Cerebrospinal fluid levels of IL-6 are decreased and correlate with cognitive status in DLB patients

In line with the many previous preclinical and clinical studies demonstrating a link
between cognition and IL-6, we show that CSF levels of IL-6 in patients with DLB significantly
and negatively correlated with MMSE score. A similar, although not significant, association
trend between these two variables was also seen in patients with AD. These findings
indicate that increased IL-6 levels are linked to cognitive disabilities in patients
with DLB and patients with AD and could be interpreted as further evidence of involvement
of neuroinflammatory processes in both of these disorders. However, our results also
show that patients with DLB displayed significantly lower CSF concentrations of IL-6
than both control subjects without dementia and patients with AD. This result inevitably
contradicts the idea that neuroinflammatory processes, or at least IL-6 secretion,
is a prominent feature of DLB pathology. Interestingly, a lack of protuberant neuroinflammatory
processes in patients with DLB, as opposed to in patients with AD, has been described
before. As noted in the Introduction section, studies imply that A?
_1–42
pathology underlies the altered IL-6 levels in patients with AD because elevated expression
of IL-6 has been found in close vicinity to A?
_1–42
plaques in these patients 19], 20]. If this scenario holds true, one would expect that patients with DLB, who frequently
also have A?
_1–42
plaques, would show elevated CSF concentration of IL-6. However, in a previous brain
tissue study, researchers compared the amount of activated microglia in patients with
AD and patients with DLB and found that, although activated microglia were present
in all cortical regions of patients with AD, the number of activated microglia in
patients with DLB did not differ from numbers found in control subjects without dementia
44]. The authors of the study pointed out that their patients with DLB, just like patients
with DLB in general 23], did not display neurofibrillary tangles (NFT) characteristic of AD and suggested
that NFT pathology rather than A?
_1–42
pathology is the underlying cause of the neuroinflammatory processes in AD 44]. This reasoning would explain the difference in IL-6 levels between patients with
AD and patients with DLB in our study, as the patients with DLB showed unaltered P-tau
concentrations (considered to be indicative of few or no NFT), whereas the P-tau levels
in the patients with AD were significantly elevated (considered to be indicative of
NFT presence).

Quantification of IL-6 in CSF from patients with DLB has (to our knowledge) been performed
only once previously. That study, in contrast to ours, demonstrated a slight, non-significant
increase in IL-6 levels 36]. The discrepancy between the previous finding and our results could be explained
by the fact that the earlier study used a different IL-6 assay, but other variables
may also be accounted for. For example the mean MMSE score of patients with DLB in
the earlier study was lower than the mean MMSE score in our DLB patient groups (18.0?±?4.8
vs 20.4?±?5.5). Admittedly, the difference in MMSE is small, but in view of our findings
demonstrating an increase in IL-6 levels in patients with lower MMSE scores, it may
well be that an analysis of CSF from DLB patient groups with lower mean MMSE scores
fails to reveal a potential disease-induced decrease in IL-6 levels. Further, the
control subjects without dementia included in our study were significantly younger
than the patients with DLB and patients with AD, which is a limitation, particularly
considering that previous studies have demonstrated an age-dependent increase of IL-6
levels in CSF 45] and serum (for review, see 46]). However, no significant correlation between age and CSF IL-6 levels was found in
our study, a finding also reported by other researchers 17], 24], 47]. The IL-6 levels were still significantly decreased in patients with DLB compared
with both control subjects without dementia and patients with AD when the IL-6 values
were analysed after age adjustment (ANCOVA). In addition, the levels of IL-6 in patients
with DLB were significantly lower than those of patients with AD, who were within
the same age range as the patients with DLB. Nevertheless, to exclude any potential
cofounding effect of age, we selected 21 age-matched DLB patient–control pairs. Analysis
of CSF from this subcohort still showed decreased levels of IL-6 in patients with
DLB compared with control subjects without dementia (p?=?0.003). From a clinical perspective it is important to point out that DLB has a
male preponderance 48] and that our DLB cohort therefore should be considered atypical. However, because
levels of IL-6 were unaffected by gender in all patient groups, we concluded that
gender preponderance did not affect the results in this study.

Further, the size of our cohorts is rather small, which undoubtedly is a limitation
of the study. Another limitation is the fact that we had access only to CSF from the
individuals included in the study and therefore were not able to analyse IL-6 levels
in serum or blood in the same individuals. Indeed, a comparison of IL-6 levels in
blood and CSF could potentially indicate whether the IL-6 source is of peripheral
or brain parenchymal nature. We would therefore like to emphasise the importance that
future ante-mortem clinical studies on the implication of IL-6 in DLB and AD pathology
should include larger cohorts as well as analysis of both CSF and serum and blood.

In recent years, levels of ?-synuclein have been proposed as a biomarker for synucleinopathy,
as patients with PD, DLB and multiple system atrophy often show decreased CSF levels
of the peptide compared with patients with AD 1], 49]–53] and controls 52], 54], 55]. It should be pointed out, however, that a couple of studies have shown contradicting
results, with patients with DLB displaying unaltered 56], 57] or even increased 58] levels of ?-synuclein compared with patients with AD. Additionally, in two studies
researchers reported unaltered ?-synuclein levels in patients with DLB compared with
controls 49], 58], although the mean ?-synuclein value in one study was lower than in controls. The
?-synuclein values in patients with DLB included in our study were significantly decreased,
which is in line with results in the majority of previous studies. Although the mean
value of ?-synuclein in patients with AD was increased, it did not reach significance.
This discrepancy seen in our results and others’ could be due to several factors,
including diagnostic criteria and cohort sample sizes. The fact that all the previously
reported CSF ?-synuclein studies used different analytical methods, including in-house
ELISAs with various antibodies, mass spectrometry and Luminex technology (Luminex,
Austin, TX, USA), is most probably also an important aspect to consider and highlights
the importance of finding a global ‘gold standard’ when analysing CSF ?-synuclein
in the future. Nevertheless it should be underscored that the majority of studies,
regardless of analytical method, point towards ?-synuclein as a promising biomarker
for distinguishing synucleinopathies such as DLB and PD with dementia from other forms
of dementia.

Our study shows a positive correlation between IL-6 and ?-synuclein levels in CSF
from patients with DLB. This finding could be viewed from the perspective that lowered
?-synuclein concentration is thought to mirror the synucleinopathy characteristic
of entrapment of ?-synuclein in the brain parenchyma. If this hypothesis holds true,
IL-6 secretion must decrease as ?-synuclein pathology increases. Such a scenario would
be in line with our results showing decreased levels of IL-6 in patients with DLB
compared with patients with AD and control subjects without dementia. The suggested
scenario, however, inevitably also contradicts the earlier results demonstrating activated
microglia and enhanced mRNA IL-6 levels adjacent to ?-synuclein occlusions in the
hippocampus of patients with DLB and patients with PD 22]. In vitro studies have additionally shown increased microglial activation 59], 60] and IL-6 secretion 59] in the presence of ?-synuclein. Moreover, studies of patients with PD, another neurodegenerative
disorder characterised by synucleinopathy, display microglial activation, increased
mRNA expression of IL-6 22], 61] and elevated CSF levels of IL-6 30], 62]. With regard to these findings, and considering the fact that we found increased
IL-6 levels in patients with DLB with very low MMSE scores, it is important not to
rule out the possibility that IL-6 plays a vital role in DLB pathology. Future studies
using animal and cell culture models are required, however, to fully understand the
impact of IL-6 on brain cell integrity and function in relation to DLB pathology.