The temporal dynamics of plasma fractalkine levels in ischemic stroke: association with clinical severity and outcome

Fractalkine (CX3CL1, FKN) is the only chemokine with the CX3C motif and was first described in 1997 [1]. It appears to be unique because it exists in a membrane-bound, as well as in a soluble, form after being cleaved mainly through a disintegrin and metalloproteinase domain-containing protein 10 (ADAM10) [2]. FKN has chemotactic and adhesive functions for inflammatory cells. Its corresponding receptor, CX3CR1, is expressed mainly on monocytes, natural killer cells, T-lymphocytes and microglia.

FKN appears on the surface of endothelial cells and neurons and has been associated with several, mainly inflammatory, pathologies such as rheumatoid diseases [3], atherosclerosis, and neurodegenerative disorders like Alzheimer’s [4] and Parkinson’s disease [5]. In the brain, the CX3CL1/CX3CR1 axis takes part in neural-microglial signaling [6‐8].

Recently, FKN has become notable in stroke research, primarily in animal models and in in vitro studies. The distinct roles of FKN and its receptor in cerebral ischemia are not yet sufficiently investigated and whether FKN adopts neuroprotective or neurotoxic functions remains controversial. Rodents treated with exogenous FKN after permanent occlusion of the middle cerebral artery (pMCAO) had smaller infarction sizes and less neurological deficits [9]. Controversially, genetically CX3CL1- and CX3CR1-deficient mice are also less susceptible to cerebral ischemia [9‐11]. In inflammatory states of the brain, FKN probably acts via modulatory effects on microglia [12, 13], which exert crucial activities in the response after ischemia. Thus, the interactions between neurons and microglia, mediated by the CX3CL1/CX3CR1 pathway could be essential in the pathology of stroke. In contrast to most other markers in stroke, FKN might, therefore, be less reactively driven by post-ischemic inflammation, but could be taking part in the process itself.

Donohue et al. recently proposed that higher FKN plasma levels are associated with better 180d-Outcome of ischemic stroke [14]. Based on these results we aimed to investigate the temporal pattern of FKN plasma levels in the hyperacute as well as long-term stage of ischemic stroke and whether FKN is independently associated with clinical outcome.

The main findings of this study are: (i) FKN levels in patients with moderate to severe ischemic stroke differ from controls at distinct time points; (ii) after ischemic stroke, the temporal profile of FKN levels differs in regard to stroke severity; (iii) dynamics of FKN are independently associated with neurological outcome; (iv) dynamics of FKN inversely correlate with the extent of brain damage as measured by S100B; and (v) the correlation between FKN and CRP and TIMP-1 that is present in healthy controls is not present in stroke patients, neither in the acute nor in the prolonged stage.

In this study, we observed an alteration of FKN levels after stroke compared to controls with CVRF, depending on stroke severity. Although we provide early levels of fractalkine, we cannot give any information about FKN dynamics in the phase before 6 h after stroke. However, at 3 d, at 7 d, and by tendency, at 90 d in patients with moderate/severe strokes, FKN levels are reduced compared to controls, while levels in patients with mild strokes are not. With this study, we are not able to answer whether FKN is released into plasma reactively, or whether it is itself intervening in the cascade of post-ischemic inflammation. At 90 days after stroke - a time point, at which post-stroke inflammation has decreased - FKN levels still differ between patients with mild and moderate/severe strokes. This might hint at an individual potential for regulating FKN. Furthermore, FKN might be rapidly regulated after the onset of stroke, since it differs at early time points after stroke depending on stroke severity.

Donohue et al. described the levels of FKN being linked to long-term outcome as measured by 180d-mRS [14]. Since the cascade of inflammation after acute ischemic stroke evolves rapidly, we aimed to investigate the association between early FKN levels and clinical outcome. Regarding the absolute levels of FKN, we could not find an association with outcome but could demonstrate a linkage only to stroke severity on admission. However, in the current study we could demonstrate that less decrease of FKN over time is related to a beneficial stroke outcome. In concordance to that, less decrease of FKN is also associated to smaller infarction sizes, represented by lower values of the damage marker S100B. This might support the assumption of an important linkage of FKN with the pathology of stroke.

Although there is some controversy about its distinct functions, the chemokine FKN has been regarded to fulfill a neuroprotective role in several in vitro and in vivo models of neurodegenerative and neuroinflammatory diseases. One hypothesis is that the interaction of the neuron-expressed FKN and its corresponding receptor, CX3CR1, which is found on microglia, leads to an inhibition of microglial destructive inflammation [4, 5, 20, 21]. Treatment of microglia with FKN in vitro, for example, provokes an inhibition of cell death mechanisms through Fas ligand-induced cell death and other pro-apoptotic mechanisms [22]. Furthermore, it has been shown in a cell culture model that survival of hippocampal neurons can be induced by the presence of FKN via the protein kinase Akt pathway [23]. High levels of FKN might also play a role in cell death after stroke.

The main differences in ∆FKN were observed for the time interval from 6 h to 3 d. This could be because FKN may take part in the recruitment of leucocytes. In a murine stroke model, it has been shown that the influx of immune cells (for example, macrophages, lymphocytes and neutrophils) peaks at 1 to 3 d after ischemia [24]. Additionally, the infiltration of macrophages in the post-ischemic brain significantly increases between day 1 and day 4 after an experimental stroke [25]. Especially the nonclassical subset of monocytes, which expresses high levels of CX3CR1, could be of particular importance in the FKN-dependent recruitment of cells. Urra et al. reported a significant proportional decrease of nonclassical monocytes at 2 d after stroke onset. They also described an inverse association of these cells with infarct size and outcome [26]. This might be in line with our findings about the inverse association of ∆FKN with infarct size and outcome. Although it has been shown that monocyte subsets are involved in the post-ischemic inflammatory processes [27], detailed investigations of monocyte recruitment in connection with FKN and other chemokines to the brain are lacking. In a murine model of myocardial infarction, Ly-6C^low monocytes, which are comparable to the human nonclassical subset, dominate a second phase of inflammatory response, which starts at some days after ischemia and have been regarded as reparative [28] (for review see: [29]). Recruitment of Ly-6C^low monocytes to the post-ischemic myocardium is executed via FKN. Also after stroke, the migration of potentially protective nonclassical monocytes might be dependent on FKN. Possibly, a neuroprotective characteristic of FKN could, therefore, also be explained by a chemotactic impact on nonclassical monocytes. Measurements of the expression of CX3CR1 on circulating leukocytes would be helpful in future studies to address this point and to investigate the association of receptor expression and plasma FKN dynamics. While the experimental data suggest a potential neuroprotective effect of FKN, its impact in patients is uncertain. So far, it even remains unclear if concentrations of FKN in plasma are accurately reflecting its actions in the central nervous system.

FKN is also present on the surface of endothelial cells and contributes to chemotaxis and invasion of inflammatory cells, such as, for example, monocytes [30, 31]. The CX3CL1/CX3CR1 axis is therefore suspected to carry out pro-atherogenetic functions, as it might contribute to formation of atherosclerotic lesions [32‐35]. In our control collective, which contains individuals with cardiovascular risk factors, CRP levels are strongly correlated to FKN, implying that there might be a linkage to cardiovascular risk in humans. Recently, FKN was suggested to be an independent predictor of cardiovascular mortality in a heart failure patient cohort [36]. As FKN levels might be distorted by atherosclerotic diseases we also included the etiology of stroke in our multivariate model. But, ∆FKN was independently associated with clinical stroke outcome.

While FKN levels correlate significantly with CRP and TIMP-1 in healthy controls, there is no association in stroke patients. The fact that the associations are not restored until 90 d after stroke may underline the long-term impact of the acute event. The finding that there are no correlations to other markers of inflammation in stroke patients is in accordance to the different specific dynamics of FKN in stroke.

Although we provide evidence of an independent association of FKN dynamics with stroke outcome, and therefore present FKN as a putative future biomarker in stroke, the interpretation of the data must be done cautiously. Considering the multifactorial pathology of ischemic stroke, the patient cohort is relatively small. For that reason, we also could not evaluate a probable effect of any treatment, such as, for example, thrombolysis. Furthermore, despite strict exclusion criteria, an overlap of pre-ischemic inflammation processes that might distort the results cannot be ruled out completely. Due to strict exclusion criteria concerning infectious diseases, we could include relatively fewer patients with severe strokes. Finally, as this is a retrospective study, measurements of CX3CR1 expression on vital leucocytes could not be done.