alexa Ceramide Profiles of Human Serum Gangliosides GM2 and GD1a exhibit Cancer-associated Alterations | OMICS International
ISSN: 2153-0637
Journal of Glycomics & Lipidomics

Like us on:

Make the best use of Scientific Research and information from our 700+ peer reviewed, Open Access Journals that operates with the help of 50,000+ Editorial Board Members and esteemed reviewers and 1000+ Scientific associations in Medical, Clinical, Pharmaceutical, Engineering, Technology and Management Fields.
Meet Inspiring Speakers and Experts at our 3000+ Global Conferenceseries Events with over 600+ Conferences, 1200+ Symposiums and 1200+ Workshops on
Medical, Pharma, Engineering, Science, Technology and Business

Ceramide Profiles of Human Serum Gangliosides GM2 and GD1a exhibit Cancer-associated Alterations

Stephan Kirsch1, Jamal Souady1, Michael Mormann1,2, Laura Bindila1,3 and JasnaPeter-Katalinić1,4*

1Institute of Medical Physics and Biophysics, University of Münster, Germany

2Institute of Hygiene, University of Münster, Germany

3Institute of Physiological Chemistry, University Medical Center Mainz, Germany

4Department of Biotechnology, University of Rijeka, Croatia

*Corresponding Author:
Jasna Peter-Katalinić
Department of Biotechnology
University of Rijeka, RadmileMatejčić 2, 51000 Rijeka, Croatia
E-mail: [email protected] hr

Received date: October 10, 2012; Accepted date: December 26, 2012; Published date: December 28, 2012

Citation: Kirsch S, Souady J, Mormann M, Bindila L, Peter-Katalinic J (2012) Ceramide Profiles of Human Serum Gangliosides GM2 and GD1a exhibit Cancerassociated Alterations. J Glycomics Lipidomics S2:005. doi: 10.4172/2153-0637.S2-005

Copyright: © 2012 Kirsch S, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Visit for more related articles at Journal of Glycomics & Lipidomics

Abstract

Human serum gangliosides were analyzed by nano-HPLC/MS to characterize their glycoforms and ceramide profiles within a single experiment. This unbiased glycosphingolipidomics approach was successfully applied to a set of clinical samples of gastric and pancreatic cancer patients and related control cohorts. Evaluation of acquired data revealed no clear differences with respect to ganglioside glycoforms among the patient groups investigated. However, significant alterations were observed within the ceramide profiles of different ganglioside structures, particularly pronounced for gangliosides GM2 and GD1a. Based on the increase of palmitic acid-comprising structures and the concurrent decrease of stearic acid-comprising ganglioside signals observed for both glycoforms, the ceramide ratio of these signals was applied for discrimination of cancer patients from control sample cohorts. Based on the performance evaluated in statistical analyses, the GM2-related ceramide ratio is proposed as a potential biomarker in particular for pancreatic cancer. Although further validation is required to determine its capability for cancer detection in mass screenings and its applicability for clinical routine, the data presented here support a promising perspective for applications.

Introduction

Various kinds of cancer, such as pancreatic and gastric carcinomas, are associated with high mortality rates that significantly increase with the disease progression at the time of diagnosis [1]. Due to the lack of clear clinical symptoms caused by the gastrointestinal types of cancer only few patients are diagnosed at early stages of the disease. Accordingly, there is an obvious demand for early diagnostic markers to increase survival rates and enable development of new treatment strategies. Glycoconjugates in general have been associated with tumor development since it was discovered, that they are often expressed as onco-developmental antigens [2]. In fact, a significant number of approved cancer biomarkers are glycoproteins being monitored in the blood for their application in diagnosis, prognosis and/or therapy selection of different types of cancer [3]. Although there is growing evidence that glycosphingolipids (GSLs) play an important role in tumor biology, they are not used yet as tumor markers. The qualitative and quantitative analysis of GSLs in routine set-up lags far behind other types of glycoconjugates or biomolecules, which limited so far their interrogation as biomarker candidates and subsequent validation. GSLs are expressed by almost all mammalian cells and consist of a polar glycan chain attached to a nonpolar ceramide moiety. The carbohydrate chain is free for interactions in the extracellular space, while the ceramide portion anchors the GSLs within the outer leaflet of the cell membrane [4]. Being involved in many developmental and cell-signalling processes [5] GSLs are also known to exhibit altered expression due to cancer development. In particular gangliosides, sialylated GSLs, are over expressed in many tumors [6] or exhibit alterations of certain structural elements such as an increased α2-6- sialylation [7] or elevated fucosylation associated with the metastatic potential of a tumor [8].

Being inserted in the plasma membrane, GSLs can be released from their host cell and taken up by another one or by a carrier protein such as blood lipoproteins [9]. This process is referred to as shedding and uptake and is observed for lipid-anchored molecules in general [10]. Gangliosides released in such processes were reported to have various in vivo effects, eg. a reduction of an immune response [11-13], a decreased threshold for angiogenic signaling [14] and an enhanced response on epidermal growth factors [15]. Accordingly, shed gangliosides are assumed to promote tumor growth and consequently influence prognosis of the disease. Due to this biological context gangliosides could be feasible biomarkers especially as shedding can affect not only the tumor microenvironment but also serological levels of GSLs. Being well accessible for sample collection the use of blood, serum or plasma is the source of choice for disease monitoring. Several studies already demonstrated increased levels of GSL-bound sialic acid in the serum of patients suffering from mammary carcinoma [16], colorectal cancer [17], and melanoma [18]. Targeted monitoring of single ganglioside structures in cancer patient sera revealed increased concentrations of GM3 and GD3 in the presence of melanoma [19] and elevated levels of GD2 in context of retinoblastoma [20]. Typically, the studies so far were either focused on the total ganglioside amount and accordingly, revealed no information about particularly overexpressed GSL-species, or were based on immunological interactions and consequently performed in a targeted manner with low-throughput, monitoring only a few GSL-species being postulated a priori as “tumor-associated”. In order to study serological GSL changes in patients suffering from pancreatic and gastric cancer we employed a recently developed method based on nano-high performance liquid chromatography (nano-HPLC) coupled to electrospray ionization mass spectrometry (ESI-MS) for GSL detection and structural elucidation [21]. This LC/ MS approach enables an unbiased monitoring of a broad spectrum of ganglioside structures with increased throughput. Moreover the analytical setup allows the discrimination of isobaric/isomeric analyte species differing with respect to their glycan chain composition as well as determination of ceramide profiles of individual GSL-glycoforms.

Materials and Methods

Reagents

Sample preparation and analysis were performed using LC/MS grade acetonitrile, methanol (Merck, Germany), chloroform, and water (Sigma-Aldrich, Germany). Ammonium acetate (Merck, Germany), Tris, and CaCl2 (Sigma-Aldrich, Germany) were of analytical grade.

Serum samples

Serum samples were collected from patients suffering from gastric carcinoma and pancreatic adenocarcinoma while serum from healthy individuals and patients suffering from non-malignant diseases of the stomach and the pancreas served as a control. Demographic data of patient cohorts are summarized in table 1. Due to the difference in mean age of the control patient group compared to diseased patient cohorts, all disease-associated differences observed in our data were investigated for any correlation with age to exclude an association to aging processes. All sera were collected and provided by RNTECH (Bucharest, Romania). Samples and patient data were collected and stored in accordance with the guidelines of the local ethical committee, compliant with corresponding European regulations.

Patient group Number Gender Age (Mean ± SD) UICC Cancer stage/ type of benign disease
Female Male
Control (CTR) 38 28 10 30.19 ± 4.4 --------
Pancreatic Cancer (PAN) 17 5 12 59.97 ± 8.86 IA (2), IB (5), IIA (3), IIB (7)
Gastric Cancer (STO) 35 8 27 63.86 ± 8.27 IB (10), II (8); IIIA (6), IIIB (4), IV (7)
Benign diseases pancreas (BP) 8 4 4 54.69 ± 14.45 pancreatitis (7), serous cyst (1)
Benign diseases stomach (BS) 10 2 8 59.01 ± 14.47 stromal tumor (4), active ulcer (3), chronic gastric ulcer (1), gastric schwannoma (1), inflammatory fibroid polyp (1)

Table 1: Demographic data of investigated patient groups. Given are the numbers of patient samples and gender, average and standard deviation for age, subgroups of patients differing in type of benign disease or in cancer stage classified according to the Union of International Cancer Control (UICC).

Sample preparation

For lipid extraction serum aliquots of 300 μl were freeze dried and reconstituted in 1.3 ml of chloroform/methanol/water (30/60/8, v/v/v). Homogenized solutions were sonicated, centrifuged and supernatants were transferred to new vials while the extraction was repeated for the remaining pellets. The obtained extracts were combined and evaporated. Co-extracted phospholipids were depleted by phospholipase C digest. For this purpose, samples were reconstituted in 600 μl digest buffer containing 10 mMTris, 10 mM CaCl2 and 2.5 U Bacillus cereus phospholipase C (Sigma-Aldrich, Germany) and incubated over night at 37°C. Further sample preparation involved two steps of solid-phase extraction (SPE). Reversed-phase SPE was performed with Sep-Pak C18 chromatography material (Waters, France) packed into glass cartridges to column volumes (CV) of 250 μl. After column equilibration, the aqueous digest solutions were applied to the cartridge. The flow-though was reapplied twice in order to ensure efficient binding of analytes. After a washing step with 6 CV of water, GSLs were eluted with 3 CV of methanol. Eluates were evaporated and reconstituted in 600 μl chloroform. The second step of SPE was performed with Iatrobead silica chromatography material (Macherey-Nagel, Germany) packed to CV of 200 μl. Following column equilibration, sample solutions were applied and collected flow-through fractions were reapplied twice in order to enhance analyte binding. Cartridges were washed with 3 CV of chloroform and chloroform/methanol (9/1, v/v), respectively. The elution of analytes was performed in four steps using (A) 3 CV chloroform/methanol (2/1, v/v), (B) 3 CV chloroform/methanol (1/2, v/v), (C) 3 CV chloroform/methanol (1/4, v/v), and (D) a combined step of 3 CV methanol and 3 CV methanol/water (4/1). All fractions were evaporated and reconstituted in 40 μl of methanol. C-fractions contain predominantly gangliosides and were used for following nano- HPLC/MS analysis while A-and B-fractions comprised a set of neutral serum GSLs (data not shown). The applied nomenclature of GSLs follows the recommendations of the IUPAC-IUB Joint Commission of Biochemical Nomenclature [22].

Nano-HPLC

Chromatography was performed with an Ultimate 3000 system (Dionex, Germany) equipped with a nanoscale (75 μm×100 mm, 5 μm) TSK gel amide 80 column (Alltech Grom GmbH, Germany). For nano- HPLC analysis the samples were dissolved in acetonitrile/methanol (7/3, v/v) and injected to the column in an amount corresponding to 1.5 μl of serum. Separation of GSLs was carried out by a linear gradient from 5 mM ammonium acetate in acetonitrile/methanol (99/1, v/v) to 5 mM ammonium acetate in methanol/water (8/2, v/v) as reported previously [21]. The eluate was flow-through admitted to MS analysis via a sheathless interface involving coated silica emitters (New Objective, MA, USA) with an inner diameter of 15 μm.

Mass spectrometry

MS analyses were performed in negative ion mode by a quadrupole time-of-flight (QTOF) mass spectrometer equipped with a nano electrospray ion source (Micromass, UK). Experiments were carried out using electrospray voltages in the range of 1.5-2.0 kV and a sampling cone potential of 60 V while a desolvation gas stream of 150 l/h was applied.

Data processing and statistical analysis

MS data acquisition and processing was performed with the MassLynx software (Micromass, UK). Extracted ion chromatograms (EIC) of detected GSL-ions were generated for individual patient samples and used for subsequent statistical analysis. To normalize data and reduce run-to-run variation absolute intensities were converted into percentage intensities relative to the sum of all monitored ganglioside signals.

The non-parametric Mann-Whitney-U test was chosen for statistical evaluation given that most of the detected GSL-signals showed no Gaussian-/normal-distribution within the different patient groups. Curves for receiver operating characteristics (ROC) were generated to investigate the capability of serum GSL signals for identification of cancer patients and their discrimination from control sample cohorts. All statistical analyses were performed with the PASW statistics 18 (SPSS Inc., IL, USA) software.

Results

Serum gangliosides

It has long been elucidated that neutral GSLs such as monohexyland lactosylceramide as well as globo-series GSLs comprise about 80% of the total serum GSL amount and the remaining 20% represent acidic ganglioside structures [23], i.e. GM3, GM2, GD3, GD1a, GD1b and GT1b with GM3 being the most abundant among these ganglioside species [24,25]. Also sialylated nLc4Cer [24] and GM1 [26] were documented to be present at minor concentrations. However, little is known about the ceramide heterogeneity of individual gangliosides. To generate a primary comprehensive list of serum gangliosides we performed nano-HPLC/MS analysis of ganglioside solutions purified from serum of a control volunteer (cf. Materials and Methods). EICs generated for the most intense signal in the ceramide pattern of each detected ganglioside molecular species (Figure 1A) illustrate the relative abundance of individual gangliosides. Gangliosides LC/MS separation was mainly influenced by the length of the glycan chain and the number of included sialic acids. Less polar GSLs such as GM3 (m/z 1151.74) and GM2 (m/z 1382.86) comprising of three and four carbohydrates, respectively, eluted with low retention times, while the last eluting ganglioside (m/z 1707.13, doubly charged) was composed of 16 carbohydrates. This structure represents the most complex GSL that we found in human serum and was to our knowledge for the first time detected in human serum. A few m/z-values gave rise to two EIC peaks. As shown previously, such elution profiles originate from a stronger retention of α2-6-sialylated gangliosides compared to α2- 3-sialylated isoforms and/or from differences in fucosylation [21]. In case of GD1 ions at m/z 917.49, which do not match one of these structural motifs, separated analytes were identified as GD1a (44 min) and the differentially sialylated isoform GD1b (46 min) by tandem- MS analyses (data not shown). With respect to detected ceramide profiles two different subgroups of analytes can be discriminated. Palmitic acid (C16:0)-containing ceramide represents the most abundant lipid structure in almost all serum gangliosides. For GM2 (m/z 1382.86), GD1 (m/z 917.49) and GT1 (m/z 1074.05) structures however, the stearic acid (C18:0)-containing ceramide is the most abundant. Associated mass spectra are displayed for the predominant ganglioside species of each group, GM3 and GD1a, in figure 1B and 1C, respectively. GM3 species contain a large set of different fatty acids ranging from C12:0 to C24:0 within their ceramide moieties where C16:0-ceramides give rise to the most intense signals, followed by C24:1, C18:0, C22:0 fatty acid-comprising structures. In contrast, GD1a exhibits a ceramide profile with fatty acid chain lengths varying between C16:0 and C24:1 where the stearic acid-containing ceramide structure is most dominant. A complete list of GSL-related ions detected in nano-HPLC/MS analysis of human serum gangliosides is given in table 2 along with retention times and structural assignments. In general the serum ganglioside composition as derived from our analysis concurs very well with the data published in literature so far. Yet, the serological GSL profile is furthered particularly with respect to the ceramide pattern occurring among each ganglioside structure. Assignment of structures to the neolacto-or ganglio-series was based on immunostaining data obtained in a parallel study. In this context, also GM1a could be identified although at lower levels as compared to its isobaric structures of IV3/6Neu5Ac-nLc4Cer (data not shown). The detected amounts of GD3 could not be monitored in an appropriate way due to the coeluting impurities covering a close m/z-range. Data derived within the screening of patient samples revealed a huge intensity variation, and accordingly we decided not to consider GD3- related signals for data normalization or statistical analysis.

glycomics-lipidomics-serum-gangliosides

Figure 1: (-) Nano-HPLC/MS analysis of human serum gangliosides.
(A) Display of extracted ion chromatograms (EIC) calculated for the most intense signal in the ceramide pattern of individual ganglioside glycoforms. Insert zooms into a set of less abundant analyte ions (upper line), a detailed list of detected ions is given in Table 2.
(B) Mass spectrum of GM3 ions eluting at min 35.
(C) Mass spectrum of GD1a ions corresponding to an elution time of 44 min. doubly charged ions are marked by#, asterisks indicate partially coeluting VI3Neu5Ac-nLc4Cer species.

m/z-value m/z-value Ion type Retentiontime [min] Structure assignment
(exp.) (theo.)    
1. Peak 2. Peak
   
       
          PI(34:2)
833.56 833.52 [M-H] 31.26
835.54 835.53 31.81 PI(34:1)
[M-H]
         
857.54 857.52 31.07 PI(36:4)
[M-H]
         
859.55 859.55 31.09 PI(36:3)
[M-H]
         
861.56 861.55 31.15 PI(36:2)
[M-H]
         
883.55 883.53 30.99 PI(38:5)
[M-H]
         
885.59 885.55 30.97 PI(38:4)
[M-H]
         
899.57 899.52 31.05 PI(h38:5)
[M-H]
         
901.56 901.54 30.97 PI(h38:4)
[M-H]
         
909.59 909.55 30.99 PI(40:6)
[M-H]
         
917.56 917.61 30.95 PI(40:2)
[M-H]
         
571.31 571.29 34.66 LysoPI(16:0)
[M-H]
         
595.31 595.29 34.64 LysoPI(18:2)
[M-H]
         
597.33 597.30 34.50 LysoPI(18:1)
[M-H]
         
599.35 599.32 34.50 LysoPI(18:0)
[M-H]
         
619.33 619.29 34.39 LysoPI(20:4)
1095.65 1095.64 35.72 GM3(d18:1/12:0)
[M-H]
         
1123.72 1123.67 35.61 GM3(d18:1/14:0)
[M-H]
         
1137.70 1137.65 35.48 GM3(t18:1/14:1)
[M-H]
         
1149.73 1149.69 35.54 GM3(d18:1/16:1)
[M-H]
         
1151.74 1151.71 35.50 GM3(d18:1/16:0)
[M-H]
         
1165.74 1165.69 35.54 GM3(t18:1/16:1)
[M-H]
         
1167.75 1167.70 36.38 GM3(t18:1/16:0)
[M-H]
         
1177.75 1177.72 35.40 GM3(d18:1/18:1)
[M-H]
         
1179.78 1179.74 35.40 GM3(d18:1/18:0)
[M-H]
         
1193.78 1193.72 35.40 GM3(t18:1/18:1)
[M-H]
         
1205.77 1205.75 35.29 GM3(d18:1/20:1)
[M-H]
         
1207.79 1207.77 35.29 GM3(d18:1/20:0)
[M-H]
         
1219.77 1219.73 35.42 GM3(t18:1/20:2)
[M-H]
         
1221.79 1221.75 35.15 GM3(t18:1/20:1)
[M-H]
         
1233.84 1233.78 35.20 GM3(d18:1/22:1)
[M-H]
         
1235.83 1235.80 35.20 GM3(d18:1/22:0)
[M-H]
         
1247.82 1247.76 35.12 GM3(t18:1/22:2)
[M-H]
         
1249.83 1249.78 35.16 GM3(t18:1/22:1)
[M-H]
         
1257.82 1257.79 35.25 GM3(d18:1/24:3)
[M-H]
         
1259.83 1259.80 35.18 GM3(d18:1/24:2)
1261.85 1261.82 35.12 GM3(d18:1/24:1)
[M-H]
         
1263.86 1263.83 35.11 GM3(d18:1/24:0)
[M-H]
         
1275.82 1275.80 35.20 GM3(t18:1/24:2)
[M-H]
         
1277.85 1277.81 36.02 GM3(t18:1/24:1)
[M-H]
         
1279.88 1279.83 35.98 GM3(t18:1/24:0)
[M-H]
         
1291.83 1291.79 35.25 GM3(t18:1/h24:2)
[M-H]
         
1293.82 1293.81 35.08 GM3(t18:1/h24:1)
[M-H]
         
1295.86 1295.82 36.54 GM3(t18:1/h24:0)
[M-H]
         
1354.84 1354.79 38.17 GM2(d18:1/16:0)
[M-H]
         
1368.84 1368.76 38.06 GM2(t18:1/16:1)
[M-H]
         
1380.84 1380.80 38.09 GM2(d18:1/18:1)
[M-H]
         
1382.86 1382.82 38.11 GM2(d18:1/18:0)
[M-H]
         
1396.83 1396.81 [M-H] 38.06 GM2(t18:1/18:1)

Table 2: Acidic lipid species detected by (-) nano-HPLC/MS analysis of purified ganglioside fractions corresponding to 1.5 μl human serum of a control volunteer patient. Analyte species printed in bold were monitored within the complete set of investigated samples. Asterisks indicate a subset of ions that was not employed for data normalization since low signal intensity hampered their reproducible detection even within the control group.

Reproducibility

Having a comprehensive map of human serum gangliosides we proceeded to search for changes associated with the onset of cancer. LC/ MS protocols for quantitative screening in different sample set would require the application of internal standards for accurate quantification. However, synthetic GSLs of similar structures and length as the serum GSLs are not routinely produced and available, and therefore the use of internal standards is currently compromised. Accordingly, we performed relative quantitation and normalized our data by calculating the percentage value for individual ions based on the sum of all monitored ganglioside signals (Table 2). To investigate the capability of this relative quantitation approach six identical serum aliquots that were pooled from serum samples of ten control volunteers were analyzed. GSLs were purified from each aliquot separately; subsequent nano-HPLC/MS analysis was performed in triplicates. Run-to-run variation was evaluated based on standard deviation of triplicates. Sample-to-sample variation was deduced from standard deviations of triplicate mean values for each investigated sample aliquot. The latter is depicted in figure 2 for each monitored ganglioside glycoform where intensities of different ceramide structures attached to the same glycan chain were summed for the diagram. Standard deviations are indicated by error bars and relative standard deviations are given by numbers on top of each column. The data displayed clearly demonstrate a good reproducibility of high abundant ganglioside structures varying in a range of about 15% of the mean value while relative standard deviations of less intense ganglioside signals dramatically increase to values of up to 90%. Also the data for run-to-run variation revealed very similar values (data not shown). In both cases the increase of relative standard deviation could be associated mainly to varying background signals of the corresponding spectra that hampered the detection of minor components. If such effects could be prevented, even the monitoring of less abundant structures seems to be possible under well-controlled experimental conditions. Nevertheless, the screening of the other serum gangliosides is straightforward since their signals are not significantly affected by variable background signals. In summary we can conclude the feasibility of our label-free method to generate accurate and reproducible results.

glycomics-lipidomics-Method-Reproducibility

Figure 2: Method Reproducibility. Normalized intensities of gangliosides detected in pooled samples from serum of ten control volunteers. Mean values and standard deviations refer to the sum of monitored ganglioside signals among each glycoform.
Superscribed letters indicate the first (a) and second (b) eluting isoform in case of isobaric structures.

Serum ganglioside levels and ceramide profiles

Since it is known that tumor cells often over express single ganglioside structures we first investigated if ganglioside serum levels correlate with the onset of cancer. For higher reproducibility the major focus of this study was set to the top 10 abundant ganglioside glycoforms. Signal intensities of different ceramide containing ganglioside ions were summed and screened for statistically significant differences. However, statistical analysis revealed only a few minor alterations among the investigated groups (Figure 3), while in total the displayed set of major serum ganglioside constituents remained at a constant level, independent from the disease status. Being only of low statistical significance, these alterations were not suited to discriminate all five patients groups. Since only minor disease-associated changes occurred in terms of ganglioside glycoform levels we decided to focus on the ceramide pattern of the analyzed gangliosides. Differential analysis of data revealed in this case an increase of palmitic acidcontaining ceramide structures in serum of cancer patients compared to the non-diseased control group, whereas especially stearic and behenic acid (C22:0)-containing structures exhibited an opposite effect. Remarkably, changes were limited to few ganglioside glycoforms and among them differentially pronounced. Ceramide profiles of GM3, IV3Neu5Ac-nLc4Cer, and IV6Neu5Ac-nLc4Cer revealed significant but minor changes (data not shown). By contrast, GD1a (Figure 4A) C16:0-ceramide containing species showed an increase of 40% in case of the pancreatic cancer group while sera of gastric cancer patients exhibited an increase of about 15% relative to the non-diseased control group. The reverse effect was observed for C18:0-ceramides where abundances were reduced by about 20% and 10% in the pancreatic and gastric cancer patient sera, respectively. Relative abundances of C20:0 and C24:1-ceramide comprising GD1a species remained unaffected. Similarly for GM2 (Figure 4C), the C16:0-ceramide structures were more elevated in the pancreatic (63%) and gastric (23%) cancer patient group compared to controls. Control sera of benign diseased patient groups showed a similar trend, though the differences were in both cases less pronounced compared to the corresponding cancer patient group. Based on the association observed for palmitic and stearic acid comprising species, the ratio of measured intensities for every sample was calculated. As documented by boxplots in figure 4B and 4D for GD1a and GM2, respectively, both gangliosides exhibit a very similar trend. The control groups show the lowest CerC16:0/CerC18:0 ratio while elevated values among the investigated cancer groups can be observed. Particularly samples obtained from pancreatic cancer patients feature a much broader spectrum of values, demonstrating a high individual variability of this sample cohort. P-values displayed within the diagrams demonstrate highly significant differences between control patients and the cancer patient groups. To further evaluate the significance of the ceramide ratio for cancer detection and consequently its applicability for cancer diagnosis a ROC-analysis was carried out. Therefore, the positive identification of cancer patients based on GM2-and GD1a-related CerC16:0/CerC18:0 ratios were evaluated in reference to the non-diseased control group only (Figure 5A), as well as the combined set of relevant benign and non-diseased control patients (Figure 5B). As demonstrated by the diagrams the GM2-related values revealed a much better performance for detection of both types of cancer when compared to the GD1a-related data. In addition, CerC16:0/CerC18:0-based cancer detection was clearly superior for the pancreatic cancer where area under the curve (AUC) values of up to 96% for the dissection of malignant and non-diseased control patients was reached. Considering additionally the benign diseased control group an AUC value of 93% could be obtained for the GM2-related ceramide ratio. Although the detection of gastric cancer was less efficient for both investigated data sets, AUC values of 85% could still be achieved for GM2.

glycomics-lipidomics-ganglioside-glycoforms

Figure 3: Normalized serum ganglioside levels of ten most abundant ganglioside glycoforms. Mean values and standard deviations refer to the sum of monitored ganglioside signals among each glycoform. Patient groups are represented by the columns as follows: pancreatic cancer (black), benign pancreatic diseases (striped), non-diseased controls (white), benign gastric diseases (checked), gastric cancer (grey).

glycomics-lipidomics-Ceramide-profiles

Figure 4: Ceramide profiles and CerC16:0/CerC18:0 ratios of serum gangliosides GD1a and GM2. Ceramide signals are normalized based on the sum of all structures detected among (A) GD1a and (C) GM2 for each sample. CerC16:0/CerC18:0 ratios of (B) GD1a and (D) GM2 of individual patients are displayed as boxplots. Patient groups are represented as follows: pancreatic cancer (black/PAN), benign pancreatic diseases (striped/BP), non-diseased controls (white/CTR), benign gastric diseases (checked/BS), gastric cancer (grey/STO). Error bars indicate standard deviations. P-values displayed in the upper panel of each diagram (panels A and C only for P≤0.01) were obtained by the non-parametric Mann-Whitney-U test.

glycomics-lipidomics-gastric-cancer

Figure 5: ROC-curves for identification of pancreatic and gastric cancer patients based on GM2-and GD1a-related ceramide ratios. (A) ROC-plots for identification of pancreatic (PAN) or gastric cancer patients (STO) combined in data sets with non-diseased controls (CTR). (B) ROC-curves for identification of pancreatic or gastric cancer patients combined in data sets with non-diseased controls and patient cohorts suffering from benign diseases of the pancreas (BP) and the stomach (BS), respectively. Curves were calculated for GM2-(black lines) and GD1a-related CerC16:0/CerC18:0 ratios (grey lines).

Discussion

Serum ganglioside composition

A great portion of serum glycosphingolipids originates from shedding processes at various cell types. Consequently this pool of biomolecules originates from different sources of the body [10]. Unfortunately, the tissue or cell type origin of single ganglioside constituents in the serum cannot be elucidated yet. Though, based on the ganglioside composition of cells that are in the close contact with blood, such as blood cells and endothelial tissues, it can be hypothesized that shedding at these cells predominantly affect GSL levels in serum. Both, erythrocytes [19] and endothelial cells [27,28] mainly comprise GM3 and IV3Neu5Ac-nLc4Cer while leucocytes were found to contain a broad spectrum of singly sialylated neolactoseries gangliosides ranging from GM3 up to XIV3Neu5Ac-nLc14Cer [29,30]. The cerebrospinal fluid that is connected to the blood via the lymphatic system [31] could be another source of serum ganglio-series gangliosides as it contains especially GD1a, GD1b, GT1b, and GQ1b [32]. Due to their reduced anchoring potential within the membrane preferably lipids with short chain fatty acids are shed [33]. Concordant with this fact, our data demonstrate that serum gangliosides exhibit a relatively high abundance of palmitic and stearic acid and only to a minor extend long chain fatty acids. Interestingly, ceramide profiles vary among the different ganglioside structures detected in serum. In case of GD1a, GD1b, GT1, and GM2 most intense signals derive from structures containing stearic acid in their ceramide moiety while in all other profiles the most abundant species contain palmitic acid. The fact that especially neuronal tissues highly express Cer(d18:1/18:0) structures [34] might support the hypothesis that the corresponding gangliosides are originally shed from neuronal tissues and are subsequently transported into the blood. Still, other tissue types exhibiting a similar ceramide synthase (CerS) expression could represent alternative sources, and accordingly concerted efforts will be needed to elucidate the serum GSL sourcing.

Cancer-associated alterations

Although tumors can induce strong alterations of serum ganglioside glycoform levels [35] data generated in the course of our current study revealed no obvious differences among the five patient groups. Due to our analytical set-up we could obtain a relative quantitative response, in which a general increase of the serum levels of all ganglioside species cannot be monitored. However, a parallel study on a similar set of patient samples employing semi-quantitative immunostaining of thin-layer separated gangliosides being consistent with our current results did not indicate such an effect (data not shown). Although total serum levels remained unchanged, unambiguous alterations in the ceramide profiles of different gangliosides could be evidenced (Figure 4). Relative abundances of C16:0-ceramide-comprising structures were elevated while those of C18:0-ceramides significantly reduced in cancer patient samples. Similar effects have been observed for the ceramide composition in human head and neck squamous carcinoma cells (HNSCC) [36] and breast cancer tissues [37]. A proposed mechanism for these changes has been associated to the expression the ceramide synthases CerS6 and CerS1 and their effect on ER-stress mediated apoptosis via activation of caspase-3 [38]. In this context, CerS6/C16:0-ceramide downregulation and CerS1/C18:0-ceramide induction was found to similarly increase programmed cell death in vitro and in vivo. Moreover increased C18-ceramide levels were found to correlate with the lymphovascular invasion and the formation of nodal metastases in HNSCC tumors [39] and thus evaluated as serum biomarkers in a recent phase II trial to monitor a chemotherapeutic response [40]. These effects were described for free ceramide structures, however, similar changes among GSL structures were to our knowledge not investigated so far. In fact, many GSL studies are still focused on the glycan chain composition only although it is well known that the ceramide portion can reveal important biological information [41]. The strong association of observed changes in ceramide profiles to gangliosides GD1a and GM2 could be related to the tissue characteristics of both investigated tumors. Representing adenocarcinomas, both tumor types derive from glandular and thus epithelial tissues. GM2 has been reported to be expressed on a variety of epithelial tissues including those of the pancreas and the stomach [42]. In addition, also the presence of GM2 in the majority of corresponding carcinomas could be evidenced. Although GD1a expression was less detailed investigated it was detected in different carcinomas or derived cell lines [43-45]. Accordingly, this association to epithelial cells could suggest that similar alterations might also occur among other (adeno) carcinomas. The ROC-analysis revealed a good capability of the GM2- related ceramide ratio for detection of pancreatic cancer. Although many biomarkers proposed so far exhibit a higher performance for the discrimination of control patients and pancreatic cancer patients, their major drawback is the lack of feasibility to discriminate between malignant and benign diseased patients [46,47]. In our study such effects were clearly less pronounced. However, due to the relatively small set of patient samples analyzed in this study further validation is necessary in order to determine the feasibility of the GM2-and GD1arelated ceramide profiles for clinical use.

Conclusions

In this current study of human serum no alterations of ganglioside glycoforms with respect to cancer and cancer type, were observed at a significant level. The unbiased nano-HPLC/MS approach used for the analysis of serum gangliosides enabled however the identification of cancer-related changes within the ceramide profiles of several ganglioside structures. To our knowledge this is the first study documenting serum level alterations of GSL fatty acid substitution with cancer. Based on the performance evaluated in ROC-analysis the GM2- related ceramide ratio represents a potential biomarker for pancreatic cancer. Discrimination of pancreatic cancer patients and non-diseased control volunteers could be accomplished at 92.1% sensitivity and 88.2% selectivity. Taking into account also benign diseased control patients’ performance dropped slightly to 89.1% sensitivity and 84.8% selectivity which is still superior to other biomarkers proposed so far. In the case of gastric cancer the GM2-related ceramide ratios revealed less specificity for cancer patient detection that renders an application for tumor detection alone unlikely.

Grant support

The study was financially supported by the European Union GLYFDIS project (project no. LSHB-CT-2006-037661, to J.P.K.) and the Deutsche Forschungsgemeinschaft within the Sonderforschungsbereich 492 (project Z2). S.K. is a fellow of the German National Academic Foundation (Studienstiftung des Deutschen Volkes).

References

Select your language of interest to view the total content in your interested language
Post your comment

Share This Article

Relevant Topics

Recommended Conferences

  • International Conference on Computational Biology and Bioinformatics Sep 05-06 2018 Tokyo, Japan
    November 19-21-2018 Bucharest, Romania
  • 3rd International Conference on Enzymology and Molecular Biology
    September 17-18, 2018 Singapore City, Singapore
  • 4th International Conference on Enzymology and Lipid Science September 17-18, 2018 Singapore Theme: Novel Advancements in Enzymology and Lipid Science
    September 17-18, 2018 Singapore City, Singapore
  • 4th International Conference on Biochemistry & Biophysics October 03-04, 2018 Los Angeles, California, USA
    October 03-04, 2018 Los Angeles, USA
  • 9th International Conference on Biopolymers and Polymer Sciences November 19-21-2018 Romania,Bucharest
    November 19-21-2018 Bucharest, Romania
  • 12th International Conference on Advancements in Bioinformatics and Drug Discovery
    November 26-27, 2018 Dublin, Ireland
  • Cellular and Molecular Mechanism Conference: Health and Disease November 29-30, 2018 Bali, Indonesia
    November 19-21-2018 Bucharest, Romania

Article Usage

  • Total views: 11956
  • [From(publication date):
    specialissue-2012 - Aug 21, 2018]
  • Breakdown by view type
  • HTML page views : 8149
  • PDF downloads : 3807
 

Post your comment

captcha   Reload  Can't read the image? click here to refresh

Peer Reviewed Journals
 
Make the best use of Scientific Research and information from our 700 + peer reviewed, Open Access Journals
International Conferences 2018-19
 
Meet Inspiring Speakers and Experts at our 3000+ Global Annual Meetings

Contact Us

Agri & Aquaculture Journals

Dr. Krish

kactakapaniyor.com

[email protected]

+1-702-714-7001Extn: 9040

Biochemistry Journals

Datta A

Taktube

[email protected]

1-702-714-7001Extn: 9037

Business & Management Journals

Ronald

porn sex

[email protected]

1-702-714-7001Extn: 9042

Chemistry Journals

Gabriel Shaw

Gaziantep Escort

[email protected]

1-702-714-7001Extn: 9040

Clinical Journals

Datta A

sikiş

[email protected]

1-702-714-7001Extn: 9037

instafollowers

James Franklin

[email protected]

1-702-714-7001Extn: 9042

Food & Nutrition Journals

Katie Wilson

[email protected]

1-702-714-7001Extn: 9042

General Science

Andrea Jason

mp3 indir

[email protected]

1-702-714-7001Extn: 9043

Genetics & Molecular Biology Journals

Anna Melissa

[email protected]

1-702-714-7001Extn: 9006

Immunology & Microbiology Journals

David Gorantl

[email protected]

1-702-714-7001Extn: 9014

Materials Science Journals

Rachle Green

[email protected]

1-702-714-7001Extn: 9039

Nursing & Health Care Journals

Stephanie Skinner

[email protected]

1-702-714-7001Extn: 9039

Medical Journals

putlockers

Nimmi Anna

[email protected]

1-702-714-7001Extn: 9038

Neuroscience & Psychology Journals

Nathan T

seks

[email protected]

1-702-714-7001Extn: 9041

Pharmaceutical Sciences Journals

Ann Jose

[email protected]

1-702-714-7001Extn: 9007

Social & Political Science Journals

Steve Harry

[email protected]

1-702-714-7001Extn: 9042

 
© 2008- 2018 OMICS International - Open Access Publisher. Best viewed in Mozilla Firefox | Google Chrome | Above IE 7.0 version
Leave Your Message 24x7