A longitudinal and cross-sectional study of plasma neurofilament light chain concentration in Charcot-Marie-Tooth disease

Advances in genetic technology and small molecule drug development have paved the way for clinical trials in Charcot-Marie-Tooth disease (CMT); however, the current FDA-approved clinical trial outcome measures are insensitive to detect a meaningful clinical response. There is, therefore, a need to identify sensitive outcome measures or clinically relevant biomarkers. The aim of this study was to further evaluate plasma neurofilament light chain (NFL) as a disease biomarker in CMT. Plasma NFL was measured using SIMOA technology in both a cross-sectional study of a US cohort of CMT patients and longitudinally over 6 years in a UK CMT cohort. In addition, plasma NFL was measured longitudinally in two mouse models of CMT2D. Plasma concentrations of NFL were increased in a US cohort of patients with CMT1B, CMT1X and CMT2A but not CMT2E compared with controls. In a separate UK cohort, over a 6-year interval, there was no significant change in plasma NFL concentration in CMT1A or HSN1, but a small but significant reduction in patients with CMT1X. Plasma NFL was increased in wild type compared to GARS C201R mice. There was no significant difference in plasma NFL in GARS P278KY compared to wild type mice. In patients with CMT1A, the small difference in cross-sectional NFL concentration vs healthy controls and the lack of change over time suggests that plasma NFL may lack sufficient sensitivity to detect a clinically meaningful treatment response in adulthood.

Developing successful treatments in preclinical models of CMT is only part of the journey in delivering therapies to patients. For a treatment to be adopted in routine clinical practice, it will need to show efficacy in clinical trials. CMT provides particular difficulties when it comes to designing clinical trials. 3 CMT is usually a lifelong disease, and even in the rapidly progressive forms, the rate of progression is slower than for other diseases such as amyotrophic lateral sclerosis. It is widely assumed that a successful treatment would be one that stops the progression of the disease, and therefore, clinical trials need to be designed with outcome measures that are able to detect a slowing in the rate of progression. A number of CMT-specific clinical outcome measures have been designed that have been validated or are undergoing validation, including the CMT neuropathy score, CMT Functional Outcome Measure, CMT Health Index and CMT Peds. [4][5][6][7][8][9] In addition, biomarkers of disease progression, such as nerve and muscle MRI are also being developed as outcome measures for clinical trials. [10][11][12] Neurofilaments are the most abundant cytoskeletal proteins in neurons of both the central and peripheral nervous systems. 13 It has been shown that plasma neurofilament light chain concentration (NFL) is increased in several neurological diseases, including CMT, where it also correlates with disease severity. [14][15][16][17][18][19][20][21] To be able to use a blood biomarker such as NFL in clinical trials, it is important to know how plasma concentrations vary over time. In this study, we replicate our previous crosssectional work in another cohort of CMT patients, investigate the change in plasma NFL over time in patients with CMT and in two mouse models of the disease.

| GARS mouse models
The generation and characterisation of the GARS P278KY and GAR-S C201R mouse models have been described previously. 22,23 All experimental procedures were conducted in accordance with animal care protocols approved by the Institutional Animal Care and Use Committee at The Jackson Laboratory. Blood samples were obtained from 5, 7, 9 and 11-week-old wild-type and GARS mice (n = 3-7 per age group) using a lancet puncture of the submandibular vein.

| Participants
Blood samples were collected prospectively between January 2017 and May 2019, with informed consent, from 27 out of 75 CMT patients who had previously donated blood for a previous study. 21 In addition, blood from 49 patients with CMT, identified and evaluated in the Inherited Neuropathy Consortium (INC) clinic in the Department of Neurology at Iowa, was also collected.
The disease severity, as measured using the Rasch modified CMT examination score, version 2 9 (hereto referred to as the weighted CMTES), was recorded at the same time that plasma was collected. The weighted CMTES is a validated outcome measure for assessing the severity of CMT. It is a composite score that includes the patient's symptoms and examination findings. 9 All patients underwent nerve conduction studies to confirm the presence of neuropathy; however, a weighted CMT neuropathy score (CMTNS) (which required neurophysiology at the same time as the clinical assessment) was only included if a nerve conduction study had been performed within 18 months of the blood sample.

| Blood sampling and sample collection and storage
All participants were evaluated in outpatient clinics, and blood samples were taken and processed within 1 hour. Blood was collected into EDTA-containing tubes and centrifuged at 20 C at 3500 rpm for 10 minutes. Plasma was then aliquoted and stored at À80 C.

| Simoa plasma NFL measurements
Plasma NFL concentration was measured using two highly correlated methods, employing the same antibodies: the in-house Simoa NFL assay that has been described in detail previously, 16

| Statistical analysis
Statistical analysis was performed using SPSS version 27.00 (IBM, New York, USA) and GraphPad Prism 9.0 (GraphPad Inc., California, USA). Correlations were assessed using Spearman's correlation coefficient. Two-tailed paired t-tests were used to compare differences in plasma NFL concentration in patients with CMT at baseline and after 6 years. One-way ANOVA with post hoc Dunnett's two-tailed t-test was used to compare differences in age and plasma NFL between CMT subtypes and controls in the Iowa cohort.

| RESULTS
There has been recent interest in the potential use of plasma NFL as a biomarker of disease progression in CMT for use in clinical trials. We, therefore, sought to examine plasma NFL concentration in a crosssectional cohort of CMT patients and longitudinally in a further cohort of patients with CMT and in two established mouse models of the disease.

| Plasma NFL concentration is increased in patients with CMT1B, CMT1X and CMT2A, but not CMT2E, compared with controls
We have previously demonstrated an increase in plasma NFL concentration in UK patients with CMT1A, CMT1X and HSN1. 21 We, therefore, sought to see if we could replicate this finding in an independent cohort of patients with CMT from the United States of America. The cohort of patients from Iowa comprised 18 patients with CMT1B, 18 with CMT1X, 4 with CMT2A and 9 with CMT2E and 25 controls (Table 1). There was no significant difference in the age of the patients with each type of CMT and controls (One-way ANOVA, P = .931) or the sex ratio (Chi-square, P = .53). Plasma NFL concentration was significantly increased in patients with CMT1B (ANOVA P < .0001, Dunnett's two-tailed t-test, P < .0001), CMT1X (P = .001) and CMT2A (P = .048) compared with controls but not in patients with CMT2E (P = .939) ( Figure 1A and Table 1). In contrast to our previous study in a UK cohort, there was no correlation between plasma NFL and the weighted CMTES and CMTNS for any of the CMT subtypes included in the study ( Figure 1B and Table 1). There was a significant correlation between plasma NFL in patients with CMT1B and the ulnar nerve Conduction Velocity (CV) (Spearman Rho = 0.876, P < .0001) and Ulnar Compound Muscle Action Potential (Rho = 0.682, P = .015) but not for patients with CMT1X or CMT2A (Table 1). There was no correlation between age of onset and plasma NFL in the CMT1B cohort (Pearson correlation coefficient, r = 0.44, P = .11).

| Plasma NFL changes with time in two mouse models of CMT
In order for plasma NFL to be of use as a biomarker of disease progression in CMT, it is necessary to know if the concentration changes with time. We have previously shown that in a mouse model of CMTX, 24 the concentration of plasma NFL rises rapidly between 2 and 3 months before falling by a third at 1 year. We, therefore, measured plasma NFL at 5, 7, 9 and 11 weeks in two mouse models of CMT2D ( Figure 2). The GARS C201R mouse is a milder model with normal life expectancy in contrast to the GARS P278KY mouse, which has a background-dependent reduced life expectancy of less than 6 months. 22 Plasma NFL was increased in wild type compared with GARS C201R mice at 5, 7 and 11 weeks, although the difference only reached statistical significance at 11 weeks (Mann-Whitney U-test, P = .01, Figure 2A). Plasma NFL concentration was also increased in wild type compared with GARS C201R mice at 11 weeks in an unrelated colony in the United Kingdom (see Figure S1). Plasma NFL was increased in the GARS P278KY mouse compared with wild type at 5 weeks, although this did not reach significance. There was no difference in plasma NFL at 7, 9 and 11 weeks consistent with the early axon loss in these mice, followed by very slow progress after 6-8 weeks of age ( Figure 2B). 22 3.3 | Plasma NFL is stable in CMT1A and HSN but not CMT1X over a 6-year period  Table 2). An analysis of follow up plasma NFL for the three major CMT subtypes revealed no significant change over 6 years in CMT1A (mean change = À2.44 pg/mL, SD = 11.5, P = .52, Figure 4A) and HSN1 (mean change = À0.69 pg/mL, SD = 1.18, P = .21, Figure 4C) but a significant reduction in CMT1X (mean change = À3.28 pg/mL, SD = 2.13, P = .01, Figure 4B)  An alternative explanation may be due to alteration of the NFL epitope recognised by the antibody used in the Simoa analysis as a result of neurofilament aggregations induced by the point mutation. 27 It is often assumed that the rate of axonal degeneration in genetic peripheral neuropathy such as CMT is constant. This is an important assumption to test, because if the rates of axonal degeneration  change with time, it may affect the timing of, or the ability to detect a significant alteration in NFL concentration in a clinical trial. To explore this further, we examined plasma NFL in two different mouse models of CMT2D, which are known to show progressive neurodegeneration over the examination period. 28 In the more severe GARS P278KY mouse, plasma NFL was highest at 5 weeks before falling to normal levels suggesting an early window of opportunity for treatment.
Plasma NFL for the GARS C201R mouse was similar to the baseline wild type concentration; however, the wild type mice showed significantly increased plasma NFL at multiple later time points. The cause for this difference is unknown but raises concern about the suitability of plasma NFL as a biomarker of axonal degeneration for trials in mouse models of this subtype of the disease. We originally speculated that the elevated plasma NFL was due to haemolysis of samples, which can result in spuriously elevated NFL concentrations (NFL is expressed in red blood cells 29 ); however, the replication of this result in a separate colony at a different time point would argue against this, although it remains a possibility.
In this study, we were also able to collect paired blood samples on 27 patients from our original CMT cohort after a 6-year interval. 21 Our analysis shows no statistically significant change in plasma NFL