First report of microplastics in the gastrointestinal tracts of North American insectivorous bats

Collection of samples for protocol testing

We received bat stomach contents collected during previous field studies as part of WNS surveillance. At the time of collection, WNS was recently introduced to NA, and large quantities of dead bats were found in caves. The samples used in this study were collected from three locations in the Northeast USA: Bennington County cave, Vermont (n = 42), Hampden County mine, Massachusetts (n = 25), and Warren County mine, New York, USA (n = 25). We assume that they were all collected from sites by researchers and know that they were all from a single species, Myotis lucifugus (little brown bat). The samples were not collected for the purpose of studying microplastics and there was no protocol in place to limit contamination at time of stomach content collection; therefore, we only used these samples to modify a method of MP extraction and quantification and cannot determine the source of MPs in these samples.

Bat gastrointestinal tract (GIT) collection

We collected bat carcasses through public health monitoring programs in Tennessee, USA. We received dead bats that had been collected throughout the state and submitted to facilities in Knox and Davidson Counties. Bats were often submitted following human or pet exposure and all were missing brain tissue from rabies tests that were performed. None of these bats were tested for the presence of WNS. We chose Eptesicus fuscus (big brown bat) as our study species because we had the highest number of carcasses of that species. Bats were frozen at −20 °C, then thawed for three hours on the day of processing and necropsied in a dedicated necropsy lab at the University of Tennessee College of Veterinary Medicine to extract the full, intact gastrointestinal tracts (GIT) from esophagus to anus. During necropsy, we placed bats on metal pans, necropsied them with small metal scissors and scalpels, and stored all extracted organs in aluminum foil. We did not rinse equipment during necropsy but kept organs intact to avoid contamination of the internal organs. External contamination of organs was addressed in the microplastic lab on the day of digestion (see below). We refroze each sample in labeled whirl pak bags until further analysis. Unlike the collection method of the stomach content samples discussed previously, this GIT collection protocol ensured reduced risk of MP contamination, thus allowing us to test the dietary pathway of exposure.

Extraction of microplastics

We adapted methods previously used to extract MPs from GITs of birds of prey (Carlin et al., 2020). We rinsed each intact GIT sample with pure deionized water (DI water) twice to remove external microplastics. This step was not conducted with the stomach contents samples as that would have washed away all the material. A solution of potassium hydroxide (KOH) was prepared from pellets (Macron Fine Chemicals, Batch No. 0000271176) and deionized water so that the concentration was 10% KOH. The KOH was not filtered so as not to introduce MPs from the container that it would be filtered into in an additional step. Instead, the solution was used in both tissue digestion and the processing of blanks to quantify contamination (see below). We weighed each GIT sample and digested tissues in the KOH solution and placed them on a shaker at 180 revolutions per minute (RPM) speed for 24 h. The amount of KOH used for each sample was determined by using an analytical scale to weigh the GIT sample and aimed to add 3 times the weight in KOH. All GITs were fully submerged in KOH despite the small volumes of solutions used. Samples remained in the KOH solution for 3–12 days depending on tissue digestion success. While Carlin et al. (2020) applied heat to the digestion of GITs of large birds of prey, we found during protocol testing that applying heat during digestion was not required, possibly due to the small sample mass of bat GITs. We used a glass vacuum filter system and glass membrane filters (Whatman, grade GF/F borosilicate glass microfiber filters, 47 mm diameter, 0.42 mm thickness) to separate out undigested from digested material. We originally used filters with a 2.7 µm pore size for stomach content samples but determined that a smaller pore size could be used compared to those used for large birds (Carlin et al., 2020) so we modified the protocol to use 0.7 µm pore size for GIT samples. We stored samples on the glass filters in closed glass petri dishes in a dark filing cabinet until we counted particles in subsequent steps.

Limiting and quantifying contamination

To avoid introducing contamination to samples, all benches were wiped down at the beginning of each session with paper towels and DI water. We wore 100% white cotton lab coats, covered all samples with aluminum foil while processing, and rinsed all glassware with pure DI water. We incorporated two procedural blanks every digestion session in the lab to quantify background contamination following the same procedures that we used to digest tissue samples (rinsing glassware, adding 10% KOH, filtering through vacuum filter, storing filters in petri dishes, and counting and characterizing particles). The procedural blanks only contained the average amount of KOH solution (in grams using an analytical scale) that was added to the previous ten GIT samples and did not contain any tissues. We attempted to limit the number of people in the lab, but it was a shared space, and we could not always control lab use.

Counting and characterizing microplastics

We examined each glass filter under a microscope (Motic, Model SMZ-171) and scanned left-to-right and up-to-down. We scanned on multiple magnification settings ranging 7.5–50X to count particles of various sizes. We placed a dot, made with a fine tipped marker, beside every piece of plastic to avoid counting particles more than once. We characterized each particle type based on the type and color (Carlin et al., 2020; Rodríguez-Seijo & Pereira, 2017). We used a hot needle method to determine if some particles were plastics. First, we would heat the needle with a flame and place it on the edge of a particle. If it melted or bent with heat, we classified it as an MP. If it did not melt or bend, we applied pressure to the particle with the needle. If the particle broke under pressure, we did not classify the particle as an MP. Chitinous insect parts that remained post-digestion were not classified as MPs. They did not melt or bend with heat, and they would often shatter under needle pressure. Although the hot needle method is sometimes considered outdated now, it has been used for nearly a decade and was a common method at the time of this study (De Witte et al., 2014). If we were not sure about a particle, we were conservative and did not classify it as an MP. We also recorded the color of each particle, as color is often characterized in MP research (Carlin et al., 2020; Masiá, Ardura & Garcia-Vazquez, 2019). We used a Moticam X3 camera and software to take pictures and measure the longest side of a subset of particles per sample.

Microplastics and age, sex, and body condition

We calculated MP concentration by taking the number of particles divided by the mass of the GIT tissue plus the added KOH solution. Microplastic concentration for procedural blanks (i.e., controls) were the number of particles divided by the mass of the KOH solution only. We log transformed MP concentrations from GITs to normalize the data for use in the statistical analysis but report raw concentrations in the results. We used analysis of variance (ANOVA) to test if MP concentrations in bat samples were significantly different than control samples. We used ANOVA to also determine if MP concentrations varied by age and sex. We graphed body mass and MP concentrations by collection month to investigate seasonal patterns of MP concentrations and bat mass. We fit a linear regression model in R version 4.3.2 to determine if MP GIT concentration had a relationship with bat mass, a proxy for estimating fat stores (McGuire et al., 2018). Bat mass was the dependent variable and MP concentration was the independent variable.

Protocol testing with bat stomach contents

Methods used previously for extracting MPs from GITs of birds of prey (Carlin et al., 2020) and adapted for this study extracted 306 MPs from 85 of the 92 M. lucifugus stomach contents. Fibers made up over half of the observed plastics (56%), followed by fragments (35%), films (4%), foams (3%), and fiber bundles (2%). Blue particles made up over half of the observed plastics (56%), followed by clear (16%), and red particles (12%). The remaining 16% included black, brown, gray, green, pink, purple, white, and yellow particles. This determined that the method was appropriate for extracting different particle types from bat samples, that applying heat during the digestion was not necessary, and that a smaller pore size filter could be used in subsequent steps with GITs.

Microplastics concentrations and characterizations

We extracted 608 plastic particles from 26 E. fuscus GITs collected from 15 Tennessee counties (Fig. 1) and 28 plastic particles from 10 control samples. The average MP concentration in GITs was 7.5 n/g ± 10.9 SD (Table 1). The majority of microplastics were fibers (n = 574; 94%; Fig. 2). We also identified nine fiber bundles, 16 fragments, seven films, and two spheres. We identified 26 fibers and two foams in our 10 control samples. Control samples had an average MP concentration of 1.2 n/g ± 0.5 SD (Table 1). All plastics in control samples were clear fibers (64%), blue fibers (18%), purple fibers (11%), or blue foams (7%). The most abundant MPs in GIT samples were clear fibers (52%), blue fibers (35%), red fibers (8%), purple fibers (2%), blue fragments (1%), clear fiber bundles (1%), and white films (1%). We also identified fibers of other colors and other plastics occurring in smaller numbers. To be conservative in our estimates of MPs, we only used fibers in the statistics because they were the easiest to tell apart from natural material and were the most frequent type occurring in bat samples.

Microplastics and age, sex, and body condition

Microplastic concentrations were significantly higher in GIT samples than control samples (Table 1; p = 0.01). Concentrations were not significantly higher in males (n = 13) than females (n = 13) in the ANOVA analysis (p = 0.08). Higher concentrations in adults (n = 18) compared with juveniles (n = 7) were also not significant (Table 2; p = 0.96). The linear model determined that MP fiber concentrations in the GIT had a significant negative association with bat body condition, with lower body mass associated with higher GIT MP concentrations (Fig. 3; p = 0.007, adjusted r² = 0.23, F_1,24 = 0.72 [Bat mass = 15.77−1.37 * log (MP concentration)]).

Eptesicus fuscus were collected in April (n = 1), May (n = 8), June (n = 4), July (n = 5), August (n = 3), September (n = 1), and November (n = 4). We did not compare body mass and MP concentrations across months statistically due to limited sample size. Mean bat body mass appeared similar in all months (12.4–14.1 g range) except November (17.5 g ± 3.7 SD; Fig. 4). Microplastic concentrations appeared lowest in April (0.7 n/g), but the sample size was a single bat. May appeared to have the highest MP concentrations (11.7 n/g ± 18.3 SD; Fig. 4).