Impacts of PCBs on grey seal pup immunity and influenza A infection vary across biological scales

Abstract

Disease outbreaks in marine species have increased, prompting concern regarding the health of coastal ecosystems. Environmental contaminants are hypothesized to contribute to disease outbreaks, as many contaminants have documented immunosuppressive effects in a variety of marine organisms. Of particular concern are marine mammals that have long life-spans, live in coastal waters, and have thick blubber, all of which leads to greater contaminant exposure and accumulation. In the Northwest Atlantic Ocean, polychlorinated biphenyls (PCBs) have been detected in grey seals (Halichoerus grypus), yet research on impacts of PCBs in this species is limited or yields contradictory results. We first validated the use of dried blood spot cards (DBS) to measure contaminants in wild grey seal pups using a paired analysis to whole blood samples (N = 30). Then, DBS was used to measure PCBs in pup blood (N = 127) and we assessed the impact of PCBs on gene expression (N = 23), cytokines (N = 87), and influenza A virus (IAV) infection status (N = 127). We found a significant correlation between the concentrations of PCBs detected by DBS and whole blood. PCBs were found to down-regulate adaptive immunity genes, but up-regulate innate immunity and defense response genes. Despite the apparent effects of PCBs at the molecular level, we observed no difference in cytokine profiles nor IAV infection status across pups with and without measurable PCBs. Our research highlights the intricacies of contaminant-induced effects on immune function and sheds light on how grey seals survive ongoing threats in their environment.

1. Introduction

Over the past several decades, an increasing number of reported disease outbreaks in coastal species have prompted concern that coastal ecosystem health may be declining (Desforges et al., 2016; Gulland & Hall, 2007), with high levels of environmental contaminants cited as one contributing factor (de Swart, Harder, et al., 1995; Ross, 2002; Ross et al., 1996). From marine mammals to humans, exposure to environmental contaminants has been associated with immunosuppression, reproductive impairment, population declines, among other health effects (I. Ahmad et al., 2023; Desforges et al., 2018; Sonne et al., 2020). Of particular concern are a group of persistent organic pollutants, polychlorinated biphenyls (PCBs), of which there are 209 different congeners, or unique chemical compounds. Their long half-lives and stability have allowed PCBs to persist in environments despite their use and production being banned or severely restricted. In addition to environmental persistence, these pollutants threaten marine wildlife due to their lipophilic nature, which allows them to bioaccumulate in organisms and biomagnify up the food chain (Faroon & Ruiz, 2016; Levin et al., 2016; O’Shea & Tanabe, 2003).

The health impacts of these contaminants are of particular concern for marine mammals, which are known to bioaccumulate some of the highest levels of contaminants due to their long life-spans, coastal residency, high trophic level, and thick blubber that acts as a repository for contaminants (Bossart, 2011; Desforges et al., 2016; Reddy et al., 2001). Over the past several decades, PCBs have been detected in global pinniped populations, including in the North Atlantic Ocean (Hobbs et al., 2002; Rigét et al., 2020; Shaw et al., 2005; Vanden Berghe et al., 2012), Baltic Sea (Nyman et al., 2002), North Pacific Ocean (Peterson et al., 2014), Arctic Ocean (Facciola et al., 2022), and Antarctic Ocean (Khairy et al., 2021). Across these bodies of water, PCBs have been found to adversely affect marine mammal immune function (e.g., de Swart et al. 1995; Desforges et al. 2016; Levin et al. 2005, 2007, 2016; Schwacke et al. 2011; Sørmo et al. 2009), the endocrine system (Bennett et al., 2021; Schwacke et al., 2011; Troisi et al., 2020; Villanger et al., 2013), and other critical biological functions (e.g., Nyman et al. 2003; Mos et al. 2007; Vanden Berghe et al. 2010). In addition, PCBs have been found to disrupt the expression of genes related to T-cell activity, cytokines, metabolism, growth and development, immune function, and more (Brown et al., 2014, 2017; Neale et al., 2005; Noël et al., 2017; Soulen et al., 2022). These prior findings set the stage for analyses of PCB-associated health effects and disease susceptibility in wild pinniped populations.

In the Northwest Atlantic, PCBs have been detected in the blubber of grey (Halichoerus grypus) and harbor seals (Phoca vitulina) (Hobbs et al., 2002; Shaw et al., 2005), two sympatric species of pinnipeds that have experienced numerous large-scale mortality events associated with infectious disease in recent years. One infectious disease, influenza A virus (IAV), was first isolated in Northwest Atlantic pinnipeds during an epidemic in 1979–1980 that resulted in the death of ~20% of the local harbor seal population (Geraci et al., 1982; Runstadler & Puryear, 2020). Since then, IAV has been implicated as the cause of disease outbreaks of varying magnitude in the Northwest Atlantic in 1982, 1991, 2011, and 2022 (Fereidouni et al., 2014; Puryear et al., 2016, 2023). The recurring nature of IAV epidemics is suggestive of the presence of a reservoir host, in which the virus can evolve prior to spilling over into susceptible populations. In the North Atlantic, grey seals are thought to fill this role as they are permissive to IAV infections without associated morbidity and mortality (Puryear et al., 2016). Grey seal populations were nearly extirpated in the mid-1900s due to intense hunting pressure and bounty programs (Wood et al., 2011), but since receiving federal protection, the population has rebounded to an estimated 27,000 animals in the Northeast US (Hayes et al., 2024). As grey seal numbers increase, so too does the potential for intra- and inter-species interactions that foster pathogen transmission in coastal waters and at shared haul out sites.

It is uncontested that PCBs can be found in Northwest Atlantic grey seals (Hobbs et al., 2002; Nyman et al., 2002), but research on the impacts of PCBs in this population is limited to sex hormones (Troisi et al., 2020) or association with exposure biomarkers (Nyman et al., 2003). Across the Atlantic, research on grey seals in Scotland suggests there is no effect of PCBs on immune-related processes such as phagocytosis or the respiratory burst, infection prevalence, or hematology and biochemical blood parameters (A. Hall et al., 1997; Hammond et al., 2005), which contrasts with studies on other pinniped species (see references above). The aim of our study was to investigate the impact of PCBs across biological scales in Northwest Atlantic grey seal pups within the context of IAV. To do so, we first validated the use of dried blood spot (DBS) cards for contaminant analysis in grey seal pups. DBS cards offer a cost-effective approach to sampling that minimizes the amount of whole blood that needs to be collected from a wild animal (Jobst et al., 2020). We then measured PCBs in grey seal pup blood and tested for associations between PCBs and gene expression, cytokine production, and IAV infection status. Due to their documented immunosuppressive properties, we hypothesized that the presence and concentration of PCBs would alter the expression of immune genes, alter the production of cytokines, and be associated with a greater probability of IAV infection status. Our findings advance our understanding of human-made contaminants and their impacts on the health of a coastal sentinel species and mechanisms that underlie environmentally modulated disease susceptibility in a proposed wild disease reservoir.

2. Materials and Methods

2.1. Sample collection

Blood samples and swabs (nasal, conjunctival, and rectal) were collected and archived during annual grey seal health assessments carried out at South Monomoy Island (Monomoy National Wildlife Refuge), Muskeget Island, and Great Point, Nantucket in Massachusetts (MA), United States between 2014 and 2020. Sampling was authorized under National Marine Fisheries Service (NMFS) research permits #10080–95 and 17670–01 and National Wildlife Refuge (NWR) System Research and Monitoring Special Use Permits 16-MNY-01 and 53514. Animal handling, biological sampling, and disease testing were conducted as previously described (Puryear et al., 2016). For this study, ~50–200ul of blood per individual was collected onto Whatman 903 protein saver cards (Cytiva, Marlborough, MA, USA) and archived at 4°C for contaminant analysis (Batterman & Chernyak, 2014). For a subset of pups, ~5–10 ml of whole blood was also collected and archived at 4°C to validate the use of DBS cards for contaminant analysis (40 CFR Part 136). Previously published serum cytokine profiles (N = 87; McCosker et al. 2025) and blood transcriptomes (N = 23; McCosker et al. 2025) from a subset of the grey seal pups used in this study enabled comparisons of measured PCBs to cytokine concentrations and gene expression.

Sex and molt stage of all pups were recorded in the field, with molt stage serving as a proxy for age in wild grey seal pups. Molt stages II (full lanugo), III (molted around face and flippers), IV (molting on body), and V (completed molting) are associated with pups that are on average 3–7 days old, 7–18 days old, 18–25 days old, and 25–40 days old, respectively (Bowen et al., 2003; Noren et al., 2008).

2.2. Dried blood spot card validation

To validate DBS cards as a sampling method for contaminant analysis, paired whole blood and DBS cards from a subset of pups were analyzed for a suite of 21 PCB congeners (IUPAC #s: 18, 28, 44, 49, 52, 66, 87, 101, 105, 118, 128, 138, 153, 170, 180, 183, 184, 187, 195, 206, 209) using the Agilent 6890 GC with a Waters Tandem Mass Spectrometer (GC/MS/MS) platform. For validation purposes, the detection (or lack thereof) of each individual PCB congener was analyzed across samples and methods. A Shapiro-Wilk test was used to determine if the number of detections per PCB congener were normally distributed, and a paired Wilcoxon-Mann-Whitney test was used to determine if the number of detections per PCB congener were different between whole blood and DBS cards. For detections of PCB congeners that were above the limit of detection (LOD) using both methods, a Pearson correlation test was used to assess the relationship between the concentration of each PCB congener in each sample measured via blood spot cards and whole blood. All statistical tests described here and below were performed in R v. 4.3.1 (R Core Team, 2024) with α = 0.05.

2.3. PCB measurement

A subset of 16 PCB congeners (IUPAC #s: 18, 44, 66, 87, 105, 118, 128, 138, 153, 180, 183, 184, 187, 195, 206, 209) were measured in the full set of DBS cards (n = 127) using the Agilent GC/MS/MS in multiple reaction monitoring mode and methods optimized by the University of Connecticut’s Center for Environmental Sciences and Engineering (Provatas et al., 2019). The samples were analyzed in five batches across which the reported LOD, or lowest concentration that can be detected by the instrument, differed due to variable signal to noise ratios during each round of measurements. For analyses of the complete dataset, we conservatively applied the maximum LOD for each of the 16 PCB congeners that were measured in all datasets and considered all measured PCB concentrations below that value as not detected (assigned a zero for further analysis). When sample size allowed, we also analyzed the batches that shared LODs separately to explore potential batch effects and report on the findings from some additional PCB congeners that were only analyzed in a subset of samples. Unless otherwise noted, the results in the main manuscript are from the analysis of the complete dataset and results from the batch-specific analyses are presented in the Supplementary Materials for comparison. Given that congeners of PCBs are typically correlated with one another and there was an excess of zero detections, we implemented a hurdle model framework, in which we evaluated the presence and concentration of ΣPCBs, as described below. The concentration of non-zero ΣPCBs were log-transformed to approximate a normal distribution.

To identify environmental and biological factors that may influence the presence and/or concentration of PCBs in the full set of DBS cards, we evaluated how ΣPCBs vary across years, sampling location, sex, and molt stage. A logistic regression was used to test for an association between the presence/absence of PCBs across years. A Pearson’s Chi-squared test, or Fisher’s Exact Test when expected counts in more than 20% of cells fell below five, was used to test for associations between PCB presence/absence and location, sex, and molt stage. Significant associations were further examined with a post-hoc pairwise Fisher’s Exact Test to test for significant differences across each pairwise comparison using the holm p-adjustment method in the R package rstatix v. 0.7.2 (Kassambara, 2023). For pups with measurable PCBs, an ANOVA with Tukey’s HSD (package rstatix) was used to test for differences in ΣPCBs across molt stages, t-tests were used to test for differences in log ΣPCBs between location and sex, and a Pearson’s correlation was used to determine if there was an association between year and ΣPCBs.

2.4. Gene expression

Blood transcriptomes from a subset of the female pups included in this study (N = 23) had been previously generated via RNA sequencing following hemolysis and removal of red blood cells, as described in McCosker et al. (2025). This subset of female samples was collected between 2016–2019, predominately from Monomoy Island (n = 21), and predominantly classified as molt stage V (n = 22). For the present study, these data were used to measure mRNA transcript abundance and test for differentially expressed genes (DEGs) between pups with (n = 3) and without (n = 20) PCBs, as conservatively assessed using the maximum study-wide LOD. Though we acknowledge that true gene expression involves several processes from genotype to phenotype (Buccitelli & Selbach, 2020), here and throughout the manuscript the term gene expression is used synonymously with mRNA transcript abundance, as is common in the field and analytical tools used for these analyses.

Analyses of gene expression were conducted using the R package edgeR (v. 4.0.12; Robinson et al. 2009). Genes that did not have read counts in at least three samples (the minimum group size) were filtered out, and counts for remaining genes were normalized based on library size. A Fisher’s exact test was used to test for DEGs between pups with and without PCBs, and genes were considered to be differentially expressed if they exhibited an uncorrected p-value <0.05 and absolute log2 fold change > 1. Using a combination of thresholds for p-value and fold change takes into account both statistical and biological significance to identify biologically meaningful genes for further investigation (McCarthy & Smyth, 2009; Xiao et al., 2014). The R package rbioapi v. 0.8.2 (Rezwani et al., 2022) was used to access the Protein Analysis Through Evolutionary Relationships (PANTHER) Classification System (Mi et al., 2019) to determine which Gene Ontology (GO) terms were significantly over-represented (Fisher’s exact test FDR < 0.05) among DEGs compared to the background list of genes that were expressed in grey seal pup blood, as described in McCosker et al. (2025). The R package WebGestaltR v. 0.4.4 (Wang et al., 2020) was used to determine if there were any significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways among differentially expressed genes.

2.5. Cytokines

Concentrations of 13 serum cytokines (GM-CSF, IFNg, IL-2, IL-6, IL-7, IL-8, IL-15, IP-10, KC-like, IL-10, IL-18, MCP-1, TNFa) had been previously quantified (pg/mL) in a subset of the grey seal pups included in this study (N = 87), as described in McCosker et al. (2025). Cytokines were quantified using the commercially available Millipore MILLIPLEX^® Canine Cytokine Magnetic Bead Panel kit according to the manufacturers’ instructions, validated for use in pinnipeds by Levin et al. (2014). Samples were diluted 1:3 and standards were prepared using a 1/4 serial dilution to generate a standard curve. If the reported cytokine concentration was extrapolated outside of the standard curve calculated by the software (19% of measurements), the value was accepted as the calculated concentration and values reported as “out of range – below asymptote of equation” (71% of measurements) were assigned a zero, following Levin et al. (2014). For the present study, cytokines that were either not detected or only detected in a single pup from this subset were removed from further analyses (GM-CSF, IP-10, MCP-1, and TNFα), and associations between the profile of the nine remaining cytokines with PCB presence and concentration were tested. Cytokines operate within complex biological networks, often influencing the production of one another. Therefore, we used multivariate statistics to (1) determine if the presence and/or concentration of cytokines differed across pups with and without measurable PCBs and (2) visualize differences among multivariate cytokine distributions between groups. To do so, we constructed dissimilarity matrices using 1 - Sørensen’s index (Sørensen, 1948) and the Bray-Curtis index to quantify dissimilarity between groups (i.e., pups with and without measurable PCBs) in cytokine presence/absence and concentrations, respectively, using the function betadiver from R package vegan v. 2.6–8 (Oksanen et al., 2024). Permutational multivariate analysis of variance (PERMANOVA) was used to compare the dissimilarity of cytokine presence/absence and concentration between pups with and without measurable PCBs using adonis2 in vegan. The function betadisper from vegan was used to model group dispersions, and the assumption of homogeneity of dispersions under adonis2 was tested using an ANOVA.

Partial-least squares discriminant analysis (PLS-DA) was used to visualize differences in cytokine profiles between groups using the R package mixOmics v. 6.26.0 (Rohart et al., 2017). PLS-DA is a multivariate supervised dimensionality-reduction technique that is aware of group identities and attempts to create latent variables that separate groups as best as possible, with each latent variable explaining some proportion of variation in cytokine profiles among pups (Barker & Rayens, 2003; Ruiz-Perez et al., 2020). PLS-DA is useful for datasets with correlated independent variables, as is the case in cytokine measurements.

2.6. IAV status

Three swabs (nasal, conjunctival, and rectal) were collected from most grey seal pups in this study (N = 124) and the remaining three pups were sampled using a subset of swabs (i.e., rectal, conjunctival and rectal, or nasal). All swabs were tested for the presence of influenza viral RNA using RT-PCR following methods previously described by Puryear et al. (2016). Briefly, swabs were stored in viral transport media (VTM), kept chilled during collection, and frozen in a liquid nitrogen dry shipper before being stored at −80°C. RNA was extracted (using Omega Mag-Bead Viral DNA/RNA Kit, Omega Bio-Tek, Norcross, GA, USA) and tested for the presence of viral RNA (using qScript XLT One-Step RT-qPCR ToughMix, Quanta Biosciences, Gaithersburg, MD, USA) using primers directed against a conserved region of the influenza A matrix gene (Spackman et al., 2002). Pups were classified as IAV+ if at least one swab was avian influenza (AI) matrix positive, using a cycle threshold (Ct) of < 45. In the Northwest Atlantic, grey seals are exposed to a variety of influenza subtypes (Puryear et al. 2016), across which immune responses may vary (Sládková & Kostolanský, 2006), but this mode of viral screening cannot distinguish between subtypes. Therefore, observations likely reflect associations with the dominant IAV subtype present in pups, or those shared across multiple subtypes. A Pearson’s Chi-squared test was used to determine if the presence of PCBs was associated with IAV status. For pups with measurable levels of PCBs, a logistic regression was used to assess the relationship between ΣPCB concentration and IAV status.

3. Results

3.1. Sample collection

PCBs were measured in 127 DBS cards and 30 whole blood samples from wild grey seal pups. Across all 127 samples that were collected between 2014 and 2020, most (64%) originated from pups born on Monomoy Island, followed by Muskeget Island (31%) and Great Point (5%). Samples were approximately balanced across sexes (55% female, 39% male, 6% unknown) and were predominately molt stages IV (18%) and V (53%), with some molt stage III (12%), one molt stage II, and 21 (17%) pups for which molt stage was not recorded. In the subset of pups born on Monomoy Island in 2017 from which whole blood was collected for contaminant analysis, samples were distributed approximately evenly across sexes (40% female, 50% male, 10% unknown) and molt stage (37% V, 27% IV, 30% III, 3% II, and 3% unknown). For a full list of grey seal pup samples and attributes, see Supplemental Table S1.

3.1. Dried blood spot card validation

In the subset of wild grey seal pups for which paired whole blood and DBS cards were collected (N = 30) and 21 PCB congeners were measured, 70% and 67%, respectively, of grey seal pups had measurable PCBs, many with detections for only one or a few congeners. Across the 21 PCB congeners measured, 13 were detected through whole blood and 10 were detected DBS cards, and there was no significant difference in the number of detections per congener between methods (Z = −1.16, p = 0.25) (Figure 1). For instances in which a PCB congener was detected in a sample through both DBS and whole blood (n = 15), there was a significant correlation (r = 0.75, df = 13, p < 0.01) in the concentrations measured between both collection methods (Figure 1, Supplemental Figure S1).

Figure 1.

Validation of dried blood spot (DBS) cards for use in measuring polychlorinated biphenyls (PCBs) in wild grey seal pups, including the proportion of samples (out of N = 30) in which each congener was measured above the limit of detection (LOD) through DBS cards and whole blood (WB; left). For instances in which a PCB congener was detected in a sample through both DBS and whole blood (n = 15), the concentration measured was significantly correlated (r = 0.75) between whole blood and DBS cards, depicted by a line with shaded 95% confidence intervals (right). Data originated from wild grey seal pups sampled on Monomoy Island, Massachusetts, USA in 2017.

3.2. PCB measurement

Using the conservative study-wide LOD across 16 PCB congeners measured in all samples, PCBs were detected in 26% (n = 33) of 127 sampled grey seal pups, including 13 of the 16 measured PCB congeners (Table 1). Individual PCB congeners were detected in 1–12 pups, with PCB 44 (9.45%) and PCB 66 (5.51%) detected the most frequently. Across all samples, PCB 87 had the greatest mean concentration (231.55 ng/g wet weight ± 23.16 SE), when considering only detections that were above LOD (Table 1). Within a single sample, total PCB concentrations within pups ranged from 27.9 to 445 ng/g wet weight.

Table 1.

Summary of 16 polychlorinated biphenyl (PCB) congener detections in wild grey seal (Halichoerus grypus) pups (N = 127). Metrics reported include the maximum limit of detection (LOD) across batches, the number of samples in which a given PCB congener was detected above the LOD, and the mean ± standard error (SE), median, and range of non-zero concentrations for each PCB congener measured. Data originate from pups sampled on Monomoy Island, Muskeget Island, and Great Point, Nantucket, Massachusetts, USA between 2014 and 2020. Batch-specific summaries can be found in Supplemental Tables S3 – S5.

Congener	LOD (ng/g wet weight)	# Samples > LOD	Mean ± SE (ng/g wet weight)	Median (Range) (ng/g wet weight)
PCB 18	25	1 (0.79%)	129.5	129.5
PCB 44	26.7	12 (9.45%)	62.15 ± 5.57	54.7 (27.9 – 102.5)
PCB 66	50	7 (5.51%)	124.14 ± 15.92	72.1 (50.8 – 445)
PCB 87	75	2 (1.57%)	231.55 ± 23.16	231.55 (109.8 – 353.3)
PCB 105	26.7	4 (3.15%)	48.15 ± 4.4	51.45 (27.9 – 61.8)
PCB 118	75	2 (1.57%)	102.75 ± 9.22	102.75 (84.7 – 120.8)
PCB 128	75	1 (0.79%)	81.2	81.2
PCB 138	100	2 (1.57%)	117.3 ± 10.46	117.3 (102 – 132.6)
PCB 153	50	4 (3.15%)	77.03 ± 7.06	70.15 (54.2 – 113.6)
PCB 180	266.7	0	NA	NA
PCB 183	266.7	0	NA	NA
PCB 184	133.3	1 (0.79%)	135.4	135.4
PCB 187	266.7	0	NA	NA
PCB 195	66.7	2 (1.57%)	76.25 ± 6.74	76.25 (75.2 – 77.3)
PCB 206	26.7	1 (0.79%)	46.7	46.7
PCB 209	66.7	3 (2.36%)	118.23 ± 10.82	131.7 (72.7 – 150.3)
ΣPCBs	-	33 (25.98%)	118.75 ± 17.67	81.2 (27.9 – 445)

Within and across PCB batches, we detected little to no significant differences in the proportion of pups with PCBs or the non-zero ΣPCBs across location, sex, and molt stage (Supplemental Table S2). Across the complete dataset, the proportion of pups with PCBs significantly declined in more recent years (X² = 11.77, df = 1, p < 0.01), although the concentration of non-zero ΣPCBs remained variable and was not significantly correlated with year (r = −0.05, df = 30, p = 0.77; Figure 2).

Figure 2.

Proportion of wild grey seal pup DBS cards with measurable levels of PCBs above the study-wide maximum limit of detection (LOD) across years (left) and the log-transformed concentration of non-zero ΣPCBs across years (right), with sample sizes noted below. For the concentration of PCBs, there was no significant correlation (r = −0.05) with year, depicted by a line with 95% confidence intervals. Data originate from pups sampled between 2014 and 2023 in Massachusetts, USA.

3.3. Gene expression

Among a subset of 23 female grey seal pups from 2016–2019 for which RNA sequencing data were previously generated, 574 genes (out of 12,067 expressed genes following removal of genes with low read counts) were found to be differentially expressed between grey seal pups with (N = 3) and without (N = 20) PCBs in the complete PCB dataset. This analysis was not re-run on PCB batch-specific datasets due to sample size limitations.

Among 318 genes that were up-regulated in pups with PCBs, there were 35 over-represented Biological Process (BP) terms, including terms related to host immune & defense processes (e.g., “innate immune response,” “defense response,” “regulation of viral genome replication”), red blood cells (e.g., “erythrocyte differentiation”), metabolic and biosynthetic processes (e.g., “tetrapyrrole biosynthetic process”), transport mechanisms (e.g., “gas transport”), among others (Figure 3). Additionally, 6 GO Cellular Component (CC) terms were over-represented among up-regulated genes that were predominately related to the cytoskeleton and plasma membrane (e.g., “cortical cytoskeleton,” “plasma membrane”). Among 256 down-regulated genes, there was one over-represented GO BP term, “adaptive immune response” and 6 over-represented GO CC terms including those related to immune activity (e.g., “immunological synapse” and “cytolytic granule”) and protein complexes (e.g., “T cell receptor complex” and “FANCM-MHF complex”; Figure 3). There were no GO Molecular Function terms found to be over-represented among either up- or down- regulated genes, nor any significantly over-represented KEGG pathways among DEGs.

Figure 3.

Significantly (FDR < 0.05) over-represented GO Biological Process and Cellular Component terms among genes that were up- (+ Fold Enrichment) or down-regulated (− Fold Enrichment) in blood samples from wild grey seal pups with measurable levels of PCBs in blood compared to pups with no measurable levels of PCBs. Data originate from pups sampled between 2014 and 2023 in Massachusetts, USA.

3.4. Cytokines

Among a subset of 87 grey seal pups in which cytokines were previously measured (McCosker et al. 2025), we found no significant difference in the multivariate distributions of cytokine profiles between pups with (n = 18) and without (n = 69) PCBs using either presence/absence (F_[1,85] = 2.70, p = 0.06) or concentration (F_[1,85] =1.23, p = 0.29) data from the complete dataset. Neither cytokine presence/absence nor concentration exhibited significantly different group dispersions (F_[1,85] = 2.74 and 1.07, p = 0.10 and 0.30, respectively). The PLS-DA was unable to discriminate between groups, as pups with and without PCBs overlapped in multivariate space, despite the first two latent variables explaining 15.7% and 14.7%, respectively, of variation in cytokine profiles (Figure 4). Five cytokines (IL-2, IL-7, IL-8, IL-15, and IL-10) were detected more frequently in pups with PCBs, three were detected less frequently in pups with PCBs (IFNg, IL-6, and KC-like), and one was detected in all pups (IL-18), despite all cytokines exhibiting roughly equal concentrations across groups (Supplemental Figures S2 – S3). Analyses of PCB batch-specific analyses yielded similar results (Supplemental Figures S7 – S9 and Figures S12 – S14).

Figure 4.

Partial least squares discriminant analysis (PLS-DA) visualization depicting variation in cytokine profiles between wild grey seal pups with (n = 18) and without (n = 69) measurable levels of PCBs in blood. The two axes represent the first two latent variables (LV) created to reduce the number of dimensions while maximizing discrimination between groups using cytokine profiles. The proportion of variation in cytokine profiles among pups explained by each LV is indicated, and ellipses depict 95% confidence intervals. Data originate from pups sampled between 2014 and 2023 in Massachusetts, USA.

3.5. IAV status

Across all 127 grey seal pups, we found no significant difference in the proportion of animals with PCBs detected above the study-wide maximum LOD in wild grey seal pup blood (X² = 0.25, df = 1, p = 0.62), nor the concentration of ΣPCBs (X² = 0.03, df = 1, p = 0.86) between pups with (n = 41) and without (n = 86) IAV (Figure 5). Among the PCB batch-specific analyses (Supplemental Figures S5, S10, S15), only one batch exhibited a significant relationship with IAV status; ΣPCBs were higher in pups with IAV (n = 5) than pups without IAV (n = 14) in the first batch of PCB data (Supplemental Figure S5). This different result may be due to differences in the samples and/or PCB congeners included in each batch.

Figure 5.

The proportion of wild grey seal pups with (n = 41) and without (n = 86) influenza A virus (IAV) that had measurable levels of PCBs in blood (left) and log-transformed ΣPCB concentrations for pups with measurable levels (right). Data originate from pups sampled between 2014 and 2023 in Massachusetts, USA..

4. Discussion

Despite the fact that their production ceased in the 1970s, PCBs persist in the environment as a result of their chemical stability and continue to bioaccumulate in higher trophic level species, such as marine mammals, due to their lipophilic nature (Faroon & Ruiz, 2016; Levin et al., 2016). We quantified PCBs in grey seal pup blood, validated the use of DBS cards for PCB analysis in pups, and assessed the relationship between PCB load and pup health across multiple biological scales: gene expression, cytokine production, and IAV infection status. We found that PCBs are associated with dysregulation of immune and defense genes, but there was little to no evidence of an association between PCBs and cytokine profiles or IAV infection status. Our findings highlight potential mechanisms that may be related to the grey seal’s ability to function as a wild disease reservoir host.

In our paired samples, the profiles and concentrations of PCBs were not significantly different between whole blood and DBS cards, validating the use of DBS for PCB analysis in grey seal pups. DBS cards were originally developed for medical applications in humans but have since been used for various applications in the wildlife health field (Hansen et al., 2014; Lehner et al., 2013; McHuron et al., 2019; Provatas et al., 2017, 2019; Shlosberg et al., 2011). The use of DBS cards requires a smaller volume of blood, thus minimizing negative impacts on live animals, and is generally simpler and cheaper to collect, store, and transport than larger volumes of blood (Jobst et al., 2020; Lehner et al., 2013). Our validation of this tool for measuring PCBs in grey seal pups can guide future wildlife pinniped health assessment work, thereby increasing the wealth of information that can be obtained with a limited amount of blood.

Using DBS cards to measure 16 PCB congeners, our study detected PCBs in 26% (n = 33) of wild grey seal pups sampled between 2014 and 2020, providing evidence that PCBs persist in Northwest Atlantic food webs. Across our study period we also detected a declining trend in the presence of PCBs, but not in the concentration of PCBs. These observations mirror those elsewhere across the globe where PCBs continue to be detected in pinnipeds, but in declining numbers. Within the past two decades, PCB detections and declining trends have been reported among Scottish and Baltic grey seals (Bennett et al., 2021; K. J. Robinson et al., 2019; Schmidt et al., 2020), in ringed seals (Pusa hispida) across the Arctic and Baltic Sea (Bjurlid et al., 2018; Facciola et al., 2022; Rigét et al., 2020), and in Weddell seals in Antarctica (Khairy et al., 2021).

Though relative changes reported within studies can be compared globally, it is more challenging to compare absolute values of PCB load between studies. Studies vary among the congeners that are measured (209 total PCB congeners exist, but many studies only measure a fraction of these), the tissue type within which PCBs are measured (e.g., blubber, liver, blood), the age class of pinniped studied (e.g., pups, juveniles, adults), and the way in which PCB data are presented in publications (e.g., individual congeners, grouped congeners, or sum PCBs). These confounding factors may in part explain differences that were observed between the PCB profiles that were reported among grey seal pup blood in the Northwest Atlantic in this study and the dominant patterns in PCB profiles reported previously among pinnipeds. Across many previous studies, PCBs 153, 138, and 180 have been reported as the most abundant congeners detected in pinnipeds (Bernt et al., 1999; Debier et al., 2003a; Khairy et al., 2021; Koistinen et al., 1997; K. J. Robinson et al., 2019; Vanden Berghe et al., 2012), similar to other wildlife species (Megson et al., 2022). In contrast, the most frequently detected congeners in grey seal pups in the present study were PCBs 44 and 66, while the greatest mean concentration of PCBs was observed in PCB 87 (Table 1). This inter-study difference may be due to differences between studies in the tissue in which PCBs were measured. PCBs with lower chlorine content (denoted by lower numbers on the scale from 1–209) are more frequently detected in serum and they are more susceptible to biotransformation whereas PCBs with higher chlorine content have greater lipophilicity and are more likely to be found in fatty tissues ((Grimm et al., 2015). Congruently, most (though not all; see Debier et al. 2003a and 2003b; Vanden Berghe et al. 2012) of the prior studies that report high concentrations of PCBs 153, 138, and 180 in pinnipeds have analyzed blubber samples.

In this study, PCBs were measured in grey seal pup blood, which represents the congeners and concentrations of PCBs that are actively circulating throughout the body and can have immediate physiological impacts (Beckmen et al., 2016). Fat stores, as well as PCBs that have been sequestered there, tend to mobilize during periods of fasting in pinnipeds that can lead to greater bioavailability of PCBs (Debier et al., 2003b). Grey seal pups undergo a period of fasting for about 10–28 days post-weaning (A. J. Hall, 1998). As the samples for this study were collected during this post-weaning period, this timing might have led to observed differences in the amount and type of PCBs detected in blood. Though the concentration of ΣPCBs measured within pups in our study varied greatly, ranging from 0 – 445 ng/g wet weight, we observed no detectable trends across molt stages (a proxy for pup age), suggesting that individual variation in PCB profiles may largely be due to the source of PCBs (their mother), rather than physiological processes within the pup.

During the post-weaning fast, grey seal pups are not yet foraging on their own, which means pup PCB profiles directly reflect maternal PCB load. Small quantities of PCBs are transferred in utero, and then greater quantities of PCBs are transferred during lactation as PCBs are mobilized in the dam during milk production (Debier et al., 2003a; Wolkers et al., 2004). The quantity of PCBs transferred can vary based on birth order, with first time mothers offloading greater concentrations than multiparous dams (Beckmen et al., 1999). There is no way of knowing if grey seal pups in the present study were born to primiparous or multiparous dams, but incorporating PCB measurements in studies of grey seal populations where mark-recapture methods are used, such as on Sable Island in Canada (den Heyer et al., 2021), may illuminate patterns in grey seal pup PCB profiles relative to birth order.

To further understand the impacts of PCBs in wild grey seal pups across biological scales, we investigated changes in gene expression, cytokine profiles, and IAV infection associated with PCBs. We found that pups with PCBs exhibited dysregulation of the immune system, and defense mechanisms more broadly. GO terms related to adaptive immune function (e.g., adaptive immune response, T-cell receptor complex) were associated with down-regulated genes, while those related to innate immuntiy (e.g., cytolytic granule, antiviral innate immune response) were associated with up-regulated genes in pups with PCBs. These findings are generally consistent with a large body of literature that documents PCB-mediated dysregulation of immune gene expression across pinnipeds (Castelli et al., 2014; Neale et al., 2005; Noël et al., 2017; Routti et al., 2010; Soulen et al., 2022) and other marine mammals (Buckman et al., 2011; Fossi et al., 2010; Mancia et al., 2014; Noël et al., 2014). Many genes associated with down-regulated immune GO terms were involved in T-cell function (e.g., CD6, CD7, IL2RB, PDCD1, TRAV genes), which has been shown to be impaired in captive harbor seals fed contaminated Baltic Sea fish (de Swart et al., 1994; de Swart, Ross, et al., 1995) and wild grey seal pups inhabiting the contaminated Baltic Sea (Sørmo et al., 2009). Further, the contaminant mixtures reflective of those found in wild killer whale (Orcinus orca) and polar bear (Ursus maritimus) blood has been found to decrease T and B cell proliferation in killer whales, bottlenose dolphins (Tursiops truncatus), harp seals (Pagophilus groenlandicus), hooded seals (Cystophora cristata), and polar bears (Desforges et al., 2016, 2017). Together, these results illustrate a general pattern of impaired cell-mediated immunity in marine mammals exposed to environmental contaminants. The present study suggests these effects can be seen in wild grey seals as early as post-weaning following exposure through colostrum.

Despite the apparent down-regulation of cell-mediated immunity, grey seal pups with PCBs exhibited an up-regulation of genes related to defense response, innate immune response, and regulation of viral replication. In addition, some cytokines (e.g., IL-2, IL-7, IL-8, IL-10, and IL-15) were detected more frequently in pups with PCBs compared to pups without PCBs. Though in contrast to several other studies (see Desforges et al., 2016 for a review), the up-regulation of immune and defense genes and cytokines could be due to PCB-stimulation of the immune system, which has been documented via increased natural killer activity across marine mammal species and via increased respiratory burst activity in harbor seals (Desforges et al., 2017; Hammond et al., 2005). These results could also be a by-product of increased infection rates in pups exposed to PCBs (Buckman et al., 2011). Among the subset of samples with gene expression and PCB data from the present study, 33.3% (5/20) of grey seal pups without PCBs were IAV+ while 66.7% (2/3) of grey seal pups with PCBs were IAV+. Prior work with these RNA sequence data has documented IAV-induced dysregulation of immune genes (McCosker et al. 2025), which could be confounding results in the present study. Interestingly, IAV is known to inhibit host protein production to redirect cellular machinery towards increasing viral replication during a process termed host shut-off (Rivas et al., 2016). It is possible that genes associated with regulation of viral replication and response to virus is an artifact of the grey seal immune response intercepting interactions between IAV and host cells, a mechanism that may contribute towards their ability to resist IAV-induced morbidity and mortality and serve as a wild disease reservoir. Future research using in vitro methods to experimentally expose grey seal blood to pathogens and/or to PCBs would be useful in disentangling molecular processes underlying grey seal response to health challenges (e.g., Bogomolni et al. 2016).

Beyond noting that several cytokines were detected more frequently in pups with PCBs, there was little to no significant evidence of an association of PCBs with cytokine profiles, the concentration or presence of individual cytokines, or IAV infection status. Cytokines are a part of both the innate and adaptive immune system, both of which exhibited evidence for dysregulated gene expression. In one case, the gene (CXCL8) coding for interleukin-8 (IL-8) showed evidence of up-regulation in pups with PCBs without resulting in an increase in IL-8 levels in the cytokine dataset. It is possible that grey seal pups are able to compensate for PCB-induced effects at the molecular level with post-translational processes, or that different immune cells have different levels of sensitivity to contaminants (Desforges et al., 2017). Timing may also play a role, as protein levels often peak later than mRNA transcript levels (Lee et al., 2011), and our RNA and cytokine samples were collected at the same time. This discrepancy could also be a result of the study design as the pups used for RNA sequencing were a subset of the pups for which cytokines were measured. Additionally, this study primarily focused only on 16 PCB congeners that were quantified across all batches in grey seal pup blood. Studies that have previously documented correlations between PCBs and specific cytokines (e.g., IL-1B in ringed seals (Brown et al., 2014; Routti et al., 2010), IL-10 in killer whales (Buckman et al., 2011)) measured 43 – 206 PCB congeners in either liver or blubber. The limited number of congeners measured in blood for this study along with the snapshot of PCB profiles obtained might also have hindered the ability to detect associations between PCBs and cytokine production.

Despite these methodological limitations, the lack of association detected between PCBs and IAV infection status could be a true biological signal. Brown et al. (2014) noted a “hockey stick” curve – PCBs had to reach a certain threshold before an effect was observed. The maximum ∑PCBs detected in a sample across the complete dataset in our study was 445 ng/g wet weight, which equals 0.05 ng/g lipid weight (ng/g wet weight / (lipid content % / 100) = ng/g lipid weight) assuming approximately 90% lipid content in gray seal blubber (Addison & Stobo, 1993). The most conservative threshold for PCBs in ringed seal (Pusa hispida) blubber was 1,370 ng/g lipid weight, far exceeding the greatest concentrations detected in our study. It is possible that the combination of decreased PCBs in the environment over time and the growth stage of grey seal pups (not yet foraging and accumulating PCBs on their own) could have resulted in a level of PCBs that has not yet reached a threshold whereby physiological effects can be seen across biological scales. This may also explain why no changes were observed in the expression of genes previously identified as biomarkers or associated with metabolism of PCBs (e.g., AHR, CYP1A1, ARNT), which is in contrast to many studies across marine mammals (Brown et al., 2014, 2017; Buckman et al., 2011; Castelli et al., 2014; Fossi et al., 2010; Kim et al., 2005; Noël et al., 2014; Routti et al., 2016; Soulen et al., 2022). Future studies that quantify PCB loads in adult grey seals in the Northwest Atlantic, although difficult logistically, would help elucidate patterns of PCB accumulation over time and potential physiological impacts.

As a wild reservoir species, grey seals have exhibited a remarkable ability to recover from near-extirpation in the Northeast USA and withstand infectious disease during outbreaks. This research suggests that while PCBs may disrupt immune and defense responses in weaned grey seal pups at the molecular level, the physiological responses at higher biological levels, including IAV infection, are less impacted. While this finding may suggest that PCBs are of minimal concern to the Northwest Atlantic pinniped population, it is also possible that PCBs might affect other components of the immune system (e.g., other cytokines (Brown et al., 2014; Neale et al., 2005), lymphocyte proliferation (Levin et al., 2005; Sørmo et al., 2009), phagocytosis (Hammond et al., 2005; Mos et al., 2006)) or other biological systems (e.g., the endocrine system (Bennett et al., 2021; Troisi et al., 2020)) that were not measured in the present study. Further, there are other contaminants of concern that could be impacting this population that should be investigated in future research, such as per- and polyfluoroalkyl substances (PFAS), polybrominated diphenyl ethers (PBDEs), and pesticides (Li et al., 2024). Understanding the mechanisms by which grey seals survive current threats in their environment will help us understand how marine mammal populations might be affected by the ongoing and emerging threats.

Highlights.

• This study validates the use of dried blood spot cards for PCB analysis in grey seal pups.

• Results suggest PCBs dysregulate immune and defense response genes.

• PCBs were not associated with changes in cytokine concentrations or influenza A infection.

Data availability

Data and code will be made available at https://github.com/christinamccosker/hg-pcb and https://christinamccosker.github.io/hg-pcb/.