Improved Methods for Oxytocin Analysis from Human Serum and Urine by Orbitrap ESI‐LCMS after SPE Purification

Abstract

IntroductionNative oxytocin (OT) levels in non‐pregnant/non‐lactating/non‐medicated (NPLM) humans are very low. The limit of detection (LOD, 10–25 pg/mL) of our previous liquid chromatography mass spectrometry (LCMS) method precluded their quantification in serum and urine. Thus, we sought to improve the LCMS sensitivity of OT measurements in these matrices by hydrophobic tagging or optimizing solid phase extraction (SPE) procedures to selectively concentrate native OT.MethodsOT was reduced then alkylated with N‐alkyl acetamide (C12, C14, C16, C18) tags or directly tagged with sulfonyl chloride reagents. Native serum OT was crashed with acetonitrile (ACN), extracted using a Sep‐Pak C18 cartridge washed 4× with increasing ACN concentrations to remove interferences, eluted with ACN, then concentrated by a flow of nitrogen gas. LCMS was performed using a Kinetex C18 column (150 × 3 mm, 2.6 μm) and Q‐Exactive orbitrap MS. Native urinary OT was crashed with ACN then extracted using a Strata‐X SPE cartridge washed 5× by gravity with increasing ACN concentrations. OT was eluted with ACN then concentrated online using a Strata‐X SPE (20 × 2.0 mm, 25 μm) coupled to a Kinetex C18 column followed by LCMS analysis using a Q‐Exactive orbitrap MS operated in ESI+ mode with targeted SIM scan. Quantification of the analytes was set within 5 ppm of the calculated exact isotope masses ([M+2H]2+: OT: 504.22551; OT‐d5: 506.74120).ResultsAlkylation with C12 and C14 yielded a di‐ and mono‐capped product, respectively, while C16 and C18 failed to react. Reactions of native OT with biphenyl sulfonyl chloride resulted in mono‐tagged adducts whereas reaction with 1‐methylimidazole sulfonyl chloride lead to mono‐ and bi‐tagged adducts in a 1:1 ratio. None of these products had better sensitivity than native OT. Thus, we decided to focus on native OT. The LOD of spiked serum using our SPE protocol was 2.5 pg/mL. The LOD in neat OT standards and spiked urine using our manual SPE and online concentrating protocol was 0.25 pg/mL and 1–5 pg/mL, respectively, depending on background noise. Applying this modified SPE method, we could detect OT consistently in one (1.6 pg/mL) and inconsistently in two lactating women’s urine.ConclusionWe were unable to improve the sensitivity of OT measurements using hydrophobic tags as previously described1. Our findings were confirmed by the Max Planck Institute for Psychiatry, (Munich, Germany), that performed the reaction on aq. OT and [Arg8]‐VP with the C12 and C14 reagents we synthesized followed by TOF based LCMS which revealed a 10‐fold lower sensitivity compared to native OT. Therefore, we chose to measure native OT after SPE purification to lower the LOD in serum and urine to allow measurement of the very low native OT (1–2 pg/mL) levels found in NPLM humans. We found that manual SPE was mandatory to improve sensitivity. Using online SPE alone, the LOD in spiked urine was > 20 pg/mL and resulted in more interferences. Modifying the previous SPE protocol for serum and urine extraction lowered the LOD to 2.5 pg/mL and 1–5 pg/mL, respectively. The increased sensitivity for urinary OT allowed OT detection in some but not all lactating women’s urine at levels lower than reported by ELISA.Support or Funding InformationP30 CA71789