Biomarker and proteome analysis of milk from dairy cows with clinical mastitis: Determining the effect of different bacterial pathogens on the response to infection

Emily L. O'Reilly, Lorenzo Viora, Francisco Malcata, P. Theo Pepler, Ruth Zadoks, Nicola Brady, Han Quang Hanh, Mark Mclaughlin, Anita Horvatić, Andrea Gelemanovic, Josipa Kules, Vladimir Mrljak, P David Eckersall

Research output: Contribution to journalArticlepeer-review

Abstract

Antimicrobial usage (AMU) could be reduced by differentiating the causative bacteria in cases of clinical mastitis (CM) as either Gram-positive or Gram-negative bacteria or identifying whether the case is culture-negative (no growth, NG) mastitis. Immunoassays for biomarker analysis and a Tandem Mass Tag (TMT) proteomic investigation were employed to identify differences between samples of milk from cows with CM caused by different bacteria. A total of 94 milk samples were collected from cows diagnosed with CM across seven farms in Scotland, categorized by severity as mild (score 1), moderate (score 2), or severe (score 3). Bovine haptoglobin (Hp), milk amyloid A (MAA), C-reactive protein (CRP), lactoferrin (LF), α-lactalbumin (LA) and cathelicidin (CATHL) were significantly higher in milk from cows with CM, regardless of culture results, than in milk from healthy cows (all P-values <0.001). Milk cathelicidin (CATHL) was evaluated using a novel ELISA technique that utilises an antibody to a peptide sequence of SSEANLYRLLELD (aa49–61) common to CATHL 1–7 isoforms. A classification tree was fitted on the six biomarkers to predict Gram-positive bacteria within mastitis severity scores 1 or 2, revealing that compared to the rest of the samples, Gram-positive samples were associated with CRP < 9.5 μg/ml and LF ≥ 325 μg/ml and MAA < 16 μg/ml. Sensitivity of the tree model was 64%, the specificity was 91%, and the overall misclassification rate was 18%. The area under the ROC curve for this tree model was 0.836 (95% bootstrap confidence interval: 0.742; 0.917). TMT proteomic analysis revealed little difference between the groups in protein abundance when the three groups (Gram-positive, Gram-negative and no growth) were compared, however when each group was compared against the entirety of the remaining samples, 28 differentially abundant protein were identified including β-lactoglobulin and ribonuclease. Whilst further research is required to draw together and refine a suitable biomarker panel and diagnostic algorithm for differentiating Gram- positive/negative and NG CM, these results have highlighted a potential panel and diagnostic decision tree. Host-derived milk biomarkers offer significant potential to refine and reduce AMU and circumvent the many challenges associated with microbiological culture, both within the lab and on the farm, while providing the added benefit of reducing turnaround time from 14 to 16 h of microbiological culture to just 15 min with a lateral flow device (LFD).
Original languageEnglish
Article number105240
JournalResearch in Veterinary Science
Volume172
Early online date29 Mar 2024
DOIs
Publication statusPublished - Jun 2024

Keywords / Materials (for Non-textual outputs)

  • Mastitis
  • Dairy biomarker
  • Acute phase
  • Classification tree diagnostic

Fingerprint

Dive into the research topics of 'Biomarker and proteome analysis of milk from dairy cows with clinical mastitis: Determining the effect of different bacterial pathogens on the response to infection'. Together they form a unique fingerprint.

Cite this