Subscriber access provided by READING UNIV
Article
A metabolomic signature of endometrial cancer Jacopo Troisi, Laura Sarno, Annamaria Landolfi, Giovanni Scala, Pasquale Martinelli, Roberta Venturella, Annalisa Di Cello, Fulvio Zullo, and Maurizio Guida J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.7b00503 • Publication Date (Web): 13 Dec 2017 Downloaded from http://pubs.acs.org on December 14, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 25
Journal of Proteome Research
1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
A metabolomic signature of endometrial cancer Jacopo TROISIa,b*, Laura SARNOc, Annamaria LANDOLFIa, Giovanni SCALAb, Pasquale MARTINELLIc, Roberta VENTURELLAd, Annalisa DI CELLOd, Fulvio ZULLOd, Maurizio GUIDAa,b
a) Department of Medicine, Surgery and Dentistry “Scuola Medica Salernitana”, University of Salerno, Baronissi (SA), Italy b) Theoreo srl, Montecorvino Pugliano (SA), Italy c) Department of Neurosciences and Reproductive and Dentistry Sciences, University of Naples Federico II, Naples (Italy) d) Unit of Obstetrics and Gynaecology, Department of Medical and Surgical Sciences, "Magna Graecia" University of Catanzaro, Catanzaro (CZ), Italy
*Corresponding
author: Jacopo Troisi, Theoreo srl – Spin-off Company of the University of
Salerno, Via degli Ulivi, 3 CAP 84090, Montecorvino Pugliano (SA) Italy. Tel/FAX +39 089 0977435 email:
[email protected] ACS Paragon Plus Environment
Journal of Proteome Research
Page 2 of 25
2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Abstract Endometrial cancer (EC) is the most common cancer of the female reproductive tract in developed Countries. At the moment, no effective screening system is available. Here, we evaluate the diagnostic performance of a serum metabolomic signature. Two enrollments were carried out: one constituted of 168 subjects: 88 with EC and 80 healthy women, was used for building the classification models. The second (used to establish the performance of the classification algorithm) was constituted of 120 subjects: 30 with EC, 30 with ovarian cancer, 10 with benign endometrial disease and 50 healthy controls. Two ensemble models were built, one with all EC vs. controls (Model I) and one in which EC patients were aggregated according to their histotype (Model II). Serum metabolomic analysis was conducted via gas chromatography-mass spectrometry, while classification was done by an ensemble learning machine. Accuracy ranged from 62% to 99% for the Model I and from 67% to 100% for the Model II. Ensemble model showed an accuracy of 100% both for Model I and II. The most important metabolites in class separation were: lactic acid, progesterone, homocysteine, 3-hydroxybutyrate, linoleic acid, stearic acid, myristic acid, threonine, and valine. The serum metabolomics signature of endometrial cancer patients is peculiar because it differs from that of healthy controls and from that of benign endometrial disease and from other gynecological cancers (such as ovarian cancer).
Keyword: Endometrial cancer; Metablomics; Partial Least Square Discriminant Analysis; Classification models; Ensemble machine learning models.
ACS Paragon Plus Environment
Page 3 of 25
Journal of Proteome Research
3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Introduction Endometrial carcinoma (EC) is the most common invasive cancer of the female genital tract and is responsible for 7% of all invasive cancers in women (excluding skin cancers).1,2 EC is a tumor of peri- and postmenopausal period, since the peak of incidence is between 55 and 65 years old. Pathological studies and molecular analyses have supported the classification of EC into two broad categories: Type I and Type II. Type I is the most frequent, with a percentage of cases greater than 80%. It is often well differentiated and undermines the proliferative endometrial glands and, therefore, is considered an endometrioid carcinoma. It mostly occurs in the context of endometrial hyperplasia and, similarly to this condition, is associated with obesity, diabetes, hypertension, infertility and unopposed estrogen stimulation. Many studies have provided further evidence to support the view that endometrial hyperplasia is a precursor of EC.
3,4
Type II EC generally occurs a decade later than type I (65-75 years) and,
unlike type I, mostly develops on a framework of endometrial atrophy. Type II represents less than 15% of EC and is poorly differentiated (G3). Its most common subtype is serous, named for the biological and morphological overlap with the ovarian cancer. Other less common histologic subtypes are clear cell carcinoma, and the malignant mixed Müllerian. While for cervical cancer there is a screening test called Papanicolau test, a screening for EC on asymptomatic premenopausal and postmenopausal women is currently not available. Studies on exocervical samples show a frequency of false negatives of about 40-50%: due to the action of vaginal environment, exfoliated endometrial cells lose the characteristics that allow differentiation of tumor cells from normal cells. On the other hand, the prognosis is closely related to the early diagnosis. Indeed, the 5-year survival rate decreases drastically from 9096% in the case of diagnosis at stage I, up to a 0-2% in the case of diagnosis at stage IV.5 Therefore, the need for a non-invasive screening that allows the identification of a high risk population and a prompt diagnosis may change the natural history of this cancer. The use of metabolomic for the study of disease biomarkers is recently spreading. Metabolites are low-molecular-weight organic and inorganic chemicals that are the substrates, ACS Paragon Plus Environment
Journal of Proteome Research
Page 4 of 25
4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
intermediates, and byproducts of enzyme-mediated biochemical reactions in the cell.6 Metabolomics is thought to have the potential to provide more fundamental and global information than genomics, proteomics, and transcriptomics and to more precisely reflect the phenotype of the cell. Several thousands of metabolites in human serum have been identified so far7 and the application of metabolomics has allowed the development of biomarkers in many diseases8 and especially for early diagnosis.9 However, the use of metabolomics in gynecology is, at the moment, limited. In this paper we described the performance of two different models based on serum metabolomic fingerprint of patients with EC: model I, discriminating the patients on the basis of presence or absence of EC; model II, discriminating patients with a diagnosis of EC on the basis of the different histotype (Type 1 or 2).
Materials and Methods Population and study design This was a prospective pilot study conducted from May 2012 to November 2016 in three different hospital in South Italy (the University hospitals of Catanzaro, Naples and Salerno). Patients with a diagnosis of EC, ovarian cancer and benign endometrial disease were included in the study. All details about patients’ enrollment and samples’ collection are reported in supplementary section S1. All sampling centers have respected the same collection protocol as reported in supplementary section S1. EC and ovarian cancer were defined on the basis of the histological examination after the surgery.10 Benign endometrial diseases included hyperplasia without atypia, endometrial polyps and abnormal bleeding, with a normal histology after surgery.11 Controls were recruited from healthy women visiting the gynecological clinics for routine controls and resulting have no endometrial, ovarian or vaginal anomalies.
ACS Paragon Plus Environment
Page 5 of 25
Journal of Proteome Research
5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Exclusion criteria were: women in fertile age and history of previous diagnosis of cancer in any other site. Two separate enrollments were conducted: •
Recruitment I (EC women compared with a control group of healthy women), used for building the case-control classification models.
•
Recruitment II: women with a diagnosis of EC, ovarian cancer, benign endometrial disease were enrolled and compared with a control group of healthy women. This cohort was used to establish the performance of the classification algorithms previously built.
Two ensemble models were built, a two-class model with all EC vs. controls (Model I) and a two-class models in which EC patients were aggregated according to their histotype (Type I or Type II). The study was approved by the ethics committee CEI “Comitato etico Campania Sud” (IRB No. 91/106416) and a written consent form was signed by each participant. At enrollment, demographic, anamnestic and clinical characteristics were collected for all the included patients and recorded on a dedicated database. Impaired fasting glucose, diabetes, hypertriglyceridemia, hypercholesterolemia and metabolic syndrome were defined according to NCEP ATP III criteria. Moreover, a clinical and gynecological examination, vaginal ultrasound, additional uterine imaging, Carbohydrate Antigen 125 (CA125), Human Epididymis Protein 4 (HE4) dosage and R.O.M.A Index12 were performed and recorded.
Sample collection Human tissue collection strictly adhered to the guidelines outlined in the Declaration of Helsinki IV edition. Blood samples of each included patient were collected immediately before hysterectomy or medical therapy, using a BD vacutainer (Becton Dickinson, Oxfordshire, UK) blood collection red tube (with no additives). After centrifugation, the sample was immediately ACS Paragon Plus Environment
Journal of Proteome Research
Page 6 of 25
6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
frozen to -80 °C until the time of analysis (see supplementary section S1 for details). All patients were asked to respect a 12-hours fast before blood collection.
Samples metabolomics analysis The metabolome extraction, purification and derivatization was carried by the MetaboPrep GC kit (Theoreo, Montecorvino Pugliano, Italy). The protocol, according to the manufacturer instructions, required the adding of 50 µL of serum sample to a microcentrifuge tube containing the extraction mixture and the internal standard. The sample and extraction mixture was vortexing at 1250 rpm for 30 seconds. The extract metabolome was centrifuge for 5 minutes at 16000 rpm keeping the temperature below the 4°C. Two hundred microliter of the upper liquid phase was removed and transferred into a microcentrifuge tube containing the purification mixture, these were vortexing at 1250 rpm for 30 seconds. A rapid centrifuge of the sample (to prevent the sediment resuspension) at 16000 rpm keeping the temperature below the 4°C allow the solid sedimentation. One hundred seventy five microliter of liquid upper phase was transfer into the glass vial and freeze-drying overnight. For the derivatization the lyophilized sampled was added of 50 µL of the first derivatization mixture and vortexing at 1200 rpm at room temperature for 90 minutes, than were added 25 µL of the second derivatization mixture and vortexing at 1200 rpm at room temperature for 90 minutes. The derivatizated metabolome was transfer in a 100 µL insert for the auto sampler injection. This was centrifuge for 5 minutes at 16000 rpm keeping the temperature below the 4°C before inject in the GC-MS. Samples of 1.8 µL from the derivatized solution were injected into the GC-MS system (GC-2010 Plus gas chromatograph coupled to a 2010 Plus single quadrupole mass spectrometer; Shimadzu Corp., Kyoto, Japan). Chromatographic separation was achieved with a 30 m 0.25 mm CP-Sil 8 CB fused silica capillary GC column with 1.00 µm film thickness from Agilent (Agilent, J&W), with helium as carrier gas. The initial oven temperature of 100 °C was held for 1 min and then raised at 4 °C/min to 320 °C with further 4 minutes of hold time. The gas flow was set to achieve a constant linear velocity of 39 cm/s and ACS Paragon Plus Environment
Page 7 of 25
Journal of Proteome Research
7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
the split flow was set to 1:5. The mass spectrometer was operated in electron impact (70 eV) in full scan mode in the interval of 35-600 m/z with a scan velocity of 3333 amu/sec and a solvent cut time of 4.5 minute. The complete GC programme duration was 60 minutes. For peak identification, the linear index difference max tolerance was set to 10, while the minimum matching for NIST library search was set to 85%. Samples were divided in batches, each containing 25 of them. Each batch was monitored with 4 controls: a blank injection, an injection of a standard mix, an injection of a pooled sample solution and a replicate injection of one sample (randomly selected from the samples in the batch). The blank consisted in an injection of 2 µL of hexane, the mix solution contained a mix of 50 molecules (organic acids, sugars, amino acids, sterols, fatty acids, vitamins, nitrogen bases, etc.), the pooled sample contained 2 µL of each of 50 randomly selected derivatized samples, while the replicated injection was made using one derivatized samples of the analyzed batch. Each batch was considered valid if the blank solution generated no detectable peak, the standard peak area ratio (normalized to the internal standard area) was within 10% of the expected value, the area ratio (normalized to the internal standard area) of the 100 major peaks of the repeated injection was within 15% of the original sample and the pooled sample was allocated in the same area of the other pooled samples, that is
