Plasma proteomic profiling of bacterial cold water disease-resistant and -susceptible rainbow trout lines and biomarker discovery

Abstract

Genetic variation for disease resistance is present in salmonid fish; however, the molecular basis is poorly understood, and biomarkers of disease susceptibility/resistance are unavailable. Previously, we selected a line of rainbow trout for high survival following standardized challenge with Flavobacterium psychrophilum (Fp), the causative agent of bacterial cold water disease. The resistant line (ARS-Fp-R) exhibits over 60 percentage points higher survival compared to a reference susceptible line (ARS-Fp-S). To gain insight into the differential host response between genetic lines, we compared the plasma proteomes from day 6 after intramuscular challenge. Pooled plasma from unhandled, PBS-injected, and Fp-injected groups were simultaneously analyzed using a TMT 6-plex label, and the relative abundance of 513 proteins was determined. Data are available via ProteomeXchange, with identifier PXD041308, and the relative protein abundance values were compared to mRNA measured from a prior, whole-body RNA-seq dataset. Our results identified a subset of differentially abundant intracellular proteins was identified, including troponin and myosin, which were not transcriptionally regulated, suggesting that these proteins were released into plasma following pathogen-induced tissue damage. A separate subset of high-abundance, secreted proteins were transcriptionally regulated in infected fish. The highest differentially expressed protein was a C1q family member (designated complement C1q-like protein 3; C1q-LP3) that was upregulated over 20-fold in the infected susceptible line while only modestly upregulated, 1.8-fold, in the infected resistant line. Validation of biomarkers was performed using immunoassays and C1q-LP3, skeletal muscle troponin C, cathelcidin 2, haptoglobin, leptin, and growth and differentiation factor 15 exhibited elevated concentration in susceptible line plasma. Complement factor H-like 1 exhibited higher abundance in the resistant line compared to the susceptible line in both control and challenged fish and thus was a baseline differentiator between lines. C1q-LP3 and STNC were elevated in Atlantic salmon plasma following experimental challenge with Fp. In summary, these findings further the understanding of the differential host response to Fp and identifies salmonid biomarkers that may have use for genetic line evaluation and on-farm health monitoring.

Keywords: Flavobacterium psychrophilum; bacterial cold water disease; biomarker; complement C1q-like protein 3; complement factor H-like 1; disease resistance.

Figure 1

Experimental design comparing plasma proteomes between genetic lines and infection status. Six groups were compared in this study and are shown by genetic line and infection status. N is the number of fish in each group contributing to the plasma pool. Proteomic changes measured by TMT 6-plex were compared to a published RNA-seq dataset from 1 g fish, and candidate biomarkers were validated by ELISA and/or SPARCL immunoassays. In the SFp group, 15 fish out of 20 initial study fish contributed to the pool.

Figure 2

Bacterial load and plasma packed cell volume on day 6 post-infection. (A) Flavobacterium psychrophilum abundance in spleen tissue and expressed as genomic equivalents per 100 ng of input sample DNA (geometric mean ± sd). (B) Whole blood packed cell volume (mean ± sd). The 90% confidence intervals for healthy fish are indicated by dotted lines (from Marancik et al. (29)). One-way ANOVA with letters indicating significant difference (p < 0.05).

Figure 3

Infection-regulated proteins with ≥2-fold change. (A) Venn diagram of proteins either shared or unique to each group: SFp/SPBS (n = 61), RFp/RPBS (n = 35), SPBS/S0 (n = 5), RPBS/R0 (n = 3). (B) Heat map of infection-regulated proteins (n = 67) specific to either RFp/RPBS or SFp/SPBS groups determined from the fold-change comparisons shown in (A). Brackets indicate broad infection regulated categories of either increased abundance (Fp ↑) or decreased abundance (Fp↓). Hierarchal clustering of both samples and proteins identifies patterns of expression and color coding indicate the following groups: unique to SFp/PBS, green; unique to RFp/RPBS, red; shared between groups, brown. Heat map: yellow increased, blue decreased fold change.

Figure 4

Comparison of the fold-change (f.c.) in protein (y-axis) and corresponding RNA (x-axis). (A) SFp/SPBS fold-change, R ^{2 =} 0.488, p < 0.0001 and (B) RFp/RPBS fold-change, R ^{2 =} 0.261, p < 0.0001. Box region marks twofold change in protein and RNA, and red colored circles indicate proteins that lack computationally predicted leader peptide and are thus likely intracellular proteins. Linear regression and 95% CI. Each slope is significantly non-zero. Individual proteins are labeled and are indicted by larger-sized circles.

Figure 5

Validation of biomarkers (mean + sd). (A) C1q-LP3, (B) STNC, (C) Haptoglobin, (D) Cath2, (E) Leptin, and (F) GDF-15. Asterisks indicate significance values (* p < 0.05; ** p < 0.01; *** p < 0.001). (G) Heat map of normalized biomarker dataset markers significantly associated with infection (q < 0.05). (H) Nearest network analysis of variables and projection in top three principal component space. (I) Spearman correlations between each variable (bottom) and significance (top). Blue = significant positive correlation, red = significant negative correlation.

Figure 6

Complement factor H-like protein domains, phylogeny, and expression differences between lines. (A) SMART domain prediction of human CFH (NP_000177.2) and eight trout proteins based on gene models from the OmykA_1.1 assembly. (B) Phylogeny of human CFH and complement factor H-related proteins and trout proteins. (C). Rainbow trout CHFL-1 ELISA. One-way ANOVA Kruskal–Wallis test with Dunn’s multiple comparisons test. Asterisks indicate significance values (** p < 0.01; *** p < 0.001). (D) TMT proteomics and relative abundance of trout CFHL-1 in plasma. (E) Day 5 cfhl-1 gene expression values from Marancik et al. (27). (F) Transcript counts in normal tissue from Swanson clonal line.

Figure 7

Biomarker elevation in commercial stocks of rainbow trout and Atlantic salmon challenged with F. psychrophilum. (A) Rainbow trout post-challenge survival; arrows indicate sampling time points D1, D3, D6, and D20 post-injection. (B) C1q-LP3 in plasma following challenge (mean ± sd). N = 10 fish per time point except for day 20 (n = 6 or 2; low and high dose, respectively). (C) Atlantic salmon C1q-LP3 at 3 and 6 days post-challenge, (D) STNC, and (E) Fp load in spleen. One-way ANOVA Kruskal–Wallis test with Dunn’s multiple comparisons test (mean and sd). Asterisks indicate significance values (* p < 0.05, ** p < 0.01).