Localized amyloids important in diseases outside the brain – lessons from the islets of Langerhans and the thoracic aorta


P. Westermark, Rudbeck Laboratory, C5, SE-751 85 Uppsala, Sweden
Fax: +46 18 55 27 39
Tel: +46 18 611 3849
E-mail: per.westermark@igp.uu.se


It has long been understood that amyloids can be lethal in systemic diseases. More recently, it has been accepted that local cerebral aggregation of the small peptide Aβ is involved in the pathogenesis of Alzheimer’s disease. Protein aggregation, with the generation of small amyloid deposits in specific organs, also occurs outside the central nervous system and often is associated with increased cell death. In this review, we discuss two lesser known but common localized amyloid fibril-forming proteins: the polypeptide hormone islet amyloid polypeptide (IAPP) and the lactadherin-derived peptide medin. IAPP aggregates and induces the depletion of islet β-cells in type 2 diabetes and in islets transplanted into type 1 diabetic subjects. Initial amyloid deposition occurs intracellularly and parts of this amyloid consist of proIAPP. Medin derived from lactadherin expressed by smooth muscle cells aggregates into amyloid in certain arteries, particularly the thoracic aortic media layer, and may have a role in the generation of the potentially lethal conditions of thoracic aortic aneurysm and dissection.


amino acid residue


C/EBP homologous protein


endoplasmic reticulum




heparan sulfate


islet amyloid polypeptide




unfolded protein response


The biochemical revolution in amyloid research, starting with the identification by Benditt et al. [1] and Glenner et al. [2] of two major amyloid fibril proteins in systemic amyloidosis (AA and AL, respectively), has provided detailed insight into a number of different diseases, representing an initially not foreseen outcome. It was understood that the amyloid fibril, which is unexpectedly similar in different diseases, is comprised of small proteins or polypeptides. The generic structure of cross β-sheet conformation with hydrogen bonding between subunits was proposed at an early date [3], although the diverse biochemical nature of amyloid fibril proteins remained unknown [4]. With time, 27 proteins have been shown to comprise major components of fibrils in deposits, fulfilling the criteria for amyloid [5]. It should be noted that proteins of several of the inclusion bodies occurring in neurodegenerative disease are not included. Initially, major interest was given to systemic amyloid forms, simply because these offered the largest amount of amyloid, making purification and analysis of the protein components possible. Subsequently, when the third major amyloid fibril protein was analyzed in the deposits in the C-cell tumor medullary carcinoma of the thyroid and found to comprise calcitonin [6], a more diverse composition of amyloid was suspected. Now, solely in localized deposits, no < 13 amyloid fibril proteins have been characterized [5] and more are expected to be discovered because some of the localized amyloids remain to be characterized chemically.

The first issue to be discussed when dealing with localized amyloid is what should be considered as an amyloid and what should not? This is self-evidently a matter of definition. With current knowledge, it is reasonable to limit the designation ‘amyloid’ to pathological in vivo deposits that fulfill certain criteria: an affinity for the dye Congo red with concomitant greenish–yellow birefringence, a typical fibrillar ultrastructure and a typical X-ray diffraction pattern. Extracellular deposits with these properties also have some additional characteristics that point to some mutual behavior. They all contain heparan sulfate (HS) and bind the plasma protein serum amyloid P-component. In accordance with these criteria, a number of protein aggregates are not included, although, in the literature, they often are referred to as amyloid. These include many of the intracellular protein aggregates found in neurodegenerative diseases. In addition, according to this strict definition, normal protein aggregates and in vitro-made fibrils are not amyloid.

It is very often obvious that amyloid deposits themselves are important in generating disease in the systemic forms. This is generally less obvious for localized amyloids. In those conditions, the amyloid deposits are commonly very small, albeit widely spread in an organ or tissue. A good example is Alzheimer’s disease where amyloid deposits appear as multiple small rounded ‘plaques’ (in reality: globules) in the cerebral cortex. Similar small amyloid deposits are known from a number of different tissues and organs and have generally been overlooked and regarded as a non-important, secondary phenomenon. A change in this view came with the characterization of the amyloid in cerebral congophilic angiopathy and Alzheimer’s ‘plaques’, where it was shown that the deposited material was a novel peptide, now designated Aβ [7,8]. This induced a direct paradigm shift, and most of the current research concerning Alzheimer’s pathogenesis and treatment is centred on Aβ and its aggregation.

Localized amyloid deposits are found not only in many different organs in association with certain disorders, but also in the same tissues in association with apparently normal aging. This fact has made it difficult to confirm the importance of the deposits in the pathogenesis of diseases. Alzheimer’s disease is no exception; although, in general, the amount of amyloid both in vessels and in the gray matter is larger in diseased individuals compared to elderly individuals without any cognitive impairment, the difference is quantitative and not qualitative. Recent research has shown that the mechanism by which amyloid formation is pathogenically important is much more complicated than just being considered the result of a mass effect. This is certainly true for most or all local amyloid deposits.

This review deals with two biochemically divergent localized amyloidoses, of which one is clearly associated with specific diseases. These comprise amyloid formed by islet amyloid polypeptide associated with type 2 diabetes and amyloid consisting of the medin fragment of the precursor lactadherin. The last form may be involved in the pathogenesis of thoracic aortic aneurysm and dissection.

Amyloid in the islets of Langerhans

Amyloid deposits are the most typical morphological islet lesion in type 2 diabetes. The lesion was described before the characteristic hallmarks of Alzheimer’s disease were identified but has attracted much less attention, although it may be a sign of a pathogenic mechanism as important as that of Aβ deposition. The islet amyloid is restricted to the endocrine pancreas and, similar to the Aβ amyloid, it is not part of a systemic deposition. The amyloid fibril consists mainly of the polypeptide hormone islet amyloid polypeptide (IAPP), which is a product of islet β-cells. IAPP is stored together with insulin and several additional peptides in the secretory vesicles, and all components are co-released at exocytosis. However, on a molar basis, the ratio between IAPP and insulin is approximately 1 : 50. The normal function of IAPP is not completely understood, although the peptide is involved in glucose homeostasis, both by paracrine and autocrine loops. High concentrations of IAPP (10 μm) have been shown to inhibit glucose-stimulated, high voltage-gated calcium channels present on the β-cell and thereby to inhibit insulin secretion [9]. IAPP may also act by its effects on cerebral receptors.

Islet amyloid in the development of diabetes

The 37 amino acid residue (aa) IAPP is expressed as an 89 aa preprohormone consisting of a 22 aa leader sequence and a 67 aa proIAPP molecule [10,11]. ProIAPP has two short flanking peptides, cleaved off at maturation at double basic aa residues, similar to (and by the same prohormone convertases) as proinsulin. Further post-translational processing of IAPP includes a C-terminal amidation and the formation of a disulfide bridge between Cys 2 and 7.

Human IAPP is an unusually amyloid-prone molecule in vitro. Particularly important is segment 20–29, which shows the strongest interspecies variation in this otherwise strongly conserved peptide. Proline residues in this segment make IAPP in many species, including the rat and mouse, incapable of fibril formation, whereas other species, such as humans, monkeys and cats, have amyloidogenic IAPP [12]. Interestingly, in these species, a type 2 diabetes-like condition also exists that indicate the pathogenic importance of IAPP in the development of this form of diabetes.

The significance of islet amyloid has been a matter of discussion for a very long time. The mechanism behind the progressive β-cell loss that occurs in patients with type 2 diabetes is unknown, although the formation of amyloid and the amyloid per se can both be responsible for this [13,14]. There is increasing experimental evidence suggesting that aggregation of IAPP is important for the progressive loss of β-cells in type 2 diabetes [15,16], as well as for the gradually impaired function of islets transplanted into type 1 diabetic patients [17,18]. These matters have been reviewed recently [14] and are not discussed in the present review. Here, we review some of the more recent results concerning the mechanisms by which IAPP amyloid develops and how this may result in the loss of islet β-cells.

Initiation of islet amyloid

β-cell stress is an important risk factor for the development of islet amyloid and, in addition to its observance in association with type 2 diabetes and in transplanted islets, islet amyloid has also been described in patients who have undergone gastrectomy [19]. Gastrectomy causes a rapid absorption of carbohydrates and can induce glucose intolerance. On the other hand, β-cell rest can have beneficial effects, including a decreased aggregation of IAPP to fibrils. Thus, early insulin treatment of patients with type 2 diabetes has been shown to be beneficial for β-cell function, whereas treatment with sulfonylurea (glibenclamide), which induces increased insulin and IAPP release, results in decreased C-peptide response [20]. In a study where eight cats were partially pancreatectomized and given either insulin or sulfonylurea, islet amyloid developed in all of the cats treated with sulfonylurea, although only in one of four cats treated with insulin [21].

It is not known how the first islet amyloid develops. IAPP is normally well controlled so that it does not form amyloid, despite its high propensity to aggregate. Insulin can prevent IAPP-aggregation in a concentration-dependent manner [22,23] and NMR analysis demonstrates that the structure formed by residues 11–18 has an important role in the interaction with insulin [24]. Therefore, it is possible that the molar ratio of the peptides is important, especially given that proIAPP and proinsulin expression are normally regulated in parallel [25]. There are several studies showing that, under certain conditions, there may be an increase in the plasma IAPP to insulin ratio. IAPP and insulin plasma levels in mice fed a diet high in fat showed a 4.5-fold increase in fasting plasma IAPP levels compared to animals fed standard chow. There was no significant difference in plasma insulin levels between the groups [26]. In addition, the ratio of prohormone to mature hormone may be important because proIAPP appears to be particularly amyloidogenic, thereby leading to cell death [27,28]. In the early phase of the development of type 2 diabetes, there is a period coincident with deteriorated hormonal secretion. During this period, a shift in the ratio between the prohormone and the biologically active hormone may be observed. In one study, proinsulin was shown to comprise 12.8 ± 0.8% of the insulin-reactive material secreted from the β-cells in healthy controls, whereas proinsulin constituted 35.9 ± 4.1% of the insulin-reactive fraction secreted in patients with type 2 diabetes [29]. There was no significant difference in the total insulin-reactive fraction between individuals with type 2 diabetes compared to healthy controls. A similar finding was recently reported with respect to proIAPP and IAPP serum levels, although the concentrations of IAPP were somewhat difficult to interpret. An increase of plasma proIAPP occurred in individuals with impaired glucose tolerance and type 2 diabetes [30]. The detected proIAPP levels in control subjects were 6.95 ± 2.5 pmol·L−1, 15.51 ± 3.74 pmol·L−1 in the impaired glucose tolerance group and 21.15 ± 4.74 pmol·L−1 in the type 2 diabetes group. A similar increase of IAPP levels was not detected, and therefore the fraction of proIAPP was increased compared to IAPP [30].

Findings with hIAPP transgenic rodent models

Because islet amyloid does not develop in the mouse and rat, it has proven difficult to study its initiation and progression. The development of human IAPP transgenic mouse [31] and rat [32] strains has helped to create models that are relevant to this important topic. We have used mice from a strain that expresses human IAPP under the regulation of the insulin promoter, and these animals are also deficient for murine IAPP [33]. Male mice from this strain develop islet amyloid when fed a high-fat diet for 12 months. In mouse islets stained for amyloid with Congo red, both extracellular and intracellular amyloid could be detected. Immunolabeling with antibodies reactive against the processing sites of proIAPP, and therefore specific for proIAPP, labeled a fibrillar aggregate within the halo region of the secretory granules of mice fed a high-fat diet (Fig. 1) [34]. Intracellular amyloid has been described in systems with rapidly developing islet amyloid [35,36]. Consequently, it is possible that IAPP amyloid formation is initiated intracellularly and that the propagation of extracellular amyloid occurs later when the intracellular deposits have led to cell death [34].

Figure 1.

 (A) Intracellular amyloid in β-cells of a human IAPP transgenic mouse. Some of the aggregates are encircled by a membrane and this suggests that they are formed intragranularly. (B) Fibrillar material is present in the halo region of the secretory granule. The fibrils are immunolabeled with IAPP antibodies, shown with 10 nm gold particles (white arrows). (C) No fibrilar aggregates are present in the halo region of secretory granules from transgenic mice fed standard show, which do not develop amyloid (black arrows).

How is endoplasmic reticulum (ER) stress linked to islet amyloid?

As indicated above, increased hormonal synthesis can lead to ER-stress and activation of the unfolded protein response (UPR). If the UPR is insufficient or if ER-stress becomes chronic, pathways to apoptosis can be activated. Overexpression of aggregation-prone proteins such as IAPP, which can form cell toxic oligomers, is suggested to cause ER-stress and initiate the UPR [37]. Different proteins can be used for monitoring ER-stress, including Bip (also known as heat-shock 70 kDa protein 5). This is an ER-luminal chaperone that assists in the folding of newly-synthesized proteins and is also important for transport of misfolded proteins from the ER to the proteasomal degradation system [38]. A second marker for ER-stress is CHOP (C/EBP homologous protein, also known as growth arrest- and DNA damage-inducible gene 153). When activated, CHOP translocates from the cytosol to the cell nucleus where it can activate apoptosis [39]. Bip and CHOP have both been detected in islets from patients with type 2 diabetes [40,41]. Oligomeric aggregates can escape the ER-compartment, allowing them to have direct effects on mitochondrial membranes, leading to the production of reactive oxygen species [42]. It should be emphasized that the processing of proIAPP into IAPP takes place in the late Golgi and in the secretory granules, and therefore it is most likely misfolded proIAPP that causes ER-stress [43]. Hull et al. [44] failed to verify the induction of ER-stress in mice during amyloid formation. Instead, they showed that extracellular amyloid induces reactive oxygen species production and cell death [44].

Misfolded proteins are ubiquitinated and degraded by the proteasomal pathway. Protein aggregates present in the cytosol are ubiquitinated but cannot be degraded by the proteasome. Instead the newly-described aggrephagy, an autophagy-related pathway, is assumed to be responsible for this degradation [45]. Aggregates destined for degradation are encircled by a double membrane, forming an autophagosome that fuses with a lysosome, and the aggregates are then degraded. If autophagy is hampered, autophagolysosomes can accumulate in the cytosol, blocking cell function and possibly leading to cell death [46]. Autophagic vacuoles have accumulated in β-cells of individuals with type 2 diabetes [47]. Neuronal driven expression of Aβ 1-42 in Drosophila melanogaster results in an autophagy-mediated neurodegeneration. The mechanism responsible for this involves the formation of dysfunctional autophagolysosomes, oxidative stress and cytoplasmic acidification caused by disintegrating autophagolysosomes [48]. We have expressed human proIAPP and human IAPP in D. melanogaster and reported that this results in amyloid [49]. When the intracellular response to protein aggregation was monitored, there were limited signs of ER-stress and activation of UPR and, instead, the autophagy pathway had a more prominent role in the handling of the misfolded proteins (Schultz SW, Gu X, Rusten TE, Alenius M, Westermark GT).

When amyloid-containing cells die, amyloid can remain in the tissue and seed further amyloid made up of IAPP secreted from the surrounding β-cells, causing the amyloid mass to grow [34].

Inflammation is linked to type 2 diabetes and β-cell stress

There is evidence that a chronic inflammatory process is partially responsible for the loss of insulin production and β-cell death in type 2 diabetes [50]. The inflammasome is a protein complex that contains caspase 1, which is responsible for cleavage and thereby activation of the proinflammatory cytokine pro-interleukin (IL)-1β into IL-1β. It was recently shown that misfolded and aggregated IAPP can activate the inflammasome in macrophages and cause IL-1β secretion, thereby triggering inflammation [51]. Macrophages are present in the islets and they can phagocytose extracellular amyloid. It remains to be determined whether IAPP aggregated within the β-cells can trigger the inflammasome reaction. This is new area of research that links protein aggregation to inflammation.

HS and islet amyloid

HS is a component present in all kinds of amyloid and is of importance in amyloidogenesis [52]. HS is present in islet amyloid [53] and HS proteoglycan accelerates the formation of amyloid fibrils from IAPP and proIAPP [54]. HS proteoglycan interaction occurs with specific residues in proIAPP and IAPP [54]. The inhibition of glycosaminoglycan synthesis reduces amyloid formation [55]. Heparanase is an enzyme that cleaves the HS chains present on the cell surface and in the extracellular matrix into shorter fragments. The overexpression of heparanase in HEK 293 cells was shown to reduce sensitivity to Aβ aggregation [56]. The development of systemic AA-amyloidosis in response to an inflammatory stimulus in mice overexpressing human heparanase resulted in a marked reduction in amyloid mass [57]. We have established a double transgenic mouse strain that overexpresses hIAPP and heparanase. When isolated islets from these mice were cultured in high glucose for 2 weeks, a significant decrease in amyloid accumulation was observed. In addition, control mice expressing hIAPP deposited 2.6-fold more amyloid compared to the double transgene (Oskarsson et al., unpublished results). It has been shown that HS interacts with IAPP and that, if this takes place on the cell surface, it can create small spots with a high peptide concentration that might initiate the aggregation of IAPP. This could be responsible for the amyloid picture observed in vivo with bundles of amyloid associated with the cell membrane [58].

Amyloid in the aortic media

Although localized aortic medial amyloid may be the most common of all human amyloid forms, it represents one of the least recognized types. It occurs in almost all individuals aged > 50–60 years [59–62]. Because the amyloid is restricted to arteries, it is regarded as a localized form, although it can be found in different parts of the body, including the central nervous system [63]. Although the amyloid is most common and most pronounced in the media of the thoracic aorta, it also affects the abdominal part, as well as larger arterial branches, particularly those in the upper part of the body [64]. The aortic media is composed of lamellar units, each consisting of two elastic lamellae with a single layer of smooth muscle cells in the middle [65]. In the aortic media, amyloid is seen in the form of thin deposits along elastic membranes or as somewhat larger lumps, often quite widely spread. However, intracellular deposits are common within smooth muscle cells. In arterial branches, such as the temporal artery where deposits are commonly found, amyloid is always most pronounced in association with the internal elastic lamina [63].

Nature of the aortic media amyloid

It was clear from immunohistochemical studies that the fibril protein in aortic media amyloid is different from the other amyloids that had been studied previously [61,62]. Purification and aa sequence analysis by Edman degradation of the main protein from three different aortas revealed a 50 aa protein with a ragged N-terminus, with species starting both before and after that of the main protein component [66]. The protein, named ‘medin’, was revealed to be an internal cleavage product of its precursor lactadherin, previously known as milk fat globule epidermal growth factor 8, as well as several other names [67]. It was originally identified in milk (human, cow, mouse) as a component of the fat globule membrane [68]. Human lactadherin is a 364 aa protein, expressed as a 387 preprotein. It contains several different domains (Fig. 2). It has an epidermal growth factor-like N-terminus containing an integrin binding RGD motif [69]. The remainder (> 300 aa) of the protein consists of two domains (C1 and C2), with homology to domains in coagulation factors V and VIII. In addition to the full-length 46 kDa protein, an N-terminally truncated variant of 30 kDa has been isolated from human milk.

Figure 2.

 The lactadherin molecule contains several domains. In the epidermal growth factor-like domain in the N-terminal region, there is a Arg-Gly-Asp sequence. The C-terminal part of the molecule consists of two discoidin-like domains (C1 and C2), with the last containing the 50 aa residue peptide medin. Reproduced with permission [92]. EGF, epidermal growth factor.

Lactadherin is expressed in many different cells, including breast epithelium, macrophages, glia cells and smooth muscle cells [66–68,70,71]. In the aorta, and at least in many other larger arteries, it is expressed by smooth muscle cells of the media layer and by endothelial cells [63]. Because it has a leader sequence, it is expected to be secreted.

Lactadherin function

A number of different effects and suggested functions have been ascribed to lactadherin. It has been found to protect the suckling from rotavirus infection [72] and can compete with coagulation factors V and VIII and thereby inhibit coagulation [73]. Lactadherin binds to integrins on macrophages through its RGD motif and to phosphatidylserine exposed on injured cells through the C2 domain [70], most likely especially through the end of that domain [74]. Accordingly, lactadherin acts as an important mechanism for the removal of apoptotic cells [75]. Lactadherin-deficient mice show reduced phagocytosis, resulting in the accumulation of apoptotic cells and debris in aortic atherosclerotic plaques [76]. Lactadherin has been suggested to play a role in dendritic cells, where it was found to be complexed with exosomes [77]. Additionally, a connection with another amyloid disease has been noted because lactadherin binds to Aβ and lactadherin-deficiency in Alzheimer’s disease may inhibit Aβ clearance [71]. If this is the case, lactadherin might also participate in the clearance of a number of other proteins.

Medin constitutes positions 245–294 of lactadherin and is situated in the C2 domain (Fig. 2). The crystal structure of the C2 domain of lactadherin has been determined previously [78,79], revealing that the structure is very similar to the corresponding domain of factor V and factor VIII, containing 19 antiparallel β strands, of which eight constitute a distorted β-barrel. There are three relatively large loops at the lower surface of the lactadherin C2 domain, which have been referred to as spikes [79]. Spikes 1 and 3, displaying solvent-exposed hydrophobic residues, were shown to be important for binding to the phosphatidyl serine. The medin motif starts immediately after spike 1 and ends at the end of spike 3. Accordingly, medin contains spike 2, which has no known binding functions [79], and spike 3, which is important for adhesion. The C2 domain of factor VIII is not the only part of the molecule that is important for membrane binding [80] and this may also be true for lactadherin.

Lactadherin and medin binding to elastin

Given the fact that lactadherin is expressed by aortic smooth muscle cells and that medin amyloid deposits are found along elastic lamine, it is natural to ask whether the protein interacts with elastic material. With the aid of several systems, we have demonstrated concentration-dependent binding of both medin and lactadherin to tropoelastin [81]. We suggested that lactadherin, similar to fibulin-5 [82–84], is important for the arrangement and anchoring of elastic fibers to cells. The results of our previous study indicate that elastin binding occurs through the medin domain [81].

There is no evidence that medin is a product of alternative splicing; rather, it appears to be enzymatically cleaved out from lactadherin. This possibility fits with the fact that the N- and C-termini of the molecule are positioned at two long loops and that the N-terminus of medin is heterogeneous [66]. Whether medin is an abnormal cleavage product or occurs normally (with possible biological function) is still unknown. Medin has not been described from other tissues expressing lactadherin. Given the common finding of intracellular amyloid in aortic smooth muscle cells, it is likely that cleavage takes place in an intracellular compartment, such as the ER or Golgi. This should also mean that the cleavage of lactadherin to release medin is independent of the binding to elastin. However, further studies are clearly required.

Amyloid fibril formation from medin

Full-length (50 aa) medin is fibrillogenic in vitro, and, at a concentration of 0.25–05 mm in water, most of the peptide was converted into fibrils within 5–7 days [85]. Similar to several other amyloid fibril proteins, the conversion of medin into fibrils proceeds via an α helical state [86,87]. Heparin enhances fibril formation and results in a shorter lifetime of the prefibrillar toxic oligomeric species [88]. Analyses with the SPOT technique [89], tango prediction [90], the 3D profile method [91] and synthetic short peptides, all indicate that the most fibrillogenic region is close to the C-terminus of medin [85], with the two phenylalanine residues at positions 43 and 48 of medin (287 and 292 of lactadherin) appearing to be particularly important [92]. There are no results available indicating that other parts of lactadherin or the full-length molecule are part of amyloid fibrils. Although the molecules are not evolutionary related, there are some interesting similarities between the amyloid-determination regions of Aβ, IAPP and medin [93,94].

The possible importance of medin amyloid in vascular diseases

Medin and giant cell arteritis

The occurrence of small amyloid deposits in the aortic media has been known for a long time but has attracted very little attention. Small deposits close to the inner elastic lamina of the temporal artery have also been commonly described [95]. Aortic media amyloid is so common that it has been claimed that everyone aged > 55 years of age has some deposits. This is an important reason why the alteration has been disregarded. Given the fact that medin amyloid is affecting the inner elastic lamina of the temporal artery, a site that is primarily affected in giant cell arteritis, we suggested that the amyloid, or the protein in another form, is involved in the development of this disease, which is suspected to have an autoimmune pathogenesis but with unknown antigens [63]. An elastin component has been suggested to be an autoimmune component and medin immunoreactivity and even amyloid was sometimes demonstrated in the giant cells. This finding comprises very indirect evidence for the participation of medin in temporal arteritis, although it is important to explore this possibility further. There is an interesting connection with the thoracic aorta, which is the most prevalent site of medin amyloid, in that giant cell arteritis of the temporal artery may be associated with giant cell aortitis, affecting the vessel’s media and with a risk of the development of aortic aneurysm or dissection [96,97]. This relationship has been questioned, however, and biopsies of the aorta are usually taken only after vascular catastrophes such as aortic dissection [98]. In addition, there is a notable association described between Aβ amyloid angiopathy and granulomatous angiitis [99]. Furthermore, some mice in a transgenic mouse model of cerebral amyloid angiopathy developed vasculitis [100]. These two last examples indicate that the vasculitis might be secondary to the amyloid, although further studies are warranted.

Medin and thoracic aortic aneurysm and dissection

Thoracic aortic aneurysm and dissection are life-threatening diseases with high acute mortality. In both conditions, the media of the aorta is weakened. Rare causes are mutations in the gene for fibrillin-1, leading to defective elastic material but, in most cases, there is no known molecular abnormality. Morphologically, ‘cystic media degeneration’ is often seen. Given the fact that medin amyloid and the defects in the thoracic aortic aneurysm and dissection are localized to the same anatomical site, it was natural to investigate a possible relationship. Unexpectedly, the amount of amyloid was significantly lower in both diseases compared to age-matched control specimens [101]. On the other hand, the amount of immune reactive medin was not lower, indicating an increased concentration of molecules in non-amyloid form. Water soluble aggregates of medin were extractable from aneurysm and dissection materials [101]. Cell toxic mechanisms of pre-fibrillar amyloid protein aggregates are assumed to be important for the tissue lesions in other forms of amyloid-related disorders, including Alzheimer’s disease and type 2 diabetes. Aggregated medin was found to be toxic to smooth muscle cells in vitro [88] and induced the increased production of metalloproteinase 2 [101], which is one of the metalloproteinases found to be upregulated in this condition.

Medin may also be deleterious to the vascular wall, particularly the aorta, in other direct ways. If medin forms from the cleavage of lactadherin already bound to elastin, this may disrupt the anchoring of smooth muscle cells to elastin. If medin forms from lactadherin that is not associated with other components (e.g. intracellularly), soluble medin should be able to compete with lactadherin for binding to elastin [81]. One of the key morphological findings in aortic dissection in older patients with Marfan syndrome (and also in those without this disease) was the loss of a connection between elastic lamellae and smooth muscle cells [102]. We suggest that medin may be an important player in the pathogenesis of thoracic aortic aneurysm and dissection (Fig. 3).

Figure 3.

 Some of the possible ways by which lactadherin (green + red) and medin (red) may be involved in the pathogenesis of aortic aneurysm and dissection. If lactadherin acts as a linker between elastin and smooth muscle cells, cleavage of the molecule (arrow) to release medin will break this linkage. Medin can also compete with lactadherin in the binding to elastin. Aggregated medin in a non-amyloid oligomeric form may be toxic to smooth muscle cells and induce apoptosis, either by an intracellular mechanism or by interaction with the cell membrane. SMC, smooth muscle cell.


The misfolding and aggregation of specific proteins is shown (or suspected) to be an important pathogenic mechanism in a great number of neurodegenerative diseases, with some being a result of the formation of characteristic amyloid deposits. It would not be unexpected that the same situation is true for many other diseases outside the central nervous system as well. Greater attention should be given to the common but overlooked small amyloid deposits seen in many organs in association with aging. Tissue injury can become massive and lead to severe dysfunction of the involved organ, irrespective of whether the toxic effects are exerted by oligomeric aggregates or mature fibrils. In the present review, we have discussed only two amyloid forms, although there are others, including amyloid derived from atrial natriuretic factor in the cardiac atria, amyloid in seminal vesicles derived from semenogelin I, and amyloid in the parathyroid glands containing an unknown protein. Future research may demonstrate that these amyloids are critical for the deleterious effects associated with the aging process.


We thank our graduate students who have performed much of the work cited in this review. Our own research has been supported by the Swedish Research Council, the Swedish Diabetes Association, the Novo Nordic Insulin Fund, the Family Ernfors Fund and the Heart Lung Foundation.