|Home | About | Journals | Submit | Contact Us | Français|
Sulfur mustard [bis(2-chloroethyl)sulfide, SM] is a well-known DNA-damaging agent that has been used in chemical warfare since World War I, and is a weapon that could potentially be used in a terrorist attack on a civilian population. Dermal exposure to high concentrations of SM produces severe, long-lasting burns. Topical exposure to high concentrations of 2-(chloroethyl) ethyl sulfide (CEES), a monofunctional analog of SM, also produces severe skin lesions in mice. Utilizing a genetically engineered mouse strain, Big Blue, that allows measurement of mutation frequencies in mouse tissues, we now show that topical treatment with much lower concentrations of CEES induces significant dose- and time-dependent increases in mutation frequency in mouse skin; the mutagenic exposures produce minimal toxicity as determined by standard histopathology and immunohistochemical analysis for cytokeratin 6 and the DNA-damage induced phosphorylation of histone H2AX (γ-H2AX). We attempted to develop a therapeutic that would inhibit the CEES-induced increase in mutation frequency in the skin. We observe that multi-dose, topical treatment with 2,6-dithiopurine (DTP), a known chemical scavenger of CEES, beginning 1 hour post-exposure to CEES, completely abolishes the CEES-induced increase in mutation frequency. These findings suggest the possibility that DTP, previously shown to be non-toxic in mice, may be useful as a therapeutic agent in accidental or malicious human exposures to SM.
Sulfur mustard [bis(2-chloroethyl)sulfide, SM] has been used as a chemical warfare agent since WWI, and is known to be toxic to the skin, eyes and respiratory tract. In particular, dermal exposure leads to blistering and ulcerative lesions that resist treatment, and heal slowly and with extensive scarring (Papirmeister et al., 1991). In several rodent models, topical SM treatment induces full-thickness burns that heal slowly (Vogt et al., 1984; Casillas et al., 1997; Smith et al., 1997). Similar lesions are also induced in mice by treatment with the less potent half-mustard analog 2-(chloroethyl) ethyl sulfide (CEES) (Isidore et al., 2007). In both humans and mice, lesions develop slowly and are not apparent for at least 12 hrs post-treatment.
Although both SM (Ogston et al., 1946) and CEES (Liu et al., 2010) hydrolyze readily in aqueous solution, several lines of evidence point to a more prolonged stability in biological systems. In an experimental model, macromolecular damage in response to mustard (globin adduct) has been shown to increase over a period of many days (Noort et al., 2002). In an accidental SM exposure incident, excretion of hydrolysis products of SM was significantly above background 7 days post exposure (Barr et al., 2008). The persistence of SM in skin tissue has been documented in a combat-related human exposure (Draschet al., 1987), and in experiments with cadaver skin (Hattersley et al., 2008), leading to the concept that mustards survive in treated skin for extended periods of time in a “reservoir” where they are protected from aqueous hydrolysis and detoxification. If this long-lived, tissue reservoir of active mustard contributes to the toxic effects in skin, then it is conceivable that post-exposure application of treatments that detoxify this reservoir represents a window of therapeutic opportunity.
It is thought that direct damage to tissue macromolecules underlies the toxic effects of mustards. Indeed, both protein and nucleic acid adducts of mustards have been described. In particular, mustard-DNA adducts are highly mutagenic, and are likely responsible for the association of chronic mustard exposure with development of respiratory tract cancers seen in several epidemiological studies (Wada et al., 1968; Easton et al., 1988; Hu et al., 2002). Low millimolar concentrations of CEES are mutagenic in human skin cell cultures (Powell et al., 2010).
We have previously shown that the non-toxic thiopurine 2,6-dithiopurine (DTP) is biologically active in mouse skin as a scavenger for an electrophilic carcinogen, 7r,8t-dihydroxy-9,10t-oxy-7,8,9,10-tetrahydrobenzo[a]pyrene, completely abolishing both DNA adduct formation and the initiation of carcinogenesis (MacLeod et al., 1991). DTP reacts facilely with a wide variety of electrophilic compounds at neutral pH in vitro (MacLeod et al., 1993; Qing et al., 1996), including CEES and 2-(chloroethyl) methyl sulfide (Liu et al., 2010). Competition experiments demonstrated that DTP was almost an order of magnitude more active in detoxifying CEES than common anti-oxidants such as glutathione and N-acetyl-cysteine (Liu et al., 2010). Correlated with this activity, we found that DTP could effectively block both cytotoxicity and mutagenicity of CEES in human skin cell cultures (Powell et al., 2010). These findings suggest the possibility that the scavenging activity of DTP may be effective in blocking the mutagenic effects of mustards in vivo. We now present evidence in a mouse skin model that moderate doses of CEES induce mutations in vivo, and that therapeutic treatment with DTP following CEES-exposure completely abolishes this mutagenic activity.
Big Blue mice were obtained from Stratagene (La Jolla, CA), and a breeding colony, homozygous for the recombinant chromosome, was maintained in a temperature and humidity controlled, AAALAC-approved facility with a 12 hr light-dark cycle. C57BL/6 mice were obtained from the NIH. All experiments were carried out under protocols approved by the IACUC of The University of Texas MD Anderson Cancer Center. Initial experiments indicated that female and male animals responded similarly to topical treatment with CEES, and all further experiments were conducted with females, 8-10 weeks of age. Two-three days prior to treatment, mice were shaved dorsally with clippers. On the day of treatment, mice were anesthetized with ketamine. All treatments were performed in a class IIB biological safety cabinet, and the treated mice were held in disposable cages in the safety cabinet for the duration of the experiment. During treatment and for the first 24 h post-treatment laboratory personnel were further protected by the use of powered air-purifying respirators.
CEES (Aldrich Chemicals) was diluted just before use in anhydrous ethanol to concentrations in the range of 100-400 mM. Anesthetized mice were treated with 20 μL of CEES stock, applied by micropipet to the shaved dorsum approximately 3.5 cm rostral to the base of the tail on the midline. The applied dose spread across the skin to form irregular regions that could be approximated as a circle with a diameter of ~1.5 cm. The mice were held for approximately 25 sec, allowing the ethanol to be absorbed or evaporated, and then returned to disposable holding cages as above. For routine treatments, 200 mM CEES was used, resulting in an applied dose of 4.0 μmol, at a dose density of ~2.3 μmol/cm2. Control mice received ethanol only.
The genome of Big Blue mice (on a C57BL/6 background) contains ~100 copies of a modified lambda phage genome that includes a mutagenesis target, the lacI gene, that allows the scoring of mutations induced by mutagenic treatments in vivo (Kohler et al., 1991). This is done by isolating genomic DNA from the appropriate target organ of the treated mice, in this case dorsal skin, excising the lambda genomes, packaging the DNA in vitro into lambda phage, and enumerating white (wild type) and blue (mutant) plaques after infection of a lawn of reporter bacteria in the presence of the beta-galactosidase substrate X-gal.
To determine mutation frequencies in the treated skin, mice treated as described above were sacrificed by CO2 asphyxiation, the dorsal skin was dissected and trimmed to a small piece containing the treated area, and frozen in liquid nitrogen. Genomic DNA was purified from the frozen skin by an overnight digestion with Proteinase K and RNase A in the presence of 1% SDS, followed by phenol:chloroform extraction and ethanol precipitation. Phage genomic sequences were excised from purified genomic DNA and packaged into lambda phage utilizing Transpack Kits obtained from Stratagene (La Jolla, CA). The lacI mutant screen was performed as suggested by the manufacturer with minimal modifications. Indicator E. coli cultures were transduced with packaged phage from each skin sample, and plated to determine the transduction efficiency. For each skin tissue sample at least 18,000 pfu (generally >70,000 pfu) were plated at an average of 300-400 pfu/15 cm Petri dish and incubated overnight at 37°C. The plates containing the transduced E.coli were left at room temperature for a period of two days to allow detection of all mutants. Mutation frequencies were calculated as the number of total blue plaques divided by the overall pfu counted per skin sample.
All immunohistochemistries were performed by the Histology and Tissue Processing Facility Core of the Center for Research on Environmental Disease. Detection of cytokeratin 6 was performed as previously described (Wang et al., 2007), using an antibody from Covance (Berkeley, CA) that does not cross-react with other epidermal cytokeratins. Anti-phospho γH2AX antibody was obtained from Cell Signaling (Danvers, MA; #2577).
DTP was synthesized as described previously (Ozola and Mikstais, 1978) or repurified from a commercially available source as described (Liu et al., 2010). Purity was at least 97% as determined by HPLC (Liu et al., 2010). For topical application, a cream commonly used for formulation of pharmaceuticals for human use, Lipoderm (LD; PCCA USA, Houston, TX) was utilized. DTP was pre-wet with propylene glycol, and thoroughly mixed with LD in an ointment mill to obtain a final concentration of 20% DTP (wt/wt). Concentrations were verified by extracting DTP from a measured amount of the cream with 0.1 M K2HPO4, and determining the amount of DTP extracted spectrophotometrically.
To determine the approximate lifetime of applied DTP in the skin of mice, 0.1cc of 20% DTP was applied to the shaved backs of C57BL/6 mice, and gently rubbed in for about 30 sec. Control mice received 0.1 cc of LD only. After various times, mice were sacrificed and the dorsal skin dissected. A dermotome was used to excise 4 samples of skin (0.4 cm diameter circles) from each mouse. Each sample was digested overnight at 37° C in 134 μL of Nuclei Lysis Solution (Promega, Madison WI) containing 90 mM EDTA and 1.5 mg/mL proteinase K, then diluted with 1000 μL 0.1 M K2HPO4. After centrifugation at 12,000 rpm for 5 min, the clear supernate was carefully removed with a Pasteur pipet, and an aliquot diluted and assayed spectrophotometrically, utilizing the characteristic 348 nm absorption band of DTP (Qing et al., 1995).
In an experiment in which a Big Blue mouse is treated with a given concentration of a mutagen, the distribution of mutants between DNA molecules in the treated skin is expected to be random, and therefore, can be described by Poisson statistics. The probability that the mutation frequency observed at a particular dose is significantly different than the background frequency can then be calculated from the Poisson distribution.
In experiments where multiple mice were used in each experimental group, the significance of a difference between groups was evaluated by Student's t-test. In both cases, p<0.05 was taken as the cutoff for statistical significance.
When liquid CEES (~8.59 M) is applied without dilution to the dorsal skin of C57BL6 mice, severe lesions are obtained within 24 h, with severe necrosis of the epidermal cells (Isidore et al., 2007). By 3-4 days, no intact interfollicular epidermal cells can be observed, and the epidermis has been replaced by eschar tissue (data not shown). To test the effects of lower, potentially mutagenic doses, we diluted CEES to either 200 or 400 mM in EtOH and applied 20 μL to the shaved dorsum of female C57BL6 mice; control mice received ethanol only. Mice were sacrificed three days later, and the dorsal skin prepared for histological analysis. Representative microscopic images of H&E-stained sections of the dorsal skin 3 days post-treatment are shown in Figure 1. At 200 mM CEES (Figure 1B) histological analysis revealed little or no change from control (Figure 1A). In particular, we did not observe significant thickening of the epidermis, inflammation, or ulcerative lesions. In contrast, 400 mM CEES (Figure 1C) induced severe hyperplasia over the entire treated region, and focal lesions characterized by severe acute inflammation, the presence of thick eschar tissue and a breakdown of the integrity of the interfollicular epidermis (IFE). In some cases, frank ulcers were seen (Figure 1D).
The presence in nuclei of a specific, phosphorylated form of histone H2A, called γ-H2AX, is often used as a biomarker of DNA damage (Dickey et al., 2009). The formation of γ-H2AX 1-3 days after treatment of mouse skin with SM has been reported (Joseph et al., 2011). We therefore also examined skin sections for γ-H2AX by immunohistochemistry (IHC). In vehicle control-treated skin, strongly labeled, γ-H2AX positive nuclei were extremely rare in the epidermis (<< 1% of the IFE cell nuclei). After treatment with 200 mM CEES the presence of nuclei stained positively for γ-H2AX in the IFE increased (Figure 2A), although the incidence was still only ~1-2% of the interfollicular nuclei. In the skin treated at 400 mM CEES, positive staining for γ-H2AX was widespread, with many positively stained nuclei seen in both the basal and suprabasal layers of the IFE (Figure 2B). At higher magnification, staining for γ -H2AX was clearly punctate as expected for DNA damage repair foci (Figure 2C; (Sedelnikova et al., 2003)). Positive nuclei were also seen in the hair follicles and in fibroblast nuclei in the dermis and hypodermis (Figure 2D; arrows and arrowheads, respectively). Cytokeratin 6 (CK6) expression in IFE is often associated with hyperplasia and epidermal irritation (Kopan and Fuchs, 1989; Rothnagel et al., 1999), whereas hair follicles normally express CK6. When examined for CK6 expression by IHC, most of the interfollicular epidermis in mice treated with 200 mM CEES was negative; normal staining of hair follicles was seen throughout. Occasional groups of keratinocytes exhibited cytoplasmic staining for CK6 (Figure 2E). No CK6 expression was observed in the interfollicular epidermis of control ethanol-treated mice (not shown). At 400 mM CEES, cytoplasmic expression of CK6 was ubiquitous in the treated, hyperplastic IFE (panel 2F). Thus, the extent of expression of these two markers was relatively modest in mice treated with 200 mM CEES, in agreement with the relative normality of the histopathology, and provided no evidence for severe DNA and tissue damage.
To optimize conditions for estimating induced mutation frequencies, we applied various concentrations of CEES to the shaved dorsal skin, and isolated genomic DNA from the treated skin 1 or 2 days following treatment. Preliminary measurements made after control treatment with ethanol only, led to an average background mutation frequency of 4 × 10-5 (range 2-6 × 10-5). This is similar to previous background measurements in Big Blue skin (de Boer et al., 1998). As shown in Figure 3, mutation frequencies measured 24 h after CEES-treatment were above background at doses of 300 and 400 mM CEES, but this increase was not statistically significant (p>0.05, Poisson distribution). In contrast, mutation frequencies measured 48 h after treatment were significantly higher than background at all doses greater than or equal to 200 mM CEES, ranging from about 11-13 × 10-5, an increase of about 3-fold above background (p<0.005, Poisson distribution). This was higher than the maximal frequencies measured at 24 h, suggesting that 24 h is not sufficient time for maximal mutation induction in keratinocytes. The conversion of CEES-induced DNA damage to mutations is thought to involve trans-lesion synthesis and is known to require DNA replication and cell division (Provost et al., 1993; Lange et al., 2011). Indeed, further experiments with a dose of 200 mM CEES showed a higher mutation frequency at 4 days following treatment (data not shown), and the rest of the study was carried out using 200 mM CEES, with genomic DNA isolation 4 days following treatment.
We wished to determine whether DTP, a known nucleophilic scavenger of CEES (Liu et al., 2010), could modify mutagenicity in this model. For initial experiments, we chose to deliver DTP by topical treatment of the CEES-exposed skin with a commercial cream used for trans-dermal delivery of pharmaceuticals in humans. The delivery vehicle used for these experiments, Lipoderm (LD), was chosen for its ability to penetrate the skin, as described by the manufacturer. Indeed, a single topically applied dose of DTP in LD rapidly entered the skin, and was cleared from the skin with a half-life of about 4 hr (Figure 4A). Since as noted above the persistence of active mustard in skin is expected to be considerably longer than 4 h, it seemed likely that multiple treatments with DTP would be needed to inhibit DNA adduct formation and mutagenesis.
Based on the preliminary pharmacokinetics, we simulated the expected levels of DTP in skin after multiple treatments. This allowed us to develop a schedule of DTP treatments that could be expected to maintain a high level of DTP in the skin. The first treatment was applied 1 hr post exposure to CEES to model the delay between exposure and treatment to be expected in a mass casualty scenario. Further treatments were given 2 hr later, then 4hr later, and then every 6 hr for a total of 13 treatments. Figure 4B presents a simulation of the relative tissue concentration of DTP for the first 9 treatments, assuming a constant half-life of 4 hrs throughout this regimen. The predicted tissue concentration of DTP remained fairly high for the duration of the treatment regime (Figure 4B); the average level of skin DTP over the 48 h simulation shown was about 98% of the maximal level seen following the initial dose.
To test the effects of DTP on CEES-induced mutagenesis in vivo, mice were treated with either EtOH or 200 mM CEES, and one hour later the 13-dose regimen was begun with either LD only or LD containing 20% DTP. Genomic DNA was isolated from the treated skin 4 days after CEES treatment, and used to determine the mutation frequency. Three independent experiments are summarized in Figure 5. The average background mutation frequency (EtOH/LD) was 4.7×10-5, similar to previous measurements without the LD treatment. Treatment with CEES followed by LD induced a mutation frequency of 13.0 × 10-5, approximately 2.8-fold higher than the control; this difference was statistically significant (t-test: p<10-5). When treatment with CEES was followed by multiple treatments with DTP, the mutation frequency was reduced to 4.4×10-5, not significantly different than the control (t-test: p>0.5). Thus, this DTP treatment regimen completely abolished the mutagenesis induced by CEES in the skin of Big Blue mice.
Induction of mutations by CEES was first measured in a bacterial system (Gilbertet al., 1975), and we have previously reported induced mutation rates in a human keratinocyte cell line (Powell et al., 2010). Mutagenicity of SM in rats has been inferred from dominant lethal analysis after exposure by inhalation (Rozmiarek et al., 1973) and gavage (Sasser et al., 1993). The results presented above clearly show that topical application of CEES induces a significant increase in mutation frequency in the skin, using a phage-based reporter system. This is the first quantitative measurement of the in vivo mutagenic action of a sulfur mustard in mammals, although such an effect was clearly expected based on measurements in vitro and in lower model organisms. Mutation frequency was close to maximal in these assays at a dose of CEES that was insufficient to induce major histopathological changes in the treated skin. Furthermore, the induced mutation frequency dropped to background levels when the CEES treatment was followed by a therapeutic regimen of 13 treatments with DTP. Since we have previously shown facile adduct formation between CEES and DTP, both in a test tube (Liu et al., 2010) and in cell culture (Powell et al., 2010), this suggests that this scavenging action may be the primary mechanism by which DTP protects against mutagenicity in this model. For this to be the case, much of the mutagenic action of CEES must have taken place after the one hour lag between CEES exposure and the first DTP treatment. As noted in the Introduction, it is thought that sulfur mustards remain stably in the skin for periods of time much longer than one hour, and are able to continue to produce macromolecular adducts several days following treatment (Noort et al., 2002).
The lack of overt pathological changes in mouse skin treated with 200 mM CEES was somewhat unexpected, given the severe toxicity shown previously with neat CEES in C57BL/6 mice (Isidore et al., 2007). We have found histopathological lesions similar to those described by Isidore et al. at an intermediate dose of 430 mM (unpublished data), including loss of all epidermis, extensive damage to the dermis and massive inflammation. However, Tewari-Singh and colleagues reported similar “full thickness” lesions in SKH-1/hairless mice after treatment with CEES at approximately 80 mM in acetone (Tewari-Singh et al., 2009). This apparent difference could be a strain dependent difference in sensitivity, or a solvent effect. Indeed, when we treated C57BL/6 mice with 80 mM CEES in acetone we also observed extensive epidermal damage (unpublished results). This suggests that acetone somehow enhances the skin toxicity of CEES compared to effects with ethanol as supporting solvent.
Exposure of mouse skin to SM in the vapor phase has recently been shown to induce γ-H2AX formation, visible 1-3 days following treatment (Joseph et al., 2011). This phosphorylation of H2AX is induced by DNA double strand breaks, mediated by the ATM/ATR/DNA-PK kinases, and plays a key role in the organization of multi-protein complexes in damaged chromatin that facilitate double strand break repair (Dickey et al., 2009). Although classically induced by agents that directly create double strand breaks such as ionizing radiation, alkylating agents are also known to induce γ-H2AX formation (Dickey et al., 2009). This is presumed to be due to the formation of double strand breaks during cellular attempts to repair the simple alkylation adducts in DNA. Thus, it is not surprising that the mono-functional alkylating agent CEES induces γ-H2AX foci (Figure 2, B-D). The presence of multiple labeled nuclei suggests that in many keratinocytes DNA repair is still not complete three days following exposure to 400 mM CEES.
Concerns have been expressed that SM could be utilized by terrorists to harm civilian populations (Saladi et al., 2006). In a terrorist attack on an urban population, it is likely that many people would be exposed to amounts of liquid sulfur mustard that produce obvious, albeit delayed skin toxicity. These patients would require immediate treatment, and extended hospitalization. However, it seems likely that a large number of people might receive lower doses that do not cause overt toxicity, and that this exposure could go undetected and untreated. Because CEES and SM react with nucleophiles by a common reaction mechanism, the present results suggest that some of these victims would receive mutagenic doses. It is well known that most mutagens are also carcinogens, and that chronic low-dose mustard exposure is associated with increased risk of cancer (Wada et al., 1968; Easton et al., 1988; Hu et al., 2002). In experimental animals, acute treatment with SM has been shown to be carcinogenic (Heston, 1953; Fox and Scott, 1980). Taken together, these results suggest the possibility that civilian victims receiving a single, low dose of SM would be at increased risk for cancer development later in life. Thus, a therapeutic intervention with DTP based on our mouse model would be likely to benefit that fraction of the victims receiving a sub-toxic dose.
We thank Drs. Wallace Baze and Donna Kusewitt for expert histopathological analysis, Dr. Howard Thames for help with statistical analysis, the staff of the Histology and Tissue Processing Facility Core for preparation of slides, and Rebecca Deen and Joi Holcomb for manuscript preparation. This work was supported by grant U01 NS058191 from NINDS (Dr. M. MacLeod) through the CounterACT Program, and by Center Grants from the NCI (P30 CA016672) and the NIEHS (P30 ES007784). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the federal government. The funding agencies had no role in study design, data collection or analysis, manuscript preparation or the decision to publish.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest Statement
The authors declare that they have no conflicts of interest.