Hypothermia as an Adjunct Therapy to Vesicant-induced Skin Injury

Cell culture

Seed cultures of J-774, CHO-K1, and HeLa cell lines were obtained from the American Type Culture Collection (Manassas, Va). J-774 cells are a mouse macrophage–derived cell line, CHO-K1 cells are a strain of the epithelial cell line derived from Chinese hamster ovary cells, and HeLa cells were originally cultured from a human cervical carcinoma. The cultures were grown in 10% foetal calf serum in Dulbecco’s Modified Eagle Medium supplemented with streptomycin (100 µg/mL) and penicillin (100 IU/mL), and the medium was changed as required. Stock cultures were closely monitored and not allowed to grow to confluency before being subcultured. Test cultures were used just before, or at, confluency. The methods for culturing human skin keratinocytes and chick embryo neurons have previously been described.^32–35

Chemical treatment

On the day of chemical treatment, the cultures were treated with the CW agent dissolved in ethanol so that the desired final agent concentration was reached at 0.25% ethanol (vol/vol). To assess cytotoxicity, alamarBlue was added (10%, vol/vol) to the cultures and allowed to incubate for the last 3 to 5 hours of the routine 24 hour (cell lines, neurons) or 48-hour (keratinocytes) treatment. The absorbances (570–600 nm) were then read on a Thermomax titerplate reader (Molecular Devices, Sunnyvale, Calif). This assay measures metabolic viability and we have found it to be a reliable indicator of cytotoxicity, yielding results similar to that obtained by more commonly used methods such as the neutral red and MTT assays.³⁴ Median lethal concentration values were determined graphically from experiments, utilizing 6 wells per data point. In all cases, values represent data from at least 3 separate experiments. Studies investigating the effects of temperature on HD or L toxicity in neurons and keratinocytes were carried out in humidified CO₂ incubators set at 25°C, 31°C, or 37°C. Munitions-grade sulfur mustard was distilled at Defence Research and Development Canada–Suffield to greater than 98% purity. Munitions-grade lewisite was distilled and the resultant product consisted of 79.9% dichloro (2-chlorovinyl) arsine (the desired lewisite product), 0.9% bis(2-chlorovinyl) chloroarsine, and 17.9% unknown.³³

Animal studies

Male hairless guinea pigs (strain Crl: IAF(HA)-hrBR) were acquired from Charles River Laboratories (St Constant, Quebec, Canada). The animals were acclimated for at least 1 week before experimental use. In conducting this research, the authors adhered to the Guide to the Care and Use of Experimental Animals and The Ethics of Animal Experimentation published by the Canadian Council on Animal Care. The methodology for HD and L skin exposure, as well as cooling agent-exposed sites, has previously been described.^32–35

Historical vesicant hypothermia research

Some of the first work to investigate the effect of lowered temperature on sulfur mustard toxicity appears to have been motivated by a then-popular theory of mustard toxicity. The “hydrochloric acid theory of mustard vesication” (reviewed in Papirmeister et al² hypothesized that the chemical hydrolysis of sulfur mustard within the cell resulted in sufficient liberation of hydrochloric acid so as to cause cytotoxicity. In a review article by Lynch et al,⁸ the authors reasoned that since the rate of sulfur mustard hydrolysis would be much decreased by reducing the temperature, this should provide relief from its toxicity. In these studies, catfish were exposed to a “quarter-saturated” solution of sulfur mustard in tap water for 5 minutes, or goldfish to a “half-saturated” solution of sulfur mustard in tap water for 10 minutes, and then transferred to either room temperature water (24°C–26°C) or cold water (8°C–10°C). Of the 10 catfish that were placed in room temperature water, all died within 28 hours, exhibiting hemorrhages on the fins, tails, and ventral surfaces within 20 hours. Five of 7 catfish placed back into the cold water survived for the 5-day experimental treatment with no signs of skin injury being observed. In the goldfish experiments, all 8 fish returned to room temperature water died within 6 days, while all 7 fish returned to the cooler water survived the 17-day test period. No skin injury was observed in the goldfish experiments. The authors concluded, after conducting further studies demonstrating that the toxicity of hydrochloric acid was not temperature dependent in these 2 species, that “these experiments on fish tend to substantiate the theory of intracellular liberation of acid.”⁸ Although the protective effects of hypothermia against sulfur mustard were demonstrated in these studies, it appears that these experiments largely summarized the work carried out with hypothermia and mustard sulfur to this point in time, since this report was reproduced in its entirety in a history of American medical countermeasures research against gas warfare published in 1926,³⁶ with no mention of any additional related work.

Between the World Wars, a great deal of CW research was carried out by a number of countries, since it was anticipated that the next major war would rely heavily on chemical weaponry. However, to our knowledge, there is no record of work carried out on the effect of lowered temperature on sulfur mustard injury. During World War II, while there was a wealth of research that supported the assertion that sulfur mustard–induced skin injury was definitively more severe in hot and humid conditions (either occluded by clothes or in tropical climates),^2,37 relatively little work appears to have been carried out on the effects of hypothermia. The focus of much of the work that did occur, appeared to use lowered temperature to decrease the amount of “fixed” (or nonextractable) sulfur mustard in the skin, an endpoint that was believed to be correlated with the severity of skin injury.³⁷ The original reporting of this work has not been possible for the authors to obtain. However, intriguing hints exist that point to the efficacy (or lack thereof) of cooling sulfur mustard–exposed tissue. Sir Frederick Banting, codiscoverer of insulin and a major supporter of Canada’s chemical biological effort during the World War II, was said to have tested the effects of ice-pack treatment on himself in 1940. He was described as undergoing a 5-minute exposure of 100 mg of sulfur mustard spread over a 6.5 × 1.5-in area on his lower right leg, before blotting the residual liquid off and applying ice. No toxicity was apparent until he had to fight a grass fire the next afternoon and the ice melted, resulting in extremely serious and painful injury. Despite this, the Canadian Medical Research Council apparently considered the ice treatment to be a success.³⁸

Figure 1. Reproduction of a page from a 1943 Chemical Warfare Laboratories (Ottawa, Ontario, Canada) physiological note referring to successful therapeutic ice-pack treatments to sulfur mustard–skin injury. Sulfur mustard is referred to as “H” while “S” refers to nitrogen mustard-2.

Additional intriguing evidence that ice-pack treatment of sulfur mustard–exposed skin might be of utility is a record of a meeting held by the Physiological Mechanisms Committee in New York City.³⁹ On the last page of this document (Fig 1), there is a short discussion on the efficacy of ice-pack treatment on sulfur mustard–exposed skin with respect to lowering the amount of nonextractable sulfur mustard, and preventing skin injury. Three years later, a review of what appears to be related work is included in a treatise published, describing the Allied CW agent research effort.⁴⁰ The chapter on mechanisms of sulphur mustard-induced skin injury³⁷ details considerable effort demonstrating the relationship between the amount of nonextractable sulfur mustard in the skin and resultant injury. It documented that in human skin, in contrast to pig and rabbit skin, sulfur mustard is almost maximally fixed within 2 minutes of exposure. The ice-pack studies subsequently described were viewed as confirmatory evidence that there was virtually no significant reservoir of unreacted sulfur mustard in human skin very shortly after exposure. In this work, humans and pigs were exposed to liquid mustard for 10 minutes, decontaminated with petroleum ether, and then ice-packs were applied to the exposure sites. The results of these experiments were described as “In man (Table 8), the ice-pack treatment did not significantly affect either the amount of fixed H as determined 24 hours after exposure or the severity of the lesion that developed. In the pig (Table 9), on the other hand, the ice-pack treatment resulted in the fixation of considerably less H than would otherwise have been the case, and in the partial or complete inhibition of visible injury development.”³⁷ Thus, it was concluded that in the pig, cooling exposed skin decreased the rate of sulfur mustard fixation, allowing it to be slowly carried away by body fluids during the cooling treatment, whereas in man, cooling was ineffective because virtually all sulfur mustard fixation occurred during exposure. Interestingly, although Renshaw chose not to include cooling as a possible treatment modality in his summary to this chapter, he referred to “Canadian experiments” as demonstrating that if thorough decontamination of sulfur mustard–exposed skin is not carried out then ice-pack treatment does reduce the severity of the skin injury. To this point, we have not been able to locate the original documentation of these studies.

Effect of hypothermia on vesicant-induced toxicity

Our interest in the potential protective effects of cooling HD-exposed tissue did not arise from historical considerations or from the hypothesis that hypothermia would prevent or ameliorate the initial lesions that HD causes to eventually cause toxicity. Rather, our motivation for examining temperature as a potential modulator of mustard toxicity was a result of work carried out with a class of drugs that conferred significant protection against HD toxicity in vitro. These compounds were nitric oxide synthase inhibitors with structural similarities to D/L-nitroarginine methyl ester D/L-NAME). Although they were shown to confer up to 300% protection against HD (in terms of median lethal concentration), this protection was not mediated through nitric oxide synthase inhibition.^41–43 While these findings were in themselves interesting, what was also intriguing was that the efficacy of L-NAME in cultures treated after HD exposure, remained maximal at 3 hours post–HD treatment, and then only slowly declined over the next 5 hours. These findings were consistent with HD rapidly initiating its effects and setting into motion metabolic events that resulted in a lesion (inhibited by L-NAME) being first expressed by approximately 3 hours post–HD treatment and fully expressed by about 8 hours post–HD treatment. This hypothesis could be easily tested by altering the incubation temperatures of HD-treated cultures and examining the effect of temperature on the efficacy of L-NAME posttreatment on HD toxicity, as a function of time. These studies were subsequently carried out. Similar studies were also carried out with L and the chelator dimercaptosuccinic acid (DMSA). Although both HD and L are vesicant in nature, they are very different in how they exert their respective toxicities. A comparison of their toxic effects was thus carried out both in vitro and in vivo.

The toxicity of the 2 agents was initially compared in cell cultures of widely varying origins (Table 1). The toxicity of L was at least 2 orders of magnitude greater than that of HD in differentiated chick embryo neurons, proliferating cultures of first passage human skin keratinocytes, CHO-K1, and HeLa cells, and 70 times more toxic in J-774 cells. The toxicities of HD and L were also assessed in studies using hairless guinea pigs (Fig 2), an animal model that has been shown to develop microblistering of the epidermis when the skin is exposed to HD vapor for 2 to 8 minutes.⁴⁴ In our laboratory, these 2 agents produced skin injury in this animal model in different fashions.^32,33 Four- and six-minute HD exposures slowly produced visible lesions that were maximal by 72 hours. In contrast, 2- to 6-minute L vapor exposures produced lesions of much greater severity. Lewisite-induced skin injury progressed very rapidly so that blood scabs were well formed at all exposure sites by 24 hours and of maximal severity by approximately 48 hours. Although the vapor pressure of L is several times higher than that of HD, it is significantly more cytotoxic and produces much more severe skin injury than HD does under similar exposure conditions.

Effect of hypothermia on sulfur mustard–induced toxicity

The effect of temperature on the toxicity of HD has been well characterized in first passage human skin keratinocyte cultures of varying differentiative states.³⁵ The toxicity of HD gradually declined as the cells were tested in proliferating, just-confluent, or post–confluent stages of culture. Nevertheless, in all cases, HD toxicity was significantly reduced by holding the cultures at 25°C, as opposed to 37°C. The effects of temperature were further examined in proliferating cultures treated with HD³² and held at 25°C, 31°C, or 37°C. At 37°C, HD toxicity was rapidly expressed and maximal by 48 hours postexposure. When the cells were incubated at 31°C, the 24-hour toxicity was significantly less than that exhibited at 37°C, but the cells then progressively lost viability over the ensuing 4 days. At 25°C, HD toxicity on day 1 was similar to that at 31°C, but increased only marginally with time so that by day 4, toxicity was still significantly less than that exhibited in cultures incubated at 31°C or 37°C (Fig 3A). Extensive confirmatory vidence of the protective effect of lowered temperature against HD toxicity was also obtained when apoptotic endpoints (morphology, DNA fragmentation) were utilized.³² The protective effects of lowered temperature were not persistent. When cultures incubated at the lowered temperature for 4 days were switched back to 37°C, the toxicity after 24 hours was similar to those of cultures incubated at 37°C for the entire 5-day time period (Fig 3A).

Figure 2. Effect of vesicant vapor exposure on hairless guinea pig skin. Vapor exposures of either distilled sulfur mustard (HD) or distilled lewisite (L) were carried out on hairless guinea pig skin for 2, 4, or 6 minutes. Results show the development of skin lesions at 7, 24, and 72 hours. Results depict the appearance of lesions on 3 animals and are representative of several different experiments.

Although the transient protective effects of cooling HD-exposed cultures were not promising with respect to obtaining positive outcomes in animal work, cooling of HD-exposed hairless guinea pig skin at 10°C for 6 hours (Fig 4) or at 5°C for 4 hours was protective in comparison with animals whose exposure sites were left at room temperature whereas similar cooling regimens at 15°C were not. Decreased, but still significant protection was also obtained if cooling (5°C) was delayed by 1 hour before 5 hours of cooling.³²

Similar, but less extensive work examining the effect of cooling HD-exposed skin was also carried out using 20-kg domestic swine.⁴⁵ In these studies, longer HD vapor exposure times (4–16 minutes) of the abdominal area were necessary to produce skin injury. The 4-hour ice-pack treatments used to obtain maximal protective effects in this animal model only required the skin to be cooled to approximately 15°C to 20°C. Cooling of HD-exposed skin consistently resulted in less injury than exposure sites left at room temperature. However, the severity of the lesions produced by HD exposure, as well as the protection afforded by cooling, appeared to be significantly more variable in swine than that observed in hairless guinea pigs. As the structure of swine skin is significantly closer to that of human skin than is guinea pig skin, this may suggest a similar variable response in humans.

Figure 3.Effect of incubation temperature on the development of distilled sulfur mustard (HD) and distilled lewisite (L) toxicity in human skin keratinocytes. Proliferating cultures of first passage neonatal human skin keratinocytes were treated with either HD (A) or L (B), and then incubated at 25°C, 31°C, or 37°C. At 1, 2, 3, or 4 days after treatment, unique test cultures were assayed for viability using alamarBlue. The HD-treated cultures assayed at 4 days, were refed, and then incubated for an additional 24 hours at 37°C before viability assessment. Data represent the mean ± SEM of three different experiments using tissue from three different donors, and was analyzed by analysis of variance and post hoc Dunnett’s Method Comparison testing. *Significantly different from the median lethal concentrationobtained at 24 hours at 37°C, P < .05.

As stated previously, our original goal in examining the effect of temperature on vesicant-induced toxicity was to ascertain whether cooling HD-treated cell cultures would result in an increase in the therapeutic window in which L-NAME was protective. This drug is effective against HD in just-confluent cultures of human skin keratinocytes. However, the extremely rapid decline in HD toxicity with culture age, especially as the cultures became confluent and differentiation was initiated, precluded reproducible protection studies in this system. We, therefore, carried out this work in differentiated primary chick embryo neuron cultures, a system that we have found to be very sensitive to the effects of HD, sensitive to the protective effects of L-NAME, and to be predictive of results obtained in human skin keratinocytes.³⁵ Lowering the incubation temperature of HD-exposed neurons significantly lengthened the period of time in which L-NAME was still of therapeutic value (Fig 5A). At 37°C, 5.0 mM L-NAME was maximally effective when added 2 hours after the neurons had been exposed to HD. This protection then gradually decreased until no significant advantage was obtained 12 hours post–HD exposure. At 25°C, maximal protection could still be observed when adding the L-NAME up to 12 hours after HD exposure. No efforts were made to duplicate these results in the hairless guinea pig model, since L-NAME was not found to be protective in this model system, probably due to difficulties in obtaining and maintaining sufficiently high concentrations at the HD target sites.³⁵

Effect of hypothermia on lewisite-induced toxicity

In contrast to the results obtained with HD, keratinocyte cultures exposed to L and incubated at 25°C rapidly lost viability with time (Fig 3B). At 37°C, L toxicity was almost maximal by 24 hours and decreased only slightly during the next 3 days. At 24 hours, in cultures incubated at 31°C, the toxicity of L was reduced (but not statistically significant) in comparison with that obtained at 37°C. However, by 48 hours, this protective effect was lost and median lethal concentration values were similar to those obtained at 37°C. The protective effect of lowered temperature was dramatic when the treated cultures were incubated at 25°C. The median lethal concentration values obtained at 24 hours were more than 12 times higher than those at 37°C. However, these protective effects were rapidly lost over the following days. Clearly, the development of HD toxicity in this culture system is more temperature dependent than that of L.

Figure 5.Effect of incubation temperature on the efficacy of protective drug treatments in cell cultures exposed to sulfur mustard (A) or lewisite (B). Mature chick embryo neuron cultures were treated with 5.0 mM L-nitroarginine methyl ester (L-NAME) at varying time intervals after HD exposure of cultures incubated at 25°C or 37°C (A) and proliferating cultures of first passage cultures of human skin keratinocytes were treated with 1.0 mM demercaptosuccinic acid (DMSA) at varying time intervals after L exposure of cultures incubated at 25°C, 31°C, or 37°C (B). Test cultures were assayed for viability using alamarBlue™. Results represent the mean ±SEM of three separate experiments using tissue from different donors, and were analyzed by analysis of variance and post hoc Tukey HSD Multiple Comparison testing. ⋆ Significantly different from the median lethal concentration of agent-only–treated cultures within the same test temperature, P < .05.

Figure 6.Effect of incubation temperature on the efficacy of protective demercaptosuccinic acid (DMSA) treatment after exposure of distilled lewisite (L) on hairless guinea pig skin. Animals were exposed to L vapor for 2, 4, or 6 minutes. At 2 minutes posttreatment, the exposure sites were treated topically with 50 µL of vehicle (DMSO) only; (A) or DMSA (50 mg/mL, B) for 1 hour. In additional treatment groups, the L exposure sites were left for 2 hours at room temperature before being treated with DMSA (50 µL of 50 mg/mL, C) for 1 hour, or the exposure sites were cooled at approximately 10°C for 2 hour before the 1-hour treatment with DMSA (D). Three animals were used for each treatment and the photographs depict the appearance of the lesions at 3 days posttreatment.

Lewisite produces much more toxicity in a shorter time frame, both in vitro and in vivo than does HD. Thus, the results of the cooling experiments utilizing hairless guinea pigs exposed to L were unexpected.³³ Not only were the lesions caused by 2 to 6 minutes L vapor exposures very significantly reduced by cooling at 10°C for as little as 30 minutes, but protection could still be obtained by cooling up to 1 hour post–L exposure. Perhaps, most surprising was that the very serious skin injury produced by L vapor exposure could almost be totally eliminated by cooling (Fig 4). Protection studies using DMSA and L were carried out in human skin keratinocytes. In contrast to the protective effects of l-NAME against HD toxicity, which are only exhibited in differentiating cultures, DMSA is also protective against L toxicity in proliferating cultures. Cultures were held at 25°C, 31°C, or 37°C exposed to L, and then treated with 1.0 mM DMSA at 1, 3, 5, or 8 hours postexposure. After DMSA treatment, all cultures were returned to a 37°C incubator for the remainder of the 48 hour L exposure period. When the cultures were DMSA-treated at 1 hour post–L exposure, significant protection against L was obtained only at 25°C, although the viability at the other 2 temperatures was elevated. After this, the protective effects of DMSA rapidly declined with time at all temperatures (Fig 5B). The relative success of this in vitro work led us to examine the possibility that hypothermia may also result in an extended DMSA therapeutic window against L in hairless guinea pigs.³³ In comparison with vehicle [dimethylsulphoxide (DMSO), Fig 6A] treatment only, a 1-hour DMSA treatment immediately after L vapor exposure resulted in very significant protection, with only minor skin injury evident at the 6-minute exposure sites (Fig 6B). Delaying the DMSA treatment by 2 hours resulted in very good protection at the 2-minute and 4-minute L exposure sites, but not at the 6-minute L exposure sites (Fig 6C). The severity of skin injury from this latter treatment was similar to that obtained when the exposure sites were cooled for 2 hours.³³ Combining these 2 treatments, so that DMSA was applied after 2 hours of cooling, totally eliminated all signs of skin injury (Fig 6D). Even close examination of the exposure sites failed to detect any difference from the surrounding tissue.