We have employed a serum peptide profiling based approach to identify serum peptide biomarkers that discriminate PE and healthy pregnant controls. 52 significant peptide biomarkers from 14 protein precursors were found and a 19-peptide biomarker panel was constructed which can diagnose PE with great sensitivity and specificity.
The differential 52 serum peptides are derived from proteins known to be involved in the pathophysiology of PE, e.g. A1AT, APO-L1, FGA, ITIH4, KNG1, SERPINA1 in acute inflammatory and defense response; APO-A4, APO-C3, APO-E, and APO-L1 in lipid metabolism; C3, C4A, FGA, and SERPINA1 in the activation of complement and coagulation responses. This might reflect the nature of PE as a multi-factorial disorder with complicated pathophysiological changes. However, little is known about the function of these peptide fragments, including their possible biological activity.
For both systemic and renal diseases, we previously hypothesized 
that naturally occurring biofluid peptide biomarkers can be the surrogates of pathophysiologies in signaling, proteolytic, and anti-proteolytic pathways. Sequence alignment analyses () of these peptides found that FGA peptides line up by forming clusters (n
4) within either the N- or C-terminal end with ladder-like truncations at the opposite ends, suggesting that there is likely disease-specific proteolytic degradation of the parent protein. The peptide biomarkers can be the derivatives of serological proteins, disease specific shedding from other organs, and/or renal-specific proteins, all of which are generated during the proteolysis that occurs in either circulation during systemic diseases or dysfunctional kidneys, and then trimmed down by exoproteases into ladder-like clusters. The discovery of the serum peptide biomarkers for PE supports the notion that PE pathophysiology or pathogenesis can lead to serum specific protein degradation patterns throughout the progression of the disease from early to late gestation. Moreover, our 19-peptide panel predicted well with comparable sensitivity and specificity at either early or late gestational age weeks, indicating its potential utility throughout the disease course and potentially in early onset of PE. This is in contrast to the established use of the sFlt-1/PIGF ratio 
, which works better in early onset but does not have sufficient statistical power to accurately predict late-onset PE.
Interestingly, we have found an ITIH4 peptide (LLGLPGPPDVPDHAAYHPF) as a PE biomarker. This peptide shares an almost identical sequence as a previously published spontaneous preterm birth (SPB) serum peptide biomarker (QLGLPGPPDVPDHAAYHPF) 
but there is a preceding amino acid sequence change from L to Q 
. Close examination of a database of common gene variations (http://snp.ims.u-tokyo.ac.jp/cgi-bin/SnpInfo.cgi?SNP_ID=IMS-JST073530
) revealed that this change is due to the single nucleotide polymorphism (SNP) in ITIH4 where a single coding nucleotide differs from A of amino acid codon cAa to T of cTa, resulting in an amino acid change from Q to L. The exact biological function of ITIH4 and its degraded serum peptide is unknown. Given that the same ITIH4 peptide is a biomarker of both PE and SPB, it is very likely that this is not a disease-process-related biomarker as PE and SPB have very different pathophysiologies.
We also recognize several limitations to our study. Proteomic profiling data were acquired from a commercial vendor with little specific information on the clinical characteristics including blood pressure at the time of delivery, baseline blood pressure, birth weight, level of proteinuria which are data that one would normally see in a study on preeclampsia. Current analysis can only be of confirmative diagnostic rather than predictive values. Samples at asymptomatic stages of pregnancy (i.e. at earlier time points) should have been examined to study the predictive value of the panel. Samples from women with other hypertensive disorders of pregnancy would be required to see if the panel differentiates between these and PE. Both of these aspects would be clinically relevant. The former in order to target intensive monitoring and preventative strategies to those at risk, the latter in order to target therapy (i.e. delivery of the baby) to those with PE whereas women with other hypertensive disorders could potentially continue with their pregnancy. In addition, there is heavy bias towards African American, Asian and Hispanic ethnicities. Robust prospective analysis of this 19-peptide panel in sufficiently powered independent samples would still be mandatory to validate this panel's clinical usefulness in PE diagnosis.
We proposed that serum peptidome biomarker analysis might be useful in diagnosing PE, however, the challenges in developing cleavage site-specific antibodies and a resultant ELISA for these peptide biomarkers make it difficult for translation into a point-of-care antibody-based assay. Technologic advances in multiple reaction monitoring (MRM) 
coupled with stable isotope dilution (SID) mass spectrometry (MS) have empowered a “universal” approach to perform quantitative assays for peptides with minimum restrictions, and the ease of assembling multiplex peptide detections in a single measurement. Using common materials and standardized protocols, the reproducibility and transferability of MRM assays between laboratories and across instrument platforms have been demonstrated 
. Therefore, in a similar fashion as the current common practice of applying MRM based newborn screening of metabolic diseases, a greater acceptance by the clinical community of SID-MRM-MS technology as a generally applicable approach for biofluid protein and peptide quantification is expected. We believe a future prospective trial of our serum peptide PE biomarker panel, using SID-MRM-MS, will lead to a quick and reliable multiplexed test which can be run routinely in the hospital setting for PE care.