Diagnosis of Amyloidosis Subtype By Laser-Capture Microdissection (LCM) and Tandem Mass Spectrometry (MS/MS) Proteomic Analysis

Aim

Correct identification of the protein that is causing amyloidosis is crucial for clinical management. Current standard laboratory methods have limited ability to detect the full range of amyloid forming proteins. We assessed the diagnostic value of LCM-MS/MS, which combines specific sampling of amyloid deposits by LCM with protein identification by MS/MS.

Methods

Biopsy specimens were referred to the Princess Alexandra Hospital Amyloidosis Centre. For all specimens, 10µm sections of formalin-fixed paraffin embedded tissue were stained with Congo Red using a standard technique. LCM was performed using an Arcturus XT instrument with an infrared capture laser. Proteins were extracted with FFPE Protein Extraction Solution (Agilent Technologies), digested with trypsin and peptides were analysed by nano-liquid chromatography-coupled MS/MS using an Agilent Chip CUBE-QTOF. Database searching was performed using Spectrum Mill (Agilent) with the NCBInr human protein database.

Results

Biopsies were received on 136 patients: there was insufficient tissue in the block in 7, repeat LCM was required in15 cases and no amyloid forming protein was identified in 8. In 121/136 (89%) an amyloid forming protein was identified. Proteins identified included immunoglobulin light chain (localised amyloid n=25, systemic AL n=45), immunoglobulin heavy chain (AH n=6), transthyretin (senile amyloid n=25, hereditary ATTR n=6), serum amyloid A (AA n=7), fibrinogen alpha chain (AFib n=2), LECT2 (ALect2 n=2), TGFb (corneal lattice amyloid n=1) and semenogelin (seminal vesicle amyloid n=2). It was not infrequent, particularly in cases of localised amyloidosis (>80% of cases), for smaller amounts of other amyloid forming proteins to be present especially immunoglobulins, transthyretin and ApoA1 (Table 1). This suggests that these inherently amyloidogenic proteins are capable of integrating within the amyloid deposit. An amyloid proteomic signature as previously defined by the Mayo Clinic (at least two of SAP, ApoE and ApoA4; Haematologica 2014;99(7):1239) was present in 76% of cases and was more likely to be found if larger amounts of amyloid could be dissected (p=0.0001).

In terms of clinical impact, amyloid typing by immunohistochemical stains had been attempted in 87 cases and reported as diagnostic in 39. Five of these were subsequently revealed by proteomic analysis to be incorrect. Overall, the clinical diagnosis of amyloid subtype was altered by proteomic analysis in 24% of cases.

Conclusion.

LCMMS/MS identifies an amyloid forming protein in ~90% of clinical biopsy samples. Amyloid deposits often contain small amounts of other amyloid forming proteins which may reflect not just contamination but co-deposition of fibrils due to a shared beta pleated sheet conformation. Because of this, results need to be interpreted in the context of full clinical information to enable correct diagnosis of amyloid subtype.