Systemic AA amyloidosis in the red fox (Vulpes vulpes)

Abstract

Amyloid A (AA) amyloidosis occurs spontaneously in many mammals and birds, but the prevalence varies considerably among different species, and even among subgroups of the same species. The Blue fox and the Gray fox seem to be resistant to the development of AA amyloidosis, while Island foxes have a high prevalence of the disease. Herein, we report on the identification of AA amyloidosis in the Red fox (Vulpes vulpes). Edman degradation and tandem MS analysis of proteolyzed amyloid protein revealed that the amyloid partly was composed of full‐length SAA. Its amino acid sequence was determined and found to consist of 111 amino acid residues. Based on inter‐species sequence comparisons we found four residue exchanges (Ser31, Lys63, Leu71, Lys72) between the Red and Blue fox SAAs. Lys63 seems unique to the Red fox SAA. We found no obvious explanation to how these exchanges might correlate with the reported differences in SAA amyloidogenicity. Furthermore, in contrast to fibrils from many other mammalian species, the isolated amyloid fibrils from Red fox did not seed AA amyloidosis in a mouse model.

Introduction

Amyloidoses comprise a large group of protein folding disorders, characterized by aggregation of a misfolded protein into about 10 nm thick β‐pleated sheet, straight fibrils of undetermined length. Many different proteins can form amyloid‐like fibrils under denaturing conditions in vitro, but only about 40 specific proteins are known to form amyloid in association with human and animal disease.1 Two main groups of amyloidosis occur, systemic and localized. The systemic amyloidoses are generally life‐threatening with deposits in many organs. The fibril protein is synthesized in its native form at one location such as the liver or bone marrow, released to the circulation and finally deposits as fibrils in a variety of tissues. Amyloid A (AA) amyloidosis was the most prevalent systemic amyloidosis in humans in the western world until efficient treatment of rheumatic diseases were introduced2 and is the predominating form of systemic amyloidosis in animals.3

A prerequisite for the development of AA amyloidosis is a high plasma concentration of the fibril precursor serum AA (SAA), which is an acute phase reactant. SAA is a highly conserved protein composed of 104 to 112 amino acid residues, which in humans folds into a four helix bundle.4 It is expressed mainly in the liver as an apolipoprotein and is normally complexed with high density lipoprotein particles in serum.5 The SAA plasma concentration can increase 1000‐fold in inflammatory conditions and reach >1 g/L.6 Of yet unknown reasons, the highly soluble and alpha‐helical SAA can aggregate into amyloid fibrils and lead to a severe disease, primarily due to renal lesions. In the amyloid aggregates, protein AA often contains predominantly of N‐terminal fragments of SAA, although full‐length SAA is usually a minor component. Human protein AA may lack one or two amino acids at the N‐terminus, and between 9 and 59 amino acid residues at the C‐terminus.7

AA amyloidosis occurs spontaneously in many mammalian and avian species, but the prevalence varies considerably even between subgroups of the same species. In dogs, AA amyloidosis is common, while the disease has never been described in the Blue fox (Alopex lagopus), and was not found in Gray foxes (Urocyon cinereoargenteus ; n = 401).8, 9 In contrast, foxes that live on the Channel Islands off the Californian coast (Urocyon littoralis) suffer from a high prevalence of AA amyloidosis, possibly due to genetic variations in the SAA gene.9

Herein, we report on the identification of AA amyloidosis in the Red fox (Vulpes vulpes). Mass spectrometry analysis confirmed that SAA, and SAA derived protein fragments, were the main constituents of the amyloid deposits and the Red fox SAA amino acid sequence was determined. Surprisingly, the isolated amyloid fibrils did not seed AA amyloidosis in a mouse model.

Results and Discussion

Histological examination revealed that two of 25 Red foxes (8%) randomly selected from the general wildlife disease surveillance cases necropsied at the Swedish National Veterinary Institute (SVA) suffered from systemic amyloidosis. The two affected foxes were found dead and suffered from severe generalized inflammatory processes. The first (V 3399/13), was a male in normal body condition that had multiple cutaneous bitemarks from another fox or canine. Cause of death was sepsis secondary to suppurative cellulitis (phlegmone), pleuritis and myositis. The second animal (V 309/14), also male, was cachectic, i.e., displayed serous fat atrophy of the bone marrow. It was an older fox judged by advanced dental attrition. The spleen, kidney and liver were firmer than normal and the spleen was mildly to moderately enlarged. The animal had suppurative wounds in the head region, likely the result of bite wounds, oral ulcers under the tongue, and gastric ulcers. Both animals had amyloid deposits in the spleen and kidney at histopathology [Fig. 1(A–C)], and one (V 309/14) also displayed characteristic deposits in the liver. None of the brain samples contained amyloid. The chronic inflammatory processes most likely led to persistent high serum level of SAA in the affected animals, a known risk factor for the development of AA amyloidosis.

Figure 1.

Histological and biochemical characterization of Red fox amyloid. Histological sections from spleen (A, B) and kidney (C) that are stained with Congo red. The amyloid deposits stain red in A and C and show birefringence between crossed polars (B). SDS‐PAGE of tissue extract (D, left lane, stained with Coomassie blue) shows distinct bands with apparent molecular weights of around 10–14 kDa. Western blot analysis of this material with antiserum against human protein AA (middle lane) shows reaction with this band but also with some bands with higher molecular masses, probably aggregation products. Western blot with antiserum A144 against a C‐terminal synthetic peptide of SAA shows reaction corresponding to the upper part of protein AA main band visualized by SDS‐PAGE and by western blot with AA5.

To determine the main protein component of the amyloid deposits we extracted the insoluble protein fraction, i.e., fibrils from the tissue samples. These fibrils were obtained by repeated (x 6) homogenization and centrifugation in 0.15 M NaCl followed by dH₂0 × 2 and lyophilization. SDS PAGE analysis of dissolved fibrils and of material from the major retarded peak after gel filtration, showed distinct bands with apparent molecular weights of around 10–12 kDa [Fig. 1(D)], which is compatible with the previously described 111‐residue SAA from Island and Blue fox.8, 9 Western blot with two different antisera identified the protein band materials as protein AA [Fig. 1(D)]. LC‐MS analysis of trypsinated extract material confirmed that the main constituent of the protein deposits was SAA (Supporting Information Table S1). To further characterize the amyloid protein, we determined its primary structure by a 4‐stage use of mass spectrometry and Edman degradation. Initially, the latter revealed the crude protein material from the gel filtration to have a blocked N‐terminus, in further analysis (see below) found to be pyro‐Glu‐converted Gln, like in Blue fox8 and other SAA proteins.

As next stage, we separately tested direct injection of unresolved tryptic or Glu‐C digests of the whole protein with a Thermo Q Exactive Plus Orbitrap instrument for MS/MS fragment analysis with subsequent Mascot 2.4 search engine assignment against the known Blue fox SAA sequence as master protein. However, this also failed to give interpretable identifications, suggesting that several differences must exist between the Blue and Red fox SAA sequences.

As third stage, we digested the Red fox SAA protein in separate batches with Glu‐C protease, Lys‐C protease, Arg‐C protease and trypsin, separated the fragments by C8‐HPLC, and submitted each of the fragments to individual sequence analysis on the Procise HT sequencer. This produced the primary structure of 60% of the Red fox sequence (positions 9–15, 19–41, 47–60, 62–76, and 92–99, see Supporting Information Fig. S1), and thereby established four amino acid residue replacements in the Red fox SAA protein, relative to the Blue fox variant: S31A, K63N, L71F, and K72L (Fig. 2).

Figure 2.

Sequence alignment of SAA from human (Homo sapiens, Uniprot ID: P0DJI8), island fox (Urocyon littorals),9 dog (Canis lupus familiaris, UniProt ID: P19708), Red fox (Vulpes vulpes) and Blue fox (Alopex lagopus).8 The location of the α‐helices determined by X‐ray crystallography is indicated by blue bars. The location of the amyloidogenic peptides identified by ZipperDB and Tango are indicated by green dots. Amino acid substitutions between the Red fox and Blue fox SAA are indicated in red. The semi‐invariant and invariant residues between all sequences are indicated by dots and asterisks, respectively.

Finally, as the fourth stage of analysis, we submitted stage‐2‐proteolytic digests (above) to the Orbitrap MS/MS analysis, but now with the 4‐residue exchanged SAA sequence as master sequence, and then received a perfect match, identifying all residues in Red fox SAA (Supporting Information Fig. S1). This confirmed the four exchanges, proving that Red fox SAA has 111 residues in total, the N‐terminus to be Pyro‐Glu converted Gln, and defining the amino acid sequence of Red fox SAA as given in Supporting Information Figure S1.

With this Red fox sequence information completed, we aligned SAA primary structures (Fig. 2), establishing that 1 of the 4 replacements, Lys63, is unique to the Red fox variant. Furthermore, this replacement is in the amyloidogenic helix 3, although outside the amyloid core segment (Fig. 2), considered to be a major determinant of SAA amyloid formation.4 One of the other replacements, Ser31, is adjacent to helix 2 and the remaining two replacements (Leu71 and Lys72) are located in between helices 2 and 3, in a region deleted in human SAA (Fig. 2).

As expected, Red fox SAA has high sequence similarities to SAA from other species. Since the Blue fox may be resistant to the development of AA amyloidosis,8 the sequence differences between these two species are of interest. Three of the four positions in which the amino acid sequence between the Red fox and Blue fox SAA differ are located in the loops between helices, as judged from sequence alignment to human SAA, for which a 3D structure is available.4 In general, loop regions are less important than secondary structure elements for tertiary interactions. Hence sequence differences in loops are more easily tolerated and less likely to significantly affect the stability of the protein. The fourth variable position is located in helix 3 that is reported to be amyloidogenic. In this position a Lys residue is found in the Red fox SAA while in SAA from other canine species, and also in human SAA, there is an Asn residue. This charge replacement may affect SAA solubility but its relevance for amyloid formation is not evident since residue identity at this position does not appear to segregate with observed ability to form amyloid. According to prediction algorithms used to identify aggregation prone peptide segments10, 11 there are two amyloidogenic peptide segments in SAA, the first in helix 1 and the second in helix 3 (Fig. 2). Both these segments are 100% identical in SAA from Red and Blue fox. Similarly, amino acid residues that constitute the core of the four helix bundle or make C‐terminal tail interface contacts4 may be important for the amyloidogenicity of SAA. These positions are conserved in all SAA sequences analyzed herein.

Long‐term elevation of circulating SAA, secondary to infectious, immune‐mediated, neoplastic or familial disease, precedes the development of AA amyloidosis. Risk factors for the development of AA amyloidosis in Island foxes are nephritis, captivity, age, and possibly genetic variations and/or an exogenous factor (since the risk of AA amyloidosis is larger in some geographic areas).9 Despite a high frequency of chronic inflammatory disease, most individuals do not develop AA amyloidosis. The reason for this and the mechanism whereby soluble, α‐helical SAA converts into insoluble, β‐sheet fibrils are incompletely understood, but cleavage of the C‐terminal part of SAA may be important since this part seems to be important for maintained tertiary structure.4 Usually, an N‐terminal SAA fragment, in human often composed of residues 1–76, is predominant in amyloid aggregates,12, 13 but there are many exceptions since e.g., the amyloid deposits in the Island foxes contained mainly full length SAA.9 Our LC‐MS/MS of the amyloid deposits resulted in generally stronger signals for fragments originating from the N‐terminal part than the C‐terminal part of SAA. This suggests that the majority of amyloid SAA was C‐terminally truncated, but other explanations like different stabilities of N‐ and C‐terminal fragment ions in the MS cannot be excluded. However, western blot results also indicated that the fibril protein consisted of fragments of SAA and that full‐length SAA only constituted a minor part [Fig. 1(D)]. In addition to SAA, apolipoprotein E (ApoE) was also identified by LC‐MS/MS (with a high Mascot score—283), suggesting that it is a constituent of the deposits (Supporting Information Fig. S1). Interestingly, ApoE has been shown to be involved in another amyloid disease, Alzheimer disease (AD). Evidence for this comes from studies showing that ApoE co‐purifies with the Aβ amyloid from AD brains14 and the ɛ4 allele has a strong genetic association with late onset AD. Thus, it is possible that ApoE could influence also the susceptibility to develop SAA amyloid in foxes. However, ApoE is a common finding with several human amyloid types and its importance here is elusive.

The lag‐time before fibril formation can be markedly reduced and the rate of amyloid formation greatly enhanced by addition of preformed fibrils or nuclei, a phenomenon known as seeding.15 Many mouse strains develop, after a fairly long lag phase (weeks or months) AA amyloidosis after a chronic inflammatory challenge. The lag phase is dramatically reduced to a few days if the animals are injected with fibril extracts from a liver of an amyloidotic mouse. This is astonishingly efficient at serial dilutions down to picograms of protein.16 Amyloid fibrils from other species can also seed AA amyloid in the inflamed mouse model by even oral intake17, 18 and it has been shown without any doubt that AA and apolipoprotein AII systemic amyloidoses are transmissible in mice.15, 16, 19 AA amyloidosis is also transmissible in minks and non‐mammalian species such as chicken. Cheetahs in captivity often suffer from AA amyloidosis that may spread via faeces20 and indicates that oral transmission may occur. To test whether the isolated fibrils from a Red fox could seed AA amyloidosis in mice, 27 mice were given a subcutaneous injection of silver nitrate to cause a sterile abscess. Thirteen animals were then given i.v. injections of amyloid fibrils that were extracted from one of the affected foxes studied here, while 14 mice were given buffer only. In the experimental mouse model of AA amyloidosis the spleen is invariably affected21, 22 and was the organ used in the present study. Cross‐seeding murine AA amyloidosis with fibrils from other species has been described by others17, 23 and we have previously shown that AA‐amyloid fibrils extracted from cow, cat, donkey, and dog act as seed for AA amyloid in mouse using the same protocol as in the present study (unpublished data). Therefore, the finding that Red fox amyloid did not seed AA amyloidosis in mice was surprising. Only one out of 13 injected mice had developed a small deposit.

In conclusion, we characterized Red fox AA amyloid and found distinct residue replacements compared to other fox species, but they do apparently not directly explain observed species differences in amyloid prevalence, or the inability of Red fox amyloid to seed AA amyloidosis in a mouse model.

Materials and Methods

Sample collection

Twenty‐five carcasses from free ranging Red foxes (Vulpes vulpes) that were sent to the Swedish Veterinary Institute, Uppsala, Sweden for necropsy were included in the study. The foxes were either found dead or euthanized due to disease or trauma (9 animals), or shot during fox hunting and sent for autopsy as a part of the general wildlife disease surveillance, or as part of the national Echinococcus surveillance program.24 The 25 foxes represent every second of the foxes submitted to the Veterinary Institute between Oct 1 2013 and May 15 2014, in total 61, that were possible to evaluate doing a full necropsy. Tissue samples from the spleen, kidney, brain, and liver were collected for histological examination.

Histology

Tissue specimens were fixed in 4% neutral buffered formaldehyde solution, dehydrated and embedded in paraffin. Sections were stained with alkaline Congo red solution and examined in a polarization microscope for amyloid or immunolabelled with a rabbit antiserum against murine SAA25 crossreacting with AA from many different mammalian species.

Isolation and characterization of amyloid

About 2 g of splenic tissue from one fox with heavy amyloid deposits were homogenized in 0.15 M NaCl followed by centrifugation at 15,000g. Supernatant was discarded and pellet material was rehomogenized as above. The procedure was repeated 6 times followed by homogenization in distilled water and centrifugation × 2. Finally, pellet material was lyophilized. For purification of the fibril protein, fibrils were delipidized, dissolved in 6 M guanidine HCl and subjected to gel filtration as described.26 Fractions containing the main retarded peak material were pooled, dialyzed exhaustively against deionized water and lyophilized.

Extracted fibrils and purified protein were analyzed by SDS‐PAGE using a 15% acrylamide gel stained with Coomassie Brilliant Blue. Western blot was performed as described27 with anti‐human AA rabbit antiserum A528 and with a rabbit antiserum against a synthetic peptide (DPNHFRPAGLPEKY) corresponding to position 91–104 of humans SAA which is identical to Red fox SAA positions 98–111.

Liquid chromatography mass spectrometry (LC‐MS)

The protein composition of the amyloid protein sample was analyzed with a nanospray LC‐MS system; Agilent 6330 ion trap coupled to an Agilent 1200 HPLC pump system with a chip cube interface using C18 reverse phase chip with integrated electrospray. Five microgram purified and trypsinated amyloid sample was dissolved in 2% aqueous acetonitrile containing 0.2% formic acid, injected into the C18 chip, separated at 0.3 µL/min with increasing concentrations of acetonitrile containing formic acid and detected in positive mode MS. LC‐settings; buffer A, 2% acetonitrile 0.2% formic acid and Buffer B, acetonitrile with 0.2% formic acid. The gradient included separation of peptides and wash and reconstitution of the column as described: 0–92 min; 3–26% acetonitrile, 92–112 min; 26–36%, 112–136 min; 36–60%, 137–141 min; 60–95%, 137–141 min; 95%, 141–143 min 95–93%; 143–147.9 min 3%. The three most abundant ions in each scan (m/z 228–1800 scan range) were subjected to tandem MS.

For protein identification, the tandem MS data was exported in mascot generic format and uploaded to the online tool “Mascot ms/ms ions search” (Matrix science). Search settings included Swiss‐prot database, max 1 missed trypsin cleavage, Mammalian taxonomy, no fixed or variable modifications, peptide charge 2+ and 3+, peptide mass tolerance ± 1.2 Da, fragment mass tolerance 0.4 Da and ESI‐TRAP instrument selection. In addition, one major peptide was identified by allowing variable modifications Gln‐>pyro‐Glu, oxidation2 and phosphorylation (S,T).

MS/MS and edman sequencer analysis

Sequencer analysis: Red fox AA protein from the final purification step (above) was concentrated with N₂, re‐solubilized in 9 M urea, diluted with 1 M ammonium bicarbonate and distilled water to 0.45 M urea, 100 mM ammonium bicarbonate. Different batches were digested at 37°C for 4 hr with a protease (trypsin or Glu‐C, Lys‐C or Arg‐C), at max ∼10 μL of a ∼0.15 mg/mL protease stock solution per AA batch of max ∼2–4 μg. Digests were then subjected to peptide separation by HPLC on Vydac C8 with a 0–60% acetonitrile gradient. Eluted material was submitted to sequencer analysis in an ABI Procise HT N‐terminal sequencer with subsequent PTH detection by HPLC in the ABI 140C Microgradient system, as described.29

LC‐electrospray MS/MS analysis: Alternatively, intact peptide digests, prepared as above, were separated with a C18 column (homemade; 25 cm, Silica Tip 360 μm OD, 75 mm ID, New Objective, Woburn, MA) at a 300 nL/min flow of 5–26% (A, 2% and B, 98% acetonitrile, both in 0.1% formic acid)) for 55 min followed by 95% B for 5 min in an Ultimate 3000 RSL Cnano LC‐MS/MS system and on‐line insertion into a Thermo Scientific Q Exactive Plus Orbitrap mass spectrometer (13 without use of the focusing device). Spectra were analyzed using the Mascot search engine v. 2.4 (Matrix Science Ltd., UK).

In vivo seeding experiments

The protein concentration in Red fox AA‐amyloid fibril extract was determined (Pierce) and diluted to 1 µg/µL 0.150 M NaCl. Female mice (6–12 weeks old) were used. Thirteen animals were injected on day 1 with 100 µg protein extract in the tail vein followed by an injection of 0.2 mL silver nitrate subcutaneously on the back. Another 14 mice were given silver nitrate but no fibrils. Administration of 1% silver nitrate, 0.1 mL was repeated to all mice on day 7, 14, and 21. Animals were sacrificed on day 23, half of the spleens were crushed and smeared between two glass slides, stained for amyloid with Congo red and examined in a polarization microscope.

Conflict of Interests

The authors report no conflicts of interest

Supporting information

Supporting Information Figure 1.

Supporting Information Table 1.