each. Following hybridization andwashing, sections were optionally counterstained withthe DNA-binding dye 4,6-diamidino-2-phenylindole(DAPI) (1 ìg/ml) and were mounted in Vectashield(Vector Laboratories, Burlingame, CA, U.S.A.) forobservation by fluorescence microscopy.

Immunohistochemistry

Anti-human p21 monoclonal antibody CP36,46 a giftof E. Harlow, was previously shown to detect rat p21 bythe use of an antigen retrieval technique.29 Staining wasachieved using the avidin–biotin–peroxidase complex(Vector Laboratories, Burlingame, CA, U.S.A.), asrecommended by the manufacturer. Sections were notcounterstained.

Detection of DNA damage by terminal transferaselabelling

For the reaction of available DNA 3* hydroxylswith terminal transferase, the published procedure31

was modified to accommodate the use of digoxigeninor Texas Red as labels rather than biotin. A mix-ture comprising 30 m Tris–HCl (pH 7·2), 140 msodium cacodylate, 1 m cobalt chloride, 0·1 m DTT,20 ì digoxigenin-dUTP (Boehringer–Mannheim,Indianapolis, IN, U.S.A.) or 8 ì Texas Red-X dUTP(Molecular Probes, Eugene, OR, U.S.A.) and 800units/ml terminal transferase (Boehringer–Mannheim)


(20 ìl per section) was added for 1 h at 37)C in ahumidified incubator. Following washing in water(two changes over 20 min), the sections were stainedfor light microscopic examination (digoxigenin) or forfluorescence microscopy (Texas Red). For digoxigenin,staining was achieved using an alkaline phosphatase-sheep anti-mouse Fab fragment conjugate (Boehringer–Mannheim) followed by a 30 min incubation with thechromogens 5-bromo-4-chloro-3-indolyl phosphate andnitro blue tetrazolium, using the procedure recom-mended by the manufacturer. The reaction was stoppedby washing sections in water. The wet sections werephotographed without counterstain.

Labelling of double-strand breaks by ligation ofbiotin-labelled hairpin oligonucleotides

A hairpin oligonucleotide that can be ligated todouble-strand breaks, as found in apoptotic cells, wasdescribed previously.33 It has a 10 bp stem region toform a hairpin with a defined double-strand end with asingle 3* A overhang.33 A loop of 20 nt was designed toaccommodate biotin labels without base-pairing in thisregion. At five places in the loop, the oligonucleotidewas synthesized with the amino modifier C6 deoxy-uridine (Glen Research, Sterling, VA, U.S.A.). Aftersynthesis of the oligonucleotide, biotin was covalentlyattached to the amino groups by reaction with biotinbis-aminohexanoyl N-hydroxysuccinimide ester (GlenResearch). The synthesis and post-synthesis bio-tinylation were performed by Synthetic Genetics Corp.(San Diego, CA, U.S.A.).

The oligonucleotide was ligated to DNA in tissuesections in situ using T4 DNA ligase. Sections weredeparaffinized with xylene and rehydrated in gradedalcohol concentrations. After washing in water, sectionswere incubated for 90 min at 65)C in 10 m sodiumcitrate, pH 6·0, and then rewashed with water.

The following procedures were performed at roomtemperature (23)C). Sections were incubated with 25 ìg/ml proteinase K (Oncor, Gaithersburg, MD, U.S.A.) inphosphate-buffered saline (PBS) for 5 min. Sections werethen rinsed thoroughly with water. A mixture of 50 mTris–HCl (pH 7·8), 10 m MgCl2, 10 m DTT, 1 mATP, 15 per cent polyethylene glycol (8000 MW, Sigma),with hairpin oligonucleotide at 35 ìg/ml and DNAT4 ligase (Boehringer–Mannheim, Indianapolis, IN,U.S.A.) at 250 units/ml was added (20 ìl per section).Sections were covered with glass coverslips and placed ina humidified box for 16 h. The sections were then washedwith several changes of water over 2 h.

Sections were stained for light microscopy or fluor-escence microscopy, or optionally were processed fordouble-labelling using terminal transferase as describedabove before visualization. For light microscopy, sec-tions were incubated with the avidin–biotin–peroxidasecomplex as described above. For fluorescence mi-croscopy, fluorescein–avidin conjugate (Vector Lab-oratories) was added at 4 ìg/ml in 50 m sodiumbicarbonate, 15 m sodium chloride, pH 8·2, for 45 min.After washing, sections were counterstained with DAPIand observed by fluorescence microscopy.

J. Pathol. 189: 119–126 (1999)


Table I—Damage to the adrenal cortex initiated by various interventions

Intervention DNA damage

p21

ApoptosismRNA

(Northern)In situ

hybridization Immunoreactivity

Rat adrenal glandsSham operation " " " " (+)Ischaemia/reperfusion of adrenal gland + + + + + + + (+)Contralateral gland " " " " (+)Hypovolaemia + + + + + + n.d. (+)Sepsis (caecal ligation and puncture) + + + + + + + (+)Acute pancreatitis + + + + n.d. n.d. (+)Zymosan + + + + n.d. n.d. (+)Acrylonitrile + + + + n.d. n.d. (+)

Human adrenal glands from organ donors + + +* + +* + +* + +* +

Tabulation of the surgical procedures used in rats in the present experiments together with data on human adrenal glands from organ donors(*human adrenal gland data for DNA damage and p21 are from ref. 29). " indicates not present; (+) to + + + indicates estimate of extent. n.d.=notdetermined.

Fig. 1—Expression of p21 in experimental injury of the rat adrenal gland, showing that a variety ofsurgical and chemical interventions induce expression of p21. Adrenal glands were obtained from animalssubjected to the indicated surgical procedures, which are described in the Materials and Methods sectionand listed in Table I. Blots of RNA from the adrenal glands from these animals were hybridized with ap21 cDNA probe and then rehybridized with a 28S ribosomal RNA probe as a loading control.CLP=caecal ligation and puncture model for sepsis; i./r.=ischaemia/reperfusion; the left adrenal gland(L) was subjected to 30 min ischaemia followed by the indicated time of recovery; RNA from thenon-ischaemic right gland (R) was also used. Combined hypovolaemia and ischaemia/reperfusion wasperformed in some animals

RESULTS

Several surgical and chemical interventions produceddamage to the rat adrenal cortex, as summarized inTable I. In these procedures, adrenal glands removedfrom the animals at death were reddened and grosslyenlarged. The glands were divided for histologicalstudies and for preparation of RNA for northernblotting. All the interventions performed caused anincrease in p21 mRNA (Fig. 1). In situ hybridizationwith an antisense RNA probe against rat p21 showed a


uniform increase in expression throughout the adrenalcortex (Fig. 2). The medulla showed much lowerexpression.

It was more difficult to demonstrate p21 expression byimmunocytochemistry because most available anti-bodies do not recognize rodent p21 in paraffin sections,even after antigen retrieval (our unpublished data). Incontrast, both human and bovine p21 are readilydetected by this method.29 However, the monoclonalantibody CP36 raised against human p2146 producesweak but specific staining for p21 in rat tissue sections.29

J. Pathol. 189: 119–126 (1999)


Fig. 2—In situ hybridization of rat adrenal gland sections with afluorescent p21 antisense RNA probe, demonstrating uniform induc-tion of p21 throughout the cortex by ischaemia/reperfusion injury. Theinner part of the cortex is shown adjacent to the medulla, which is notstained by the probe. Adrenal glands were subjected to ischaemia, asdescribed in the text, followed by 8 h of recovery (sections a and b) orwere from a control animal (c). a was photographed using a greenemission fluorescence filter to show the fluorescein-labelled p21 probe.Nuclei appear unstained. b and c were photographed with a dualwavelength blue/green emission filter showing both the hybridizedprobe and DAPI-stained nuclei. Note that the cytoplasmic greenfluorescence corresponding to hybridization of the probe is brighterthan the nuclei in b, whereas in the control section (c) the nucleiare brighter than the cytoplasm. The very bright objects in all threeimages are red blood cells, which exhibit yellow autofluorescence.#100


Fig. 3—Immunohistochemical demonstration of p21 protein in nucleiof rat adrenocortical cells. p21 was detected using a monoclonalantibody against human p21 and was visualized by peroxidase stain-ing. a is a section from a control animal; b is from an animal withexperimental sepsis (caecal ligation and puncture) at 8 h. #800

Fig. 4—Comparison of DNA damage and apoptosis in rat adrenalglands. DNA damage was assessed by terminal transferase labelling ofstrand breaks and apoptosis was assessed by ligation-mediated label-ling of cells with double-strand breaks in DNA. a, b, and c are terminaltransferase-labelled sections of rat adrenal glands from animals sub-jected to ischaemia followed by 8 h of recovery (a), or CLP at 8 h (b) orfrom a control animal (c). a*, b*, and c* are serial sections from the sameglands labelled by ligation with a hairpin oligonucleotide probe andvisualized as described in the Materials and Methods section. #200

Use of this antibody showed increased nuclear expres-sion of p21 throughout the adrenal cortex in sepsis andother types of damage (Fig. 3). The adrenal medulla wasnot stained.

The treatments used in these experiments producedclear evidence of DNA damage. This was shown bylabelling available 3* ends of DNA at the termini ofsingle-strand breaks using terminal transferase (Fig. 4).As previously shown, this method did not label mostcells in adrenals from control animals.29 The time takenfor the formation of the reaction product of the alkalinephosphatase reaction was greater than that convention-ally used to label strand breaks in apoptotic cells (30 minversus 4 min; see ref. 29).

The extent of apoptosis was much less than that ofDNA damage. Apoptotic cells were visualized by liga-tion of a hairpin oligonucleotide probe to double-strand

J. Pathol. 189: 119–126 (1999)


breaks.32,33 In serial sections from experimental animals,occasional single cells labelled by ligation of oligo-nucleotide probes were surrounded by much largernumbers of cells with single-strand breaks (Fig. 4). Thenumber of apoptotic cells encountered was too low toestablish whether it was increased by the treatmentslisted in Table I.

To confirm that the nuclei stained by double-strandDNA ligation were undergoing apoptosis, it was neces-sary to show details of nuclear morphology, which isdifficult in sections stained by chromogen deposition.32

Therefore, in some sections the oligonucleotide probewas visualized using fluorescence microscopy (Fig. 5).Individual stained nuclei showed characteristic conden-sations of chromatin.33 These nuclei also stainedbrightly when visualized with DNA-binding dyes, as isalso characteristic of apoptotic cells (Fig. 5).

A similar discrepancy in the abundance of single-strand breaks in DNA and double-strand breaks wasobserved in human adrenal glands previously shown tohave large numbers of p21-positive cells.29 When humanadrenal sections were subjected to dual labelling forDNA damage using terminal transferase and apoptosisusing hairpin oligonucleotide ligation, apoptotic cellswere observed to be more abundant in glands, such asthat shown in Figs 6 and 7, which had been demon-strated to have large numbers of cells expressingp21.29 However, there were far fewer cells labelled byoligonucleotide ligation than by terminal transferase(Fig. 6).

Fig. 5—Visualization of an apoptotic cell in a rat adrenal section (CLP8 h). A nucleus with characteristic morphology associated withapoptosis is labelled by ligation of a hairpin oligonucleotide probe.(a) Ultraviolet excitation showing nuclei stained with DAPI; (b) thesame microscope field observed by green fluorescence, showing theoligonucleotide probe ligated to the section. #500


Fig. 6—Reassessment of DNA damage and apoptosis in an adrenalgland from a human organ donor, confirming that terminal transferaselabels many cells that are not undergoing apoptosis. Adrenal glandsections were from an 8-year-old male donor, as previously reported.29

Dual labelling was performed to show single-strand breaks in DNA(terminal transferase labelling with Texas Red-dUTP) and double-strand breaks characteristic of apoptosis (ligation of hairpin oligo-nucleotide probe visualized with avidin-fluorescein). (a) The sectionobserved with ultraviolet excitation showing all nuclei stained withDAPI; (b) the same microscope field observed with green fluorescence,showing a nucleus with double-strand breaks; (c) the same fieldobserved with red fluorescence showing nuclei with single-strandbreaks. #500

DISCUSSION

In these experiments we studied the reaction of therat adrenal cortex to experimental injury. A commonreaction of DNA damage, as demonstrated by single-strand breaks accessible to terminal transferase, and byinduction of p21, as demonstrated by northern blotting,in situ hybridization, and immunocytochemistry, wasprovoked by a variety of surgical and chemical inter-ventions, other than a sham operation. We previouslyshowed evidence of increased p53 in ischaemia/reperfusion, one of the conditions studied here.29 Unfor-tunately, the available antibodies to p53 do not detectwild-type rat p53 with high sensitivity, although a fewlabelled cells may be observed, as we previously demon-strated.29 The currently studied interventions thatraise p21 expression also produced weak staining for

J. Pathol. 189: 119–126 (1999)


nuclear p53 (not shown). Together, these observationsallow the conclusion that p21 is induced by a transcrip-tional mechanism in the rat adrenal cortex under thesecircumstances. Although p21 has multiple functions,17,47

a possible role of p21 in these events would be to preventcell division until damaged DNA has been repaired.3

The similarity of the effects of a wide variety ofsurgical interventions and chemical agents suggests acommon pathogenesis not specifically related to theparticular events initiating the injury. A likely commontarget for various forms of adrenal injury is the micro-vasculature. In many tissues, shock, sepsis, ischaemia/reperfusion injury, and various pharmacologicalconditions involve leakiness of capillaries (clinical capil-lary leak syndrome48). Extreme fluctuations in perfusionof the adrenal gland resulting from shock lead todisseminated intravascular coagulation.49 Endotoxinscause fibrin deposition in the adrenals and kidneys,but not in other organs, correlating with increases inplasminogen activator inhibitor-I and tissue factorand decreases in urokinase plasminogen activator.50

Thrombosis may result in slowing of blood flow andpooling of blood. Capillary endothelial cells aredamaged by cytokines that circulate at high levels inshock and sepsis, causing the endothelium to becomeleaky.49 Extravasation of red blood cells and leukocytesis observed in the adrenal glands in shock and sepsis.18,19

DNA damage in adrenocortical cells might result fromexposure of the cells to a cytokine that can cause DNAstrand breaks, or to blood proteins that are not normallyin direct contact with adrenocortical cells, or to oxygen-radical generating systems, such as granulocytes.51,52

Oxygen radicals and their products cause extensivestrand breaks in target cell DNA.53

In human adrenal glands, DNA damage wasaccompanied by both p21 expression and apoptosis, aswe previously reported29 and have confirmed here usinga specific assay for double-strand breaks. However, rat


adrenal glands had a very low level of apoptosis underboth control and experimental conditions. Because thesame tissues showed evidence of extensive DNA damageby terminal transferase labelling, we confirm that thelatter method cannot be used in isolation as a methodfor assessing apoptosis.32,33

The reason for the occurrence of apoptosis in humanbut not rat glands is not clear. Moreover, it is not clearwhy apoptosis is confined to a small number of cells inthe human adrenal cortex, although it may be that thosecells have sustained a greater level of DNA damage. It islikely that p21 elevation and apoptosis are independenteffects of DNA damage and not linked;3 in numeroussystems, the relationship between p21 and apoptosis iscomplex and no consensus has been reached.54,55 Onesimple model is that low levels of DNA damage createa growth arrest signal, via p21, which allows repair ofthe damage, and that higher levels of damage triggerthe apoptotic response, even in a cell in which p21 hasalready been induced.29

In severe adrenal injury, haemorrhage causes massivedestruction of the gland with possibly fatal conse-quences.19 A non-lethal form of damage appears to beassociated with a temporary, but prolonged loss of thesynthesis of androgens by the adrenal glands.56,57 Thishas been observed in patients with burns and in patientsundergoing various surgical procedures.58,59 Androgensare synthesized by the inner zone of the adrenal cortex,the zona reticularis, which may be more vulnerable todamage because it is on the venous end of the capillarybed.56,57 More extensive observations on a wide varietyof human adrenal glands from donors of different ageswill be required to show whether apoptosis is morefrequent in the reticularis than in other zones. We havesuggested that a mild, chronic form of damage to thehuman adrenal gland could result from periodic fluctua-tions in blood flow over a normal life span. Duringageing, this may cause progressive changes in adreno-cortical structure and function.56,57 The present modelof experimental injury of the rat adrenal cortex willfacilitate the investigation of the role of organ DNAdamage and of mediators of the response to DNAdamage, such as p21, both in acute medical conditionsand in chronic responses during ageing.

ACKNOWLEDGEMENT

This work was supported by grants AG 12287 andAG 13633 from the National Institute on Aging.

Fig. 7—Distribution of apoptotic cells labelled by ligation with ahairpin oligonucleotide probe in human adrenal cortex (same tissuesample as studied in Fig. 6). The capsule is visible at the top of thesection; medullary tissue is not present in this part of the gland. Theprobe was detected with avidin-Texas Red and the section wascounterstained with DAPI. The section was photographed with redfluorescence and the DAPI-stained nuclei are faintly visible in thebackground. #50

REFERENCES

1. Lane DP. p53, guardian of the genome. Nature 1992; 358: 15–16.2. Wyllie AH. Apoptosis. Death gets a brake. Nature 1994; 369: 272–273.3. Canman CE, Chen C-Y, Lee M-H, Kastan MB. DNA damage responses:

p53 induction, cell cycle perturbations, and apoptosis. Cold Spring HarborSymp Quant Biol 1994; 59: 277–286.

4. Yeargin J, Haas M. Elevated levels of wild-type p53 induced by radio-labeling of cells leads to apoptosis or sustained growth arrest. Curr Biol1995; 5: 423–431.

5. Johnson M, Dimitrov D, Vojta PJ, et al. Evidence for a p53-independentpathway for upregulation of SDI1/CIP1/WAF1/p21 RNA in human cells.Mol Carcinog 1994; 11: 59–64.

J. Pathol. 189: 119–126 (1999)


6. Di Leonardo A, Linke SP, Clarkin K, Wahl GM. DNA damage triggers aprolonged p53-dependent G1 arrest and long-term induction of Cip1 innormal human fibroblasts. Genes Dev 1994; 8: 2540–2551.

7. El-Deiry WS, Tokino T, Waldman T, et al. Topological control ofp21WAF1/CIP1 expression in normal and neoplastic tissues. Cancer Res 1995;55: 2910–2919.

8. Dulic V, Kaufmann WK, Wilson SJ, et al. p53-dependent inhibition ofcyclin-dependent kinase activities in human fibroblasts during radiation-induced G1 arrest. Cell 1994; 76: 1013–1023.

9. Macleod KF, Sherry N, Hannon G, et al. p53-dependent and independentexpression of p21 during cell growth, differentiation, and DNA damage.Genes Dev 1995; 9: 935–944.

10. El-Deiry WS, Tokino T, Velculescu VE, et al. WAF1, a potential mediatorof p53 tumor suppression. Cell 1993; 75: 817–825.

11. Harper JW, Adami GR, Wei N, Keyomarsi K, Elledge SJ. The p21Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependentkinases. Cell 1993; 75: 805–816.

12. Noda A, Ning Y, Venable SF, Pereira-Smith OM, Smith JR. Cloning ofsenescent cell-derived inhibitors of DNA synthesis using an expressionscreen. Exp Cell Res 1994; 211: 90–98.

13. Deng C, Zhang P, Harper JW, Elledge SJ, Leder P. Mice lackingp21CIP1/WAF1 undergo normal development, but are defective in G1checkpoint control. Cell 1995; 82: 675–684.

14. Brugarolas J, Chandrasekaran C, Gordon JI, Beach D, Jacks T, HannonGJ. Radiation-induced cell cycle arrest compromised by p21 deficiency.Nature 1995; 377: 552–557.

15. Megyesi J, Udvarhelyi N, Safirstein RL, Price PM. The p53-independentactivation of transcription of p21WAF1/CIP1/SDI1 after acute renal failure.Am J Physiol 1996; 271: F1211–1216.

16. Morrissey JJ, Ishidoya S, McCracken R, Klahr S. Control of p53 andp21WAF1 expression during unilateral ureteral obstruction. Kidney Int 1996;50: S84–S92.

17. Labaer J, Garrett MD, Stevenson LF, et al. New functional activities for thep21 family of cdk inhibitors. Genes Dev 1997; 11: 847–862.

18. Siegel LM, Grinspoon SK, Garvey GJ, Bilezikian JP. Sepsis and adrenalfunction. Trends Endocrinol Metab 1994; 5: 324–328.

19. Rao RH. Bilateral massive adrenal hemorrhage. Med Clin N Am 1995; 79:107–129.

20. Arnason JA, Graziano FM. Adrenal insufficiency in the antiphospholipidantibody syndrome. Semin Arthritis Rheumatism 1995; 25: 109–116.

21. Streeten DHP. Adrenal hemorrhage. Endocrinologist 1996; 6: 277–284.22. Hoerr N. The cells of the suprarenal cortex in the guinea-pig. Their reaction

to injury and their replacement. Am J Anat 1931; 48: 139–160.23. Yarrington JT. Chemically induced adrenocortical lesions. In: Jones TC,

Mohr U, Hunt RD, eds. Monographs on Pathology of LaboratoryAnimals—Endocrine System. Berlin: Springer-Verlag, 1983: 69–75.

24. Colby HD, Eacho PI. Chemical-induced adrenal injury: role of metabolicactivation. In: Thomas JA, Korach KS, McLachlan JA, eds. EndocrineToxicology. New York: Raven, 1985: 35–66.

25. Szabo S, Lippe IT. Adrenal gland: chemically induced structural andfunctional changes in the cortex. Toxicol Pathol 1989; 17: 317–329.

26. Bohm M. Adrenal, cutaneous and myocardial lesions in fulminating endo-toxinemia (Waterhouse–Friderichsen syndrome). Pathol Res Pract 1982;174: 92–105.

27. Kuhajda FP, Hutchins GM. Adrenal cortico-medullary junction necrosis: amorphologic marker for hypotension. Am Heart J 1979; 98: 294–297.

28. Porter JM, Muscato K, Patrick JR. Adrenal hemorrhage: a comparison oftraumatic and nontraumatic deaths. J Natl Med Assoc 1995; 87: 569–571.

29. Didenko VV, Wang X, Yang L, Hornsby PJ. Expression ofp21WAF1/CIP1/SDI1 and p53 in apoptotic cells in the adrenal cortex andinduction by ischemia/reperfusion injury. J Clin Invest 1996; 97: 1723–1731.

30. Shoemaker WC, Peitzman AB, Bellamy R, et al. Resuscitation from severehemorrhage. Crit Care Med 1996; 24: S12–S23.

31. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed celldeath in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol1992; 119: 493–501.

32. Didenko VV, Hornsby PJ. Presence of double-strand breaks with single-base 3* overhangs in cells undergoing apoptosis but not necrosis. J Cell Biol1996; 135: 1369–1376.

33. Didenko VV, Tunstead JR, Hornsby PJ. Biotin-labeled hairpin oligo-nucleotides. Probes to detect double-strand breaks in DNA in apoptoticcells. Am J Pathol 1998; 152: 897–902.

34. Dong Z, Saikumar P, Weinberg JM, Venkatachalam MA. InternucleosomalDNA cleavage triggered by plasma membrane damage during necrotic cell


death: involvement of serine but not cysteine proteases. Am J Pathol 1997;151: 1205–1213.

35. Kovacs K, Carroll R, Tapp E. Temporary ischaemia of the rat adrenalgland. J Pathol Bacteriol 1966; 91: 235–240.

36. Wichterman KA, Baue AE, Chaudry IH. Sepsis and septic shock—a reviewof laboratory models and a proposal. J Surg Res 1980; 29: 189–201.

37. Aho HJ, Koskensalo SM, Nevalainen TJ. Experimental pancreatitis in therat. Sodium taurocholate-induced acute haemorrhagic pancreatitis. Scand JGastroenterol 1980; 15: 411–416.

38. Szabo S, Huttner I, Kovacs K, Horvath E, Szabo D, Horner HC.Pathogenesis of experimental adrenal hemorrhagic necrosis (apoplexy).Ultrastructural, biochemical, neuropharmacologic, and blood coagulationstudies with acrylonitrile in the rat. Lab Invest 1980; 42: 533–546.

39. von Asmuth EJ, Maessen JG, van der Linden CJ, Buurman WA. Tumournecrosis factor á (TNF-á) and interleukin 6 in a zymosan-induced shockmodel. Scand J Immunol 1990; 32: 313–319.

40. Southard JH. Development and optimization of preservation solutions. In:Lemasters JJ, Oliver C, eds. Cell Biology of Trauma. Boca Raton: CRCPress, 1995; 289–297.

41. Parker SB, Eichele G, Zhang P, et al. p53-independent expression of p21Cip1

in muscle and other terminally differentiating cells. Science 1995; 267:1024–1027.

42. Naseeruddin SA, Hornsby PJ. Regulation of 11â- and 17á-hydroxylases incultured bovine adrenocortical cells: cyclic adenosine 3*,5*-monophosphate,insulin-like growth factor-I, and activators of protein kinase C.Endocrinology 1990; 127: 1673–1681.

43. Didenko VV, Hornsby PJ. A quantitative luminescence assay for non-radioactive nucleic acid probes. J Histochem Cytochem 1996; 44: 657–660.

44. Yang L, Didenko VV, Noda A, et al. Increased expression of p21Sdi1 inadrenocortical cells when they are placed in culture. Exp Cell Res 1995; 221:126–131.

45. Malek L, Sooknanan R, Compton J. Nucleic acid sequence-based amplifi-cation. In: Isaac PG, ed. Methods in Molecular Biology, Vol. 28: Protocolsfor Nucleic Acid Analysis. Totowa, NJ: Humana Press, 1994; 253.

46. Zhu L, Harlow E, Dynlacht BD. p107 uses a p21CIP1-related domain tobind cyclin-cdk2 and regulate interactions with E2F. Genes Dev 1995; 9:1740–1752.

47. Shim J, Lee H, Park J, Kim H, Choi E-J. A non-enzymatic p21 proteininhibitor of stress-activated protein kinases. Nature 1996; 381: 804–807.

48. DeFouw DO, Meyenhofer MF, Rizzo VJ, Henry CBS. Ultrastructuralaspects of capillary barrier function after injury. In: Zikria BA, Oz MO,Carlson RW, eds. Reperfusion Injuries and Clinical Capillary LeakSyndrome. Armonk, NY: Futura Publishing Company, 1994; 3–30.

49. Hinshaw LB. Sepsis/septic shock: participation of the microcirculation: anabbreviated review. Crit Care Med 1996; 24: 1072–1078.

50. Yamamoto K, Loskutoff DJ. Fibrin deposition in tissues from endotoxin-treated mice correlates with decreases in the expression of urokinase-typebut not tissue-type plasminogen activator. J Clin Invest 1996; 97: 2440–2451.

51. Baehner RL, Boxer LA, Ingraham LM. Reduced oxygen by-products andwhite blood cells. In: Pryor WA, ed. Free Radicals in Biology, Vol. 5. NewYork: Academic Press, 1982; 91–114.

52. Suematsu M, Tsuchiya T. Oxidative stress in ischemia–reperfusion: Reilly’sirritation syndrome revisited. In: Montagnier L, Olivier R, Pasquier C, eds.Oxidative Stress in Cancer, AIDS, and Neurodegenerative Diseases. NewYork: Marcel Dekker, 1998; 503–505.

53. Ward JF, Blakely WF, Joner EI. Mammalian cells are not killed by DNAsingle-strand breaks caused by hydroxyl radicals from hydrogen peroxide.Radiat Res 1985; 103: 383–392.

54. Wang J, Walsh K. Resistance to apoptosis conferred by CDK inhibitorsduring myocyte differentiation. Science 1996; 273: 359–361.

55. Duttaroy A, Qian JF, Smith JS, Wang E. Up-regulated p21Cip1 expressionis part of the regulation quantitatively controlling serum deprivation-induced apoptosis. J Cell Biochem 1997; 64: 434–446.

56. Hornsby PJ. Biosynthesis of DHEAS by the human adrenal cortex and itsage-related decline. Ann N Y Acad Sci 1995; 774: 29–46.

57. Hornsby PJ. The biosynthesis of DHEA by the adrenal cortex and itsage-related decline. In: Watson RR, ed. Health Promotion and Aging. TheRole of Dehydroepiandrosterone (DHEA). Reading: Harwood AcademicPublishers, 1999; 1–13.

58. Semple CG, Gray CE, Beastall GH. Adrenal androgens and illness. ActaEndocrinol (Copenhagen) 1987; 116: 155–160.

59. Lindh A, Carlstrom K, Eklund J, Wilking N. Serum steroids and prolactinduring and after major surgical trauma. Acta Anaesthesiol Scand 1992; 36:119–124.

J. Pathol. 189: 119–126 (1999)

Documents

DNA damage and p21WAF1/CIP1/SDI1 in experimental injury of the rat adrenal cortex and trauma-associated damage of the human adrenal cortex