Skin TDP-43 pathology as a candidate biomarker for predicting amyotrophic lateral sclerosis decades prior to motor symptom onset

Abstract

The recognition that disease-associated proteinopathies can manifest in peripheral organs outside the central nervous system preceding the onset of neurological symptoms, has transformed our understanding of Parkinson's disease, in wide terms of pathogenesis, detection and diagnosis. For amyotrophic lateral sclerosis, non-motor symptoms, and non-central nervous system pathologies are gaining increased recognition but remain incompletely understood. Here, using a TDP-43 RNA aptamer and a Stathmin-2 cryptic exon transcript BaseScope™ ISH probe, we identify widespread peripheral organ TDP-43 pathology prior to motor symptom onset in a discovery cohort of ante-mortem tissues from people who went on to develop ALS. Peripheral organs exhibiting both TDP-43 toxic gain- and loss-of function include muscle, lymph node, gallbladder, colon and with notably high incidence, skin. Given the accessibility of skin as a readily biopsiable tissue, representing a promising substrate for the detection of disease-associated proteinopathies and the development of minimally invasive biomarkers, we established an extended cohort of ante-mortem skin samples for TDP-43 pathology validation and further investigation. In skin biopsies taken during life from 17 individuals who went on to develop ALS we identify TDP-43 pathology from all 17 individuals in a wide distribution of anatomical sites, up to 26.5 years before ALS diagnosis - a presymptomatic period comparable to that observed for skin α-synucleinopathy in Parkinson's disease. TDP-43 pathology was most abundant in skin biopsies from the back and shoulder, with sweat and sebaceous glands showing the highest involvement. TDP-43 pathology was also associated with structural changes. As skin α-synucleinopathy has been established as a biomarker for both the detection of Parkinson's disease and the differentiation of Parkinson's disease from multiple system atrophy, we propose that skin TDP-43 likewise holds diagnostic and discrimination potential for diseases characterised by TDP-43 proteinopathy.

Keywords: TDP-43; biomarker; early detection; non-CNS; non-motor; presymptomatic; skin; sweat glands.

Figure 1.. Improved sensitivity of TDP-43^APT facilitates detection of TDP-43 pathology in discovery cohort of ALS pre-symptomatic peripheral tissues.

A. Schematic demonstrating organs with evidence of TDP-43 pathology, detected (left), and those with no evidence of TDP-43 pathology (right). Cell types affected by TDP-43 pathology are also detailed in the inset box in the left panel. B. Schematic demonstrating range of time from ante-mortem sampling to symptom onset across organs that show evidence of TDP-43 pathology.

Figure 2.. TDP-43^APT consistently detects TDP-43 pathology in skin biopsies with improved sensitivity compared to antibody approaches in pre-symptomatic ALS

A. Photomicrographs taken at 40x magnification stained with TDP-43^APT with DAB chromogen and counterstained with haematoxylin. Images show examples of a single case that had no evidence of TDP-43^APT pathology (left) and images from two other cases that showed evidence of TDP-43^APT pathology within the dermis (white arrowheads) of individuals who later went on to develop ALS. Scale bar = 20 μm. B. Images of skin biopsies from one individual who had a C9orf72 mutation. The image on the left, taken the same year as symptom onset, shows evidence of TDP-43^APT pathology (white arrowheads) in the nuclei and cytoplasm of sebocytes of the sebaceous gland within the dermis (–3). The image shows the same site biopsied 36 months prior to diagnosis, with no evidence of TDP-43^APT pathology (4). Scale bar = 40 μm C. Image taken from area highlighted in (B) showing the presence of STMN-2 cryptic exons (indicated by red arrowheads) in the peripheral nerve bundle adjacent to the affected sebaceous gland. Scale bar = 50 μm.

Figure 3.. Muscle biopsy reveals TDP-43 pathology in myocytes and surrounding neurovascular bundles in pre-symptomatic ALS.

A. Photomicrographs taken at 40x magnification stained with pTDP-43 antibody (left image) and TDP-43^APT (right image) with DAB chromogen and counterstained with haematoxylin. Scale bar = 50 μm. Images demonstrate lack of immunoreactivity for pTDP-43 antibody, but strong immunoreactivity for TDP-43^APT staining. B. Photomicrographs taken at 20x magnification stained with TDP-43^APT (left image) with 4x optical zoom in right image. These images are taken from the second muscle biopsy indicating that TDP-43^APT pathology can be seen in both individuals sampled within this cohort. White arrowheads indicate TDP-43^APT pathology; scale bar = 50 μm. C. Photomicrographs taken at 20x magnification stained with TDP-43^APT (left image) with optical zoom in right image. Scale bar = 75 μm. The image demonstrates TDP-43^APT pathology within the neurovascular bundle in between fascicles. White arrowheads indicate TDP-43^APT pathology within the peripheral nerve, and blue arrowheads indicate pathology within the endothelial cells of the adjacent blood vessels. D. The same region imaged in (C), except stained with in situ hybridisation probes directed at STMN-2 cryptic exon, which is only seen in the context of TDP-43 loss-of-function. Here TDP-43 loss-of-function can be visualised within the peripheral nerve bundle (red arrowheads indicate individual mRNA containing the STMN-2 cryptic exon. Scale bar = 50 μm.

Figure 4.. Blood vessel involvement by TDP-43 pathology in pre-symptomatic ALS.

Photomicrographs taken at 20x magnification stained with TDP-43^APT with DAB chromogen and counterstained with haematoxylin. Images demonstrate examples of neurovascular bundles with and without TDP-43^APT pathology from each of the affected organ systems. Scale bar = 50 μm. Red arrowheads indicate arteries, blue arrowheads indicate veins, orange arrowheads indicate peripheral nerves. White arrowheads indicate TDP-43^APT pathology.

Figure 5.. Aggregate- and vacuole-laden macrophages observed across non-CNS tissues in pre-symptomatic ALS

Photomicrographs taken at 40x magnification stained with TDP-43^APT with DAB chromogen and counterstained with haematoxylin. Scale bar = 50 μm in the left panel and 20 μm in the right panel. Black box in left panel represents area of optical zoom in right panel. Amongst non-CNS tissues we observed the presence of abnormal macrophage morphologies (white arrowhead) including aggregate- and vacuole-laden macrophages, in gastrointestinal tract, skin, and lymph node, consistent with histological evidence of lysosomal dysfunction.

Figure 6.. Widespread presymptomatic skin TDP-43 pathology revealed from extended validation cohort of skin biopsies up to 26.5 years prior to motor symptom onset.

A. Anatomical illustrations of the wide-range of anatomical sites from which TDP-43 pathology was detected from skin. B. Timeline representing years prior to ALS diagnosis that all 25 TDP-43 pathology positive skin biopsies had been taken, ranging from 0.58 years to 26.57 years, with anatomical sites of biopsies.

Figure 7.. Sweat and sebaceous glands show highest TDP-43 pathology.

TDP-43 pathology burden across peripheral neurovascular bundles and adnexal structural and its anatomical and temporal variation in sweat glands is illustrated here in the extended validation cohort. A. Histological images demonstrate the predominant location of TDP-43^APT pathology in this extended presymptomatic cohort. Scale bar = 50 μm. B. Proportion of dermal structures positive for TDP-43 pathology. Each bar represents the total number of biopsies in which sweat glands, sebaceous glands, peripheral nerves, and blood vessels were present. Bars show the proportion of these structures with detectable TDP-43 pathology (blue) versus those without pathology (pink). Sweat glands were TDP-43 positive in 93% of cases, sebaceous glands in 58%, and both peripheral nerves and blood vessels in 86%. C. Histology images illustrating the methodology of TDP-43 pathology burden scoring in sweat glands by superpixels-based segmentation. This method groups pixel similarity between different cellular populations and provides information pertaining to DAB intensity, defined as a ‘H-score’. Scale bar = 50 μm. D. Anatomical illustration of sites where skin biopsies included sweat glands, with representative histological images. These regions included the scalp, cheek, shoulder, forearm, back, arm, chest, neck and lip. E. Violin plots for TDP-43 pathology burden across anatomical regions for which skin biopsies contain sweat glands. TDP-43 pathology burden was greatest in the sweat glands of the shoulder and back and lowest in that of the chest Anatomical regions are colour-coded in correspondence with Figure 7D., previously. F. Violin plots for TDP-43 pathology burden for sun exposed (forearm, lip, neck, scalp and cheek), and non-sun exposed (arm, shoulder and back), anatomical sites with representative p-value (p=0.0037; Mann Whitney-U test) showing that pathology burden was significantly greater in non-sun exposed skin regions. G. TDP-43 pathology burden (H-score) of sweat glands from a range of anatomical sites was plotted against time to ALS diagnosis. Heterogeneous trajectories were observed, with some regions (e.g., back, shoulder) consistently showing high burden, while others (e.g., chest, lip, arm) varied between patients and proximity to symptom onset.

Figure 8.. Skin collagen density shows a very strong linear relationship with TDP-43 pathology.

Scatterplot illustrating the strong linear relationship between skin collagen density index (CD-index) and TDP-43 pathology burden amongst skin biopsies (n=23 biopsies from 15 individuals) taken from the validation cohort. Skin collagen density explains 93% of the variation in TDP-43 pathology burden (p<0.001).