Role of plasmacytoid dendritic cells and other accessory cells in the activation of human natural killer cells by herpes simplex virus type 1 Die Rolle plasmazytoider dendritischer Zellen und anderer akzessorischer Zellen in der Aktivierung humaner natürlicher Killer-Zellen durch Herpes-Simplex-Virus-1 Der Naturwissenschaftlichen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg zur Erlangung des Doktorgrades Dr. rer. nat. vorgelegt von Karin Petra Vogel aus Nürnberg
101
Embed
Role of plasmacytoid dendritic cells and other accessory ... · Role of plasmacytoid dendritic cells and other accessory cells in the activation of human natural killer cells by herpes
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Role of plasmacytoid dendritic cells and other accessory cells in the
activation of human natural killer cells by herpes simplex virus type 1
Die Rolle plasmazytoider dendritischer Zellen und anderer akzessorischer
Zellen in der Aktivierung humaner natürlicher Killer-Zellen durch
Herpes-Simplex-Virus-1
Der Naturwissenschaftlichen Fakultät
der Friedrich-Alexander-Universität
Erlangen-Nürnberg
zur
Erlangung des Doktorgrades Dr. rer. nat.
vorgelegt von
Karin Petra Vogel
aus Nürnberg
Als Dissertation genehmigt von der Naturwissenschaftlichen Fakultät
der Friedrich-Alexander-Universität Erlangen-Nürnberg
Tag der mündlichen Prüfung 23.01.2015
Vorsitzender des Promotionsorgans Prof. Dr. Jörn Wilms
IFN-, IL-2, IL-10, IL-4, IL-5 or TNF- within supernatants. Significant IL-6 secretion
compared to the mock control was induced by CpG-A (p<0.01) (data not shown), while IL-8
was secreted in all samples including the mock control (FIG. 11). Two cytokines, namely
IL-1 and TNF-, were significantly increased in HSVINF-stimulated PBMC compared to the
mock control and to HSVUV-stimulated PBMC (p<0.05). These findings suggest that IL-1
and TNF- might be involved in the stimulation of NK cell activation and in particular NK
cell effector functions by HSVINF.
FIG. 11. Infectious HSV-1 induces secretion of IL-1 and TNF-. PBMC were stimulated for 18h with
CpG-A, UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as control, and
supernatants were analyzed for cytokines using an 11plex bead array (Affymetrix eBioscience). Secretion of
IL-8, IL-1, and TNF- within PBMC (pg/ml), given as mean and standard error of seven independent
experiments. #
p<0.05 vs. mock; * p<0.05 as indicated (Tukey HSD).
Results
44
In order to investigate the effect of PDC-derived type I IFN, and also of IL-1 and TNF-,
more closely, neutralization experiments were conducted. PBMC were stimulated with
CpG-A, HSVUV and HSVINF in the presence of antibodies against IFN-R, IL-1 and TNF-
or the respective isotype controls. At 12h p.s. NK cell CD69 up-regulation, IFN- secretion
and degranulation as well as IFN- secretion within PBMC were determined. Neutralization
of TNF- significantly decreased CD69 up-regulation induced by CpG-A, HSVUV and
HSVINF (p<0.01) (FIG. 12A), and it also significantly reduced HSVINF-induced IFN-
secretion (p<0.05) (FIG. 12B), while it did not affect NK cell degranulation (FIG. 12C).
Blocking of IFN-R significantly diminished CpG-A- and HSVUV-induced CD69 up-
regulation (p=0.05 and p<0.01, respectively), but had only a minimal effect on HSVINF-
induced CD69 up-regulation (FIG. 12A) and no inhibitory effect on NK cell effector
functions (FIG. 12B, C). In fact, neutralization of the IFN-R even increased IFN- secretion,
although not significantly (FIG. 12B). In contrast, neutralization of IL-1 did neither
influence NK cell activation nor NK cell effector functions. IFN- secretion was reduced by
neutralization of IFN-R, in consistence with the known autokrine loop (Marie et al., 1998),
and interestingly, also by neutralization of TNF- as well as IL-1 (FIG. 12D). Reduction of
IFN- levels was distinct, although only significant for CpG-A (p<0.05 for TNF- and
IFN-R). Simultaneous neutralization of TNF- and IFN-R did not result in increased
effects on HSVINF-induced CD69 up-regulation (FIG. 13A), degranulation (FIG. 13C) or
IFN- secretion (FIG. 13D), while the increase of IFN- secretion observed after IFN-R
neutralization was abolished by combination of both antibodies (FIG. 13B). These findings
indicate a crucial role for TNF- in HSVINF-induced NK cell activation and IFN- secretion,
whereas it is negligible in HSVINF-induced NK cell degranulation. Besides, all three cytokines
seem to be required for the secretion of large amounts of IFN- upon stimulation with either
CpG-A or HSV-1. Furthermore, combined neutralization of TNF- and IFN-R suggests
opposed functions of TNF- and type I IFN in IFN- induction by HSVINF.
Results
45
FIG. 12. TNF- plays a major role in HSV-1-induced NK cell activation. PBMC were stimulated for 12h
with CpG-A, UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as control, in
the presence of neutralizing antibodies against IL-1 (IL-1) and TNF- (TNF-), and an isotype control
(IgG1), against the IFN-/ receptor (IFN-R), and an isotype control (IgG2a). Cells were analyzed by flow
cytometry (FACS), supernatants by enzyme-linked immunosorbent assay (ELISA). A - C. NK cells were gated
as CD56-positive CD3- and CD14-negative population and analyzed for CD69 expression (A), IFN- secretion
(B) and CD107a surface expression (C) (%). D. IFN-2a/2b (IFN-) secretion within PBMC (pg/ml). All values
are given as mean and standard error of five independent experiments. * p≤0.05, ** p≤0.01 as indicated
(Student’s t-test).
Results
46
FIG. 13. Simultaneous neutralization of TNF- and IFN-R does not increase the inhibitory effect of
TNF- neutralization. PBMC were stimulated for 12h with CpG-A, UV-inactivated (HSVUV) and infectious
(HSVINF) HSV-1, or left unstimulated (mock) as control, in the presence of neutralizing antibodies against
TNF- (TNF-), and an isotype control (IgG1), against the IFN-/ receptor (IFN-R), and an isotype
control (IgG2a) or both TNF- and IFN-R, and both isotype controls. Cells were analyzed by flow cytometry
(FACS), supernatants by enzyme-linked immunosorbent assay (ELISA). A - C. NK cells were gated as CD56-
positive CD3- and CD14-negative population and analyzed for CD69 expression (A), IFN- secretion (B) and
CD107a surface expression (C) (%). D. IFN-2a/2b (IFN-) secretion within PBMC (pg/ml). All values are
given as mean and standard error of three independent experiments.
Results
47
5.5 Monocytes contribute to HSV-1-induced TNF- production
Since TNF- seemed to be the key cytokine in HSVINF-induced NK cell activation within
PBMC, we were interested in which cell populations might be responsible for TNF-
production and performed a Cytokine Secretion Assay (Miltenyi Biotec) to detect
TNF--secreting cells within PBMC. We analyzed TNF- secretion within seven different
cell populations, namely PDC, monocytes, B cells, NK cells, T cells, CD4+ T cells and
CD8+ T cells, upon stimulation with CpG-A, HSVUV and HSVINF. We first looked at TNF-
secretion within the individual cell populations and could identify PDC and monocytes as
major TNF- sources with significant secretion upon stimulation with CpG-A (p<0.01),
HSVUV (p<0.01 for PDC and p<0.05 for monocytes) and HSVINF compared to mock (p<0.01)
(FIG. 14A). In addition, CpG-A stimulated TNF- secretion within the B cell population
(p<0.01), while HSVINF induced TNF- secretion within B cells (p<0.01), NK cells (p<0.01),
T cells (p<0.05), CD4+ T cells (p<0.01) and CD8
+ T cells (p<0.05). A significant
difference in TNF- secretion between HSVUV and HSVINF stimulation was observed within
monocytes (p<0.01). We next decided to identify total TNF- secretion within PBMC. For
this purpose we multiplied TNF- secretion within each cell population (FIG. 14A) with the
frequency of the respective cell population within PBMC (FIG. 14B), resulting in each cell
population’s TNF- secretion within PBMC, and combined TNF- secretion of all cell
populations, to get total TNF- secretion within PBMC (FIG. 14C). Interestingly, only
CpG-A and HSVINF induced significant overall TNF- secretion within PBMC compared to
the mock control (p<0.01), and HSVINF-induced TNF- secretion also differed significantly
from HSVUV-induced TNF- secretion (p<0.05), confirming the results of the bead array
(FIG. 11). Notably, monocytes appeared to be key producers of TNF- in this analysis
(FIG. 14C).
Results
48
FIG. 14. Monocytes contribute to HSV-1-induced TNF- production. PBMC were stimulated for 18h with
CpG-A, UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as control, and
analyzed by flow cytometry (FACS). PDC were gated as CD304-positive CD3- and CD14-negative, monocytes
as CD14-positive, B cells as CD19-positive CD3- and CD14-negative, NK cells as CD56-positive CD3- and
CD14-negative, T cells as CD3- and TCR-positive CD14-negative, CD4+ T cells as CD3- and CD4-positive
CD14-negative, and CD8+ T cells as CD3- and CD8-positive CD14-negative population. A. TNF--secreting
cells (%) within the respective individual cell population. B. Frequency of each cell population (%) within
PBMC. C. TNF--secreting cells (%) within PBMC. All values are given as mean (A - C) and standard error (A,
C) of five independent experiments. #
p<0.05, ##
p<0.01 vs. mock; * p<0.05, ** p<0.01 as indicated (Tukey
HSD).
Results
49
5.6 Monocytes can be infected by HSV-1
Next, we decided to engage in HSV-1 infection experiments, because NK cells are known to
recognize infected cells as target cells (Vivier, 2006). Monocyte activation seemed to be
particularly influenced by HSV-1 infectivity, and several working groups could already
demonstrate infection of mononuclear phagocytes (Daniels et al., 1978; Albers et al., 1989).
Since preliminary experiments of PBMC infected with a virus isolate (HSVGFP) expressing a
GFP-VP22 fusion protein also hinted at monocytes as HSV-1 target cells within PBMC (data
not shown), we conducted infection experiments with isolated monocytes. We noticed that
monocytes, which had been purified using magnetic beads specific for CD14, were only in
part positive for CD14 after being in cell culture for 24h or 48h (FIG. 15). Labeling of freshly
isolated monocytes evidenced a purity of about 95%, and staining of monocytes for lineage
markers as well as specific phagocyte markers demonstrated a contamination by other cells of
less than 5% after cultivation (data not shown). Thus, monocytes appear to down-regulate
CD14 when being cultured.
FIG. 15. Monocytes down-regulate CD14 upon cultivation. Monocytes were purified by magnetic-activated
cell sorting (MACS) using CD14-coupled beads and analyzed by flow cytometry (FACS) for CD14 expression.
Representative FACS plot of monocytes immediately post purification (p.p.) and 24h and 48h p.p.
Monocytes were infected with HSVGFP and analyzed for green fluorescence. HSVINF was used
as infectious non-fluorescent and HSVUV as non-infectious non-fluorescent control virus. In
fact, monocytes infected with HSVGFP exhibited significant green fluorescence compared to
the mock control and the two non-fluorescent viruses HSVUV and HSVINF at 24h (p<0.01) and
48h post infection (p.i.) (p<0.01 for mock and HSVUV, n.s. for HSVINF), demonstrating
infection of monocytes (FIG. 16A). The percentage of HSVGFP-infected monocytes declined
Results
50
from 24h to 48h p.i. , indicating rather abortive than productive HSV-1 infection, in
concordance with observations of several working groups (Daniels et al., 1978; Albers et al.,
1989; Bruun et al., 1998). In order to determine productivity of monocyte infection we
analyzed supernatants of HSV-1-infected monocytes and HSV-1-infected human foreskin
fibroblasts (HFF) as control cells for HSV-1 DNA, using quantitative PCR. Cells were
infected with HSVGFP and with another HSV-1 variant (HSVd106S), which is infectious, but
non-replicative (Liu et al., 2009). Infected cells were cultured for up to 5 days (FIG. 16B).
HSV-1 DNA increased over time in supernatants of HSVGFP-infected HFF, whereas it
declined in supernatants of HSVd106S-infected HFF, corresponding to the replication capacities
of HSVGFP and HSVd106S. In contrast, HSV-1 DNA dropped in supernatants of HSVd106S- as
well as HSVGFP-infected monocytes. These results confirm non-productive infection of
monocytes by HSV-1.
FIG. 16. Monocytes are non-productively infected by HSV-1. A. Purified monocytes were infected with
UV-inactivated (HSVUV), infectious (HSVINF), and infectious GFP-expressing (HSVGFP) HSV-1, or left
uninfected (mock) as control, and analyzed by flow cytometry (FACS) for green fluorescence (%) 24h and 48h
post infection (p.i.). Values are given as mean and standard error of eleven independent experiments. ##
p<0.01
vs. mock; ** p<0.01 as indicated (Tukey HSD). B. Purified monocytes of three different donors and human
foreskin fibroblasts (HFF) were infected with infectious GFP-expressing (HSVGFP) and an infectious but
replication-deficient GFP-expressing (HSVd106S) HSV-1 and cultivated for different time periods. Supernatants
of the indicated time points were analyzed by quantitative PCR for viral load (copies/ml). Values of monocytes
are given as mean of three different donors.
Results
51
5.7 Monocytes up-regulate MHC-I upon exposure to infectious HSV-1
Since HSV-1 has been published to down-regulate MHC-I molecules via ICP47 (Hill et al.,
1995; Früh et al., 1995), which might be responsible for HSVINF-induced NK cell activation
(Huard and Früh, 2000), we checked for expression of classical HLA-ABC and non-classical
HLA-E. Monocytes inoculated with HSVINF and HSVGFP exhibited significant HLA-ABC up-
regulation compared to the mock control at 24h and 48h p.i. (p<0.01), with rising kinetics
from 24h to 48h p.i. (p<0.01) (FIG. 17A). In contrast, HSVUV did not induce HLA-ABC up-
regulation, but behaved like the mock control with significant differences to both HSVINF and
HSVGFP at 24h and 48h p.i. (p<0.01). HLA-E was regulated in a similar manner to
HLA-ABC, with overall up-regulation being induced by HSVINF and HSVGFP (FIG. 17B).
Although plotting green fluorescence against HLA-ABC and HLA-E expression indicated
MHC-I down-regulation in few infected monocytes, overall up-regulation of MHC-I in
monocyte cultures was much more distinct (FIG. 18).
FIG. 17. Monocytes up-regulate MHC-I upon exposure to infectious HSV-1. Purified monocytes were
infected with UV-inactivated (HSVUV), infectious (HSVINF), and infectious GFP-expressing (HSVGFP) HSV-1, or
left uninfected (mock) as control, and analyzed by flow cytometry (FACS) for MHC-I expression (MFI) 24h and
48h post infection (p.i.). A. Fold change of HLA-ABC expression, given as mean and standard error of eleven
independent experiments. ##
p<0.01 vs. mock; ** p<0.01 as indicated (Tukey HSD). B. Fold change of HLA-E
expression, given as mean of two (HSVUV, HSVINF) and three (HSVGFP) independent experiments.
Results
52
FIG. 18. Monocytes are infected by HSV-1 and up-regulate MHC-I upon exposure to infectious HSV-1. Purified monocytes were infected with UV-inactivated (HSVUV), infectious (HSVINF), and infectious
GFP-expressing (HSVGFP) HSV-1, or left uninfected (mock) as control, and analyzed by flow cytometry (FACS)
for green fluorescence and MHC-I expression 24h and 48h post infection (p.i.). Representative FACS plot of
green fluorescence (GFP) and HLA-ABC and HLA-E expression.
Since type I IFN are known to induce up-regulation of MHC-I molecules (Samuel, 2001), we
checked monocyte supernatants for INF- and observed reproducible secretion only by
monocytes inoculated with HSVINF and HSVGFP (p<0.05 for HSVGFP vs. mock and vs.
HSVUV, at 24h and 48h p.i.), but not HSVUV (FIG. 19).
Results
53
FIG. 19. Infectious HSV-1 induces IFN- secretion by monocytes. Purified monocytes were infected with
UV-inactivated (HSVUV), infectious (HSVINF), and infectious GFP-expressing (HSVGFP) HSV-1, or left
uninfected (mock) as control, and supernatants were analyzed for IFN-2a/2b using enzyme-linked
immunosorbent assay (ELISA). IFN-2a/2b (IFN-) secretion at 24h and 48h post infection (p.i.), given as
mean and standard error of ten independent experiments. #
p<0.05 vs. mock; * p<0.05 as indicated (Tukey
HSD).
In order to test the hypothesis, that type I IFN were responsible for HLA-ABCE up-regulation
after HSV-1 infection, we performed neutralization experiments, where we infected
monocytes with HSVGFP in the presence of IFN-R and the isotype control (IgG2a).
Comparing IFN-R with IgG2a revealed distinct effects of type I IFN on monocyte infection
as well as HLA-ABCE regulation (FIG. 20).
Results
54
FIG. 20. Type I IFN suppress HSV-1 infection of monocytes and induce up-regulation of MHC-I. Purified
monocytes were infected with infectious GFP-expressing HSV-1 (HSVGFP) in the presence of a neutralizing
antibody against the IFN-/ receptor (IFN-R) and an isotype control (IgG2a) and analyzed by flow
cytometry (FACS) for green fluorescence and MHC-I expression 24h and 48h post infection (p.i.).
Representative FACS plot of green fluorescence (GFP) and HLA-ABC and HLA-E expression.
Neutralization of IFN-R increased monocyte infection at 24h (p=0.05) and 48h (p<0.05) p.i.
(FIG. 21A) and prevented up-regulation of HLA-ABC at 24h (n.s.) and 48h (p<0.05) p.i.
(FIG. 21B) as well as HLA-E at 24h (p<0.05) and 48h (p<0.01) p.i. (FIG. 21C). Furthermore,
IFN-R neutralization significantly diminished IFN- secretion at 24h and 48h p.i. (p<0.05)
(FIG. 21D), once again confirming the positive feedback loop for IFN- production (Marie et
al., 1998) (FIG. 12D). These results prove type I IFN as cause of HSV-1-induced HLA-ABCE
up-regulation by monocytes and further propose type I IFN as potential restriction factors for
productive HSV-1 infection and replication in monocytes.
Results
55
FIG. 21. Type I IFN suppress HSV-1 infection of monocytes, induce up-regulation of MHC-I and trigger
IFN- secretion. Purified monocytes were infected with infectious GFP-expressing HSV-1 (HSVGFP) in the
presence of a neutralizing antibody against the IFN-/ receptor (IFN-R) and an isotype control (IgG2a), or
left uninfected (mock) as control, and analyzed by flow cytometry (FACS) 24h and 48h post infection (p.i.),
supernatants were analyzed using enzyme-linked immunosorbent assay (ELISA). A. Green fluorescent
monocytes (%). B. Fold change of HLA-ABC expression (MFI). C. Fold change of HLA-E expression (MFI). D.
IFN-2a/2b (IFN-) secretion. All values are given as mean and standard error of three independent
experiments. * p≤0.05, ** p≤0.01 as indicated (Student’s t-test).
Results
56
5.8 HSVd106S affects monocytes similar to HSVGFP
GFP is coupled to the tegument protein VP22 in HSVGFP and is thus present within viral
particles, so HSVGFP particles themselves fluoresce. Consequently, green fluorescing
monocytes may not be infected monocytes expressing GFP, but monocytes with fluorescing
viral particles sticking to them. We therefore repeated our infection experiments with
HSVd106S, which carries the GFP gene under the control of a human cytomegalovirus
(HCMV) promoter. HSVd106S particles do not fluoresce, so monocytes can only fluoresce
when they have been infected and express GFP. Infection of monocytes with HSVd106S had
effects similar to infection with HSVGFP. HSVd106S induced significant fluorescence at 24h p.i.
(p<0.05) (FIG. 22A), that was lost at 48h p.i. HSVd106S infection induced up-regulation of
HLA-ABC at 24h and 48h p.i. (p<0.01) (FIG. 22B) and of HLA-E at 24h (p<0.05) and 48h
p.i. (FIG. 22C). HSVd106S-infected monocytes secreted even higher amounts of IFN- than
HSVGFP-infected monocytes (FIG. 22D). Similarities in fluorescence as well as IFN-
induction and HLA-ABCE up-regulation induced by both viruses argue for an actual infection
of monocytes by HSV-1.
Results
57
FIG. 22. HSVd106S affects monocytes similar to HSVGFP. Purified monocytes were infected with infectious
GFP-expressing (HSVGFP) and an infectious but replication-deficient GFP-expressing (HSVd106S) HSV-1, or left
uninfected (mock) as control, and analyzed by flow cytometry (FACS) 24h and 48h post infection (p.i.),
supernatants were analyzed using enzyme-linked immunosorbent assay (ELISA). A. Green fluorescent
monocytes (%). B. Fold change of HLA-ABC expression (MFI). C. Fold change of HLA-E expression (MFI). D.
IFN-2a/2b (IFN-) secretion. All values are given as mean and standard error of six (A, D), five (B), and three
(C) independent experiments. # p≤0.05,
## p≤0.01 HSVd106S vs. mock (Student’s t-test).
Results
58
Altogether, infection experiments demonstrated monocytes as target cells for HSVINF,
suggesting them as crucial cell population in HSVINF-induced NK cell activation not only via
TNF- secretion, but via recognition of infected monocytes by NK cells. MHC-I down-
regulation by infected monocytes is a possible mechanism, yet according to our studies
unlikely. Furthermore, up-regulation of MHC class I polypeptide-related sequence (MIC) A
or B, which would be recognized by activating NK cell receptors (Vivier, 2006), could be
excluded in preliminary experiments (FIG. 23).
FIG. 23. HSV-1 does not induce up-regulation of MHC class I polypeptide-related sequence (MIC)A or B.
Purified monocytes were infected with infectious GFP-expressing HSV-1 (HSVGFP), or left uninfected (mock) as
control, and analyzed by flow cytometry (FACS) for MICA/B expression 24h post infection (p.i.).
Representative FACS plot of MICA/B expression.
Results
59
5.9 Monocytes mediate NK cell effector functions upon HSV-1
infection within the PBMC context
In order to investigate the actual contribution of monocytes and also of PDC to
HSVINF-induced NK cell activation within the PBMC context, we conducted cell depletion
experiments, comparing non-depleted PBMC with monocyte- or PDC-depleted PBMC.
Depletion of monocytes as well as PDC decreased CD69 up-regulation, although to a variable
extent (FIG. 24A): CpG-A-induced CD69 up-regulation was significantly reduced only by
monocyte depletion (p<0.05), while it was diminished by both monocyte and PDC depletion
in the case of HSVUV (p<0.01) and HSVINF (p<0.01 and p<0.05, respectively). For HSVINF
stimulation the inhibitory effect of monocyte depletion was significantly stronger than the
effect of PDC depletion (p<0.01), which argues for monocytes to be more important in NK
cell activation than PDC when HSV-1 is infectious. HSVINF-induced NK cell effector
functions were both affected by cell depletion in the same manner. While depletion of PDC
had no effect on either effector function, depletion of monocytes prevented both IFN-
secretion (FIG. 24B) and degranulation (FIG. 24C). These results confirm PDC as important
cell population in NK cell activation by HSV-1, and they furthermore reveal monocytes as
key accessory cells in HSVINF-caused NK cell activation, and as indispensable cell population
for the induction of NK cell effector functions within the PBMC context. Interestingly, both
cell populations seem to be crucial for CpG-A- as well as HSV-1-stimulated IFN-
production (FIG. 24D). Depletion of monocytes as well as PDC reduced secretion of IFN-
induced by CpG-A, HSVUV (p<0.01 for monocyte depletion and p<0.05 for PDC depletion)
and HSVINF (p<0.01).
Results
60
FIG. 24. Monocytes mediate NK cell effector functions upon HSV-1 infection within the PBMC context.
PBMC were left non-depleted (PBMC) or were depleted of monocytes (PBMC monocytes) or of PDC
(PBMC PDC) and stimulated for 12h with CpG-A, UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1,
or left unstimulated (mock) as control. Cells were analyzed by flow cytometry (FACS), supernatants by enzyme-
linked immunosorbent assay (ELISA). A - C. NK cells were gated as CD56-positive CD3- and CD14-negative
population and analyzed for CD69 expression (A), IFN- secretion (B) and CD107a surface expression (C) (%).
D. IFN-2a/2b (IFN-) secretion within PBMC (pg/ml). All values are given as mean and standard error of eight
independent experiments. * p<0.05, ** p<0.01 as indicated (Tukey HSD).
Results
61
5.10 PDC serve as crucial accessory cell population in NK cell
activation by HSV-1-infected HFF
Donaghy et al. demonstrated the presence of PDC within recurrent genital herpes lesions and
their co-localization with NK cells (Donaghy et al., 2009), so we wanted to investigate the
role of PDC as accessory cell population in NK cell activation within infected tissue. In order
to simulate the situation of NK cell activation within tissue, we conducted experiments, in
which we inoculated NK cells with HSVUV and HSVINF and co-cultivated them with human
foreskin fibroblasts (HFF) in the absence and in the presence of PDC. NK cell-PDC ratios in
assays were adjusted to their physiological ratio within PBMC of the respective donors. At
24h p.s. , CD69 up-regulation on NK cells was measured. Clearly, HSV-1 does not activate
NK cells in a direct manner, since NK cells did not up-regulate CD69 in response to either
HSVUV or HSVINF, when cultured alone (FIG. 25A). Stimulation of NK cells with HSVUV and
in particular HSVINF in the presence of PDC led to moderate but not significant CD69
up-regulation. NK cells stimulated with HSV-infected HFF alone also slightly up-regulated
CD69, however, CD69 up-regulation was not significant. In contrast, when co-cultivated with
both HFF and PDC, NK cells stimulated with HSVUV as well as HSVINF significantly up-
regulated CD69 compared to the mock control (p<0.01). Checking IFN- levels, we observed
that PDC secreted considerably more IFN- in the presence of HFF (FIG. 25B). These results
indicate PDC as important accessory cells for NK cell activation within HSV-1-infected
tissue, possibly via secretion of type I IFN, and further suggest a dependence of PDC on a
sufficient cell density, and hence possible interactions with other cells, within the cell culture
to secrete high amounts of IFN- in response to HSV-1 stimulation, as observed by
Rönnblom et al. (Rönnblom et al., 1988).
Results
62
FIG. 25. PDC serve as crucial accessory cell population in NK cell activation by HSV-1-infected HFF.
Purified NK cells were cultivated alone or together with purified PDC, with HFF, or with HFF and PDC, and
stimulated for 24h with UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as
control. Cells were analyzed by flow cytometry (FACS), supernatants by enzyme-linked immunosorbent assay
(ELISA). A. NK cells were gated as CD56-positive CD3- and CD14-negative population and analyzed for CD69
expression (%). B. IFN-2a/2b (IFN-) secretion within cell culture (pg/ml). All values are given as mean and
standard error of three independent experiments. ##
p<0.01 vs. mock (Tukey HSD).
Interestingly, NK cells co-cultivated with HSV-infected HFF expressed less CD56 than
HSV-stimulated NK cells cultured without HFF. The decrease in CD56 expression was even
more obvious, when PDC were present (FIG. 26A). This effect was caused only by HSVINF,
not by HSVUV (FIG. 26B). CD56 expression on NK cells stimulated with HSVINF was
significantly lower compared to mock (p<0.05) and HSVUV (p<0.05). Apparently, HSV
infection of and / or replication within HFF induces NK cells to down-regulate CD56, the
effect being boosted by PDC.
Results
63
FIG. 26. NK cells co-cultivated with HSV-1-infected HFF down-regulate CD56. Purified NK cells were
cultivated alone or together with purified PDC, with HFF, or with HFF and PDC, and stimulated for 24h with
UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as control. NK cells were
gated as CD56-positive CD3- and CD14-negative population and analyzed by flow cytometry (FACS) for CD56
expression (MFI). A. Representative FACS plot of CD56 expression on NK cells after stimulation with HSVINF.
B. Fold change of CD56 expression, given as mean and standard error of three independent experiments.
# p<0.05 vs. mock; * p<0.05 as indicated (Tukey HSD).
Results
64
5.11 PDC supernatants inhibit HSV-1 replication in HFF
Type I IFN are known to lead to an antiviral state of virus-infected and -susceptible cells
(ISAACS and LINDENMANN, 1957). Since PDC are key producers of type I IFN (Siegal et
al., 1999; Cella et al., 1999) and furthermore have been shown to suppress HSV-2 replication
upon vaginal infection (Lund et al., 2006), we decided to examine the potential of PDC-SN to
inhibit HSV-1 replication in HFF. For this purpose, HFF were infected with HSVGFP at an
MOI of 0.001 and 0.01 and cultivated for 24h and 48h in the absence and presence of
PDC-SN corresponding to IFN-2a/2b levels of 20pg/ml and 200pg/ml. Infection rates were
determined via GFP expression in HFF. The first observation we made was the wide range of
infection rates in HFF cultured without PDC-SN (FIG. 28A), varying for MOI 0.001 between
0.1% and 2.6% at 24h and between 0.5% and 82.0% at 48h, for MOI 0.01 between 0.1% and
19.5% at 24h and between 19.3% and 99.4% at 48h p.i. Obviously, productivity of HSV-1
replication depends on the current state and condition of the infected cell, which seems to be
variable for HFF. However, when infected HFF were cultured in the presence of PDC-SN,
infection rates were reduced compared to infection rates in the absence of PDC-SN (FIG. 27).
Relative reduction of infection rates was significant at 24h p.i. for MOI 0.01 at 20pg/ml and
200pg/ml (p<0.05) (FIG. 28B), at 48h p.i. for MOI 0.001 at 200pg/ml (p<0.05) and for MOI
0.01 at 20pg/ml and 200pg/ml (p<0.01) (FIG. 28C). These results evidence the potential of
PDC to directly inhibit HSV-1 replication in target cells via secretion of antiviral cytokines,
most likely type I IFN.
Results
65
FIG. 27. HSV-1 replication in HFF decreases in the presence of PDC supernatants. HFF were infected with
infectious GFP-expressing HSV-1 (HSVGFP) at a MOI of 0.001 and a MOI of 0.01, in the absence of PDC
supernatants (w/o PDC-SN) and in the presence of PDC supernatants containing 20pg/ml and 200pg/ml
IFN-2a/2b (IFN-), or left uninfected (mock) as control, and analyzed by flow cytometry (FACS) for green
fluorescence 24h and 48h post infection (p.i.). Representative FACS plot of green fluorescence (GFP).
Results
66
FIG. 28. PDC supernatants inhibit HSV-1 replication in HFF. HFF were infected with infectious
GFP-expressing HSV-1 (HSVGFP) at a MOI of 0.001 and a MOI of 0.01, in the absence of PDC supernatants
(w/o PDC-SN) and in the presence of PDC supernatants containing 20pg/ml and 200pg/ml IFN-2a/2b (IFN-),
or left uninfected (mock) as control, and analyzed by flow cytometry (FACS) for green fluorescence 24h and
48h post infection (p.i.). A. HSV-1 replication in HFF w/o PDC-SN, shown as green fluorescent HFF (%) 24h
and 48h p.i. B and C. HSV-1 replication in the presence of PDC-SN, shown as fold change of green fluorescent
HFF (%) compared to infection w/o PDC-SN 24h (B) and 48h (C) p.i. All values are given as mean and standard
error of three independent experiments. #
p<0.05, ##
p<0.01 vs. w/o PDC-SN (Tukey HSD).
Results
67
5.12 PDC-NK cell interactions are hampered in an HIV-1-infected
woman suffering from persisting genital ulcers
Human immunodeficiency virus type 1 (HIV-1) infection leads to a decrease in numbers as
well as function of PDC, leading to reduced IFN- secretion (Feldman et al., 2001; Schmidt
et al., 2005; Schmidt et al., 2006). It furthermore causes a defective crosstalk between PDC
and NK cells via functional defects of PDC and also NK cells (Reitano et al., 2009), on which
antiretroviral therapy has only minimal effects (Benlahrech et al., 2011). We therefore
analyzed PDC and NK cell activation in PBMC of an African woman infected with HIV-1
and suffering from immune reconstitution inflammatory syndrome (IRIS). Three months after
viral load decline and CD4+ T cell increase due to successful antiretroviral treatment she
developed painful genital ulcers due to HSV-2, which were only temporarily resolved by
several courses of aciclovir, topical application of imiquimod and a radical bilateral
vulvectomy (Strehl et al., 2012). Repeated virological analysis of the hyperproliferative
lesions revealed human papilloma virus type 54 (HPV-54) infection in addition to HSV-2
infection.
PBMC of this patient were stimulated with CpG-A, a TLR-7 agonist (S-27609), HSVUV and
HSVINF and analyzed at 18h p.s. for IFN- secretion, expression of markers for PDC
migration (CCR7), activation (CD80) and maturation (CD83) as well as NK cell activation
(CD69). Up-regulation of CCR7, CD80 and CD83 on stimulated but also on mock-cultivated
PDC suggested pre-stimulation of PDC in vivo (FIG. 29A). Activation of the patient’s NK
cells upon stimulation was severely impaired (FIG. 29B), as well as IFN- secretion within
PBMC (FIG. 29C), compared to a healthy control donor. These results suggest that impaired
IFN- production by PDC and subsequently reduced activation of NK cells contributed to the
patient’s disease.
Silencing of peripheral IFN- responses in HIV-1 infection has been associated with
enhanced interaction of CD40 on PDC with CD40 ligand (CD40L), a co-stimulatory
molecule, which is up-regulated upon immune activation (Donhauser et al., 2012). CD40L
levels transiently increase with the CD4+ T cell recovery upon antiretroviral therapy, which
might boost IFN-susceptible opportunistic infections in IRIS. Therefore, we retrospectively
analyzed levels of soluble (s)CD40L in the plasma of our patient (P1) and four other patients
Results
68
suffering from opportunistic infections (P2 - P5). Indeed, sCD40L levels in eight consecutive
plasma samples of P1 after initiation of antiretroviral treatment were significantly higher than
in cross-sectional samples of 52 untreated HIV-1-infected patients (p<0.001) (Donhauser et
al., 2012), and they were also higher than in the samples of P2 - P5 (P<0.01) (FIG. 29D).
Thus, suppression of TLR-7- and TLR-9-induced IFN- production by elevated sCD40L
levels may have contributed to the unusual and treatment-refractory genital ulcers in P1.
Alternatively, enhanced sCD40L levels may reflect prior in vivo stimulation caused by HSV-2
and HPV-54 infections. Altogether, these data indicate important interactions of PDC and NK
cells, which are hampered in immunosuppressed individuals and thus may lead to inefficient
control of persistent viral infections, such as human papilloma virus and herpes simplex virus
infections.
Results
69
FIG. 29. PDC-NK cell interactions are hampered in an HIV-1-infected woman suffering from persisting
genital ulcers. A. Surface marker expression on PDC of the investigated patient. PBMC were stimulated for 18h
with CpG-A, S-27609, UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as
control. Cells were analyzed by flow cytometry (FACS) for marker expression (%) indicating PDC migration
(CCR7), activation (CD80) and maturation (CD83) and compared to freshly isolated PBMC (baseline). B and C.
PBMC of the patient and a healthy control donor (control) were stimulated for 18h with CpG-A, S-27609,
UV-inactivated (HSVUV) and infectious (HSVINF) HSV-1, or left unstimulated (mock) as control. NK cells were
gated as CD56-positive CD3- and CD14-negative population and analyzed for CD69 expression (%) by flow
cytometry (FACS) (B), supernatants were analyzed for IFN-2a/2b (IFN-) secretion within PBMC (pg/ml) by
enzyme-linked immunosorbent assay (ELISA) (C). D. Plasma samples of the investigated patient (P1) and four
other IRIS cases (P2 - P5) were analyzed for sCD40L levels (ng/ml) by enzyme-linked immunosorbent assay
(ELISA). ** p<0.01 P2 - P5 vs. P1 (Tukey HSD).
Discussion
70
6 Discussion
Our analysis of human NK cell activation and induction of effector functions by different
stimuli revealed HSV-1 as potent and fast inducer of NK cell activation within the PBMC
context. Interestingly, we observed several differences concerning the stimulating potential
between HSVINF and HSVUV, first, on the level of NK cell activation and second, on the level
of cytokine secretion within PBMC (FIG. 30).
HSVINF induced overall NK cell activation significantly faster than HSVUV (FIG. 6B) and also
caused significantly stronger activation within the CD56bright
NK cell subset (FIG. 7D). In
addition, only HSVINF, and not HSVUV, induced significant NK cell IFN- secretion (FIG. 8B)
and degranulation (FIG. 8C). These results support the finding of Fitzgerald-Bocarsly and
colleagues that HSV-1-inoculated HFF were only lysed by human PBMC when the virus used
was infectious (Fitzgerald-Bocarsly et al., 1991), but stand in contrast to another study by
Ahmad et al., in which NK activity of HSV-1-stimulated human PBMC was similar for
infectious and UV-inactivated virus (Ahmad et al., 2000). This discrepancy could be caused
by the use of different methods. Ahmad et al. investigated the increase of basic lytic activity
against the NK cell target K562, whereas Fitzgerald-Bocarsly et al. analyzed lysis of HFF,
and we detected induction of degranulation in the absence of cytotoxicity-inducing target
cells.
Although both HSVINF and HSVUV caused secretion of exceedingly high amounts of IFN-
within PBMC, HSVINF-induced IFN- secretion was significantly faster and slightly stronger
than HSVUV-induced IFN- secretion (FIG. 9). Furthermore, in our study HSVINF stimulated
significant secretion of IL-1 and TNF-, while HSVUV failed to induce these two cytokines
within PBMC (FIG. 11), contradicting an early study published by Gosselin et al., who saw
similar TNF- induction in human PBMC by infectious and UV-irradiated HSV-1 (Gosselin
et al., 1992). Our observations indicate an effect of viral infectivity on the induction of pro-
inflammatory cytokines as well as on NK cell activation within the PBMC context and
suggest the necessity for viral infectivity in the induction of NK cell effector functions
(FIG. 30), either via induction of different cytokines, or via direct recognition of virus-
infected cells by NK cells as targets, or both.
Discussion
71
FIG. 30. Infectious and UV-inactivated HSV-1 exhibit different stimulation potentials within the PBMC
context. Both infectious and UV-inactivated HSV-1 are able to induce IFN-2a/2b (IFN-) secretion within
PBMC and CD69 up-regulation on NK cells, whereas only infectious HSV-1 causes secretion of TNF- and
IL-1 within PBMC and NK cell effector functions IFN- secretion and degranulation indicated by CD107a
surface expression. These findings suggest the importance of viral infectivity in complete activation of effector
NK cells by HSV-1.
NK cell effector functions were evenly distributed between the two NK cell subsets
(FIG. 8A), contradicting a previous concept of strict classification of CD56dim
and CD56bright
NK cells into a mainly cytotoxic and a major cytokine secreting subset, respectively (Cooper
et al., 2001a). However, Vivier proposed to rather define the CD56dim
and CD56bright
NK cell
subsets as “target cell responsive” and “cytokine responsive”, respectively, both possessing
the ability for cytotoxicity as well as cytokine secretion, depending on the stimulus (Vivier,
2006). Recently, De Maria et al. described CD56dim
NK cells as rapid producers of IFN-
upon antibody-mediated stimulation of natural killer receptors (De et al., 2011). We report
Discussion
72
here for the first time that infectious HSV-1 is a potent stimulus for IFN- secretion by
CD56dim
NK cells within the PBMC context.
INF- has been published to be of importance in HSV-induced NK cell activation (Gill et al.,
2011; Feldman et al., 1992) and several groups identified it as main cytokine in the induction
of NK cell activation after stimulation of human PDC with influenza virus, CpG, and poly
(I:C) (Benlahrech et al., 2009; Gerosa et al., 2005; Marshall et al., 2006; Romagnani et al.,
2005). Stimulation of purified NK cells with supernatants of HSVINF-stimulated PDC
(PDC-SN) demonstrated time- and dose-dependent induction of CD69 on NK cells by
PDC-SN (FIG. 10A). Interestingly, NK cell activation occurred in two phases, suggesting that
the initial and subsequent CD69 up-regulation were induced by two different mechanisms.
Further studies are required to identify the underlying mechanisms. The fact that
neutralization of the IFN-/ receptor significantly decreased PDC-SN-induced NK cell
activation (FIG. 10C) proves type I IFN as key cytokines in PDC-induced NK cell activation
after stimulation with HSV-1 (FIG. 31), consistent with PDC-induced NK cell activation after
stimulation with CpG-A (Benlahrech et al., 2009; Gerosa et al., 2005; Marshall et al., 2006;
Romagnani et al., 2005). However, NK cell activation by PDC-SN was slightly stronger than
NK cell activation by rhIFN- (FIG. 10B), and other groups described the involvement of
further cytokines, in particular TNF-, which collaborated with IFN- in PDC-induced NK
cell activation (Gerosa et al., 2005; Marshall et al., 2006; Romagnani et al., 2005). Actually,
we detected TNF- not only in PDC supernatants (data not shown), but also in PBMC
supernatants (FIG. 11).
Discussion
73
FIG. 31. PDC-dependent NK cell activation by HSV-1 is mediated by type I IFN. HSV-1-stimulated PDC
secrete high amounts of IFN-2a/2b (IFN-) and other type I IFN. Stimulation of purified NK cells with PDC
supernatants leads to CD69 up-regulation which is inhibited by a neutralizing antibody against the IFN-/
receptor. This identifies type I IFN as key cytokines in PDC-dependent NK cell activation by HSV-1.
TNF- has been shown to play an essential role in HSV infection in vivo. TNF- knockout
mice exhibited decreased survival rates in acute corneal HSV-1 infections and increased
reactivation rates after UV light stimulation (Minami et al., 2002), and lethal encephalitis after
intranasal HSV-1 infection (Sergerie et al., 2007). In our studies we demonstrated TNF- as
critical cytokine for CpG-A- and HSV-1-induced human NK cell activation (FIG. 12A) and
also for HSVINF-caused IFN- secretion (FIG. 12B) within the PBMC context. Cooper et al.
described the ability of IL-1 to co-stimulate IFN- production of CD56bright
NK cells together
with IL-12 or, in particular, IL-15 (Cooper et al., 2001b), however, we did not detect any
direct influence of IL-1 on NK cell activation or induction of effector functions
(FIG. 12A, B, C).
Type I IFN, which played a major role in PDC-mediated NK cell activation (FIG. 10),
appeared to be less important within the PBMC context. IFN-R neutralization only
significantly inhibited NK cell activation induced by CpG-A and HSVUV, but not by HSVINF
(FIG. 12A), and had no influence on degranulation (FIG. 12C). On the contrary, HSVINF-
induced IFN- secretion was increased by IFN-R neutralization (FIG. 12B), suggesting a
rather inhibitory influence of high amounts of type I IFN on NK cell IFN- secretion.
Similarly, Cousens et al. observed inhibition of murine IL-12 and subsequently IFN-
Discussion
74
secretion by type I IFN upon viral infection and bacterial stimulation (Cousens et al., 1997).
IFN- secretion after simultaneous neutralization of TNF- and IFN-R (FIG. 13B) points to
opposed functions of TNF- and type I IFN in IFN- induction by HSVINF, namely that
TNF- induces, and type I IFN rather inhibit HSV-1-induced IFN- production by NK cells.
We could observe a strict dependence of IFN- levels on type I IFN upon stimulation of
PBMC with GpG-A and HSV-1 (FIG. 12D) and also upon infection of monocytes with
HSV-1 (FIG. 21D), which is in concordance with the already published autocrine loop (Marie
et al., 1998). Interestingly, TNF- and IL-1 also strongly influenced IFN- secretion
(FIG. 12D). This suggests that secretion of high amounts of IFN- demands a positive
feedback loop consisting of a tight crosstalk of IFN--producing cells with each other and / or
other immune cells via production of type I IFN, TNF-, and IL-1 (FIG 32). Actually,
TNF- has been shown to induce secretion of low amounts of type I IFN, particularly IFN-,
in both human and mouse macrophages (Yarilina et al., 2008). Jimbo et al. demonstrated
IL-1 to be involved in a positive feedback loop increasing its own secretion by intervertebral
disc cells and also secretion of other inflammatory mediators like IL-6 and cyclooxygenase
(COX)-2 (Jimbo et al., 2005).
Discussion
75
FIG. 32. Secretion of high amounts of IFN-2a/2b (IFN-) within PBMC demands a positive feedback
loop involving type I IFN, TNF- and IL-1. HSV-1-induced secretion of IFN- is greatly diminished by
neutralization of the IFN-/ receptor as well as TNF- and IL-1, suggesting that not only type I IFN, but also
TNF- and IL-1 are involved in a positive feedback loop leading to secretion of high IFN-2a/2b (IFN-)
amounts after HSV-1 simulation of PBMC.
TNF- secretion assays revealed PDC and monocytes as potent TNF- sources upon
stimulation with HSVINF (FIG. 14A). Considering the much higher frequency of monocytes
within PBMC (FIG. 14B), monocytes have to be regarded as the most numerous TNF-
producers in the blood. Obviously, TNF- secretion by PDC did not depend on viral
infectivity, in contrast to monocytic TNF- secretion, which was significantly increased by
viral infectivity (FIG. 14A). Interestingly, with 16% the percentage of TNF--secreting
monocytes (FIG. 14A) was clearly higher than the percentage of infected monocytes of 2%
(FIG. 16A), which argues against TNF- secretion mainly by infected monocytes. It rather
suggests that infection of few monocytes stimulates a number of uninfected bystander
monocytes to secrete TNF-. The difference in the induction of TNF- secretion that we
observed between HSVINF and HSVUV on the PBMC level (FIG. 14C) could explain why
HSVUV was unable to mediate significant IFN- secretion by NK cells; neutralization of
TNF- significantly diminished NK cell IFN- secretion induced by HSVINF (FIG. 12B).
Discussion
76
These observations indicate that induction of certain NK cell effector functions, like IFN-
secretion, requires the secretion of pro-inflammatory cytokines within PBMC, which in turn
depends on viral infectivity. Induction of other NK cell effector functions, like cytotoxicity
indicated by degranulation, is independent at least from the cytokines investigated in our
study, but also depends on viral infectivity.
Ahmad et al. reported IL-15 as crucial cytokine in HSV-1-induced NK activity of human
PBMC, but in their study, infectivity of HSV-1 was not required for the induction of NK
activity (Ahmad et al., 2000), and in a later study Ahmad et al. showed that HSV-1-induced
up-regulation of IL-15 gene expression in monocytic cells was independent of viral infectivity
(Ahmad et al., 2007). We tested PBMC supernatants for secreted IL-15, but did not detect any
after HSV-1 stimulation (data not shown). Furthermore, the observations in the two other
studies would preclude IL-15 as critical factor for NK cell effector functions, since in our
studies only HSVINF induced NK cell effector functions, whereas HSVUV failed to induce
them (FIG 8). IL-18, which we did not investigate, might play a role in HSVINF-induced NK
cell activation and effector functions, since it has already been demonstrated to contribute to
NK cell activation in HSV-1-infected mice (Barr et al., 2007; Reading et al., 2007).
Infection experiments evidenced monocytes as target cells for HSV-1 which are infected, yet
to a very small percentage and without allowing productive viral replication (FIG. 16)
(FIG. 33), in concordance to prior studies of monocyte infection by HSV-1 (Bruun et al.,
1998; Daniels et al., 1978). Ineffective infection of and replication in monocytes was in part
due to monocytic type I IFN secretion, since IFN-R blocking significantly increased
infection rates in purified cells, in particular 48h p.i. (FIG. 21A).
Similarly, type I IFN were published to suppress HSV-1 replication in vitro in Vero cells,
HEp-2 cells, and fibroblasts (Härle et al., 2001; Noisakran et al., 2000) and to play a crucial
role in resistance to HSV infections in vivo (Dupuis et al., 2003; Casrouge et al., 2006).
Another restriction factor for HSV-1 replication in monocytes might be SAM domain and HD
domain-containing protein 1 (SAMHD1), which was recently shown to inhibit HSV-1
replication in differentiated macrophage cell lines (Kim et al., 2013).
Interestingly, IFN- secretion by monocytes depended on HSV-1 infectivity (FIG. 19), in
concordance to the observation of Melchjorsen et al., that cytokine induction by HSV in
Discussion
77
human monocyte-derived cells is dependent on virus replication (Melchjorsen et al., 2006).
This stands in contrast to IFN- secretion by PDC, which is induced by HSVINF as well as
HSVUV (Schuster et al., 2010) (FIG. 9). This difference is probably due to the fact that HSV-1
is able to infect monocytes (FIG. 16A) (Bruun et al., 1998; Daniels et al., 1978), but not PDC
(Schuster et al., 2010), and due to different recognition molecules involved; TLR-9 is
responsible for HSV-1 recognition in PDC (Krug et al., 2004), but not in monocytes, where
melanoma differentiation-associated protein 5 (MDA5) was shown to be the primary mediator
of HSV-1 recognition in macrophages (Melchjorsen et al., 2010). MDA5, a retinoic acid-
inducible gene (RIG)-I-related protein, senses viral RNA with a helicase domain and mediates
the induction of an antiviral response within the infected cell (Yoneyama et al., 2005). Since
the presence of herpesviral RNA requires the initiation of viral replication and therefore viral
infectivity, UV-inactivated HSV-1 should not be recognized by MDA5, explaining the lack of
IFN- induction in monocytes by HSVUV (FIG. 19). Interestingly, HSVd106S induced stronger
IFN- secretion than both infectious wildtype isolates (FIG. 22D). This might be due to a
deletion within the ICP27 gene of HSVd106S (Liu et al., 2009). Melchjorsen et al. determined
ICP27 as a factor counteracting cytokine induction in monocyte-derived cells by HSV
(Melchjorsen et al., 2006).
Discussion
78
FIG. 33. Monocytes are non-productively infected by HSV-1 and up-regulate MHC-I in a type I IFN-
dependent manner. HSV-1 is able to infect monocytes, but without production of new viral particles. HSV-1-
infected monocytes secrete low amounts of IFN-2a/2b (IFN-). Neutralization of the IFN-/ receptor
prevents MHC-I up-regulation induced by infectious HSV-1 (HSVINF) and increases infection rates in
monocytes. These data demonstrate non-productive infection of monocytes by HSV-1 leading to type I IFN-
dependent up-regulation of MHC-I. Furthermore, type I IFN prove to be involved in preventing productive
HSV-1 infection of and replication in monocytes.
Depletion experiments confirmed PDC as crucial IFN- source (FIG. 24D), and furthermore
as potent mediators of CpG-A- and HSV-induced NK cell activation (FIG. 24A) within the
PBMC context. PDC-induced NK cell activation was at least in part due to IFN- production
(FIG. 10, FIG. 12A, FIG. 24D). However, NK cell effector functions did not depend on PDC
(FIG. 24B, C), which stands in contrast to an early study, that indicated a supporting role for
the so called “IFN-producing cells (IPC)” in NK cell-mediated lysis of HSV-1-infected
fibroblasts (Feldman et al., 1992). In addition, our studies showed the importance of
monocytes in NK cell activation (FIG. 24A) and also in IFN- secretion (FIG. 24D).
Monocytes may account for high IFN- levels in different ways. They could contribute
directly by secretion of IFN- itself, as observed for infected monocytes by us (FIG. 19) and
also by others (Linnavuori and Hovi, 1983). However, isolated monocytes only reacted with
IFN- production to infectious, not to UV-inactivated HSV-1, while depletion of monocytes
from PBMC diminished IFN- levels upon stimulation with both HSVINF and HSVUV.
Another possibility would be an indirect contribution of monocytes via secretion of IL-1 and
TNF-, thereby further stimulating CpG-A- and HSV-1-induced IFN- secretion by PDC.
Cytokine neutralization experiments (FIG. 12D) suggest this way of monocyte involvement in
Discussion
79
IFN- production. Furthermore, early studies suggested the dependence of an IFN- response
to HSV-1 on close contact and interactions of IFN-producing cells with other cells within the
cell culture (Rönnblom et al., 1988) and the potential of PBMC-derived cytokines to enhance
HSV-1-induced IFN- secretion by IFN-producing cells (Cederblad and Alm, 1990). Also,
Megjugorac et al. investigated interactions between PDC and HSV-infected monocyte-
derived (mo) DC and could demonstrate induction of IFN- secretion from PDC by HSV-
infected moDC (Megjugorac et al., 2007). Most importantly, we identified monocytes as
indispensable cell population in the induction of NK cell effector functions by HSVINF within
the PBMC context (FIG. 24B, C). Our findings within PBMC appear similar to a study of
PDC-induced NK cell activation, in which NK cell CD69 up-regulation and IFN- production
were induced by soluble factors, whereas degranulation and cytotoxicity were only observed
after direct contact with CpG-stimulated PDC (Benlahrech et al., 2009). In contrast to
Benlahrech et al. we did not find IFN- as major soluble factor for NK cell IFN- production,
but could determine TNF- secretion as important mechanism in the induction of IFN-
(FIG. 12B), which is probably due to the fact that our analyses were conducted with whole
PBMC, not purified PDC and NK cells. The exact process, in which degranulation was
induced, remained elusive.
It is very possible that infected monocytes are directly recognized by NK cells as target cells.
NK cell activation and induction of effector functions could be mediated through various
possible mechanisms. Induction of NK cell cytotoxicity via down-regulation of HLA-C
molecules on productively infected cells was demonstrated for both HSV-1 and HSV-2
(Elboim et al., 2013; Huard and Früh, 2000). Yet, in our studies we observed an overall up-
regulation of HLA-ABC and HLA-E on monocytes inoculated with HSV-1 (FIG. 17,
FIG. 18). MHC-I up-regulation was due to monocytic IFN- secretion (FIG. 19), as
demonstrated by neutralization experiments (FIG. 20, FIG. 21B, C). Unfortunately, we were
not able to investigate HLA-A, HLA-B and HLA-C separately, because no specific antibodies
are available. But the fact that only a minority of HSVGFP-infected monocytes showed
decreased MHC-I expression (FIG. 18) makes recognition of infected monocytes via MHC-I
down-regulation very unlikely. MICA has been shown to be up-regulated on TLR-stimulated
monocytes (Kloss et al., 2008), but MICA/MICB was not induced by HSV-1 in our
Discussion
80
experiments (FIG. 23). Thus, we could exclude direct NK cell activation by infected
monocytes via expression of these stress molecules.
Fitzgerald-Bocarsly et al. demonstrated the expression of immediate early genes as sufficient
to induce NK cell-mediated lysis of HSV-1-infected fibroblasts (Fitzgerald-Bocarsly et al.,
1991), and Chisholm et al. identified ICP0 as effective to trigger lysis of HSV-1-infected cells
by NK cells via the natural cytotoxicity receptors (NCR) NKp30, NKp44, and NKp46
(Chisholm et al., 2007). However, the molecules induced by ICP0 and serving as ligands to
the NCR were not identified in this study. A possible candidate might be B7H6, a molecule
expressed on tumor cells that triggers NK cell cytotoxicity and cytokine secretion via
interaction with NKp30 (Brandt et al., 2009), which was shown to be induced on human
monocytes upon stimulation with TLR ligands and pro-inflammatory cytokines such as IL-1
and TNF- (Matta et al., 2013). Other molecules involved in NK cell-monocyte/macrophage
interaction could be macrophage-expressed CD48 and NK cell-expressed 2B4, which are
involved in NK cell activation after LPS stimulation (Nedvetzki et al., 2007), or macrophage-
expressed activation-induced C-type lectin (AICL) and NK cell-expressed NKp80, which are
involved in NK cell activation after TLR ligand stimulation (Welte et al., 2006), or the
glucocorticoid-induced tumor necrosis factor receptor-ligand (GITRL), which was described
to be involved in the induction of NK cell cytotoxicity by CpG-stimulated PDC (Hanabuchi et
al., 2006) and was shown to be induced on monocytes by staphylococcal enterotoxin B
(Cardona et al., 2006).
While PDC did not mediate NK cell effector functions within the PBMC context (FIG. 24B,
C), they influenced NK cell activation (FIG. 24A). Outside the PBMC context they were
indispensable for NK cell activation, since HSV-1 did not activate purified NK cells directly
(FIG. 25A) (FIG. 34). This observation stands in contrast to a study by Kim et al., where HSV
glycoprotein (g)D peptides directly activated NK cells (Kim et al., 2012). In our co-culture
experiments with HSV-exposed HFF and PDC, neither cell type induced significant CD69 up-
regulation on NK cells by its own. Only in combination, HSV-exposed HFF and PDC
succeeded to strongly activate NK cells (FIG. 25A), which was either caused by synergistic
effects of both cell types on NK cells or mediated by IFN-, which was produced by PDC in
high amounts only in the presence of HFF (FIG. 25B). Dependence of high IFN- production
Discussion
81
by PDC on co-culture with HFF might, similar to IFN- production within PBMC
(FIG. 24D), be due to the need of PDC for close contact to and possible interactions with
other cell populations (Rönnblom et al., 1988). Interestingly, HSV-1-infected HFF somehow
caused down-regulation of CD56 on NK cells (FIG 34), with the effect being enhanced by
PDC (FIG. 26). The significance of this finding remains unclear.
FIG. 34. NK cell activation by HSV-1-infected HFF depends on PDC. HSV-1 does not activate NK cells in a
direct manner, but demands the presence of cells that are stimulated (PDC) or infected (HFF) by HSV-1. HSV-1-
infected HFF induce significant CD69 up-regulation on NK cells only in the presence of PDC, possibly in a type
I IFN-dependent manner, suggesting PDC as important accessory cells for NK cell activation within HSV-1-
infected tissue. Interestingly, NK cells stimulated with HSV-1-infected HFF down-regulate CD56, particularly in
the presence of PDC.
Besides influencing immune responses to HSV-1, PDC might also play a crucial role in
HSV-1 infection by inhibiting viral replication in HSV-1-susceptible cells via secretion of
antiviral cytokines, thereby limiting spread of virus progeny and protecting tissue from
immense damage. We observed an inhibitory effect of PDC-SN in infection experiments with
HFF (FIG. 35). Addition of PDC-SN to HSVGFP-infected HFF clearly decreased HSV-1
replication, evident from reduced green fluorescence of HFF (FIG. 27, FIG. 28B, C).
Discussion
82
Inhibition of HSV-1 replication was most likely mediated by type I IFN, as observed in
monocytes (FIG. 21A), and also demonstrated by others (Härle et al., 2001; Noisakran et al.,
2000).
FIG. 35. PDC supernatants inhibit HSV-1 replication in HFF. Viral infection of and replication in HFF are
diminished in the presence of supernatants derived from HSV-1-stimulated PDC. The inhibitory effect of PDC
supernatants is possibly mediated by type I IFN. These findings evidence the role of PDC as suppressors of
spread of HSV-1 infection within tissue by suppressing viral replication in HSV-1-susceptible cells via secretion
of antiviral cytokines.
Studies of stimulation of PDC and NK cells from HIV-1-infected individuals propose a role
for defective PDC-NK cell interactions in HIV-1-induced immune suppression (Conry et al.,
2009; Reitano et al., 2009), allowing opportunistic or IRIS-related infections. In our case
study, an HIV-1-infected patient suffering from vaginal hyperproliferative lesions due to
HSV-2 and HPV-54 infections exhibited severe functional deficits of PDC as well as NK cells
within the PBMC context. NK cells were only minimally activated by HSV-1 (FIG. 29B)
which appeared to be due to impaired IFN- secretion by PDC (FIG. 29C). Impaired IFN-
secretion upon stimulation with HSV-1 and TLR-7 and TLR-9 agonists might have been
caused by increased CD40-CD40L interactions, as demonstrated by Donhauser et al.
(Donhauser et al., 2012), since sCD40L levels were elevated in the patient (FIG. 29D). Our
findings suggest a role for impaired PDC-NK cell interactions in the severe and treatment-
refractory course of disease in the patient, emphasizing the importance of PDC-NK cell
crosstalk for efficient control of herpesviral infections.
Discussion
83
Altogether, our data propose a model in which the induction of high IFN- levels by HSV-1
within PBMC demands a tight crosstalk between PDC and monocytes involving a positive
feedback loop influenced by type I IFN, TNF- and IL-1 (FIG. 32). Secretion of IL-1 does
not directly influence NK cells, whereas type I IFN and TNF- secreted by both PDC and
monocytes mediate NK cell activation (FIG. 36). IFN- secretion and NK cell activation
alone do not depend on HSV-1 infectivity, whereas only HSVINF, not HSVUV, further induces
NK cell INF- secretion as well as degranulation (FIG. 30). Monocytes, in contrast to PDC,
play a key role in the induction of both NK cell effector functions. HSV-1-induced IFN-
secretion by NK cells is independent of type I IFN, but involves TNF- (FIG. 36), which is
only produced in sufficient amounts in response to HSVINF, not HSVUV. In contrast, NK cell
degranulation is independent of all three cytokines tested and either involves other cytokines
produced by monocytes or is mediated by direct cell:cell interactions between NK cells and
monocytes (FIG. 36). Presumably, NK cells recognize HSV-1-infected monocytes as target
cells via mechanisms other than monocytic MHC-I down-regulation or MICA/MICB
expression. While monocytes are particularly important in the activation of effector NK cells,
PDC appear to contribute to immune control early during infection by protecting HSV-1-
susceptible tissue as they suppress viral replication via secreted antiviral cytokines, probably
type I IFN, and therefore limit viral spread (FIG. 36).
Our data may stimulate further studies investigating cell surface molecules as well as
cytokines involved in the crosstalk between PDC, monocytes and NK cells. Deciphering the
mechanisms that induce functional effector NK cells is important as all three cell types are
among the first cells to infiltrate herpetic lesions and thereby may contribute to the efficient
control of primary and recurrent herpes simplex virus infections.
Discussion
84
FIG. 36. Monocytes mediate HSV-1-induced activation of effector NK cells, while PDC limit HSV-1
replication within infected tissue. While depletion of PDC as well as monocytes greatly diminishes secretion of
IFN-2a/2b (IFN-) and CD69 up-regulation on NK cells, only depletion of monocytes prevents NK cell
effector functions IFN- secretion and degranulation, identifying monocytes as crucial accessory cell population
for HSV-1-induced activation of effector NK cells within PBMC. Both type I IFN and TNF- are involved in
CD69 up-regulation, whereas only TNF- impacts IFN- secretion. Degranulation is independent of type I IFN,
TNF- and IL-1 and is possibly mediated by infected monocytes via direct cell contact. The ability of PDC
supernatants to inhibit HSV-1 replication in HFF proves PDC as important cell population in the limitation of
spread of infection within tissue via secretion of antiviral cytokines.
Abbreviations
85
7 Abbreviations
Abbreviation Full length spelling
AC accessory cell(s)
ADCC antibody-dependent cellular cytotoxicity
AICL activation-induced C-type lectin
APC allophycocyanin
ARN acute retinal necrosis
B bone marrow-derived
BDCA blood dendritic cell antigen
BSA bovine serum albumin
C Celsius
CCL chemokine (C-C motif) ligand
CD cluster of differentiation
cm centimeter
CXCL chemokine (C-X-C motif) ligand
Cy cyanine
D day(s)
DMEM Dulbecco`s Modified Eagle Medium
DNA deoxyribonucleic acid
DNAM DNAX accessory molecule
DPBS phosphate buffered saline without calcium or magnesium