470
DOI 10.1002/prca.201500083
Proteomics Clin. Appl. 2016, 10, 470–484
REVIEW
Integration of metabolomics and proteomics in multiple sclerosis: From biomarkers discovery to personalized medicine Piero Del Boccio1,2∗ , Claudia Rossi1,2∗ , Maria di Ioia2,3 , Ilaria Cicalini1,2 , Paolo Sacchetta1,2 and Damiana Pieragostino1,2 1
Department of Medical Oral and Biotechnological Sciences, University “G. d’Annunzio” of Chieti- Pescara, Chieti, Italy 2 Analytical Biochemistry and Proteomics Unit, Research Centre on Aging (Ce.S.I), University “G. d’Annunzio” of Chieti-Pescara, Chieti, Italy 3 Department of Neurosciences and Imaging, University “G. d’Annunzio” of Chieti-Pescara, Chieti, Italy Personalized medicine is the science of individualized prevention and therapy. In the last decade, advances in high-throughput approaches allowed the development of proteomic and metabolomic studies in evaluating the association of genetic and phenotypic variability with disease sensitivity and analgesic response. These considerations have more value in case of multiple sclerosis (MuS), a multifactorial disease with high heterogeneity in clinical course and treatment response. In this review, we reported and updated about proteomic and metabolomic studies for the research of new candidate biomarkers in MuS, and difficulties in their clinical applications. We focused especially on the description of both “omics” approaches that, once integrated, may synergically describe pathophysiology conditions. To prove this assumption, we rebuilt interaction between proteins and metabolites described in the literature as potential biomarkers for MuS, and a pathway analysis of these molecules was performed. The result of such speculation demonstrated a strong convergence of proteomic and metabolomic results in this field, showing also a poorness of available tools for incorporating “omics” approaches. In conclusion, the integration of Metabolomics and Proteomics may allow a more complete characterization of such a heterogeneous disease, providing further insights into personalized healthcare.
Received: July 30, 2015 Revised: November 17, 2015 Accepted: December 30, 2015
Keywords: Biomarkers / Metabolomics / Multiple sclerosis / Personalized medicine / Proteomics
1 Correspondence: Dr. Damiana Pieragostino, Department of Medical Oral and Biotechnological Sciences, University “G. d’Annunzio” Chieti-Pescara, Chieti, Italy E-mail:
[email protected] Fax: +39 0871 541598 Abbreviation: BDNF, brain-derived neurotrophic factor; CH3L1, chitinase-3-like protein 1; CIS, clinically isolated syndrome; CSF, cerebrospinal fluid; CXCL13, chemokine ligand 13; DMDs, disease-modifying drugs; EDSS, expanded disability status scale; GFAP, glial fibrillary acidic protein; HP, haptaglobin; IFN-, interferon beta; IL-12, interleukin 12; IPA, ingenuity pathway analysis; JCV, John Cunningham virus; MMP9, matrix metalloproteinase 9; MRI, magnetic resonance imaging; MuS, multiple sclerosis; NAbs, neutralizing antibodies; NTZ, natalizumab; OCBs, oligoclonal bands; PC, phosphatidyl choline; PML, progressive multifocal leukoencephalopathy; RRMuS, relapsing-remitting MuS; C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Multiple sclerosis and personalized medicine
Personalized medicine is the science of individualized prevention and therapy. The notion “one size fits all” has been replaced by the idea of patient-tailored healthcare [1]. Based on the developments made during the past decade, it is feasible to predict that a personalized approach for treating neurological disorders will become more widely applicable in the coming years [2]. Researchers conducted genetic, proteomic, and metabolomic studies evaluating the association of genetic and phenotypic variability with disease sensitivity and SPMuS, secondary progressive MuS; TNF, tumor necrosis factors; TTR, transthyretin; VTDB, vitamin D binding protein ∗ These authors contributed equally to this work. Colour Online: See the article online to view Figs. 1 and 3 in colour.
www.clinical.proteomics-journal.com
471
Proteomics Clin. Appl. 2016, 10, 470–484
analgesic response [3]. These considerations have more value in case of multiple sclerosis (MuS), which is a multifactorial disease. Nowadays, the clinical and laboratory parameters considered for MuS can be divided into multiple categories: diagnostic, prognostic, disease activity, predictive, treatment monitoring, and treatment selection. Obviously, the last mentioned categories are most important for generating a specific therapy, and are the biomarkers will be considered for this review. MuS is a chronic inflammatory disease with an autoimmune pathogenesis. Its causes are unknown but it is widely accepted by the scientific community that several environmental factors in genetically predisposed subjects lead to an abnormal activation of the immune system against myelin of the CNS. The disease is highly heterogeneous and can have various pathological courses. In most cases, it has a relapsingremitting MuS (RRMuS) that, in 50% of cases, evolves into a secondary progressive MuS (SPMuS) after about 10 years [4]. For RRMuS patients, nowadays several disease-modifying drugs (DMDs), divided into first- and second-line therapies, are available with different efficacy and safety profiles. The conventional first-line therapies are interferon beta 1a (IFN 1a), IFN- 1b, and glatiramer acetate. These therapies are injectable and have a good safety profile. More recently, oral first-line therapies, such as, teriflunomide and dimethyl fumarate, have emerged, which are more convenient for the patients, but with a safety profile not fully known since long time [5]. These drugs have different mechanisms of action, similar efficacy, and, as already mentioned, different safety profiles. When a first-line therapy fails, and/or when MuS has an aggressive course, a second-line therapy is recommended. The second-line DMDs approved for MuS are fingolimod, natalizumab (NTZ), and alemtuzumab. All these drugs are immunosuppressive, more efficacious but with greater risks than first-line DMDs [4, 6]. Response to treatment can be different in each patient and extremely difficult to predict. Thus, personalized medicine in MuS will be increasingly feasible because of new and validated biomarkers obtained, especially through the definition of response to treatment and risk associated with it. One of the crucial aspects of this disease is the transition from RRMuS to SPMuS. RRMuS is well characterized by the presence of inflammation expressed as clinical relapses and activity to a magnetic resonance imaging (MRI) scan. These types of MuS are, generally, well controlled by therapeutic options now available. In contrast, SPMuS is characterized by a prevalence of neurodegeneration, axonal loss, and brain atrophy, which begin early and progress insidiously; up to now no drugs are really effective on these processes. So, the discovery and validation of prognosis and predictive biomarkers of therapeutic response will be the next future challenge in therapeutic decision-making. These results will lead to a better tailored therapy for each patient. In this context, the integration of proteomic and metabolomic knowledge may provide a great help, since protein expressions and metabolic pathways are the consequence of the interaction between genetic and environmental factors, which are very useful for C 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
studying multifactorial diseases such as MuS. To better discuss the needs and inability to achieve personalized medicine in MuS, we described the criteria by which MuS is, currently, evaluated in clinical practice.
2
Clinical evaluation of MuS
With the 2010 revised McDonald criteria for MuS diagnosis, the demonstration of dissemination in space and time of CNS inflammation was completely entrusted to MRI studies [7]. The lumbar puncture is still used in clinical procedure in case of suspected inflammatory demyelinating diseases, since the presence of oligoclonal bands (OCBs) in cerebrospinal fluid (CSF) is considered very sensitive, even if not specific of MuS. Other biological molecules can also be found in the CSF, which include neurofilament, myelin basic protein, glial fibrillary acidic protein (GFAP), tau, neuronal cell adhesion molecule, and the growth-associated protein (GAP-439), but their diagnostic value is still to be determined [8]. Up to now, through epidemiological studies, clinicians based their prognostic evaluation on clinical and radiological indicators. For example, there are some clinical markers of good prognosis, such as female gender, young age of onset, optic neuritis or isolated sensory symptoms at debut, long time between the first and second relapse, no accumulation of disability after 5 years, few lesions at baseline MRI. Conversely, an elevated load of lesion as well as lesions located in the spinal cord are associated with worse evolution [9]. Patients in early phase, without a definite diagnosis, but with CSF OCBs, are more likely to develop definite MuS compared to patients with a normal CSF [8]. MRI measures of axonal damage, as black holes in T1 sequences and brain atrophy, correlate with clinical disability. The optic coherent tomography provides information on axonal damage and can be used to monitor it. Potential prognostic markers are immunoglobulins M (IgM) [10] and neurofilaments in CSF [11, 12], antimyelin [13], and chitinase-3-like protein 1 (CH3L1) [14] in blood and CSF [15]. These markers have to be validated before their use in clinical practice [9]. CH3L1 is the only protein discovered with proteomics in CSF and in blood, which has been validated and suggested as a prognostic marker [16]; we will discuss it later. Actually, there are some interesting articles on the evaluation of clinical effects of DMDs therapies [17,18]. Therapeutic response in patients treated with DMDs is usually evaluated considering the reduction in relapse rate, the number of new or enlarged T2 lesions, the number of gadolinium enhancing lesions, and the score of expanded disability status scale (EDSS) [19]. For patients treated with IFN-, which is a protein potentially immunogenic, it is possible to detect neutralizing antibodies (NAbs) [20]. When positive, NAbs can reduce drug’s efficacy in proportion to their titer. For high titer (>1:100), a therapeutic alternative should be evaluated. The positivity of NAbs predicts MRI and clinical activity, so precocious identification of NAbs-positive patients and the www.clinical.proteomics-journal.com
472
P. Del Boccio et al.
switch to alternative treatments may improve the response toward treatment [21]. Myxovirus resistance protein A is a molecule induced after IFN- injection, which is used to evaluate its biological activity, especially in doubtful cases [22]. The types of IFN- are differently immunogenic, depending on molecule and route of administration. IFN- 1b is more immunogenic than IFN- 1a, and the subcutaneous administration has a higher risk than the intramuscular administration. For low titer (1.5), have an increased risk of PML compared to patients with lower antiJCV antibody index (
