Ultrahigh resolution profiles lead to more detailed serum peptidome signatures of pancreatic cancer

Clinical Perspective on Proteomic and Glycomic Biomarkers for Diagnosis, Prognosis, and Prediction of Pancreatic Cancer

Pancreatic ductal adenocarcinoma (PDAC) is known as a highly aggressive malignant disease. Prognosis for patients is notoriously poor, despite improvements in surgical techniques and new (neo)adjuvant chemotherapy regimens. Early detection of PDAC may increase the overall survival. It is furthermore foreseen that precision medicine will provide improved prognostic stratification and prediction of therapeutic response. In this review, omics-based discovery efforts are presented that aim for novel diagnostic and prognostic biomarkers of PDAC. For this purpose, we systematically evaluated the literature published between 1999 and 2020 with a focus on protein- and protein-glycosylation biomarkers in pancreatic cancer patients. Besides genomic and transcriptomic approaches, mass spectrometry (MS)-based proteomics and glycomics of blood- and tissue-derived samples from PDAC patients have yielded new candidates with biomarker potential. However, for reasons discussed in this review, the validation and clinical translation of these candidate markers has not been successful. Consequently, there has been a change of mindset from initial efforts to identify new unimarkers into the current hypothesis that a combination of biomarkers better suits a diagnostic or prognostic panel. With continuing development of current research methods and available techniques combined with careful study designs, new biomarkers could contribute to improved detection, prognosis, and prediction of pancreatic cancer.

Bayesian variable selection logistic regression with paired proteomic measurements

We explore the problem of variable selection in a case‐control setting with mass spectrometry proteomic data consisting of paired measurements. Each pair corresponds to a distinct isotope cluster and each component within pair represents a summary of isotopic expression based on either the intensity or the shape of the cluster. Our objective is to identify a collection of isotope clusters associated with the disease outcome and at the same time assess the predictive added‐value of shape beyond intensity while maintaining predictive performance. We propose a Bayesian model that exploits the paired structure of our data and utilizes prior information on the relative predictive power of each source by introducing multiple layers of selection. This allows us to make simultaneous inference on which are the most informative pairs and for which—and to what extent—shape has a complementary value in separating the two groups. We evaluate the Bayesian model on pancreatic cancer data. Results from the fitted model show that most predictive potential is achieved with a subset of just six (out of 1289) pairs while the contribution of the intensity components is much higher than the shape components. To demonstrate how the method behaves under a controlled setting we consider a simulation study. Results from this study indicate that the proposed approach can successfully select the truly predictive pairs and accurately estimate the effects of both components although, in some cases, the model tends to overestimate the inclusion probability of the second component.

Adapting censored regression methods to adjust for the limit of detection in the calibration of diagnostic rules for clinical mass spectrometry proteomic data

In this paper, we consider the problem of calibrating diagnostic rules based on high-resolution mass spectrometry data subject to the limit of detection. The limit of detection is related to the limitation of instruments in measuring low-concentration proteins. As a consequence, peak intensities below the limit of detection are often reported as missing during the quantification step of proteomic analysis. We propose the use of censored data methodology to handle spectral measurements within the presence of limit of detection, recognizing that those have been left-censored for low-abundance proteins. We replace the set of incomplete spectral measurements with estimates of the expected intensity and use those as input to a prediction model. To correct for lack of information and measurement uncertainty, we combine this approach with borrowing of information through the addition of an individual-specific random effect formulation. We present different modalities of using the above formulation for prediction purposes and show how it may also allow for variable selection. We evaluate the proposed methods by comparing their predictive performance with the one achieved using the complete information as well as alternative methods to deal with the limit of detection.

Accounting for isotopic clustering in Fourier transform mass spectrometry data analysis for clinical diagnostic studies

Mass spectrometry based clinical proteomics has emerged as a powerful tool for high-throughput protein profiling and biomarker discovery. Recent improvements in mass spectrometry technology have boosted the potential of proteomic studies in biomedical research. However, the complexity of the proteomic expression introduces new statistical challenges in summarizing and analyzing the acquired data. Statistical methods for optimally processing proteomic data are currently a growing field of research. In this paper we present simple, yet appropriate methods to preprocess, summarize and analyze high-throughput MALDI-FTICR mass spectrometry data, collected in a case-control fashion, while dealing with the statistical challenges that accompany such data. The known statistical properties of the isotopic distribution of the peptide molecules are used to preprocess the spectra and translate the proteomic expression into a condensed data set. Information on either the intensity level or the shape of the identified isotopic clusters is used to derive summary measures on which diagnostic rules for disease status allocation will be based. Results indicate that both the shape of the identified isotopic clusters and the overall intensity level carry information on the class outcome and can be used to predict the presence or absence of the disease.

Application of a Serum Protein Signature for Pancreatic Cancer to Separate Cases from Controls in a Pancreatic Surveillance Cohort

BACKGROUND

Pancreatic cancer (PC) surveillance is currently offered to individuals with a genetic predisposition to PC, but routinely used radiological screening modalities are not entirely reliable in detecting early-stage PC or its precursor lesions. We recently identified a discriminating PC biomarker signature in a sporadic patient cohort. In this study, we investigated if protein profiling can accurately distinguish PC from non-PC in a pancreatic surveillance cohort of genetically predisposed individuals.

METHODS

Serum samples of 66 individuals with a CDKN2A germline mutation who participated in the pancreatic surveillance program (5 cases, 61 controls) were obtained following a standardized protocol. After sample clean-up, peptide and protein profiles were obtained on an ultrahigh-resolution matrix-assisted laser desorption/ionization-Fourier transform ion cyclotron resonance mass spectrometry platform. A discriminant score for each sample was calculated with a previously designed prediction rule, and the median discriminant scores of cases and controls were compared. Individuals with precursor lesions of PC (n = 4) and individuals with a recent diagnosis of melanoma (n = 4) were also separately considered.

RESULTS

Cases had a higher median discriminant score than controls (0.26 vs 0.016; P = .001). The only individual with pathologically confirmed precursor lesions of PC could also be clearly distinguished from controls, and having a (recent) medical history of melanoma did not influence the protein signatures.

CONCLUSIONS

Peptide and protein signatures are able to accurately distinguish PC cases from controls in a pancreatic surveillance setting. Mass spectrometry-based protein profiling therefore seems to be a promising candidate for implementation in the pancreatic surveillance program as an additional screening modality.

Typing Pseudomonas aeruginosa Isolates with Ultrahigh Resolution MALDI-FTICR Mass Spectrometry

The introduction of standardized matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) platforms in the medical microbiological practice has revolutionized the way microbial species identification is performed on a daily basis. To a large extent, this is due to the ease of operation. Acquired spectra are compared to profiles obtained from cultured colonies present in a reference spectra database. It is fast and reliable, and costs are low compared to previous diagnostic approaches. However, the low resolution and dynamic range of the MALDI-TOF profiles have shown limited applicability for the discrimination of different bacterial strains, as achieved with typing based on genetic markers. This is pivotal in cases where certain strains are associated with, e.g., virulence or antibiotic resistance. Ultrahigh resolution MALDI-FTICR MS allows the measurement of small proteins at isotopic resolution and can be used to analyze complex mixtures with increased dynamic range and higher precision than MALDI-TOF MS, while still generating results in a similar time frame. Here, we propose to use ultrahigh resolution 15T MALDI-Fourier transform ion cyclotron resonance (FTICR) MS to discriminate clinically relevant bacterial strains after species identification performed by MALDI-TOF MS. We used a collection of well characterized Pseudomonas aeruginosa strains, featuring distinct antibiotic resistance profiles, and isolates obtained during hospital outbreaks. Following cluster analysis based on amplification fragment length polymorphism (AFLP), these strains were grouped into three different clusters. The same clusters were obtained using protein profiles generated by MALDI-FTICR MS. Subsequent intact protein analysis by electrospray ionization (ESI)-collision-induced dissociation (CID)-FTICR MS was applied to identify protein isoforms that contribute to the separation of the different clusters, illustrating the additional advantage of this analytical platform.

Developments in FTICR-MS and Its Potential for Body Fluid Signatures

Fourier transform mass spectrometry (FTMS) is the method of choice for measurements that require ultra-high resolution. The establishment of Fourier transform ion cyclotron resonance (FTICR) MS, the availability of biomolecular ionization techniques and the introduction of the Orbitrap™ mass spectrometer have widened the number of FTMS-applications enormously. One recent example involves clinical proteomics using FTICR-MS to discover and validate protein biomarker signatures in body fluids such as serum or plasma. These biological samples are highly complex in terms of the type and number of components, their concentration range, and the structural identity of each species, and thus require extensive sample cleanup and chromatographic separation procedures. Clearly, such an elaborate and multi-step sample preparation process hampers high-throughput analysis of large clinical cohorts. A final MS read-out at ultra-high resolution enables the analysis of a more complex sample and can thus simplify upfront fractionations. To this end, FTICR-MS offers superior ultra-high resolving power with accurate and precise mass-to-charge ratio (m/z) measurement of a high number of peptides and small proteins (up to 20 kDa) at isotopic resolution over a wide mass range, and furthermore includes a wide variety of fragmentation strategies to characterize protein sequence and structure, including post-translational modifications (PTMs). In our laboratory, we have successfully applied FTICR “next-generation” peptide profiles with the purpose of cancer disease classifications. Here we will review a number of developments and innovations in FTICR-MS that have resulted in robust and routine procedures aiming for ultra-high resolution signatures of clinical samples, exemplified with state-of-the-art examples for serum and saliva.

Solid-phase extraction strategies to surmount body fluid sample complexity in high-throughput mass spectrometry-based proteomics

For large-scale and standardized applications in mass spectrometry- (MS-) based proteomics automation of each step is essential. Here we present high-throughput sample preparation solutions for balancing the speed of current MS-acquisitions and the time needed for analytical workup of body fluids. The discussed workflows reduce body fluid sample complexity and apply for both bottom-up proteomics experiments and top-down protein characterization approaches. Various sample preparation methods that involve solid-phase extraction (SPE) including affinity enrichment strategies have been automated. Obtained peptide and protein fractions can be mass analyzed by direct infusion into an electrospray ionization (ESI) source or by means of matrix-assisted laser desorption ionization (MALDI) without further need of time-consuming liquid chromatography (LC) separations.