Serum proteomic analysis of sex differences during an acute low back pain episode

Growing evidence suggests that pain and injury engage biological processes and systems that extend beyond the nervous system. The molecular processes driving the transition from acute to chronic low back pain (LBP) remain poorly understood. Here, we explore the serum proteomic profile of male and female participants during an acute LBP episode (N=59). Differential protein expression during the acute-stage of LBP was compared between participants with resolved LBP at three months and those with chronic or recurrent LBP at three-months. Proteome-wide analysis using mass-spectrometry identified 216 proteins confidently. Sex differences in protein abundance changes were evident upon inspection of fold changes. Multivariable data analysis identified 21 serum proteins during the acute episode that correctly classified 93% of males and 23 serum proteins that correctly classified 90% of females with ongoing LBP at three months. Most of the differentially expressed proteins during acute LBP were involved in immune, inflammatory, complement or coagulation responses. Overall, males who recovered from an acute LBP episode had a greater tendency to demonstrate an upregulated immune-driven inflammatory response during acute LBP, whilst women with pain resolution demonstrated a tendency towards a downregulated immune response. Taken together, this study provides data to suggest biological processes during an acute LBP episode may contribute to resolution, or persistence, of LBP symptoms at three months, however, these processes differ between sexes. This work provides an early foundation for future research exploring strategies targeting distinct immune system processes in men and women that may interfere with the transition from acute to chronic LBP. 

Methods
Sample preparation
Peripheral venous blood was drawn into serum tubes (BD, SST II Advance) through venepuncture of the median cubital vein at baseline assessment. The sample was clotted (30 min, room temperature) and then separated by centrifugation (2500 rpm, 15 min). Samples were pipetted into 50 μL aliquots and immediately stored at -80°C. After thawing, de-identified serums from all 59 participants were prepared by digesting 3µl of serum (57µg ul-1 +/-7µg) in 50µl of 50mM AMBIC, 2M urea, 10mM DTT at pH 8 using trypsin at 25°C for 16 hours in a 1:100 enzyme to protein ratio. Digestion was halted by acidification. Serum peptides were fractionated using hydrophobic interaction chromatography (HILIC) according to the manufacturer's protocol (PolyLC Inc, MD, USA) with the additional parameter of decreasing solvent-releasing increasingly hydrophilic peptides. Five fractions with decreasing solvent were prepared (unbound fraction, 25%,50%,70%, and 100% sequential solvent extraction) with peptides released by 70% acetonitrile in 15mM ammonium acetate fraction being evaluated further.

Mass spectrometry of samples
Digested and fractionated peptides were reconstituted in 5μL 0.1% formic acid and separated by nano-LC using an Ultimate 3000 HPLC and autosampler (Dionex, Amsterdam, Netherlands). The sample, (0.6μL from 5μl), was loaded onto a micro C18 pre-column (300 μm×5mm, Dionex) with H2O:CH3CN (98:2, 0.1% TFA) at 10μL min−1. After washing, the pre-column was switched (Valco 10 port valve, Dionex) into line with a fritless nanocolumn (75μm id×12 cm) containing reverse phase C18 media (1.9μm, 120 ̊A, Dr. Maisch GmbH HPLC). Peptides were eluted using a linear gradient of H2O:CH3CN (98:2, 0.1% formic acid) to H2O:CH3CN (64:36, 0.1% formic acid) at 250nlmin−1 over 90 min. The QExactive (Thermo Electron, Bremen,Germany) mass spectrometer (MS) was run in DDA mode as previously described (58).

Protein identification relative quantitation
Protein dataset-peak lists were generated from raw files using Mascot Daemon v2.5.1 (Matrix Science, London, UK, www.matrixscience.com). All MS/MS spectra were searched against Swissprot (downloaded February 2018; 556,568 sequences for protein identification with the following criteria: 1) species, Human; 2) allowed 1 missed cleavage; 3) variable modifications, Oxidation (M), phosphorylation (S, T, Y); 4) peptide tolerance, ±5 ppm; 5) Fragment tolerance, ±0.05 Da; 6) peptide charge+2 and +3; and 7) enzyme specificity, semi-tryptic. A decoy database search was also performed. Only proteins identified from the Swissprot database, controlled by the Benjamini-Hochberg procedure for multiple comparisons, with two or more unique peptides were included in further analysis.