Microplastic Ingestion Induces Size-Specific Effects in Japanese Quail

Plastic pollution can pose a threat to birds. Yet, little is known about the sublethal effects of ingested microplastics (MP), and the effects of MP < 1 mm in birds remain unknown. This study therefore aimed at evaluating the toxicity of environmentally relevant polypropylene and polyethylene particles collected in the Norwegian coast in growing Japanese quail (Coturnix japonica). Birds were orally exposed to 600 mg MP over 5 weeks, covering small (<125 μm) and large (3 mm) MP, both separately and in a mixture. We evaluated multiple sublethal endpoints in quail, including oxidative stress, cytokine levels, blood-biochemical parameters, and reproductive hormones in blood, as well as body mass. Exposure to small MP significantly induced the activities of the antioxidant enzymes catalase, glutathione-S-transferase, and glutathione peroxidase. Exposure to large MP increased the levels of aspartate aminotransferase (liver parameter) and decreased 17β-estradiol levels in females. Body mass was not directly affected by MP ingestion; however, quail exposed to small MP and a mixture of large and small MP had a different growth rate compared to control quail. Our study used similar levels of MP as ingested by wild birds and demonstrated size-dependent effects of MP that can result in sublethal effects in avifauna.


Collection of plastic
Several plastic items were collected from islets in Mausund, Norway (63°52'N, 8°39'E) in March 2020. Fig. S1 shows the environment where the plastic items were collected.

Plastic cleaning and preparation
The collected items (Fig. S2) were gently washed with ultrapure water and further analyzed for polymer type.

Plastic selection for the experiment
The polymer type was determined using a confocal Raman spectrometer (alpha300 R, WITec) and spectra were confirmed by comparison to the ST-Japan polymer database (L60002). Three items were selected according to unambiguous polymer identification, high abundance in marine environments, and a large volume available for performing the experiment.

Fig. S5
Plastic item 3 selected for the experiment and its RAMAN spectrum (red line) aligned with that of the polypropylene from the database (blue line). This item was identified as polypropylene (PP).

Plastic preparation for the experiment
Two different size classes were prepared: < 125 μm and 3 mm particles. First, the three plastic items were cut into smaller pieces with scissors. Then, half of the pieces were prepared by punching out 3 mm circular disks using a leather hole-puncher (Fig. S6). The other half of the plastic was cryo-milled to obtain particles < 125 μm (Fig. S7). The cryo-milling was performed through ca. 30 cycles of 2 min immersion in liquid nitrogen followed by 2 min of milling (Retsch Mixer Mill 400) at a frequency of 30 Hz until visual inspection indicated the desired size. The resulting powder was sieved using a vibratory sieve shaker (Retsch AS 200 basic) with a 125 μm sieve, and all MP < 125 μm were collected.

Quail husbandry
In October 2020, 160 quail eggs were purchased from a local farm (Trøndelag, Norway) and kept in incubators (type 180, America A/S, 94 Thisted, Denmark and J. Hemel, Verl, Germany) for 16 days at 38 °C and 50-60 % humidity with automatic egg-turning (90° per hour). The last 2 days before hatching (day 16-18), the egg rotation was stopped, and humidity was increased to 70-75 %. Fiftysix chicks hatched successfully and, once dried (24h post-hatch), they were moved to cages lined with paper and equipped with a cardboard box shelter to increase animal welfare. Eight custom-built wooden cages (1.1 m x 0.9 m) with wire-mesh lids at a height of 30 cm were used. An infrared heat lamp set at 37 °C created a gradient of heat in each cage. The temperature of the heating lamp was decreased by 2-3 °C each week until quail reached 4 weeks of age when the lamps were removed.
All cages were maintained in the same room at 22-25 °C with a light-dark cycle of 14-10 h. At week 2, cages were lined with wood chip bedding to encourage natural behaviors such as foraging and dust bathing. Quail were marked with flexible leg bands shortly after hatching for individual identification, which were replaced with larger sizes as the quail grew.

Plastic exposure and recovery
To administer the plastic powder to the quail, the desired weight of the powder was placed in a ceramic dish where finely ground poultry food and water were added (amount determined visually) in order to form soft pellets of the desired consistency. Finally, to facilitate administration to the quail, the pellets were divided into smaller pellets (Fig. S12).

Fig. S12
The different polymer types in a powder form (left). After grinded food and water were added to create the desired texture of the pellet (middle). The final small pellets created to facilitate oral administration to the quail (right).
Both 3 mm particles and powder pellets were administered orally to the quail. The 3 mm particles were carefully hand-fed using blunt forceps (Fig. S13), while the powder mixed with food and water was hand-fed using small spoons. To avoid bias for the oral administration, control quail were also captured and hand-fed food pellets following the same procedure.  Table S1. After euthanasia, the stomach tract of the birds was opened to count the remaining 3 mm particles ( Fig. S14).

Blood processing
Blood collected in heparinized Eppendorf tubes was stored on ice until it could be centrifuged (max. 5 min) to separate the plasma and red blood cells (RBC) fractions (4,000 g for 5 min at 4 °C). Plasma was then collected and transferred into a new tube and stored at -80 °C, while RBC samples were washed with saline solution (NaCl 9 mg/mL) and centrifuged again (4,000 g for 5 min at 4 °C) before storing the supernatant at -80 °C.

Oxidative stress parameters
The activities of CAT, GST and GPx were determined as described in Mennillo et al. 2 . Briefly, CAT activity was determined based on the peroxidation of methanol in the presence of hydrogen peroxide (H2O2). The formaldehyde produced was quantified using a colorimetric assay with 4amino-3-hydrasino-5-mercapto-1,2,4-triazole (Purpald) as the chromogen, and the absorbance was read at 540 nm. GST was determined using 1-chloro-2,4-dinitrobenzene (CDNB) as a substrate and absorbance was read at 340 nm. An extinction coefficient of 9.6 mM -1 cm -1 was used to calculate GST activity. The activity of GPx was determined by calculating the rate of NADPH oxidation coupled with glutathione reductase and tert-butyl hydroperoxide (12 mM) and the absorbance was read at 340 nm. The specific activity was determined using an extinction coefficient of 6.22 mM -1 cm -1 . The activities of CAT, GST and GPx were then divided by the total protein concentration (mg mL -1 ) determined by the Bradford assay. We used bovine serum albumin (BSA) to prepare the calibration curve and, after adding dye solution (Coomasie blue) to protein samples, the absorption was measured at 595 nm. SOD was quantified using a ®Cayman SOD kit (Ann Arbor, MI, USA) following the instructions of the assay. The plate was read at 450 nm of absorbance. The activity of SOD is given in U L -1 . The TBARS assay used the measurement of malondialdehyde (MDA) to construct the standard curve and the absorption was read at 530 nm. All enzymes and protein determinations were done using a Synergy HT microplate reader from Bio-Tek Instruments Inc. (Winnoski, Vermont, USA) for absorbance and fluorescence readings.
All samples were assayed in duplicate or triplicate. Therefore, we used two 96-well plates for each assay. The intra and inter-assay coefficient of variation (CV, mean % ± SE) are determined based on the final concentration for each replicate. The inter-assay CV is calculated based on 2 -4 samples. Table S2. Mean % ± SE intra-and inter-assay coefficient of variation (CV).

Cytokines
All samples were assayed in duplicate in TNF-α and 20 samples in IL-1β. We performed the analyses in two 96-well plates for TNF-α and one plate for IL-1β. The intra and inter-assay CV (mean % ± SE) are determined based on the final concentration for each replicate. The inter-assay CV for TNF-α is calculated based on 4 samples. Table S3. Mean % ± SE intra-and inter-assay coefficient of variation (CV).

Reproductive hormones
After high-speed centrifugation to remove the lipid phase, plasma samples were extracted twice with diethyl-ether using 4 times the sample volume (1200 μL), frozen in liquid nitrogen, and the supernatant was subsequently collected. After evaporation of the organic phase overnight, samples were reconstituted with 300 μL of assay buffer and frozen at -80 °C for analyses.
All samples were assayed in duplicate. The sample size was 23 for testosterone and 32 for 17βestradiol. We performed one assay for each reproductive hormone. The intra-assay CV (mean % ± SE) are determined based on the final concentration for each replicate. Table S4. Mean % ± SE intra-and inter-assay coefficient of variation (CV).

Statistical Output
The following tables show AICc values for models explaining variation in the studied parameters. The best-supported models (ΔAICc<2), those in bold, were selected for further analysis. Dependent variables included treatment (control, treatment 1, treatment 2, treatment 3), sex, body mass, and cage (C1-C8) distribution. The sample size was 55 quail. As sex was an influential factor in some parameters (HSI, Ht, TG, CK, AST, Chol, Glu, TNF-α), the same models were run again with the sexes separated. Testosterone (T) was only performed for males (n = 23) while 17-β estradiol (E2) only for females (n = 32). Table S7. AICc values for models explaining body mass (BM; mg) variation.
When evaluating differences in body mass, the best model included the interaction between treatment and week, sex, and cage (Table S7). Sex was found not-significant (Table S8), while cage was an influential factor, specially cage 2 (Table S8). To confirm that the interaction between week and treatment was not influenced by cage 2, we rerun the model excluding it and the results obtained were the same (Table S9). Since we found a significant interaction between week and treatment, resulting in different slopes in the growth of the quail, we also investigated potential statistical differences in the body mass of quail each week per separate (Table S10).  Table S11. AICc values for models explaining variation in HSI (Hepatosomatic Index). We found that T1 quail tended to show higher Ht than control quail (post-hoc t = 2.12, P = 0.09), but this could instead be due to arbitrary slight dehydration 3 than an effect of MP. When analyzed separately by sex, no further differences were found (Table S13). Similar to AST (detailed in the main Manuscript), CK levels were also higher in males than females (F1 = 7.83, P < 0.01; t = 2.79, P < 0.01; Fig S18), and when analyzed separately by sex, only males showed a difference between treatment groups (F3 = 3.84, P = 0.03) with T2 and T3 males showing lower CK levels than control males (both P < 0.05). Because the pattern of CK was not the same as AST, we could rule out a possible muscular damage.