In this cross-sectional study we demonstrated higher plasma NfL levels in SLE patients compared with age- and sex-matched healthy controls. This was the case for both NPSLE and non-NPSLE patients. Our findings indicate a higher overall degree of neuronal damage in SLE patients compared with healthy controls independently of ongoing symptoms attributed to SLE involvement of the nervous system. This finding is in line with the larger volumes of WML in SLE patients, with or without NPSLE, compared with healthy controls (7). In addition, the finding is consistent with another study demonstrating higher levels of CSF NfL in SLE patients without overt NP involvement compared with healthy controls (17).
In our study, plasma or CSF NfL concentrations did not differ when comparing SLE patients with or without a history of NPSLE, defined according to the SLICC A, SLICC B and ACR models. A recent study demonstrated higher plasma NfL levels in 45 SLE patients with a history of “focal CNS involvement” (defined as the ACR NPSLE manifestations cerebrovascular disease, seizures, myelopathy, aseptic meningitis, movement disorder, and demyelinating syndrome) compared with 75 non-NPSLE patients (20). In our ACR NPSLE group, we had a lower number of patients (12) with “focal CNS involvement”, and no significant differences of NfL concentrations were seen between these groups (data not shown). Similarly, in another study, SLE patients with overt CNS involvement had a 7-fold increase of CSF NfL concentrations compared with SLE patients without overt CNS involvement, and the levels decreased after cyclophosphamide treatment (17). Our study was not designed to specifically evaluate SLE patients with acute onset NPSLE events, however, very high plasma NfL concentrations may indicate acute central nervous tissue degeneration. In our SLE group, a small minority of patients had new onset of major NPSLE manifestations; the patient displaying the highest level of plasma NfL concentration had acute myelitis (plasma NfL level 20.5 pg/ml as opposed to the mean plasma NfL levels of 5.5 pg/ml in healthy controls). Other study designs are needed to assess the value of plasma NfL in acute NPSLE diagnostics. It is challenging to discern ongoing NPSLE activity from prior events with the present diagnostic methods, and this could be an explanation to the lack of association between NfL concentrations and NPSLE. Furthermore, NfL is a biomarker of neuronal damage and cannot be used to distinguish nervous tissue involvement in SLE versus a co-morbid disorder involving damage to nervous tissue. Thus, patients in the non-NPSLE groups studied herein, may have a low-grade neuronal affliction related to SLE without displaying symptoms that lead to classification as NPSLE (30). Patients may have a co-morbid neurological or psychiatric disorder unrelated to SLE leading to elevated NfL levels, although this was not an issue in our study. Overall, knowledge is insufficient whether the autoimmune neuroinflammatory process in NPSLE can be chronic or intermittent. In future studies, we aim to clarify this current lack of knowledge by analyzing longitudinal measurements of NfL.
Overall organ damage was independently associated with plasma NfL in the multivariate analysis. The SLICC/ACR-DI has been associated with several components of SLE, including disease activity and specific markers of active disease, socioeconomic conditions, hypertension and medication, and thus encompasses a combination of factors of importance for prognosis (31). Our results also demonstrated a possible association between higher NfL concentrations in patients with low complement factor C3 levels, or with a history of renal manifestations or anti-dsDNA positivity. Low levels of complement factors indicate complement consumption, which is a marker of SLE disease activity and is included in the SLEDAI-2K (32). More than half of our SLE patients had levels of complement factor 3 below the lower limit of normal at the time of the study, and although disease activity in general was low (median SLEDAI-2K score 2), two-thirds had at least one point on the SLEDAI-2K. This finding suggests immunological SLE-activity in the majority of our patients, although clinically low- or inactive. This low-key activity is reflected by interferon activity in clinically quiescent SLE seen in other studies (33). Our results are in accordance with a recent study demonstrating associations between plasma NfL concentrations and SLE-related organ damage and disease activity (19). In that study, however, higher NfL levels were associated with higher SLEDAI-2K score, and not with complement levels or anti-dsDNA antibodies. Although only 3% of our patients had eGFR < 60 ml/min/1.73m2, we demonstrated a possible association between plasma NfL levels and plasma creatinine levels, however, the association was not significant in the multivariate model. The association between plasma NfL and creatinine has been demonstrated in two previous studies in SLE patients, and in a study of older adults and patients with diabetes (19, 20, 34). This finding underlines the importance of including renal function when interpreting plasma NfL concentrations. Hypothetically, an SLE phenotype with renal involvement may constitute a subgroup of lupus patients that is at risk of increased neuroaxonal damage, driven by active disease reflected by anti-dsDNA antibodies and ongoing complement activation, resulting in increased NfL concentrations and MRI abnormalities, as well as overall organ damage. Our findings would be in line with a process not solely depending on a clinically clear-cut neuropsychiatric involvement within the current definitions. Mechanisms for neuronal damage in SLE are not known, although SLE disease activity, medication, hypertension and small vessel vasculopathy have been implicated (10, 11). Renal involvement in SLE is associated with higher disease activity and organ damage, hypertension, accelerated arteriosclerosis, and development of cardiovascular disease, which is consistent with the NfL-results herein (35, 36). Blood levels of NfL can indeed be used as a marker of ongoing subclinical cerebral small vessel disease assessed by silent MRI lesions (37). Emerging evidence suggests a role of the complement system in both neuroprotection and neuropathology, and future studies may explore the possible direct or indirect consequences of complement activity in the nervous system in SLE (38). Neuroimaging abnormalities and cognitive dysfunction have been described to be associated with aPL, however, we were not able to demonstrate any associations between aPL and NfL concentrations in this study (39). Certainly, future studies are needed to investigate whether increased NfL in SLE during follow-up reflects disease activity over time, as well as neuronal damage and future cognitive decline, and if normalization of NfL-concentrations through treatment results in prevention of this outcome.
Higher plasma NfL concentrations were associated with larger total CSF volumes when adjusted for age and head size. Larger CSF volumes can indicate a decrease in brain volume and larger periventricular WML volumes which might, in turn, indicate a higher degree of cerebral small vessel disease (40).
We demonstrated that higher plasma NfL concentrations in SLE patients may be associated with lower scores of simple attention, and with cognitive impairment of verbal memory and psychomotor speed. These findings are in accordance with previous findings showing that higher CSF and plasma NfL concentrations in SLE patients were associated with impairment of psychomotor speed or motor function (18, 19).
Both higher CSF and plasma levels of NfL correlated with age, which is in accordance with previous studies (14, 15). We demonstrated that plasma NfL was strongly correlated with CSF NfL in SLE patients. This important finding may facilitate the assessment of neuroaxonal damage in SLE patients in daily clinical practice, as well as in future studies, as peripheral blood testing compared to lumbar puncture is less invasive and more convenient. In addition, using blood samples will increase the facilitation to obtain comparable samples from control groups.
The limitations of this study include the relatively small group sizes and our findings need to be confirmed in larger studies. To reduce age-related MRI abnormalities and cognitive decline, and to reduce study group heterogeneity, we included only females and only patients under an upper age-threshold. Therefore, our conclusions are limited to this group. Educational level is a potential unadjusted true confounder due to the possibility that the healthy controls are comprised of health-care personnel and may be more educated than the SLE patients, however, we did not have data on educational level of our subjects. We did not have creatinine or eGFR levels of the healthy controls and could not adjust for these variables when comparing plasma NfL levels between SLE patients and controls. Also, the design of the study is cross-sectional restricting causality assumptions. Studying consecutive patients resulted in low overall SLE disease activity and NP symptoms were not new-onset, making it difficult to investigate associations with disease activity or acute NPSLE. On the other hand, the design of the study makes it possible to study the chronical aspect of the disease.
In conclusion, in this study we demonstrated higher concentrations of plasma NfL in SLE patients, with or without symptoms of nervous system involvement, compared with healthy controls, indicating a higher extent of neuronal damage in SLE patients. Furthermore, in a population of SLE patients predominantly comprised by low SLE disease activity and normal kidney function, we demonstrated a possible association between higher plasma NfL concentrations in patients with low complement C3 levels and ongoing treatment, compatible with an SLE phenotype with renal involvement. This may indicate a higher degree of neuronal breakdown in patients with active disease, also without overt clinical symptoms. A higher degree of SLE-related organ damage was independently associated with higher plasma NfL concentrations, further supporting that a more severe disease in general is of importance, and a correlation between NfL levels and CSF volumes was demonstrated. Further longitudinal studies are needed to assess whether the lupus phenotype composed of glomerulonephritis and higher organ damage, mediated by chronic lupus inflammation, are more susceptible to neuronal damage reflected by higher NfL levels, and consequently MRI abnormalities and cognitive dysfunction. NfL may serve as an indicator of neuronal damage in SLE patients in further studies.