Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS

Olivia M. Rifai; James Longden; Judi O'Shaughnessy; Michael D.E. Sewell; Judith Pate; Karina McDade; Michael J.D. Daniels; Sharon Abrahams; Siddharthan Chandran; Barry W. McColl; Christopher R. Sibley; Jenna M. Gregory

doi:10.1002/path.6008

Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS

Olivia M. Rifai^* (Corresponding Author), James Longden, Judi O'Shaughnessy, Michael D.E. Sewell, Judith Pate, Karina McDade, Michael J.D. Daniels, Sharon Abrahams, Siddharthan Chandran, Barry W. McColl, Christopher R. Sibley, Jenna M. Gregory^* (Corresponding Author)

^*Corresponding author for this work

University of Edinburgh

Research output: Contribution to journal › Article › peer-review

8 Citations (Scopus)

Abstract

Clinical heterogeneity observed across patients with amyotrophic lateral sclerosis (ALS) is a known complicating factor in identifying potential therapeutics, even within cohorts with the same mutation, such as C9orf72 hexanucleotide repeat expansions (HRE). Thus, further understanding of pathways underlying this heterogeneity is essential for appropriate ALS trial stratification and the meaningful assessment of clinical outcomes. It has been shown that both inflammation and protein misfolding can influence ALS pathogenesis, such as the manifestation or severity of motor or cognitive symptoms. However, there has yet to be a systematic and quantitative assessment of immunohistochemical markers to interrogate the potential relevance of these pathways in an unbiased manner. To investigate this, we extensively characterised features of commonly used glial activation and protein misfolding stains in thousands of images of post-mortem tissue from a heterogeneous cohort of deeply clinically profiled patients with a C9orf72 HRE. Using a random forest model to assess these features, we show that microglial staining features are the most accurate classifiers of disease status in our panel and that clinicopathological relationships exist between microglial activation status, TDP-43 pathology, and language dysfunction. Furthermore, we detected spatially resolved changes in FUS staining, suggesting that liquid-liquid phase shift of this aggregation-prone RNA-binding protein may be important in ALS caused by a C9orf72 HRE. Interestingly, no one feature alone significantly impacted the predictiveness of the model, indicating that the collective examination of all features, or a combination of several features, is what allows the model to be predictive. Our findings provide further support to the hypothesis of dysfunctional immune regulation and proteostasis in the pathogenesis of C9-ALS and provide a framework for digital analysis of commonly used neuropathological stains as a tool to enrich our understanding of clinicopathological relationships between and within cohorts.

Original language	English
Pages (from-to)	366-381
Number of pages	16
Journal	The Journal of pathology
Volume	258
Issue number	4
Early online date	20 Oct 2022
DOIs	https://doi.org/10.1002/path.6008
Publication status	Published - Dec 2022

Bibliographical note

Acknowledgments
We gratefully acknowledge Professor Tom Gillingwater for his helpful comments and support.
This work would not have been possible without the resources of the Edinburgh Brain Bank, and the people with ALS and their families who have generously donated tissue. This research was funded in part by a studentship from the Wellcome Trust (108890/Z/15/Z) to OMR and MDES, a Pathological Society and Jean Shanks foundation grant (217CHA R46564) to JMG and JO, and a Sir Henry Dale fellowship jointly funded by the Wellcome Trust and the Royal Society
(215454/Z/19/Z) to CRS.

Data Availability Statement

The datasets supporting the conclusions of this article are included in this published article and its supplementary material, or available in the figshare repository (raw images, digital pathology features, random forest data),
https://figshare.com/projects/Random_forest_modelling_and_neuropathological_features_of_a_
large_cohort_of_C9orf72-ALS_all_raw_data_/128222 [69].
Link to raw images: https://figshare.com/articles/figure/C9-ALS_and_control_images/17145896
Link to digital pathology features: https://doi.org/10.6084/m9.figshare.17145902.v1
Link to random forest results:
https://figshare.com/articles/dataset/Random_forest_results/17145905
The SD numbers of cases from the Edinburgh Brain Bank included in the study are available upon request.

Keywords

Amyotrophic lateral sclerosis
frontotemporal dementia
C9orf72
neuroinflammation
microglia
post-mortem tissue
TDP-43
FUS
machine learning
digital pathology

Access to Document

10.1002/path.6008Licence: CC BY

Rifai_etal_JPath_RandomForestModelling_VOR
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. https://creativecommons.org/licenses/by/4.0/
Final published version, 24.4 MBLicence: CC BY

https://onlinelibrary.wiley.com/doi/abs/10.1002/path.6008Licence: CC BY

Cite this

Rifai, O. M., Longden, J., O'Shaughnessy, J., Sewell, M. D. E., Pate, J., McDade, K., Daniels, M. J. D., Abrahams, S., Chandran, S., McColl, B. W., Sibley, C. R., & Gregory, J. M. (2022). Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS. The Journal of pathology, 258(4), 366-381. https://doi.org/10.1002/path.6008

Rifai, OM, Longden, J, O'Shaughnessy, J, Sewell, MDE, Pate, J, McDade, K, Daniels, MJD, Abrahams, S, Chandran, S, McColl, BW, Sibley, CR & Gregory, JM 2022, 'Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS', The Journal of pathology, vol. 258, no. 4, pp. 366-381. https://doi.org/10.1002/path.6008

@article{50e184098db04a78b3f0449c73eec28a,

title = "Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS",

abstract = " Clinical heterogeneity observed across patients with amyotrophic lateral sclerosis (ALS) is a known complicating factor in identifying potential therapeutics, even within cohorts with the same mutation, such as C9orf72 hexanucleotide repeat expansions (HRE). Thus, further understanding of pathways underlying this heterogeneity is essential for appropriate ALS trial stratification and the meaningful assessment of clinical outcomes. It has been shown that both inflammation and protein misfolding can influence ALS pathogenesis, such as the manifestation or severity of motor or cognitive symptoms. However, there has yet to be a systematic and quantitative assessment of immunohistochemical markers to interrogate the potential relevance of these pathways in an unbiased manner. To investigate this, we extensively characterised features of commonly used glial activation and protein misfolding stains in thousands of images of post-mortem tissue from a heterogeneous cohort of deeply clinically profiled patients with a C9orf72 HRE. Using a random forest model to assess these features, we show that microglial staining features are the most accurate classifiers of disease status in our panel and that clinicopathological relationships exist between microglial activation status, TDP-43 pathology, and language dysfunction. Furthermore, we detected spatially resolved changes in FUS staining, suggesting that liquid-liquid phase shift of this aggregation-prone RNA-binding protein may be important in ALS caused by a C9orf72 HRE. Interestingly, no one feature alone significantly impacted the predictiveness of the model, indicating that the collective examination of all features, or a combination of several features, is what allows the model to be predictive. Our findings provide further support to the hypothesis of dysfunctional immune regulation and proteostasis in the pathogenesis of C9-ALS and provide a framework for digital analysis of commonly used neuropathological stains as a tool to enrich our understanding of clinicopathological relationships between and within cohorts.",

keywords = "Amyotrophic lateral sclerosis, frontotemporal dementia, C9orf72, neuroinflammation, microglia, post-mortem tissue, TDP-43, FUS, machine learning, digital pathology",

author = "Rifai, {Olivia M.} and James Longden and Judi O'Shaughnessy and Sewell, {Michael D.E.} and Judith Pate and Karina McDade and Daniels, {Michael J.D.} and Sharon Abrahams and Siddharthan Chandran and McColl, {Barry W.} and Sibley, {Christopher R.} and Gregory, {Jenna M.}",

note = "Acknowledgments We gratefully acknowledge Professor Tom Gillingwater for his helpful comments and support. This work would not have been possible without the resources of the Edinburgh Brain Bank, and the people with ALS and their families who have generously donated tissue. This research was funded in part by a studentship from the Wellcome Trust (108890/Z/15/Z) to OMR and MDES, a Pathological Society and Jean Shanks foundation grant (217CHA R46564) to JMG and JO, and a Sir Henry Dale fellowship jointly funded by the Wellcome Trust and the Royal Society (215454/Z/19/Z) to CRS.",

year = "2022",

month = dec,

doi = "10.1002/path.6008",

language = "English",

volume = "258",

pages = "366--381",

journal = "The Journal of pathology",

issn = "0022-3417",

publisher = "Wiley",

number = "4",

}

TY - JOUR

T1 - Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS

AU - Rifai, Olivia M.

AU - Longden, James

AU - O'Shaughnessy, Judi

AU - Sewell, Michael D.E.

AU - Pate, Judith

AU - McDade, Karina

AU - Daniels, Michael J.D.

AU - Abrahams, Sharon

AU - Chandran, Siddharthan

AU - McColl, Barry W.

AU - Sibley, Christopher R.

AU - Gregory, Jenna M.

N1 - Acknowledgments We gratefully acknowledge Professor Tom Gillingwater for his helpful comments and support. This work would not have been possible without the resources of the Edinburgh Brain Bank, and the people with ALS and their families who have generously donated tissue. This research was funded in part by a studentship from the Wellcome Trust (108890/Z/15/Z) to OMR and MDES, a Pathological Society and Jean Shanks foundation grant (217CHA R46564) to JMG and JO, and a Sir Henry Dale fellowship jointly funded by the Wellcome Trust and the Royal Society (215454/Z/19/Z) to CRS.

PY - 2022/12

Y1 - 2022/12

N2 - Clinical heterogeneity observed across patients with amyotrophic lateral sclerosis (ALS) is a known complicating factor in identifying potential therapeutics, even within cohorts with the same mutation, such as C9orf72 hexanucleotide repeat expansions (HRE). Thus, further understanding of pathways underlying this heterogeneity is essential for appropriate ALS trial stratification and the meaningful assessment of clinical outcomes. It has been shown that both inflammation and protein misfolding can influence ALS pathogenesis, such as the manifestation or severity of motor or cognitive symptoms. However, there has yet to be a systematic and quantitative assessment of immunohistochemical markers to interrogate the potential relevance of these pathways in an unbiased manner. To investigate this, we extensively characterised features of commonly used glial activation and protein misfolding stains in thousands of images of post-mortem tissue from a heterogeneous cohort of deeply clinically profiled patients with a C9orf72 HRE. Using a random forest model to assess these features, we show that microglial staining features are the most accurate classifiers of disease status in our panel and that clinicopathological relationships exist between microglial activation status, TDP-43 pathology, and language dysfunction. Furthermore, we detected spatially resolved changes in FUS staining, suggesting that liquid-liquid phase shift of this aggregation-prone RNA-binding protein may be important in ALS caused by a C9orf72 HRE. Interestingly, no one feature alone significantly impacted the predictiveness of the model, indicating that the collective examination of all features, or a combination of several features, is what allows the model to be predictive. Our findings provide further support to the hypothesis of dysfunctional immune regulation and proteostasis in the pathogenesis of C9-ALS and provide a framework for digital analysis of commonly used neuropathological stains as a tool to enrich our understanding of clinicopathological relationships between and within cohorts.

AB - Clinical heterogeneity observed across patients with amyotrophic lateral sclerosis (ALS) is a known complicating factor in identifying potential therapeutics, even within cohorts with the same mutation, such as C9orf72 hexanucleotide repeat expansions (HRE). Thus, further understanding of pathways underlying this heterogeneity is essential for appropriate ALS trial stratification and the meaningful assessment of clinical outcomes. It has been shown that both inflammation and protein misfolding can influence ALS pathogenesis, such as the manifestation or severity of motor or cognitive symptoms. However, there has yet to be a systematic and quantitative assessment of immunohistochemical markers to interrogate the potential relevance of these pathways in an unbiased manner. To investigate this, we extensively characterised features of commonly used glial activation and protein misfolding stains in thousands of images of post-mortem tissue from a heterogeneous cohort of deeply clinically profiled patients with a C9orf72 HRE. Using a random forest model to assess these features, we show that microglial staining features are the most accurate classifiers of disease status in our panel and that clinicopathological relationships exist between microglial activation status, TDP-43 pathology, and language dysfunction. Furthermore, we detected spatially resolved changes in FUS staining, suggesting that liquid-liquid phase shift of this aggregation-prone RNA-binding protein may be important in ALS caused by a C9orf72 HRE. Interestingly, no one feature alone significantly impacted the predictiveness of the model, indicating that the collective examination of all features, or a combination of several features, is what allows the model to be predictive. Our findings provide further support to the hypothesis of dysfunctional immune regulation and proteostasis in the pathogenesis of C9-ALS and provide a framework for digital analysis of commonly used neuropathological stains as a tool to enrich our understanding of clinicopathological relationships between and within cohorts.

KW - Amyotrophic lateral sclerosis

KW - frontotemporal dementia

KW - C9orf72

KW - neuroinflammation

KW - microglia

KW - post-mortem tissue

KW - TDP-43

KW - FUS

KW - machine learning

KW - digital pathology

U2 - 10.1002/path.6008

DO - 10.1002/path.6008

M3 - Article

SN - 0022-3417

VL - 258

SP - 366

EP - 381

JO - The Journal of pathology

JF - The Journal of pathology

IS - 4

ER -

Random forest modelling demonstrates microglial and protein misfolding features to be key phenotypic markers in C9orf72-ALS

Abstract

Bibliographical note

Data Availability Statement

Keywords

Access to Document

Fingerprint

Cite this