Discussion
To determine the effect of CdA on murine microglia, we examined
phenotypic and functional properties of naïve, LPS-, and IL-4-stimulated
microglia treated with different concentrations of CdA for 24 hours. We
found that CdA only affects stimulated and not naïve microglia.
Furthermore, different concentrations of CdA influenced different
aspects of microglia function. CdA concentrations putatively overlapping
with CSF concentrations in humans (0.1-1 µM) significantly reduced the
granularity, phagocytotic ability, and motility of LPS-stimulated
microglia, while higher CdA concentration (10 µM) increased the mRNA
expression of pro- and anti-inflammatory molecules, but not protein
secretion. In addition, we found that activation of microglia increased
the expression of DCK responsible for the activation of the prodrug.
CdA can cross the BBB and thereby potentially exerts an effect on CNS
resident microglia (2, 3). The maximum plasma concentration of CdA after
administration of a 10 mg tablet is in the range of 0.07-0.1 µM (23). As
the CSF concentration is 25% of the plasma concentration (2, 3), the
maximum CSF concentration attainable is approximately 0.019-0.025 µM.
The active form of CdA is CdA triphosphate, which is the phosphorylated
form of the drug (2, 3). The first step of this process is catalyzed by
deoxycytidine kinase (DCK), the activity of which is 10 times lower in
mice than humans (17, 18). The maximum CdA concentration detected in the
CSF of patients thus may correspond to murine concentrations in the
range of 0.1-1 µM. Therefore, we applied CdA concentrations of 0.01-10
µM in this study of murine microglia.
We first investigated whether the applied CdA concentrations were toxic
to microglia. We found that the viability of microglia was not affected
during the short culture conditions, similar to a previous study that
applied the intravenous compound to rat microglia cultures (16).
Therefore, we considered that using these concentrations for short time
stimulation would not have a depleting effect on microglia.
LPS stimulation resulted in a morphological change of microglia
characterized by an increase in SSC and thus a more granular morphology,
as expected. Co-treatment with CdA 0.1 µM showed a tendency towards
decreased granularity suggesting that CdA may have effect on the
functional properties of microglia.
Therefore, we examined the phagocytotic ability and motility of
microglia. We found that LPS stimulation significantly increased the
phagocytotic ability of microglia, which correlated with the
pro-inflammatory phenotype induced by LPS. Such increased phagocytotic
ability of LPS-stimulated microglia was decreased by CdA in the range of
0.1-1 µM. Furthermore, the decrease in phagocytosis by CdA corresponds
to the observed decreased granularity of microglia. CdA had no effect on
naive microglia, which is consistent with findings of former studies
performed on rat microglia (16), and murine dendritic cells (24).
Microglia motility is important for maintaining CNS homeostasis via
interaction with other cells and scanning of the CNS environment (8, 11,
25, 26). We measured the random migration by (i) track displacement,
i.e. the distance from start to end point, and (ii) total distance
travelled that includes twists and turns and describes the patrolling of
migrating cells. LPS stimulation significantly increased track
displacement and showed a tendency towards increased total distance
travelled. This is consistent with the acquisition of effector functions
induced by LPS and a tendency towards increased patrolling of the
surroundings. In contrast to random migration, LPS decreased chemotactic
migration of rat microglia (27). We also found that CdA (0.1-1 µM) did
not affect the motility of resting microglia.
Observing changes in morphology, phagocytotic ability and migration of
LPS-stimulated microglia, we next examined how CdA alters gene
expression and protein secretion of microglia stimulated with LPS and
with anti-inflammatory IL-4. LPS increased the expression of pro- and
anti-inflammatory markers consistent with a recently published study
(27). Co-treatment with CdA in the high concentration (10 µM) increased
the gene expression of both pro- and anti-inflammatory markers, such as
IL-1β, TNF, iNOS, Arg1, and TNFR2 suggesting a potentiating effect on
activated microglia. Interestingly, only the gene expression of
anti-inflammatory TNFR2, but not the pro-inflammatory TNFR1 (28) was
potentiated by CdA. The increased protein secretion of IL-1β, TNF, and
IL-10 induced by LPS was not further increased by CdA (0.1-10 µM)
consistent with a previous study on rat microglia (16). This suggests
that CdA only affects gene expression but not protein secretion. We may
speculate that accumulation of CdA in microglia may interfere with RNA
processing besides interference with DNA replication and this may alter
translational mechanisms.
We also found that IL-4 stimulation resulted in an increased gene
expression only of Arg1 but it was not affected by CdA co-treatment.
This could indicate that IL-4 is a less potent stimulus in the
short-term, and microglia may not be sufficiently activated for the
potentiating effect of CdA. In summary, CdA in the applied concentration
did not have effect on cytokine secretion of activated or resting
microglia in short-term primary cultures. The observed effect on gene
expression was achieved by higher concentrations of CdA than may be
expected in the human CSF during treatment.
Activation of microglia by LPS and IL-4 also increased the gene
expression of DCK that phosphorylates CdA to its active form. This may
suggest that activated microglia may increase its ability to respond to
CdA. In lymphocytes, CdA interferes with DNA synthesis and repair
resulting in depletion of both proliferating and resting cells (2, 3).
The mechanism by which CdA affects microglia functions remains to be
determined, and our data do not indicate microglia depletion in the
short term. A previous study also suggested that an additional mechanism
beside depletion may be responsible for the effects on microglia, such
as reduced proliferation (16). In dendritic cells, the function of CdA
is partially independent from its phosphorylation (24); but it may not
be true in microglia, as the effect of CdA on proliferation was lost
upon addition of a competitive substrate of DCK (16).
In summary, we found that CdA used in a concentration that may be
reached in the human CSF during approved treatment of MS, affected
activated but not naïve microglia; this may be related to the
upregulated gene expression of DCK upon activation. Co-treatment with
CdA reduced the phagocytotic ability and random migration of activated
microglia, while viability was not affected among the applied
conditions. Although gene expression of both pro- and anti-inflammatory
molecules was potentiated by CdA, it was induced only by higher
concentration than should be expected in the CSF, and protein secretion
was not altered. CdA also potentiated the gene expression of
anti-inflammatory TNFR2. These data suggest that beside depletion of
peripheral lymphocytes, CdA may induce additional effects particularly
in CNS resident cells that contribute to its beneficial effects in MS.