This study highlights critical insights into the roles of sGFAP and sNFL in ATTRv. Our data documented increased sGFAP levels in both pre-symptomatic subjects and symptomatic patients compared to HCs, while elevated sNfL levels were found exclusively in symptomatic patients. This suggests that sGFAP elevation marks the disease from its preclinical stages, while sNfL increase corresponds with the presence of neurological manifestations. To the best of our knowledge, this study is the first to document significantly elevated sGFAP levels in both ATTRv patients and pre-symptomatic subjects.

The interpretation of GFAP elevation in the preclinical stage of the disease requires consideration of its origins within the nervous system. sGFAP, primarily recognized as a marker of astrocytic damage and activation within CNS, may also derive from the PNS and the ENS. However, the exact contribution of this production remains unclear. Investigating the different circulating isoforms of GFAP, a capability not yet available, could provide significant clarity on this aspect.

Existing evidence suggests that astrocytic activation in ATTRv begins as early as the pre-symptomatic stage [23], potentially driven by amyloid deposits in the CNS even before the onset of polyneuropathy. This activation follows the progressive stages of TTR-amyloidosis in the CNS, which increase in extent as the disease progresses [5]. Initially, TTR deposition is observed in the leptomeninges and subarachnoid meningeal vessels, with pronounced involvement in the brainstem and spinal cord. As the disease advances to the second stage, amyloid deposition extends to subpial cortical regions and becomes more frequent in perforating cortical vessels. In the final stage, the pathology spreads to subependymal regions and basal ganglia vessels near the ependymal lining. Notably, subpial TTR amyloid deposits are associated with astrocytosis, highlighting that astrocytic activation and damage are crucial features of ATTRv pathology and may significantly contribute to the observed elevation in serum GFAP concentrations.

Additionally, the elevation of sGFAP in pre-symptomatic subjects may also derive from the PNS and/or the ENS. GFAP expression in these systems is well-documented, particularly in Schwann cells and enteric glial cells [18]. GFAP is expressed in immature Schwann cells and subsequently downregulated in myelin-forming Schwann cells [24, 25], whereas non-myelinating Schwann cells retain GFAP expression and exhibit functional similarities to astrocytes. These non-myelinating Schwann cells could release GFAP into the bloodstream in case of dedifferentiation or mild (subclinical) nerve injury. Furthermore, glial cells in the ENS expressing GFAP may represent another potential source of sGFAP elevation. This sub-epithelial glia plays key trophic and neuromodulatory roles by supporting intestinal neurons and epithelial cells, regulating gut motility, and mediating neurotransmitter signaling. Given the high prevalence of gastrointestinal manifestations in ATTRv [26], it is plausible to hypothesize that early alterations in the ENS among TTR variants carriers could contribute to the observed sGFAP elevation.

The significance of the elevated sNfL levels in patients with ATTRv, which are absent in pre-symptomatic subjects, warrants specific consideration. sNfL is the most widely recognized biomarker of neuroaxonal damage, with its serum concentration reflecting contributions from both the CNS and the PNS [27, 28]. Notably, sNfL has been reliably shown to distinguish symptomatic patients with ATTRv polyneuropathy from pre-symptomatic subjects, a finding confirmed by our study [29]. Although sNfL is commonly utilized to monitor the clinical progression of patients with ATTRv polyneuropathy [30], definitive evidence attributing its entire concentration exclusively to the PNS is lacking [28]. Indeed, the pathological involvement of the CNS in ATTRv has been extensively documented and reported soon after the first description of the disease [31, 32]. Given the observed elevation of GFAP levels in our patient cohort, it is plausible to hypothesize that a portion of the sNfL may originate from the CNS, although the exact proportion remains undetermined. Future research should explore neurofilament isoforms with regional specificity to better localize and characterize the sources of sNfL elevation [28]. Interestingly, among these isoforms, peripherin seems a very promising biomarker of neuroaxonal damage involving the PNS and its serum concentrations have been already explored for this purpose [33].

Another important finding of our study is the higher sGFAP levels documented in female HCs compared to males. This aligns with previous research showing elevated plasma GFAP concentrations in cognitively unimpaired women compared to men [34]. Similar sex differences, characterized by higher peripheral blood GFAP levels in females, have also been described in Alzheimer's disease [34], Parkinson's disease [35], and in traumatic brain injury [36]. Sex-based variations in astrocyte number, differentiation, and function have been extensively documented [37–40]. Moreover, inflammation impacts the blood–brain barrier (BBB) differently between sexes. The differential expression of tight junction proteins and inflammatory markers between males and females underscores the importance of sex-specific mechanisms in neuroinflammation and BBB permeability [41]. These findings carry significant implications for the statistical analysis of scientific studies investigating sGFAP, as sex appears to be a significant covariate that must be considered for accurate interpretation of measurements.

The lack of significant differences in sGFAP levels between pre-symptomatic subjects and symptomatic patients might suggest that sGFAP serves more as a marker of disease presence rather than progression. In further support of this, no significant correlation was detected between sGFAP concentrations and years from PADO in the pre-symptomatic phase. However, several factors could contribute to this observation. One potential explanation is the sample size of our study, which might have limited the statistical power to detect subtle differences between these groups. Another possible contributing factor is the disparity in sex distribution between the pre-symptomatic group and the symptomatic patients, with a higher proportion of females in the former (51%) compared to the latter group (23%). Considering that sGFAP levels, as previously demonstrated, are generally higher in females, this sex imbalance may mask any potential differences attributable to disease progression. Future studies with larger and more balanced cohorts should explore these factors to better understand the dynamics of sGFAP levels across the stages of ATTRv.

A potential limitation of our study was the age disparity among the cohorts of HCs, presymptomatic subjects, and patients. Specifically, the patients were older than the subjects in the other cohorts. To address this limitation, we conducted all relevant analyses of sGFAP and sNfL levels while controlling for age as a covariate. Furthermore, we validated our findings by selecting a subgroup of older HCs who were age-matched with the ATTRv patients, thereby eliminating age-related differences.

Finally, the positive correlation between neurological impairment, as assessed by NIS, and sNfL and sGFAP concentrations supports their utility as biomarkers of disease severity. The association between sNfL levels and clinical disability has been previously demonstrated by our group using an alternative method for the assessment of sNfL [11]. Similarly, the relationship between sGFAP and clinical measures of disability has been described in peripheral neuropathy [42]. To the best of our knowledge, this study provides the first evidence of such a correlation in ATTRv patients.