Wild rats as urban detectives for latent sources of asbestos contamination

Based on a large body of evidence asbestos minerals have been classified as carcinogens. Despite the Italian ban on asbestos in 1992 and the subsequent remediation activities, latent sources of contamination may still represent a hazard where asbestos were particularly used. Using wild rats as sentinel animals, this study aimed at uncovering sites with the greatest potential for non-occupational exposure to asbestos in the city of Casale Monferrato (Piedmont Region, Italy), where the largest Italian manufacturing plant of asbestos-cement had been active. During the study period (2013-2015) a total of 40 wild rats were captured from 16 sampling capture points. The lungs of wild rats have been investigated by using scanning electron microscopy (SEM) with energy dispersive spectroscopy (EDS). The SEM-EDS detected the presence of asbestos fibers (tremolite/actinolite, amosite, and chrysotile) in rats' lungs from 11 sampling points. The hypothetical rats' home-range and the observed site-specific concentration of asbestos fibers per gram of dry lung tissue were used to identify areas to be targeted by additional search of latent sources of asbestos. In conclusion, our results showed that the use of wild rats as sentinel animals may effectively integrate the strategies currently in use to reduce the exposure to asbestos.


INTRODUCTION 28 29
According to international laws 'asbestos' is a set of six asbestiform silicates, five belonging to the 30 mineralogical group of amphiboles (actinolite, amosite or grunerite asbestos, anthophyllite, 31 crocidolite and tremolite) and one to serpentine group (chrysotile) (Dir. 2009/148/EC). 32 In the past, nearly all over the world, asbestos have been widely employed in many composite 33 a) A regular grid of squares of 200 meters per side was placed on the map of the city (Fig. 1,  96 OpenStreetMap® cartographic image processed with QGIS System). The length of the side of the 97 squares was defined considering the rat behaviour and the home range usually covered around a 98 permanent den (Gardner-Santana et al., 2009). Initially the priority was given to areas with buildings 99 built in 60's and 70's ("60/70 area"), perimeter area located S and SW of the old town, i.e. dating 100 back to the period of greater asbestos production and high densities of population. A control location, 101 to check the ability of rat capturing of the research team, has been added in a suburban area (~2.5 km 102 from Casale Monferrato) where the presence of rats had been reported by private citizens. 103 b) Multiple capture traps (Ekomille, Ekommerce SRL, Atessa, Italy) were selected as sampling tools. 104 These devices allow multiple and continuous catches, until 10-12 wild rats are captured discouraging 105 other animal species to enter the trap. Traps were used following the instructions of the construction 106 companies using the appropriate personal protective equipment. It was planned that the available 107 traps (1 to 3) should have been placed as close as possible to the centre of squares for about 3 weeks 108 or less if the number of captured rats had been satisfactory (initially a target of 5 rats per trap had 109 been fixed). After the maximum 3-week period the traps were moved to other squares in the grid. c) The rats had to be individually subjected to necropsy and histopathology. Later, the lung material 114 at each trapping sites was to be pooled and analysed, keeping separate the material from rats weighing 115 100 g or more (heavy rats: H) from that of smaller rats (light rats: L), regardless of sex or species. 116 The "100 g" cut-off was then used as a proxy for age at exposure to asbestos fibres: the assumption 117 was that younger animals (body-weight < 100 g), given their shorter lifespan, would have had fewer 118 opportunities to inhale and accumulate asbestos. 119 After the first year of sampling, the number of sampling points with successful captures and the 120 number of captured rats were much lower than expected. To improve capture performance, advice 121 from rodent control companies was obtained and an enhanced cooperation was searched from citizens 122 reporting rats. Therefore, the revised procedures about the traps' management were: site selection 123 based on information from residents, new types of bait, multiple traps per sampling point. Additional 124 30 individual-capture traps (hereinafter called as "snap traps") were incorporated into the sampling.

Necropsy 131
The captured rats were sexed and weighed. A necropsy was performed according to a standardised 132 protocol in order to detect the presence of any lesions referable to zoonotic diseases (as post-capture 133 safeguarding of the health of the research team) and to collect tissue portions for further examination. 134 Spleen, liver and kidneys were collected and tested for Francisella tularensis and Leptospira spp. by 135 PCR (Forsman et al., 1994;…). Lungs were sampled and fixed in 4% buffered formaldehyde 136 solution. The left lung was subjected to histopathology whereas the right lung was subjected to 137 electron microscopy. Heart, spleen, liver and kidneys were also sampled for histopathological 138 examination to evaluate the general health status of the animals. 139 Optica Milano S.p.A, Milan, Italy), in order to easily detect, respectively, the presence of ferruginous 152 corpuscles due to asbestos and the presence and the extent of pulmonary fibrosis. 153 The Perls staining is suitable to stain in blue iron-ferric on tissue sections but is not able to stain the 154 iron-ferrous and the iron bound to haemoglobin, ferritin and pigments due to use of acid formalin. As 155 positive control for the Perls staining was used a sample of lung tissue from rats subjected to 156 inhalation of amosite. The result of Perls staining was expressed as positive or negative. 157 The Masson staining allows the visualization of muscular fibres in red, collagenous tissue in blue and 158 erythrocytes in yellow. An ad hoc grading score was used for the evaluation of presence/absence of 159 pulmonary fibrosis and its extent, performing a semi-quantitative evaluation: Negative (absence of 160 appreciable pulmonary fibrosis), Positive with grade I (minor fibrosis), Positive with grade II

Scanning electron microscopy 164
The pools of lung samples were prepared for scanning electron microscopy with energy dispersive 165 spectrometry (SEM-EDS) and examined to evaluate the presence of asbestos fibres and their 166 concentration according the protocol described by Belluso et al. (2006). By the SEM-EDS it is difficult to distinguish correctly between chrysotile and asbestiform antigorite 177 (a non-asbestos), both species belonging to the serpentine group (Fornero et al., 2009). When it was 178 not possible to classify the fibre as chrysotile sensu stricto, in conservative mode and considering that 179 the final objective is to locate residual sources of asbestos to sanitize the area, for our analysis we 180 assumed that these fibres belonging to the asbestos group chrysotile. Besides, our assumption is 181 supported by the fact that there is no deposit or natural outcrops of these minerals were reported near 182 Casale Monferrato; therefore, it was reasonable to assume that these fibres derive from anthropogenic 183 manufacturing linked to asbestos.

Necropsy and PCR 222
During the necropsy, no macroscopic lesions were found in any captured rat. All PCR analyses 223 conducted on the samples were negative for F. tularensis and Leptospira spp. It was not possible to 224 perform any laboratory analysis on two of the captured rats since, at a first macroscopic evaluation, 225 they were in very poor conditions of conservation. 226 Among the 11 pools positive to asbestos, one contained the 2 rats with histopathological diagnosis 251 for asbestosis; one rat with probable asbestosis (i.e., Perls and Masson stains positive but without 252 identifying ferruginous bodies) was present in each of other 3 pools. 253 The mean concentration of asbestos fibres was 30,136 ff/gdw (n = 11 pools, sd = 27,126 ff/gdw). 254 LAF showed higher concentration (n = 10, mean = 30,600, sd = 26,336 ff/gdw) compared with SAF 255 (n = 5, mean = 5,100, sd = 0 ff/gdw). 256 Regardless of the length of the asbestos fibres, tremolite/actinolite (including tremolite s.s.) was the 257 most common species of asbestos found in the lungs of rats, both in absolute number of fibres (43 258 fibres in 9 pools, 16 rats) and per gram of dry tissue (mean = 24,367 ff/gdw), followed by amosite (9 259 fibres in 2 pools, 7 rats; 22,950 ff/gdw) and chrysotile/asbestiform antigorite (including those 260 properly identified as chrysotile: 9 fibres in 6 pools, 12 rats; 11,050 ff/gdw). 261 The highest asbestos load has been detected in the pools of the sampling points F2_1 (2 pools, 1 H 262 and 1 L, mean = 58,650 ff/gdw) and F2_3 (81,600 ff/gdw) (Fig. 2).
observed an overlap of the respective buffer areas (i.e., hypothetical home-range around the sampling 266 point), consistent with the hypothesis of an asbestos hot-spot in this area (Fig. 3). 267 268

DISCUSSION AND CONCLUSIONS 269
Our study confirms the rat (R. norvegicus and R. rattus) as a helpful sentinel of remaining asbestos; 270 moreover, our results suggest that at least one geographically defined location in the urban area of 271 Casale Monferrato could represent a probable hot-spot. Our approach is based on the observation that 272 the murine lung tissue acts as a filter for mineral fibres, among which the asbestos fibres are 273 distinguishable and can be also classified by species. 274 The heterogeneity of the mineral species found and the marked differences between the 275 concentrations detected make it possible to process the data and search for the causes (sources) of 276 these differences, valid for territorial realities even diverse from Casale Monferrato. 277 Three types of asbestos were detected in the lung tissue of the captured rats: chrysotile (the most used 278 asbestos at the Eternit plant), tremolite/actinolite, and amosite. Interesting to note is the high load of 279 tremolite/actinolite asbestos, which were not commercially used in ACM to a significant extent (e.g., 280 in the preparation of fibre cement), and the absence of crocidolite fibres. 281 The high load of tremolite/actinolite asbestos in the rat lungs could be explained by the following traps, the number of caught rats was less than expected. The reasons for this can be found in the innate 295 ability of rats to warn of the danger of traps, in the periodic rodent control activities carried out in the 296 urban area, and in the low effectiveness of traps. Multiple capture traps allow for long-term 297 monitoring with acceptable management costs (i.e., personnel effort to check the traps) and it is very 298 effective when the spread of rats is epidemic. However, given that periodic rodent control activities 299 have reduced the murine population in the urban area, it is reasonable to assume that the presence of 300 wild rats is widespread but without very large colonies, except in certain areas. 301 In contrast, snap traps appeared to be more effective, but these require higher management costs to 302 ensure daily, or twice-daily, controls. 303 Therefore, the skill and the effectiveness in the capture are crucial to successfully complete the study 304 and were a main stone in the planning of the study. Our experience suggests to considering all the 305 information about the territory, the density of murine population and to place several and different 306 kind of trap devices. For example, we suggest using the snap traps for short intensive periods and the 307 multiple capture traps for long periods, especially with high murine population density. 308 The methodology proposed by this study represents a preliminary screening step that, in situations 309 where the presence of latent sources of asbestos is suspected, significantly reduces the geographic 310 magnitude of the area to be investigated. Given a radius of 100 m around the sampling point, the 311 exact identification of the latent source requires a further and more detailed search. At this stage, 312 sentinel animals may still be useful, e.g., other traps could be placed at a shorter distance and 313 additional research strategies applied in the suspect area, such as the administration to local 314 inhabitants of an ad hoc questionnaire to help in the detection of the putative latent source. 315 Furthermore, the search could be guided by the type of asbestos found in the lungs of locally captured 316 rats, that is, particularly by looking at the ACM that might contain them or the sites where they were 317 known to be located. Finally, to improve the search, the assistance of rodent experts could help to 318 track the footpaths and burrows of rats, which would make it possible to better determinate the 319 rodents' home-range and thus the search area. 320 As conclusion, we can say that the approach presented in this study can be used in the framework of 321 public health campaigns to address asbestos removal activity. 322 323 SEM numbers in bold were positive pools for asbestos fibres. 334 *: two rats have been excluded by the histopathologic and SEM-EDS investigations since they were in poor conditions of conservation at a first 335 macroscopic evaluation. 336 **: one rat has been excluded by the SEM-EDS investigations because the quality of the sample did not allow it. 337 # and § : analysed in the same pool. 338