- Email: [email protected]
Inhibition of lipopolysaccharide induced acute inflammation in lung by chlorination
Journal of Hazardous Materials 303 (2016) 131–136
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Inhibition of lipopolysaccharide induced acute inflammation in lung by chlorination Jinshan Zhang a , Jinling Xue a , Bi Xu a , Jiani Xie a , Juan Qiao b,∗ , Yun Lu a,∗ a b
Environmental Simulation and Pollution Control State Key Joint Laboratory, School of Environment, Tsinghua University, Beijing 100084, China Department of Chemistry, Tsinghua University, Beijing 100084, China
h i g h l i g h t s • • • •
Chlorination is effective to reduce the inflammation inducing capacity of LPS in lung. LAL-detected endotoxin activity is not correlated to the potency of inflammation induction. Alkyl chain of LPS was chlorinated in chlorination process. LPS aggregate size decreases after chlorination.
a r t i c l e
i n f o
Article history: Received 17 June 2015 Received in revised form 25 September 2015 Accepted 12 October 2015 Available online 22 October 2015 Keywords: Lipopolysaccharide Chlorine Inhalation exposure Aggregate size Reclaimed water
a b s t r a c t Lipopolysaccharide (LPS, also called endotoxin) is a pro-inflammatory constituent of gram negative bacteria and cyanobacteria, which causes a potential health risk in the process of routine urban application of reclaimed water, such as car wash, irrigation, scenic water refilling, etc. Previous studies indicated that the common disinfection treatment, chlorination, has little effect on endotoxin activity removal measured by Limulus amebocyte lysate (LAL) assay. However, in this study, significant decrease of acute inflammatory effects was observed in mouse lung, while LAL assay still presented a moderate increase of endotoxin activity. To explore the possible mechanisms, the nuclear magnetic resonance (NMR) results showed the chlorination happened in alkyl chain of LPS molecules, which could affect the interaction between LPS and LPS-binding protein. Also the size of LPS aggregates was found to drop significantly after treatment, which could be another results of chlorination caused polarity change. In conclusion, our observation demonstrated that chlorination is effective to reduce the LPS induced inflammation in lung, and it is recommended to use health effect-based methods to assess risk removal of water treatment technologies. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Wastewater reclamation and reuse is important in sustainable development and water resource conservation strategy. The ratio of wastewater reuse is increasing rapidly and it has been used for flushing toilet, car wash, irrigation, refilling scenic water body, etc, which are closely related to human health [1]. However, there are many risk factors identified from reclaimed water contact, in which the microbial pathogens and their derivatives are the most commonly discussed [2]. Lipopolysaccharide (LPS) as a pro-inflammatory constituent of gram negative bacteria and some
∗ Corresponding authors. E-mail addresses: [email protected] (J. Qiao), [email protected] (Y. Lu). http://dx.doi.org/10.1016/j.jhazmat.2015.10.024 0304-3894/© 2015 Elsevier B.V. All rights reserved.
cyanobacteria, is released by cell multiplication, death, and lysis. It is relatively heat stable and similar in structure regardless of source [3,4], which can cause serious health issues including fever, inflammation, asthma, diarrhea, shortness of breath, intravascular coagulation and even death [5–7]. LPS is composed of lipid A, core oligosaccharide and O antigens. The toxicity of LPS is associated with the structure of lipid A whereas the immunogenicity is dependent on polysaccharide component [6,8,9]. The size of LPS monomer is ranging from 10 to 20 kDa, and it can form aggregates with high stability like micelle or vesicle with the size over 1000 kDa due to their hydrophobic interaction in aquatic systems [10,11]. It is known that only aggregates of LPS are biologically active [12]. The exposure routes to LPS include inhalation, intravenous injection and oral intake. In the urban use of reclaimed water, inhalation of aerosolized water droplets is the main exposure. Endotoxin activity in the secondary
132
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
effluent from a multi-country investigation report ranges from 201 to 20100 EU/mL, and the average level is 2994 EU/mL. Although the average endotoxin activity reduced to 1584 EU/mL after advanced treatments, the endotoxin activity still ranges 3–19700 EU/mL in the reclaimed water [13]. In order to minimize the health risk of exposure to reclaimed water containing LPS, microbial pathogens and other chemical compounds, the tertiary treatments are crucial to the safety of wastewater reuse [14], such as chlorination, ozone disinfection, UV irradiation, ultrafiltration and so on. Chlorination is one of the most widely used disinfection methods in wastewater reclamation to guarantee water quality due to its easiness to operate and low cost [15]. Endotoxin inactivation by chlorine had been studied previously. Anderson et al. [4] examined chlorination effect on tap water spiked with LPS standard (Escherichia coli strain O55:B5) and found that chlorine can hardly remove the endotoxin activity of LPS with the maximum inactivation rate of 1.4(EU)/mLh at residual chlorine from 2 mg/L up to 100 mg/L with reaction time more than 100 h. Rapala et al. [3] indicated that chlorination had no effect on endotoxin activity by studying endotoxin removal during drinking water treatment. Gehr et al. [16] found that endotoxin activity in low endotoxin-containing water samples (7–10 EU/mL) can both be moderately decreased or increased by chlorination. Can et al. [17] investigated endotoxin activity following different treatment process steps and indicated that endotoxin increased after chlorination, these were caused by bacteria lysis and endotoxin release. Similar results were also found in secondary effluent [9]. However, all these studied used LAL assay to assess treatment effects. One thing, which is worth of attention, is that the validity of using the in vitro LAL assay for assessing the inactivation efficiency of LPS has not been verified, since chemical treatment of LPS can cause changes at molecular level. In this study, we measured the chlorinated LPS activities by LAL assay and a mouse inhalation exposure model simultaneously. LAL assay showed moderate increase of endotoxin activity, while lung inflammation was significantly reduced by chlorination, which indicates that chlorination can reduce the health effect caused by LPS inhalation and LAL is not suitable to evaluate the treatment efficiency. The mechanism of LPS inactivation by chlorine was also investigated. 2. Materials and methods 2.1. Animals The study protocol was approved by the Institutional Animal Care and Use Committee of Tsinghua University and was performed according to the recommendations in the Guide for the Care and Use of Laboratory Animals. We made our best efforts to minimize suffering and the number of experimental animals. Male ICR mice aged 8–9 weeks were purchased from Vital River company (Beijing, China) and kept in a barriered facility for experimental animals in Tsinghua University. 2.2. Inhalation exposure HOPE-med 8050 dynamic inhalation system (HOPE Ltd, Tianjin, China) was used to provide an exposure chamber with a constant concentration of water aerosol flow. The aerosol concentration versus samples fluid rate was calibrated by an atmospheric sampler (Casella, UK) with drying columns. The water pumping rate was set to 0.7 mL/min, and the water aerosol concentration in the chamber was about 9 mg/L. Other parameters are as follows: carrier gas flow velocity 3.00 m3 /h, temperature 25 ◦ C, and exposure time 6 h.
2.3. Bronchoalveolar lavage fluid (BALF) and cells counting Three hours after inhalation exposure, mice were sacrificed by cervical dislocation. In our previous study, three-hour has been found to be a time point for a peak of cellular response induced by LPS. BALF was collected by cannulating the upper part of trachea using syringe in 2 mL PBS (pH7.4) containing 0.5% bovine serum albumin. 300 L BALF was centrifuged at 800 rcf for 5 min at 4 ◦ C, and resuspended in 1.5 mL dilution reagent (supplementary reagent of animal hematology analyzer, Mindray, China) for the concentration of white blood cells (WBC) analysis by animal hematology analyzer (Mindray, China). The rest BALF was also centrifuged, and resuspended again in 10 L PBS containing 0.5% BSA for smear. For smear, every 5 L sample was spread on a slide (Shitai Ltd., China), dried in the hood, and then stained by Wright–Giemsa method. Briefly, the slides were soaked into methanol for 15 min and dried in the air. Then they were stained by Wright solution for 3 min, Wright/PBS (PBS pH 6.6, volume ratio 1:1) solution for another 3 min, rinsed thoroughly with PBS (pH 6.6). Next, the slides were stained by Giemsa solution for 10 min, rinsed thoroughly by deionized water. Finally, the slides were dried and mounted for microscopy (Leica, Germany). 300 to 600 white cells were counted for each sample to calculate the proportion of polymorphonuclear neutrophils (PMNs).
2.4. LAL assay Standard Limulus Amebocyte Lysate assay kits were purchased from Xiamen Houshiji Ltd., China, and performed according to the manufacturer’s instructions.
2.5. Chlorination of LPS LPS (L2637, L2880) purchased from Sigma–Aldrich was extracted from E. coli serotype O55:B55, and dissolved in pyrogenfree water at high concentration, and stored at 4 ◦ C for further use. Synthetic LPS containing water was prepared by spiking LPS stock solution to make the concentration about 2000 EU/mL, and was adjusted to pH 7.0 by adding phosphate buffer. Chlorination was carried out by adding sodium hypochlorite stock solution (15 g/L as available chlorine) except for the chlorine-free control and concentration of available chlorine was detected by chlorineSpectrophotometric method (HANNA, Italy) before use. At the end of the pre-determined reaction time, excessive sodium thiosulfate (2 times the maximal consumption of chlorine) was added to neutralize residual chlorine, the whole process was performed in the dark. The chlorine dosage ranged from 0 to 50 mg/L and the reaction time ranged from 5 to 240 min were selected. Then, the water samples were ready for LAL assay and inhalation exposure experiments. The chlorination and quenching process did not change the pH, which was maintained at 7.17 ± 0.02. 2.6. Scanning electron microscopy (SEM) and 1 H NMR spectroscopy The concentrations of LPS in synthetic water samples were 0.5 mg/mL for SEM and 1 mg/mL for NMR spectroscopy. Two samples before and after chlorination (50 mg/L free chlorine, reaction time was 4 h for SEM and 15 min for NMR) were prepared and dialyzed to deionized water by 7KD dialysis cassette (Pierce Biotechnology, USA). After dialysis, the samples were freeze-dried for SEM (Hitachi, Japan). For NMR spectroscopy, 2 mg freeze-dried samples were dissolved in 0.5 mL D2 O, and performed in the Ana-
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
133
lytical Center of Tsinghua University, the NMR model is ECA-600, JEOL Ltd. Japan. 2.7. Aggregate size determination The concentration of LPS in synthetic water samples used for aggregate size determination was 0.01 mg/mL (81000 EU/mL detected by LAL). Aggregate size was examined by Nano-particle size and Zeta potential analyzer (Beckman Coulter, USA).
3. Results 3.1. Endotoxin activity and inflammatory effects of chlorinated LPS As mentioned above, the removal of endotoxin activity by chlorination had some inconsistency in the literatures [3,4,9,16,17]. In this study, the endotoxin activity of LPS cannot be inactivated by chlorine from the LAL assay results (Fig. 1a). Instead, chlorination increased endotoxin activity moderately with all tested chlorine concentration. However, when tested in mouse inhalation model, consistently significant decrease of about 60% of inflammatory effect was observed in BALF by counting the percentage of PMNs (Fig. 1b). It is interesting that the decrease of inflammation stopped at a certain level regardless of quantity of chlorine, and only less than 5% of chlorine was consumed after 4 h treatment at 20 mg/L and 50 mg/L levels. This result suggests that chlorination can reduce the inflammation inducing ability of LPS, but cannot remove it completely. The possible effects of disinfection reagents were also tested (control C1), which was proved to be negligible (Fig. 1b). The concentration of WBC in BALF also decreased after chlorination, which further demonstrated the inhibition of inflammation induction (Fig. 1c). When it was confirmed that chlorination can inhibit the inflammation inducing ability of LPS, the next important question is that whether the LPS can be inactivated at common chlorination condition, since 4 h are usually much longer than the real treatment time. It was found that the inflammatory response decreased rapidly at the first 15 min, and reached a critical level at 60 min. And this pattern is exactly opposite to the results of LAL assay (Fig. 2a). The WBC concentration in BALF also dropped starting from the reaction time of 5 min, and no obvious further decrease was observed with longer treatment (Fig. 2b). 3.2. Chemical composition of LPS after chlorination It is well known that chlorine treatment generates chlorinated disinfection products. It should be one of the main reasons that affect the inflammation induction. First, infrared spectrum was taken from the freeze–dried samples, but no obvious peak change was observed (data not shown). Energy dispersive spectrometer was used to analyze the chemical composition of dialyzed and freeze-dried samples. And the atom ratio between carbon and chlorine is about 200:1 after 4 h reaction (Table 1), which means the chlorination happened but only in a few sensitive sites of LPS. To figure out the positions of these sensitive sites, 1 H NMR spectroscopy was performed. There are three main groups of hydrogen peaks (Fig. 3). According to the analysis of a previous report [18], the hydrogens between 1 and 2 ppm should be on the alkyl chains, while the hydrogens with chemical shift ranged from 3.3 to 4.5 ppm should be on saccharide rings. If the peak area was normalized by the group 1, both the hydrogen quantities of group 2 and 3 were decreased, which means both hydrogens on alkyl chains and saccharide rings were replaced by chlorines. The chlorination on alkyl
Fig. 1. Chlorination inhibits inflammation inducing capacity of LPS. (a) LAL results of endotoxin activity after 4 h treatment of different doses of free chlorine, 10, 20, 30, and 50 mg/L, respectively. (b) The percentage of PMN in total cells of BALF. Mice were exposed for 6 h and sacrificed 3 h after exposure. (c) Total WBC concentrations in BALF. (C1, ultrapure water chlorination by 50 mg/L chlorine without additional LPS; C2, unexposed; n = 5 mice/group). The values are means ± standard deviation (SD). The asterisks represent endotoxin activity and white blood cells of chlorinated samples are significantly different from that of untreated LPS exposed group (one way ANOVA, p < 0.05).
chains can affect its binding activity to LPS-binding protein (LBP) and its ability to form aggregates. 3.3. Decrease of LPS aggregate size after chlorination Since the alky chain was chlorinated as mentioned ahead, the aggregation of LPS monomers may also change. Therefore, the average aggregate size was tested before and after chlorination. It was found that the aggregate size rapidly decreased within the first
134
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
Fig. 2. Chlorination effects occurs at early treatment stage. (a) LAL result of LPS level, PMN percentage and (b) relevant WBC concentration in BALF versus contact time at the presence of 20 mg/L free chlorine. The values are means ± standard deviation (SD). The asterisks represent endotoxin activity and white blood cells of chlorinated samples are significantly different from that of untreated LPS exposed group (one way ANOVA, p < 0.05) (n = 5 mice/group).
Fig. 3. One dimensional 1 H spectroscopy with integration values of peak areas by making the peak at chemical shift of 1.08 ppm as a reference. (a) LPS before chlorination and (b) with 15 min chlorination at 50 mg/L chlorine.
Table 1 Element analysis of LPS by EDS. Element
C O Na Mg P S Cl Ca Totals a
4 h of chlorinationa
Control Weight (%)
Atomic (%)
Weight (%)
Atomic (%)
55.60 35.04 0.92 0.47 5.20 0.71 0.00 2.06 100.00
65.02 30.76 0.56 0.27 2.36 0.31 0.00 0.72
53.56 40.20 0.45 0.27 3.16 0.26 0.79 1.30 100.00
62.21 35.06 0.28 0.15 1.42 0.11 0.31 0.45
0.5 mg/mL LPS react with 50 mg/L free chlorine.
15 min of chlorination and then kept stable (Fig. 4), which is consistent with the rapid decrease of the inflammatory responses at early chlorine treatment (Fig. 2). The microstructure of dialyzed and freeze–dried samples were also observed by scanning electronic microscope. But there is no obvious difference caused by chlorination (Fig. A.1). Most particles are at micrometer scale, which is much larger than normal LPS aggregate. It is unknown whether the dialysis and freeze–drying caused the change of LPS aggregation. 4. Discussion Chlorination is the most common disinfection method to guarantee the safety of the reclaimed water usage, and is often the only
Fig. 4. LPS aggregation size versus contact time at chlorine dose of 50 mg/L (LPS concentration was 0.01 mg/mL). The values are means ± standard deviation (SD). The asterisks represent the aggregate sizes of chlorinated samples are significantly different from that of untreated LPS sample (one way ANOVA, p < 0.05).
affordable disinfection in drinking water treatment of the developing countries due to its low cost, simplicity for measurement and control, and a good residual effect [19,20]. However, it was believed that chlorination has no significant effect on endotoxin activity removal in previous studies which was detected mainly by LAL assay as mentioned above. In this study, we discovered that chlorination can significantly inhibit inflammation inducing ability of LPS, when the endotoxin activity detected by LAL was slightly increased after treatment.
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
The difference between LAL results and inflammatory responses in mice may be due to the diversity of reaction mechanism. LAL is the most widespread method in pyrogen test due to its rapid, sensitive and simple [21]. It is an LPS-induced coagulation reaction with amebocyte lysate, which is initiated by proenzyme factor C binding of LPS and finally visualized with chromogenic substrate [16,17,22]. However, in mammals, LPS can be recognized by the innate immune systems as one of the representative pathogen-associated molecular patterns (PAMPs). LPS-binding protein (LBP), which presents on the surface of airway, combines with LPS through high affinity with Lipid A region and transfers LPS monomers from aggregates to CD14. The formed complex binds to TLR4-MD2 receptor and then induces signal transduction across the cell membrane with final activation of a set of immune protective genes, which results in increase of PMN in the lung [6,23,24]. Thus, the initial LPS binding reactions are different in two systems, which are possibly sensitive to chemical or physical change of LPS. The chlorination on alkyl chain of LPS may affect the LPS-LBP binding, which could be the first reason for reduced inflammation induction. Furthermore, the chlorination should also cause the shrink of aggregate size. The aggregation is governed by non-polar interactions between alkyl chains in lipid A portion. The addition of chlorine atom in alkyl chain can increase the polarity of the molecule, which may lead to smaller aggregates. Since the LPS monomer was reported to have no endotoxin activity [12], which makes us hypothesize that the aggregate size could also be one of the factors affecting endotoxin activity or inflammation inducing activity. LAL assay is obviously not sensitive to the change of aggregate size according to our results, while this size may be crucial to innate immunity. The innate immune system recognizes repetitive arrays of proteins, carbohydrates, and lipids on bacterial surface, which are so called ‘molecular patterns’. But how big the ‘pattern’ can effectively initiate innate immune responses is unclear. It is only known that the immune system evolves to recognize large enough patterns to avoid hypersensitivity to some molecules instead of real pathogens. Therefore, the smaller the LPS aggregates are, the less inflammation may be induced. For LAL assay, decreased aggregate size means more aggregates formed, which could be the cause of increase endotoxin activity. This result is similar to the conclusion of Huang et al. [9], who found that endotoxin activity increased significantly after chlorination in pure Gram-negative bacteria solution, in which we can regard bacterium as a large LPS aggregate and they were break into small aggregates after treatment. It should be emphasized that the risk of LPS in water will be changed by different treatments even when the LPS quantity is not changed in this case. Thus, the assessment method for risk removal of treatment technologies should not only rely on quantity-measurement assay, such as LAL assay. The health effectbased examination is necessary for evaluation. Other evidences of the limitation of LAL assay is from the inconsistency between the endotoxin activity and the induction of cytokine secretion by different natural LPSs from bacteria in aquatic environment [25,26]. They suggested that the LPS of some indigenous bacteria from aquatic environment may have stronger inflammation inducing ability than that of E. coli at the same levels of endotoxin activity. Therefore, inflammatory potency test was also recommended for measure the potential risks caused by LPS. Final thing is that chlorination cannot completely inhibit the inflammation inducing ability of LPS. This is due to the weak oxidative potency of chlorine, and it was demonstrated by the low chlorinated site number on LPS and low chlorine consumption during chlorination. As most of the disinfection technologies cannot completely remove LPS [13], ultrafiltration may be the most effective way to control the endotoxic risk in reclaimed water.
135
5. Conclusions In this study, our results reveal that chlorination is effective to reduce the acute inflammation inducing capacity of LPS in lung by adding chlorine to the alky chain of LPS molecule and decreasing LPS aggregate size. Endotoxin activity detected by LAL assay is not correlated with the inflammation induction capacity of chlorinated LPS, and it is recommended to use health effect-based methods to evaluate risk removal of water treatment technologies. Acknowledgement This study was funded by the National High Technology Research & Development Program of China (863 Program, 2013AA065205). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat.2015.10. 024. References [1] D. Chang, Z. Ma, Wastewater reclamation and reuse in Beijing: influence factors and policy implications, Desalination 297 (2012) 72–78. [2] S. Toze, Reuse of effluent water—benefits and risks, Agric. Water Manage. 80 (1) (2006) 147–159. [3] J. Rapala, K. Lahti, L.A. Räsänen, A.L. Esala, S.I. Niemelä, K. Sivonen, Endotoxins associated with cyanobacteria and their removal during drinking water treatment, Water Res. 36 (10) (2002) 2627–2635. [4] W.B. Anderson, C.I. Mayfield, D.G. Dixon, P.M. Huck, Endotoxin inactivation by selected drinking water treatment oxidants, Water Res. 37 (19) (2003) 4553–4560. [5] W.B. Anderson, R.M. Slawson, C.I. Mayfield, Review/synthèse a review of drinking-water-associated endotoxin, including potential routes of human exposure, Can. J. Microbiol. 48 (7) (2002) 567–587. [6] V. Liebers, M. Raulf-Heimsoth, T. Brüning, Health effects due to endotoxin inhalation (review), Arch. Toxicol. 82 (4) (2008) 203–210. [7] O. Michel, Systemic and local airways inflammatory response to endotoxin, Toxicology 152 (1) (2000) 25–30. [8] W.B. Anderson, P.M. Huck, D.G. Dixon, C.I. Mayfield, Endotoxin inactivation in water by using medium-pressure UV lamps, Appl. Environ. Microbiol. 69 (5) (2003) 3002–3004. [9] H. Huang, Q.Y. Wu, Y. Yang, H.Y. Hu, Effect of chlorination on endotoxin activities in secondary sewage effluent and typical Gram-negative bacteria, Water Res. 45 (16) (2011) 4751–4757. [10] D. Petsch, F.B. Anspach, Endotoxin removal from protein solutions, J. Biotechnol. 76 (2) (2000) 97–119. [11] M.B. Gorbet, M.V. Sefton, Endotoxin the uninvited guest, Biomaterials 26 (34) (2005) 6811–6817. [12] M. Mueller, B. Lindner, S. Kusumoto, K. Fukase, A.B. Schromm, U. Seydel, Aggregates are the biologically active units of endotoxin, J. Biol. Chem. 279 (25) (2004) 26307–26313. [13] H. Huang, H.Y. Hu, Q.Y. Wu, Review on concentration of endotoxin in water and its removal effect in water treatment processes, J. Environ. Hyg. 3 (3) (2013) 273–277. [14] M. Montemayor, A. Costan, F. Lucena, J. Jofre, J. Munoz, E. Dalmau, L. Sala, The combined performance of UV light and chlorine during reclaimed water disinfection, Water Sci. Technol. 57 (6) (2008) 935–940. [15] D. Ma, B. Gao, D. Hou, Y. Wang, Q. Yue, Q. Li, Evaluation of a submerged membrane bioreactor (SMBR) coupled with chlorine disinfection for municipal wastewater treatment and reuse, Desalination 313 (2013) 134–139. [16] R. Gehr, S.P. Uribe, I.F.D.S. Baptista, B. Mazer, Concentrations of endotoxins in waters around the island of Montreal, and treatment options, Water Qual. Res. J. Can. 43 (4) (2008) 291–303. [17] Z. Can, L. Wenjun, S. Wen, Z. Minglu, Q. Lingjia, L. Cuiping, T. Fang, Endotoxin contamination and control in surface water sources and a drinking water treatment plant in Beijing, China, Water Res. 47 (11) (2013) 3591–3599. [18] U. Zähringer, S. Ittig, B. Lindner, H. Moll, U. Schombel, N. Gisch, G.R. Cornelis, NMR-based Structural analysis of the complete rough-type lipopolysaccharide isolated from Capnocytophaga canimorsus, J. Biol. Chem. 289 (34) (2014) 23963–23976. [19] H. Galal-Gorchev, Chlorine in water disinfection, Pure Appl. Chem. 68 (9) (1996) 1731–1735. [20] D.J. Nozaic, Chlorine: is it really so bad and what are the alternatives? Water SA 30 (5) (2004) 18–24.
136
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
[21] K. Takayama, N. Qureshi, C.R. Raetz, E. Ribi, J. Peterson, J.L. Cantrell, A.G. Johnson, Influence of fine structure of lipid A on Limulus amebocyte lysate clotting and toxic activities, Infect. Immun. 45 (2) (1984) 350–355. [22] J.L. Ding, B. Ho, A new era in pyrogen testing, Trends Biotechnol. 19 (8) (2001) 277–281. [23] S.I. Miller, R.K. Ernst, M.W. Bader, LPS, TLR4 and infectious disease diversity, Nat. Rev. Microbiol. 3 (1) (2005) 36–46. [24] S.Y. Seong, P. Matzinger, Hydrophobicity: an ancient damage-associated molecular pattern that initiates innate immune responses, Nat. Rev. Immunol. 4 (6) (2004) 469–478.
[25] Y. Ohkouchi, S. Tajima, M. Nomura, S. Itoh, Inflammatory responses and potencies of various lipopolysaccharides from bacteria and cyanobacteria in aquatic environments and water supply systems, Toxicon 97 (2015) 23–31. [26] Y. Ohkouchi, S. Tajima, M. Nomura, et al., Comparison of inflammatory responses in human cells caused by lipopolysaccharides from E. coli and from indigenous bacteria in aquatic environment, J. Environ. Sci. Health A 47 (13) (2012) 1966–1974.
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Inhibition of lipopolysaccharide induced acute inflammation in lung by chlorination Jinshan Zhang a , Jinling Xue a , Bi Xu a , Jiani Xie a , Juan Qiao b,∗ , Yun Lu a,∗ a b
Environmental Simulation and Pollution Control State Key Joint Laboratory, School of Environment, Tsinghua University, Beijing 100084, China Department of Chemistry, Tsinghua University, Beijing 100084, China
h i g h l i g h t s • • • •
Chlorination is effective to reduce the inflammation inducing capacity of LPS in lung. LAL-detected endotoxin activity is not correlated to the potency of inflammation induction. Alkyl chain of LPS was chlorinated in chlorination process. LPS aggregate size decreases after chlorination.
a r t i c l e
i n f o
Article history: Received 17 June 2015 Received in revised form 25 September 2015 Accepted 12 October 2015 Available online 22 October 2015 Keywords: Lipopolysaccharide Chlorine Inhalation exposure Aggregate size Reclaimed water
a b s t r a c t Lipopolysaccharide (LPS, also called endotoxin) is a pro-inflammatory constituent of gram negative bacteria and cyanobacteria, which causes a potential health risk in the process of routine urban application of reclaimed water, such as car wash, irrigation, scenic water refilling, etc. Previous studies indicated that the common disinfection treatment, chlorination, has little effect on endotoxin activity removal measured by Limulus amebocyte lysate (LAL) assay. However, in this study, significant decrease of acute inflammatory effects was observed in mouse lung, while LAL assay still presented a moderate increase of endotoxin activity. To explore the possible mechanisms, the nuclear magnetic resonance (NMR) results showed the chlorination happened in alkyl chain of LPS molecules, which could affect the interaction between LPS and LPS-binding protein. Also the size of LPS aggregates was found to drop significantly after treatment, which could be another results of chlorination caused polarity change. In conclusion, our observation demonstrated that chlorination is effective to reduce the LPS induced inflammation in lung, and it is recommended to use health effect-based methods to assess risk removal of water treatment technologies. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Wastewater reclamation and reuse is important in sustainable development and water resource conservation strategy. The ratio of wastewater reuse is increasing rapidly and it has been used for flushing toilet, car wash, irrigation, refilling scenic water body, etc, which are closely related to human health [1]. However, there are many risk factors identified from reclaimed water contact, in which the microbial pathogens and their derivatives are the most commonly discussed [2]. Lipopolysaccharide (LPS) as a pro-inflammatory constituent of gram negative bacteria and some
∗ Corresponding authors. E-mail addresses: [email protected] (J. Qiao), [email protected] (Y. Lu). http://dx.doi.org/10.1016/j.jhazmat.2015.10.024 0304-3894/© 2015 Elsevier B.V. All rights reserved.
cyanobacteria, is released by cell multiplication, death, and lysis. It is relatively heat stable and similar in structure regardless of source [3,4], which can cause serious health issues including fever, inflammation, asthma, diarrhea, shortness of breath, intravascular coagulation and even death [5–7]. LPS is composed of lipid A, core oligosaccharide and O antigens. The toxicity of LPS is associated with the structure of lipid A whereas the immunogenicity is dependent on polysaccharide component [6,8,9]. The size of LPS monomer is ranging from 10 to 20 kDa, and it can form aggregates with high stability like micelle or vesicle with the size over 1000 kDa due to their hydrophobic interaction in aquatic systems [10,11]. It is known that only aggregates of LPS are biologically active [12]. The exposure routes to LPS include inhalation, intravenous injection and oral intake. In the urban use of reclaimed water, inhalation of aerosolized water droplets is the main exposure. Endotoxin activity in the secondary
132
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
effluent from a multi-country investigation report ranges from 201 to 20100 EU/mL, and the average level is 2994 EU/mL. Although the average endotoxin activity reduced to 1584 EU/mL after advanced treatments, the endotoxin activity still ranges 3–19700 EU/mL in the reclaimed water [13]. In order to minimize the health risk of exposure to reclaimed water containing LPS, microbial pathogens and other chemical compounds, the tertiary treatments are crucial to the safety of wastewater reuse [14], such as chlorination, ozone disinfection, UV irradiation, ultrafiltration and so on. Chlorination is one of the most widely used disinfection methods in wastewater reclamation to guarantee water quality due to its easiness to operate and low cost [15]. Endotoxin inactivation by chlorine had been studied previously. Anderson et al. [4] examined chlorination effect on tap water spiked with LPS standard (Escherichia coli strain O55:B5) and found that chlorine can hardly remove the endotoxin activity of LPS with the maximum inactivation rate of 1.4(EU)/mLh at residual chlorine from 2 mg/L up to 100 mg/L with reaction time more than 100 h. Rapala et al. [3] indicated that chlorination had no effect on endotoxin activity by studying endotoxin removal during drinking water treatment. Gehr et al. [16] found that endotoxin activity in low endotoxin-containing water samples (7–10 EU/mL) can both be moderately decreased or increased by chlorination. Can et al. [17] investigated endotoxin activity following different treatment process steps and indicated that endotoxin increased after chlorination, these were caused by bacteria lysis and endotoxin release. Similar results were also found in secondary effluent [9]. However, all these studied used LAL assay to assess treatment effects. One thing, which is worth of attention, is that the validity of using the in vitro LAL assay for assessing the inactivation efficiency of LPS has not been verified, since chemical treatment of LPS can cause changes at molecular level. In this study, we measured the chlorinated LPS activities by LAL assay and a mouse inhalation exposure model simultaneously. LAL assay showed moderate increase of endotoxin activity, while lung inflammation was significantly reduced by chlorination, which indicates that chlorination can reduce the health effect caused by LPS inhalation and LAL is not suitable to evaluate the treatment efficiency. The mechanism of LPS inactivation by chlorine was also investigated. 2. Materials and methods 2.1. Animals The study protocol was approved by the Institutional Animal Care and Use Committee of Tsinghua University and was performed according to the recommendations in the Guide for the Care and Use of Laboratory Animals. We made our best efforts to minimize suffering and the number of experimental animals. Male ICR mice aged 8–9 weeks were purchased from Vital River company (Beijing, China) and kept in a barriered facility for experimental animals in Tsinghua University. 2.2. Inhalation exposure HOPE-med 8050 dynamic inhalation system (HOPE Ltd, Tianjin, China) was used to provide an exposure chamber with a constant concentration of water aerosol flow. The aerosol concentration versus samples fluid rate was calibrated by an atmospheric sampler (Casella, UK) with drying columns. The water pumping rate was set to 0.7 mL/min, and the water aerosol concentration in the chamber was about 9 mg/L. Other parameters are as follows: carrier gas flow velocity 3.00 m3 /h, temperature 25 ◦ C, and exposure time 6 h.
2.3. Bronchoalveolar lavage fluid (BALF) and cells counting Three hours after inhalation exposure, mice were sacrificed by cervical dislocation. In our previous study, three-hour has been found to be a time point for a peak of cellular response induced by LPS. BALF was collected by cannulating the upper part of trachea using syringe in 2 mL PBS (pH7.4) containing 0.5% bovine serum albumin. 300 L BALF was centrifuged at 800 rcf for 5 min at 4 ◦ C, and resuspended in 1.5 mL dilution reagent (supplementary reagent of animal hematology analyzer, Mindray, China) for the concentration of white blood cells (WBC) analysis by animal hematology analyzer (Mindray, China). The rest BALF was also centrifuged, and resuspended again in 10 L PBS containing 0.5% BSA for smear. For smear, every 5 L sample was spread on a slide (Shitai Ltd., China), dried in the hood, and then stained by Wright–Giemsa method. Briefly, the slides were soaked into methanol for 15 min and dried in the air. Then they were stained by Wright solution for 3 min, Wright/PBS (PBS pH 6.6, volume ratio 1:1) solution for another 3 min, rinsed thoroughly with PBS (pH 6.6). Next, the slides were stained by Giemsa solution for 10 min, rinsed thoroughly by deionized water. Finally, the slides were dried and mounted for microscopy (Leica, Germany). 300 to 600 white cells were counted for each sample to calculate the proportion of polymorphonuclear neutrophils (PMNs).
2.4. LAL assay Standard Limulus Amebocyte Lysate assay kits were purchased from Xiamen Houshiji Ltd., China, and performed according to the manufacturer’s instructions.
2.5. Chlorination of LPS LPS (L2637, L2880) purchased from Sigma–Aldrich was extracted from E. coli serotype O55:B55, and dissolved in pyrogenfree water at high concentration, and stored at 4 ◦ C for further use. Synthetic LPS containing water was prepared by spiking LPS stock solution to make the concentration about 2000 EU/mL, and was adjusted to pH 7.0 by adding phosphate buffer. Chlorination was carried out by adding sodium hypochlorite stock solution (15 g/L as available chlorine) except for the chlorine-free control and concentration of available chlorine was detected by chlorineSpectrophotometric method (HANNA, Italy) before use. At the end of the pre-determined reaction time, excessive sodium thiosulfate (2 times the maximal consumption of chlorine) was added to neutralize residual chlorine, the whole process was performed in the dark. The chlorine dosage ranged from 0 to 50 mg/L and the reaction time ranged from 5 to 240 min were selected. Then, the water samples were ready for LAL assay and inhalation exposure experiments. The chlorination and quenching process did not change the pH, which was maintained at 7.17 ± 0.02. 2.6. Scanning electron microscopy (SEM) and 1 H NMR spectroscopy The concentrations of LPS in synthetic water samples were 0.5 mg/mL for SEM and 1 mg/mL for NMR spectroscopy. Two samples before and after chlorination (50 mg/L free chlorine, reaction time was 4 h for SEM and 15 min for NMR) were prepared and dialyzed to deionized water by 7KD dialysis cassette (Pierce Biotechnology, USA). After dialysis, the samples were freeze-dried for SEM (Hitachi, Japan). For NMR spectroscopy, 2 mg freeze-dried samples were dissolved in 0.5 mL D2 O, and performed in the Ana-
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
133
lytical Center of Tsinghua University, the NMR model is ECA-600, JEOL Ltd. Japan. 2.7. Aggregate size determination The concentration of LPS in synthetic water samples used for aggregate size determination was 0.01 mg/mL (81000 EU/mL detected by LAL). Aggregate size was examined by Nano-particle size and Zeta potential analyzer (Beckman Coulter, USA).
3. Results 3.1. Endotoxin activity and inflammatory effects of chlorinated LPS As mentioned above, the removal of endotoxin activity by chlorination had some inconsistency in the literatures [3,4,9,16,17]. In this study, the endotoxin activity of LPS cannot be inactivated by chlorine from the LAL assay results (Fig. 1a). Instead, chlorination increased endotoxin activity moderately with all tested chlorine concentration. However, when tested in mouse inhalation model, consistently significant decrease of about 60% of inflammatory effect was observed in BALF by counting the percentage of PMNs (Fig. 1b). It is interesting that the decrease of inflammation stopped at a certain level regardless of quantity of chlorine, and only less than 5% of chlorine was consumed after 4 h treatment at 20 mg/L and 50 mg/L levels. This result suggests that chlorination can reduce the inflammation inducing ability of LPS, but cannot remove it completely. The possible effects of disinfection reagents were also tested (control C1), which was proved to be negligible (Fig. 1b). The concentration of WBC in BALF also decreased after chlorination, which further demonstrated the inhibition of inflammation induction (Fig. 1c). When it was confirmed that chlorination can inhibit the inflammation inducing ability of LPS, the next important question is that whether the LPS can be inactivated at common chlorination condition, since 4 h are usually much longer than the real treatment time. It was found that the inflammatory response decreased rapidly at the first 15 min, and reached a critical level at 60 min. And this pattern is exactly opposite to the results of LAL assay (Fig. 2a). The WBC concentration in BALF also dropped starting from the reaction time of 5 min, and no obvious further decrease was observed with longer treatment (Fig. 2b). 3.2. Chemical composition of LPS after chlorination It is well known that chlorine treatment generates chlorinated disinfection products. It should be one of the main reasons that affect the inflammation induction. First, infrared spectrum was taken from the freeze–dried samples, but no obvious peak change was observed (data not shown). Energy dispersive spectrometer was used to analyze the chemical composition of dialyzed and freeze-dried samples. And the atom ratio between carbon and chlorine is about 200:1 after 4 h reaction (Table 1), which means the chlorination happened but only in a few sensitive sites of LPS. To figure out the positions of these sensitive sites, 1 H NMR spectroscopy was performed. There are three main groups of hydrogen peaks (Fig. 3). According to the analysis of a previous report [18], the hydrogens between 1 and 2 ppm should be on the alkyl chains, while the hydrogens with chemical shift ranged from 3.3 to 4.5 ppm should be on saccharide rings. If the peak area was normalized by the group 1, both the hydrogen quantities of group 2 and 3 were decreased, which means both hydrogens on alkyl chains and saccharide rings were replaced by chlorines. The chlorination on alkyl
Fig. 1. Chlorination inhibits inflammation inducing capacity of LPS. (a) LAL results of endotoxin activity after 4 h treatment of different doses of free chlorine, 10, 20, 30, and 50 mg/L, respectively. (b) The percentage of PMN in total cells of BALF. Mice were exposed for 6 h and sacrificed 3 h after exposure. (c) Total WBC concentrations in BALF. (C1, ultrapure water chlorination by 50 mg/L chlorine without additional LPS; C2, unexposed; n = 5 mice/group). The values are means ± standard deviation (SD). The asterisks represent endotoxin activity and white blood cells of chlorinated samples are significantly different from that of untreated LPS exposed group (one way ANOVA, p < 0.05).
chains can affect its binding activity to LPS-binding protein (LBP) and its ability to form aggregates. 3.3. Decrease of LPS aggregate size after chlorination Since the alky chain was chlorinated as mentioned ahead, the aggregation of LPS monomers may also change. Therefore, the average aggregate size was tested before and after chlorination. It was found that the aggregate size rapidly decreased within the first
134
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
Fig. 2. Chlorination effects occurs at early treatment stage. (a) LAL result of LPS level, PMN percentage and (b) relevant WBC concentration in BALF versus contact time at the presence of 20 mg/L free chlorine. The values are means ± standard deviation (SD). The asterisks represent endotoxin activity and white blood cells of chlorinated samples are significantly different from that of untreated LPS exposed group (one way ANOVA, p < 0.05) (n = 5 mice/group).
Fig. 3. One dimensional 1 H spectroscopy with integration values of peak areas by making the peak at chemical shift of 1.08 ppm as a reference. (a) LPS before chlorination and (b) with 15 min chlorination at 50 mg/L chlorine.
Table 1 Element analysis of LPS by EDS. Element
C O Na Mg P S Cl Ca Totals a
4 h of chlorinationa
Control Weight (%)
Atomic (%)
Weight (%)
Atomic (%)
55.60 35.04 0.92 0.47 5.20 0.71 0.00 2.06 100.00
65.02 30.76 0.56 0.27 2.36 0.31 0.00 0.72
53.56 40.20 0.45 0.27 3.16 0.26 0.79 1.30 100.00
62.21 35.06 0.28 0.15 1.42 0.11 0.31 0.45
0.5 mg/mL LPS react with 50 mg/L free chlorine.
15 min of chlorination and then kept stable (Fig. 4), which is consistent with the rapid decrease of the inflammatory responses at early chlorine treatment (Fig. 2). The microstructure of dialyzed and freeze–dried samples were also observed by scanning electronic microscope. But there is no obvious difference caused by chlorination (Fig. A.1). Most particles are at micrometer scale, which is much larger than normal LPS aggregate. It is unknown whether the dialysis and freeze–drying caused the change of LPS aggregation. 4. Discussion Chlorination is the most common disinfection method to guarantee the safety of the reclaimed water usage, and is often the only
Fig. 4. LPS aggregation size versus contact time at chlorine dose of 50 mg/L (LPS concentration was 0.01 mg/mL). The values are means ± standard deviation (SD). The asterisks represent the aggregate sizes of chlorinated samples are significantly different from that of untreated LPS sample (one way ANOVA, p < 0.05).
affordable disinfection in drinking water treatment of the developing countries due to its low cost, simplicity for measurement and control, and a good residual effect [19,20]. However, it was believed that chlorination has no significant effect on endotoxin activity removal in previous studies which was detected mainly by LAL assay as mentioned above. In this study, we discovered that chlorination can significantly inhibit inflammation inducing ability of LPS, when the endotoxin activity detected by LAL was slightly increased after treatment.
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
The difference between LAL results and inflammatory responses in mice may be due to the diversity of reaction mechanism. LAL is the most widespread method in pyrogen test due to its rapid, sensitive and simple [21]. It is an LPS-induced coagulation reaction with amebocyte lysate, which is initiated by proenzyme factor C binding of LPS and finally visualized with chromogenic substrate [16,17,22]. However, in mammals, LPS can be recognized by the innate immune systems as one of the representative pathogen-associated molecular patterns (PAMPs). LPS-binding protein (LBP), which presents on the surface of airway, combines with LPS through high affinity with Lipid A region and transfers LPS monomers from aggregates to CD14. The formed complex binds to TLR4-MD2 receptor and then induces signal transduction across the cell membrane with final activation of a set of immune protective genes, which results in increase of PMN in the lung [6,23,24]. Thus, the initial LPS binding reactions are different in two systems, which are possibly sensitive to chemical or physical change of LPS. The chlorination on alkyl chain of LPS may affect the LPS-LBP binding, which could be the first reason for reduced inflammation induction. Furthermore, the chlorination should also cause the shrink of aggregate size. The aggregation is governed by non-polar interactions between alkyl chains in lipid A portion. The addition of chlorine atom in alkyl chain can increase the polarity of the molecule, which may lead to smaller aggregates. Since the LPS monomer was reported to have no endotoxin activity [12], which makes us hypothesize that the aggregate size could also be one of the factors affecting endotoxin activity or inflammation inducing activity. LAL assay is obviously not sensitive to the change of aggregate size according to our results, while this size may be crucial to innate immunity. The innate immune system recognizes repetitive arrays of proteins, carbohydrates, and lipids on bacterial surface, which are so called ‘molecular patterns’. But how big the ‘pattern’ can effectively initiate innate immune responses is unclear. It is only known that the immune system evolves to recognize large enough patterns to avoid hypersensitivity to some molecules instead of real pathogens. Therefore, the smaller the LPS aggregates are, the less inflammation may be induced. For LAL assay, decreased aggregate size means more aggregates formed, which could be the cause of increase endotoxin activity. This result is similar to the conclusion of Huang et al. [9], who found that endotoxin activity increased significantly after chlorination in pure Gram-negative bacteria solution, in which we can regard bacterium as a large LPS aggregate and they were break into small aggregates after treatment. It should be emphasized that the risk of LPS in water will be changed by different treatments even when the LPS quantity is not changed in this case. Thus, the assessment method for risk removal of treatment technologies should not only rely on quantity-measurement assay, such as LAL assay. The health effectbased examination is necessary for evaluation. Other evidences of the limitation of LAL assay is from the inconsistency between the endotoxin activity and the induction of cytokine secretion by different natural LPSs from bacteria in aquatic environment [25,26]. They suggested that the LPS of some indigenous bacteria from aquatic environment may have stronger inflammation inducing ability than that of E. coli at the same levels of endotoxin activity. Therefore, inflammatory potency test was also recommended for measure the potential risks caused by LPS. Final thing is that chlorination cannot completely inhibit the inflammation inducing ability of LPS. This is due to the weak oxidative potency of chlorine, and it was demonstrated by the low chlorinated site number on LPS and low chlorine consumption during chlorination. As most of the disinfection technologies cannot completely remove LPS [13], ultrafiltration may be the most effective way to control the endotoxic risk in reclaimed water.
135
5. Conclusions In this study, our results reveal that chlorination is effective to reduce the acute inflammation inducing capacity of LPS in lung by adding chlorine to the alky chain of LPS molecule and decreasing LPS aggregate size. Endotoxin activity detected by LAL assay is not correlated with the inflammation induction capacity of chlorinated LPS, and it is recommended to use health effect-based methods to evaluate risk removal of water treatment technologies. Acknowledgement This study was funded by the National High Technology Research & Development Program of China (863 Program, 2013AA065205). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat.2015.10. 024. References [1] D. Chang, Z. Ma, Wastewater reclamation and reuse in Beijing: influence factors and policy implications, Desalination 297 (2012) 72–78. [2] S. Toze, Reuse of effluent water—benefits and risks, Agric. Water Manage. 80 (1) (2006) 147–159. [3] J. Rapala, K. Lahti, L.A. Räsänen, A.L. Esala, S.I. Niemelä, K. Sivonen, Endotoxins associated with cyanobacteria and their removal during drinking water treatment, Water Res. 36 (10) (2002) 2627–2635. [4] W.B. Anderson, C.I. Mayfield, D.G. Dixon, P.M. Huck, Endotoxin inactivation by selected drinking water treatment oxidants, Water Res. 37 (19) (2003) 4553–4560. [5] W.B. Anderson, R.M. Slawson, C.I. Mayfield, Review/synthèse a review of drinking-water-associated endotoxin, including potential routes of human exposure, Can. J. Microbiol. 48 (7) (2002) 567–587. [6] V. Liebers, M. Raulf-Heimsoth, T. Brüning, Health effects due to endotoxin inhalation (review), Arch. Toxicol. 82 (4) (2008) 203–210. [7] O. Michel, Systemic and local airways inflammatory response to endotoxin, Toxicology 152 (1) (2000) 25–30. [8] W.B. Anderson, P.M. Huck, D.G. Dixon, C.I. Mayfield, Endotoxin inactivation in water by using medium-pressure UV lamps, Appl. Environ. Microbiol. 69 (5) (2003) 3002–3004. [9] H. Huang, Q.Y. Wu, Y. Yang, H.Y. Hu, Effect of chlorination on endotoxin activities in secondary sewage effluent and typical Gram-negative bacteria, Water Res. 45 (16) (2011) 4751–4757. [10] D. Petsch, F.B. Anspach, Endotoxin removal from protein solutions, J. Biotechnol. 76 (2) (2000) 97–119. [11] M.B. Gorbet, M.V. Sefton, Endotoxin the uninvited guest, Biomaterials 26 (34) (2005) 6811–6817. [12] M. Mueller, B. Lindner, S. Kusumoto, K. Fukase, A.B. Schromm, U. Seydel, Aggregates are the biologically active units of endotoxin, J. Biol. Chem. 279 (25) (2004) 26307–26313. [13] H. Huang, H.Y. Hu, Q.Y. Wu, Review on concentration of endotoxin in water and its removal effect in water treatment processes, J. Environ. Hyg. 3 (3) (2013) 273–277. [14] M. Montemayor, A. Costan, F. Lucena, J. Jofre, J. Munoz, E. Dalmau, L. Sala, The combined performance of UV light and chlorine during reclaimed water disinfection, Water Sci. Technol. 57 (6) (2008) 935–940. [15] D. Ma, B. Gao, D. Hou, Y. Wang, Q. Yue, Q. Li, Evaluation of a submerged membrane bioreactor (SMBR) coupled with chlorine disinfection for municipal wastewater treatment and reuse, Desalination 313 (2013) 134–139. [16] R. Gehr, S.P. Uribe, I.F.D.S. Baptista, B. Mazer, Concentrations of endotoxins in waters around the island of Montreal, and treatment options, Water Qual. Res. J. Can. 43 (4) (2008) 291–303. [17] Z. Can, L. Wenjun, S. Wen, Z. Minglu, Q. Lingjia, L. Cuiping, T. Fang, Endotoxin contamination and control in surface water sources and a drinking water treatment plant in Beijing, China, Water Res. 47 (11) (2013) 3591–3599. [18] U. Zähringer, S. Ittig, B. Lindner, H. Moll, U. Schombel, N. Gisch, G.R. Cornelis, NMR-based Structural analysis of the complete rough-type lipopolysaccharide isolated from Capnocytophaga canimorsus, J. Biol. Chem. 289 (34) (2014) 23963–23976. [19] H. Galal-Gorchev, Chlorine in water disinfection, Pure Appl. Chem. 68 (9) (1996) 1731–1735. [20] D.J. Nozaic, Chlorine: is it really so bad and what are the alternatives? Water SA 30 (5) (2004) 18–24.
136
J. Zhang et al. / Journal of Hazardous Materials 303 (2016) 131–136
[21] K. Takayama, N. Qureshi, C.R. Raetz, E. Ribi, J. Peterson, J.L. Cantrell, A.G. Johnson, Influence of fine structure of lipid A on Limulus amebocyte lysate clotting and toxic activities, Infect. Immun. 45 (2) (1984) 350–355. [22] J.L. Ding, B. Ho, A new era in pyrogen testing, Trends Biotechnol. 19 (8) (2001) 277–281. [23] S.I. Miller, R.K. Ernst, M.W. Bader, LPS, TLR4 and infectious disease diversity, Nat. Rev. Microbiol. 3 (1) (2005) 36–46. [24] S.Y. Seong, P. Matzinger, Hydrophobicity: an ancient damage-associated molecular pattern that initiates innate immune responses, Nat. Rev. Immunol. 4 (6) (2004) 469–478.
[25] Y. Ohkouchi, S. Tajima, M. Nomura, S. Itoh, Inflammatory responses and potencies of various lipopolysaccharides from bacteria and cyanobacteria in aquatic environments and water supply systems, Toxicon 97 (2015) 23–31. [26] Y. Ohkouchi, S. Tajima, M. Nomura, et al., Comparison of inflammatory responses in human cells caused by lipopolysaccharides from E. coli and from indigenous bacteria in aquatic environment, J. Environ. Sci. Health A 47 (13) (2012) 1966–1974.