How do we improve treatments for patients with amyloidosis using proteomics?

KEYWORDS:

• Amyloidosis
• amyloid typing
• biomarkers
• early diagnosis
• novel treatments

1. Introduction

Amyloidoses are conditions characterized by deposition of autologous proteins as extracellular fibrillar aggregates. This class of diseases is heterogeneous and in continuous expansion, and 36 distinct proteins are currently known to cause amyloidoses in humans [1,2]. Amyloid typing consists in defining the nature of the fibrillar protein; this is a key diagnostic step because the various forms, despite similar ultrastructural appearance, are treated with completely different approaches. These latter can range from chemotherapy, in the case of immunoglobulin light chain amyloidosis (AL), to transplantation of the organ producing the amyloidogenic protein, in hereditary forms including transthyretin amyloidosis (ATTR) [3]. In addition, potential novel therapies are under study for specific amyloidosis types, such as antisense oligonucleotides to reduce the precursor synthesis in ATTR and antiamyloid antibodies to clear the deposits in AL [3]. Clearly, errors or delays in the diagnostic phase can translate into potentially dramatic consequences, due to wrong treatment choices or missed therapeutic opportunities. In recent years, proteomics has been introduced as a clinically useful tool for the management of amyloidoses [2,4,5]. In a short time since its first applications, it has gained impressive relevance as a method for amyloid protein identification without the use of antibodies, which are sometimes unreliable in this context [5,6]. Methods for analyzing the whole-tissue proteome, especially in the case of fat tissue, or the amyloid deposits isolated by laser microdissection have been developed [4,5,7–9]. There are multiple ways in which proteomics may help improve the care of patients with amyloidoses: (1) by ameliorating the diagnostic sensitivity and specificity; (2) by identifying novel amyloidoses; (3) by indicating novel candidate biomarkers; and (4) by providing better knowledge of the molecular bases of these diseases. The major acquisitions regarding each of these aspects have been recently reviewed [5].

2. Amelioration of the diagnostic phase

Currently, the major clinical impact of proteomics for improving amyloid care stands in its ability to unequivocally define the disease type. Today, mass spectrometry (MS)-based amyloid typing is indeed a unique example of an accredited clinical proteomic assay, employed in the routine clinical workup [5]. This approach significantly increases the positive identification of the amyloid fibril protein compared to traditional immunohistochemistry [10]; in the context of validation studies performed by a major referral center, it was reported to have 100% typing sensitivity and specificity [4], although these performances need to be further defined through larger clinical assessments. Proteomic amyloid identification requires minimal amounts of tissue and can be performed on formalin-fixed paraffin-embedded samples, as those typically acquired for pathological examination. The principal limitations currently stand in the facts that the technology and the specific expertise for translating the proteomic data into a clinical diagnostic report are available only at selected centers and that multicenter validation studies are still missing. A collective effort for technique standardization and comparison of the performances of proteomics with those of alternative, highly reliable typing methods such as immuno-electron microscopy [11] are now advocated [6,10].

3. Novel forms of amyloidoses are continuously discovered

Novel forms of human amyloidoses have recently been discovered, and proteomics has played an important role in this setting [12]. The ability to identify and partially sequence proteins without the need for fibril purification greatly facilitates diagnosis of new types from patients’ biopsies. In addition, and in contrast with immunohistochemistry, which requires a piori selection of the antibody panel, proteomics allows detecting proteins whose nature is not hypothesized in advance. Today, MS is certainly the most convenient and reliable method for spotting novel amyloidoses. Classification of a new disease type needs, however, definitive confirmation by biochemical study of the extracted fibrils [1]. Moreover, thanks to proteomic methods for the analysis of routine and archived samples, the landscape of the various amyloidoses has been significantly redefined, with important clinical and therapeutic implications [1,5]. For example, previously unrecognized forms such as ALect2 were shown to have an unexpectedly high prevalence in western countries [4].

4. Sensitivity of proteomics for early detection of amyloid

An important issue in the management of amyloid diseases is to achieve early and sensitive diagnosis. This is crucial to grant timely access to established or experimental treatments, thus preventing the progressive deposition of amyloid and deterioration of organ function. Recent work suggests that proteomic analysis of adipose tissue – the site of choice for evaluating the presence of fibrils in candidate patients – may be more sensitive than current methods based on Congo red staining [9]. Assessment of deposits is achieved through the identification of a set of amyloid-associated proteins (apolipoprotein E, serum amyloid P-component, and apolipoprotein A-IV), whose simultaneous occurrence is highly indicative of fibrils. This application is promising and might allow diagnosing the disease at early stages, before the formation of conspicuous amounts of aggregates. Increasing the diagnostic performance is especially important for those forms, such as ATTR, in which Congo red staining of fat tissue has lower diagnostic sensitivity [11], and in classes of individuals at risk, such as carriers of amyloidogenic mutations, or MGUS patients with abnormal κ/λ free light chains (LCs) ratio or with LV6-57 gene usage (at risk for AL) [13,14]. The putative higher performance of proteomics compared with histology is likely related to the intrinsic analytical sensitivity of modern MS, which allows detecting minute amounts of proteins. The clinical implications of a positive MS diagnosis in the absence of positive Congo red staining, however, are not yet completely clear. It is also intriguing to speculate that the presence of aggregate-associated proteins, in the absence of Congo red positivity, may subtend the existence of pre-fibrillar aggregates in vivo; indeed, this hypothesis, which would open entirely novel scenarios, has not yet been demonstrated.

5. Need for novel biomarkers of amyloid organ dysfunction and treatment response

There are other ways in which proteomics could be of aid in the management of amyloid diseases, namely through the identification of novel organ dysfunction biomarkers and through the evaluation of the amyloidogenic precursor [5]. Biomarkers are now a mainstay for assessing amyloid organ damage (especially regarding the heart and kidney), prognosis, and response to treatment [3]; novel molecules, however, are still needed to specifically differentiate amyloidosis from other conditions and to sensitively monitor active proteotoxic damage in some affected organs, such as kidney and nervous system. The way ahead, in this field, is still long, but encouraging evidence is available. Proteomic analysis of fat tissue, for example, has disclosed several resident proteins whose levels are altered in presence of the disease [5,15]. Evaluation of the specific proteome changes in the major target sites would be a powerful way to identify altered species correlated with disease presence and severity. In addition, the molecular profile of the fibrillar deposits might provide useful information to predict sensitivity to antiamyloid therapies and to allow monitoring their efficacy [3]. It will be especially interesting to study the distribution of amyloid-associated proteins, either working as chaperones (e.g. clusterin) or protecting fibrils from degradation (e.g. serum amyloid P), under the hypothesis that these species may modulate the biological properties of the deposits. Novel prognostic and diagnostic information could also derive from the proteomic characterization of the deposited precursor. In fact, some biochemical features, such as fragmentation in the case of TTR, are already known to influence clinical presentation, including affinity for diagnostic tracers, organ involvement, and age at onset [2,16,17]. In the case of immunoglobulin LCs, high-throughput proteomic identification of clonotypic peptides allows assigning germ-line gene usage in the majority of patients; this, in turn, has an impact on disease pathophysiology and tissue tropism [2,14]. For genetically determined forms such as ATTR, deposited amyloidogenic variants can be identified (using classical searching algorithms and augmented databases for known mutations or tools such as de novo sequencing algorithms for novel substitutions) [5,18]. This has important implications, since the different variants can have distinct organ tropism and prognosis [5]. Moreover, MS is being increasingly exploited for evaluating the amyloidogenic precursors at diagnosis and during follow-up. In the case of monoclonal LCs, for example, monitoring the mass of each patient’s protein increases the specificity, with analytical sensitivities comparable or superior to those of the single current clinical assays, namely serum protein electrophoresis and immunofixation and immunonephelometry [19–21]. Studies are ongoing to validate the quantification of the amyloid LC by MS.

6. How can we make the best use of proteomics for improving amyloidosis treatment?

Overall, proteomics is a powerful tool to identify the amyloid proteins, the amyloid-associated species, and proteins correlated with organ dysfunction. Its potential for improving the clinical management and treatment is manifold, but this approach is conceptually and practically distinct from the classical procedures used in the clinic, pathology, and clinical biochemistry laboratories, and consensus guidelines still need to be drawn for translating the proteome profile into a validated instrument for the physician’s practice. Controversies sometimes exist in the clinical interpretation of the protein lists, and the coexistence of multiple potentially amyloidogenic proteins opens the issue of differentiating between serum contamination and actual co-deposition of different precursors, which now appears to be a more diffuse phenomenon than once thought [2], with critical implications for therapeutic management.

Prospectively, proteomics is going to improve treatment for amyloidoses because it is increasing our understanding of the disease mechanisms at multiple levels. The evidence disclosed by this approach is reshaping our knowledge regarding both the epidemiology and the pathogenesis of these conditions. No other tools can provide, in a single analysis, a comparable wealth of information, and the progressive availability of bioinformatic instruments will likely allow digging even deeper in data set repositories, provided that meaningful questions are formulated.

Declaration of interest

G. Merlini was supported by Associazione Italiana per la Ricerca sul Cancro special program ‘5 per mille’ (number 9965); Fondazione Cariplo (2013-0964); the Italian Ministry of Health, research target project ‘Cardiac amyloidosis: molecular mechanism and innovative therapies for a challenging aging-related cardiomyopathy’ (RF-2013-02355259). F. Lavatelli was supported by Fondazione Cariplo (2016-0489). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

This manuscript was supported by Associazione Italiana per la Ricerca sul Cancro, Fondazione Cariplo and the Italian Ministry of Health.