Amyloidosis is an uncommon disease in which a substance called amyloid, composed of fibrous protein, accumulates in tissues and organs, including the liver, kidneys, tongue, spleen, and heart.


Amyloidosis may occur for no known reason, in which case it is known as primary amyloidosis; more commonly, it is a complication of some other disease, and in such cases it is called secondary. Conditions that may lead to amyloidosis include multiple myeloma (a cancer of bone marrow), rheumatoid arthritis, familial Mediterranean fever, tuberculosis, and other longstanding infections such as chronic osteomyelitis (bone infection). Amyloid is also deposited in the brain in Alzheimer’s disease. Small deposits of amyloid are a normal feature of aging.

Symptoms and signs 

The symptoms of amyloidosis vary, depending on the organs affected and the duration of the condition. Affected organs typically become enlarged. An accumulation of amyloid in the heart may result in arrhythmias (disturbances of the heart rate or rhythm) and heart failure (reduced pumping efficiency of the heart). If the stomach and intestines are affected, symptoms such as diarrhoea may develop, and the lining of these organs may become ulcerated. Primary amyloidosis is often characterized by deposits of amyloid in the skin, which appear as slightly raised, waxy spots. Deposits of amyloid in the kidneys may cause kidney failure, which can be fatal.


There is no treatment for the removal of amyloid deposits. However, it is possible to halt the progression of secondary amyloidosis by treatment of the underlying disorder.

Amyloidosis in detail - technical

Topics covered:

  • Essentials
  • Clinical amyloidosis
  • Amyloid fibrils
  • Diagnosis and monitoring of amyloidosis
  • Management of amyloidosis
  • Further reading


Amyloidosis is the clinical condition caused by extracellular deposition of amyloid in the tissues. Amyloid deposits are composed of amyloid fibrils, abnormal insoluble protein fibres formed by misfolding of their normally soluble precursors. About 25 different proteins can form clinically or pathologically significant amyloid fibrils in vivo as a result of either acquired or hereditary abnormalities. Small, focal, clinically silent amyloid deposits in the brain, heart, seminal vesicles, and joints are a universal accompaniment of ageing. However, clinically important amyloid deposits usually accumulate progressively, disrupting the structure and function of affected tissues and lead inexorably to organ failure and death. No treatment yet exists which can specifically clear amyloid deposits, but intervention which reduces the availability of the amyloid fibril precursor proteins may lead to amyloid regression with clinical benefit.

Pathology—amyloid fibrils of all types are similar: straight, rigid, and non branching; of indeterminate length and 10–15 nm in diameter; and with their subunit proteins arranged in a stack of twisted antiparallel β-pleated sheets. The fibrils bind Congo red dye producing pathognomonic green birefringence when viewed in polarized light, and the protein type can be identified by immunostaining or proteomic analysis. Amyloid deposits always contain a non fibrillar plasma glycoprotein, amyloid P component, the universal presence of which is the basis for use of radioisotope-labelled serum amyloid P component as a diagnostic tracer.

Clinicopathological correlation—amyloid may be deposited in any tissue of the body, including blood vessels walls and connective tissue matrix; clinical manifestations are correspondingly diverse. Although there are some typical clinical presentations related to fibril type, there are many forms of amyloidosis in which there is little or no concordance between the fibril protein, or the genotype of its precursor, and the clinical phenotype. Identification of the amyloid fibril protein is always essential for appropriate clinical management.

Specific types of amyloidosis

Reactive systemic (AA) amyloidosis—fibrils composed of AA protein derived from the acute phase protein, serum amyloid A protein (SAA). Occurs as a complication of any chronic inflammatory disorders (e.g. rheumatoid arthritis) or infections in which SAA concentrations are persistently increased. Most commonly presents with proteinuria and/or organomegaly, e.g. hepatosplenomegaly: nephrotic syndrome may develop before progression to endstage renal failure. Treatment is directed towards the underlying condition, aiming to reduce SAA values to normal.

Monoclonal immunoglobulin light chain (AL) amyloidosis—fibrils consists of all or part of the variable (VL) domain of monoclonal immunoglobulin light chains. May complicate any B-cell dyscrasia but most cases are associated with otherwise ‘benign’ monoclonal gammopathy. Highly variable idiotypic disease but characteristic presentations include involvement of the heart (restrictive cardiomyopathy), kidneys (proteinuria, renal failure), gut (motility disorders, malabsorption), tongue (macroglossia), and nerves (painful sensory polyneuropathy). Treatment is cytotoxic chemotherapy aimed at elimination or suppression of the causative B–cell clone, as for myeloma.

Hereditary systemic amyloidoses—include familial amyloid polyneuropathy, which is caused by mutations in the gene for the plasma protein transthyretin and characterized by progressive peripheral and autonomic neuropathy and varying degrees of visceral involvement.

Clinical amyloidosis

It is useful for diagnosis, management, and the development of new treatments to distinguish clearly between amyloidosis, in which extracellular amyloid deposits are the unequivocal cause of tissue damage and disease, and other diseases in which amyloid deposits of unknown pathogenic significance are present in the tissues. In systemic amyloidosis, in which amyloid can accumulate in the viscera, blood vessels, and connective tissue throughout the body (except within the brain substance itself), the deposits definitely cause the clinical disease. By contrast, amyloid deposits in the brain and cerebral blood vessels are a central part of the pathology of Alzheimer’s disease, the fourth most common cause of death in the Western world, and amyloid is present in the islets of Langerhans of the pancreas in all patients with type 2 diabetes; however, the extent to which these local amyloid deposits are responsible for disease, if at all, is uncertain.

Systemic amyloidosis is responsible for about 1 in 1000 of all deaths in developed countries, and is a serious and important disease because it is often difficult to diagnose, it is usually fatal, and its management is complex and costly. Most systemic amyloidosis is a complication of other underlying primary conditions, which include monoclonal gammopathies of all types, chronic inflammatory disorders, and dialysis for endstage renal failure. Hereditary amyloidosis is very rare, except in a few geographic foci, but its diversity is remarkable. It is important because of its poor prognosis, the complexity of clinical management, the difficult genetic issues involved, and its considerable value as a model for understanding the pathogenesis of amyloid deposition.

Although there are some correlations between fibril protein type and clinical manifestations, there are also many forms of acquired and hereditary amyloidosis in which there is little or no concordance between the fibril protein, or the genotype of its precursor, and the clinical phenotype (Tables 1 and 2). There are evidently genetic and/or environmental factors, distinct from the amyloid fibril protein itself, that determine whether, when, and where clinically significant amyloid deposits form. The nature of these important determinants of amyloidogenesis is obscure. Furthermore, the mechanisms by which amyloid deposition causes disease are poorly understood. While a heavy amyloid load is invariably a bad sign, there may be a poor correlation between the local amount of amyloid and the level of organ dysfunction. Active deposition of new amyloid is often associated with accelerated deterioration compared with stable, long-standing deposits. Nascent or newly formed amyloid fibrils generated in vitro are also cytotoxic to cultured cells, whereas aged or ex vivo fibrils are generally inert, but it is not known if or how this relates to effects in vivo. In most forms of systemic amyloidosis there is overwhelming evidence that tissue damage and resultant disease are caused by the physical presence and accumulation of amyloid deposits, and not by cytotoxicity of the amyloidogenic proteins or their prefibrillar aggregates.

Table 1 Acquired amyloidosis syndromes
Clinical syndrome Fibril protein
Systemic AL amyloidosis, associated with immunocyte dyscrasia, myeloma, monoclonal gammopathy, occult dyscrasia AL derived from monoclonal immunoglobulin light chains
Local nodular AL amyloidosis (skin, respiratory tract, urogenital tract, etc.) associated with focal immunocyte dyscrasia AL derived from monoclonal immunoglobulin light chains
Reactive systemic AA amyloidosis, associated with chronic active diseases AA derived from SAA
Senile systemic amyloidosis Transthyretin derived from plasma transthyretin
Sporadic cerebral amyloid angiopathy Aβ derived from APP
Haemodialysis-associated amyloidosis; localized to osteoarticular tissues or systemic β2-microglobulin derived from high plasma levels
Primary localized cutaneous amyloid (macular, papular) ? Keratin-derived
Ocular amyloid (cornea, conjunctiva) Not known
Orbital amyloid AL or AH derived from monoclonal Ig

Abbreviations: AL, monoclonal immunoglobulin light chain; AA, amyloid A; SAA, serum amyloid A protein; APP, amyloid precursor protein; AH, monoclonal immunoglobulin heavy chain fragment; Ig, immunoglobulin.

Reactive systemic (AA) amyloidosis

Associated conditions

AA amyloidosis occurs in association with chronic inflammatory disorders, chronic local or systemic microbial infections, and, occasionally, malignant neoplasms. In western Europe and the United States of America the most frequent predisposing conditions are idiopathic rheumatic diseases (Table 3). Amyloidosis complicates up to 10% of cases of rheumatoid arthritis and juvenile inflammatory arthritis, although, for reasons that are not clear, the incidence is lower in the United States than in Europe. Amyloidosis is exceptionally rare in systemic lupus erythematosus, related connective tissue diseases, and in ulcerative colitis, by contrast with Crohn’s disease; probably because in lupus and ulcerative colitis there is a blunted acute phase response of serum amyloid A protein, the precursor of AA amyloid fibrils. Tuberculosis and leprosy are important causes of AA amyloidosis, particularly where these infections are endemic. Chronic osteomyelitis, bronchiectasis, chronically infected burns, and decubitus ulcers, as well as the chronic pyelonephritis of paraplegic patients, are other well-recognized associations (see Table 3). Hodgkin’s disease and renal carcinoma, which often cause fever, other systemic symptoms, and a major acute phase response, are the malignancies most commonly associated with systemic AA amyloidosis.

Clinical features

AA amyloid involves the viscera, but may be widely distributed without causing clinical symptoms. More than 90% of patients present with nonselective proteinuria resulting from glomerular deposition, and nephrotic syndrome may develop before progression to endstage renal failure. Haematuria, isolated tubular defects, nephrogenic diabetes insipidus, and diffuse renal calcification rarely occur. Kidney size is usually normal, but may be enlarged, or, in advanced cases, reduced. End-stage chronic renal failure is the cause of death in 40 to 60% of cases, but acute renal failure may be precipitated by hypotension and/or salt and water depletion following surgery, excessive use of diuretics, or intercurrent infec tion, and may be associated with renal vein thrombosis. The second most common presentation is with organ enlargement, such as hepatosplenomegaly or thyroid goitre, with or without overt renal abnormality, but in any case amyloid deposits are almost always widespread at the time of presentation. Involvement of the heart and gastrointestinal tract is frequent, but rarely causes functional impairment.

Table 2 Hereditary amyloidosis syndromes
Clinical syndrome Fibril protein
Predominant peripheral nerve involvement, familial amyloid polyneuropathy. Autosomal dominant Transthyretin variants (commonly Met30, over 80 described)
A-I N-terminal fragment of variant Arg26
Predominant cranial nerve involvement with lattice corneal dystrophy. Autosomal dominant Gelsolin, fragment of variants Asn187 or Tyr187
Oculoleptomeningeal amyloidosis. Autosomal dominant Transthyretin variants
Non-neuropathic, prominent visceral involvement (Ostertag-type). Autosomal dominant Apolipoprotein A-I N-terminal fragment of variants
Lysozyme variants
Fibrinogen α-chain variants
Predominant cardiac involvement, no clinical neuropathy. Autosomal dominant Transthyretin variants
  • Hereditary cerebral haemorrhage with amyloidosis (cerebral amyloid angiopathy). Autosomal dominant:
  •  Icelandic type (major asymptomatic systemic amyloid also present)
Cystatin C, fragment of variant Glu68
 Dutch type Aβ derived from APP variant Gln693
Familial Mediterranean fever, prominent renal involvement. Autosomal recessive AA derived from SAA
Muckle–Well’s syndrome and other heriditary periodic fevers. Autosomal recessive (see Chapter 12.12.2) AA derived from SAA
Cutaneous deposits (bullous, papular, pustulodermal) Not known

Amino acids: Met, methionine; Arg, arginine; Asn, asparagine; Tyr, tyrosine; Glu, glutamic acid; Gln, glutamine.

Abbreviations: Aβ, amyloid β APP, amyloid precursor protein; SAA, serum amyloid A protein.

AA amyloidosis may become clinically evident early in the course of associated disease, but the incidence increases with the duration of the primary condition. The mean duration of chronic rheumatic diseases, such as rheumatoid arthritis, ankylosing spondylitis, or juvenile rheumatoid arthritis, before amyloid is diagnosed is 12 to 14 years, although they can present much sooner. For most patients the prognosis is closely related to the degree of renal involvement and the effectiveness of treatment for the underlying inflammatory condition. In the presence of persistent uncontrolled inflammation, 50% of patients with AA amyloidosis die within 5 years of diagnosis; however, if the acute phase response can be consistently suppressed proteinuria can cease, renal function can be retained, and the prognosis is much better. The availability of chronic haemodialysis and transplantation prevents early death from uraemia per se, but amyloid deposition in extrarenal tissues may be responsible for a less favourable prognosis than for some other causes of endstage renal failure.

Table 3 Conditions associated with reactive systemic amyloid A amyloidosis
Chronic inflammatory disorders
Rheumatoid arthritis
Juvenile inflammatory arthritis
Ankylosing spondylitis
Psoriasis and psoriatic arthropathy
Reiter’s syndrome
Adult Still’s disease
Behçet’s syndrome
Crohn’s disease
Chronic microbial infections
Decubitus ulcers
Chronic pyelonephritis in paraplegics
Whipple’s disease
Malignant neoplasms
Hodgkin’s disease
Renal carcinoma
Carcinomas of gut, lung, urogenital tract
Basal cell carcinoma
Hairy cell leukaemia

Amyloidosis associated with immunocyte dyscrasia: monoclonal immunoglobulin light chain (AL) amyloidosis

Associated conditions

Monoclonal immunoglobulin light chain (AL) amyloidosis may complicate almost any dyscrasia of cells of the B-lymphocyte lineage, including multiple myeloma, malignant lymphomas, and macroglobulinaemia, but most cases are associated with otherwise benign monoclonal gammopathy. Amyloidosis occurs in up to 15% of cases of myeloma, in a lower proportion of other malignant B-cell disorders, and probably in fewer than 5% of patients with a benign monoclonal gammopathies, which are, of course, much more common than myeloma. In some cases, deposition of AL amyloid may be the only evidence of the B-cell dyscrasia.

A monoclonal paraprotein or free light chains can be detected in the serum or urine of about 90% of patients with AL amyloidosis, while in the remaining 10% of cases detection of immunoglobulin gene rearrangement in the bone marrow or peripheral blood sometimes confirms a monoclonal gammopathy. The paraprotein may also appear after presentation and diagnosis of the amyloidosis, and subnormal levels of some or all serum immunoglobulins, or increased numbers of marrow plasma cells, may provide less direct clues to the underlying aetiology. Until recently it has been the practice to diagnose apparently primary cases of amyloidosis, with no previous predisposing inflammatory condition or family history of amyloidosis, as AL type by exclusion. However, it has now been recognized that autosomal dominant hereditary non-neuropathic amyloidosis, particularly that caused by variant fibrinogen α-chain, may be poorly penetrant and of late onset, so there may be no family history. The coincident occurrence of a monoclonal gammopathy may then be gravely misleading, and it is essential to exclude by genotyping all known amyloidogenic mutations, and to seek positive immunohistochemical or biochemical identification of the amyloid fibril protein in all cases.

Clinical features

AL amyloidosis occurs equally in men and women, usually over the age of 50 years, but occasionally in young adults. It has a lifetime incidence (and is the cause of death) of between 0.5 and 1 in 1000 individuals in the United Kingdom. The clinical manifestations are protean, as virtually any tissue other than the brain may be directly involved. Uraemia, heart failure, or other effects of amyloidosis usually cause death within a year of diagnosis, unless the underlying B-cell clone is effectively suppressed.

The heart is affected in 90% of patients with AL amyloidosis. In 30% of these, restrictive cardiomyopathy is the presenting feature and in up to 50% of these patients it is fatal. Other cardiac presentations include arrhythmias and angina. Measurement of circulating brain natriuretic peptide provides a sensitive index of cardiac dysfunction in cardiac AL amyloidosis, and often shows rapid improvement when there is a clonal response to cytotoxic chemotherapy. This suggests that the amyloidogenic light chains may themselves have intrinsic cardiotoxicity, in addition to their deposition as amyloid fibrils. Renal AL amyloidosis has the same manifestations as renal AA amyloidosis, but the prognosis is worse. Gut involvement may cause disturbances of motility (often secondary to autonomic neuropathy), malabsorption, perforation, haemorrhage, or obstruction. Macroglossia occurs rarely, but is almost pathognomonic. Hyposplenism sometimes occurs in both AA and AL amyloidosis. Painful sensory polyneuropathy, with early loss of pain and temperature sensation, followed later by motor deficits, is seen in 10 to 20% of cases, and carpal tunnel syndrome in 20%. Autonomic neuropathy leading to orthostatic hypotension, impotence, and gastrointestinal disturbances may occur alone or together with the peripheral neuropathy, and has a very poor prognosis. Skin involvement takes the form of papules, nodules, and plaques, usually on the face and upper trunk, and involvement of dermal blood vessels results in purpura, occurring either spontaneously or after minimal trauma, and is very common. Articular amyloid is rare, but may mimic acute polyarticular rheumatoid arthritis, or it may present as asymmetrical arthritis affecting the hip or shoulder. Infiltration of the glenohumeral joint and surrounding soft tissues occasionally produces the characteristic ‘shoulder pad’ sign. A rare but serious manifestation of AL amyloidosis is an acquired bleeding diathesis that may be associated with deficiency of factor X, and sometimes also factor IX, or with increased fibrinolysis. It does not occur in AA amyloidosis, although in both AL and AA disease there may be serious bleeding in the absence of any identifiable coagulation factor deficiency.

Senile amyloidosis

Some amyloid is seen in all autopsies on individuals over 80 years of age, but it is not known whether this contributes to the ageing process or whether it is an epiphenomenon that becomes clinically important only when it is extensive.

Senile systemic (cardiac) amyloidosis

Up to 25% of older people have microscopic, clinically silent systemic deposits of transthyretin amyloid involving the walls of the heart and blood vessels, smooth and striated muscle, fat tissue, renal papillae, and alveolar walls. By contrast with most other forms of systemic amyloidosis (including hereditary transthyretin amyloid caused by point mutations in the transthyretin gene), the spleen and renal glomeruli are rarely affected. The brain is not involved. Occasionally, more extensive deposits in the heart, affecting the ventricles and atria and situated in the interstitium and vessel walls, cause significant impairment of cardiac function and may be fatal. The transthyretin involved is usually of the normal wild type, but cases with transthyretin variants have been described that may be hereditary. The isoleucine 122 variant, which occurs in about 1.3 million African Americans, including about 13 000 individuals homozygous for this polymorphism, is associated with a greatly increased risk of senile cardiac amyloidosis, suggesting that this condition may be substantially underdiagnosed; cases of congestive cardiac failure in older people presumably being assumed to reflect coronary artery disease.

Senile focal amyloidosis

Microscopic and clinically silent amyloid deposits of different fibril types, localized to particular tissues, are very commonly present in older people. Deposits of β-protein (see below) as amyloid in cerebral blood vessels and intracerebral plaques seen in normal older brains may or may not have been the harbinger of Alzheimer’s disease had the patient survived long enough. Amyloid deposits composed of apolipoprotein A-I are present in most osteoarthritic joints at surgery or autopsy, usually in close association with calcium pyrophosphate deposits, and affect the articular cartilage and joint capsule. However, the significance of this age-associated articular amyloid, the amount of which is correlated with neither the presence nor clinical severity of osteoarthritis, is not known. The corpora amylacea of the prostate are composed of β2-microglobulin amyloid fibrils. Amyloid in the seminal vesicles is derived from semenogelin I, an exocrine secretory product of the vesicle cells. Isolated deposits of cardiac atrial amyloid consist of atrial natriuretic peptide. The focal amyloid deposits commonly present in atheromatous plaques of older subjects are of two types: containing fibrils either composed of medin, a fragment of lactadherin, or the N-terminal fragment of apolipoprotein A-I.

Table 4 Cerebral amyloidosis
Age-related amyloid angiopathy with or without intracerebral deposits
  • Hereditary amyloid angiopathy of meningeal and cortical vessels associated with cerebral haemorrhage:
  •  Icelandic type
  •  Dutch type
Hereditary amyloid angiopathy affecting the entire central nervous system
Alzheimer’s disease: sporadic, familial, or associated with Down’s syndrome
  • Cerebral amyloid associated with prion disease:
  •  Sporadic spongiform encephalopathy, Creutzfeldt–Jacob disease, variant Creutzfeldt–Jakob disease
  •  Familial prion disease, familial Creutzfeldt–Jacob disease, GSS syndrome atypical familial prion disease
Familial oculoleptomeningeal amyloidosis

Cerebral amyloid

The brain is a very common and important site of amyloid deposition (Table 4), although, possibly because of the blood–brain barrier, there are never any deposits in the cerebral parenchyma itself in any form of acquired systemic visceral amyloidosis. However, cerebrovascular transthyretin amyloid may occur in familial amyloid polyneuropathy resulting from the most common transthyretin variant (methionine for valine at residue 30), and oculoleptomeningeal amyloidosis is caused by other very rare transthyretin variants. The common and major forms of brain amyloid are confined to the brain and cerebral blood vessels, with the single exception of cystatin C amyloid in hereditary cerebral haemorrhage with amyloidosis, Icelandic type, in which there are major, though clinically silent, systemic deposits.

Alzheimer’s disease

By far the most frequent and important type of amyloid in the brain is that related to Alzheimer’s disease, which is the most common cause of dementia and affects more than 3 million individuals in the United States of America and a corresponding proportion of other Western populations. It is generally a disease of older people, and its prevalence is therefore increasing. The clinical differential diagnosis of senile dementia and the positive identification of Alzheimer’s disease are difficult and often of limited precision in life. However, intracerebral and cerebrovascular amyloid deposits are hallmarks of the neuropathological diagnosis.

The amyloid fibrils are composed of β-protein (Aβ), a 39- to 43-residue cleavage product of the large amyloid precursor protein. The vast majority of cases of Alzheimer’s disease are sporadic, but there are also families with an autosomal dominant pattern of inheritance and usually early onset. In about 20 families there are causative mutations in the APP gene for amyloid precursor protein on chromosome 21, and most other kindreds have mutations in the genes for presenilin 1 (chromosome 14) and presenilin 2 (chromosome 1). All these mutations are associated with increased production from amyloid precursor protein of Aβ1–42, the most amyloidogenic form of Aβ. Since all individuals with Down’s syndrome (trisomy 21) develop Alzheimer’s disease if they survive into their forties, there is evidently a close link between amyloid precursor protein, Aβ overproduction, Aβ amyloidosis, and the pathogenesis of Alzheimer’s disease. However, it remains unclear whether or how Aβ per se, or the amyloid fibrils that it forms, contribute to the neuronal loss that underlies the dementia.

Synthetic Aβ fibrils formed in vitro are markedly cytotoxic, and cause the death of cultured cells by apoptosis and necrosis. Although it is not clear to what extent these findings reflect phenomena that may be responsible for neurodegeneration in vivo, there is increasing evidence, from both transgenic mouse models of Alzheimer’s disease and in vivo intracerebral injection of different molecular conformations of Aβ, that small oligomeric prefibrillar aggregates of Aβ are associated with and cause cognitive dysfunction. There is controversy about the correlation between the severity of dementia in Alzheimer’s disease and the extent of amyloid angiopathy and plaques. Nevertheless, the fact that patients with Alzheimer’s disease caused by amyloid precursor protein and presenilin mutations have exactly the same neuropathology as sporadic cases, including tangles, argues strongly that the amyloid precursor protein and Aβ pathway can be of primary pathogenetic significance.

In addition to the Aβ deposits in the brains of patients with Alzheimer’s disease and Down’s syndrome, there are also extensive ‘amorphous’ deposits throughout the brain. These do not stain with Congo red, and are detectable only by immunohistochemical staining. Their significance is unknown. They apparently precede the appearance of histochemically identifiable amyloid, but are not necessarily the precursor of it because they are present in areas such as the cerebellum in which Aβ is never seen. The nonfibrillar, nonamyloid protein apolipoprotein E is demonstrable in many amyloid deposits, including those of Alzheimer’s disease. The APOE4 gene (chromosome 19), encoding one of the three isoforms of this apolipoprotein, is strongly associated with a predisposition to develop Alzheimer’s disease and with increased amounts of amyloid in the brain, but the underlying mechanisms are unknown.

Another neuropathological feature of Alzheimer’s disease, and some other neurodegenerative conditions, is the neurofibrillary tangle located intracellularly within neuronal cell bodies and processes. These tangles have a characteristic ultrastructural morphology of paired helical filaments, and are composed of an abnormally phosphorylated form of the normal neurofilament protein, tau. They bind Congo red and then give the pathognomonic green birefringence of amyloid when viewed in polarized light. Although their electronmicroscopic ultrastructure is completely different from that of amyloid fibrils, the most recent review of amyloid nomenclature includes them as amyloid.

Senile cerebral amyloidosis and amyloid (congophilic cerebral) angiopathy

Up to 60% of the brains of nondemented older individuals contain Aβ in the cerebral blood vessels, and there may also be focal intracerebral Aβ plaques. These deposits are usually clinically silent and may or may not have been harbingers of Alzheimer’s disease, had the patients survived long enough. Sometimes the amyloid angiopathy is more extensive, and is increasingly recognized as an important cause of cerebral haemorrhage and stroke, to be distinguished from atherosclerotic cerebrovascular disease.

Hereditary cerebral haemorrhage with amyloidosis: hereditary cerebral amyloid angiopathy
Icelandic type (OMIM#604312)

Cerebrovascular amyloid deposits composed of a fragment of a genetic variant of cystatin C are responsible for recurrent major cerebral haemorrhages starting in early adult life in members of families originating in western Iceland. There is autosomal dominant inheritance and appreciable, but clinically silent, amyloid deposits are present in the spleen, lymph nodes, and skin. There is no extravascular amyloid in the brain, and the neurological deficits, often including dementia, of surviving patients are compatible with their cerebrovascular pathology.

Dutch type (OMIM#605714)

In families originating from a small region on the coast of the Netherlands the autosomal dominant inheritance of a genetic variant of Aβ, which is deposited as cerebrovascular amyloid, results in recurrent normotensive cerebral haemorrhages starting in middle age. There are also amorphous Aβ deposits in the brain and early senile plaques, without congophilic amyloid cores. Multi-infarct dementia occurs in survivors, but some patients become demented in the absence of stroke. Amyloid outside the brain has not been reported.

Cerebral amyloid associated with prion disease

The neuropathology of a group of progressive, invariably fatal spongiform encephalopathies sometimes, but certainly not always, includes intracerebral amyloid plaques. These diseases are transmissible and in some cases hereditary. The sporadic and familial Creutzfeldt–Jacob disease, the familial Gerstmann–Sträussler–Scheinker syndrome, and kuru are caused by prions (PrPSc), conformational isoforms of the normal physiological cellular prion protein (PrPC). The human diseases are closely related to the animal diseases scrapie of sheep and goats, transmissible encephalopathy of mink, elk, and male deer, and bovine spongiform encephalopathy. Variant Creutzfeldt–Jacob disease is apparently the result of transmission of bovine spongiform encephalopathy to humans.

The significance of amyloid per se in these disorders is not clear because it is not always histologically detectable, and in some disorders is not seen, e.g. fatal familial insomnia and bovine spongiform encephalopathy (which is apparently a result of the transmission of ovine scrapie to cattle). When scrapie or its human counterparts are transmitted to experimental animals by inoculation of affected brain tissue, the development of intracerebral amyloid depends on the strain of infectious agent and the genetic background of the recipient. Even when amyloid is present in the brain it is not seen elsewhere, e.g. in the spleen, although the latter is a rich source of the infective agent. However, when the infective agent is exhaustively and highly purified from brain or spleen it forms typical congophilic amyloid fibrils composed of the proteinase-resistant subunit PrPSc, and when amyloid deposits are present in affected brains they immunostain with antiprion antibodies.

The amyloid fibril protein is thus directly related to the cause of the encephalopathy, but histologically demonstrable amyloid deposition is evidently not necessary for the expression of disease. Indeed, recent work in transgenic and knockout mouse strains clearly demonstrates both that prion amyloid deposition is not a necessary condition for the development of transmissible spongiform encephalopathy, and that expression of the normal cellular isoform, PrPC, is absolutely required. Neuronal damage may perhaps be caused by a cytotoxic interaction between prefibrillar PrPSc aggregates and the normal PrPC, or indeed by other mechanisms entirely. This is a different situation from the extracerebral amyloidosis, and from cystatin C and nonhereditary cerebral amyloid angiopathies, in which amyloid deposition is invariably present when there is clinical disease, and is unequivocally the cause of tissue damage.

Hereditary systemic amyloidosis

Familial amyloid polyneuropathy (hereditary transthyretin amyloidosis) (OMIM#176300)

Familial amyloid polyneuropathy is an autosomal dominant syndrome with onset at any time from the second decade onwards. It is characterized by progressive peripheral and autonomic neuropathy and varying degrees of visceral involvement especially affecting the vitreous of the eye, the heart, kidneys, thyroid, and adrenals. There are usually amyloid deposits throughout the body involving the walls of blood vessels and the connective tissue matrix; the pathology is due to these deposits. Apart from major foci in Portugal, Japan, and Sweden, familial amyloid polyneuropathy has been reported in most populations throughout the world. There is considerable variation in the age of onset, rate of progression, and involvement of different systems; although within families the pattern is usually quite consistent. There is remorseless progression and the disorder is invariably fatal. Death results from the effects and complications of peripheral and/or autonomic neuropathy, or from cardiac or renal failure.

Familial amyloid polyneuropathy is caused by mutations in the gene for the plasma protein transthyretin (formerly known as prealbumin). The most frequent of these causes a methionine for valine substitution at position 30 in the mature protein, but more than 80 amyloidogenic mutations have been described. There is often little correlation between the underlying mutation and the clinical phenotype, which is evidently determined by other genetic and possibly also environmental factors, although in a few cases certain mutations are uniquely associated with particularly aggressive or relatively organ-limited disease. The amyloidogenic transthyretin mutations are not always penetrant, and asymptomatic methionine-30 homozygotes over the age of 60 years have been reported. Rare kindreds with the apolipoprotein A-I arginine 26 variant, which usually causes nonneuropathic amyloidosis, may present with prominent peripheral neuropathy resembling transthyretin familial amyloid polyneuropathy (OMIM*107680).

Familial amyloid polyneuropathy with predominant cranial neuropathy (OMIM*137350)

Originally described in Finland, but now reported in other populations, this autosomal dominant hereditary amyloidosis presents in adult life with cranial neuropathy, lattice corneal dystrophy, and distal peripheral neuropathy. There may be skin, renal, and cardiac manifestations and microscopic amyloid deposits are widely distributed in connective tissue and blood vessel walls; life expectancy approaches normal. The amyloid fibrils are derived from variants of the actin-modulating protein gelsolin, encoded by point mutations. Individuals homozygous for these mutations have severe renal amyloidosis in addition to the usual neuropathy.

Nonneuropathic systemic amyloidosis (OMIM#105200)

In this rare autosomal dominant syndrome of major systemic amyloidosis without clinical evidence of neuropathy, the patterns of organ involvement and overall clinical phenotype vary between families. The kidneys are often the most severely affected organ, leading to hypertension and renal failure, but the heart, spleen, liver, bowel, connective tissue, and exocrine glands may all be involved. Following clinical presentation there is inexorable progression to death or organ failure requiring transplantation. Clinical presentation is usually in early adulthood, although in a few kindreds it may be as late as the sixth decade. The amyloid proteins so far identified are genetic variants of apolipoprotein A-I and A-II, lysozyme, and the fibrinogen α-chain.

Cardiac amyloidosis

Cardiac amyloidosis without overt involvement of other viscera or neuropathy, progressing inexorably to death, is associated with certain transthyretin gene mutations and is inherited in an autosomal dominant manner with variable penetrance (see Table 3). By far the most common variant is transthyretin iseoleucine 122, which occurs in 4% of African Americans and frequently causes cardiac amyloidosis from the sixth decade onwards.

Familial Mediterranean fever (OMIM#249100)

Familial Mediterranean fever is an autosomal recessive autoinflammatory disorder caused by mutations in the gene MEFV on chromosome 16 that encodes a neutrophil-specific protein of unknown function called pyrin or marenostrin (see Chapter 12.12.2). The disease is characterized by recurrent episodes of fever, abdominal pain, pleurisy, or arthritis, and predominantly occurs in non-Ashkenazi Jews, Armenians, Anatolian Turks, and Levantine Arabs. Among Sephardi Jews of North African origin, and in the other populations (except Armenians and, to a lesser extent, Ashkenazi Jews), untreated familial Mediterranean fever is eventually complicated in a high proportion of cases by typical systemic AA amyloidosis. Furthermore, some patients with familial Mediterranean fever present with AA amyloidosis before they have experienced any symptoms, and this is consistent with the recent finding that a substantial acute phase plasma protein response is frequently present, even in asymptomatic individuals. The variable incidence of amyloidosis in patients with familial Mediterranean fever from different populations is not wholly explained by their specific pyrin gene mutations, and is another illustration of the unknown genetic determinants of clinical amyloidosis.

Haemodialysis-associated amyloidosis

Almost all patients with endstage renal failure who are maintained on haemodialysis for more than 5 years develop amyloid deposits composed of β2-microglobulin. These deposits are predominantly osteoarticular and are associated with carpal tunnel syndrome, large-joint pain and stiffness, soft-tissue masses, bone cysts, and pathological fractures. Renal tubular amyloid concretions may also form. The serious clinical problems associated with β2-microglobulin amyloidosis constitute the major cause of morbidity in patients on long-term dialysis. Furthermore, in some patients more extensive deposition occurs, most commonly in the spleen but also in other organs, and a few cases of death associated with systemic β2-microglobulin amyloid have been reported. The β2-microglobulin is derived from the high plasma concentrations that develop in renal insufficiency and are not cleared by dialysis. This type of amyloidosis also occurs in patients on continuous ambulatory peritoneal dialysis, and has even been reported in a few patients with chronic renal failure who have never been dialysed. Improved clearance of β2-microglobulin by current dialysis membranes and procedures encouragingly seems to be reducing the incidence of this form of amyloidosis.

Endocrine amyloid

Many tumours of APUD cells that produce peptide hormones have amyloid deposits in their stroma. These are probably composed of the hormone peptides; in the case of medullary carcinoma of the thyroid the fibril subunits are derived from procalcitonin. In insulinomas the amyloid fibril protein is a novel peptide first identified in that site and subsequently shown to be the fibril protein in the amyloid of the islets of Langerhans seen in type 2 (maturity onset) diabetes. This peptide is called islet amyloid polypeptide, or amylin, and shows appreciable homology with calcitonin gene-related peptide. Amyloid of this type is an almost universal feature of the pancreatic islets in type 2 diabetes, and becomes more extensive with increasing duration and severity of the disease. Although the amyloid itself is probably not initially responsible for the metabolic defect in this form of diabetes, it is likely that progressive amyloid deposition, leading to islet destruction, subsequently does contribute to the pathogenesis. The possible hormonal or other role of islet amyloid polypeptide itself, which is produced by the islet β-cells, is also not yet clear.

Rare localized amyloidosis syndromes

Amyloid deposits localized to the skin occur in both acquired and hereditary forms of amyloidosis. Primary localized cutaneous amyloidosis presents in adult life as macular or papular lesions, the fibrils of which may be derived from keratin. Hereditary cutaneous amyloid lesions are rare, of unknown fibril type, and are sometimes associated with other, nonamyloid, multisystem disorders. Amyloid deposits in the eye cause local problems in the cornea (corneal lattice dystrophy) or conjunctiva, while orbital amyloid presents as mass lesions that can disrupt eye movement and the structure of the orbit. In one such case the fibril protein has been identified as a fragment of IgG heavy chain. Lactoferrin and keratoepithelin have been identified as the amyloid fibril proteins in different cases of corneal amyloidosis.

Localized foci of AL amyloid can occur anywhere in the body in the absence of systemic AL amyloidosis, the most common sites being the skin, upper airways and respiratory tract, and the urogenital tract. They may be associated with a local plasmacytoma or B-cell lymphoma producing a monoclonal immunoglobulin, but often the cells, which must be present to produce the amyloidogenic protein, are scattered inconspicuously in the affected tissue. The clinical problems caused by these space-occupying amyloidomas are usually cured by surgical resection, but this is not always possible.

Amyloid fibrils

Regardless of their very diverse protein subunits, amyloid fibrils of different types are remarkably similar: straight, rigid, nonbranching, of indeterminate length, and 10 to 15 nm in diameter. They are insoluble in physiological solutions, relatively resistant to proteolysis, and bind Congo red dye, producing pathognomonic green birefringence when viewed in polarized light. Electron microscopy reveals that each fibril consists of two or more protofilaments, the precise number varying with the fibril type. The X-ray diffraction patterns of all the different ex vivo amyloid fibrils, and of synthetic fibrils formed in vitro, that have been studied demonstrate the presence of a common core structure within the filaments: the subunit proteins are arranged in a stack of twisted antiparallel β-pleated sheets lying with their long axes perpendicular to the long axis of the fibril.

Recent observations have shown that many different proteins, including molecules totally unrelated to amyloidosis in vivo, can be refolded after denaturation in vitro to form typical, stable, congophilic cross-fibrils. Although it is not clear why only the 26 known amyloidogenic proteins adopt the amyloid fold and persist as fibrils in vivo, a major unifying theme that is currently emerging is that, in all cases studied, the precursors are relatively destabilized. Even under physiological or other conditions they may encounter in vivo, they adopt partly unfolded states that involve the loss of tertiary or higher-order structure. These readily aggregate, with retention of β-sheet secondary structure, into protofilaments and fibrils. Once the process has started, seeding may also play an important facilitating role, so that amyloid deposition may progress exponentially as expansion of the amyloid template captures further precursor molecules.

Amyloid fibril proteins and their precursors

Immunoglobulin light chain

AL proteins are derived from the N-terminal region of monoclonal immunoglobulin light chains and consist of all or part of the variable domain. Intact light chains may occasionally be found, and the molecular weight therefore varies between about 8 and 30 kDa. The light chain of the monoclonal paraprotein is either identical to, or clearly the precursor of, AL isolated from the amyloid deposits.

AL is more commonly derived from λ chains than from κ chains, despite the fact that κ chains predominate among both normal immunoglobulins and the paraprotein products of immunocyte dyscrasias. A new λ-chain subgroup, λVI, was initially identified as an AL protein in two cases of immunocyte dyscrasia-associated amyloidosis, before it had been recognized in any other form, and it has subsequently been observed in many more cases of AL amyloidosis. Furthermore, there is increasing evidence from sequence analysis of Bence Jones proteins of both κ and λ type from patients with AL amyloidosis, and of AL proteins themselves, that these polypeptides contain unique amino acid replacements or insertions compared with nonamyloid monoclonal light chains. In some cases these changes involve the replacement of hydrophilic framework residues by hydrophobic residues, changes likely to promote aggregation and insolubilization; in others the monoclonal light chains from amyloid patients have been directly demonstrated to have decreased solubility and a greater propensity for precipitation than control nonamyloid proteins. The inherent amyloidogenicity of particular monoclonal light chains has been elegantly confirmed in an in vivo model in which isolated Bence Jones proteins were injected into mice. Animals receiving light chains from AL amyloid patients developed typical amyloid deposits composed of the human protein, whereas animals receiving light chains from myeloma patients without amyloid did not.

Amyloid A (AA)

The AA protein is a single nonglycosylated polypeptide chain usually of mass 8000 Da and containing 76 residues corresponding to the N-terminal portion of the 104-residue serum amyloid A protein (SAA). Smaller and larger AA fragments, even some whole SAA molecules, have also been reported in AA fibrils. SAA is an apolipoprotein of high density lipoprotein particles, and is the polymorphic product of a set of genes located on the short arm of chromosome 11. It is highly conserved in evolution and is a major acute phase reactant in all species in which it has been studied. Most of the SAA in plasma is produced by hepatocytes, in which the synthesis is under transcriptional regulation by cytokines (especially interleukin 1, interleukin 6, and tumour necrosis factor) acting via NF-κB-like transcription factors, and possibly others. After secretion it is rapidly associated with high density lipoproteins, from which it displaces apolipoprotein A-I. The circulating concentration can rise from normal levels of up to 3 mg/litre to over 1000 mg/litre within 24 to 48 h of an acute stimulus, and with ongoing chronic inflammation the level may remain persistently high. Certain isoforms of SAA, the products of different genes, are predominantly synthesized elsewhere in the body by macrophages, adipocytes, and certain other cells. Although they also associate with high density lipoproteins, their acute phase synthesis is stimulated differently and they presumably have different functions. There is also a closely related family of high density lipoprotein trace apoproteins that are not acute phase reactants; they have been designated constitutive SAAs, although they do not form amyloid.

The precursor of amyloid fibril AA protein is circulating SAA, from which amyloid fibril AA protein is derived by proteolytic cleavage. Such cleavage can be produced by macrophages and by a variety of proteinases, but since further cleavage of AA is readily demonstrable in vitro it is not clear why the AA peptide persists in amyloid. Furthermore, it is not known whether the process of AA fibril generation involves cleavage of SAA before and/or after aggregation of monomers. Persistent overproduction of SAA causing sustained high circulating levels is a necessary condition for deposition of AA amyloid, but it is not known why only some individuals in this state develop amyloidosis. In mice, only one of the three major isoforms of SAA is the precursor of AA in amyloid fibrils. Human SAA isoforms are more complex, but homozygosity for particular types seems to favour amyloidogenesis, although there may also be ethnic differences.

The normal functions of SAA are not known, although modulating effects on reverse cholesterol transport and on lipid function in the microenvironment of inflammatory foci have been proposed. A protein, homologous with SAA, produced by rabbit fibroblasts has been reported to act as an autocrine stimulator of collagenase production in vitro. Other reports of potent cell-regulatory functions of isolated denatured delipidated SAA have yet to be confirmed in physiological preparations of SAA-rich high density lipoproteins. Regardless of its physiological role, the behaviour of SAA as an exquisitely sensitive acute phase protein with an enormous dynamic range makes it an extremely valuable empirical clinical marker. It can be used objectively to monitor the extent and activity of infective, inflammatory, necrotic, and neoplastic disease. Furthermore, routine monitoring of SAA should be an integral part of the management of all patients with AA amyloidosis or disorders predisposing to it, as control of the primary inflammatory process in order to reduce SAA production is essential if amyloidosis is to be halted, enabled to regress, or prevented. Automated immunoassay systems for SAA are available that meet a World Health Organization international reference standard.


Transthyretin, formerly known as prealbumin, is a normal nonglycosylated plasma protein with a relative molecular mass of 55 044 Da. It is composed of four identical noncovalently associated subunits, each of 127 amino acids. It is produced by hepatocytes and the choroid plexus, and is a significant negative acute phase protein. Each tetrameric molecule is able to bind a single thyroxine or triodothyronine molecule and up to 15% of circulating thyroid hormone is transported in this way. Transthyretin also forms a 1:1 molecular complex with retinol-binding protein, which transports vitamin A.

Transthyretin is encoded by a single-copy gene, but is appreciably polymorphic and more than 90 different point mutations encoding single residue substitutions have been identified. Normal wild type transthyretin is an inherently amyloidogenic protein that forms the fibrils in senile systemic amyloidosis. In vitro exposure to reduced pH is sufficient to generate transthyretin amyloid fibrils from the pure protein. Most of the variant forms of transthyretin have been associated with hereditary amyloidosis, and show decreased stability in vitro compared with the wild type. Transgenic mice expressing variant human transthyretin with a methionine 30 substitution sometimes develop systemic amyloid deposits, though unfortunately not in the peripheral nerves, even when the transgene is expressed in the choroid plexus and transthyretin amyloid is deposited in the meninges and choroid plexus. This is another example of the important unknown factors, other than the presence of an amyloidogenic protein itself, that determine where and when clinical amyloidosis develops.

Individuals heterozygous for transthyretin mutations have a mixture of wild type and variant transthyretin monomers in their circulating transthyretin, and if they develop amyloidosis both forms are often present, although the variant may predominate in the amyloid fibrils. Although cleavage fragments of transthyretin are commonly present, intact transthyretin subunits are also found and fibrillogenesis does not depend on an initial proteolytic step.

Amyloid beta (Aβ)

The fibril protein in the intracerebral and cerebrovascular amyloid of Alzheimer’s disease, Down’s syndrome, and hereditary amyloid angiopathy of the Dutch type is a 39- to 43-residue sequence derived by proteolysis from a precursor protein of high molecular weight, the amyloid precursor protein (APP), encoded on the long arm of chromosome 21. Several isoforms of APP are generated by alternative splicing of transcripts from the 19-exon gene, yielding three major forms: APP695, APP751, and APP770. These are each single-chain, multidomain glycoproteins with the 47 residues of the C-terminal lying within the cytoplasm, a 25-residue membrane-spanning region, and the rest of the molecule lying extracellularly. APP751 and APP770 contain a 56-residue Kunitz-type serine proteinase inhibitor domain encoded by exon 7.

Following glycosylation and membrane insertion APPs are cleaved extracellularly, close to the transmembrane sequence, by so-called APP secretase activity. This releases, in the case of APP751 and APP770, a molecule known as proteinase nexin II, which avidly binds factor XIa, trypsin, and chymotrypsin, as well as epidermal growth factor-binding protein and the γ subunit of nerve growth factor. The predominant species of mRNA found in the brain encodes APP695, which lacks the proteinase inhibitor domain, while mRNA for APP751 is the most abundant in other tissues. Despite this, 85% of secreted APP in the brain is proteinase nexin II. Interestingly, APP secreted by a glial cell line is substantially glycosylated with chondroitin sulphate glycosaminoglycan chains. APP also undergoes high-affinity interactions with heparan sulphate. These observations suggest that APP may have important functions in cell adhesion, cell migration, and modulation of growth-factor activities. APP proteinase nexin II is present in and released by platelets, and probably functions in the clotting cascade.

The amyloidogenic peptide Aβ, encoded by parts of exons 16 and 17, corresponds to the part of the APP sequence that extends from within the cell membrane into the extracellular space. Secretase cleavage of APP to release the soluble form cannot therefore generate intact Aβ itself, or larger fragments containing it. However, there is an alternative processing pathway for APP, in which it is taken up whole by lysosomes and cleaved to yield fragments that do contain the whole Aβ sequence. Furthermore, APP cleaved at the N-terminus of Aβ, and soluble Aβ itself, are normally produced by cell lines and by mixed brain cells in culture, and are present in the cerebrospinal fluid. However, the source of the Aβ in the intracerebral amorphous deposits, and that which aggregates as amyloid fibrils in the brain and cerebral blood vessels, is still not known. The 42-residue form of Aβ (Aβ1–42) is markedly the most amyloidogenic, and all the mutations in the APP and presenilin genes that are associated with hereditary Alzheimer’s disease result in increased production of this amyloid. Increased availability of the precursor is thus responsible for amyloidogenesis, but the pathogenesis of neuronal damage and dementia remain unclear.

Cystatin C

Cystatin C (formerly called γ-trace) is an inhibitor of cysteine proteinases, including cathepsin B, H, and L. It is encoded by a gene on chromosome 20 and consists of a single nonglycosylated polypeptide chain of 120 residues. It is present in all major human biological fluids at concentrations compatible with a significant physiological role in proteinase inhibition. The normal concentration in cerebrospinal fluid is 6.5 mg/litre (range 2.7–13.7), but is much lower (2.7 mg/litre, range 1.0–4.7) in patients with the Icelandic type of hereditary cerebral amyloid angiopathy, in whom fragments of the glutamine 68 genetic variant of cystatin C form the amyloid fibrils. This reduced concentration is diagnostically useful, and is evident even in presymptomatic carriers of the cystatin C gene mutation.

The point mutation that causes the disease encodes a glutamine for leucine substitution in the mature protein, and the amyloid fibril protein consists of the C-terminal 110 residues of the variant. This N-terminally truncated form is not detectable in the cerebrospinal fluid of affected patients, suggesting that cleavage takes place either in close proximity to fibril deposition or after the fibrils have formed. The variant cystatin C is less stable than the wild type and readily forms fibrils in vitro. It is not known whether cerebral haemorrhage in cystatin C amyloidosis is caused simply by the damaging effects of vascular amyloid deposition, or whether deficiency in inhibitory capacity for cysteine proteinases also plays a part.


Gelsolin is a widely distributed 90 kDa cytoplasmic protein that binds actin monomers, nucleates actin filament growth, and severs actin filaments. Alternative transcriptional initiation and message processing from a single gene on chromosome 9 are responsible for the synthesis of a secreted form of gelsolin (93 kDa), which circulates in the plasma at a concentration of about 200 mg/litre. Its function in the blood is not known, but may be related to the clearance of actin filaments released by dying cells. In the Finnish type of hereditary amyloidosis the amyloid fibril protein is a 71-residue fragment of variant gelsolin, with asparagine substituted for aspartic acid at position 15 (corresponding to residue 187 of the mature molecule), and the same mutation has been discovered in affected kindreds from different ethnic backgrounds. In one Danish family with the same phenotype there is a different mutation at the same nucleotide, predicting a tyrosine for aspartic acid substitution at residue 187. Synthetic and recombinant peptides that include the asparagine for aspartic acid substitution at residue 187 are less soluble than the wild type sequence and readily form amyloid fibrils in vitro.

Apolipoprotein A-I and A-II

Apolipoprotein A-I is the most abundant apolipoprotein among the high density lipoprotein particles, and participates in their central function of reverse cholesterol transport from the periphery to the liver. Apolipoprotein A-I variants are extremely rare, and may be phenotypically silent or may affect lipid metabolism. However, 15 different variants of apolipoprotein A-I, including single- and multiple-residue substitutions and deletions, have been associated with amyloidosis. These are inherited in an autosomal dominant manner and are usually highly penetrant, but there are marked variations in the age and manner of presentation, even within the same family and in different kindreds with the same mutation.

The amyloid fibril protein consists, in all cases studied, of the first 90 or so N-terminal residues, even when the causative variant residue(s) are more distal. Wild type apolipoprotein A-I is also amyloidogenic, forming the deposits associated with atheromatous plaques in older people; the various amyloidogenic mutations presumably encode sequence changes that render apolipoprotein A-I less stable and/or more liable to cleavage that yields the fibrillogenic N-terminal fragment. Predominantly renal amyloidosis has also been described in a handful of families in association with several different mutations in a normal stop codon in the gene for apolipoprotein A-II, which results in a peptide extension from residue 78.


Lysozyme is the classic bacteriolytic enzyme of external secretions, discovered by Fleming in 1922. It is also present at high concentration within articular cartilage and in the granules of polymorphs, and is the major secreted product of macrophages. Lysozymes are present in most organisms in which they have been sought, although their physiological role is not always clear. The complete structures of hen egg-white and human lysozymes are known to atomic resolution, and their catalytic mechanism, epitopes, folding, and other aspects of their structure–function relationship have been analysed exhaustively. This contrasts with the absence of detailed three-dimensional structural information on any other amyloid fibril protein or its precursor, except transthyretin and β2-microglobulin.

Lysozyme, unlike transthyretin and β2-microglobulin, is not inherently amyloidogenic, and is therefore a valuable model for the investigation of amyloid fibrillogenesis. There is only one copy of the lysozyme gene in the human genome, and no disease is associated with lysozyme other than amyloidosis. The first mutations identified to cause amyloidosis were substitution of threonine for isoleucine at residue 56 in one family, and histidine for aspartic acid at residue 67 in another. These dramatic changes in residues that are extremely conserved throughout the lysozyme and related α-lactalbumin protein families destabilize the native fold, so that the variants readily adopt partly unfolded states, even under physiological conditions, and spontaneously aggregate in vitro, and evidently also in vivo, into amyloid fibrils. Several further amyloidogenic variants of lysozyme have recently been described.

Islet amyloid polypeptide

Islet amyloid polypeptide (amylin; IAPP) is a 37-residue molecule encoded by a gene on chromosome 12 and with 46% sequence homology to the neuropeptide calcitonin gene-related peptide. Islet amyloid polypeptide is produced in the β-cells of the pancreatic islets of Langerhans, and is stored in and released from their secretory granules together with insulin. It has been reported to modulate insulin release and to induce peripheral insulin resistance, vasodilatation, and lowering of plasma calcium, but neither its physiological role nor its contribution to diabetes are yet known.

The amyloidogenicity of islet amyloid polypeptide depends on the amino acid sequence between residues 20 and 29, as shown by in vitro fibrillogenesis with synthetic peptides. The synthetic decapeptide IAPP20–29, and even the hexapeptide IAPP25–29, form amyloid-like fibrils in vitro, whereas other islet amyloid polypeptide fragments do not. There is also a correlation between conservation of this sequence and deposition of the amyloid in the islets of diabetic animals of different species. However, the role of the amyloid in diabetogenesis remains to be established. In the degu, a South American rodent, spontaneous diabetes is associated with islet amyloid composed of insulin, and xenogeneic insulin can also form amyloid in humans at the site of repeated therapeutic insulin injections.


β2-Microglobulin is a nonglycosylated, nonpolymorphic, single-chain protein of 99 residues, with a single intrachain disulphide bridge and a relative molecular mass of 11 815, encoded by a single gene on chromosome 15. It becomes noncovalently associated with the heavy chain of major histocompatibility class I antigens, and is required for transport and expression of the major histocompatibility complex (MHC) at the cell surface. Amino acid sequence homology places β2-microglobulin in the superfamily that includes immunoglobulins, T-cell receptor α- and β-chains, Thy-1 (CD90), MHC class I and II molecules, secretory component, etc. Its three-dimensional structure is a typical β-barrel with two antiparallel pleated sheets comprising three and four strands, respectively, and closely resembles an immunoglobulin domain.

β2-Microglobulin is produced by lymphoid and a variety of other cells, in which it stabilizes the structure and function of MHC class I antigens at the cell surface. When these complexes are shed by cleavage of the heavy chain at the cell surface, free β2-microglobulin is released. The circulating concentration of β2-microglobulin is 1 to 2 mg/litre and the protein is rapidly cleared by glomerular filtration and then catabolized in the proximal renal tubule. Impairment of renal function is associated with retention of β2-microglobulin and increased circulating levels because there is no other site for its catabolism. Daily production of β2-microglobulin is about 200 mg, and in patients in endstage renal failure on haemodialysis, plasma β2-microglobulin levels rise to and remain at levels of about 40 to 70 mg/litre. Isolated unaltered β2-microglobulin can form amyloid-like fibrils itself in vitro, and most studies of ex vivo β2-microglobulin fibrils show the whole intact molecule to be the major subunit, although fragments and altered forms of β2-microglobulin have also been reported.


Amyloidotic organs contain more glycosaminoglycans than normal tissues, and at least some of this is a tightly bound integral part of the amyloid fibrils. These fibril-associated glycosaminoglycans are heparan sulphate and dermatan sulphate in all forms of amyloid that have been investigated. Fibrils isolated by water extraction and separated from other tissue components contain 1 to 2% by weight glycosaminoglycan, none of which is covalently associated with the fibril protein. Interestingly, in systemic AA and AL amyloidosis, the only forms in which this has been studied so far, there is markedly restricted heterogeneity of the glycosaminoglycan chains, suggesting that particular subclasses of heparan and dermatan sulphates are involved. Immunohistochemical studies demonstrate the presence of proteoglycan core proteins in all amyloid deposits, and that these are closely related to fibrils at the ultrastructural level. However, in isolated fibril preparations much of the glycosaminoglycan material is free carbohydrate chains, and it is not yet clear whether this represents aberrant glycosaminoglycan metabolism related to amyloidosis or is just an artefact of postmortem degradation of core protein.

The significance of glycosaminoglycans in amyloid remains unclear, but their universal presence, intimate relationship with the fibrils, and restricted heterogeneity all suggest that they may be important. Glycosaminoglycans are known to participate in the organization of some normal structural proteins into fibrils and they may have comparable fibrillogenic effects on certain amyloid fibril precursor proteins. Furthermore, the glycosaminoglycans on amyloid fibrils may be ligands to which serum amyloid P component, another universal constituent of amyloid deposits, binds.

Amyloid P component and serum amyloid P component

Amyloid deposits in all different forms of the disease, both in humans and in animals, contain the nonfibrillar glycoprotein amyloid P component. Amyloid P component is identical to and derived from the normal circulating plasma protein, serum amyloid P component, a member of the pentraxin protein family that includes C-reactive protein. Human serum amyloid P component is secreted only by hepatocytes, is a trace constituent of plasma (women: mean 24 mg/litre, range 8–55, men: mean 32 mg/litre, range 12–50), and is not an acute phase reactant. Nevertheless, apart from the fibrils themselves, amyloid P component is always by far the most abundant protein in all amyloid deposits.

Serum amyloid P component consists of five identical noncovalently associated subunits, each with a molecular mass of 25 462 Da, arranged in a pentameric disc-like ring. The tertiary fold of the subunit is dominated by antiparallel β-sheets, forming a flattened β-barrel with jellyroll topology and a core of hydrophobic side chains. This is the lectin fold, shared with a variety of other animal, plant, and bacterial carbohydrate-binding proteins (lectins). Serum amyloid P component is a calcium-dependent ligand-binding protein; its best-defined specificity is for the 4,6-cyclic pyruvate acetal of β-D-galactose, but it also binds avidly and specifically to DNA, chromatin, glycosaminoglycans (particularly heparan and dermatan sulphates), and to all known types of amyloid fibrils. The latter interaction is responsible for the unique specific accumulation of serum amyloid P component in amyloid deposits. Aggregated, but not native, serum amyloid P component also binds specifically to C4-binding protein and fibronectin from plasma, although serum amyloid P component is not complexed with any other protein in the circulation. In addition to being a plasma protein, human serum amyloid P component is also a normal constituent of certain extracellular matrix structures. It is covalently associated with collagen and/or other matrix components in the lamina rara interna of the human glomerular basement membrane, and is present on the microfibrillar mantle of elastin fibres throughout the body.

Although no deficiency of serum amyloid P component has been described, and it has been stably conserved in evolution, its physiological function remains unclear. There is a single copy of its gene on chromosome 1, no polymorphism of the amino acid sequence, and the single biantennary oligosaccharide chain attached to asparagine at residue 32 is the most invariant glycan of any known glycoprotein. Studies of serum amyloid P component knockout mice have shown that serum amyloid P component is involved in host resistance to some infections, and contributes to the pathogenesis of others, but these animals are otherwise healthy and have a normal lifespan.

The serum amyloid P component molecule is highly resistant to proteolysis and, although not itself a proteinase inhibitor, its binding to amyloid fibrils in vitro protects them against proteolysis. Once bound to amyloid fibrils in vivo, serum amyloid P component persists for very prolonged periods and is not catabolized at all, by contrast with its rapid clearance from the plasma (half-life 24 h) and prompt catabolism in the liver. These observations suggest that serum amyloid P component may contribute to the persistence of amyloid deposits in vivo; serum amyloid P component knockout mice show retarded and reduced induction of experimental AA amyloidosis, confirming that serum amyloid P component is significantly involved in the pathogenesis of amyloidosis.

Other proteins in amyloid deposits

A number of plasma proteins, other than the fibril proteins themselves and serum amyloid P component, have been detected immunohistochemically in some amyloid deposits. These include α1-antichymotrypsin, some complement components, apolipoprotein E, and various proteins of the extracellular matrix and basement membrane. None of these is as universal, abundant, or selective as serum amyloid P component, and their role, if any, in the pathogenesis or effects of amyloid deposition is not known.

Diagnosis and monitoring of amyloidosis


Until recently, amyloidosis was an exclusively histological diagnosis, and green birefringence of deposits stained with Congo red and viewed in polarized light remains the gold standard. Furthermore, immunohistochemical staining of amyloid-containing tissue is the simplest method for identifying the type of amyloid fibril present. However, biopsies provide extremely small samples and therefore can never provide information on the extent, localization, progression, or regression of amyloid deposits. A major advance in clinical amyloidosis has been the development of radiolabelled serum amyloid P component as a specific tracer for amyloid. Combined scintigraphic imaging and metabolic analysis using labelled serum amyloid P component have provided a wealth of new information on the natural history of many different forms of amyloid and their response to treatment.

Histochemical diagnosis of amyloidosis


Amyloid may be an incidental finding on biopsy of the kidneys, liver, heart, bowel, peripheral nerves, lymph nodes, skin, thyroid, or bone marrow. When amyloidosis is suspected clinically, biopsy of the rectum or subcutaneous fat is the least invasive. Amyloid is present at these sites in more than 90% of cases of systemic AA or AL amyloidosis. Alternatively, a clinically affected tissue may be biopsied directly.

Congo red and other histochemical stains

Many cotton dyes, fluorochromes, and metachromatic stains have been used, but Congo red staining, and its resultant green birefringence when viewed with high-intensity polarized light, is the pathognomonic histochemical test for amyloidosis. The stain is unstable and must be freshly prepared every 2 months or less in alkaline alcoholic solution. It is critical to have a section thickness of 5 to 10 µm and include in every staining run a positive control tissue containing modest amounts of amyloid.


Although many amyloid fibril proteins can be identified immunohistochemically, the demonstration of amyloidogenic proteins in tissues does not, on its own, establish the presence of amyloid. Congo red staining and green birefringence are always required, and immunostaining may then enable the amyloid to be classified. Antibodies to serum amyloid A protein are commercially available and always stain AA deposits, similarly with anti-β2-microglobulin antisera and haemodialysis-associated amyloid. In AL amyloid the deposits are stainable with standard antisera to κ or λ immunoglobulin light chains in only about one-half of cases, probably because the light-chain fragment in the fibrils is usually the N-terminal variable domain, which is largely unique for each monoclonal protein. Immunohistochemical staining of transthyretin, Aβ, and prion protein amyloid may require pretreatment of sections with formic acid or alkaline guanidine, or deglycosylation.

Electron microscopy

Amyloid fibrils cannot always be convincingly identified ultrastructurally, and electron microscopy alone is not sufficient to confirm the diagnosis of amyloidosis.

Problems of histological diagnosis

The tissue sample must be adequate (e.g. the inclusion of submucosal vessels in a rectal biopsy specimen), and failure to find amyloid does not exclude the diagnosis. The unavoidable sampling problem means that biopsy cannot reveal the extent or distribution of amyloid. Experience with Congo red staining is required if clinically important false negative and false positive results are to be avoided. Immunohistochemical staining requires positive and negative controls, including demonstration of the specificity of staining by absorption of positive antisera with isolated pure antigens. The recent development of powerful proteomic analysis methods applicable to histological sections suggests that, in specialist centres, direct identification of all amyloid fibril proteins may soon be possible on microscopic tissue samples.

Nonhistological investigations

Two-dimensional echocardiography showing small, concentrically hypertrophied ventricles, generally impaired contraction, dilated atria, homogeneously echogenic valves, and ‘sparkling’ echodensity of ventricular walls is virtually diagnostic of cardiac amyloidosis. However, clinically significant restrictive diastolic impairment may be difficult to detect, even by comprehensive Doppler echocardiography and other functional studies. Cardiac magnetic resonance imaging appearances are also characteristic in amyloidosis and coupled with late gadolinium enhancement can be diagnostic.

In cases of known or suspected hereditary amyloidosis, the gene defect must be characterized. If amyloidotic tissue is available the fibril protein may be known and the corresponding gene can then be studied, but if no tissue containing amyloid is available, screening of the genes for known amyloidogenic proteins must be undertaken.

Biochemical and immunochemical tests exist for screening the plasma for amyloidogenic variant protein products of mutant genes, e.g. for transthyretin and apolipoprotein A-I variants, but molecular genetic analysis of DNA is easier to perform and is the most direct approach. However, regardless of the DNA results, it is desirable, if possible, directly to identify the respective protein in the amyloid.

Serum amyloid P component as a specific tracer in amyloidosis

The universal presence in amyloid deposits of amyloid P component, derived from circulating serum amyloid P component, is the basis for the use of radioisotope-labelled serum amyloid P component as a diagnostic tracer in amyloidosis. No localization or retention of labelled serum amyloid P component occurs in healthy subjects or in patients with diseases other than amyloidosis. Radioiodinated serum amyloid P component has a short half-life (24 h) in the plasma and is rapidly catabolized, with complete excretion of the iodinated breakdown products in the urine. However, in patients with systemic or localized extracerebral amyloidosis, the tracer rapidly and specifically localizes to the deposits, in proportion to the quantity of amyloid present, and persists there without breakdown or modification. Highly purified serum amyloid P component, isolated from donor plasma according to pharmaceutical current good manufacturing practice, is oxidatively iodinated under conditions that preserve its function intact. The medium-energy, short half-life, pure γ-emitter 123I is used for scintigraphic imaging, and the long half-life isotope 125I is used for metabolic studies. The dose of radioactivity administered (<4 mSv) is well within accepted safety limits and more than 10 000 studies have been completed without any adverse effects. In addition to high-resolution scintigraphs, the uptake of tracer into various organs can be precisely quantified and, together with highly reproducible metabolic data on the plasma clearance and whole-body retention of activity, the progression or regression of amyloid can be monitored serially and quantitatively.

Important observations regarding amyloid include the following: the different distribution of amyloid in different forms of the disease; amyloid in anatomical sites not available for biopsy (adrenals, spleen); major systemic deposits of forms of amyloid previously thought to be organ-limited; a poor correlation between the quantity of amyloid present in a given organ and the level of organ dysfunction; a nonhomogeneous distribution of amyloid within individual organs; and evidence for rapid progression and sometimes regression of amyloid deposits with different rates in different organs. Examples of major regression of amyloidosis, when it has been possible to reduce or eliminate the supply of fibril precursor, are very encouraging. Studies with labelled serum amyloid P component thus make a valuable contribution to the diagnosis and management of patients with systemic amyloidosis, and in the United Kingdom these are routinely available for all known or suspected cases of amyloidosis in the National Health Service National Amyloidosis Centre at the Royal Free Hospital, London.

Management of amyloidosis

Although no treatments yet exist that specifically promote the mobilization of amyloid, there have been substantial recent advances in the management of systemic amyloidosis, in particular active measures to support failing organ function while attempts are made to reduce the supply of the amyloid fibril precursor protein. Serial serum amyloid P component scintigraphy in more than 2000 patients with various forms of amyloid has shown that control of the primary disease process, or removal of the source of the amyloidogenic precursor, often results in regression of existing deposits and recovery or preservation of organ function. This strongly supports aggressive intervention, and relatively toxic drug regimes or other radical approaches can be justified by the poor prognosis. Such an approach, leading to reduced morbidity and improved survival, was the basis for the establishment of the National Health Service National Amyloid Centre. However, clinical improvement in amyloidosis is often delayed long after the underlying disorder has remitted, reflecting the very gradual regression of the deposits that is now recognized to occur in most patients. Continuing production of the amyloid precursor protein should be monitored as closely as possible in the long term, to determine the requirement for and intensity of treatment for the underlying primary condition. In AA amyloidosis this involves frequent estimation of the plasma SAA level, and in AL amyloidosis it requires monitoring of the serum free light-chain concentration or other markers of the underlying monoclonal plasma-cell proliferation.

The treatment of AA amyloidosis ranges from potent anti-inflammatory and immunosuppressive drugs in patients with rheumatoid arthritis, to lifelong prophylactic colchicine in familial Mediterranean fever, and surgery in conditions such as refractory osteomyelitis and the tumours of Castleman’s disease. The biological cytokine-inhibiting agents antitumour necrosis factor and recombinant interleukin 1 receptor antagonist can induce rapid and complete remission of inflammatory activity in many patients with rheumatoid or juvenile idiopathic arthritis, and those with inherited periodic fever syndromes.

Treatment of AL amyloidosis is based on that for myeloma, although the plasma-cell dyscrasias in AL amyloidosis are often very subtle. Prolonged low-intensity cytotoxic regimes, such as oral melphalan and prednisolone, are beneficial in about 20% of patients. More dose-intensive chemotherapy regimes, such as oral melphalan and dexamethasone, and thalidomide-based regimes, notably oral cyclophosphamide, thalidomide and dexamethasone, are associated with responses in more than 50% of patients, along with manageable toxicity. High-dose chemotherapy with autologous peripheral blood stem-cell transplantation may be associated with even higher response rates, although procedural mortality is high in individuals with multiple amyloidotic organ involvement, especially patients with autonomic neuropathy, severe cardiac amyloidosis, or a history of gastrointestinal bleeding, and in those aged over about 60 years.

The disabling arthralgia of β2-microglobulin amyloidosis may partially respond to nonsteroidal anti-inflammatory drugs or corticosteroids, but even the most severe symptoms usually rapidly vanish following renal transplantation. The basis for this remarkable clinical response is unclear, since although transplantation rapidly restores normal β2-microglobulin metabolism, regression of β2-microglobulin amyloid may not be evident for many years.

Hepatic transplantation is effective in familial amyloid polyneuropathy associated with transthyretin gene mutations, since the variant amyloidogenic protein is produced mainly in the liver. Outcome has proved best among younger patients with the methionine-30 substitution, though even in this group the peripheral neuropathy usually only stabilizes. Unfortunately, paradoxical progression of established cardiac amyloidosis with wild type transthyretin has been observed in many older patients, particularly those with nonmethionine-30 substitutions. Important questions therefore remain about the selection of patients and timing of the procedure but, so far, early intervention seems advisable. On a similar basis, hepatic transplantation has also been successfully undertaken in some patients with hereditary fibrinogen α-chain and apolipoprotein A-I amyloidosis.

Supportive therapy remains critical in systemic amyloidosis, with the potential for delaying target organ failure, maintaining quality of life, and prolonging survival while the underlying process can be treated. Rigorous control of hypertension is vital in renal amyloidosis. Surgical resection of amyloidotic tissue is occasionally beneficial but, in general, a conservative approach to surgery, anaesthesia, and other invasive procedures is advisable. Should any such procedure be undertaken, meticulous attention to blood pressure and fluid balance is essential. Amyloidotic tissues may heal poorly and are liable to bleed. Diuretics and vasoactive drugs should be used cautiously in cardiac amyloidosis because they can reduce cardiac output substantially. Dysrhythmias may respond to conventional pharmacological therapy or to pacing. Replacement of vital organ function, notably dialysis, may be necessary, and cardiac, renal, and liver transplant procedures have a role in selected cases.

Finally, a number of different therapies aimed specifically at inhibiting the formation of amyloid fibrils or promoting fibril regression are currently under development, and some are already being evaluated clinically. The latter include approaches directed at precursor protein production, glycosaminoglycans, serum amyloid P component, preventing aberrant protein folding, and various kinds of immunotherapy, offering hope that amyloidosis may become more readily treatable.

Further reading

Booth DR, et al. (1997). Instability, unfolding and aggregation of human lysozyme variants underlying amyloid fibrillogenesis. Nature, 385, 787–93.
Drüeke TB (1998). Dialysis-related amyloidosis. Nephrol Dial Transplant, 13 Suppl 1, 58–64.
Gertz MA, Merlini G, Treon SP (2004). Amyloidosis and Waldenstrom’s macroglobulinemia. Hematology Am Soc Hematol Educ Program, 257–82.
Hardy J (1997). Amyloid, the presenilins and Alzheimer’s disease. Trends Neurosci, 20, 154–9.
Hawkins PN (2002). Serum amyloid P component scintigraphy for diagnosis and monitoring amyloidosis. Curr Opin Nephrol Hypertens, 11, 649–55.
Hawkins PN, Lavender JP, Pepys MB (1990). Evaluation of systemic amyloidosis by scintigraphy with 123I-labeled serum amyloid P component. N Engl J Med, 323, 508–13. 
Kyle RA, Gertz MA (1995). Primary systemic amyloidosis: clinical and laboratory features in 474 cases. Semin Hematol, 32, 45–9.
Lachmann HJ, et al. (2002). Misdiagnosis of hereditary amyloidosis as AL (primary) amyloidosis. N Engl J Med, 346, 1786–91.
Lachmann HJ, et al. (2003). Outcome in systemic AL amyloidosis in relation to changes in concentration of circulating free immunoglobulin light chains following chemotherapy. Br J Haematol, 122, 78–84.
Lachmann HJ, et al. (2007). Natural history and outcome in systemic AA amyloidosis. N Engl J Med, 356, 2361–2371.
Pepys MB (2006). Amyloidosis. Annu Rev Med, 57, 223–41.
Westermark P et al. (2007). A primer of amyloid nomenclature. Amyloid, 14, 179–83.