Cyanine3 (cy3)

A 3-mm diameter skin biopsy was taken under subcutaneous anaesthesia on the ventroradial side of the proximal phalanx of the index finger innervated by the median nerve. The biopsy was fixed in fresh periodate-lysine-paraformaldehyde for 30 minutes. The tissue was then washed in phosphate buffer and stored for 2 to 3 days in sucrose in phosphate buffer. After embedding in optimal cutting temperature gel, the tissue was frozen and stored at −80°C. Staining was performed using a previously described free-floating method,^{46 (link)} using protein gene product 9.5 (PGP 9.5 Ultraclone, Isle of Wight, United Kingdom, 1:1000; Zytomed, Berlin, Germany 1:200) and myelin basic protein (Abcam, Cambridge, United Kingdom 1:500) as primary antibodies and Cy3 (Stratech, Ely, United Kingdom 1:1000) and Alexa Fluor 488 (Abcam, 1: 500) as secondary antibodies.
Intraepidermal nerve fibre density (IENFD) was quantified in 50-μm skin sections using an Axio LSM 700 microscope with an Observer Z1 imaging system (Zeiss, Cambridge, United Kingdom) by determining the amount of fibres per millimeter epidermis according to current guidelines.^{30 (link)} We also quantified dermal innervation by evaluating the number of Meissner corpuscles per millimeter epidermis, the percentage of PGP⁺ dermal nerve bundles containing MBP, and the mean nodal length as previously reported.^{46 (link)}

Sections (50 μm) were cut on a cryostat and stained using a previously described protocol (Schmid et al., 2014 (link)). Serial biopsies of the same patient were processed simultaneously to minimize variability. Sections were blocked with 5% fish gelatine before incubation with the primary antibodies for myelin basic protein (MBP, Abcam, 1:500) and protein gene product 9.5 (PGP, Ultraclone, 1:1000) overnight. Secondary antibodies were also incubated overnight (Alexa 488 1:1000, Abcam; Cy3, Stratech 1:500). The PGP antibody from Ultraclone was discontinued during our study and replaced with the PGP antibody from Zytomed (1:200) after successful benchmarking.
IENFD was established by counting three sections per participant using an Axio LSM 700 microscope (Zeiss) (Lauria et al., 2010 (link)). Sections were then photographed at 20× magnification (Observer Z1, Zeiss) and epidermal length was measured with ImageJ (NIH, USA). Average IENFD was expressed as fibres/mm epidermis. Intra-rater reliability to determine IENFD was examined by processing and counting n = 23 skin samples twice, 1 week apart. Intraclass correlation coefficient (ICC) revealed almost perfect intra-rater agreement for IENFD [ICC(2,1) 0.913 (95% confidence interval 0.791–0.963), P < 0.0001].