Under pressure—exploring partner changes, physiological responses and telomere dynamics in northern gannets across varying breeding conditions

Study site and species

We conducted fieldwork on Bonaventure Island (48°30′N, 64°09′W) located in the Parc national de l’Île-Bonaventure-et-du-Rocher-Percé in the Gulf of St. Lawrence (Québec, Canada), the largest gannet colony in Canada, where a long-term study started in 2008. This colony was monitored annually from 2008 to 2019 for partnership status and breeding success. Gannets in this population are known to exhibit a highly variable divorce rate (22 ± 3%, range: 13–46%), which is influenced by the previous year’s breeding success. During the same period, breeding success remained low and variable, ranging from 3% to 60%, with an average of 32 ± 5% (Pelletier & Guillemette, 2022).

Fieldwork

Between 2017 and 2019, we captured gannets once or twice during the breeding season (incubation and chick rearing periods) using a noose pole. We recorded the body mass (BM) of the birds, and when two BM measurements were taken over the season, we calculated the body mass variation (BMvar, averaging 27 ± 7 days between two measures). All bird capture and handling methods were approved by the Animal Care Committee (ACC) of the Université du Québec à Rimouski and complied with the guidelines of the Canadian Council on Animal Care (CCAC; CPA-49-12-102, CPA-65-16-177). We marked each bird with a U.S. Fish and Wildlife Service steel ring with an alphanumeric code and colored plastic band (permit numbers SC25, RE-27, 10704) with approval from Société des établissements de plein air du Québec (permit numbers PNIBRP-2008-001 to PNIBRP-2019-001). We identified established partnerships in the set of monitored nests as previously described (Pelletier et al., 2023; Pelletier & Guillemette, 2022).

We plucked breast feathers (for sex determination) and we took blood samples (less than 5 ml, approx. 0.16% volume/weight) as described in Pelletier et al. (2023). In the field, blood samples were divided in two tubes, one for blood smears and hematocrit preparation, and another one was centrifuged (5 min; 1,700× g). Two capillary tubes were centrifuged during 5 min at 1,300× g for hematocrit reading. Two or three blood smears were prepared in the field using the two-slide wedge technique, fixed 2 min in absolute methanol and air dried prior to storage (Clark, Boardman & Raidal, 2009). From the centrifuged tube, plasma and blood cells phases were divided in cryotubes and we kept them in liquid nitrogen (−196 °C) for up to 6–7 days and sorted them on dry ice (−78 °C) before storing them in a −80 °C freezer until biomarkers analyses. The birds were immediately released near their nesting site after blood sampling. Handling time was kept to a maximum of 15 min and the blood sample was taken within the first 5 min. None of the birds displayed any immediate harmful effects.

Telomere length

Red blood cells were centrifuged from blood samples and stored previously at −80 °C. DNA was extracted with DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). Agarose gel electrophoresis was used to determine the integrity of genomic DNA samples and DNA was checked using the ratio of absorbance (A₂₆₀/A₂₈₀) on a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA). The telomere length assay was adapted from the published original protocol (Cawthon, 2002), from the modified protocol for birds (Criscuolo et al., 2009), and from the qPCR protocol developed for the Roche LightCycler 480 (LC480) (Jodczyk et al., 2015). This approach allows for the relative quantification of the mean telomere length in blood cells by assessing the quantity of telomere repeats (T) within a sample of genomic DNA, in relation to the amount of a single copy reference gene (S). This relationship is then expressed as a T/S ratio. For the telomeric DNA and the single copy reference gene, qPCR was conducted in duplicate and in isolated reactions using the LC480. Care was taken to ensure that the T and S reactions for each individual sample maintained the same relative position within the 96-well plate format. Primers for the telomere PCR were telg (5′-ACACTAAGGTTTGGGTTTGGGTTTGGGTTTGGGTTAGTGT-3′) and telc (5′-TGTTAGGTATCCCTATCCCTATCCCTATCCCTATCCCTAACA-3′) (Cawthon, 2009). Primers for the single copy reference gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (Atema, Oers & Verhulst, 2013) were GAPDH-F_sula3 (5′-CCTGTTCATCACTGCCAGGT-3′) and GAPDH-R_sula3 (5′-AGCATCCCATACACCCTCC A-3′). The GAPDH primers were designed by Primer Premier 5.0 (Premier Biosoft International, Palo Alto, CA, USA). Amplicon lengths were 160 bp, the GC content was 55%, and the melting temperature was between 61.13 °C and 62.63 °C. We compared the primer sequences to known gene databases at NCBI, using the BLAST option. BLAST of the 2 primers, primer F and primer R identified GAPDH gene of Sula nebouxxi (blue-footed booby) with E-value of 0.33 or less. All the primers were synthesized and purified by Sigma Life Science (Ontario, Canada). We performed qPCR for both GAPDH and telomeres using 20 ng of DNA per reaction (Criscuolo et al., 2009; Jodczyk et al., 2015). The telomere primers telc/telg were used at a concentration of 300 nM and primers GAPDH-F_sula3/GAPDH-R_sula3 at 900 nM for a final volume of 20 µl containing 10 µl of LightCycler^® 480 SYBR Green I qPCR Master Mix.

The thermal cycling profile for the telomere PCR consisted of the following steps: 95 °C for 12 min followed by 40 cycles (3-step) of 95 °C for 10 s, 62 ° C for 10 s and 72 °C for 5 s. GAPDH PCR started with 12 min at 95 °C followed by 40 cycles (3-step) of 95 °C for 10 s, 58 °C for 10 s and 72 °C for 6 s. Melt curve analysis was performed at the end of each run to verify the specificity of PCR amplification products. Each 96-well plates included non-target controls and serial dilutions (with 80 ng, 20 ng, 5 ng, 1.25 ng and 0.3 ng of DNA per well), run in triplicate, of DNA from the same reference bird (the internal calibrator), which was used to generate a reference curve to control for the amplifying efficiency of the qPCR and to calibrate the threshold Ct. Mean values per plate were used to calculate relative T/S ratios of target individual relative to the reference individual using the formula 2^−ΔΔCt , where [2^{C_t(telomeres)}/2^C_t(GAPDH)]⁻¹. T/S ratios were used to estimate the average relative telomere length (TL) and TL_year1−TL_year0 for the annual telomere rate of change (TROC). Efficiencies of qPCR reactions were between 96–113% for the five telomeres plates (T plates) and between 98–107% for the five GAPDH plates (S plates). Mean intraplate coefficient of variation was 0.5 ± 0.7% for the Ct values of GAPDH assays and 0.9 ± 0.7% for the Ct values of telomere assays, and interplate coefficient of variation was 3.8% (T plates) and 2.2% (S plates) for △C_t values of the reference sample. Intraclass correlation coefficient for intra- and inter- runs (ICC; based on 2 samples repeated over the 96-well plates) were 0.975 and 0.929 for the five T plates, and 0.968 and 0.910 for the five S plates, respectively (Eisenberg, 2016; Verhulst et al., 2016).

Oxidative stress and inflammation

We measured two biomarkers commonly used of oxidative damage (thiobarbituric acid reactive substances (TBARS) and 8-hydroxy-2′-deoxyguanosine (OHdG)), one biomarker of total antioxidant capacity (TAC) in plasma, two biomarkers of peroxidation susceptibility (peroxidation index of plasma (PI_p) and blood cells (PI_c)) and three biomarkers of inflammation susceptibility (ω6/ω3 ratio in plasma and blood cells, and plasma globulins (GLOB) concentration, see next subsection). The method for measuring TBARS and TAC in gannets has been described elsewhere in depth (Pelletier et al., 2023). We corrected TBARS for triglycerides concentration in plasma (TBARSt) (Perez-Rodriguez et al., 2015).

Oxidative DNA damage was assessed by measuring the amount of OHdG in plasma, using commercial kits (DNA/RNA oxidative damage ELISA Kit; Cayman Chemical, Ann Arbor, MI, USA). We used kit instructions for measurement of OHdG in plasma (in triplicate) with a dilution of 1:50 (10 µL of plasma into 490 µL of ELISA buffer). The 96-well plates were read in EnVision Multimode Microplate Reader at a wavelength of 412 nm. Mean ± SD intraplate coefficient of variation was 4.2 ± 3.9% and interplate coefficient of variation was 5.0%.

PI_p and PI_c were determined for plasma and blood cells . Fatty acid (FA) transmethylation protocol was adapted from Lepage & Roy (1984) and Ekström et al. (2017) for plasma and blood cells. We performed trans-methylation of 50 µL of plasma (or blood cells) with 500 µL of toluene, 3 ml of a 12% sulfuric acid (H₂SO₄) methanol solution directly in test vials, at 90 °C in a dry bath for 1 h. Vials were vortexed before the dry bath, after 30 min and at the end of the procedure and were subsequently cooled at room temperature. At the end, we added 3 ml of nano water and 500 µL of hexane (for the fatty acids methyl ester (FAME) extraction), followed by centrifugation (3,000×g) for 10 min at room temperature. We collected the supernatant, containing FAME in toluene-hexane, in a test tube. Using gas chromatography (Trace Ultra 100, Thermo Fisher Scientific, Waltham, MA, USA) equipped with a 60 m × 0.25 mm i.d. capillary column (DB-23, Agilent Technologies Canada, Mississauga, ON, Canada), we separated and quantified the FAMEs. We utilized helium as the carrier gas, maintaining a constant pressure of 230 kPa, and set the vaporization temperature at 230 °C with a split injection of 100 ml min⁻¹. The programming began at 50 °C, then rose to 140 °C at a heating rate of 25 °C min⁻¹, followed by an increase to 195 °C at a heating rate of 3 °C min⁻¹, and finally increased at a rate of 4 °C min⁻¹ up until 225 °C, which we maintained for 5 min. We identified the individual methyl esters by comparing them with known standards (Supelco 37 Component FAME mix). The proportion of each FA was calculated by dividing each peak area with the sum of all FA peak areas in each individual. Proportions of ω6 and ω3 fatty acids were cumulated to evaluate inflammation susceptibility in plasma and blood cells. PI, was calculated with cumulated percentages of unsaturated fatty acids per categories according to Hulbert et al. (2007) (values without units): ${\displaystyle PI=\sum \left[\left(0.025\times \text{\%}monoenoics\right)+1\times \left(\text{\%}dienoics\right)+2\times \left(\text{\%}trienoics\right)+4\times \left(\text{\%}tetraenoics\right)+6\times \left(\text{\%}pentaenoics\right)+8\times \left(\text{\%}hexaenoics\right)\right]}$

where % monoenoics = cumulated percentage of monounsaturated fatty acids (MUFA, with one double bond) and % dienoics, trienoics, tetraenoics, pentaenoics, hexaenoics are cumulated percentages of PUFA with two, three, four or six double bonds, respectively.

Hematological biomarkers

The fixed and dried blood smears were stained during the weeks following blood sampling using a Wright-Giemsa modified staining protocol (Woronzoff-Dashkoff, 2002). Blood smears were examined by only one observer with a compound microscope at 400 and 1,000× magnification. Leukocytes were identified as heterophils (H), lymphocytes (L), eosinophils, monocytes, or basophils (Campbell, 2015). At least 100 white blood cells were counted per slide in duplicates or triplicates. We calculated HL ratios as a hematological index of stress (Davis, Maney & Maerz, 2008). This ratio has been shown to scale positively with plasma levels of corticosterone, the primary vertebrate stress hormone (Davis, Maney & Maerz, 2008; Gross & Siegel, 1983; Ots & Horak, 1996; Vleck et al., 2000). Means ± SD intraindividual coefficient of variation for H count, L count and HL ratio was 6.3 ± 4.9%, 9.0 ± 7.0% and 12.3 ± 8.9%, respectively. Hematocrit (HCT) was determined in the field (in duplicates) after centrifugation. HCT was divided by plasma total protein concentration to remove the effect of hemoconcentration when there is dehydration (possibly caused by elevated temperatures at colony) and subsequent artificial increase in HCT. Mean ± SD intraindividual coefficient of variation was 0.4 ± 0.9%.

Nutritional status and muscle damage

We evaluated nutritional status (fasting/starvation levels) by measuring glucose (GLU), triglycerides (TRIG), uric acid (URIC), total protein (TP), albumin (ALB), and globulins (GLOB) concentrations in plasma and ALB/GLOB ratio (A/G) with IDEXX Catalyst Dx™. In addition, the study of variations in ALB, GLOB and A/G ratios is also important for health evaluation in birds. For example, a decrease in the A/G ratio may indicate malnutrition, inflammatory conditions, immune response or liver dysfunction, while a decrease may indicate dehydration or immunodeficiency (Lumeij, 2008). We assessed muscle damage by determining creatine kinase (CK) activity in plasma using the same device. Using a commercial colorimetric assay kit (Beta-Hydroxybutyrate (Ketone Body) Colorimetric Assay Kit, Cayman Chemical Assay, USA) and an EnVision Multimode Microplate Reader (with readings at 450 nm), we also measured beta-hydroxybutyrate (BHB), a ketone body. This serves as a marker of fatty acid catabolism during fasting (Vleck & Vleck, 2002). Pelletier et al. (2023) detailed these methods.

Data analysis

Telomere length

TL did not vary with partnership status (F_1,76 = 0.62, p = 0.621), but varied with year (F_2,67 = 20.75, p < 0.0001). TL was similar in 2017 and 2018 (t₆₇ =−1.80, p = 0.180), but TL was 137% higher in 2017 (t₆₉ = 3.29, p = 0.004) and 279% higher in 2018 compared to 2019 (t₅₂ = 6.42, p < 0.0001). The annual TROC varied only with year (t₁₂ = −6.69, p < 0.0001) and breeding stage (t₃₄ = 2.06, p = 0.047) but not with partnership status (t₁₆ = 0.24, p = 0.812). TROC was lower during the unfavorable year, with a huge effect size (Cohen’s d =  − 3.86) and was 126% higher (lower reduction in TL) during the incubation period compared to the chick-rearing period. Consequently, during a favorable year, gannets preserved their telomere length on average (Fig. 2A). Based on the biomarkers highlighted in the longitudinal analyses (Fig. 1), we explored relationships between each biomarker and the TROC using a nonparametric Spearman’s rank correlation coefficient to detect trends. Our results revealed two weak correlations between OHdG and TROC (Fig. 2B, Spearman’s ρ = −0.422, p = 0.002) and between HL ratio and TROC (Fig. 2C, Spearman’s ρ = −0.466, p <0.001). These correlations indicate that individuals that show greater telomere reduction are also those with greater oxidative DNA damage and a higher HL ratio in the second year of measurement in longitudinal monitoring.

Oxidative stress and inflammation

The annual difference in TBARSt was 60% lower in retained mates (p = 0.047) than changed mates (Fig. 1) and 88% lower during the incubation period than the chick rearing period (p = 0.015, Fig. 3A). The difference in OHdG concentration between subsequent years was not affected by partnership status (t₃₇ = 0.72, p =0.478), but varied according to years (t₃₇ = 2.04, p =0.049) with higher differences between 2018 and 2019 in retained mates (Fig. 1, mean ±95% CI did not include zero, lower 95% CI = 0.257). The annual difference in TAC in plasma varied with partnership status (t₃₅ = 2.48, p =0.020), year (t₃₅ =3.92, p =0.0004) and URIC difference (t₃₅ = 2.87, p =0.007). Retained mates exhibited 151% higher TAC than changed mates, and TAC increased more in bad year (2018 to 2019, 390% more, with a very large effect size, Cohen’s d = 1.32). TAC difference was negative in 2018 compared to 2017 for all individuals (Fig. 1), but only retained mates increased TAC in 2019 (zero not included in 95% CI, lower 95% CI = 0.060).

The annual difference in ω6/ ω3 fatty acids ratio in plasma (ω6/ω3_p) was best explained by the null model (no model including partnership status was selected, because none was different from the null model (p > 0.05) and no model including partnership status had reasonable AICc value around 2 or less. Moreover, the annual difference in PI_p did not vary according to partnership status (t₁₁ = 0.64, p = 0.530) and year (t₁₁ = −1.19, p = 0.259). With the blood cells, the annual difference in ω6/ω3_c did not vary with partnership status (t₃₆ = 0.84, p = 0.405) but varied with year (t₃₆ = −3.24, p = 0.003, Fig. 1). The ω6/ω3_c difference was 70% lower in 2018–2019 than 2017–2018, with a large effect size (Cohen’s d = −1.08). No model including the dependent variables was selected for PI_c, because none was different from the null model (p > 0.05) and no model including partnership status and year had reasonable AICc value (around 2 or less).

Body mass

There was a trend for 60% higher annual difference in BM for retained mates (t₄₇ = 1.89, p = 0.064) but all means ± 95% CI included zero, suggesting that annual individual difference in BM was not influenced by environmental conditions (Fig. 1), but BM was 115% higher during incubation than during chick-rearing (Fig. 3B). Annual difference in BMvar was 106% higher in retained mates (t₉ = 3.92, p = 0.004) and BMvar was 231% higher during bad years (2019) compared to a good year (2018) (t₇ = 6.87, p < 0.001, with a huge effect size, Cohen’s d = 5.20).

Hematology

The annual difference in HL ratio did not vary with partnership status (t₄₆ = 0.28, p =0.780), but varied with year (t₄₆ = 5.46, p = 0.018). There was a trend for a decrease in HL ratio between 2017 and 2018 and a trend for an increase between 2018 and 2019, but 95% CI included zero in both years (Fig. 1). The annual differences in HCT/TP where 77% lower in bad year (2019) compared to increase in a good year (2018) (t₄₂ = −3.45, p =0.001). Differences in HCT/TP tended to be lower in changed mates (t₄₂ = 1.82, p =0.077). HCT/TP increased during the favorable year for retained mates and decreased for changed mates during the most unfavorable year (Fig. 1).

Nutritional status and muscle damage

The annual difference in GLU concentration in plasma did not vary with partnership status (t₄₃ = −1.19, p =0.242) but varied with breeding stage (t₄₃ = 2.04, p = 0.047) and year (t₄₃ = 3.34, p = 0.002). GLU difference increase more (+180%) during incubation compared to chick rearing period and was 341% higher in 2018-2019 than 2017-2018 (Fig. 1). During favorable years, both partnership status decreased GLU concentration in plasma (zero not included in 95% CI).

The annual difference in TRIG concentration in plasma did not vary with partnership status (t₄₃ = −1.18, p =0.246) but varied with breeding stage (t₄₃ = 2.25, p = 0.029). TRIG difference was 154% higher during incubation than during chick-rearing (Fig. 3C).

No model including the partnership status was selected for the annual difference in BHB concentration in plasma, since none including this parameter had reasonable AICc (around 2 and less). The best model included breeding stage and year (χ² = 4.03, df = 1, p =0.045). BHB difference varied with year (t₁₅ = −2.16, p = 0.047) and breeding stage (t₃₂ = 2.13, p = 0.041). BHB difference was 52% lower during unfavorable year and 165% during the incubation period.

The annual difference in TP concentration in plasma varied with partnership status (t₄₂ = −2.18, p =0.035), with year (t₄₂ = 3.21, p = 0.002) but did not vary with breeding stage (t₄₂ = 1.86, p = 0.070). TP decreased in retained mates only during favorable years (Fig. 1) and increased during the unfavorable year in changed mates (with large effect size: Cohen’s d = 0.99). The annual difference in ALB concentration in plasma did not vary with partnership status (t₄₃ = −1.63, p = 0.111) but varied with year (t₄₃ = 2.22, p =0.032) with higher difference between 2018 and 2019 where both, retained and changed mates, increased ALB during bad year (Fig. 1). The annual difference in plasma GLOB concentration varied with partnership status (t₄₂ = −2.10, p = 0.042), and year (t₄₂ = 3.20, p = 0.003), and there was a trend with breeding stage (t₄₂ = 1.91, p =0.063). In the favorable year, GLOB decreased for retained mates and in the unfavorable year, GLOB increased for changed mates. The annual difference in A/G ratio did not vary with partnership status (t₄₃ = 1.35, p = 0.184) but varied with year (t₄₃ = −2.90, p = 0.006). In the favorable year, A/G ratio increased for retained mates and in the unfavorable year, A/G ratio decreased for changed mates.

The annual difference in URIC concentration in plasma did not vary with partnership status (t₄₃ = −0.39, p = 0.698) but it varied with year (t₄₃ = 3.48, p = 0.001). URIC decreased during the favorable year and increased during the unfavorable year (with a large effect size, Cohen’s d = 1.06, Fig. 1).

The annual difference in CK activity in plasma did not vary with partnership status (t₄₃ = −1.27, p = 0.210) but it varied with year (t₄₃ = 4.36, p = 0.0001). In 2018, both partnership status decreased CK, but in 2019, only changed mates showed an increase in CK.

Behavioral flexibility through partnership status

Divorce, as a behavioral strategy, has been demonstrated to mitigate the effects of low food abundance, subsequently enhancing breeding success in the year following a failure (Pelletier & Guillemette, 2022). A recent study further illuminated this phenomenon among gannets, revealing that those that paired with new partners augmented their parental effort during breeding seasons. Specifically, they increased their foraging effort, both in distance and range, during unfavorable years (Pelletier et al., 2023). This behavioral flexibility serves as a mechanism to counter some of the stressors precipitated by environmental fluctuations. However, these coping strategies are not without cost. They may adversely impact an individual’s physiological condition and, consequently, affect future reproductive success. In our study, we provide a detailed quantification of these physiological costs, shedding a light on the intricate balance between immediate benefits and potential long-term consequences.

Telomeres and reproduction costs

Telomere length, serving as an integrative marker of an individual’s life stress and stress-coping mechanisms, and individual quality (Angelier et al., 2019), was examined in our longitudinal study. We found no differences in telomere length between retained and changed mates, suggesting both types of parents might not differ in ‘quality’. Interestingly, gannets’ telomeres were stabilized during the favorable year and shortened during unfavorable breeding conditions when parental effort increased, aligning with life history theory (Stearns, 1992) and observations of telomere attrition being influenced by seasonal variations and non-permanent states (Criscuolo et al., 2020; Tissier et al., 2022; Viblanc et al., 2022). These results are also in line with other recent studies (Pepke et al., 2022; Pepke et al., 2023), which have shown that the variation due to year effects were larger than that due to all other environmental, parental or genetic factors included, suggesting that individual heterogeneity in telomere dynamics is mainly driven by annual environmental stochasticity. This may implied that increased effort during poor feeding years (Pelletier et al., 2023) is associated with shortened telomeres, and a reduction in life expectancy, as in black-legged kittiwakes (Rissa tridactyla) (Benowitz-Fredericks et al., 2022).

Despite observing higher levels of lipid oxidative damage and lower antioxidant capacity in birds that changed mates, no differences were found in DNA oxidative damage and telomere length between changed and retained mates. This suggests that stress due to parental compensation after a mate change might be counterbalanced by an endogenous or exogenous protective mechanism, perhaps similar to the telomerase activity described by Haussmann et al. (2007), which could prevent the negative impact on telomeres. Partnership status, therefore, does not seem to affect telomere length and dynamics, a surprising result that prompts further exploration of protective mechanisms during breeding or non-breeding seasons.

Our study underscores that annual breeding conditions have more pronounced effects on telomere dynamics than re-mating behavior in gannets. We observed that a favorable breeding year stabilizes telomere length, while an unfavorable year contributes to telomere shortening. This is in line with previous research showing that environmental rearing conditions can influence both offspring and parents’ telomere length (Viblanc et al., 2020), and a recent meta-analysis linking exposure to stressors with shorter telomeres (Chatelain, Drobniak & Szulkin, 2020). The results gathered across three very contrasting years substantiate these findings, emphasizing the need for future studies to further explore the relationships between telomere length, breeding conditions, and potentially unexplored biological mechanisms.

Physiological status

Our study did not report differences in terms of stress (evaluated with the HL ratio) according to partnership status. However, within our three-years sample scheme, the HL ratio increased and lymphocytes decreased during the unfavorable year (Fig. 1), indicating higher stress during unfavorable breeding conditions. Furthermore, oxygen-carrying capacity was not related to partnership status, instead showing a relationship in the between-year comparison. In medium-vs.-good years comparison, retained mates increased HCT and, at the opposite, in good-vs.-bad years comparisons, changed mates decreased HCT. As HCT is a key determinant of oxygen transport capacity and exercise performance (Böning, Maassen & Pries, 2011; Yap et al., 2018), increased HCT during favorable breeding conditions suggests that birds are more adept at performing endurance flights than during unfavorable conditions. Birds typically increase erythropoiesis and hematocrit rates before and during migratory flights (Fudickar et al., 2016), or in response to an increased workload due to brood enlargement (Hõrak, Ots & Murumägi, 1998). The study’s findings suggest that gannets may be better equipped to support metabolic activities such as foraging flights during favorable breeding conditions due to the observed increase in hematocrit levels.

Workload induced by foraging flights or protective behavior could partly drive oxidative balance. In our study, sex was not included in any of the best models produced with the set of biomarkers used, but we found that short-term plasma oxidative damage was higher (with higher TBARSt) in changed mates. However, with a medium-term biomarker test (OHdG), there was no effect of partnership status. The increase in parental effort promoted by partnership status changes (Pelletier et al., 2023) does not appear to destabilize the oxidative balance long enough to cause a divergence in cumulative oxidative DNA damage. The differences in DNA damage observed in longitudinal tests between intermediate-vs.-good and good-vs.-bad years comparisons suggest that environmental conditions may have a greater impact on oxidative balance from a medium- to long-term perspective. Our study suggests that increased foraging effort during unfavorable breeding conditions and after a mate change may affect the non-enzymatic antioxidant capacity of gannet plasma. This phenomenon has been observed in small passerines, such as zebra finches (Taeniopygia guttata), where increased parental effort led to decreased plasma antioxidant capacity and increased oxidative stress (Alonso-Alvarez et al., 2004).

Our study also suggests that changes in breeding conditions may lead to annual variations in antioxidant capacity, which are partially influenced by partnership status and associated parental effort. Gannets, regardless their pairing behavior, exhibited a decrease in total antioxidant capacity (TAC) in favorable years and an increase in TAC in unfavorable years (only in retained mates), indicating a flexible physiological response to cope with oxidative stress during periods of high metabolic activity and reduced food availability. Conversely, the changed mates did not adjust their non-enzymatic antioxidant capacity in the blood to cope with the increased metabolic demands during the unfavorable breeding conditions. King penguins (Aptenodytes patagonicus) have been shown to resist stress by upregulating endogenous antioxidant defenses (glutathione antioxidant system) and decreasing mitochondrial efficiency (Stier et al., 2019) and we suspect that differences in stress markers associated with partnership status in gannets may be related to regulation of antioxidant capacity.

The difference in peroxidation index in plasma (PI_p) decreased between medium-vs.-good years to good-vs.-bad years, suggests an increase in oxidative damage susceptibility during unfavorable years. PI_p is influenced by the composition of fatty acids (FAs) obtained mainly via dietary sources in seabirds (Käkelä et al., 2007; Käkelä et al., 2009). In our study, changed and retained mates share the same diet, but further analysis of plasma and prey fatty acid profiles could potentially confirm this (Budge, Iverson & Koopman, 2006; Iverson et al., 2004; Karnovsky, Hobson & Iverson, 2012).

The ratio of omega-6 to omega-3 fatty acids, which is a marker of inflammation susceptibility, varied between years but was not affected by partnership status. In blood cells, the ratio decreased during the unfavorable year, which suggests a higher intake of fish rich in omega-3 PUFAs or fewer prey items rich in omega-6 PUFAs. Eicosanoids, which are hormone-like lipids that modulate inflammatory and immune responses, are derived from omega-6 and omega-3 fatty acids. Omega-6 fatty acids have pro-inflammatory effects, while omega-3 fatty acids inhibit inflammation (Larsson et al., 2004; Rose & Connolly, 1999). The decrease in the omega-6/omega-3 ratio during the unfavorable year (with higher intake of fish rich in ω3 PUFAs like mackerel) could potentially provide some protection against the excess inflammation associated with increased metabolism.

In our investigation, we were unable to uncover definitive evidence pointing to elevated nutritional stress correlated with partnership status, as examined through the lenses of carbohydrate, lipid, and protein metabolism. Nevertheless, nuances in protein metabolism did manifest as discernible disparities in both nutritional stress and inflammation between changed and retained mates. To elucidate, total protein (TP) and globulin (GLOB) concentrations were found to be elevated in changed mates, and the levels of TP, GLOB, and albumin (ALB) were observed to rise during the unfavorable year. Intriguingly, there was an absence of variation in protein catabolism, as signified by uric acid levels, associated with partnership status. However, we did identify an escalation during the unfavorable year (Pelletier et al., 2023). Elevated TP and ALB levels can act as indicators of protein reserves, such as those seen during during hyperphagia before migration (Allison, 1955; Jenni-Eiermann & Jenni, 1998; Mori & George, 1978), but these levels can alternatively be reflective of dehydration (Boyd, 1981) or chronic inflammatory states, characterized by an increase in α, β or γ globulin (GLOB) concentrations and oftentimes, a decrease in ALB (Całkosiński et al., 2016). Our findings propose that changed mates might be subject to an escalation in inflammation, particularly during the unfavorable year, in spite of a reduction in susceptibility to inflammation (as illustrated by the ω6/ω3 results previously detailed). Experiments conducted on laboratory rats have documented declines in ALB concentration during acute inflammation and augmentations in various globulins as a response to either chronic inflammation or infection (Całkosiński et al., 2016). Therefore, it may be posited that the amplified GLOB levels in changed mates could have an underlying connection to chronic inflammation (Gustafsson et al., 1994; Martinez et al., 2003).

Our study also revealed that during unfavorable years, breeding gannets showed an increase in protein catabolism, as indicated by the rise in uric acid (URIC) levels (Mori & George, 1978). However, their levels were lower than those reported for other piscivorous marine birds (1,190–1,784 µmol L⁻¹, Newman, Piatt & White (1997)) but higher than values reported for other avian species (178–595 µmol L⁻¹, Campbell & Grant (2010)). In contrast, there was no significant variation in triglycerides (TRIG) and β-hydroxybutyrate (BHB) levels, suggesting that we cannot assess the nutritional status of gannets based on these markers alone (Jenni-Eiermann & Jenni, 1998). Nevertheless, plasma activity of creatine kinase (CK) indicated that annual breeding conditions influenced muscle damage in both partnership statuses, with gannets decreasing CK levels during favorable years and increasing them during unfavorable years. These results are in accordance with the highest foraging effort associated with flight (Guglielmo, Piersma & Williams, 2001; Swanson & Thomas, 2007). This observation could likely explain the increased protein catabolism observed during unfavorable years for reproduction.

The comparison of annual differences in glucose concentration (GLU) may provide insights into potential connections between changes in parental effort and impacts on telomere length, though our study did not observe a direct correlation between GLU and TL (or TROC). GLU levels are typically regulated by hormonal control and kept within narrow limits, as their maintenance is critical for the bird’s central nervous system (Braun & Sweazea, 2008; Totzke et al., 1999). Our results suggest that GLU variation in plasma may be indicative of more global consequences of stress levels than nutritional status level. This hypothesis is further supported by the HL ratio results. During favorable years, GLU levels and HL ratios were lower, and they increased during unfavorable years. The toxicity of glucose is related to damage from reactive oxygen species (ROS) through the oxidation of glucose, non-enzymatic glycation of proteins, and the production of advanced glycation end (AGEs) products, which are implicated in aging (Holmes & Austad, 1995). Rates of ROS production and protein glycation are known to be proportional to glucose concentration (Bunn et al., 1976) , and modest plasma glucose levels could be beneficial in reducing its apparent toxicity. Lower GLU levels might assist birds, which typically have 2-4x higher GLU concentration than similar-sized mammals (Holmes & Austad, 1995), to reduce glucose toxicity. While the reduced GLU levels observed in 2018 suggest that glucose regulation could be a contributing factor to telomere length stability measured in that year, the absence of direct associations between TL and GLU in our study limits the certainty of this connection. Moreover, the increase in GLU levels observed in 2019, combined with increases in TP and GLOB levels as signs of inflammation, could potentially explain telomere dynamics through hyperglycemia-related oxidative stress and inflammatory processes that accelerate age-related telomere attrition (Koliada, Krasnenkov & Vaiserman, 2015), though further research require to substantiate this relationship.