Search

Supplementary Table S2. Scaffold 'Sample view' of all proteins and protein clusters identified in protein band 1 to 5.

Supplementary Figure 1. Standard curve of rabbit monoclonal antibody to recombinant survivin using DAC-ELISA (A) and sandwich ELISA (B). Supplementary Figure 2. Antibody response to recombinant survivin was measured in seven NSCLC sera previously reported to be positive for survivin antibodies and in seven healthy non-smoking control sera. Supplementary Figure 3. Standard curve of rabbit anti-survivin spiked in non-smoking control sera using sandwich ELISA. Supplementary Figure 4. Standard curve of rabbit anti-survivin spiked in NSCLC patient sera using sandwich ELISA. Supplementary Figure 5. Standard curve of human autoantibodies to HuD using sandwich ELISA. Supplementary Figure 6. Results of sandwich ELISA were confirmed by western blot analysis.

The high mortality rate in lung cancer is largely attributable to late diagnosis. Case–control studies suggest that autoantibodies to the survivin protein are potential biomarkers for early diagnosis. We tested the hypothesis that sandwich ELISA can detect autoantibodies to survivin before radiologic diagnosis in patients with early-stage non–small cell lung cancer (NSCLC). Because previous studies assayed survivin autoantibodies with the direct antigen-coating ELISA (DAC-ELISA), we first compared that assay with the sandwich ELISA. Based on the more robust results from the sandwich ELISA, we used it to measure survivin autoantibodies in the serum of 100 individuals from a well-controlled population study [the Dutch–Belgian Lung Cancer Screening Trial (NELSON) trial] composed of current and former smokers (50 patients with NSCLC, both before and after diagnosis, and 50 matched, smoking-habit control subjects), and another 50 healthy nonsmoking control subjects. We found no difference in specific autoantibodies to survivin in NSCLC patients, although nonspecific median optical densities were 24% higher (P < 0.001) in both NSCLC patients and smokers, than in healthy nonsmokers. Finally, we confirmed the ELISA results with Western blot analysis of recombinant and endogenous survivin (HEK-293), which showed no anti-survivin reactivity in patient sera. We conclude that specific anti-survivin autoantibody reactivity is most likely not present in sera before or after diagnosis. Autoantibody studies benefit from a comparison to a well-controlled population, stratified for smoking habit. Cancer Immunol Res; 4(2); 165–72. ©2015 AACR.

Lung cancer has the highest mortality rate among cancer patients in the world, in particular because most patients are only diagnosed at an advanced and non-curable stage. Computed tomography (CT) screening on high-risk individuals has shown that early detection could reduce the mortality rate. However, the still high false-positive rate of CT screening may harm healthy individuals because of unnecessary follow-up scans and invasive follow-up procedures. Alternatively, false-negative and indeterminate results may harm patients due to the delayed diagnosis and treatment of lung cancer. Non-invasive biomarkers, complementary to CT screening, could lower the false-positive and false-negative rate of CT screening at baseline and thereby reduce the number of patients that need follow-up and diagnose patients at an earlier stage of lung cancer. Lung cancer tissue generates lung cancer-associated proteins to which the immune system might produce high-affinity autoantibodies. This autoantibody response to tumor-associated antigens starts during early-stage lung cancer and may endure over years. Identification of tumor-associated antigens or the corresponding autoantibodies in body fluids as potential non-invasive biomarkers could thus be an effective approach for early detection and monitoring of lung cancer. In this review we provide an overview of differentially expressed protein, antigen and autoantibody biomarkers that combined with CT imaging might be of clinical use for early detection of lung cancer.

Sera from lung cancer patients contain antibodies against tumor-associated antigens. Specific amino acid sequences of the complementarity-determining regions (CDRs) in the antigen-binding fragment (Fab) of these antibodies have potential as lung cancer biomarkers. Detection and identification of CDRs by mass spectrometry can significantly be improved by reduction of the complexity of the immunoglobulin molecule. Our aim was to molecular dissect IgG into κ and λ fragments to reduce the complexity and thereby identify substantially more CDRs than by just total Fab isolation. We purified Fab, Fab-κ, Fab-λ, κ and λ light chains from serum from 10 stage I lung adenocarcinoma patients and 10 matched controls from the current and former smokers. After purification, the immunoglobulin fragments were enzymatically digested and measured by high-resolution mass spectrometry. Finally, we compared the number of CDRs identified in these immunoglobulin fragments with that in the Fab fragments. Twice as many CDRs were identified when Fab-κ, Fab-λ, κ and λ (3330) were combined than in the Fab fraction (1663) alone. The number of CDRs and κ:λ ratio was statistically similar in both cases and controls. Molecular dissection of IgG identifies significantly more CDRs, which increases the likelihood of finding lung cancer-related CDR sequences.

In cancer and autoimmune diseases, immunoglobulins with a specific molecular signature that could potentially be used as diagnostic or prognostic markers are released into body fluids. An immunomics approach based on this phenomenon relies on the ability to identify the specific amino acid sequences of the complementarity-determining regions (CDR) of these immunoglobulins, which in turn depends on the level of accuracy, resolution, and sensitivity that can be achieved by advanced mass spectrometry. Reproducible isolation and sequencing of antibody fragments (e.g., Fab) by high-resolution mass spectrometry (MS) from seven healthy donors revealed 43 217 MS signals: 225 could be associated with CDR1 peptides, 513 with CDR2 peptides, and 19 with CDR3 peptides. Seventeen percent of the 43 217 MS signals did not overlap between the seven donors. The Fab isolation method used is reproducible and fast, with a high yield. It provides only one Fab sample fraction for subsequent characterization by high-resolution MS. In 17% and 4% of these seven healthy donors, qualitative (presence/absence) and quantitative (intensity) differences in Fab fragments could be demonstrated, respectively. From these results, we conclude that the identification of a CDR signature as biomarker for autoimmune diseases and cancer without prior knowledge of the antigen is feasible.

Late diagnosis of lung cancer is still the main reason for high mortality rates in lung cancer. Lung cancer is a heterogeneous disease which induces an immune response to different tumor antigens. Several methods for searching autoantibodies have been described that are based on known purified antigen panels. The aim of our study is to find evidence that parts of the antigen-binding-domain of antibodies are shared among lung cancer patients. This was investigated by a novel approach based on sequencing antigen-binding-fragments (Fab) of immunoglobulins using proteomic techniques without the need of previously known antigen panels. From serum of 93 participants of the NELSON trial IgG was isolated and subsequently digested into Fab and Fc. Fab was purified from the digested mixture by SDS-PAGE. The Fab containing gel-bands were excised, tryptic digested and measured on a nano-LC-Orbitrap-Mass-spectrometry system. Multivariate analysis of the mass spectrometry data by linear canonical discriminant analysis combined with stepwise logistic regression resulted in a 12-antibody-peptide model which was able to distinguish lung cancer patients from controls in a high risk population with a sensitivity of 84% and specificity of 90%. With our Fab-purification combined Orbitrap-mass-spectrometry approach, we found peptides from the variable-parts of antibodies which are shared among lung cancer patients.