An Evaluation of Renal Tubular DNA Laddering in Response to Oxygen Deprivation and Oxidant Injury1

M. Iwata, D. Myerson, B. Torok-Storb, and R.A. Zager

cols. However, neither protocol induced ethidium

M. Iwata, B. Torok-Storb, R.A. Zager, Department of bromide- or tdt-detectable DNA laddering. It was Medicine, University of Washington and the Fred concluded that: ( 1) minimal DNA laddering develops Hutchinson Cancer Research Center, Seattle, WA postischemia, and this change is reliably detected only by the tdt method; (2) it correlates with the D. Myerson, Department of Pathology, University of Washington and the Fred Hutchinson Cancer Re- morphologic expression of tubular necrosis, not ap- search Center, Seathe, WA optosis; and (3) in vitro oxidative- and energy deple- tion-medlated proximal tubular cell death can be (J. Am. Soc. Nephrol. 1994; 5:1307-1313) dissociated from DNA ladder formation.

Key Words: Apoptosis. endonuclease. hydrogen peroxide. ABSTRACT ischemia. hypoxia It has recently been suggested that endonuclease activation and/or apoptosis, possibly triggered by poptosis is an internally regulated program by oxidant stress, are Important pathogenetic mecha- A which cells can be eliminated from tissues in a nisms in oxygen deprivation/reoxygenatlon-induced controlled fashion in response to a variety of stimuli proximal tubular cell death. To explore this possibility, (reviewed in References 1-5). It has been widely pro- DNA laddering,” a characteristic feature of these posed as a mechanism for the elimination of cells processes, was sought in: ( 1) postlschemic rat kidneys during development, for the removal of senescent

(25- or 40-mm arterial clamping; 0, 1, 4, 8, 24, and 48 h cells, and for the control of excessive proliferation and 6 days reflow); (2) posthypoxic isolated rat prox- during tissue regeneration. It may also be Induced by imal tubular segments and (3) cultured human kid- a variety of disease processes (e.g. , neurodegenerative ney proximal tubular cells (HK-2) subjected either to disorders, radiochemotherapy, oxidative stress), by energy depletion plus Ca2 overload (antimycin A selected regulatory hormones (e.g. , glucocorticoid ad- plus 2-deoxyglucose plus Ca2 ionophore A23187), or dition to lymphocytes), or by the withdrawal of specific trophins (e.g. , interleukins, estrogen. testosterone) to H202-induced cell death. DNA was subsequently (1-6). Irrespective of the stimulus, strikingly similar extracted, electrophoresed through agarose gels, morphologic changes result, most notably chromatin and visualized with ethidium bromide or Southern condensation and nucleolar degeneration, nuclear blotting. To maximize ladder detection, DNA samples compaction, and decreased cell volume (3). Subse- were also end-labeled with 32P dideoxyadenosine quently, the involved cells fragment into apoptotic triphosphate with terminal deoxynucleotidyl trans- bodies” that are rapidly eliminated, either by phago- ferase (tat), followed by electrophoresis. None of the cytosis or by in situ degeneration. From a pathogenetic postischemlc DNA samples demonstrated any lad- standpoint, apoptosis is believed to differ markedly dering by either ethidium bromide staining or South- from ischemic on toxic cell death (7): the latter are em analysis (apoptotic lymphocyte DNA was a post- externally, not internally, controlled: they usually af- five control). However, trace laddering was apparent fect large numbers of contiguous cells in a relatively nonspecific fashion; and early cell swelling, not con- by the tat technique, commencing at 1 h of reflow, densation, results. peaking at 24 h, and resolving slowly thereafter. This Although the precise mechanisms that mediate ap- finding correlated with the morphologic expression of optosis are poorly defined, a characteristic feature of it tubular necrosis, not apoptosis. Hypoxia/reoxygen- is endonuclease activation, which results In the deg- ation caused proximal tubular segment death (44 to radation of genomic DNA at intennucleosomal linker 64%), and HK-2 cells were slowly killed by both the regions. This process generates 180- to 185-base-pain H2O2 and the energy depletion/Ca24-Ioadlng proto- fragments, which by agarose gel electrophoresis, yield a “laddered” DNA appearance (e.g., References 4 and 1 ReceIved March 25, 1994. Accepted June 15, 1994. 6). Because the morphologic changes of apoptosls are 2 corresponcience to Dr. R.A. Zoger, Fred Hutchinson Cancer Research Center, extremely transient and, thus, are easily missed, DNA 1 124 ColumbIa Street, Seattle, WA 98104. laddening” has become widely accepted as a marker 1046.6673/0506-1307$03.00/0 of apoptotic cell death. Journal of the American SOCiety of Nephrology Copyright C 1994 by the American SOCIetY of Nephrology In 1992, Schumer et aL reported that DNA,

Journal of the American Society of Nephrology 1307 Renal Tubular DNA Laddenng

extracted from rat kidneys 1 2 to 48 h postischemia ing served as negative controls. As a positive control, (5-, 30-, 45-mm arterial clamping), generated charac- laddered DNA was obtained from 2B4 lymphocytes, which teristic apoptotic ladders, as assessed by agarose gel were rendered apoptotic by exposure to 1 M dexarnethasone electrophoresis with ethidium bromide staining (8). for 24 h (N = two preparations) (12). In an effort to correlate the results with the morphologic On the basis of those observations, it has been sug- appearance ofthe kidneys from which the DNAwas obtained, gested that apoptosis may be triggered by ischemia, coronal kidney slices were taken after each of the reflow potentially contributing to the development of post- periods, they were fixed by immersion in either 10% buffered ischemlc tubular cell death and, hence, acute renal formalin or methyl Carnoy’s, and subsequently, 5-sm sec- failure (ARF) (8,9). To gain support for this intriguing tions were cut and stained with hematoxylin and eosin. They possibility, our laboratory recently used a newly de- were examined by a pathologist (D. Myerson) for the appear- scribed histocytochemical technique (10) to ascertain ance of tubular necrosis and apoptosis (as defined above) whether DNA damage can be observed in situ in without knowledge of the specific experimental protocol that postischemic rat kidneys (1 1). ThIs method uses for- had been undertaken. main-fixed tissues that are incubated with terminal In Vitro 02 Deprivation Injury: Isolated Proximal deoxynucleotidyl transferase (tdt) and biotinylated de- Tubular Segment Experiments oxyunidine, the latter being incorporated into the 3’-OH ends of DNA, exposed with DNA breaks (10). Since the onset of tubular cell death cannot be precisely Subsequently, the deoxyunidlne signal is amplified timed in vivo by morphologic assessments, It was impossible with avidin penoxidase. allowing for its detection by to discern from the in viva experiments whether DNA frag- mentation precedes the loss of cell viability, as would be conventional light microscopy. The results of that expected If It were helping to mediate tubular cell death. To study indicated that DNA breaks can be found within gain possible insights Into this Issue, DNA laddening in postischemic rat kidneys; however, they appeared to response to hypoxic-reoxygenation injury was sought In be limited to those tubular cells that were already isolated rat proximal tubular segments (FF5), because the overtly necrotic, suggesting that the DNA changes loss of cell viability can be precisely timed in this system by were a result, rather than a mediator, of postischemic lactate dehydrogenase (LDH) release. Three sets of FF5, tubular cell death. Furthermore, although DNA prepared as previously described (13). were each divided into breaks were observed in that study, in situ detection three equal aliquots as follows: ( 1) continuous oxygenation with the tdt method does not necessarily mean that (95% 02-5% CO2 incubation for 45 mm); (2) hypoxic injury (95% N2-5% CO2 incubation for 30 raIn): and (3) 30 mm of characteristic apoptotic laddering had resulted. In hypoxia followed by 15 mm ofreoxygenation, each performed other words, the DNA changes could have reflected as previously described (13). At the completion of each nonspecffic DNA degradation, rather than endonucle- incubation, percent LDH release was determined (the ase-mediated ‘apoptotic” death. amount of total tubular LDH released into the PTS bathing As recently suggested by Bonventre (9), the docu- media) (13). Finally, the FF5 suspension was pelleted by mentation of apoptosls, or DNA laddening, in early centrifugation, and the DNA was extracted and analyzed with postischemic renal tissues could substantially alter ethidium bromide-stained gels and the tdt method, as de- our current thinking about the mechanisms of post- scrlbed below. ischemic ARF. Therefore, the goals of this study were Energy Depletion-Induced Cell Death in to confirm the presence of laddered DNA in postis- chemic rat kidneys by a variety of methods and to Cultured Human Proximal Tubular Cells: further explore its pathogenetic significance with Evaluation of DNA Ladder Formation three in vitro models of proximal tubular cell death. To further assess whether DNA fragmentation is an early pathogenetic factor in energy depletion-induced tubular cell METHODS death, additional experiments were performed with a newly described cultured human proximal tubular cell line, HK-2 In Vivo lschemia/Reperfusion Experiments (14). The cells were grown in flasks in keratinocyte-serum Male Sprague Dawley rats (175 to 275 g) were anesthetized free medium (K-SFM) as previously described (14). When with pentobarbital (30 to 40 mg/kg Ip), a midline laparotomy confluent, they were trypsinized (0.05%: followed by fetal calf was performed, and the renal arteries were then occluded serum quenchlngJ, and the detached cells were recovered by with smooth vascular clamps. To create different degrees of centnifugation and then plated in six-well plates at a density ischemlc tissue damage. the left and right kidneys ofeach rat of -0. 15 million cells per well, each containing 3 mL of were occluded for 25 and 40 mm, respectively. After the K-SFM. The cells were cultured overnight at 37#{176}Cwith 5% ischemic insults were completed, the kidneys were either CO2. and then, they were incubated either under control immediately removed without allowing reperfusion or vail- culture conditions or in the presence of a mitochondrial able lengths of reflow were permitted (1, 4, 8, 24, 48 h or 6 respiratory inhibitor (7.5 M antimycin A lOOx stock solu- days: N = two to four rats for each time frame). Rats tion. in 1% ethanol) plus 20 mM 2-deoxyglucose (a compet- undergoing 8 h or more of reperfusion were allowed to itive inhibitor of glucose uptake). Because pilot studies have recover from anesthesia, with free food and water access shown that the combination ofantimycin A plus 2-deoxyglu- being provided. At the appropriate reperfusion time, the cose Induces -90% AlP depletion in HK-2 cells, but without kidneys were resected, and DNA was isolated and analyzed a loss of cell viabifity, the cells were concomitantly incubated with ethidlum bromide-stained agarose gels, Southern anal- with a calcium lonophore (5 M A23187; Sigma Chemical ysis, and/or the tdt method, as described below. DNA ex- Co., St. Louis, MO), which does cause lethal HK-2 injury, but tracted from normal rat kidneys during each tissue process- only in the presence ofan energy depletion state. After 1-, 4-,

1308 Volume 5 - Number 6 - 1994 Iwata et al

8-, 24-, and 48-h incubations, the cell viability of the control apoptotic lymphocytes were also analyzed by Southern anal- and challenged cells was determined In representative wells ysis in an effort to optimize DNA ladder detection. These gels by rital dye exclusion (the percentage ofcells with nuclei that (N = 2 each) were transferred to nylon and subjected to stained positive within 5 mm of 50 pg/mL ethidlum bromide blotting hybridization under low-stringency conditions with addition: total cell numbers were determined by counter- 32P-labeled rat or human genomic DNA probes (20). They staining with 15 g/mL acridine orange in India ink). The were exposed for 3 days on X-OMT AR-5 ifim (Kodak, Roch- cells in the remaining wells were mechanically detached from ester, NY) , with the use of an intensifying screen to maximize the plastic and centrIfuged; the DNA was isolated and ana- ladder detection. lyzed with ethidlum bromide-stained gels and the tdt The tdt method was used to analyze at least two DNA method, as described below. samples obtained from each of the following protocols: nor- mal and ischemic/postischemlc kidneys (25 and 40 mm of H202-Induced Oxidant Stress of HK-2 Cells: ischemla/no reflow: or 1, 4, or 24 h or 6 days ofreflow): three Evaluation of DNA Fragmentation complete sets of VT’S aliquots: and all HK-2 cell experiments. Approximately 2 g of purified DNA was quantitatively end- Because postischemlc tubular necrosis has been reported labeled with 32P dideoxyadenosine tniphosphate (Amersham, to be mediated, in part, by Iron-dependent oxidant stress Arlington Heights, IL). with tdt (Boehringer Mannheim, Indi- (e.g., References 15 and 16) and because oxidant stress can anapolis, IN), according to the manufacturer’s instructions. clearly trigger DNA damage (e.g.. References 6 and 17). the Subsequently, the labeled DNA samples were electropho- following experiments were undertaken to ascertain whether resed through 2% agarose gels. dried under a vacuum, and the Iron-dependent oxidative death of HK-2 cells correlates autoradlographed. with endonuclease-mediated double-stranded DNA breaks. To this end, the effects ofH2O2 on HK-2 cell viability and DNA RESULTS AND DISCUSSION ladder formation were assessed, because it has previously been shown that H2O2 induces HK-2 cell death via an Irrespective of the length of ischemia (25 or 40 min) Iron-dependent mechanism (14). Plates of HK-2 cells were or reperfusion (0 mm to 6 days), none of the kidney prepared as noted above. and then, they were incubated samples obtained from the in viva experiments ap- under control conditions or in the presence of 5 mM H202 for peared degraded or generated apoptotic ladders. as 8, 24, or 48 h. After the completion of each incubation time, assessed by the ethidium bromide-stained gels (Fig- percent cell viabifity was assessed in control and H202- ure 1, top). This indicates that, at the very most, only exposed wells by ethidium bromide staining, as denoted above. Cells in the remaining wells were used to harvest DNA samples for subsequent electrophoretic analysis, as detailed 25 mm of ischemia 40 mm of ischemia below.

. reperfusion PTS-l reperfusion PTS-2 DNA Isolation and Analysis .-i n e eiM.e Total DNA was isolated from whole kidney tissues, FF5, and HK-2 cells by the technique ofTilly and Hsueh (18). In brief, the tissues were snap frozen in liquid nitrogen and stored at -80#{176}C.Subsequently, they were suspended in 4 vol of a homogenizing buffer (0. 1M NaCl, 0.2 M sucrose, 10 mM EDTA, 0. 1 M Ths: pH 8.0) and homogenized with a Polytron tissue dispenser. One-tenth volume of 10% sodium dodecyl sulfate was added and incubated at 65#{176}Cfor 30 mm. One- third volume of 4 M potassium acetate was added, the B homogenate was mixed and incubated for 1 h at 4#{176}C,and the cellular debris and protein precipitates were pelleted by centrifugation (5,000 x g for 10 mm: 4#{176}C).The supernatant was extracted two times with 2 vol ofphenol:chloroform (1:1), followed by reextraction with 2 vol of chloroform. The aque- ous phase was collected: the DNA was precipitated with the addition of3 vol ofchffled ethanol and then held at -80#{176}Cfor more than 1 h. The DNAwas pelleted by microfuging (14,000 Figure 1 . (A and C) Ethidlum bromide-staIned gels contain- rpm X 30 mm: 4#{176}C);it was then washed with 70% ethanol ing DNA from control kidneys, from kidneys that were 8, 24, and air dried for 10 mm. The material was resuspended in 50 or 48 h or 6 days postischemla. and from PTS subjected to L ofTE (10 mM Ths (pH 8.0J, 1 mM EDTA) and incubated at continuous oxygenation (oxy) or hypoxia/reoxygenatlon 37#{176}Cfor1 h with 1 ML of DNAase-free RNAase. The phenol: (hyp) injury. None of these samples demonstrated laddered chloroform extractions were then repeated. Finally, the DNA was precipitated with ethanol and solublllzed in 25 jL of DNA. In contrast, DNA from dexamethasone-freated 214 one-tenth strength TE. DNA concentrations were determined lymphocytes generated a characteristic ladder pattern, by absorbance at 260 nm. serving as a positive control. M, molecular weight markers DNA samples (either 2 or 5 gJ were electrophoresed (range, 220 to 12,200 base pairs). (B and D) Southern anal- through 1.6% agarose containing ethidium bromide (19), yses of the above positioned gels. Slight DNA degradation Is with laddered phage DNA (220 to 12,200 base pairs) being seen in one sample (24 h post-25 mm of ischemia), but no used as molecular weight markers. Gels containing samples laddering was apparent. In contrast, laddering was seen in from normal kidneys, postischemic kidneys (8, 24, or 48 h or the samples obtained from dexamethasone-freated 214 6 days after 25 or 40 rain of ischemla), the FF5, and the lymphocytes.

Journal of the American Society of Nephrology 1309 Renal Tubular DNA Laddering

trace amounts of laddered DNA develop in the after- math of ischemic renal injury. These negative C ethidium bromide results cannot be readily ascribed C to a technical artifact because laddening was observed 25 mln/0 with this method when DNA samples from apoptotic 40 min/0 25 mm/i h lymphocytes were studied (Figure 1 , top). Of note, Schumer et aL previously reported extensive DNA 40 mm/i h laddering in postischemic rat kidneys using ethidlium bromide-stained gels (8). Thus, our findings stand in contrast to their results. The reason for this discrep- min/4 h ancy remains unknown. However the fact that they min/4 h found prominent laddening even in kidneys subjected to as little as 5 mm of ischemia and that 25 min of C lschemia is generally required to induce any in vivo C

tubular cell death raises a question as to whether 25 min/24 h those results predominantly represented in viva, ver- sus ex viva, DNA degradation. 40 min/24 h In a further attempt to document DNA laddering in the aftermath of postischemic kidney injury, DNA from our in vivo experiments were also subjected to 25 min/6 D Southern analysis. Again, no clear laddering was de- tected, even by this method (Figure 1, bottom). In contrast, DNA extracted from apoptotic lymphocytes 40 min/6 D did generate ladders by this technique, even when the sample applied to the gel was reduced from 5 to 0.2 p.g Figure 2. tdt analysis of DNA extracted from normal kidneys of DNA. The fact that no laddering was seen when this (C) and from kidneys subjected to ischemia and variable reduced amount of DNA was analyzed with ethidium lengths of reperfuslon (the lengths of lschemla/reperfusion bromide-stained gels confirmed the greater sensitivity are given as minutes/hours). Trace amounts of laddered of the Southern technique. Thus, the results with DNA were apparent at 1 h of reflow, with increasing amounts Southern analysis substantiated the ethidium bro- being seen at 4 and 24 h. By 6 days postlschemia, the vast mide results, indicating that, at most, only trivial majority of laddered DNA had resolved. Greater amounts of amounts of laddered DNA develop post-renal is- laddered DNA were seen in the kidneys subjected to the chemia. 40-mm, versus the 25-mm, ischemic insult at 4 h or more of Given the completely negative results with ethidium reflow. (Of note, only 3-OH ends/breaks react with tdt. Thus, bromide-stained gels and Southern analyses, the intact DNA generates a much weaker signal than DNA that even more sensitive tdt method was used to define has fragmented. Therefore, the absolute percentage of frog- whether any laddered DNA existed within our post- mented DNA cannot be directly calculated by densifomefry ischemic kidney samples. The fact that we had prey- and the relative proportions of fragmented versus nonfrag- ously documented in situ postischemic DNA damage mented DNA are not accurately reflected). by this technique (1 1) suggested to us that It should have sufficient sensitivity to prove whether or not that damage reflected laddered DNA. Neither the normal during early reperfusion (1 to 4 h), increases progres- kidney samples nor those extracted from kidneys sively over the first 24 h, and resolves slowly theneaf- subjected to ischemia/no reflow demonstrated any ten; and (4) its extent appears to reflect the amount of fragmented DNA. However, by 1 h of reflow. trace postischemic tissue damage. laddering was seen in some of the samples, and the To discern the temporal relationship between tdt- amount increased progressively over the ensuing 23 h detectable laddening and the onset of tubular morpho- (Figure 2). By 6 days of reflow, a time correlating with logic injury, renal histologic sections were examined renal functional recovery (e.g., References 1 1 and 21), from 1 to 24 h postischemia. Proximal tubular cell the vast majority of the tdt-detectable DNA damage injury was clearly discernible by 1 h of reflow (brush had disappeared. Finally, the extent of laddening border membrane blebbing and shedding), and overt seemed to correlate with the severity of ischemic tubular necrosis was first seen at 4 h of reflow, injury, because at 4 h and 24 h and 6 days of reflow, reaching its maximum 20 h later. Thus, the onset and the changes were noticeably greater in samples ex- extent of the laddening appeared to reflect the degree tracted from kidneys that had undergone the 40 min, of tubular injury, suggesting that it may have been a versus the 25 mm, ischemic insult. Thus, the tdt secondary consequence of that injury, rather than experiments indicated that: ( 1) laddered DNA does, in being a primary mediator ofit. Indeed, we reached this fact, develop in postischemic rat kidneys; (2) the same conclusion in our former study of the in situ tdt amount generated is exceedingly small, because only analysis of postischemic kidneys (1 1). In that study, the tdt method could detect it; (3) it commences tdt-detectable DNA fragmentation was seen only in

1310 Volume 5 - Number 6 ‘ 1994 Iwata et al

cells that were overtly necrotic, and not in their mor- degree of energy depletion alone, a superimposed phologically intact counterparts, suggesting that the Ca2 ionophore challenge was used (22), which only DNA changes occurred as a secondary response to causes HK-2 cell death when they are in an energy- lethal cell damage. Of note, none of the 1- to 24-h depleted state. As shown in Figure 4, this combined postischemic kidney sections demonstrated clear challenge induced a time-dependent model of HK-2 morphologic evidence of apoptosis. As we previously cell death. However, at no time during the course of reported, apoptotic cells can be observed within post- the experiments did DNA laddening result (Figure 5). ischemic kidneys (1 1). However, they appear predom- Thus, these findings completely support the conclu- inantly during the late recovery period (5 to 6 days sions from the isolated tubule experiments by proving postischemia), they are very infrequent and largely that energy depletion/Ca2 overload-mediated cell confined to occasional nodules of regenerating epithe- ha, and they undoubtedly reflect a pathway for the elimination ofredundant cells, produced as part of the oxy regenerative response (1 1). Thus, although apoptosis hyp may be an Important event in postischemic ARF, it hyp/reoxy seems more likely that its dominant role is during the oxy recovery, rather than the initiation, phase of injury. hyp To further gauge the temporal relationship between Figure 3. tdt analysis of Isolated PTS DNA subjected to con- DNA laddering and the loss of tubular cell viabifity tlnuous oxygenation (oxy), hypoxia (hyp). or hypoxla/ (thereby assessing a possible cause and effect rela- reoxygenation injury (hyp/reoxy). tdt-detectable ladders tionship), the isolated FFS experiments were per- were apparent in all oxygenated allquots, but the amount of formed. Of note, because the loss of cell viabifity in the laddenng was not Increased In response to hypoxla/ this system can be precisely timed by LDH release, reoxygenatlon. these experiments should permit a precise determina- tion of whether DNA laddening precedes, or results from, tubular cell death. Those studies revealed the 0) following noteworthy results: (1) clear DNA ladders E could not be documented in oxygenated or hypoxic/ posthypoxic FF5 aliquots when assessed by ethidium 0 100 bromide staining or Southern analysis (Figure 1); (2) E 80 tdt-detectable laddering was observed in all continu- ‘D ___i, ----- ---- .--. Controls .c: ‘ .... #{149}‘H2O2 ously oxygenated FF5 (Figure 3), undoubtedly reflect- w 60 -1 -----. 22 ATP depletion/Ca ing DNA damage sustained during the tubular isola- 0) C ‘ ..... ionophore tion process. (Of note, we have documented this same ‘O 40 finding in P1’S isolated in another laboratory IT.J. C,) x “ Burke, Denver, COl, confirming that this isolation uJ 20 \ - damage is not unique to our laboratory); and (3) U) 0) n ,...... - hypoxia/reoxygenation induced 44 to 64% tubular I 0 0 12 24 36 48 cell death (LDH release), and despite this, there was Time in Hours absolutely no increase in the amount of tdt-detectable laddered DNA over that seen in the oxygenated con- Figure 4. TIme course of HK-2 cell death in the H202 (8. 24, trols (10 to 15% LDH release) (Figure 3). Thus, these and 48 h) and the energy depletion + Ca2 ionophore (1,4, results clearly demonstrate that hypoxia/reoxygen- 8, 24, and 48 h) experiments. Cell vIabIlIty was assessed by ation-induced tubular cell death can occur without vital dye (ethidium bromide) exclusion (see Text). Note the developing double-stranded DNA breaks. gradual onset of cell death, reaching 100% at 24 h. DespIte this slow evolution of Injury, no DNA laddering became It should be noted that interpretation of the above apparent (see Figures 5 and 6). FF5 results is somewhat complicated by the fact that baseline tdt-detectable DNA breaks were observed even in those tubules not subjected to hypoxic injury. Because of this, it was not possible to completely dissociate hypoxia/reoxygenation cell death from lad- C lh dered DNA. Thus, we next studied the evolution of 4h energy depletion-induced proximal tubular cell death 8h in HK-2 cells, because no background DNA damage 24h exists in this cell culture system. To this end, HK-2 48h cells were incubated with a mitochondrial respiratory inhibitor (antimycin A) plus an inhibitor of glucose Figure 5. tdt analysis of normal HK-2 cells and HK-2 cells uptake (2-deoxyglucose), a protocol shown in pilot subjected to energy depletion + Ca2 lonophore treatment. experiments to induce ‘-‘90% AlP depletion. However, No DNA Iadderlng developed at any time during the course because lethal cell injury does not result from this of these experiments. C. control incubation with K-SFM.

Journal of the American Society of Nephrology 1311 Renal Tubular DNA Laddering

dens, the pathogenetic significance of this finding a C 48h remains open to interpretation. This is because the bulk of the DNA damage appears concomitantly with . the morphologic appearance of necrosis, rather than 448h proceeding it (suggesting that it may be a secondary, rather than being a primary event), and because the in Figure 6. tdt analysis of HK-2 cell DNA under control incuba- vitro studies presented here demonstrate that both ton conditions and after 8, 24. and 48 h of H202 Incubation. energy depletion/Ca2 overload- and oxidant-trig- No laddering was apparent, despite 100% cell death at both gered proximal tubular cell death can develop in the 24 and 48 h. The amount of DNA applied was progressively absence of laddered DNA. This does not mean that increased at each time (from 2 up to 5 g. as depicted by DNA damage is irrelevant to these forms of injury. the Increasingly broader bands of intact DNA near the top of However, we do believe that double-stranded DNA the gel). Despite this, no DNA laddering were detected. breaks are not necessarily critical to the initiation of them in all circumstances, as demonstrated by this report. death can, indeed, occur In the absence of endonucle- ase-medlated double-stranded DNA breaks. ACKNOWLEDGMENTS The final goal of this study was to assess whether We thank Ms. K. Burkhart and Mr. B. Schlmpf for their expert technical assistance and Thomas Burke, Ph.D. , University of Cob- H2O2-induced/fron-dependent HK-2 cell killing oc- rado Health Sciences Center, for providing us with samples of proxi- curs via an endonuclease-dependent pathway. The mal tubular segments. This work was supported by a Grant DK- reason for addressing this issue is that we did docu- 38432 from the NIH. ment tdt-detectable DNA breaks during early vascular reperfuslon (Figure 1 ; Reference 1 1), a time of pur- REFERENCES

ported critical iron-dependent oxidant tissue stress 1 . Ellis RE, Yuan J, Horvitz R: Mechanisms and functions (1 5, 16). Thus, had we been able to document HK-2 of cell death. Annu Rev Cell Biol 1991:7:663-698. 2. Collins MKL, Lopez Rivas A: The control of apoptosis in DNA laddering with oxidant injury, this would provide mammalian cells. Trends Blochem Sd 1993:18:307- a possible pathogenetic insight into both our past (11) 309. and these in viva results. To this end, HK-2 cells were 3. Mends MJ, Wyllie All: Apoptosis. The role of endonucle- incubated with 5 mM H202, which induced gradual ase activation. Am J Pathol 1990; 136:593-608. 4. Wyffle AH: Glucocorucoid-induced thymocyte apoptosis cell death over the course of the experiment (Figure 4). Is associated with endogenous endonuclease activation. However, absolutely no DNA laddening resulted (Fig- Nature (Lond) 1980:284:555-556. ure 6), suggesting that a nonapoptotic pathway was 5. Kerr JFR, Wyllie AH, Currie AR: Apoptosis: A basic biological phenomenon with wide ranging implications In involved. Of note, H2O2 has previously been reported tissue kinetics. Br J Cancer 1972:26:239-257. to trigger endonuclease-dependent cell death in LLC- 6. Ueda N, Shah SV: Endonuclease-induced DNA damage PK1 cells, another proximal tubular cell line (6). Thus, and cell death in oxidant injury to renal tubular epithe- lial cells. J Clin Invest 1992:90:2593-2597. our HK-2 results appear discordant with those data. 7. Searle R, Kerr JFR, Bishop CJ: Necrosis and apoptosis: Although the reason for this discrepancy remains Distinct modes ofcell death with fundlmentally different unknown, it indicates that pathways of H2O2-induced significance. Pathol Ann 1982:17:229-259. 8. Schumer M, Colombel MC, Sawczuk IS, et aL: Morpho- injury may vary considerably, depending on the cell logic, biochemical, and molecular evidence of apoptosis line used. Thus, results obtained with one should be during the reperfusion phase after brief periods of renal interpreted with caution. It should be noted, however, lschemla. Am J Pathol 1992:140:831-838. 9. Bonventre JV: Mechanisms of ischernic acute renal fail- that HK-2 cells are not completely resistant to apop- ure. Kidney mt 1993:43:1160-1 178. totic/endonuclease-mediated cell death, because tdt- 10. Gavriell Y, Sherman Y, Ben-Sasson SA: Identification of detectable ladders do develop when they are main- programmed cell death in situ via specffic labeling of tamed under conditions of growth factor and nutrIent nuclear DNA fragmentation. J Cell Biol 1992:19:493- 501. withdrawal (unpublished data from this laboratory). 1 1 . Zager RA, Fuerstenberg SM, Baehr PH, Myerson D, Thus, the negative results with our H2O2 and energy Torok-Storb B: Antioxidant prophylaxis: An evaluation depletion/cell Ca2-loading experiments cannot sim- of potential effects on recovery from post-Ischemic acute renal failure. J Am Soc Nephrol 1994:4:1588-1597. ply be ascribed to an absence of the pathways neces- 12. Zacharchuk CM, Mercep M, Chakrabortl PK, Simons sary for inducing laddered DNA. SS, Ashwell JD: Programmed T lymphocyte death. Cell In conclusion, this study indicates that: ( 1) trace activation- and steroid-induced pathways are mutually exclusive. J Immunol 1990; 145:4037-4045. amounts of DNA laddening can be documented in the 13. Zager RA, Schlmpf BA, Gmur DJ, Burke TJ: Phospho- aftermath of in viva ischemic renal Injury. However, lipase A2 activity can protect renal tubules from oxygen the amount generated appears to be far too low to deprivation injury. Proc Nail Acad Sci USA 1993:90: permit reliable detection by routine ethidium bro- 8297-8301. 14. Ryan MJ, Johnson G, Kirk J, Fuerstenberg SM, Zager mide-stained agarose gels or even Southern analysis. RA, Torok-Storb B: HK-2: An immortalized proximal Thus, If one wishes to document DNA laddening in tubule epithelial cell line from normal adult human postischemic kidneys, the highly sensitive tdt tech- kidney. Kidney mt 1994:45:48-57. 15. Paller MS Hoidal JR Ferris TF: Oxygen free radicals in nique is likely to be required; and (2) despite the isehemic acute renal failure in the rat. J Clin Invest appearance of tdt-detectable postischemic DNA lad- 1984;74: 1156-1164.

1312 Volume 5 . Number #{243}#{149}1994 Iwata et al

16. Paller MS, Hedlund BE: Role of Iron in postischemlc 20. Ausubel FM, Brent R, Kingston RE, et aL. Eds. Current renal injury in the rat. Kidney Int 1988:34:474-480. Protocols in Molecular Biology. New York: John Wiley 17. Golconda MS. Ueda N, Shah SV: Evidence suggesting and Sons: 1987:2.9.1-2.9.15. that iron and calcium are Interrelated in oxidant- 2 1 . Zager RA, Baltes LA, Sharma HM, Jurkowitz MS: Re- induced DNA damage. Kidney mt 1993:44:1228-1234. sponses of the ischemic acute renal failure kidney to 18. Tilly JH, Hsueh AJW: Microscale autoradlographic additional Ischemlc events. Kidney 1984;26:689- method for the qualitative and quantitative analysis of mt 700. apoptotic DNA fragmentation. J Cell Physiol 1993:154: 519-526. 22. Weinberg JM, Venkatachalam MA, Roeser NF, Davis JA, 19. Sambrook J, Fritsch EF, Maniatis T: Molecular Varani J, Johnson LI: Amino acid protection of cultured Cloning2nd Ed. Cold Spring Harbor, NY: Cold Spring kidney tubule cells against calcium ionophore-induced Harbor Laboratory Press; 1989:6.9-6.19. lethal cell Injury. Lab Invest 1991:65:671-678.

From F)’anz HoWeirc Dance of Death. Death is presenting the flask of urine from the elderly patient to the physician. Originally published in Djons in 1538. Thisflgure was reproducedfrom Holbein’s Dance ofDeath and Bible Woodcuts, The Sylvan Press, New York, 1 947, p. 26. From the collection of the Clendening Library of the Histonj ofMedicine, University of Kansas.

Journal of the American Society of Nephrology 1313