Elucidating mechanisms of chlorine toxicity: reaction kinetics, thermodynamics, and physiological implications

Giuseppe L. Squadrito, Edward M. Postlethwait, Sadis Matalon


Industrial and transport accidents, accidental releases during recreational swimming pool water treatment, household accidents due to mixing bleach with acidic cleaners, and, in recent years, usage of chlorine during war and in acts of terror, all contribute to the general and elevated state of alert with regard to chlorine gas. We here describe chemical and physical properties of Cl2 that are relevant to its chemical reactivity with biological molecules, including water-soluble small-molecular-weight antioxidants, amino acid residues in proteins, and amino-phospholipids such as phosphatidylethanolamine and phosphatidylserine that are present in the lining fluid layers covering the airways and alveolar spaces. We further conduct a Cl2 penetration analysis to assess how far Cl2 can penetrate the surface of the lung before it reacts with water or biological substrate molecules. Our results strongly suggest that Cl2 will predominantly react directly with biological molecules in the lung epithelial lining fluid, such as low-molecular-weight antioxidants, and that the hydrolysis of Cl2 to HOCl (and HCl) can be important only when these biological molecules have been depleted by direct chemical reaction with Cl2. The results from this theoretical analysis are then used for the assessment of the potential benefits of adjuvant antioxidant therapy in the mitigation of lung injury due to inhalation of Cl2 and are compared with recent experimental results.

  • hypochlorite
  • lung epithelial lining fluid
  • antioxidants
  • exposure
  • therapy

the toxicity of inhaled chlorine (Cl2) and the treatment and mitigation of chlorine-induced lung injury have been of interest due to industrial and transport accidents (37, 84), accidental releases during recreational swimming pool water treatment (5, 15, 55), and household accidents due to mixing bleach with acidic cleaners (10, 17, 35, 46). In recent years, Cl2 has received renewed attention because of its use during war and in acts of terror (12). It is striking that so little attention has been given to study the mechanism of action of Cl2, and the genesis and development of the lung injury in terms of the chemical reactivity of Cl2, especially when one considers the morbidity and mortality that exposures to Cl2 gas have produced and can still produce in the future. At present, therapy for Cl2 inhalation injury consists in alleviating pulmonary symptoms (73). Only recently, a mechanism-oriented, chemically specific approach to prophylactic and postexposure therapy is being employed (45, 74) wherein antioxidant replenishment as well as various agents to restore compromised alveolar function (such as surfactant and ion transport) are being used to counteract Cl2 toxicity and to decrease morbidity. This review represents an attempt to elucidate the transient species generated during exposure to Cl2 and their modes of action. We believe such investigation will be conducive to new and improved mechanism-oriented therapeutic strategies.

We here describe chemical and physical properties of Cl2 that are relevant to the chemical reactivity of Cl2 with biological molecules present in the fluid layers covering the airways and alveolar spaces. This analysis is then used for the assessment of the potential benefits of adjuvant antioxidant therapy in the mitigation of lung injury due to inhalation of Cl2 and is compared against recent experimental results.

The biological response to a Cl2 exposure depends on the concentration of Cl2 and the duration of the exposure in a manner described by the generalized Haber's law (52). In addition, one must consider individual susceptibility and that the injury will progressively extend to more distal sites as Cl2 concentrations increase (8, 57). Chlorine more rapidly damages the wet mucosal tissues and thus it seems appropriate to first discuss the aqueous solubility of Cl2 and its reaction with water. It is interesting to note that histopathological analysis may not reveal the true extent of tissue injury compared with the results from physiological and biochemical analyses [e.g., bronchoalveolar lavage protein (45), surfactant function (45), antioxidant levels (45), and sodium-dependent transport (74)] that suggest underlying injury may be more serious than histopathological analysis alone would suggest.

The Solubility of Chlorine in Water

At ambient temperature, Cl2 is an oxidant gas that is moderately soluble in water, being approximately five times more soluble than are ozone (O3) or nitrogen dioxide (NO2), two common oxidant gases of much environmental interest (2, 14, 31, 70, 75). However, Cl2 rapidly reacts with water, contrasting sharply with O3 and NO2 in this regard. Because of the high reactivity of Cl2 toward water, reactive uptake of Cl2 by aqueous solutions is favored relative to other oxidant gases, such as O3 and NO2, that do not react with water under physiologically relevant conditions to any significant extent. Thus, it becomes important to distinguish between the physical solubility and the reactive absorption of Cl2.

The Reaction of Chlorine With Water

Chlorine reacts with water according to Eq. 1: Cl2+H2Ok1k1HOCl+H++ClKH2O=1.8×103M2 (1)

The reaction of Cl2 with water is not a benign reaction that merely scavenges Cl2 as it penetrates the aqueous milieu. Although Eq. 1 partially destroys Cl2, it results in the formation of hypochlorous acid (HOCl), yet another powerful oxidant, and the strong acid hydrochloric acid (HCl). HOCl is a weak acid with pKa = 7.54 at 25°C and pKa = 7.45 at 37°C (53); for the surface of the lung, where the pH = 7.0, ∼74% of the hypochlorite would exist as HOCl and 26% as OCl, if equilibrium was achieved.

The value for equilibrium constant KH2O for Eq. 1 is small and is somewhat deceptive because it may suggest the equilibrium will lie to the left of Eq. 1. However, when one considers that pH = 7.0 and [Cl] = 0.1 M for the surface of the lung, and that these values will remain relatively unchanged as a result of Cl2 exposures, one determines: [HOCl]/[Cl2] = KH2O/[H+][Cl] = 1.8 × 10−3 M2/(1.0 × 10−7 M)(0.1 M) = 1.8 × 105, indicating that >99.999% of the Cl2 that is absorbed within the respiratory tract will be converted to HOCl and OCl, at equilibrium, and if Cl2 were to react only with water.

The reaction kinetics for the hydrolysis of Cl2 is relatively fast and was studied using stopped flow spectrophotometry (87). Thus, using data published by Wang and Margerum (87), and correcting for temperature effects, one can compute k1 and k−1 for Eq. 1 as 61.7 s−1 and 34.6 × 10−3 M−2s−1, respectively, at 37°C. With these values, and assuming the pH and [Cl] will remain unchanged at pH = 7.0 and [Cl] = 0.1 M, respectively, a computer simulation using the software package Gepasi v. 3.30 (50) of the reaction kinetics associated with Eq. 1 and the acid dissociation of HOCl indicates that equilibria will be established in ∼11 ms. However, as discussed shortly, lung epithelial lining fluid (ELF; airway and alveolar) contains high concentrations of reactive biological molecules, in particular low-molecular-weight antioxidants, which are potential targets for reaction with Cl2 and may compete with the hydrolysis of Cl2 by reacting with Cl2 before it can undergo hydrolysis.

The toxicity of Cl2 is generally attributed to hypochlorite formed as a result of Cl2 hydrolysis according to Eq. 1 (1, 19, 26, 33, 56, 58, 89). (In this article, the term hypochlorite is used to refer to the sum of HOCl and its conjugate base OCl present in prototropic equilibrium, i.e., the stoichiometric concentration of hypochlorite. When a discussion is specifically limited to one of these species, its formula will be given.) The concomitant formation of the strong acid hydrochloric acid (HCl) is not considered important to the mechanism of Cl2 toxicity because of the large ELF buffering capacity that can, in most cases, neutralize the acid insult. The ELF has an appreciable bicarbonate concentration (11 mM) (48), and the large volume of lung blood flow serves a source to resupply buffering agents. HCl is nearly 33 times less irritating than Cl2 (7), supporting that HCl is not pivotal in the mechanism of toxicity of Cl2.

The Direct Reaction of Cl2 with Biological Molecules on the Surface of the Respiratory Tract

Direct, fast reactions of Cl2 with biological molecules present on the respiratory tract surfaces are possible, but certain conditions must be met for these reactions to compete with the rate of Cl2 hydrolysis. Once Cl2 penetrates into the aqueous ELF, reactions of Cl2 with biological molecules will prevail when kapp × [S] ≥ k1 = 61.7 s−1, where kapp is the apparent second order rate constant for the reaction of Cl2 with biological molecule S, at pH 7, and k1 is the pseudo-first order rate constant for the forward reaction in Eq. 1. (It becomes necessary to employ apparent rate constants, kapp, to account for differences in reactivities for acid and base forms of ionizable substrates. kapp represents the global reactivity of a substrate S and is a function of pH and pKa. Calculations are done at pH 7.0, which is close to the pH of the ELF. This will be discussed in more detail below.) Thus, the ability of a biological molecule S to compete with Cl2 hydrolysis is a balance between its kapp (its intrinsic reactivity towards Cl2) and its concentration in the physiological compartment where the reactions occur. Due to the law of mass action, large values of kapp and high [S] will trend to favor reaction of Cl2 with S over its hydrolysis. The line depicted in Fig. 1 represents the pairs of kapp and [S] for which the rate of reaction equals the rate of Cl2 hydrolysis. For a given value of kapp, any value [S] above this line represents a scenario where the Cl2 rate of reaction with the biological molecule S exceeds the rate of hydrolysis. It is apparent that there are neither physical nor biological limitations for the rate of reaction of Cl2 with S to exceed the rate of hydrolysis of Cl2. For example, for values of kapp smaller than the limit for two species to diffuse together in aqueous solution (∼1 × 1010 M−1s−1) but larger than 1 × 105 M−1s−1, the minimum required [S] for reaction to compete with hydrolysis will vary between ∼6 nM and 0.6 mM. This range of concentrations for biological molecules is not particularly unusual in a fluid like the lung ELF. It appears we are the first to consider that direct reaction of Cl2 with biological molecules may indeed occur on the surface of the respiratory tract (for the current and contrasting view, see, for example, Refs. 1, 19, 26, 33, 56, 58, 89).

Fig. 1.

The boundary where the rate of reaction of Cl2 with a substrate S equals the rate of hydrolysis (log [S] = 1.79 − log kapp). Substrates with a product of rate constant (kapp) and concentration [S] that are below the rate of hydrolysis of Cl2 (61.7 s−1) fall below the line. Conservative estimates based on the reactivities for HOCl for ascorbate (squares), glutathione (triangles), and urate (circles), common small-molecular-weight water-soluble antioxidants found in lung epithelial lining fluid of humans and rats, are also shown on this reactivity map. In aqueous solution, Cl2 is typically more reactive than HOCl.

It is clear that Cl2 reactions with ELF biological molecules need not approach the limit imposed by aqueous diffusion rates to compete with the hydrolysis of Cl2. We now proceed to assess whether or not the rate constants for these reactions are sufficiently large for the lung ELF biological molecules, at the concentrations they are present in the ELF, to be able to compete with water for Cl2. Surprisingly little is known of the chemical reactivity and rates of reaction of aqueous molecular (solute) Cl2 with biological compounds. The reaction kinetics of some simple amines with Cl2 have been studied by Margerum et al. (47) and by Matte et al. (49) (Table 1). Unprotonated amines (including simple primary and secondary amines and amino acids) react with large kapp (in the range 108−109 M−1s−1) while their protonated counterparts are relatively unreactive (47, 49). The fraction of the amine that is unprotonated depends on the pKa of the amine, and thus, the apparent k2, kapp at pH 7, for some amines that have been studied can be calculated from their k2 and their degree of deprotonation [the fraction f = Ka/(Ka + [H+])] as shown in Table 1. The value for kapp = k2 × [Ka/(Ka + [H+])] also is sensitive to the pKa. The amines with smaller pKa values show the larger populations of unprotonated amine for the same pH value, and for similarly substituted amines, they also show the higher reactivity. Values of kapp for the amines listed in Table 1 range from 3.0 × 105 M−1s−1 for dimethylamine to 1.4 × 108 M−1s−1 for glycylglycine. The amino-terminal function in peptides and proteins is especially reactive toward Cl2 due to its low pKa. The products from the reaction of Cl2 with amines are the corresponding chloramines. Based on the magnitude of kapp and the concentrations necessary to out-compete hydrolysis (Fig. 1), this limited data set suggests the rates of reactions of Cl2 with lung ELF biological molecules that contain simple, and especially terminal, amine functionalities may out-compete Cl2 hydrolysis.

View this table:
Table 1.

Apparent rate constants for the reaction of amines with aqueous Cl2 at pH 7 and 25.0°C

The rates of reaction of Cl2 with a limited number of N-substituted amides and ureas also have been studied (see Supplemental Table S1; Supplemental data for this article is available online at the AJP-Lung web site.). These substrates provide important information of the general reactivity of these functional groups in various peptides, proteins, or other nitrogenous compounds that may be present in the lung ELF. Owing to the electron-withdrawing carbonyl functionality, the N-center in amides is significantly less reactive than in amines suggesting that amide functionalities in lung ELF biological molecules may not out-compete the rate of hydrolysis of Cl2. Biological molecules in the lung ELF that contain functionalities similar to N-alkyl-substituted ureas are more reactive than simple amides and may in some cases approach the rate of hydrolysis of Cl2.

No other nucleophiles have been studied, but it is likely that the reactivity of similarly substituted thiolates (RS), thiols (RSH), and unprotonated amines (RNH2) for reaction with Cl2 follows the order RS > RSH ≈ RNH2, as is predicted by reactivity parameters (18) and by comparison to the reactivity of these thiols and amines for ozone (69). Thus, one would expect thiolate and thiol functionalities in lung ELF biological molecules to become rapidly chlorinated by Cl2.

Comparison of the Reactivities of Cl2 and HOCl for Biological Molecules

Contrary to Cl2, the chemical reactivity and rate constants for the reactions of HOCl with a large number of substrates have been studied and have been recently reviewed (24). The reactivity of substrates of biological interest is briefly reviewed in the following section. Since the order of reactivity Cl+ > H2OCl+ > Cl2 > Cl2O > HOCl > ROCl > H2NCl > OCl is widely accepted (23, 79) and a larger number of rate constants for HOCl reactions with various substrates has been measured, the rate constants for the reactions of HOCl can be taken as lower limits for those for Cl2 with the same substrates. It can be seen in Table 2 that this relationship holds for the limited number of amines that have been studied, and k2 for Cl2 is always larger than k2 for HOCl, and similarly, kapp for Cl2 is always larger than kapp for HOCl (24, 47). This order of reactivity (i.e., that Cl2 reacts more rapidly than HOCl) is also observed for the N-substituted amides and ureas (80).

View this table:
Table 2

Comparison of the second order and apparent rate constants (both in M−1s−1) for the reaction of amines with aqueous Cl2 and HOCl at pH 7 and 25.0°C

The Reaction of HOCl with Biologically Relevant Molecules on the Surface of the Respiratory Tract

The case of small-molecular-weight antioxidants.

The lung ELF has high concentrations of the small-molecular-weight antioxidants ascorbic (AA) and uric acids (UA), and glutathione (GSH), and these antioxidants are very reactive towards HOCl (29, 60, 62) (Table 3) and Cl2 (since, as we explained above, Cl2 is at least as reactive as HOCl).

View this table:
Table 3.

Second order rate constants and apparent second order rate constants

Rats and humans have different relative abundance of these antioxidants (Table 4). Thus, the Sprague-Dawley rat contains on average 25 times more AA than the human, and 2.6 times less UA (probably due, in part, to the presence of urate acid oxidase in the rat and its absence in humans).

View this table:
Table 4.

Comparison of human and Sprague-Dawley rat bronchoalveolar respiratory tract lining fluid antioxidant profiles

It is interesting to note that these differences suggest that the efficacy of specifically formulated antioxidant cocktails may differ with species. The values given in Table 4 represent concentrations averaged over an extended lung surface since the bronchoalveolar lavage from which they were determined mixes the ELF from large regions of the posttracheal respiratory tract. Actual concentrations may vary with anatomical location (20, 42). During low-level Cl2 exposures (≤3 ppm), the tissue injury is largely limited to the nasal epithelium, but the inflammation extends to the alveoli at 9 ppm (8), suggesting that significant postnasal breakthrough begins just above the 3-ppm level. We are primarily concerned about acute exposures to >>10 ppm Cl2, and this is the reason we base our calculations on average lower respiratory tract ELF concentrations. However, as we propose is true for the lower respiratory tract ELF, the reactions of Cl2 with antioxidants influence reactive absorption of Cl2 by the upper respiratory tract ELF. In this regard, for example, significant concentrations of ascorbate and urate are also found in the nasal ELF (20, 42). Antioxidants that are present in the ELF at proximal sites initially further facilitate Cl2 reactive uptake at these sites, but antioxidants are present in finite quantities, and the scrubbing efficiency gradually subsides with time, resulting in increased longitudinal Cl2 penetration, especially for high Cl2 concentrations exposures.

Competition Between Reaction of Cl2 with Small-Molecular-Weight Antioxidants and the Hydrolysis of Cl2

The rate constants for reaction of small-molecular-weight antioxidants with Cl2 are unknown. However, the competition between reaction and hydrolysis for Cl2 can be conservatively estimated using the rate constants for HOCl for the antioxidants shown in Table 3 since Cl2 is more reactive than HOCl, and also considering that the lung ELF contains many other reactive substrates in addition to antioxidants. Thus, when the antioxidants are present in large excess with respect to Cl2 and the buffering capacity is large, the pseudo-first order rate constant for reaction with antioxidants can be calculated as kantioxidants=kUA[UA]+kGSH[GSH]+kAA[AA] (2) where the rate constants for individual antioxidants are their values of k2 given in Table 3.

For Cl2, in humans, kantioxidants > (3 × 105 M−1s−1)(207 × 10−6 M) + (4 × 107 M−1s−1)(109 × 10−6 M) + (1 × 107 M−1s−1)(40 × 10−6 M) = 62 + 4.4 × 103 + 4.0 × 102 = 4.9 × 103 s−1. Similarly, in rats, kantioxidants > (3 × 105 M−1s−1)(81 × 10−6 M) + (4 × 107 M−1s−1)(43 × 10−6 M) + (1 × 107 M−1s−1)(1,004 × 10−6 M) = 24 + 1.7 × 103 + 1.0 × 104 = 1.2 × 104 s−1.

One can now compare kantioxidants to the pseudo-first order rate constant for hydrolysis of Cl2 (61.7 s−1; see The Reaction of Chlorine with Water). Thus, for humans and rats, respectively, kantioxidants will be 79 and 190 times larger than the rate constant for hydrolysis of Cl2, strongly suggesting that reactions with substrates in the lung ELF can out-compete hydrolysis and that hydrolysis of Cl2 will become important only when reactive substrates had been depleted below critical thresholds, and contrary to the common assumption that Cl2 reacts with water to form HOCl (and HCl) (1, 19, 26, 33, 56, 58, 89). It is also interesting to note that while AA will be reacting with most of the Cl2 or HOCl in the rat lung ELF, for humans, the most important antioxidant is GSH. It is important to stress here that these are conservative estimates and that Cl2 will in all likelihood be even more reactive toward antioxidants compared with hydrolysis than these HOCl-based estimates indicate. In addition, it is important to recognize that the relative contributions of each antioxidant has been calculated for HOCl and that because Cl2 is more reactive than HOCl, the values of k2 for Cl2 will be larger and will trend to become more similar in value as they approach the diffusion-limited rate. Thus, for Cl2, the concentration of an antioxidant plays a relatively more important role than k2 in determining its contribution to kantioxidants than in the case of HOCl.

The calculations above were performed for typical antioxidant concentrations found in the lung ELF, but the antioxidant concentrations in the nose, for example, may be quite different from those found in the lung ELF. In this regard, the human nose ELF is known to contain much lower GSH concentrations than the lung ELF (83). These at times large longitudinal concentration gradients will affect local dose as well as the local reaction products.

The Case of Nitrite

Chemical cross talk between hypochlorite and nitrite to form nitryl chloride (NO2Cl), a nitrating species, has been hypothesized within the context of Cl2 toxicology (26). It was speculated that hypochlorite will react with nitrite, but that proposal lacked a kinetic analysis. For this interaction to be significant within the context of the lung surface compartment, the reaction of hypochlorite with nitrite would have to out-compete the reactions of hypochlorite with lung ELF small-molecular-weight antioxidants. The reaction kinetics of hypochlorite with nitrite were studied near ELF pH by Panasenko et al. (59) who reported a second order rate constant of 7.4 ± 1.3 × 103 M−1s−1 at pH 7.2 and 25°C. Concentrations of nitrite in the lung ELF of healthy and asthmatic subjects have been reported (27). In healthy individuals, the lung ELF nitrite concentrations range from ∼5 up to 20 μM, whereas nitrite in asthmatic subjects can be as high as 90 μM. Using the rate constant reported by Panasenko et al. and the highest nitrite concentration, one can now calculate that knitrite for hypochlorite can be as high as (7.4 × 03 M−1s−1)(9.0 × 10−5 M) = 0.66 s−1, a value that is much lower than the values for kantioxidants we calculated in the previous section, suggesting that nitrite will not effectively compete with lung ELF antioxidants for reaction with hypochlorite. Our prediction agrees with observations by Whiteman et al. (88), who found insignificant 3-nitrotyrosine formation when cells and cell lysates were exposed to both hypochlorite and nitrite. The experiments conducted by Whiteman et al. reveal that the reaction of hypochlorite with nitrite is slow compared with the reactions of hypochlorite with other biological compounds and/or that the reaction product from the reaction of hypochlorite with nitrite, NO2Cl, is an inefficient nitrating species under these conditions. It must be noted, however, that Panasenko et al. provided minimal detail with regard to their reaction kinetics experiments and data treatment, or the chemical reaction mechanisms that may operate, or how they related to the reaction rate law expression. In this regard, complex mechanisms with unwieldy rate laws have been proposed by other investigators (11, 43, 47). Lahoutifard et al. (43), for example, propose the decomposition of NO2Cl is rate-determining, and it remains to be investigated if the assumptions and procedures employed by Panasenko et al. can be validated.

Frenzel et al. (30) reported a second order rate constant for the reaction of Cl2 with nitrite of 2.6 × 106 M−1s−1. This value can be compared with the rate constant for the reaction of hypochlorite with nitrite at pH 7.2 obtained by Panasenko et al. (59), after correcting for the ionization of HOCl, 1.1 × 104 M−1s−1, indicating that Cl2 is more reactive than HOCl and in agreement with the order of reactivity we gave above (see Comparison of the Reactivities of Cl2 and HOCl for Biological Molecules). Using the highest reported lung ELF nitrite concentration, one can calculate that knitrite for Cl2 can be as high as (2.6 × 106 M−1s−1)(9.0 × 10−5 M) = 2.3 × 102 s−1, suggesting that the rate of reaction of Cl2 with nitrite can be faster than the rate of Cl2 hydrolysis under pathological conditions that produce high nitrite concentrations in the lung ELF. However, knitrite is just 4.7% and 1.9% of the minimum estimates for kantioxidants for reaction with Cl2 that we calculated above for humans and rats, respectively (see Competition Between Reaction of Cl2 with Small-Molecular-Weight Antioxidants and the Hydrolysis of Cl2), suggesting that even under these conditions, it is unlikely that Cl2 reaction with nitrite can proceed to a significant extent in the presence of lung ELF antioxidants. In this regard, we observed postexposure nitrite administration mitigates Cl2 inhalation injury (71), suggesting the beneficial effects of nitrite we observe are independent of the mostly detrimental formation of NO2Cl that may arise from the reaction of nitrite with HOCl produced from neutrophil myeloperoxidase, in accordance with our calculations.

The Case of Lipids

Unsaturated lipids are often considered to be likely targets for reaction of HOCl (see, for example, Ref. 1). Indeed, the reaction of alkenes with HOCl to form halohydrins is a classic example of electrophilic addition in organic chemistry. However, this reaction has a very small rate constant, and although when brought together alkenes and HOCl will eventually react with one another, it is unlikely that the unsaturated hydrocarbon moieties in lipids will be able to compete with other more reactive biological substrates for Cl2. Pattison et al. (62) estimate the rate constant for reaction of hypochlorite with a double bond in a phospholipid fatty acyl chain to be ∼10 M−1s−1 at pH 7, a value that can be compared, for example, to the much larger rate constants (kapp) for antioxidants given in Table 3. Taking into consideration that lung ELF antioxidants are present in large concentrations (see Table 4), one can conclude that it is very unlikely that HOCl will be able to react directly with unsaturated hydrocarbon moieties in lipids to any significant extent in the lung ELF. We were unable to find any relevant rate constants for the reaction of Cl2 with substrates containing alkenyl residues, but we expect, as we concluded for HOCl, that the reaction of Cl2 with unsaturated hydrocarbon moieties in lipids will also be too slow to compete with reactions with antioxidants. In this regard, Cl2 and HOCl contrast sharply with ozone (O3) and with nitrogen dioxide (NO2). In the case of O3, this reaction is significant in the lung ELF because O3 reacts with double bonds in the fatty acid chains of lipids with a rate constant of ∼1 × 106 M−1s−1 (32, 66, 68), and this rate constant is ∼10,000 times larger than the corresponding rate constant for hypochlorite at pH 7 (10 M−1s−1). NO2 also reacts rapidly with polyunsaturated fatty acid moieties, with a rate constant of ∼2 × 105 M−1s−1 (67).

It is perhaps less obvious, but some lipids carry polar and more reactive functional groups in their head groups than the alkenyl groups found in their non-polar moieties. An example is the primary amine functional group that occurs in phosphatidylethanolamine, phosphatidylserine, sphingosine, and sphinganine. Of these amino-phospholipids, phosphatidylethanolamine and phosphatidylserine are generally the more abundant and thus have a higher probability to react with Cl2 or HOCl. Together, phosphatidylethanolamine and phosphatidylserine can account for up to 9.4% of the total phospholipids in the lung ELF (40, 54, 72). Humans have ∼20 ml of lung ELF, which contains ∼500 mg of phospholipids. Assuming an average molecular weight of 800 atomic mass units, one arrives at a combined bulk concentration of 3 mM for these amino-phospholipids in the ELF. As can be seen in Tables 1 and 2, the reactivity of primary amines is modulated by their pKa. The pKa of the amine groups in phosphatidylethanolamine and phosphatidylserine are 9.6 and 9.8, respectively (81), suggesting that amino-phospholipids reactivity for Cl2 and hypochlorite will be similar to that of α-alanine. This implies that the amine groups in phosphatidylethanolamine and phosphatidylserine will have kapp of ∼2 × 106 M−1s−1 and ∼4 × 104 M−1s−1 for Cl2 and hypochlorite, respectively (see Tables 1 and 2), at pH 7, the pH of the lung ELF. The overall pseudo-first order rate constants can then be calculated from kamino-phospholipids = (2 × 106 M−1s−1)(0.003 M) = 6 × 103 s−1 for Cl2, and kamino-phospholipids = (4 × 104 M−1s−1)(0.003 M) = 120 s−1 for hypochlorite.

Pseudo-first order rate constants for the reaction of water-soluble small-molecular-weight antioxidants and amino-phospholipids with Cl2 in human lung ELF are compared in Table 5. It can be seen that kamino-phospholipids for Cl2 lies within the range of the minimum estimates for the water-soluble small-molecular-weight antioxidants and that the reaction of Cl2 with the amino-phospholipids may indeed occur. Moreover, at the air-fluid interface, the alveolar ELF is covered by a monolayer of phospholipids, and any amino-phospholipids that may be present in this lipid monolayer would be more likely to react with Cl2. Tracheal aspirates have been reported to contain a larger proportion of amino-phospholipids compared with bronchoalveolar lavage (13), but because the conductive airways contain less surfactant, phospholipid chloramines formation is expected to be less important in the conductive airways.

View this table:
Table 5.

Pseudo-first order rate constants for the reaction of water-soluble small-molecular-weight antioxidants and amino-phospholipids with Cl2 in the human lung ELF

With regard to the toxicological implications for the formation of organic chloramines, these compounds are usually found to be cytotoxic (90). Biologically relevant chemical reactions of chlorine transfer by several organic chloramines had been observed (51, 61, 63, 76, 78), and although phospholipid chloramines per se had not yet been studied, we believe phospholipid chloramines may undergo similar chemical reactions and cytotoxic activities. In analogy to the chemistry of organic chloramines, phospholipid chloramines may transfer the chlorine atom to nucleophilic centers in biological molecules or decompose to form modified phospholipid with aldehydic groups on the polar head groups. In this regard, for example, chlorine transfer is thought to occur from histidine and lysine chloramine residues in lysozyme and insulin (61), and from various organic chloramines to thiol residues in α1-antitrypsin, transthyretin, and albumin (78), to peroxiredoxin 2 (76), and to creatine kinase and glyceraldehyde-3-phosphate dehydrogenase (63), and to methionine residues in IκB (51), in several cases resulting in loss of function or inactivation. The antioxidants ascorbate and glutathione were found to partially repair chlorinative damage (21, 64) and may help tissues detoxify from organic chloramines. The biological activities of phospholipids with aldehydic groups on their polar head groups, which are products expected to be formed from the spontaneous decomposition of the parent phospholipid chloramines, a process that has been recently confirmed for some N-chloroaminophospholipids (28, 41), are unknown. However, these compounds may be bioactive in light of the biological activities that had been reported for similar phospholipid aldehydes containing the aldehyde function on the acyl residue attached to the sn-2 position of the phospholipid (22, 38).

The Case of Proteins

Contrary to protein damage by reactive free radicals, which is unselective and widespread, hypochlorite reacts in quite a selective manner with amino acids in proteins (62). Since hypochlorite reacts easily but only with a relatively small number of amino acids, specific protein reactivity will strongly depend on protein amino acid composition and protein concentration. Amino acid reactivity follows the order methionine > cysteine > histidine ≈ N-terminal amine > tryptophan > lysine > tyrosine > arginine, with the rate constants for this group of amino acids spanning over six orders of magnitude. All the other amino acids are quite unreactive toward hypochlorite. Interestingly, disulfide bonds are also reactive centers for hypochlorite, with a rate constant slightly larger than that of histidine.

Relation Between the Mass of Inhaled Cl2 and the Antioxidants Pool in the Lung ELF

Antioxidants are present in the lung ELF at high concentrations and are very reactive toward Cl2 and HOCl. For these reasons, antioxidants will scavenge part of the Cl2 and HOCl, as these species penetrate and travel across the lung ELF. It is therefore of interest to estimate the antioxidant capacity that the lung ELF has to scavenge Cl2 and HOCl (Table 6).

View this table:
Table 6.

Approximate moles of antioxidants (sum of uric acid, glutathione, and ascorbic acid) present in human and rat lung ELF

Assuming ideal gas behavior, a gas mixture that contains Cl2 at room temperature (22°C) and atmospheric pressure will contain 41.3 nmol Cl2·ppm−1·l−1. This value allows the calculation of the moles of Cl2 inhaled for any inhaled volume and Cl2 concentration from using Eq. 3: nmoles of Cl2=(41.3nmol Cl2·ppm1·l1)×(Inhaled volume in liters)×(Cl2concentration in ppm) (3)

The same value also allows the calculation of the inhaled volume that contains the number of nanomoles of Cl2 that is equal to the nanomoles of antioxidants present in the lung ELF using Eq. 4: Inhaled volume=(antioxidants in nmol)×(41.3nmol Cl2·ppm1·l1)1×(Cl2concentration in ppm)1 (4)

As an example, for a concentration of Cl2 of 400 ppm, and using the antioxidant concentrations in Table 6, the inhaled volumes that contain and amount of Cl2 that is equimolar to antioxidants in the lung ELF are 0.43 and 0.0068 l, for humans and rat, respectively. These inhaled volumes correspond approximately to 1 and 5 breaths for human and rat, respectively. Experimental data indicate that exposure of rats to Cl2 for 30 min results in significant depletion of ascorbate in the BAL (from 17 to 2.5 μM) (45). In reconciling these theoretical calculations with experiment, it is important to recognize that part of the inhaled Cl2 may react from the nasal epithelium to the trachea, and we did not consider reactive uptake of Cl2 at these loci. Furthermore, the depletion of antioxidants will likely be uneven with Cl2 mass transfer also depending on local antioxidant concentrations, reactive gas fluid dynamics parameters, the distribution of ventilation, and the proximal to distal decrease in Cl2 concentration. Antioxidants may be released from the underlying tissue to compensate for the depletion caused by their reactions with Cl2. In addition, damaged cells may release intracellular antioxidants into the lung ELF so that the ELF may not be depleted of antioxidants in the short term.

Products from the Reaction of Cl2 or HOCl with the Principal Small-Molecular-Weight Water-Soluble Antioxidants

Cl2 and HOCl are expected to afford similar reaction products, but Cl2 reacts faster and is less selective than HOCl. Thus, because of these differences in reactivity and selectivity, when several substrates are available for reaction, HOCl can be more discriminant than Cl2. For example, values of k2 for reaction of HOCl with glutathione (GSH), AA, and UA (Table 3) relate as GSH:AA:UA (400:100:3), and we can expect the corresponding values for Cl2 to be more closely related than for HOCl. Furthermore, when more than one reactive center exists on the same biomolecule molecules (as in GSH), again, HOCl will be more discriminatory than Cl2. For example, two reactive centers exist in GSH, the thiol group on the cysteine residue that will yield a sulfenyl chloride, and the free α-amino group of the γ-glutamyl residue that will yield the corresponding chloramine (reaction at amide bond nitrogen will be negligible). The reactivity of the thiol group in GSH for hypochlorite can be estimated from data for cysteine obtained by Armesto et al. (3) while the reactivity of the γ-glutamyl residue α-amino group for hypochlorite can be estimated to be similar to the reactivity of alanine (24) (Table 7). The overall kapp estimated this way for the reaction of GSH with hypochlorite agrees well with the experimental value [estimated kapp = 1.8 × 107 M−1s−1 (Table 7); experimental kapp = 3 × 107 M−1s−1 (Table 3)]. The reactivity of GSH at pH 7 is largely due (>99%) to its thiol group because it has a larger k2 than the γ-glutamyl residue α-amino group, and also has a larger molar fraction of the reactive thiolate form compared with the unprotonated amino group at this pH. Thus, from the reaction of hypochlorite with GSH at pH 7, one should expect >99% yield of the glutathione sulfenyl chloride (GSCl) as the primary reaction product and just a trace of the glutathione chloramine. It is also interesting to note that most of the reactivity at pH 7 is due to HOCl and not to OCl.

View this table:
Table 7.

Analysis of the reaction kinetics of the 2 reactive centers in glutathione

There is no experimental data on reactivity of Cl2 for thiol groups, but k2 for the reaction of the glutathione thiolate with Cl2 is expected to be larger than the corresponding k2 for HOCl but will not exceed 1 × 1010 M−1s−1 (the diffusion-limited rate constant). Correcting for the ionization of the thiol group in GSH results on kapp = 2.0 × 108 M−1s−1. As is true for hypochlorite, the reactivity of γ-glutamyl residue α-amino group in GSH for Cl2 is also expected to be similar to that of alanine. Thus, we take k2 = 1.0 × 109 M−1s−1 (Table 1) for the reaction of Cl2 with this amino group in GSH. Correcting for the ionization of the amino group results in kapp = 2.5 × 106 M−1s−1. Thus, when Cl2 is the chlorinating agent, one should expect ∼99% yield of the sulfenyl chloride GSCl and 1% of the GSH chloramine. Our calculations predict the GSH chloramine is a minor product for both Cl2 and HOCl, but Cl2 affords a higher yield than HOCl.

Ascorbic acid reacts with HOCl or Cl2 by a two-electron oxidation mechanism yielding primarily dehydroascorbic acid (DHA). The reaction appears to proceed via the formation of the 2-deoxy-l-ascorbyl ester of HOCl, which then undergoes rapid spontaneous hydrolysis to DHA. The one-electron oxidation of ascorbic acid does not occur to any significant extent, and the ascorbyl radicals are produced only in trace yields (29). The two-electron oxidation of ascorbate occurs because HOCl is a strong two-electron oxidant [Eo′(HOCl/H2O, Cl) = 1.28 V] but only a relatively weak one-electron oxidant [Eo′(HOCl/HOCl·−] = 0.24 V; Eo′(HOCl/HO·, Cl) = 0.26 V; Eo′(HOCl, H+/H2O, Cl·) = 0.16 V; standard reduction potentials at pH 7 were calculated from data in Refs. 77 and 86. The same trend is reflected on the reduction potentials of aqueous Cl2 [(Eo′(Cl2/2Cl) = 1.40 V vs. Eo′(Cl2/Cl2·−) = 0.70 V]; standard reduction potentials at pH 7 were calculated from data in Refs. 77 and 86. Impaired cellular capacity to reduce and recycle DHA may result in rapid local AA depletion due to the spontaneous and irreversible hydrolysis of DHA to 2,3-diketogulonic acid at neutral pH.

The reaction of UA with HOCl affords allantoin in ∼35% yield together with traces of parabanic acid and other products (82). Regrettably, nearly 60% of the mass balance remains as undefined. As is the case for AA, the reactions of UA with HOCl or with Cl2 also appear to be, at least in part, two-electron oxidations that proceed through the formation of 5-chloroisourate, an unstable intermediate that first hydrolyzes to form 5-hydroxyisourate and finally to allantoin (4, 39). Primary and secondary products from the reaction of Cl2 or HOCl with the principal small-molecular-weight water-soluble antioxidants are shown in Fig. 2.

Fig. 2.

Products from the reaction of Cl2 or HOCl with the principal small-molecular-weight water-soluble antioxidants and representative phosphatidylethanolamine and phosphatidylserine.

Penetration Distance of Cl2 into the Lung ELF

Approximate diffusion distances (axial penetration into the lung ELF), λ, for a reactive species, such as Cl2, that diffuses with a diffusion coefficient, D, may be estimated from the Einstein-Smoluchowski equation (Eq. 5) for the time, t, that would take Cl2 to decay to a significant extent due to its chemical reactions, as has been done for other oxidants (6, 65). λ=2Dt (5)

The diffusion coefficient D for Cl2 in the lung ELF is unknown but it is probably between the corresponding values for water and the cytosol. In this regard, one can estimate a value of ∼2 × 10−5 cm2s−1 for water at 37°C by extrapolating data from Himmelblau (34) while Verkman (85) estimates that the cytosolic value is ∼25% of the value for water. Thus, we here use 1 × 10−5 cm2s−1 for the diffusion coefficient of Cl2 in the lung ELF.

The time it takes for 90% of the Cl2 initial concentration to decay, t90, can be calculated from t90 = ln10/k, where k is the first order or pseudo-first order rate constant for chemical reactions of Cl2 in the medium of interest. In Table 8, values of t90 for water, rat, and human lung ELF were calculated according to the respective rate constants in these media (see above for the derivation of these rate constants). Similarly, diffusion distances, λ, for Cl2 in these media were for these values of t90 calculated using Eq. 5. For comparison, also shown in Table 8 are diffusion distances calculated using Eq. 6, an expression derived for a point source producing Cl2 at a constant rate in a homogenous reactive medium, as has been done previously to estimate diffusion distances for nitric oxide in tissue (25, 44). λ=ln[Cl2]0[Cl2]λ·2Dln2k (6)

View this table:
Table 8.

Estimated times for 90% decay for Cl2 and diffusion distances

In pure water, the decay of Cl2 is slower than in aqueous solutions that contain the principal water-soluble low-molecular-weight antioxidants at concentrations found in the human and rat lung ELF. Consequently, the two simple models we used predict the longer diffusion distances, 9 and 16 μm, respectively, in pure water. It is interesting to note that these distances are much larger than the alveolar ELF thickness (9), suggesting that Cl2 may indeed reach and chemically react with biological molecules in the underlying tissue strata. These results are in conflict with the popular belief that Cl2 entirely hydrolyzes to HOCl (and relatively innocuous HCl) in the lung ELF and then goes on to inflict tissue damage (1, 19, 26, 33, 56, 58, 89).

We determined that it would be incorrect to assume that Cl2 will hydrolyze in the lung ELF because reaction of Cl2 with the lung ELF small-molecular-weight antioxidants is much faster than the hydrolysis of Cl2 (see Competition Between Reaction of Cl2 with Small-Molecular-Weight Antioxidants and the Hydrolysis of Cl2). Furthermore, should Cl2 undergo hydrolysis, Cl2 would be sufficiently long-lived to be able to traverse the lung ELF, reaching to the underlying tissue. Thus, the paradigm that Cl2 hydrolyzes in the lung ELF is inconsistent with reaction/diffusion kinetics and thus is unlikely to be correct. When one considers that Cl2 reacts with lung ELF antioxidants, both models predict shorter penetration distances, suggesting that significant reaction will occur before Cl2 can traverse the lung ELF.

The results from our simple “Cl2 penetration” analysis add to our reaction kinetics analysis that predicts chemical reactions of Cl2 with antioxidants in the lung ELF are much faster than Cl2 hydrolysis, and thus, the direct reactions of Cl2 with biological molecules must also be considered to better understand the mechanism related to Cl2 toxicology. Cl2 can react directly with biological molecules, thereby bypassing the formation of HOCl, and this has additional important mechanistic implications. Cl2 and HOCl are expected to give the same reaction product for a biological molecule with a single reactive center; for example, both Cl2 and HOCl will afford the same organic chloramine from a given organic amine. However, we predict Cl2 will react faster and less selectively than HOCl, and thus, the reaction product profile when many biological molecules compete for reaction with Cl2 or HOCl may vary widely.

Antioxidant Therapy and the Mitigation of Lung Injury Due to Inhalation of Cl2

Above we estimated that lung ELF antioxidants can be severely depleted in a few breaths when [Cl2] exceeds 100 ppm. Lung ELF antioxidants are not just a mere first line of defense but also play important biochemical roles in cell signaling, in terminating free radical chain reactions, and have been found to partially repair HOCl-related chemical modifications to biological molecules (21, 64). Together, these calculations and observations provide a rationale for antioxidant therapy in mitigating lung injury from Cl2 exposure.

In a recent study (45, 91), we confirmed exposures of Sprague-Dawley rats to 184 ppm or 400 ppm Cl2 for 30-min decreased ascorbate detected in bronchoalveolar lavage fluid (BALF) as well as in lung tissue, with effects more pronounced after 400 ppm exposure. Moreover, preexposure administration of an antioxidant cocktail ameliorated the effects of Cl2 as measured by smaller decrements in BALF ascorbate and PaO2 (alveolar Po2) and reduced lung permeability as determined by BALF protein. Further studies are underway regarding various antioxidant formulations as well as administration routes.

Extending Antioxidant Therapy to Phosgene Exposures

Existing symptom-oriented therapy guidelines recommend similar approaches to treat chlorine and phosgene (COCl2) exposures. However, it is important to note that the chemistries of Cl2 and COCl2 are very different. For example, as described herein, Cl2 and HOCl are oxidants and chlorinating reagents, whereas COCl2 is a bifunctional acylating reagent that is capable of cross-linking proteins (16, 36) that does not oxidize or chlorinate biological molecules. Thus, although exposures to COCl2 will result in lung ELF GSH depletion and antioxidant therapy may help restore GSH levels and redox status, the approaches that are described in this article do not necessarily apply to COCl2.


Our studies strongly suggest that Cl2 will predominantly react directly with biological molecules in the lung ELF, such as small-molecular-weight antioxidants, and that the hydrolysis of Cl2 to HOCl (and HCl) can be important only when these biological molecules have been depleted by direct chemical reaction with Cl2. Thus, our contention contrasts sharply with the general belief that HOCl is the obligatory transient oxidant that is formed in the lung ELF and mediates the toxicity of Cl2. We find such proposal open to question on two grounds: 1) the reaction of water-soluble small-molecular-weight antioxidants with Cl2 out-competes Cl2 hydrolysis, and 2) Cl2 hydrolysis is sufficiently slow that Cl2 would be able to traverse an aqueous layer of the lung ELF thickness, in particular in the alveolar region, and reach the underlying tissue strata.

A competition kinetics approach was undertaken to rank the lung ELF small-molecular-weight antioxidants according to their reactivity for Cl2 and predict the reaction products that will be formed. Consistent with our predicted efficient chemical reactions of inhaled Cl2 with lung ELF small-molecular-weight antioxidants, we demonstrated that preexposure administration of an antioxidant cocktail ameliorated the effects of Cl2. We are currently investigating the postexposure efficacy of various antioxidant cocktails. Antioxidant therapy may help restore the lung ELF antioxidant redox status back to preexposure levels and partially repair chlorine-induced covalent modifications.


This research is supported by the CounterACT Program, National Institutes of Health (NIH) Office of the Director, and NIH Grants 5U01ES-0115676-04, 5U54ES-017218-02, and 3U54ES-017218-02S1. We also acknowledge support from NIH Grants R01-HL-054696 and P01-ES-11617.


S. Matalon served as a consultant for Sepracor and received an honorarium for participating in their annual scientific meeting. He has received industry-sponsored grants from Talecris, Inspire Pharmaceuticals, and Sepracor.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
View Abstract