The expression of the chemokine receptor CCR5 in tick-borne encephalitis

Tick-borne encephalitis (TBE) is caused by a tick-borne encephalitis virus (TBEV) of Flavivirus genus (family Flaviviridae), transmitted by Ixodes ticks. It is endemic in the temperate zone of Asia and Eastern and Central Europe, where several thousand cases are reported annually, including over 200 cases in Poland [1‐3]. Infection with the European TBEV type is often asymptomatic or results in a mild, flu-like disease. The second phase of the infection, characterized by central nervous system (cns) involvement, occurs in a minority of cases and presents an uncomplicated meningitis, meningoencephalitis of highly variable severity or meningoencephalomyelitis [4‐8]. This wide range of clinical presentations suggests that host factors contribute to the susceptibility to the disease and to its severity. The outcome of the infection may be decided at several points, including the local virus replication and spread from the tick-bite site, the peripheral response immune/inflammatory, the penetration of the blood-brain barrier, and finally, the type and extent of the response within cns [9]. Nervous tissue pathology in TBE is complex and involves both the direct cytopathic effect of TBEV on the infected neural cells and the secondary immune-mediated damage [9‐13].

The cerebrospinal fluid (csf) pleocytosis is a hallmark of the intrathecal inflammatory response during cns infection. In TBE, the lymphoid pleocytosis is relatively low and dominated by Th CD4+ lymphocytes, mostly of Th1 subset, with addition of T CD8+ cells [14‐16]. The role played by different cell populations, including T CD4+ and T CD8+ lymphocytes, is debatable [17]. Both animal experiments and autopsy studies in fatal human TBE cases suggest Th CD4+ lymphocytes are indispensable for the TBEV control and elimination while Tc CD8+ cells contribute to cns immunopathology [13, 18]. Leukocyte migration into cns is driven by cytokines of the chemokine family, and a pattern and timing of expression of particular chemokines and their receptors determines the composition of the infiltrate [19, 20]. Chemokine receptor 5 (CCR5) is a receptor for chemokine ligand 3 (CCL3, previously named MIP-1α), CCL4 (MIP-1β), and CCL5 (RANTES) expressed on memory type and activated Th1 CD4+ lymphocytes, T CD8+ lymphocytes, and monocytes, and involved in Th1-type inflammatory/immune response [21‐25]. Its expression is not increased directly on lymphocyte activation but is up-regulated by pro-inflammatory and Th1-related cytokines (interleukin 2—IL-2, IL-12, tumor necrosis factor alpha—TNF-α, interferon gamma—IFN-γ) and may be decreased in anti-inflammatory/Th2 environment [21, 24, 25]. Several studies show increased CCR5 expression by lymphocytes in the inflammatory csf [26‐30], but the data on its role in flavivirus encephalitis are mostly indirect.

The CCR5Δ32 deletion in CCR5 gene results in a synthesis of a truncated protein and the lack of functional CCR5 in homozygotes [24, 31]. In CCR5Δ32 heterozygotes, both the constitutive expression of CCR5 and its induction by pro-inflammatory stimuli are significantly reduced [24, 25]. The lack of CCR5 confers no evident pathology, probably due to a redundancy within the chemokine ligand/receptor network, but it hampers protective responses against certain viral and parasitic pathogens in animal models [31‐34]. In humans, impaired CCR5 function has been suggested as a risk factor of flavivirus infections. A case of a severe viscerotropic disease after vaccination with an attenuated yellow fever vaccine has been described in a patient heterozygous for both CCR5Δ32 and a mutation in a promoter region of the CCL5 gene [35]. CCR5Δ32 homozygocity has been linked to the increased susceptibility to West Nile virus (WNV), which is a neurotropic flavivirus related to TBEV and responsible for a similarly wide spectrum of clinical presentations, from asymptomatic through mild febrile disease to severe encephalitis [36, 37]. Epidemiologic studies have shown an increased risk of the clinically overt disease in CCR5Δ32 homozygotes infected with WNV, which has been hypothesized to result from the impaired lymphocyte influx into csf [38‐40]. This findings prompted studies on CCR5Δ32 distribution in TBE patients, which have shown the increased risk of TBEV meningitis/meningoencephalitis in CCR5Δ32 bearers [41, 42]. However, the majority of patients with TBE are the wild type allele homozygotes, pointing to a role of additional host factors in defining the susceptibility to TBEV. These factors, for example, individually variable expression of inflammatory and immunoregulatory cytokines, might act either independently of CCR5 pathway or by influencing the CCR5 expression. The latter is highly variable in peripheral blood T lymphocytes in CCR5 wt/wt homozygotes and probably determined in a multi-factorial way [25]. One of the proposed mechanisms influencing an individual level of CCR5 expression is its down-regulation by its ligands, described in vitro for CCL5 [43] and as a correlation in vivo for CCL3 [44].

To clarify the role of CCR5 expression and CCR5 genotype in TBE, we have attempted to evaluate (1) if CCR5 expression in the activated Th lymphocyte population is altered in TBE and/or associated with its clinical presentation, (2) if there is a difference in the baseline expression of CCR5 in TBE patients and the general population, (3) how CCR5 expression in TBE is related to the concentrations of its ligand chemokines, and (4) how the CCR5 and its ligands expression in TBE associate with the CCR5Δ32 allele.

Although an important role of CCR5 in flavivirus encephalitis has been suggested by both animal models and epidemiologic studies, to our knowledge, its expression has not been analyzed in a patient group with respect to the clinical presentation and genetic background so far. To assess its role in TBE, we have studied CCR5 in activated Th lymphocytes (CD3+CD4+CD45RO+) in csf, a cell population that is relatively accessible, reflects the intrathecal inflammatory/immune response, is important for TBEV clearance from cns, and which appears to be preferentially recruited into csf of TBE patients. Interestingly, in our pilot experiments, we have found a tendency for a higher proportion of CD3+CD8+ versus CD3+CD4+ lymphocytes in csf lymphoid cell population in patients with the more severe manifestation of TBE, which is consistent with the hypothesis that cytotoxic T cells are involved in TBE-related immunopathology while Th cells we chose for the study are protective [13, 18].

CCR5 is commonly expressed by intrathecal Th lymphocytes in both infectious (neuroborreliosis) and non-infectious (multiple sclerosis, MS) cns inflammation, usually alongside CXCR3 receptor for chemokines CXCL9 and CXCL10, but its role remains debatable [26, 28‐30]. In MS, CXCR3 is considered responsible for a lymphocyte migration, while the presence of CCR5 is secondary to its expression on the activated (CD45RO+) lymphocytes preferentially recruited to the inflammatory csf [28, 29]. High CCR5 expression has been observed on csf CD4+ and CD8+ lymphocytes in neuroborreliosis, where it was also co-expressed with CXCR3 [26, 27, 30]. According to Giunti et al., CCR5 does not contribute significantly to lymphocyte migration into csf in neuroborreliosis but may be up-regulated on activated Th lymphocytes intrathecally and facilitate their subsequent migration to the nervous tissue [26]. In viral encephalitis, the role of CXCR3 and CCR5 has been studied mainly in animal models and shows differences depending on virus type and strain virulence [15, 20, 47]. The lack of CCR5 expression has no apparent consequences in mice infected with lymphocytic choriomeningitis virus [23] and paradoxically increases leukocyte influx and intrathecal inflammation in murine model of herpes simplex type 1 encephalitis [48]. On the other hand, CCR5 seems to be functional and active alongside CXCR3 in WNV-encephalitis model [19, 32]. CCR5-negative knock-out mice have a normal inflammatory/immune response in the periphery, but leukocyte migration into cns is impaired, accompanied by an inability to clear virus from the brain tissue and an increased mortality. The expression of other chemokine receptors (CCR1, CCR2, CXCR3) and of chemokines (CXCL9, CXCL10, CCL2, CCL3, CCL4, and CCL5) is increased but is unable to compensate for the lack of CCR5, suggesting that the CCR5 role in the infection with neurotropic flavivirus is non-redundant and distinct from encephalitides of other etiology [32]. However, CXCR3 has been found on csf Th lymphocytes from TBE patients, and its ligand CXCL10 has been detected simultaneously in csf by Lepej et al., so the contribution from this receptor remains probable too [16].

In our group of TBEV-infected patients, an activated Th lymphocyte population in csf was enriched in CCR5-expressing cells, suggesting that either they were preferentially attracted into csf or that, alternatively, CCR5 was up-regulated on activated Th lymphocytes in the pro-inflammatory csf environment, possibly influencing their further trafficking, as suggested by Giunti et al. for neuroborreliosis [26]. Both possibilities are not mutually exclusive and point to CCR5 involvement in TBE pathogenesis. Because of a suspected particular role of CCR5 in flavivirus infections, we expected its expression to be higher in TBE compared to other forms of meningitis. This was, however, not the case as the CCR5 csf expression was higher in neuroborreliosis than in TBE. The feature specific for TBE was a tendency for a continuing increase of CCR5 expression at least between the time of hospital admission and early convalescent period 2 weeks later, contrasting with the decrease in other patient groups. These trends are consistent with a relatively low initial lymphocyte count accompanied by the prolonged inflammatory changes in TBE csf, described in previous studies and observable in our study group as well [49].

Vividly increased expression of CCL5 in brains of mice with WNV-encephalitis and TBE suggests it as a main CCR5 ligand in this setting [32, 50, 51]. Our previous results pointed to a possible role of two of CCR5 ligands, CCL5 (RANTES) and CCL3 (MIP-1α) in TBE pathogenesis, although their chemotactic gradient towards csf could not be proven [52, 53]. In the recent study, Palus et al. found no increased CCL3, CCL4, and CCL5 concentrations in serum of TBE patients, which is in agreement with our current results [54]. On the other hand, we have confirmed the concentration of CCL5, but not CCL3 and CCL4, to be significantly increased in the csf of the pilot group of TBE patients. When subsequently studied in a larger patient group, CCL5 csf concentration was increased in comparison with healthy controls, correlated with the csf inflammatory parameters and tended to correlate with the csf Th CCR5 expression. When we compared CCL5 expression between TBE, neuroborreliosis, and other aseptic meningitis patients, the differences between these groups were analogous to the trends observed for CCR5 expression: in TBE, the initial CCL5 concentration was low in comparison with neuroborreliosis, but its increase was characteristically protracted into the convalescent phase. These data are consistent with CCL5 and CCR5 cooperating in driving lymphocyte migration into csf in TBE and influencing the course of the intrathecal inflammation. CCL5 concentration was relatively high in TBE patients with meningoencephalomyelitis, who also had a tendency for a higher intrathecal CCR5 expression and pleocytosis. The other studies also show the correlation of the intrathecal concentrations of CCR5 ligands with a severity of neurologic involvement. The mouse strains genetically susceptible to TBEV present with a higher expression of CCL3, CCL4, and CCL5 in the brain parenchyma during experimental TBE compared to relatively resistant strains [55], and high CCL5 concentration in patients infected with Japanese encephalitis virus is associated with mortality [56]. Interestingly, in a patient with a flu-like TBEV infection, a concentration of CCL5 in csf was higher than in any of the healthy controls, without any signs of concomitant or subsequent intrathecal inflammation, suggesting CCL5 intrathecal up-regulation even without the clinically overt cns infection.

Another line of evidence supporting the role of CCR5 in flavivirus infection comes from the studies comparing CCR5 genotypes in patients with healthy subjects. According to Glass et al. and Lim et al., CCR5Δ32 homozygotes have a several-fold increased risk of a symptomatic WNV infection [37, 38]. The data suggest that almost all infected CCR5Δ32 homozygotes develop a symptomatic disease versus a minority of the general population [19]. The increased frequency of CCR5Δ32 allele and CCR5Δ32/CCR5Δ32 genotype has been detected in Lithuanian patients with TBE [41, 42], but not in patients from Novosibirsk area (Asiatic part of Russia) likely infected with the Siberian TBEV subtype [7, 57]. The discrepancy between the European and Siberian TBE studies could result from the influence of genetic background in distinct human populations or from the differences between two separate TBEV types. The main difference between the results of WNV and European TBEV studies was the susceptibility of CCR5Δ32/wt heterozygotes. No association of CCR5Δ32 heterozogosity with WNV infection was observed, suggesting that even a reduced expression of CCR5 is sufficient for protection [39], but in the European patients, increased TBE risk was associated with CCR5Δ32 heterozogosity, as if the quantitatively changed CCR5 expression was sufficient to alter susceptibility to TBEV [41]. The reasons for that difference are unclear, especially that mechanisms by which CCR5 deficiency could influence susceptibility to flavivirus in humans has not been studied. The simple explanation in agreement with the animal models is an impaired lymphocyte influx into cns, resulting in a more significant and clinically manifest neuroinfection [39]. However, this would also favor clinically more severe cns involvement, which has not been confirmed [38, 39, 42]. In the contrary, a large epidemiologic study of blood donors has shown that the CCR5Δ32 correlates with the risk of any clinically overt WNV-related disease (including mild flu-like infections) versus asymptomatic infection and not specifically with meningitis/encephalitis [40]. That suggests that CCR5Δ32 associates with the impairment of the early peripheral response to WNV and TBEV resulting in the occurrence of the symptomatic disease, while its connection with neuroinvasion and intrathecal response in unclear.

Consistent with the previous studies, we have not observed an evident difference in the clinical presentation between TBE patients with CCR5Δ32/wt and wt/wt genotype neither in the group presented here nor in a larger population of genotyped patients from our center (unpublished data). There was also no difference in csf inflammatory parameters between CCR5Δ32/wt and wt/wt TBE patients. That suggests that either (1) CCR5 does not play a role in driving lymphocyte migration into cns in TBE, (2) CCR5 drives lymphocyte migration in TBE but its impaired expression may be compensated by alternative pathways, or (3) CCR5 drives lymphocyte migration in TBE, but its expression in CCR5Δ32/wt is induced to a level sufficient to fulfill that role. We attempted to clarify this by comparing a phenotypic CCR5 expression in TBE patients with both genotypes. As expected, the peripheral Th lymphocyte CCR5 expression in CCR5Δ32/wt heterozygotes was lower than in wt/wt homozygotes both in healthy controls and in TBE patients on admission to hospital, but in TBE, the difference became insignificant in the early convalescent period. This may be an artifact due to a small number of CCR5Δ32/wt samples, but if confirmed in a larger group of patients, it could also suggest a compensatory CCR5 up-regulation in response to TBEV challenge. More importantly, the CCR5 expression in csf obtained from three CCR5Δ32/wt TBE patients was several-fold increased in comparison with blood and reached the lower end of a value range found in wt/wt homozygotes. This suggest that the expression of CCR5 and/or its ligands was up-regulated to a level sufficient for a Th lymphocyte migration to cns and effective virus control, consistent with a normal pleocytosis and an unremarkable clinical presentation in CCR5Δ32/wt patients. CCL5 level in csf was not increased in CCR5Δ32/wt versus wt/wt genotype, so it seems not directly involved in a compensation for an impaired CCR5 expression. As CCR5Δ32/wt patients seem to have a normal intrathecal response to TBE, the increased susceptibility to TBEV associated with this genotype should depend on the impairment of the early, peripheral response, during the first febrile phase or even locally at the tick-bite site, which agrees with the conclusions from epidemiologic observations by Lim et al., who have suggested a similar mechanism for an increased susceptibility to WNV in CCR5Δ32/CCR5Δ32 homozygotes [40].

CCR5Δ32 variant is present in a small minority of TBE patients, including 2.3 % of CCR5Δ32 homozygotes in the Lithuanian TBE group [41] and only 7 CCR5Δ32/wt heterozygotes in our study, so additional predisposing factors must be involved in the remaining cases. As evident in our results as well, the expression of CCR5 in peripheral blood T lymphocytes in CCR5 wt/wt homozygotes is highly variable. Several common genetic polymorphisms associated with that variability have been identified in the promoter region of CCR5 and in the genes for CCR5 ligands, CCL5 and CCL3L1, the latter probably influencing CCR5 expression indirectly, through their association with the CCL3 and CCL5 synthesis [44, 58, 59]. We have hypothesized that the variability of CCR5 expression associated with factors other than CCR5Δ32 allele should have a similar clinical effect and contribute to a variable response to TBEV in the CCR5 wt/wt population. To verify that we have studied CCR5 in the peripheral blood Th cells in patients 6–8 weeks after the hospital admission for TBE, after the normalization of the clinical signs and symptoms and the inflammatory parameters in the periphery, we consider it a proxy of the constitutive baseline level. In fact, as we detected no hint of any dynamics of the peripheral CCR5 expression in CCR5 wt/wt homozygotes in the course of TBE, the values measured in the neurologic phase could be considered close to the baseline too. However, the median expression at any of the time points studied did not differ from the results in healthy controls and in other patient groups, which is especially striking in the face of a high individual variability. It could be speculated that TBE patients had originally lower CCR5 peripheral expression than controls, masked by a prolonged up-regulation lasting from the onset of meningitis/encephalitis to the late convalescent period. This assumption could be definitely dismissed only by a systematic study of patients with a history of TBE at still later time points, months to years after infection, to check if there is any sign of the return to a hypothetical lower baseline level. Alternatively, a study of CCR5 expression patients in a first phase of TBE before the onset of meningitis and the ones with a mild flu-like TBEV infection would be informative as for its role in the early response but is difficult to conceive as these patients are rarely diagnosed and hospitalized. Our observations with that respect are fragmentary: in three patients in whom peripheral blood lymphocyte CCR5 was measured in the first phase of the infection and who later developed meningitis, it was within the range of values found in other groups, as it was in three patients with a mild flu-like infection without cns involvement, not giving evidence for any inter-group variability or dynamical change of CCR5 expression before the cns invasion. At present, there is no data to support the association between the low constitutive CCR5 expression in the peripheral blood-activated Th cells and the susceptibility to TBEV in persons with a CCR5 wt/wt genotype.

Our results support the pathogenetic role of CCR5 and its ligand CCL5 in the neurologic phase of TBEV infection. The expression of CCR5 in activated Th lymphocytes was not particularly high in TBE patients comparing with other viral and borrelial meningitis but was relatively long-lasting and correlated with other parameters of the intrathecal inflammation. Our study did not address the role of other chemokine receptors, which may be co-expressed and cooperate with CCR5. The possible co-expression of CCR5 and CXCR3 in the Th cell population and their relative importance for the lymphocyte migration warrants a further study, which could further clarify the role of CCR5 in this setting as well.

We did not detect evidently impaired CCR5 expression or any features of altered intrathecal inflammatory response in TBE patients heterozygous for CCR5Δ32 allele, and we infer that the CCR5 expression is induced to a level adequate for the effective lymphocyte migration to cns in them. The observed increased incidence of TBE in CCR5Δ32 bearers may depend on the impaired early peripheral response to TBEV, possibly even at the tick-bite site before the further virus spread, which requires further study. However, we found no evidence that, besides a relatively rare CCR5Δ32 variant, a low baseline expression of CCR5 in the peripheral blood lymphocytes may be a marker of an increased susceptibility to flavivirus.