Fibrinogen alpha amyloidosis: insights from proteomics

Abstract

Introduction:

Systemic amyloidosis is a diverse group of diseases that, although rare, pose a serious health issue and can lead to organ failure and death. Amyloid typing is essential in determining the causative protein and initiating proper treatment. Mass spectrometry-based proteomics is currently the most sensitive and accurate means of typing amyloid.

Areas Covered:

Amyloidosis can be systemic or localized, acquired or hereditary, and can affect any organ or tissue. Diagnosis requires biopsy, histological analysis and typing of the causative protein to determine treatment. The kidneys are the most commonly effected organ in systemic disease. Fibrinogen alpha chain amyloidosis (AFib) is the most prevalent form of hereditary renal amyloidosis. Select mutations in the fibrinogen Aα (FGA) gene lead to AFib.

Expert Commentary:

Mass spectrometry is currently the most specific and sensitive method for amyloid typing. Identification of the mutated fibrinogen alpha chain can be difficult in the case of “private” frame shift mutations, which dramatically change the sequences of the expressed fibrinogen alpha chain. A combination of expert pathologist review, mass spectrometry, and gene sequencing can allow for confident diagnosis and determination of the fibrinogen alpha chain mutated sequence.

1. Introduction

Fibrinogen alpha chain amyloidosis (AFib) is the most common form of hereditary renal amyloidosis in the US [1]. Although it dominantly effects the kidneys it should be considered a systemic disease that may lead to neurological and cardiac dysfunction [2]. Treatment for AFib includes dialysis and eventually kidney or kidney and liver transplantation. Fibrinogen is produced by the liver, so the mutated form of the protein will continue to be produced and deposited unless the organ is transplanted.

Amyloidosis is not a single disease, but a group of disease that all involve the deposition of misfolded protein in extra-cellular regions. There are 36 known causative proteins and identifying the precursor protein is critical to providing proper treatment and avoiding mistreatment. Currently, the use of mass spectrometry for protein identification is the best method for clinical amyloid typing. In this review we discuss amyloidosis diagnosis in general and AFib amyloidosis diagnosis more specifically to highlight the benefits of proteomics applications in the clinical setting and review the limitations that still remain.

2. Fibrinogen

Fibrinogen is a 340 kDa soluble glycoprotein and an essential component of blood coagulation cascade. It is produced in the liver at a rate of 1.7 – 5 grams per day [3]. Approximately 75% of the protein is secreted into plasma and the rest goes to the lymph and interstitial fluid [4].

2.1. Structure of fibrinogen

Fibrinogen is a dimer of trimers, each trimer is composed of one Aα, one Bβ, and one γ chain. The three chains are encoded by three closely linked genes (FGA, FGB, and FBG) that sit in the same 50 kb region between q23 and q32 on chromosome 4 [5]. Expression of the three genes shares a common regulatory pathway and this coordinated control results in near equal levels of mRNA for each chain per hepatocyte [6]. Translation and post translational modification of the chains is also highly coordinated.

Fibrinogen assembly begins with formation of a trimer composed of one copy of each chain. Then two linear symmetrical trimers are linked to one another through non-covalent and disulfide bonds at the amino terminals. The fibrinogen hexamer is a long flexible protein array with three nodules (Figure 1) [7]. Each molecular contains two outer D domains connected to a central E domain by coiled-coil segments [8, 9]. D domains are composed of the carboxy termini of Bβ and γ chains of one trimer. The E domain contains the N-termini of all 6 chains. There are also two compact Aα C-terminal domains that are connected to the hexamer with flexible unstructured connector regions. Aα C-terminal domains interact with each other and with the central E domain.

Figure 1.

The structure of a fibrinogen molecule based on crystal structure 3GHG showing D, E and alpha C regions [7].

2.2. Biological Role of Fibrinogen

Fibrinogen is vital to hemostasis and serves as an acute phase reactant. In the event of trauma, pregnancy, or tissue inflammation fibrinogen levels can increase 2 – 10 fold and fibrinogen is converted into the insoluble protein fibrin [10]. Conversion of fibrinogen to fibrin causes a major conformational change in which the Aα C-terminal domains dissociate from the central region and become available for intermolecular interactions [11]. Thrombin, a serine protease, starts the conversion of fibrinogen to fibrin by cleaving portions of the N-termini of Aα and Bβ chains generating fibrin strands and two small polypeptides. Fibrin strands polymerize and are crosslinked to other fibrin strands forming a molecular scaffold to support platelet aggregation in blood clot formation [12, 13]. Additionally, during acute phase reactant response fibrinogen it can also function as a bridging molecule in cell-cell interactions [8].

2.3. Fibrinogen mutations linked to disease

Mutations, both autosomal dominant and recessive, in the fibrinogen chain genes can cause a series of disorders most of which are related to clotting [10, 14, 15]. Congenital afibrinogenemia is a rare autosomal recessive inherited disorder which usually involves a non-functional mutation in both the maternal and paternal copies of either the FGA, FGB, or FBG genes [16, 17]. These individuals experience frequent and sometimes life-threatening episodes of bleeding and or thrombosis due to a lack of fibrinogen. Congenital hypofibrinogenemia is also a rare inherited disorder, but individuals only have a non-functional mutation in one of the two parental FGA, FGB, or FBG genes [18]. Blood may not clot normally due to a reduced level of fibrinogen and the lower the fibrinogen levels the more symptomatic. Congenital dysfibrinogenemia is an autosomal dominant inherited disorder in which fibrinogen is composed of a dysfunctional protein made by a mutated gene plus a normal fibrinogen made by a normal gene [19]. Fibrinogen levels appear normal by immunological measurements, but when measured by clot formation methods levels are approximately 50%. This disorder has a reduced penetrance and only some individuals show symptoms of abnormal bleeding or thrombosis.

Hereditary fibrinogen alpha chain amyloidosis is a rare autosomal dominant inherited disorder resulting from a mutation in one of the two copies of the FGA gene. The mutated fibrinogen alpha chain can circulate in the individuals blood and accumulate in the kidneys forming amyloid fibrils which leads to organ dysfunction and can be fatal. What determines if a FGA mutation results in amyloidosis or one of the other diseases is not known, but recent studies have started to answer these questions.

3. Amyloidosis

Amyloidoses are a group of diseases that result from the systemic or localized deposition of amyloid fibrils in the extra-cellular spaces of tissues causing organ dysfunction and potentially death. These diseases can affect any organ and are often underdiagnosed. Amyloid fibrils are insoluble protein aggregates that form when a protein becomes misfolded. There are 36 biochemically diverse proteins that are accepted to cause amyloidosis, but the common feature of these proteins is the propensity to form beta pleated sheets under appropriate conditions (Table 1) [20]. The beta pleated sheets align in an antiparallel fashion and through hydrogen bond interactions form long protofilaments. Protofilaments interact via their side chains to form fibrils [21]. These rigid fibrils resist proteolysis in the affected tissues and organs [22, 23].

Table 1.

Amyloidosis subtypes and their precursor proteins.

Amyloid Type	Precursor Protein	System or Localized	Acquired or Hereditary	Target Organs
AL	Immunoglobulin light chain	S, L	A, H	All organs, usually except CNS
AH	Immunoglobulin heavy chain	S, L	A	All organs except CNS
AA	Serum amyloid A	S	A	All organs except CNS
ATTR	Transthyretin, wild type	S	A, H	Heart mainly in males, Lung, Ligaments, Tenosynovium, PNS, ANS, heart, eye, leptomen
Aβ2M	β2-Microglobulin, wild type	S	A, H	Musculoskeletal System, ANS
AApoAI	Apolipoprotein A I, variants	S	H	Heart, liver, kidney, PNS, testis, larynx (C terminal variants), skin (C terminal variants)
AApoAII	Apolipoprotein A II, variants	S	H	Kidney
AApoAIV	Apolipoprotein A IV, wild type	S	A	Kidney medulla and systemic
AApoCII	Apolipoprotein C II, variants	S	H	Kidney
AApoCIII	Apolipoprotein C III, variants	S	H	Kidney
Agel	Gelsolin, variants	S	H	PNS, cornea
ALys	Lysozyme, variants	S	H	Kidney
ALECT2	Leukocyte Chemotactic Factor-2	S	A	Kidney, primarily
AFib	Fibrinogen α, variants	S	H	Kidney, primarily
ACys	Cystatin C, variants	S	H	PNS, skin
ABri	ABriPP, variants	S	H	CNS
ADan	ADanPP, variants	L	H	CNS
Aβ	Aβ protein precursor	L	A, H	CNS
AαSyn	α-Synuclein	L	A	CNS
ATau	Tau	L	A	CNS
APrP	Prion protein	L, S	A, H	CJD, fatal insomnia, GSS syndrome
ACal	(Pro)calcitonin	L	A	C-cell thyroid tumors
AIAPP	Islet amyloid polypeptide	L	A	Islets of Langerhans, insulinomas
AANF	Atrial natriuretic factor	L	A	Cardiac atria
APro	Prolactin	L	A	Pituitary prolactinomas, aging pituitary
AIns	Insulin	L	A	Iatrogenic, local injection
ASPC	Lung surfactant protein	L	A	Lung
AGal7	Galectin 7	L	A	Skin
ACor	Corneodesmosin	L	A	Cornified epithelia, hair follicles
AMed	Lactadherin	L	A	Senile aortic media
AKer	Kerato-epithelin	L	A	Cornea, hereditary
ALac	Lactoferrin	L	A	Cornea
AOAAP	Odontogenic ameloblast-associated protein	L	A	Odontogenic tumors
ASem1	Semenogelin 1	L	A	Vesicula seminalis
AEnf	Enfurvitide	L	A	Iatrogenic
ACatK	Cathepsin K	L	A	Tumor associated

Situations that have been determined to cause a protein to form amyloid fibrils in vivo include a prolonged increase in protein concentration, proteolytic remodeling, and instability due to an intrinsic factors or mutation [22, 24, 25]. These events coupled with a breakdown of the protein quality control system, which is vital to protein homeostasis, can result in amyloid fibril formation and deposition. The optimal conditions for amyloid formation could be why this group of diseases predominantly affect older patients as their chaperone systems are commonly weaker [25].

Besides the abundant fibril protein, amyloid deposits also commonly consist of proteoglycans, glycosaminoglycans, serum amyloid P component (SAP), and several apolipoproteins [22, 26]. The glycans stabilize the fibrils through interactions with the extracellular matrix [22]. SAP binds to amyloid fibrils and is key to the resistance of amyloid deposit degradation [27].

Amyloid formation and deposition clearly result in tissue and organ damage. As amyloid deposits accumulate tissue architecture is disrupted and healthy tissue is displaced, which interferes with the function of the affected organs [28]. However, it is believed the protofibrils exert a more toxic effect than the mature fibrils [24]. This could explain why patients that respond to therapy show clinical improvement despite unchanged amyloid deposits [25].

Amyloidosis is a highly heterogeneous disease. As mentioned above there are 36 human amyloid proteins of which 14 are only found in systemic disease, 19 only in localized amyloidosis, and 3 that can be either localized or systemic. In addition, iatrogenic amyloidosis, which is localized, can occur at the site of injection when a patient is using a protein or peptide drug subcutaneously [29, 30, 31]. Amyloidosis subtypes are named for the fibril protein where the letter A precedes an abbreviation for the precursor protein. For example, in the case amyloidosis caused by deposition of immunoglobulin (Ig) light chains the fibril protein would be AL and the disease would be AL amyloidosis [20].

Amyloidosis can be hereditary or acquired [32, 33]. In the case of hereditary disease, the variant information, such as the substitution or deletion in the native protein, is added to the disease subtype name. For example, amyloidosis caused by fibril protein transthyretin with a single amino acid mutation of V30 to M would be ATTR V30M amyloidosis.

In some cases, a precursor protein can cause either acquired or hereditary disease. For example, wild type transthyretin is the precursor protein for acquired/senile ATTR amyloidosis which usually occurs in older patients and a hereditary form of ATTR amyloidosis (ATTRv) is caused by inherited mutated forms of transthyretin [34]. Hereditary amyloidoses are caused by inheritance of a mutated gene in an autosomal dominant fashion. There are eleven proteins that have known variant forms associated with amyloidosis, these are transthyretin [35], apolipoprotein AI [36], apolipoprotein AII [37], apolipoprotein C II [38], apolipoprotein CIII [39], fibrinogen alpha chain [40], lysozyme [41], gelsolin [42], beta-2-microglobulin [43], cystatin C [44], and immunoglobulin light chain [45].

3.5. Renal Amyloidosis

The most common organ involved in systemic amyloidosis is the kidney. The most common forms of systemic disease in the US are AL amyloidosis (~85 %) and serum amyloid A protein amyloidosis (AA amyloidosis). AL amyloidosis is associated with an underlying clonal plasma cell disease, but a true neoplasm is only diagnosed in less than 10-15 % of patients [46]. AA amyloidosis is the result of a usually prolonged inflammation due to a chronic inflammatory disease or chronic infection [25, 47]. The prominence of disease subtypes varies depending on the country or region of the world.

These are therefore the most common forms of renal amyloidosis and affect ~ 90% of all renal amyloidosis patients. The remaining patients have less common forms of amyloidosis including leukocyte chemotactic factor 2 (ALECT2) and hereditary forms of amyloidosis such as fibrinogen alpha chain (AFib) , Apo AI, Apo AII, Apo AIV (AApo AI/AII, AIV), transthyretin (ATTR), lysozyme (ALys), gelsolin (AGel), and leukocyte chemotactic factor 2 (ALECT2) [48].

Patients with amyloid deposits in the kidneys will commonly present with proteinuria with or without renal insufficiency. Renal involvement is often the first clinically detectable manifestation of systemic disease. Deposits can occur in any renal compartment, but glomerular deposits are the most common. Amyloid deposits in the interstitium and extraglomerular vessels can also occur.

4. Diagnosis of Amyloidosis

Diagnosis of amyloidosis requires a biopsy and histological evaluation. The procedures for amyloidosis diagnosis can be performed on any tissue or an abdominal fat pad aspirate rather than an invasive biopsy [49].

First, if amyloidosis is suspected, the histological sections of the specimen are stained using Congo red dye. Amyloid deposits have an orange-red color under bright light and produce apple-green birefringence under polarized light (Figure 2) [50]. Congo red positive regions can also be identified by fluorescence microscopy using a Texas Red filter. The use of fluorescence microscopy has been proposed as an additional step to assist in confident diagnosis (Figure 3) [51, 52, 53, 54].

Figure 2.

A section of a kidney biopsy with extensive fibrinogen alpha chain amyloid deposits in the glomerular regions. In the H&E stained section (A) and the Congo red stained section (B) an arrow indicates one of the glomeruli with nearly complete replacement of healthy tissue by amyloid deposits. Amyloid deposits are salmon colored in the Congo red stained section.

Figure 3.

A Congo red stained glomerular region of a kidney with extensive fibrinogen alpha chain amyloid deposits viewed by bright field microscopy (A) and viewed by fluorescence microscopy (B). Under fluorescence the amyloid deposits are bright red and easily selected for laser microdissection. The green line around the glomeruli indicates this region was selected for microdissection, protein extraction, digestion, and analysis by mass spectrometry

Unfortunately, Congo red staining is not a routine stain and is only done when amyloidosis is suspected. Amyloidosis could be missed if the stain is not performed and this is one reason the diseases are likely under diagnosed. In addition, Congo red can be technically challenging, and overstaining can yield false positive results. Over decades there have been many studies conducted to optimize Congo red staining in order to minimize false positive and false negatives [55, 56]. However, the results of these studies have been at times contradictory and Congo red continues to be a special stain that requires a experienced histotechnologist to perform [57].

In renal amyloidosis on H&E stained sections of kidney amyloid deposits can usually be identified by their “hyaline” appearance (Figure 2) [58]. Suspected amyloid biopsies still require a Congo red stain to confirm the presence of amyloid. Early stage deposits may not be detectable without the Congo red stain, in those situations electron microscopy may be of help.

In addition to Congo red staining, immunohistochemical (IHC) staining using antibodies against proteins present in all amyloidosis deposits including apolipoprotein E and SAP is used as part of the initial amyloidosis diagnosis process [59]. However, this method can be limited by antibody availability, cross-reactivity, and sensitivity of the antibodies depending upon the abundance of these general amyloid markers in different deposits [60]. In general, Congo red staining with multiple visualization methods and other special stains are used together for a confident diagnosis of amyloidosis.

4.1. Amyloid Type Determination

Once a pathologist has determined that there is amyloid deposition in a tissue the next step is determination of the causative protein. As discussed, amyloidosis is a heterogeneous group of discrete diseases, but subtypes differ in pathogenesis and approach to clinical management. Determining the amyloidogenic protein is critical to ensure patients receive the proper treatment. The most widely used technique for determining the amyloidosis type is IHC using antibodies against the known causative proteins because IHC is a technique used in most pathology laboratories [61, 62, 63, 64, 65]. However, IHC is limited often due to a lack of available antibodies, lack of sensitivity, or cross-reactivity. Additionally, in the case of hereditary amyloidosis mutation sites may disrupt the epitope region resulting in a false negative.

Approximately a decade ago a routine method for the application of mass spectrometry-based proteomics was utilized for the diagnosis and typing of amyloidosis [66]. This method is currently the most specific and sensitive way to determine amyloidosis type [67].

5. Mass Spectrometry-based Amyloid Typing

5.1. FFPE Tissue Proteomics

In the early 2000’s mass spectrometry was utilized for the typing of amyloidosis from formalin fixed paraffin embedded (FFPE) tissue specimen. FFPE tissue had not been widely used for proteomics studies due to suboptimal methods for protein extraction, concerns about modifications introduced by formalin fixation, instability or loss of post translational modifications during tissue preparation, the unknown effect of storage time on the tissue, and the variability in tissue specimen.

Multiple studies have shown that storage time ranging from less than 1 to 15 years has no significant effect on the protein level and types of proteins extracted from FFPE tissue [68, 69]. Additionally, analyses of paired fresh frozen and FFPE tissue samples have found no significant difference in the results [70, 71].Advancements in tissue preparation have allowed for the extraction of protein from small amounts of FFPE tissue [70, 72, 73, 74]. Quantitative analysis of post translational modifications on proteins extracted from FFPE tissue has also been successful [75]. Many of these topics and the recent advances in the use of FFPE tissue in proteomics have been reviewed in depth previously [76, 77, 78, 79, 80].

The integration of proteomics into clinical diagnostics has been limited, but with the advancements in FFPE proteomics the development and validation of clinical proteomic tests is growing. FFPE blocks are the most widely available source of tissue for diagnostic testing. As outlined by others standardization of FFPE preanalytical, analytical, and post analytical workflows could increase the utility of proteomics in the clinical setting [81].

5.2. Amyloid Typing with Mass Spectrometry

The earliest methods of amyloid typing by mass spectrometry required 6 – 30 sections of an FFPE tissue sample and took multiple weeks for the sample preparation [82, 83]. Following extraction of the protein from the FFPE tissue sections reverse phase liquid chromatography separation was used and fractions corresponding to a UV signal were collected. Those fractions were digested into peptides and prepared for liquid chromatography tandem mass spectrometry (LCMS).

Approximately, a decade ago a method was developed by Dogan and colleagues that coupled laser microdissection (LMD) to LCMS for the typing of amyloid in FFPE tissue samples [66]. Using this method the amyloid type was identified in 100% of a 50 case training set and 98% of a validation set of 41 cardiac amyloidosis cases [67]. All 91 cases used had been well-characterized by gold standard clinical workflows. The addition of LMD allowed for the enrichment of the amyloid deposits and resulted in a more sensitive and specific typing method.

Briefly, FFPE tissue sections are cut onto DIRECTOR™ slides for ease of laser microdissection and stained with Congo red dye. The Congo red positive amyloid deposit regions are identified under bright field and fluorescence microscopy and then laser microdissected from the slide directly into the cap of a microcentrifuge tube containing the sample preparation buffer (Figure 3). A minimum of 50,000 μm² of tissue is collected in triplicate. Tissue samples are heated to reverse formalin induced crosslinking and sonicated to homogenize the tissue and extract protein. Extracted proteins are reduced, alkylated and digested with trypsin to generated LCMS amenable peptides.

Peptides are loaded onto a reverse phase nano flow column and gradient eluted into a Orbitrap mass spectrometer. The resulting data is searched using a custom annotated human database. The search includes potential modifications introduced during fixation such as formyl, methyl, and methylol. Results are organized so that proteins are listed in order from highest to lowest number of high confidence peptide spectral matches (Table 3). The most abundant amyloidogenic protein identified in all replicates determines the subtype. Presence and level of apolipoprotein E and serum amyloid P based on peptide spectral matches are also evaluated as these are general protein markers for all types of amyloidosis.

Table 3.

Example results from a set of six AFib amyloidosis patients. The table shows the number of peptide spectral matches per protein for each patient and displays the 14 most abundant proteins detected. In general, the more peptide spectral matches the more abundant the protein. Amyloid protein markers serum amyloid P and apolipoprotein E are detected in all samples at a relatively high level. Fibrinogen alpha chain is one of the most abundant proteins and the most abundant recognized causative protein.

Protein Name	# Peptide Spectral Matches
Fibrinogen alpha chain (wild type)	45	35	67	27	67	44
Vitronectin	65	31	30	29	37	44
Complement C3	48	20	31	4	21	53
Serum albumin	11	7	8	10	19	92
Serum amyloid P-component	26	15	28	18	15	41
Apolipoprotein E	31	24	21	9	19	21
Fibrinogen alpha chain (E545V mutation)	25	15	20	12	10	22
Complement factor related protein H1	23	12	14	6	15	14
Keratin, type I cytoskeletal 10		60
Actin, cytoplasmic	12	3	9	2	9	17
Clusterin	20	10	7			24
Complement factor H related protein 5	9	8	11		12	18
Complement component C9	13	10			7	20
Vimentin	10	5	4	7	8	11

This method is a clinically validated test used for amyloid protein typing. Vrana et al found that 98% (40/41) of cases analyzed for amyloidosis classification by this method were accurately typed [67]. In comparison IHC was only able to correctly classify 42% of those cases. Another retrospective study demonstrated the utility of this method for amyloidosis typing on 127 renal biopsies [84].

In the past decade similar methods have been implemented at other institutions and applied to many types of tissue including bone marrow, fat pad aspirates, and a variety of organ biopsies [48, 85, 86]. Although the cost to establish mass spectrometry based amyloid typing is high the long-term cost per test for the patient is low in comparison to the patient and more precise. Large studies comparing LMD coupled with LCMS with standard diagnosis methods including IHC have shown the former to be more sensitive and specific. LCMS is direct detection of the peptides generated by enzymatic digestion of the proteins within the amyloid deposit. IHC is dependent upon antibodies of varying specificity and sensitivity. Application of the LMD LCMS to amyloid typing has resulted in the identification of novel amyloid types [30, 39, 87] and aided in the discovery of previously unknown hereditary mutations linked to amyloidosis [43, 88].

6. AFib Amyloidosis

AFib amyloidosis is a hereditary form of amyloidosis caused by certain mutations in the FGA gene [40]. There are no reported cases of amyloidosis involving the other fibrinogen chains. Although the disease is rare it is the most prevalent form of hereditary amyloidosis with renal involvement. AFib amyloidosis has primarily been considered a renal disease, but there is evidence of visceral, vascular, cardiac, and neurologic involvement, which suggests it be considered a systemic disease [2].

As with all forms of amyloidosis effecting the kidneys, patients presenting with AFib amyloidosis will present with proteinuria, but also hypertension. Individuals with AFib amyloidosis often experience rapid deterioration to end stage renal failure and dependence on dialysis can occur in 1 – 5 years [89]. AFib has a very characteristic histology in that there is near replacement of the glomeruli by amyloid and no involvement of the interstitial and vascular regions [58, 89].

6.1. AFib Amyloidosis Mutations

Protein aggregate extracts have shown a fragment of the Aα- chain consisting of residues 500 – 580 [40, 90]. The protease sensitive nature of the Aα- chain carboxy terminus may facilitate progression of amyloid deposition. All known Aα- chain mutations linked to amyloidosis are within residues 517 – 555 and there are approximately 15 reported mutations to date (Table 2) [40, 91, 92, 93, 94, 95, 96].

Table 2.

Amyloidogenic Fibrinogen Aα Chain Mutations

Name	Sequence Variant (mRNA)	Codon Change	Location	Ethnic Group	References
Met517_Phe521delinsGlnSerfs	c.1606_1620 delATGTTAGGA GAGTTT insCA	Frame shifting mutation	5' end of exon5	Korean	[97]
Gly519Glufs*30	c.1611delA	Frame shifting mutation	5'end of exon5	French	[108]
Phe521Leufs*28	c.1620delT	Frame shifting mutation	5' end of exon5	French	[109]
Phe521Serfs*27	c.1619_1622delTTGT	Frame shifting mutation	5' end of exon5	Arab	[108]
Val522Alafs*27	c.1622delT	Frame shifting mutation	5' end of exon5	French	[94]
Glu524Glufs*25	c.1629delG	Frame shifting mutation	5' end of exon5	American	[96]
Glu524Lys	c.1627G>A	GAG>AAG	5' end of exon5	Caucasian	[108]
Thr525Thrfs*24	c.1632delT	Frame shifting mutation	5' end of exon5	Chinese	[89]
Glu526Lys	c.1633G>A	GAG>AAG	5' end of exon5	Russian	[108]
Glu526Val	c.1634A>T	GAG>GTG	5' end of exon5	Northern European	[95]
Thr538Lys	c.1670C>A	ACA>AAA	5' end of exon5	Chinese	[89]
Glu540Val	c.1676A>T	GAA>GTA	5' end of exon5	German	[89]
Pro552His	c.1712C>A	CCT>CAT	5' end of exon5	Afro-Caribbean	[89]
Arg554Leu	c.1718G>T	CGT>CTT	5' end of exon5	Peruvian, African American	[40]
Gly555Phe	c.1720_1721delinsTT	GGT>TTT	5' end of exon5	Norwegian	[108]

The most common mutation is a substitution of valine by glutamic acid at position 526 [95]. This missense mutation has been identified in families around the world. In comparison all of the amyloidogenic frameshift mutations that have been reported are isolated to a single family [88, 94, 96, 97, 98]. For a full list of reported mutations involved in AFib amyloidosis refer to Table 2.

Garnier et al, found an amyloid-prone motif in the sequences of Aα- chain frameshift variants. The VLITL motif located in the C-terminal end of the mutant proteins was conserved across all known frameshift variants. Synthetic mutated Aα- chain peptides were generated with and without the VLITL motif and subjected to fibril induction. The peptide solutions were analyzed by fluorescence and electron microscopy to determine if fibrils formed. Only those sequences with the VLITL motif resulted in fibril formation indicating this motif is key to amyloid formation.

There are no reported cases of AFib amyloidosis that involve the wild type or full-length form of the protein. Additionally, the mutant Aα- chain has not been detected in the plasma of AFib amyloidosis patients [94]. The reasons for this are not understood, but it has been proposed that the metabolism of the mutated protein is faster in comparison to the wild type protein.

AFib mutations have variable penetrance and a history of renal disease and amyloidosis is absent in approximately 50% of patients so amyloid typing and gene sequencing are vital for diagnosis [89]. Following gene sequencing to confirm FGA gene mutation the presence of the specific sequence in the amyloid deposit can be confirmed in the mass spectrometric data.

6.2. AFib Amyloidosis Treatment

Kidney transplant has been the long-standing treatment for patients with AFib amyloidosis who progress to end stage renal disease. However, since the mutated protein continues to be produced by the liver, damage to other tissues will continue and amyloid deposition on the renal graft will likely occur [2, 99]. A combined kidney-liver transplant has been performed on patients with AFib amyloidosis. The combined transplant has been shown to halt disease progression. It has been suggested that early liver transplant or kidney-liver transplant before renal failure be considered [2]

As has been previously discussed generally, accurate amyloidosis typing is critical to ensure patients receive the proper treatment. In the case of renal amyloidosis this is more complicated because of the large number of amyloidosis diseases that manifest in the kidney.

7. Limitations of Amyloidosis Typing by Mass Spectrometry

7.1. Identification of Mutant Causative Proteins

Although amyloidosis typing by mass spectrometry has greatly improved diagnosis it is not a perfect solution. One major limitation of mass spectrometry-based proteomics that is often discussed is its dependence upon a well curated database. Mass spectra was searched against a database using one or more of several available algorithms. However, if a protein expressed differs in amino acid sequence from the sequence in the database or that protein is not in the database at all it will not be identified.

This has led to the development of proteogenomics in which proteomics is integrated with next generation sequencing and transcriptomics to generate databases specific to the cell line, organism, or patient being studied [100]. This approach is ideal for in-depth studies of patient or disease specific proteomes but requires resources and time for sequencing.

In many situations a completely unique database is not required. Some proteomics software interfaces now make it possible to use a standard proteome database, but search for single amino acid mutations in the same way post translational modifications are considered [101]. Other proteomics software is capable of partially de novo sequencing peptides taking into consideration mutations [102].

In amyloidosis typing the data is first searched against a curated human database. This database can include known mutated sequences for amyloidogenic proteins. The initial search results are analyzed to confirm the presence of serum amyloid P and apolipoprotein E and determine the most abundant amyloidogenic protein. If the amyloidosis subtype is hereditary or could be hereditary the LCMS data can be searched again using a focused database which only contains the protein sequences identified in the initial analysis. The second search will incorporate either partial de novo sequencing or all single amino acid mutations as post translational modifications.

This could lead to the identification of a missense mutations which are common in many hereditary amyloidosis including AFib and ATTR amyloidosis. However, with frameshift mutations which are also common in AFib amyloidosis, these approaches would not reveal the mutation. For these reasons sequencing of the FGA gene following determination of AFib amyloidosis and then confirmation of deposition of the sequence by mass spectrometry is the best way to determine the theoretical sequence and then confirm expression.

7.2. Interference of Non-Causative Proteins

Another common issue with amyloid protein typing is the presence of other known causative proteins or very abundant non-amyloidogenic proteins that can limit the dynamic range of the analysis. Known amyloidogenic proteins other than the true causative protein could be detected in the specimen if laser microdissection is not precise, if there are large cell populations intermixed with the amyloid deposits, or if that protein is simply present in the surrounding tissue. Additionally, it is not uncommon to have some blood contamination despite careful laser microdissection of the amyloid deposits depending upon the biopsy. When there is blood contamination hemoglobin will be present at very high levels and fibrinogen alpha will also be detected and could be mistaken for the causative protein. However, if fibrinogen beta and gamma are also present, then the fibrinogen alpha is likely from blood contamination and not the causative protein.

There have been efforts to implement more complex algorithms for the evaluation of amyloid typing mass spectrometry data to automate data analysis and improve confidence of amyloid protein typing. Recently, Taylor et al. developed and tested an algorithm to improve the success rate of diagnosing AFib amyloidosis by proteomic analysis [103]. Mass spectrometry data from 1001 Congo red positive specimen was analyzed using this algorithm. They utilized the search algorithm MASCOT with a modified database that included known fibrinogen alpha amyloidogenic mutation sequences. The first data filtering step excluded proteins with a score below 80 and less than 2 unique peptides. Any cases lacking at least 2 of the 3 general markers serum amyloid P, apolipoprotein E, and apolipoprotein A IV were then excluded. To be classified as AFib amyloidosis the fibrinogen alpha MASCOT score had to be greater than the sum of the scores of fibrinogen beta and gamma. This step was included to minimize false positives due to blood contamination. Finally, the detection of a pathogenic variant of fibrinogen alpha was required for diagnosis as AFib. This method tried to address the issues of protein contamination and identifications of mutants. Although this method would miss unknown variants this is a major step in the right direction regarding more intelligent analysis of the amyloid proteomics data.

8. Expert Opinion

Amyloid typing by LCMS has made a significant impact on the lives of patients with amyloidosis or those suspected of having amyloidosis for approximately a decade. Through a relatively simple proteomics workflow the causative protein can be identified, and proper treatment can be administered. Since its first clinical implementation thousands of amyloid deposits have been analyzed by mass spectrometry for clinical reporting [29, 30, 31, 43, 48, 87, 88].

The clinical utility of amyloid typing in FFPE tissue by LCMS proves that mass spectrometry-based proteomics can be used in the clinic. Development of this test was motivated by the need of pathologists for a way to determine the causative protein. Collaboration between pathologists with a strong background in amyloidosis and proteomics specialists resulted in a validated clinical test.

The large datasets that result from years of testing can be analyzed retrospectively to gain a stronger understanding of the amyloid microenvironment. Differences in the proteome dependent upon amyloidosis type or tissue could be determined, which would help to refine the assay and increase the specificity of the analysis. This can include robust background proteomes for different FFPE tissue types that could allow for the development of algorithms that would increase the confidence of subtype determination as shown by Taylor et al. in the case of AFib amyloidosis.

In addition, no major efforts have been made to determine if there are any protein post translational modifications that play a role in the amyloid formation mechanisms. These may include glycan modifications. Besides proteins there are known to be proteoglycans and glycosaminoglycans present in amyloid deposits and could be characterized using mass spectrometry [104]. The roles of these biomolecules are not well understood and characterization of amyloid deposit glycomics could reveal additional information about how fibrils are formed and persist. There is likely much interplay between the protein and glycan components in amyloidosis.

More generally, the implementation of clinically validated tests that utilize mass spectrometry-based proteomics has been very limited. In comparison genomics and more recently transcriptomics has flourished in the clinical setting. There is a great need and desire for proteomics to move into the clinical sphere because the pathophysiology of disease is not completely explainable by genetic and transcriptional changes.

Mass spectrometry technology has advanced to the point that more than 12,000 proteins can be identified and quantified from a complex biological sample [105, 106]. The sensitivity of the instrumentation allows for complex analyses to be performed on very small amounts of material. These techniques have been applied to large sets of clinical samples for biomarker discovery, but translation of those targets to clinically robust tests has not occurred [107]. In many cases novel candidates are identified, but not scrutinized for their diagnostic or prognostic utility.

In the short and long term, it is vital that scientists in the field of proteomics work closely with clinicians and pathologists to determine where there are unmet needs. Proteomics specialists must determine what kind of tests are required by clinical colleagues such as diagnostic, prognostic, or monitoring. Also, is the presence of a protein or modification alone meaningful or is it the relative abundance of a species that is telling. It is also important that there is an understanding of what the outcome of a positive test would be for the patient because the outcomes can range from further testing to invasive procedures.

One simple example of a space that MS has already shown success in the clinical laboratory is in the replacement of IHC in cases where reliable antibodies are not available for important diagnostic markers. In many of those cases multiplexed targeted MS based qualitative or quantitative assays using FFPE would be very powerful and provide direct detection of proteins of interest. Implementation of proteomics in the clinic has vast potential and there are both simple and complex tests that will have high clinical impact.

Article Highlights.

• Systemic amyloidoses are a rare group of diseases that result from the systemic or localized deposition of amyloid fibrils in the extra-cellular spaces of tissues causing organ dysfunction and potentially death.

• Fibrinogen alpha chain amyloidosis (AFib) is the most common form of hereditary renal amyloidosis in the US.

• AFib mutations have variable penetrance and a familial history of renal disease and/or amyloidosis is absent in approximately 50% of patients so amyloid typing and gene sequencing are vital for diagnosis.

• Amyloid typing by LCMS has made a significant impact on the lives of patients with amyloidosis or those suspected of having amyloidosis for approximately a decade and shows the potential power of proteomics in clinical settings.

• In the short- and long-term implementation of proteomics within the clinical setting, it is vital that scientists in the field of proteomics work closely with clinicians and pathologists to determine where there are unmet needs.