Novel High-Throughput Deoxyribonuclease 1 Assay

Abstract

Deoxyribonuclease I (DNase I), the most active and abundant apoptotic endonuclease in mammals, is known to mediate toxic, hypoxic, and radiation injuries to the cell. Neither inhibitors of DNase I nor high-throughput methods for screening of high-volume chemical libraries in search of DNase I inhibitors are, however, available. To overcome this problem, we developed a high-throughput DNase I assay. The assay is optimized for a 96-well plate format and based on the increase of fluorescence intensity when fluorophore-labeled oligonucleotide is degraded by the DNase. The assay is highly sensitive to DNase I compared to other endonucleases, reliable (Z’ ≥ 0.5), and operationally simple, and it has low operator, intraassay, and interassay variability. The assay was used to screen a chemical library, and several potential DNase I inhibitors were identified. After comparison, 2 hit compounds were selected and shown to protect against cisplatin-induced kidney cell death in vitro. This assay will be suitable for identifying inhibitors of DNase I and, potentially, other endonucleases.

Introduction

Double-strand DNA breaks accumulated beyond DNA repair capacity of the cell are considered “a point of no return” when cell death becomes inevitable.¹ DNA fragmentation is an enzymatic process mediated by an arsenal of DNA-degrading apoptotic endonucleases, including deoxyribonuclease I (DNase I),^2,3 DNase II,⁴ their homologs,^5–7 caspase-activated DNase (CAD),⁸ and endonuclease G (EndoG).^9,10 All apoptotic endonucleases catalyze the same reaction of hydrolysis of phosphodiester bonds. Endonucleases are expressed in all cells and tissues, with marked variation of endonuclease spectra between tissues. Full activation of endonucleases in cells degrades DNA into short oligo- and mononucleotides within a few hours. Studies show that inactivation of cytotoxic endonucleases by gene knockout prior to cell or tissue injury is highly protective, confirming that DNA breaks by endonucleases mediate cell death.^11–14 Another role of apoptotic endonucleases is to protect cells from foreign DNA, such as oligonucleotides from dead cells, gene delivery plasmids, or viral DNA.¹⁵ The search for new endonuclease inhibitors may broaden therapeutic opportunities for regulation of cell death before and after organ injuries, and for the improvement of cancer therapy or gene therapy.

The most abundant apoptotic endonuclease, DNase I, is also the most active endonuclease in mammalian cells.^3,11 It is found in all tissues, although the specific activity of the enzyme varies. It accounts for ~40% to 99% of the total endonuclease activity in most organs, blood, and urine. Kidney tubular and salivary gland epithelial cells express the highest quantities of the enzyme.^3,11 DNase I is a secreted enzyme that cleaves DNA in the digestive system (intestine, pancreas, and salivary glands) and urinary tract. In addition to this extracellular role of the secreted enzyme, in all cells, intracellular DNase I has the responsibility of degrading DNA immediately before and (mostly) after cell death. Being an endonuclease, DNase I cleaves phosphodiester bonds within a polynucleotide chain of double- or single- stranded DNA down to short 3′OH/5′P-end oligonucleotides. It requires Ca²⁺ and Mg²⁺ for its activity, is inhibited by Zn²⁺, and has a neutral pH optimum. Inside the cell, the enzyme is localized in the endoplasmic reticulum, perinuclear space, and cytoplasm.¹⁶

In the absence of tissue injury, a gene knockout of DNase I in CD-1 mice shows no phenotype, indicating that even complete inactivation of DNase I is harmless. DNase I knockout is, however, cytoprotective during toxin-induced injuries to the kidney or liver, or after multi-organ injury caused by total body irradiation.^11,13,17 Nonetheless, we have been unable to capitalize on these findings to improve research tools or patient care, due to the lack of availability of pharmaceutical agents for DNase I inhibition.

The ongoing search for DNase I inhibitors to confer tissue protection against injuries would be significantly improved if a high-throughput screening (HTS) method was available to search through chemical libraries for potential DNase I inhibitors. Current DNase assays are based on the hyperchromicity of degraded DNA,¹⁸ its decreased viscosity,¹⁹ radioactive labeling of 3′OH ends,²⁰ changes of electrophoretic mobility of DNA after degradation,¹⁷ or decreased ability of degraded DNA to bind intercalators.²¹ These methods often have low sensitivity,^18,19,21 take hours to perform,^17,20,21 or cannot be adapted to a microplate format.^17,19–21 Many of them are not specific to endonucleases,¹⁸ and others are hazardous²⁰ or not quantitative.^17,21 Consequently, currently available DNase assays have no practical utility for the HTS of large numbers of chemical compounds.

To overcome this problem, we have developed a new HTS DNase I activity assay that uses a self-quenched fluorescent reporter DNA probe. This assay has the potential to be developed into a fully automated assay that can be used for HTS and used for further studies to identify new preclinical leads.

Materials and Methods

Chemical Library

The chemical library used in the current study contained 1040 compounds that were small molecules (≤500 MW) designed as drug-like compounds for oral delivery, and characterized with regard to structure, purity, and physicochemical properties. Each compound was dissolved in DMSO of high-performance liquid chromatography grade to afford a 10 mM solution. The drug 10 mM solutions were then loaded into 96-well master plates (Thermo, Rochester, NY), 80 compounds per plate. The drug solutions in each master plate well were then serially diluted with DMSO to afford 1 in 10 dilutions of the master plates.

DNase Activity Probe

The substrate used in the high-throughput DNase I screening assay was customized short Cy5.5-labeled DNA oligonucleotide probe AB259.3²² with the sequence 5′-[Cy5.5]AACACTCCGATGAGTGTAGAATGT[Cy5.5][3P]-3′ (Fig. 1A). It forms a hairpin structure with the fluorophores at both ends in close proximity so that they quench each other. When the substrate is degraded by DNase I or other endonucleases, the fluorophores are separated and increase fluorescence signal to the detectable level. The protruding end is the likely preferable target for DNase activity, whereas the hairpin adds stability and allows one to test that the probe has no breaks prior to the experiment by melting–annealing.

Figure 1.

Elaboration of the high-throughput screening (HTS) deoxyribonuclease I (DNase I) assay conditions. (A) A schematic diagram for the substrate, Cy5.5-end-labeled oligonucleotide, and principle of the assay. (B) The effects of the concentrations of calcium, magnesium, and their combinations. (C) Determination of the linear time range of the reaction. (D) Elaboration of the reaction temperature. The change in product/min (pmole/min) for the substrate (0.25 μM) cleavage by DNase I (1.72 nM) was monitored within 20 min. Data represent mean ± standard deviations; n = 4.

High-Throughput DNase I Screening Assay

A reaction mixture was prepared in white 96-well plates (Costar, Corning, NY) as follows: 0.25 μM Cy5.5-labeled oligonucleotide probe AB259.3²², 0.1 mM CaCl₂, 0.3 mM MgCl₂, 10 mM Tris-HCl, pH 7.4, 1 μl compound in DMSO, and nuclease-free water to provide a total volume of 100 μl. The background (negative control) and uninhibited DNase I samples were measured with DMSO only, or DMSO with recombinant human DNase I (1.72 nM) (rhDNase I, Pulmozyme; Genentech, South San Francisco, CA). After the addition of DNase I, fluorescence intensity was kinetically measured on a Bio-Tek Synergy 4.0 plate reader (Bio-Tek, Winooski, VT) at 37 °C, and mean velocity (mRFU/min) within 20 min (if not specified otherwise) was automatically calculated by the plate reader. The background was subtracted prior to the calculation of DNase I activity. The percentage of DNase I activity was calculated using Equation 1:

$$\text{DNase I activity}(\mathrm{\%})=(\text{mean velocity of a compound}/\text{mean velocity of DMSO})\times 100$$ (1)

In similar assays, recombinant murine EndoG (produced in-house) was used at a concentration of 0.14 μM in 0.1 mM MgCl₂, 10 mM Tris-HCl, pH 7.4; and DNase II (Worthington, Lakewood, NJ) (3.32 nM) was tested in 100 mM sodium citrate buffer, pH 5.0. For evaluation of the quality of the assay, Z’ values were calculated using Equation 2:

$$\mathrm{Z}’=1-(3{\text{SD}}_{\mathrm{C}}+3{\text{SD}}_{\mathrm{B}})/({\mathrm{M}}_{\mathrm{C}}-{\mathrm{M}}_{\mathrm{B}})$$ (2)

where M = mean value; SD = standard deviation; C = control; and B = background.²³

Plasmid Incision Assay

A reaction mixture was prepared containing 1μg pECFP plasmid DNA, 2 mM CaCl₂, 5 mM MgCl₂, 10 mM Tris-HCl, pH 7.4, and 0.5 mM dithiothreitol. Test compound (1 μl) in DMSO was added to a desired final concentration (at final concentration of DMSO of 1%). DNase I was then added to a final concentration of 0.86 pM, and the reaction was incubated for 1 h at 37 °C. The reaction was terminated by the addition of 2 μl of 10mM Tris-HCl, pH 7.4, 1% sodium dodecyl sulfate, 25 mM ethylenediaminetetraacetic acid (EDTA), and 7.2 mM bromophenol blue. The samples were run in a 1% agarose gel in Tris–acetate– EDTA buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8), at 7 V/cm for 35 min, and DNA was stained with ethidium bromide. An EagleEye scanning densitometer (Stratagene, La Jolla, CA) was used to quantify the relative amount of endonuclease-treated plasmid DNA present as covalently closed circular (supercoiled) DNA, open circular DNA, or linear DNA, or in a digested form. One unit was defined as the amount of endonuclease capable of converting 1 μg of covalently closed supercoiled plasmid DNA to open circular, linear, or digested DNA in 1 h at 37 °C. ImigeJ1.44p (US National Institutes of Health, Bethesda, MD) was used to quantitate gel image. The gel image was set at an 8-bit mode prior to quantification, and supercoiled DNA bands were selected and plotted followed by measurements of each peak area.

Cell Culture

Normal rat tubular epithelial NRK-52E cells (ATCC, Manassas, VA) were grown in Dulbecco’s Modified Eagle’s Medium (DMEM; ATCC) supplemented with 5% fetal bovine serum at 5% CO₂/95% air in a humidified atmosphere at 37 °C, fed at intervals of 48–72 h, and used within 1 d after confluence.

Cell Death Assay

To determine their cytoprotective effect, potential DNase I inhibitors were examined in the lactate dehydrogenase (LDH) release assay (CytoTox96 Non-Radioactive Cytotoxicity assay kit; Promega, Madison, WI). NRK-52E cells (8000–10,000 per well) were grown in 96-well plates at 37 °C for 24 h followed by 2 h incubation in the presence of serial dilutions of the potential DNase I inhibitors. Cisplatin (60 μM) was then added to the cells, and after 24 h incubation, LDH release was measured as described previously.²⁴

Cell Extraction

Cells were grown to ~80% confluence in a 10 mm culture dish. Medium was aspirated, and the cells were rinsed once with ice-cold phosphate buffered saline (PBS), pH 7.4. The lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100) was added, and the cells were incubated on ice for 10 min. The sample was then disintegrated on ice by ultrasonication with 2 cycles each of 15 s duration at 25 kHz followed by centrifugation at 13,000 ×g for 10 min at 4 °C. The supernatant was collected and stored at −80 °C prior to use.

Other Enzyme Assays

LDH was measured using the CytoTox 96 Non-Radioactive Cytotoxicity Assay kit (Promega), as described above. A Protease Fluorescent Detection kit, an SOD Determination kit, and a Ribonuclease A Detection kit, all from Sigma-Aldrich (St. Louis, MO), were each used according to the manufacturer’s instructions. Although the SOD determination kit allowed measurement of all SODs, our cell extract preparation suggested that mostly SOD1 activity was measured.

Immunocytochemistry and TUNEL Assay

Cells were fixed with 5% formalin, and immunostaining was performed as described previously.²⁵ The cells were exposed with rabbit anti-EndoG at 1:200 dilution (Millipore, Billerica, MA) in dilution buffer (0.5% bovine serum albumin, 0.05% Tween-20, PBS). The primary antibodies were detected with 1:400 diluted antirabbit IgG-AlexaFluor 647 conjugates (Invitrogen, Carlsbad, CA). Control staining was performed by substituting the primary antibody with the dilution buffer. Cells were then subjected to terminal deoxynucleotidyl transferase mediated dUTP nick end labeling (TUNEL) staining using the In Situ Cell Death Detection Kit (Roche Diagnostics, Indianapolis, IN). Cell were rinsed and counterstained with 4,6-diamidino-2-phenylindole (DAPI) for DNA. Cells were then mounted under cover slips with a Prolong Antifade kit (Invitrogen), and images were acquired using the Olympus IX-51 inverted microscope (Olympus America, Center Valley, PA) equipped with a Hamamatsu ORCA-ER monochrome camera (Hamamatsu Photonics K.K., Hamamatsu City, Japan).

Real-Time Reverse Transcriptase (RT)-PCR

Total RNA from NRK-52E cells was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA), followed by DNA removal using the RNase-free DNase kit (Qiagen). Total RNA (2 μg) was reverse transcribed in a 50 μl reaction using the GeneAmp Gold RNA PCR Core Kit (Life Technologies, Grand Island, NY) according to the manufacturer’s protocol, followed by real-time RT-PCR in a 20 μl reaction using Bio-Rad (Hercules, CA). Reaction mix was prepared using Sso Advanced SYBR Green qPCR Supermix (Bio-Rad) according to the manufacturer’s recommendations. The specific primer sets for EndoG and 18S were as follows: EndoG sense: 5′-GCATGCCTGGAACAACCTTGAGAA-3′; antisense: 5′-TGCCTCCAGGATCAACACCTTGAA-3′; 18s sense: 5′-GGATCCATTGGAGGGCAAGT-3′; and antisense: 5′-ACGAGCTTTTTAACTGCAGCAA-3′. The fluorescence was measured at the end of the annealing step. The melting curve analyses were performed at the end of the reaction (after the 45th cycle) between 60 °C and 95 °C to assess the quality of the final PCR products. The threshold cycle (Ct) values were calculated. The standard curve of the reaction effectiveness was performed using the serially diluted (5 points) mixture of all experimental cDNA samples in duplicates for EndoG and 18s RNA separately. Normalization of EndoG values was performed, and the data were presented as EndoG mRNA expression.

Image Analysis

Image analysis was performed using SlideBook 4.2 software (Olympus America). For quantification, 10 independent fields of view were collected per each subset of the experiment, and mean optical density (MOD) was recorded for used channels. Red color fluorescence was used to mask EndoG in the entire cell population in images. Blue and green colors were used to mask nuclei stained with DAPI and DNA damage detected by TUNEL, respectively.

Statistics

Statistical analysis was performed with Student t test. Results were expressed as mean ± SDs. P < 0.05 was considered significant. Bonferroni modification of the Student t test was applied for the P calculation when appropriate.

Results

Optimization of Reaction Conditions for the Assay

To determine optimal conditions for the assay, we tested the best combination of Ca²⁺ and Mg²⁺ ions, DNase I concentration, time, and temperature of the reaction. Our analysis showed that the highest activity of DNase I could be achieved in the presence of 0.1 mM CaCl₂ and 0.3 mM MgCl₂ (Fig. 1B). In this condition, we tested various DNase I concentrations, and the resulting 1.72 nM DNase I was chosen because higher concentrations significantly decreased the linearity of the assay and increased its variability. The saturation of the reaction was achieved at 30 min (Fig. 1C), so only the linear range from 0 to 20 min was used in subsequent experiments. In a linear regression test, R² from 0 to 20 min was 0.947, indicating statistical linearity (P < 0.0001). The assay was also examined at different temperatures. As the temperature increased, the degradation of oligonucleotides by DNase I elevated and reached saturation at temperatures higher than 37 °C (Fig. 1D). Thus, a temperature of 37 °C was chosen for further experimentation. We also determined the K_m value for the substrate, which was 0.404 μM, according to the Lineweaver–Burk plot.

Determination of IC₅₀ Values of the Known DNase I Inhibitors

Although DNase I inhibitors with therapeutic utility are unknown, several reported DNase I inhibitors are useful for in vitro applications. Therefore, we determined whether our HTS DNase I assay could measure DNase inhibition by these inhibitors. The assay effectively measured the IC₅₀ values of 4 previously reported inhibitors: ZnCl₂ (18 μM), Zn-3,5-diisopropylsalicylic acid (Zn-DIPS; 5.5 μM), EDTA (190 μM), and G-actin (64.8 nM) (Fig. 2).

Figure 2.

Testing the inhibition of deoxyribonuclease I (DNase I) by previously known in vitro inhibitors. To generate dose–response curves, the change in product/min (pmole/min) by DNase I (1.72 nM) was monitored throughout a series of concentrations for each inhibitor. Data represent mean ± standard deviations; n = 4.

Specificity of the HTS Assay for DNase I Compared to Other Endonucleases

To determine the degree to which the HTS DNase I assay is sensitive to different endonucleases, varying concentrations of recombinant DNase I, DNase II, and EndoG were tested by this assay. The substrate titration at fixed concentrations of enzymes was performed, and the data are presented in Figure 3. The k_cat/K_m values of DNase I and EndoG acting on the same substrate were 1.105×10⁶ M⁻¹sec⁻¹ and 0.819×10⁶ M⁻¹sec⁻¹, respectively. It should be noted that the k_cat/K_m value of EndoG is not well defined due to the poor activity on the substrate. DNase II had no activity, and its k_cat/K_m value could not be determined. Thus, the data showed that the assay is more specific to DNase I than toward the 2 other endonucleases.

Figure 3.

Specificity of the high-throughput screening (HTS) assay to deoxyribonuclease I (DNase I). A Michaelis–Menten curve for each of 3 endonucleases (1.72 nM each) was generated for the assay to afford kinetic analysis. k_cat/K_m values of DNase I and endonuclease G (EndoG) acting on the same substrate were determined as 1.105×10⁶ M⁻¹sec⁻¹ and 0.819×10⁶ M⁻¹sec⁻¹, respectively. DNase II had no activity, and its k_cat/K_m value could not be determined. K_m value for the substrate was determined as 0.404 μM. Data represent mean ± standard deviations; n = 4.

Variability of the Assay

Variability of the HTS assay was determined by calculating the coefficient of variation (CV) from independent plates and operators. Two 96-well plates were prepared by 2 operators, and the mean velocity (mRFU/min) was measured to calculate interassay, intraassay, and operator variability, as described in the Methods section. For intraassay variability, 4 sets of experiment in the same plate were performed by the same operator. The results presented in Supplement Table 1 showed that all of the variabilities, including enzyme-catalyzed (signal-max) and non-enzyme-catalyzed (signal-min) ones, were less than 10%, indicating the variability of the assay is within the commonly acceptable range.

Screening of a Chemical Library to Identify DNase I Inhibitors

The assay was tested for screening of a chemical library of 1040 chemical compounds at 2 different final concentrations, 1 μM and 10 μM, against DNase I as described in the Methods section. The final concentration of DMSO in the test solution was 1%. As determined by pilot experiments (data not shown), this amount of DMSO did not interfere with DNase I activity. At 10 μM, 45 potential DNase I inhibitors that produced more than 40% inhibition of DNase I activity were identified (Fig. 4A). Usually, the threefold standard deviation (3×SD) is used to define a threshold. In our case, an SD of a total of 13 plates at 10 μM was 16.03%. Therefore, 3×SD will be 48.1%, meaning that 51.9% of inhibition would be our cutoff, resulting in 23 hits. Considering the small size of the library, however, we chose a slightly lower cutoff of 40% of inhibition, resulting in 45 hits, because we did not want to lose any potential hits. Five of these 45 compounds also inhibited DNase I activity at 1 μM concentration (Fig. 4B). The quality parameter of the assay suitability for HTS and reliability (i.e., the Z’ factor) was calculated according to Zhang et al.²³ at both concentration ranges to ensure that Z’ factor values are 0.5 or higher (Suppl. Table 2).

Figure 4.

Identification of deoxyribonuclease I (DNase I) inhibitor candidates by screening of the chemical library of 1040 compounds. The library was screened at (A) 10 μM and (B) 1 μM concentrations, affording 45 and 5 hits, respectively. The middle line indicates the average signal, and the 2 other lines are 25% higher and lower than the middle line.

Validation of the Inhibitors

To validate the inhibition activities of the candidate compounds, a plasmid incision assay (PIA) was used as a secondary assay to evaluate the 25 most potent compounds (Suppl. Fig. 1). Other compounds had limited solubility in water, which did not allow determination of their IC₅₀ values. Due to the low solubility, they were also expected to have low cell permeability. Thirteen compounds confirmed by PIA assay (Suppl. Fig. 1A) were further tested for IC₅₀ values using the HTS assay. Eleven of these compounds were partially soluble in water; however, the solubility was not sufficient to cover the entire range of the tested concentrations. These compounds were left aside. Therefore, the search resulted in 2 compounds that seemed to have unlimited solubility in water, JR-132 (1,4-phenylene-bis-aminoguanidine hydrochloride) and IG-17 (1,3-phenylene-bis-aminoguanidine hydrochloride) (Fig. 5A and 5B), with IC₅₀ values against DNase I of 2.73 μM and 8.93 μM, respectively (Fig. 5C and 5D). Both compounds were classified as substituted arylidenaminoguanidine analogs with a common structural scaffold. According to Lineweaver–Burk plots for the inhibitors, they were identified as competitive inhibitors (Suppl. Fig. 2).

Figure 5.

Chemical formulas and evaluation of the specificity of the inhibitors to apoptotic endonucleases. (A) JR-132 (1,4-phenylene-bis-aminoguanidine hydrochloride). (B) IG-17 (1,3-phenylene-bis-aminoguanidine hydrochloride). Dose–response curves of (C) JR-132 and (D) IG-17 were performed with deoxyribonuclease I (DNase I; 1.72 nM), deoxyribonuclease II (DNase II; 3.32 nM), and endonuclease G (EndoG; 0.14 μM) to determine IC₅₀ values, which are shown in (E). Data represent mean ± standard deviations; n = 4.

Specificity of the Inhibitors

Because the assay was adapted to be specific for identifying DNase I inhibitors, and used DNase I as the test enzyme, it was important to determine that the resulting hits would in fact be specific to DNase I. To determine this, the IC₅₀ values for the DNase I inhibitors, JR-132 and IG-17, were compared with their IC₅₀ values against 2 other apoptotic endonucleases, DNase II and EndoG. These experiments demonstrated that the inhibitors were somewhat more specific toward DNase I than for the 2 other DNases; however, the difference was not great (Fig. 5C–E). Steep dose– response curves (Hill slopes), observed in some cases in this experiment, can be explained by the potential presence of multiple binding sites in the enzymes. We next tested whether the 2 inhibitors are active toward bacterial nucleases, micrococcal nuclease, and Benzonase. Although the HTS assay was unable to measure the activity of these nucleases, PIA demonstrated the absence of inhibition at 10 μM concentration (Suppl. Fig. 3). When tested for specificity against 4 other enzymes—RNase A, protease, LDH, and SOD—no inhibition activity was observed with any of the test compounds at concentrations of 1 μM and 100 μM (Suppl. Fig. 4). It was concluded that the above inhibitors were specific for apoptotic endonucleases.

Cytoprotection by the Inhibitors against Cisplatin Toxicity

Apoptotic endonucleases, in particular DNase I and EndoG, have been shown to be important for cisplatin toxicity to kidney tubular epithelial cells.^11,26 To test the cytoprotective activities of the 2 arylidenaminoguanidine inhibitors, rat kidney tubular epithelial NRK-52E cells were exposed with cisplatin (60 μM) in the presence and absence of the inhibitors. Cell death was measured by using the LDH release assay. In these experiments, both inhibitors showed some protection against cisplatin-induced cell death (Fig. 6A and 6B). The IC₅₀ value for JR-132 and IG-17 could not be determined because at the highest concentration, these inhibitors exhibited significant cytotoxicity. We also observed marked inhibition of cisplatin-induced DNA fragmentation measured by using the TUNEL assay (Fig. 6C). Knowing that DNase I inhibition may lead to decreased expression of EndoG,¹² we also measured expression of EndoG protein using quantitative immunohistochemistry (Fig. 6D) and expression of EndoG mRNA (Suppl. Fig. 5). This experiment showed that EndoG expression was significantly decreased at both protein and mRNA levels by JR-132 to the level of control. Also, IG-17 inhibited EndoG expression to lower than the control level, indicating that this compound inhibits even non-cisplatin-induced DNA fragmentation.

Figure 6.

Bioprotection of cultured NRK-52E cells against cisplatin (60 μM) by the inhibitors. Cell death was measured by the lactate dehydrogenase (LDH) release assay after 24 h incubation with cisplatin in the presence of varying concentrations of (A) JR-132 or (B) IG-17. Dashed lines show control level of LDH release in nontreated cells. (C) Inhibition of cell death–associated DNA fragmentation by inhibitors at 100 μM after 24 h exposure with cisplatin as measured by TUNEL assay. (D) Endonuclease G (EndoG) protein expression measured by mean intensity of immunohistochemistry. Data represent mean ± standard deviations; n = 4. *p < 0.05.

Discussion

To discover new DNase inhibitors, it will be necessary to carry out HTS screening of large chemical libraries. Current methods for assaying endonuclease activity are, however, too insensitive, lengthy, and imprecise to serve as HTS assays for screening of large chemical libraries. To overcome this problem, we developed a new HTS DNase activity assay that uses a proprietary self-quenched fluorescent reporter DNA probe, AB259.3.²² This assay has been shown to be sensitive, rapid, and precise for measuring DNase I activity in vitro in an HTS assay format for chemical library screening. The assay is based on a mixture of a limited number of individual solutions, and, if desired, it would be amenable to full automation. The assay works with other endonucleases (e.g., EndoG), and, if necessary, the reaction conditions can be adjusted to these enzymes.

Implementation of this new assay in the screening of a library of 1040 compounds has identified 2 new DNase I inhibitors, JR-132 and IG-17, with a similar structural scaffold. These compounds are the most potent DNase I inhibitors known today. Even though the assay was designed to be specific to DNase I and used recombinant DNase I for testing, the 2 inhibitors were also able to suppress the activity of 2 other apoptotic endonucleases, EndoG and DNase II, albeit to a lesser extent. This broad specificity may be due to the use of a relatively small library of molecules, which would not allow the structural diversity needed for the identification of more specific inhibitors of DNase I, or may be due to structural similarities of the inhibitory binding sites between the target DNA enzymes. Our latest studies have determined that apoptotic endonucleases are linked in a network, in which DNase I activates other endonucleases, such as EndoG, in particular. Therefore, this broad specificity may not preclude the use of these inhibitors for cell and tissue protection.

DNase I is a recently identified mechanistic target for protection against tissue injury. Inactivation of DNase I is protective against kidney injury and injuries to other organs.^11–13,17 Genetic inactivation of DNase I protects the kidney from cisplatin, rhabdomyolysis, and ischemia–reperfusion (I/R)-induced toxicity.^11,12,26,27 DNase I gene knockout also protects the liver from acetaminophen toxicity¹³ and ameliorates injury from total body gamma irradiation to radio-sensitive organs (e.g., intestine, spleen, and bone marrow) in mice.¹⁷

Small-molecule inhibitors of DNase I judged suitable for in vivo use (i.e., nontoxic and water soluble) are not currently available. An endogenous specific inhibitor of DNase I is G-actin (globular β-actin), which inactivates DNase I by sequestration.³ G-actin is, however, ineffective as a therapeutic agent because it is a protein and thus cannot be easily delivered into cells. Among small-molecular-weight compounds, aurintricarboxylic acid (ATA) has been identified as a nonspecific inhibitor of all endonucleases and promotes survival of cultured cells. ATA, however, has 2 major properties that limit its use in vivo: its complete absence of specificity that results in broad inhibition of many RNA and DNA enzymes, and its toxicity in vivo at doses required to inhibit endonucleases. Two other compounds have been described as inhibitors of DNase I,^28,29 but their inhibitory concentrations (10 mM) are rather high, and in vivo utilities have not been demonstrated.

Another reported DNase I inhibitor, Zn-DIPS, has been shown to provide radioprotection in mice comparable to that obtained from gene inactivation of DNase I.¹⁷ The specificity of Zn-DIPS for DNase I is questionable, however. Most likely, it acts as a delivery system for the zinc ion, which is a common inhibitor of many endonucleases, as well as other enzymes. Interestingly, Zn-DIPS provided bioprotection of radiosensitive organs in mice against gamma radiation when administered before or after the insult.¹⁷ The latter finding was expected because peak endonuclease activity after irradiation or other reactive oxygen species (ROS)-mediated injuries usually follows the insult by 14–18 h. These findings provide proof-of-principle that inhibition of DNase I is bioprotective if used before and/or within a few hours after injury. A short-term inhibition of DNase I is expected to produce no damage to organs because even permanent inactivation of DNase I in CD-1 mice by gene knockout is harmless.¹¹

A comparison between the current HTS assay and other reported assays best suited for measuring DNase I activity (Suppl. Table 3) shows that our HTS assay is the fastest and the most adaptable to a 96-well plate format. The precision and productivity of our HTS assay exceed those of all other available assays. It is very sensitive and allows measurement of DNase I activity in a 96-well plate format within 25 min. The assay can be further improved by automation and by the use of smaller multiwell plates. The assay followed by hit verification and in vitro testing represents an entirely new, innovative approach that will promote the development of specific and effective DNase inhibitors for the therapy of tissue injury. These inhibitors may have broad applicability for the treatment of kidney cell injury, because DNA breaks are the key event precipitating renal cell death resulting from diverse pathogenic insults. For the same reason, these inhibitors will likely be a potential therapy to treat or prevent injuries of other organs.

In summary, we have developed a HTS DNase I assay that has been used in the search for new DNase I inhibitors. The methodology described in this study constitutes the first HTS assay to detect endonuclease activity by chemical library screening with the added possibility of full automation. Two inhibitors have been discovered that are more potent than all previously known DNase I inhibitors, and they can potentially be used for tissue protection. Similar assays for other endonucleases that can be used to search for inhibitors and activators of these enzymes are currently being investigated.