14 Amyloid

Introduction

History

Amyloidosis is a disorder of protein folding, in which normally soluble proteins accumulate in the tissues as abnormal insoluble fibrils (or filaments), thus disrupting their function. The deposited proteins, generically known as amyloid, damage the structure and function of the tissues and so cause serious disease which is often fatal when it affects major organs.

The term amyloid means ‘starch-like’ and was first used in botany to describe the starchy cellulose material found in plants. It was identified by a botanical test for carbohydrate through its reaction with iodine. Starch gives a deep blue color when iodine is applied, whereas cellulose gives a violet color. This reaction was adapted by Virchow 1853, 1854) to tissue containing amyloid, which also gives a violet coloration and was often used in early studies of amyloidosis. Amyloid is still, to date, identified by its characteristic histological staining reactions in tissue.

At autopsy the cut surface of amyloid-affected organs may show a ‘waxy’ texture and was often described by early pathologists as ‘waxy degeneration’ or ‘lardaceous disease’ (von Rokitansky 1842). During the nineteenth century this appearance was a frequent post-mortem finding in patients suffering from chronic inflammatory diseases such as tuberculosis (TB), but amyloidosis itself was rarely diagnosed during the patient’s lifetime.

Friedrich and Kekule (1859) came to the conclusion that ‘amyloid livers and spleen’ were of albuminous nature due to the high proportion of nitrogen. Subsequent analyses confirmed that amyloid was predominately of a proteinaceous nature with a variable amount (about 1–5%) of various mucopolysaccharides. Knowledge of the exact composition was at that time hampered by the inability to separate amyloid from the normal extracellular ground substance.

Crystal violet and Congo red dyes were used in early histological techniques for the demonstration of amyloid (Bennhold 1922), but it wasn’t until 1925 that the dichroic effect of Congo red stained amyloid was noted (Neubert 1925) when viewed under polarized light. Subsequently, in 1927, the optical activity of Congo red stained amyloid giving the unique ‘apple green birefringence’ was described (Divery and Florkin 1927). This phrase is still in use today to describe the birefringence and dichroism that amyloid displays when viewed under crossed polarized light, although Howie et al. (2009) suggested the correct phrase should be of ‘anomalous colors’, which describes all the colors Congo red stained amyloid can show – from yellow to green to blue.

With the arrival of electron microscopy it was observed that amyloid had a unique fibrillary ultrastructure independent of anatomical site, quite different to any other ultrastructural fibrils described before. Eanes and Glenner (1968), using X-ray diffraction, showed that the protein within the fibrils was arranged in an anti-parallel β-pleated sheet (Fig. 14.1).

Figure 14.1 Schematic diagram of amyloid structure.

Since then it has been shown that all amyloid fibrils share a common β-core structure with polypeptides chains running perpendicular to the fibril long axis, regardless of the particular protein from which they are formed (Sunde 1997).

With the development of techniques for amyloid fibril extraction (Pras et al. 1968; Glenner et al. 1972) it was confirmed that the bulk of amyloid deposits were composed of protein. They also contain up to 15% of a non-fibrillary glycoprotein known as amyloid P component (AP), derived from and identical to the normal circulating plasma protein serum amyloid P component (SAP). SAP is a calcium-dependent, ligand-binding protein that forms a normal component of basement membranes and elastic fibers, and may have a function related to its binding to glycosaminoglycans (GAGs), fibronectin and other cellular components (Pepys & Baltz 1983). It belongs to the pentraxin family and becomes specifically and highly concentrated in amyloid deposits of all types. This binding to amyloid fibrils is used in patients with amyloidosis for radiolabeled SAP scintigraphy, a diagnostic technique that is used for quantitative monitoring of amyloid deposits. (Hawkins 1990). The generic SAP ligand on amyloid fibrils remains uncharacterized (Pepys 1992).

The composition of the GAGs includes various sulfates, heparin, chrondition and dermatan, which may be involved in amyloidogenesis. The presence of the carbohydrate moieties of GAGs provides a possible explanation for the staining reaction in some of the histological methods used for amyloid detection.

Ultrastructure

As mentioned earlier, electron microscopy (EM) played an important role in the identification of the composition of amyloid, showing its unique fibrillary arrangement. To this day, EM is one of the methods for identifying amyloid. Amyloid deposits appear as masses of extracellular, non-branched filaments usually in a random orientation, though occasionally in parallel arrays of a few fibrils. Each fibril consists of two electron-dense filaments 2.5–3.5 nm in diameter, separated by a 2.5 nm space, giving a total diameter of 8–10 nm, with variable length of up to several microns (µm) (Cohen & Calkins 1959; Glenner 1981) (Fig. 14.2a, b).

Figure 14.2a An electron micrograph of kidney from a case of amyloidosis with renal involvement showing amyloid (A), epithelial cytoplasm (Epi C), basement membrane (BM) and an endothelial cell (Endo C).

Figure 14.2b A higher magnification of Figure 14.2a where amyloid fibrils can be clearly seen.

Classification

There have been many attempts to identify and classify amyloid proteins in order to collate the endless variety of clinical manifestations, histopathological appearances and associated pathology. Until 1980, the main classification used for amyloid was that of Reimann et al. (1935), who divided amyloid types into four categories:

1. Primary amyloid occurring spontaneously in the absence of an apparent predisposing illness. It affects organs and tissues such as heart, muscle, skin, and tongue.

2. Secondary amyloid occurring in patients with chronic infective disease, such as syphilis and TB. Later, inflammatory diseases such as rheumatoid arthritis were also included in this group.

3. Tumor-associated amyloid.

4. Myeloma-associated amyloid.

As more and more amyloid-forming proteins were identified, this classification became obsolete and Husby et al. (1990) proposed guidelines for a better classification based on the identity of the amyloid fibril protein, which was adopted by the World Health Organization-International Union of Immununological Societies (WHO-IUIS) and forms the currently accepted classification of amyloid.

Today there are about 25 different unrelated amyloid-forming proteins (Table 14.1). Amyloidosis nomenclature uses the letter A to designate amyloid followed by an abbreviation of the name of the fibril protein.

Pathogenesis

The process that causes proteins to become involved in amyloid formation, converting them from normal functioning proteins into inert amyloid deposits is the focus of much research. There is little to connect the different types of proteins involved (Merlini 2003). In certain amyloid types, only a limited portion of the amyloid protein precursor forms the fibril – as is the case in Alzheimer’s disease; several amyloid proteins are rich in β-pleated sheet conformation in their native form; prion protein contains no β-pleated sheet and it is the formation of β-pleated sheets de novo that is the pathological event of the prion diseases. In some amyloidosis, the whole precursor protein may be involved, or there may be proteolysis of the precursor protein with liberation of a smaller amyloidogenic fragment as with AβPP. In transthyretin (TTR) amyloidosis, the circulating protein is a tetramer and a vital pathological event appears to be release of the monomer. The pathogenesis of amyloid has been recently reviewed (Sipe 2005).

Amyloidosis is considered to belong to the category of conformational diseases, because the pathological protein aggregation is due to the reduced stability and a strong propensity to acquire more than one conformation. It is thought that such a grouping helps to provide an understanding of the etiology or episodic onset of these diseases, and opens the prospect for common approaches to therapeutic stratagems in the same way that recognition of bacteria as the causative agents of many infections allowed the idea of antibiotics being useful in all such conditions, or of steroid therapy being of potential use for all inflammatory conditions (Carrell & Lomas 1997; Carrell & Gooptu 1998). In this context it is interesting to note the development of ‘designer’ peptides that bind to Aβ and to prion protein, preventing, and even reversing, the conformational change responsible for the respective disease processes (Soto et al. 2000). This concept is becoming increasingly accepted, and it is becoming evident that amyloid is but one, albeit definable, subgroup within a larger group of misfolded or altered protein deposits which are associated with human disease (Table 14.2).

Table 14.2 Protein conformation diseases

Conditions	Affected proteins	Associated diseases
Amyloidosis	25 known in humans – see Table 14.1
Serpinopathies	α₁-Antitrypsin	α₁-Antitrypsin storage disease
	neuroserpin
Hemoglobinopathies	Hemoglobulin	Sickle cell anemia
		Drug and aging induced inclusion body hemolysis
Lewy body diseases	α-Synuclein	Parkinson’s disease
Neuronal inclusion bodies	Tau	Alzheimer’s disease
		Pick’s disease
		Progressive supranuclear palsy
	Superoxide dismutase	Motor neuron disease, AML
	Ferritin	Familial neurodegenerative disorder
Hirano bodies	Actin	Alzheimer’s disease
Polyglutamine repeats	Huntingtin	Huntington’s disease
	Ataxin	Spinocerebellar ataxias
	Androgen receptor	Spinomuscular atrophies
Prion diseases	Prion protein	Creutzfeldt-Jacob disease (CJD)
		Variant CJD
		Gerstmann-Straussler-Scheinker
		Kuru
		Fatal familial insomnia
		Japanese CAA

Amyloidosis

In AL amyloidosis, previously known as ‘primary amyloidosis’, abnormal proteins, monoclonal light chains, are produced by plasma cells or B-cells and form amyloid deposits in the tissues. These can be either of kappa (κ) or lambda (λ) isotypes. Systemic AL amyloidosis is the most common form of clinical amyloid disease in developed countries and causes the most fatalities. In systemic amyloidosis, deposits can be present in any or all of the viscera, connective tissue and blood vessels walls, although intracerebral amyloid deposits are never found (Pepys 2006).

AL amyloidosis can also be localized when it is restricted to a particular organ or tissue, and usually has a benign prognosis. In the skin, the deposits cause benign lumps and can be excised or left untreated, but localized amyloid deposits are also common in the bladder and in pulmonary tissue, often causing obstruction which can lead to complications.

There are five possible types of heavy chain amyloidosis (AH); G, A, D, E and M – these are rare but not unknown.

AA amyloidosis, previously known as ‘secondary amyloidosis’, is an occasional complication of chronic infection and inflammatory conditions, characterized by an acute phase response in which production of serum amyloid A protein (SAA) is increased. SAA is an apolipoprotein produced in the liver and is an acute phase protein which is synthesized at increased levels in patients with diseases such as rheumatoid arthritis, TB, Crohn’s disease, familial Mediterranean fever (FMF) and other hereditary periodic fevers. Chronically high levels of SAA are a prerequisite for development of AA amyloidosis.

Hereditary systemic amyloidosis is a rare disorder that is difficult to treat and often fatal. This group of disorders occurs in small clusters around the world and has an autosomal dominant pattern of inheritance. Its most common cause is a mutation in the TTR (prealbumin) gene, which affects about 10,000 individuals worldwide. Over 100 mutations, most of which are amyloidogenic, are known in TTR. The major features of hereditary TTR amyloidosis include severe and ultimately fatal peripheral and/or autonomic neuropathy (familial amyloid polyneuropathy, FAP); cardiac involvement is also common. Other mutations associated with the disorder are those in the genes encoding apolipoproteins AI and AII, fibrinogen A α-chain, gelsolin, lysozyme, cystatin C and β-protein. In all these forms the variant protein is deposited as amyloid fibrils predominantly in the abdominal viscera, although cardiac and nerve involvement can occur.

All amyloidogenic mutations are dominant, but can display both variable penetrance and expressivity. Thus there may be marked differences in age of onset, amyloid deposition and clinical presentation, not only between families but also within families with the same mutation. In contrast to AA amyloidosis, where the concentration of the amyloid fibril protein SAA is raised but of normal structure, AL and hereditary amyloidosis are associated with abnormal protein which is inclined to refold as a β-pleated sheet, resulting in amyloidosis.

LECT2 amyloidosis was discovered while studying proteins with leukocyte chemotactic activity (Yamagoe 1996). The pathogenesis of LECT2 remains to be understood, but there is no evidence that LECT2 is an inherited condition though the first cases reported were Hispanic patients (Larsen et al. 2010).

Other diseases in which amyloid occurs

As outlined by Pepys in 2006, amyloid is a histological feature of Alzheimer’s disease and type 2 diabetes mellitus but, unlike systemic amyloidosis, it is not known whether the amyloid causes these diseases. In Alzheimer’s there is an abundance of intracerebral amyloid deposits composed of β-protein, though there is poor correlation between the quantity of amyloid and the cognitive impairment. However, mutations that result in abundant deposition of β-protein as amyloid may result in early-onset Alzheimer’s disease.

In patients with type 2 diabetes, amyloid is frequently found in the pancreatic islets of Langerhans, though this is not universal in all islets or in all patients with type 2 diabetes. As well as this, in diabetic patients amyloid can also be found at the site of insulin injection as a result of the injected insulin adopting a fibillary conformation in vitro (when subjected to certain physical or chemical stimuli such as heat or acidity), causing a localized cutaneous lump (Lonsdale-Eccles et al. 2009).

It is also frequently cited that transmissible spongiform encephalopathy (TSE) is an example of amyloidosis, although in fact amyloid has never been determined histopathologically in brains of patients with the disease, nor in cows with bovine spongiform encephalopathy (BSE).

There are other protein misfolding diseases that are not amyloid or amyloidosis; these diseases produce abnormal aggregates of proteins, for example Lewy bodies in Parkinson’s disease and polyglutamine repeats present in Huntington’s disease. However, it is unclear whether the removal of these aggregates alters the disease process.

Another protein deposition disease which is often confused with amyloid is light chain deposition disease (LCDD) which has a similar histological appearance. However LCDD deposits lack the affinity for Congo red stain and do not produce the characteristic green birefringence of amyloid under cross-polarized light. Under EM, LCDD deposits are granular, which aides the distinction (Gibbs et al. 2011).

Diagnosis

There is an increased awareness of amyloidosis nowadays and more patients are being recognized. However some patients are still overlooked. The diagnosis requires presence of amyloid in a tissue, and the gold standard technique is Congo red histology although EM may aid diagnosis. Biopsies are usually taken to investigate organ dysfunction, for example of the kidneys in nephrotic patients or of sural nerves in familial polyneuropathies. Amyloid is present in up to 90% of rectal and/or subcutaneous fat biopsies in systemic AA or AL amyloidosis; so much so that rectal biopsies or fine needle aspirates of subcutaneous tissue used to be the main method of screening (Westermark & Senkvist 1973; Pepys 1992). Nowadays techniques have improved greatly such that cardiac biopsies, after a suggestive echocardiogram, are becoming more popular. It must be noted that a negative biopsy does not exclude the possibility of amyloidosis due to sample selection and site. In rectal biopsies, amyloid is usually found in the walls of submucosal vessels, so if the full thickness of the muscularis is not obtained the deposits will go undetected.

The use of SAP scintigraphy allows in vivo diagnosis as well as the monitoring of progression and regression of the amyloid deposits with treatment (Hawkins et al. 1993; Hawkins 1994). Unfortunately SAP scintigraphy is unable to visualize amyloid within cardiac tissue because the heart is a moving organ. The bone scanning method DPD scintigraphy (^99mTc-3,3-diphosphono-1,2-propanodicarboxylic acid) was serendipitously discovered to have high affinity for cardiac TTR amyloid, and is currently under evaluation at the National Amyloidosis Centre as a method for visualizing cardiac involvement.

Differentiation between different amyloid types

With the recognition that different proteins form amyloid and are associated with different clinical syndromes came the need to identify particular fibril types histologically. Furthermore, treatment of amyloidosis is entirely type-specific; hence the correct identification of the fibril type is indispensable in clinical practice.

Methods of section pretreatment using trypsin or potassium permanganate before Congo red staining were devised (Wright et al. 1976, 1977). After such pretreatment some amyloids lose their affinity for Congo red, most notably AA amyloid, whereas AL amyloid is resistant. These methods were always equivocal in practice and have been rendered obsolete by the use of immunohistochemistry and other techniques to identify the particular amyloid fibril protein specifically and reliably.

It is vitally important to discriminate between AL, hereditary amyloidosis and AA amyloidosis, as their treatments are entirely different. AL treatment is aimed at ablating the B-cell clone responsible for the amyloidogenic free light chain production using cytotoxic drugs. Hereditary amyloidosis treatment sometimes involves organ transplantation. Therapy for patients with AA amyloidosis involves measures to reduce SAA production by treating the cause of the underlying inflammation.

The tools available today to differentiate the amyloid type include direct assessment of the fibril type by immunohistochemistry, proteomics and, occasionally, fibril sequencing. Indirect, but very helpful, investigations include searches for monoclonal immunoglobulins using conventional electrophoresis and immunoassays of serum and urine, the serum free light chain assay, assessment of the hepatic acute phase response by measuring SAA and CRP and, where indicated, genetic sequencing of genes known to be associated with hereditary amyloidosis or the periodic fever syndromes. All these techniques are frequently employed for a patient with amyloidosis.

Demonstration

In hematoxylin and eosin (H&E) stained sections amyloid appears as an amorphous, eosinophilic, extracellular, faintly refractive substance that sometimes displays green birefringence under polarized light. However, it should be noted that collagen also has this appearance under polarized light in a H&E-stained section. Amyloid can also be weakly birefringent using a powerful light source when stained with periodic acid-Schiff. Whilst large deposits of amyloid can be observed with a H&E-stained section, the small deposits, for example in vessels in rectal or bone marrow samples, may be missed.

Dyes used for the demonstration of amyloid are compounds developed by the textile industry. This includes Congo red, which was developed as the first direct cotton dye in 1884 and has been ‘re-invented’ many times in the search for a differential method for the detection of amyloid (Puchtler et al. 1964). As in all histopathological methods, they are often performed on tissues that have been formalin fixed and processed in to paraffin wax. Sometimes the samples can be left standing in fixative for long periods of time, which may make the staining less sensitive and intense. Control sections must be used in all staining methods, and in the demonstration of amyloid they should be cut as they are needed as they can lose their reactivity if stored for long periods.

Congo red

The molecular formula for Congo red is C₃₂H₂₂N₆Na₂O₆S₂ (Fig. 14.3).

Figure 14.3 Chemical structure of Congo red dye.

Congo red is a symmetrical sulfonated azo dye containing a hydrophobic center with two phenyl groups bound by a diphenyl bond to give a linear molecule that is largely hydrophobic (Turnell & Finch 1992). Congo red is also a fluorescent dye (Puchtler 1965), although is not specific for amyloid. Two factors are important to the Congo red-amyloid reaction: the linearity of the dye molecule and the β-pleated sheet configuration. If the spatial configuration of either is altered, even though the chemical groupings are left intact, the reaction fails. Furthermore, the Congo red-mediated positive birefringence of amyloid implies that the dye molecules are arranged in a parallel fashion (Romhányi 1971). Recent work confirms the long-held belief that the Congo red molecule intercalates between two protein moieties at the interface between two adjacent antiparallel β-pleated sheets by disrupting the hydrogen bonds that are responsible for maintaining the β-sheet polymer, yet allowing maintenance of the integrity of the structure by the formation of new hydrogen bonds between protein and dye (Carter & Chou 1998).

Since its introduction on tissue sections by Bennhold (1992), Congo red staining for amyloid and its subsequent green birefringence when viewed by polarized light has become the gold standard. Although it was many years before the exact staining mechanism was understood, it is now well established that staining of amyloid by Congo red is due to hydrogen bonding between the Congo red dye and the β-pleated sheet in a highly orientated linear and parallel manner on the amyloid fibrils (Glenner et al. 1980). Any tissue component that binds Congo red in a linear way also exhibits green birefringence in polarized light. As well as amyloid, dense collagen fibers can also bind Congo red dye in this fashion, often meaning that formalin-fixed tissue gives false positives. By using an alkaline Congo red method this phenomenon is reduced (Puchtler et al. 1962). However Romhányi (1971) used 1% aqueous Congo red and claimed that if the tissue sections were mounted in gum arabic this problem is overcome. Bély et al. (2006) adapted Romhányi’s original method, using a long deparaffinization step of up to 5 days, together with a longer incubation in Congo red. This technique has shown that the amyloid has a stronger affinity to Congo red and therefore can be seen as more sensitive and selective. Many different tissue structures will also stain with 1% aqueous Congo red, and so it must be used under strict controlled conditions using known amyloid-positive sections in conjunction.

The specificity of Congo red staining of amyloid can also be increased by using an alcoholic method combined with high ion strength and high pH. Puchtler et al. (1962) combines all these aspects, giving a superior method to demonstrate amyloid and green birefringence under polarized light.

Recent comparison of several Congo red staining methods made during a run of the UK NEQUAS histology external quality control scheme found that Highman’s method gave the highest scores. At the referral center quoted above, they get a number of biopsies that are either false positive or false negative. Consequently, they recommend the Putchler method since the lack of a differentiation step means there is less intervention by the operator. It should be noted that false positives and false negatives can be caused by other factors even in this technique; for example, in very thin tissue sections.