Immunohistochemical staining of tissue biopsies is a standard technique used for typing amyloid. Orfila et al. used the immunofluorescence technique on abdominal fat tissue aspirates and detected five of nine samples with AA amyloid deposits [55] and Halloush et al. detected with immunoperoxidase technique three of five samples with AL and ATTR amyloid deposits [56]. Immunohistochemical staining of subcutaneous fat tissue, however, is fairly difficult due to nonspecific reactions [7]. Therefore, Westermark developed different immunochemical methods for typing amyloid in fat tissue. The initial method was a double immunodiffusion method that satisfactorily identified some patients with AA amyloidosis [4, 7]. Later his group developed an enzyme-linked immunosorbent assay (ELISA) that was useful for typing AA, AL, and ATTR amyloidosis in 14 of 15 cases [8]. Recently he described a method based on Western blot analysis combined with specific amyloid fibril antibodies that was successful in typing 32 of 35 patients with AA, AL, and ATTR amyloidosis [49] and all of 33 patients with ATTR amyloidosis [57]. Kaplan et al. also used immunochemical methods to type amyloid successfully in three of four patients with AL-kappa, in five of six patients with AL-lambda, and in one patient with AA amyloid [58].

Because of unsatisfactory typing of amyloid in fat aspirates using immunohistochemistry, Arbustini et al. developed immunoelectron microscopy for typing amyloid of patients with cardiac amyloidosis. This appeared to be a promising technique, because in all 15 patients the amyloid types involved were identified: eight samples with AL-lambda, two with AL-kappa, three with ATTR, two with AApoAI, and none with AA amyloid [42]. In a recent study of this group, immunoelectron microscopy was used for typing amyloid in abdominal fat tissue in 423 patients. The specific type of amyloidosis was correctly identified in 99 % of the cases. An abstract of this study has been presented at the XIVth International Symposium on Amyloidosis in Indianapolis in April 2014 [59, 60].

Our group in Groningen also developed and used immunochemical methods for measuring the concentrations in fat tissue of the four major amyloid proteins, i.e., amyloid A protein, TTR, light chain kappa, and light chain lambda. See Table 16.2 for the panel of cutoff values we use for fat aspiration samples in our center. An ELISA was used to measure the concentration of amyloid A protein in fat tissue in a group of 24 patients with AA amyloidosis and a group of 72 controls, including 25 patients with AL or ATTR amyloidosis and 25 patients with chronic inflammation. The upper limit of the 99 % CI of controls (11.6 ng/mg tissue) was chosen as cutoff level [10]. In a large collaborative international study, this method was validated and identified 129 of 154 patients with AA amyloidosis, whereas only 3 of 354 controls (87 AL, 30 ATTR, and 27 localized amyloidosis and 210 non-amyloidosis patients) were identified incorrectly [61]. Thus, in this study the sensitivity of detecting AA amyloidosis was 84 % (95 % CI 77–89 %), specificity was 99 % (95 % CI 98–100 %), PPV was 98 % (95 % CI 94–100 %), and NPV was 93 % (95 % CI 90–96 %). Men had lower amyloid A protein concentrations in fat than women (Fig. 16.1a). Patients with familial Mediterranean fever (FMF) had lower values than patients with arthritis or other inflammatory diseases (Fig. 16.1b). The low yield of fat tissue aspirates for the detection of AA amyloidosis in patients with FMF was already observed and discussed by Tishler et al. in 1988: they even did not detect amyloid in any of the 15 FMF patients studied [62]. However, Congo red-positive amyloid deposits were detected in 20 fat tissue samples (80 %) and the amyloid A protein concentration was elevated in 13 samples (about 50 %) of the 25 FMF patients we studied [61]. In a study of Egyptian patients with long-standing rheumatoid arthritis, quantification of the amyloid A protein concentration in fat tissue appeared to be a useful screening tool for the detection of AA amyloidosis [32]. Subcutaneous fat tissue is a major source of the acute-phase reactant serum amyloid A protein (SAA), the protein precursor of AA protein [63]. Extensive washing of fat tissue, however, was effective to prevent unwanted contamination and retention of SAA from blood in fat tissue extracts of controls. Although indeed a positive correlation between the SAA concentration in blood and amyloid A protein concentration in fat tissue was present in controls, the magnitude of the effect was very small, resulting in maximum amyloid A protein concentrations in fat tissue largely within the reference range of controls [61].

An ELISA is also used in our center to measure the concentration of TTR in fat tissue. The upper limit of the 99 % CI of controls is 4.36 ng/mg fat tissue and this value has been chosen as cutoff value. Quantification of free light chains by nephelometry is used to measure light chain kappa (95 % CI upper limit 24.4 ng/mg fat tissue), light chain lambda (95 % CI upper limit 10.6 ng/mg fat tissue), and the resulting κ/λ ratio (95 % CI reference range 0.49–5.86). See also Table 16.2 [64, 65].

Antibody-based detection of amyloid has the advantage of fast, inexpensive, and specific detection and amplification of minute amounts of a particular protein in an enormous background ocean of other proteins and tissue components. However, disadvantages are sometimes considerable: limited reactivity with commercial antibodies (false-negative), “contamination” by serum proteins (false-positive), the increasing number of other types of amyloidosis that are recognized, the issue of combined deposits (more than one type of amyloid protein present), and finally, experience and principles in the interpretation of results [66]. The consequences of mistyping amyloid may lead to a wrong choice of treatment and may be harmful for the patient. This is the main reason we wanted to develop in our center typing methods with high specificity (≥99 %) resulting in a high PPV. If a particular type is then successfully identified in a sample, the result should be virtually beyond doubt. The practical consequence of very high specificity on the other hand is relatively low sensitivity with corresponding low NPV. However, it is our opinion that it is easier to manage cases in which typing is not successful: these inconclusive or negative cases should be diagnosed as such and considered for evaluation by reference laboratories that have experience in amyloid typing using a wider antibody panel as well as the capacity to apply methods that are more sophisticated [66].

Typing of amyloid should be performed with confidence [67]: amyloid typing should result in certainty of both presence of a particular type of amyloid and absence of all other (relevant) types of amyloid. Chemical typing of the amino acid sequence of the amyloid protein involved has been the gold standard for a long time. Until recently, macropurification techniques were needed for isolation of these proteins from large amounts of tissue obtained at necropsy or after surgery, at least 10–30 g of amyloid-rich tissue [68, 69]. With some modifications, Westermark et al. were able to determine the amino acid sequence of the amyloid protein isolated from only 1 g of amyloid-rich fat tissue [70, 71] and 10 years later from an even lower amount of fat tissue [72].

New methodological approaches, including the search for appropriate extraction solvents and separation strategies, have been developed and tested in order to design efficient procedures for small scale isolation and purification of these amyloid proteins [73–75]. Kaplan et al. were among the first to investigate these micropurification techniques and they found that the strategy for small scale purification of amyloid proteins depends largely on the type of tissue sample (fresh or formalin fixed), sample size, amyloid content of the tissue, and its chemical nature [75]. Abdominal fat tissue aspirates were among the tissues they studied [58]. These techniques were further improved using different mass spectroscopic techniques in the following years [51, 76–79]. Vrana et al. introduced laser microdissection to concentrate the amyloid content in extracts from tissue samples of patients with AA, ATTR, AL-kappa, and AL-lambda amyloidosis [79]. Her group recently applied mass spectrometry-based proteomics to type amyloidosis in fat aspirates [51]. In the validation study, the presence of amyloid was detected with high sensitivity (88 %) and specificity (96 %). In 366 Congo red-positive cases, mass spectrometry analysis successfully identified an amyloid subtype in 330 cases, yielding a sensitivity of 90 %. The specific diagnoses were AL-lambda (58 %), AL-kappa (16 %), ATTR (12 %), AA (1 %), AGel (1 %), AIns (1 %), and ALys (1 %). A universal amyloid proteome signature (i.e., the universal presence of apoAIV, ApoE, and SAP) was identified with high sensitivity and specificity, similar to that of Congo red staining [51].

Lavatelli et al. developed a proteomics methodology utilizing two-dimensional (2D) polyacrylamide gel electrophoresis followed by matrix-assisted laser desorption/ionization mass spectrometry and peptide mass fingerprinting to directly characterize amyloid deposits in abdominal subcutaneous fat obtained by fine needle aspiration from seven patients with AL amyloidosis, two with ATTR amyloidosis, and seven controls. Striking differences in the 2D gel proteomes of adipose tissue were observed between controls and patients and between the two types of patients with distinct, additional spots present in the patient specimens that could be assigned as the amyloidogenic proteins in full-length and truncated forms [80]. The 2D-tandem MS technique was further developed (called Multidimensional Protein Identification Technology or shortly MudPIT) and was used to identify reliably 26 patients with the most common (AL, AA, and ATTR) types of amyloidosis who had grade 3+ or 4+ amyloid in fat tissue [81]. In another study of 30 patients with systemic amyloidosis, comprehensive protein profiles of human amyloid-affected adipose tissue from patients and its control (nonamyloid affected) counterpart were acquired. In particular, extracellular matrix (ECM), protein folding, lipid metabolism, and mitochondrial functions were among the most affected structural/functional pathways [82]. These proteomic techniques have great potential to diagnose and type amyloid, but also to clarify all kind of molecular mechanisms involved in amyloidosis [82–84].