Temporal impact of metritis and systemic antibiotics on immune cell function inferred from the blood leukocyte transcriptome

Description

Approximately one-third of dairy cows develop metritis within two weeks postpartum, a uterine infection associated with reduced fertility and increased risk of reproductive culling. Antibiotics improve clinical cure and milk production following metritis but do not consistently increase subsequent fertility postpartum compared with untreated control. To better understand why antibiotic treatment improves clinical recovery but not fertility, we tracked changes over time in blood leukocyte gene expression to examine circulating immune cell function during the resolution of metritis in antibiotic-treated and untreated cows. Twenty-three primiparous Holstein cows diagnosed with metritis (n = 11) or classified as healthy controls (n = 12) between 7 and 10 d postpartum were enrolled. Cows were assigned to receive systemic ceftiofur (1.25 g/d for 3 d) or remain untreated in a 2 × 2 factorial design. Blood samples were collected at diagnosis (wk 1), and at 1 and 3 wk after treatment (wk 2 and wk 4) postpartum. The blood leukocyte RNA was isolated and sequenced and differentially expressed genes were analyzed for functional enrichment and immune cell type. Healthy cows exhibited enrichment of B cell activation and receptor signaling pathways, whereas metritis cows showed upregulation of inflammatory and innate immune responses, including monocyte and neutrophil markers. In untreated cows (healthy versus metritis), transcriptomic differences in blood leukocytes were resolved by wk 4 postpartum. In contrast, antibiotic-treated metritis cows displayed a distinct immune trajectory. At wk 2 postpartum (one wk after antibiotic treatment), cytotoxic T cell and natural killer cell gene expression signatures increased, differing clearly from untreated metritis cows. By wk 4, a second wave of inflammatory cell-associated differentially expressed genes emerged exclusively in antibiotic-treated metritis cows, suggesting prolonged immune cell activation. These findings demonstrate that antibiotic treatment reshapes systemic immune dynamics during recovery from metritis. The altered sequence (e.g., early cytotoxic activation followed by extended inflammation) may explain why antibiotic-induced clinical cure in metritis cows does not consistently translate into improved fertility.

Steps to reproduce

The MU Genomics Technology Core performed RNA library preparation and sequencing. Stranded RNA-seq libraries were constructed using the poly-A enrichment method with the Illumina Stranded mRNA Prep (Illumina, San Diego, CA, USA). Each purified library was sequenced on a NovaSeq 6000 sequencer (Illumina) using a single NovaSeq S2-PE100 flow cell with an average read count of approximately 60 million paired-end reads per sample (range: 46–78 million). Raw paired-end reads (2 × 100 bp) were trimmed using cutadapt (v3.2; Martin, 2011). Illumina TruSeq adapter sequences were removed from the 3′ ends of R1 (AGATCGGAAGAGCACACGTCTGAACTCCAGTCA) and R2 (AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT) reads. Poly-G tails arising from empty flow cell tiles on the NovaSeq 6000 platform were trimmed from both reads, and ambiguous N bases were removed from read termini. Reads shorter than 10 bp following trimming were discarded. Ribosomal RNA content was low across all samples (average 2.4%; range: 1.0–4.8%), confirming effective poly-A enrichment. A genome index was built from the Ensembl Bos taurus ARS-UCD1.2 reference genome using HISAT2-build (v2.1.0) with default parameters. Trimmed paired-end reads were then aligned to this index using HISAT2 [(v2.1.0; (Kim et al., 2015)] with default parameters, achieving an average alignment rate of 94.1% per sample (range: 92.0–95.5%). Subread's featureCounts [v2.0.0; (Liao et al., 2014)] was used to quantify reads mapped to exon sequences of annotated genes using the Ensembl Bos taurus ARS-UCD1.2 gene annotation file (release 102), with paired-end mode enabled (-p) and default parameters for all remaining settings. An average of 66.4% of aligned reads per sample were successfully assigned to annotated genes (range: 62.0–70.8%), corresponding to an average of 42.2 million read pairs per sample. The resulting count matrix was then filtered using the filterByExpr function with default parameters and normalized using the trimmed mean of M-values (TMM) method. Differentially expressed genes (DEG) were identified using edgeR [v3.36.0; (Chen et al., 2016)] by fitting a negative binomial generalized linear model with quasi-likelihood F-tests and robust dispersion estimation. A transcript was considered differentially expressed if the FDR-adjusted P-value was < 0.05.

Institutions

• University of Missouri

  Missouri, Columbia

Categories

Funders

• Eunice Kennedy Shriver National Institute of Child Health and Human Development

  National Institutes of Health

  Bethesda

  Grant ID: R01HD092254

Licence