Growth factors play numerous functions during embryogenesis and in different pathological conditions. PDGF consists of a family of five dimeric ligands (PDGF-AA, -AB, -BB, -CC, and -DD) interacting with two different receptors subtypes, as PDGF-alpha and beta [
28,
29]. In MS, it has been proposed that PDGF may exert beneficial effects through different mechanisms, although the main pathophysiological processes are not completely identified.
In this study, we show that in a group of MS patients classified at the time of diagnosis as CIS and early RR-MS, higher CSF PDGF concentrations were associated with a beneficial effect on RFS, without affecting other clinical and radiological indexes of disease activity. In particular, as the number of patients with MRI progression in the observational period did not differ in the three groups, PDGF seems not to influence the aggressiveness of the disease. In addition, clinical disease activity is not influenced by PDGF levels as the total number of relapses during the observation time did not differ between the three PDGF groups. The finding that in our patients an anti-inflammatory milieu is associated to higher CSF PDGF levels is in line with the role of this neurotrophin in the recovery phases after acute inflammation. Indeed, the lack of negative correlation with pro-inflammatory cytokines could suggest that PDGF is not associated to a direct anti-inflammatory action and explain why disease activity is not influenced by PDGF levels. In particular, the beneficial effect on RFS is independent of first-line treatment and of indirect markers of disease activity such as escalation to second-line treatment, making unlikely the possibility that different disease course may explain our results. Indeed, our findings are in line with the hypothesis that PDGF expression during the recovery phases could promote clinical compensation of brain damage. Experimental evidence demonstrated that PDGF could enhance LTP induction, modulating the expression of genes involved in transcription-dependent synaptic plasticity [
30]. The demonstration that LTP could be crucial for clinical recovery after acute brain lesion first came from preclinical studies [
31]. It has been shown that in patients with RR-MS, higher CSF PDGF concentrations were associated to enhanced LTP-like plasticity [
15,
16,
32] and better recovery after a relapse [
16]. In addition, in patients presenting with Gd+ lesions on MRI, higher CSF PDGF levels were more frequently found in clinically silent patients compared with relapsing patients [
15]. Overall, it is likely that LTP-like plasticity enhancement mediates the clinical beneficial effect of PDGF.
Members of the PDGF family have been associated to neuroprotective effects in different conditions, including excitotoxicity [
33], energy deprivation and oxidative injury [
34], toxicity associated to human immunodeficiency virus protein [
30], and after brain ischemia [
35]. To exclude the possibility that beneficial role of PDGF could be mediated by neuroprotective effects, we explored the influence of PDGF levels on measures of disease progression. As no significant correlation emerged with PI, MSSS, and EDSS progression, it is possible that PDGF concentrations at the time of diagnosis may not influence disease severity. We also explored whether PDGF concentrations influence the levels of some well-known biomarkers of neurodegeneration and neuronal damage such as amyloid-β, tau, and NFL proteins. CSF levels of amyloid-β and tau protein, a molecule involved in the stabilization of microtubules, were found altered in different neurodegenerative diseases [
36]. Notably, in MS patients, altered amyloid-β metabolism was reported during acute inflammation [
37]. Furthermore, NFL proteins are cytoskeleton components representing a marker of axonal damage in MS [
38]. In MS, axonal damage is not only due to neurodegeneration but also to axonal transections occurring in acute demyelinating lesions [
39]. Accordingly, increased CSF levels of NFL and tau proteins have been found in MS patients, and high CSF levels of axonal cytoskeletal proteins correlate with EDSS [
40]. Axonal damage can follow acute demyelination as suggested by the findings that elevated CSF levels of NFL and tau proteins can be observed throughout the disease course and not confined to the late stages [
41]. In our patients, no significant correlation emerged between PDGF and the CSF levels of amyloid-β, tau protein, and NFL protein. The lack of association between PDGF and the examined biomarkers suggests that the beneficial effect of PDGF may not be ascribed to a different degree of neuronal damage. It is worth noting that in our cohort of patients, CSF levels of biomarkers of neurodegeneration were assessed at the time of diagnosis; therefore, the lack of association could be due to the relatively short disease duration. As in the subgroup of patients with acute inflammation and radiological activity at the time of CSF withdrawal, no significant correlation was found between PDGF levels and tau protein and NFL protein; it seems unlikely that PDGF levels in the post-acute phases may influence the subsequent building up of axonal damage. Other limitations of the present study include the lack of prospective radiological measures of neurodegeneration (i.e., gray matter cortical atrophy); nevertheless, the lack of difference in disease severity between the three PDGF groups makes unlikely the possibility that different degree of neurodegeneration may explain the results.