Inflammation is the first line defense mechanism against infection, tissue damage or cancer development. However, inappropriate inflammatory response may also trigger diseases. The quantification of inflammatory proteins is essential to distinguish between harmful and beneficial immune response. Currently used immunoanalytical assays may suffer specificity issues due to antigen–antibody interaction and possible cross-reactivity of antibody with other protein species. In addition, immunoanalytical assays typically require invasive blood sampling and additional logistics; they are relatively costly and highly challenging to multiplex. We present a multiplex assay based on selected reaction monitoring (SRM) for quantification of seven acute phase proteins (i.e. SAA1, SAA2-isoform1, SAA4, CRP, A1AT isoform1, A1AG1, A1AG2) and the adaptive immunity effector IGHA1 in dried blood spots (DBS). This type of samples is readily available from all human subjects including newborns. The study utilizes proteotypic isotopically labeled peptides with trypsin-cleavable tag and presents optimized and reproducible workflow and several important practical remarks regarding quantitative SRM assays development. The panel of inflammatory proteins was quantified with sequence specificity capable to differentiate protein isoforms with intra- and inter-day precision (<16.4% CV; <14.3% CV respectively). Quantitative results were correlated with immunonephelometric assay (typically >0.9 Pearson's R).

2.3 Sample standardization and trypsin enzymatic digestion A 3 mm DBS punches from 15 study subjects were extracted using 150 ul of 50 mM ammonium bicarbonate buffer (ABB) and 100 ul aliquots of each sample were pooled together to generate a QC sample. The total protein concentration in DBS extracts was determined using conventional protein assays (i.e. Bradford and BCA). Bradford reagent was prepared according to original publication38. Coomasie Brilliant Blue G-250 (cat. #20279) purchased from Sigma Aldrich (Merck, Darmstadt, Germany), phosphoric acid 85% (cat. #19200) purchased from Penta (Prague, Czech Republic) and ethanol for gas chromatography (cat. #102371) purchased from Merck (Darmstadt, Germany). BCA protein assay (Thermo Fisher, cat. #23227) was used according to manufacturer protocol. Dilution series of albumin was prepared in 50 mM ABB and used to construct calibration curves. The dilution series was prepared at seven concentration levels in the range of 31.3 - 2000 mg/l and at six concentration levels in the range of 31.3 - 1000 mg/l in case of BCA assay and Bradford assay, respectively. Both assays were performed on 96-well plate format where 200 l of Bradford or BCA reagent was added to 25 l of standard solution or pooled QC sample. Spectral absorbance measured at wavelengths of 562nm and 595 nm for BCA and Bradford assay respectively. The relative total protein concentration in DBS extract was determined at five dilution levels (10-, 20-, 50-, 80- and 100-fold) to accurately determine total protein concentration and to adjust the amount of trypsin (Trypsin Gold, Mass Spectrometry Grade, Promega, WI, USA) to maintain constant trypsin to protein ratio of approx.1:50. Trypsin enzymatic digestion performed at 37 °C in total volume of 43 l in a Para film sealed micro centrifuge Eppendorf tube (0.5 ml, cat. # 0030108094). The time-course optimization of enzymatic digestion (4, 6, 8, 16 and 24 h) for representative peptides detectable in QC sample shown in Figure S-1. 2.4 Sample processing of dried blood spots A typically used DBS punches measure 1/8” (≈3.2 mm) in diameter and contain approx. 3.2 ul of whole blood39. We used metric 3 mm DBS puncher, thus assuming 3 µL of whole blood per analysis. A single DBS punch placed into a micro centrifuge Eppendorf tube (1.5 ml, cat. # 0030108116) was extracted adding 150 l of buffer. In this study, we used two buffers for comparative analysis i) 50 mM ABB and ii) 50 mM ABB with 5 g/l sodium deoxycholate (SDC). The micro centrifuge tube with DBS punch and extraction buffer was placed onto a platform shaker (200 rpm, 1 h, RT). An aliquot (30 l) of the extract was transferred into a clean micro centrifuge tube and spiked with 10 l of TQLTM Peptides Mixture, concentrations are listed in Table S-2. Subsequently, 3 l of trypsin stock solution (1 g/l) were added to the DBS extract maintaining trypsin-to-total protein ratio of approx. 1:50 for reproducible enzymatic digestion of proteins. Finally, the micro centrifuge tube was sealed with a paraffin film and placed into an incubator (200 rpm, 16 h, 37°C). Enzymatic digestion was terminated adding 300 l of 2% FA in water pH<3. The digested sample was desalted using Bond Elut C18 200 mg cartridges (Agilent Technologies, USA). SPE cartridges were primed according to manufacturer recommendation: 1 ml methanol, 1 ml of 50:50 2% FA in water:ACN (v/v) pH<3, 2 ml of 2% FA in water pH<3. Next, the sample was loaded onto cartridge and washed with 2 ml of 2% FA in water pH<3. Peptides were eluted from the cartridge with 50:50 2% FA in water:ACN (v/v) pH<3. The eluent solvent was removed under a stream of nitrogen, re-dissolved in 50 l 95:5 0.1% FA in water:ACN and analyzed using UHPLC-SRM. 2.5 SRM protein assay design The general strategy of multiplex SRM protein assay development was previously described by authors40. Specifically, protein accession number and protein annotation were according to NextProt database (www.nextprot.org) and the selection of respective proteotypic peptides was under guidance of SRMAtlas compendium (www.srmatlas.org). Surrogate peptides were curated for specific quantification of SAA1/2, SAA1, SAA2-1, SAA4, IGHA1, CRP, A1AT-1, A1AG1, A1AG2 and IGHA1 proteoforms (Table 1). For optimal SRM assay sensitivity, the selection of proteotypic peptides was further restricted to the sequence length between 8 20 amino acids and experimental relative intensity reported in SRMAtlas was considered. Peptides containing cysteine residue were excluded to avoid reduction and alkylation steps. All target peptides were verified for potential occurrence of isoforms and PTMs using NextProt database and double-checked using Peptide uniqueness checker tool in NextProt. Relative frequencies of natural amino acid polymorphism was investigated in dbSNP database (www.ncbi.nlm.nih.gov/projects/SNP/) and only peptide sequences with SNP frequency <5% were further considered for the quantitative assay. An exception is peptide LTGHGAEDSLADQAANK of SAA2-1 with SNP frequency of 17%; functioning as SNP quantifier complementary to peptide GAEDSLADQAANK. At minimum two proteotypic peptides were selected per each protein and their relative intensity was verified experimentally using commercially available stable isotope labeled (SIL) peptides (Spike Tides LTM) from JPT Peptide Technologies GmbH, Berlin, Germany (Table S-1). SIL peptides were spiked into each sample to unambiguously identify native target peptides. In next step, peptides with the highest relative intensity of SRM signal per mole were purchased as SIL peptides with trypsin cleavable tag (TCT), commercially labelled Spike Tides TQLTM (JPT, Berlin, Germany). All sequences of signature SIL and SIL-TCT peptides and their corresponding SRM transitions used for quantification of inflammatory proteins are listed in Supporting Information, Tables S-1, S2 and S3. 2.6 Liquid chromatography and mass spectrometry experimental conditions Samples were injected (5 µL) on UHPLC system (InfinityTM 1260 Agilent Technologies, USA) equipped with a reversed phase analytical column (C18 Peptide CSH; 1.7 m, 2.1 mm i.d. x 100 mm; cat. #186006937) from Waters (Milford, MA). The column was thermostated at 40 °C. Mobile phase flowrate was 0.3 ml/min using buffer A (0.1% FA in water) and buffer B (0.1% FA in ACN). The gradient elution program (0-30.9 min) with re-equilibration step (31-35 min) was set as follows: 0.0 min 5% B; 25 min 30% B; 25.5 min 95% B; 30.9 min 95% B; 31 min 5% B; 35 min 5% B. The UHPLC system was coupled to a triple quadrupole mass spectrometer Agilent 6495A (Agilent Technologies, USA) with a standard-flow Jet Stream electrospray source operated in positive ion mode with capillary voltage of 3.5 kV. Additional ion source parameters were as follows: gas flow rate 11 L/min at 130 °C, sheath gas pressure 25 PSI at 400 °C, and nozzle voltage 500 V. Data acquisition was in dynamic SRM mode with a scheduled retention time window of 1.5 min centered on experimentally determined retention time of each peptide. For selectivity sufficient for unambiguous peptide identification, 3 5 SRM qualifier transitions were monitored per peptide and a single best performing SRM transition was used for quantification. The corresponding set of SRM transitions was recorded for both native and isotopically labelled peptide standards, i.e. 84 transitions in total were monitored per analysis (Table S-3). 2.7 SRM protein assay quantitative performance The matrix-matched calibration curves were generated adding SIL-TCT peptides into DBS extracts to determine the linearity range, limit of detection (LOD) and limit of quantification (LOQ). The concentration range covered protein concentrations expected in DBS extracts, typically across 1-3 orders of magnitude. A dilution series consisted of 6-10 concentration levels, each measured in 3-5 technical replicates. Peptide calibration standards were combined into two separate mixed solutions according to their abundance in blood and available concentration of SIL-TCT peptides stock solution (10 M). Mixture 1 (5-plex) contained SIL-TCT peptide surrogates A1AT-1, CRP, SAA1, SAA2-1 and SAA4 proteins and the dilution series at up to 9 concentration levels covered up to 3 orders of magnitude, due to rapid changes in blood abundance of CRP and SAA. Mixture 2 (3-plex) contained SIL-TCT peptides corresponding to high abundant serum proteins (i.e. IGHA1, A1AG1 and A1AG2) and in this case calibration curve was normalized to native peptides present in DBS extract. Mixture 2 dilution series consisted of up to 10 concentration levels due to high abundance of these proteins. QC samples used to generate matrix-matched calibration curves were prepared by extracting DBS punches from all 15 individuals using two different extraction solvents for comparison: i) 150 l of 50 mM ABB and ii) 150 l of 50 mM ABB with 5 g/l SDC. An aliquot (100 l) of each extract was pooled together and 30 l of pooled extract was spiked with 10 l a mixture of SIL-TCT peptides at various concentration levels to cover concentration ranges listed in Table 2 and Table S-4 for extraction to ABB buffer and ABB/SDC buffer respectively. Samples were further processed as described previously in section 2.4 i.e. subjected to enzymatic digestion, SPE and UHPLC-SRM analysis. The average molecular weight of a protein (Table 2) was based on sequence listed in UniProt database, molecular weight of SAA1/2 was an average of SAA1 and SAA2 protein sequences. 2.8 SRM protein assay data analysis Skyline software (Version 3.6.0.10162; MacCoss Lab, Uni. of Washington, WA, USA) was used to generate and modify SRM method, for qualitative analysis and data visual inspection. Quantitative results (i.e. integrated peak areas) were produced using commercial MassHunter Quantitative Analysis software (Agilent Technologies, USA) and further processed in Microsoft Excel (Microsoft Office Professional Plus 2013, USA). Statistical analyses were performed in software Statistica (Version 13.0.159.8). 2.9 Immunonephelometric determination of proteins Freshly frozen serum samples collected according to a protocol described in section 2.2 were transferred to St. Anne´s hospital, Department of Laboratory Immunology and briefly stored at 80 °C until analysis. Proteins of SAA family (without further isoform specification) were analyzed in serum by particle-enhanced immunonephelometry technique on BN II system (Siemens Healthcare GmbH, Munich, Germany) utilizing Siemens N Latex SAA kit (cat. #OQMP11) and SCS cleaner (cat. #OQUB19). The remaining serum inflammatory proteins (i.e. CRP, A1AT and A1AG) and IGHA1 were measured on immunonephelometer Immage 800 (Beckman Coulter, Pasadena, CA) using commercial kits (Beckman Coulter, Cal1 lot. #M504730, Cal2 lot #M501048, Cal5 Plus #M501548) and quality control samples (Beckman Coulter, Vigil Protein Controls, lot #M506601, lot #M506602 and Vigil Serology Controls, lot #M507512, #M507513) according to a standard protocol from the manufacturer. The fully automated analysis was done on calibrated instrument after running commercial QC samples.

Dried blood spots (DBS) and serum samples were collected during a single morning session from 15 healthy volunteers (8 women and 7 men) with approval of the Ethics Committee of CELSPAC: TNG (CELSPAC/EK/4/2016) at University Hospital Brno. Capillary blood DBS was collected using a lancet from the subject’s finger and captured on Whatman 903 protein saver cards. Generated blood spots were air-dried at room temperature for 3 h, consequently stored at 80 °C until analysis. Serum samples and whole blood samples were from blood collected by venipuncture into 9 ml serum tubes and blood tubes with EDTA. Blood was immediately aliquoted and stored at 80 °C until analysis. Collected blood in serum tubes was clotted and then centrifuged (10 min, 2500g, 20 °C). Serum was aliquoted (500 ul), immediately frozen and stored at 80 °C until analysis. The serum sample from an individual #14 was not possible to collect due to inaccessible peripheral veins.