Phenotype Prediction Using Cell-Free DNA Fragmentation as a Predictor of Clinical Severity in Neurofibromatosis Type 1 (NF1)
Cancer Genetics and Therapeutics
-
Primary Categories:
- Laboratory Genetics
-
Secondary Categories:
- Laboratory Genetics
Introduction:
Neurofibromatosis type 1 (NF1) is an autosomal dominant neurocutaneous condition characterized by skin pigmentation, skeletal anomalies and cancer, and is associated with pathogenic variants in the NF1 gene. While NF1 is fully penetrant, symptoms range from mild freckling to the development of malignant peripheral nerve sheath tumors (MPNSTs) , the leading cause of death in NF1, indicating variable expressivity. Most patients develop benign neurofibromas, which progress into aggressive MPNSTs in 8-13% of patients. The severity of a patient’s condition is assessed based on symptom presentation and cancer history; however, there are currently no methods to predict a patient’s risk of malignant transformation. To address this, we use liquid biopsies to obtain cell-free DNA (cfDNA), a novel avenue that permits more frequent testing using DNA fragments circulating in blood plasma. We apply “fragmentomics” analysis to identify participants at higher risk of malignant transformation for closer clinical follow-up.
Methods:
To develop strategies for stratifying participants across different levels of clinical severity, we sequenced plasma cfDNA from 54 NF1 participants and 26 healthy controls using 30-70X whole-genome sequencing (WGS). Each participant was assessed by six independent clinicians at the time of cfDNA collection. Severity scores (0 – 12) were assigned to each participant, and scores of 9+ were considered high-severity. To link cfDNA profiles to severity, we investigated cfDNA fragment size, fragment end motifs, fragment score, and nucleosome positioning as possible biomarkers of clinician-evaluated severity scores. To generate a tissue type-specific nucleosome positioning reference, we used 12 external human NF1–associated MPNST samples sequenced using ATAC-Seq.
Results:
We extracted nucleosomal signatures from plasma cfDNA, and a depth of coverage of 30X was required to differentiate NF1 carriers from healthy non-NF1 controls (p < 0.0001). This association was not significant when the data were downsampled to 1X coverage (p = 0.381). We did not detect substantial somatic copy number alterations in our cohortand ruled out the possibility of a generalized loosening in chromatin accessibility using the cfDNA fragment sizes. Thus, we determined that the observed disruption in nucleosome positioning is more likely to be an alteration in the chromatin profiles. cfDNA nucleosomal signatures were also able to differentiate between high- and low-severity participants (p = 0.001), as well as between the high- and moderate-severity participants (p = 0.007). It did not separate moderate- from low-severity participants (p = 0.986). Nucleosome positioning in cancer-negative participants was also dysregulated compared to healthy controls (p = 0.0002), suggesting underlying disease processes without full progression to cancer. To leverage nucleosome positioning inferred using cfDNA to identify early MPNSTs, we mapped cfDNA coverage to a reference for open chromatin regions known from ATAC-seq data from MPNST that exceeded 95% peak saturation. NF1 patients demonstrated less pronounced nucleosome protection of DNA compared to healthy non-NF1 participants, particularly those with an active MPNST at the time of blood collection (p < 0.0001). This indicates disruptions in nucleosome placement that could reflect changes in chromatin structure associated with MPNST progression.
Conclusion:
We were able to demonstrate that NF1 leads to a global difference in cfDNA fragmentation, and that nucleosome-dependent signatures have clinical utility as a diagnostic tool. cfDNA fragment profiles correlated with severity assessments and may fill an important gap in linking between harboring an NF1 variant and projected outcomes for each patient. Further work is needed to refine cfDNA-based severity scores and establish optimal thresholds to identify high-severity NF1 patients at risk of developing an MPNST as well as early MPNSTS. This approach may also serve to tailor the frequency of cfDNA surveillance and may have broader applications in the management of hereditary cancer syndromes.
Neurofibromatosis type 1 (NF1) is an autosomal dominant neurocutaneous condition characterized by skin pigmentation, skeletal anomalies and cancer, and is associated with pathogenic variants in the NF1 gene. While NF1 is fully penetrant, symptoms range from mild freckling to the development of malignant peripheral nerve sheath tumors (MPNSTs) , the leading cause of death in NF1, indicating variable expressivity. Most patients develop benign neurofibromas, which progress into aggressive MPNSTs in 8-13% of patients. The severity of a patient’s condition is assessed based on symptom presentation and cancer history; however, there are currently no methods to predict a patient’s risk of malignant transformation. To address this, we use liquid biopsies to obtain cell-free DNA (cfDNA), a novel avenue that permits more frequent testing using DNA fragments circulating in blood plasma. We apply “fragmentomics” analysis to identify participants at higher risk of malignant transformation for closer clinical follow-up.
Methods:
To develop strategies for stratifying participants across different levels of clinical severity, we sequenced plasma cfDNA from 54 NF1 participants and 26 healthy controls using 30-70X whole-genome sequencing (WGS). Each participant was assessed by six independent clinicians at the time of cfDNA collection. Severity scores (0 – 12) were assigned to each participant, and scores of 9+ were considered high-severity. To link cfDNA profiles to severity, we investigated cfDNA fragment size, fragment end motifs, fragment score, and nucleosome positioning as possible biomarkers of clinician-evaluated severity scores. To generate a tissue type-specific nucleosome positioning reference, we used 12 external human NF1–associated MPNST samples sequenced using ATAC-Seq.
Results:
We extracted nucleosomal signatures from plasma cfDNA, and a depth of coverage of 30X was required to differentiate NF1 carriers from healthy non-NF1 controls (p < 0.0001). This association was not significant when the data were downsampled to 1X coverage (p = 0.381). We did not detect substantial somatic copy number alterations in our cohortand ruled out the possibility of a generalized loosening in chromatin accessibility using the cfDNA fragment sizes. Thus, we determined that the observed disruption in nucleosome positioning is more likely to be an alteration in the chromatin profiles. cfDNA nucleosomal signatures were also able to differentiate between high- and low-severity participants (p = 0.001), as well as between the high- and moderate-severity participants (p = 0.007). It did not separate moderate- from low-severity participants (p = 0.986). Nucleosome positioning in cancer-negative participants was also dysregulated compared to healthy controls (p = 0.0002), suggesting underlying disease processes without full progression to cancer. To leverage nucleosome positioning inferred using cfDNA to identify early MPNSTs, we mapped cfDNA coverage to a reference for open chromatin regions known from ATAC-seq data from MPNST that exceeded 95% peak saturation. NF1 patients demonstrated less pronounced nucleosome protection of DNA compared to healthy non-NF1 participants, particularly those with an active MPNST at the time of blood collection (p < 0.0001). This indicates disruptions in nucleosome placement that could reflect changes in chromatin structure associated with MPNST progression.
Conclusion:
We were able to demonstrate that NF1 leads to a global difference in cfDNA fragmentation, and that nucleosome-dependent signatures have clinical utility as a diagnostic tool. cfDNA fragment profiles correlated with severity assessments and may fill an important gap in linking between harboring an NF1 variant and projected outcomes for each patient. Further work is needed to refine cfDNA-based severity scores and establish optimal thresholds to identify high-severity NF1 patients at risk of developing an MPNST as well as early MPNSTS. This approach may also serve to tailor the frequency of cfDNA surveillance and may have broader applications in the management of hereditary cancer syndromes.