of April 6, 2018.This information is current as

PhenotypeHuman Trigeminal Ganglia Have an APC Neuron-Interacting Satellite Glial Cells in

G. M. VerjansAngelique Poot, Albert D. M. E. Osterhaus and Georges M. Monique van Velzen, Jon D. Laman, Alex KleinJan,

http://www.jimmunol.org/content/183/4/2456doi: 10.4049/jimmunol.0900890July 2009;

2009; 183:2456-2461; Prepublished online 27J Immunol 

Referenceshttp://www.jimmunol.org/content/183/4/2456.full#ref-list-1

, 17 of which you can access for free at: cites 44 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2009 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Neuron-Interacting Satellite Glial Cells in Human TrigeminalGanglia Have an APC Phenotype1

Monique van Velzen,* Jon D. Laman,†‡ Alex KleinJan,§ Angelique Poot,*Albert D. M. E. Osterhaus,* and Georges M. G. M. Verjans2*

Satellite glial cells (SGC) in sensory ganglia tightly envelop the neuronal cell body to form discrete anatomical units. This type ofglial cell is considered neuroectoderm-derived and provides physical support to neuron somata. There are scattered hints in theliterature suggesting that SGC have an immune-related function within sensory ganglia. In this study, we addressed the hypothesisthat SGC are tissue-resident APC. The immune phenotype and function of a large series (n � 40) of human trigeminal ganglia(TG) were assessed by detailed flow cytometry, in situ analyses, and functional in vitro assays. Human TG-resident SGC (TG-SGC) uniformly expressed the common leukocyte marker CD45, albeit at lower levels compared with infiltrating T cells, and themacrophage markers CD14, CD68, and CD11b. In addition, TG-SGC expressed the myeloid dendritic cell (DC) marker CD11c,the T cell costimulatory molecules CD40, CD54, CD80, and CD86 and MHC class II. However, the mature DC marker CD83 wasabsent on TG-SGC. Functionally, TG-SGC phagocytosed fluorescent bacteria, but were unable to induce an allogeneic MLR.Finally, TG-infiltrating T cells expressed the T cell inhibitory molecules CD94/NKG2A and PD-1, and the interacting TG-SGCexpressed the cognate ligands HLA-E and PD-L1, respectively. In conclusion, the data demonstrate that human TG-SGC have aunique leukocyte phenotype, with features of both macrophages and immature myeloid DC, indicating that they have a role asTG-resident APC with potential T cell modulatory properties. The Journal of Immunology, 2009, 183: 2456–2461.

S ensory ganglia are part of the peripheral nervous system.They contain cell bodies of sensory neurons establishingthe connection between the periphery and CNS. Sensory

ganglia lack a blood-nerve barrier and enclose a high number ofsatellite glial cells (SGC)3 (1–3). SGC are considered to be neu-roectoderm-derived and involved in the maintenance of sensoryneuron homeostasis by regulating extracellular ion and nutrientlevels within sensory ganglia (2). In contrast to CNS-resident glialcells, like astrocytes and microglia, SGC have a distinct interactionwith neurons (2, 3). They directly associate with the neuronalsoma, so that each neuronal cell body is completely surrounded bya sheet of several SGC providing physical support and a protectivebarrier (3). The numerous fine invaginations between the neuronand SGC sheath illustrate their intimate association (2, 3). Uponmechanical injury to sensory neurons, SGC undergo morphologi-cal changes, proliferate, and up-regulate a variety of growth fac-tors, cytokines, and the glial marker glial fibrillary acidic protein(2, 4, 5).

Human �-herpesviruses, like HSV, are a common threat tohuman sensory ganglia. HSV establishes a lifelong latent infec-tion in neurons within sensory ganglia, predominantly the tri-

geminal ganglion (TG), and reactivates intermittently (6). Re-cent studies in mice and humans emphasized the importance ofinfiltrating T cells to control latent HSV infections in sensoryganglia (7–9). Virus-specific T cells are directly juxtaposed tolatently infected neurons, produce cytokines and cytolytic ef-fector molecules, but do not induce neuronal damage (7, 8,10 –12). Current data suggest that the neurons themselves orhitherto unrecognized resident cells in latently infected sensoryganglia induce and coordinate this nonpathogenic chronic T cellresponse (8, 10 –12).

In this study, we addressed the hypothesis that SGC are tissue-resident APC. The availability of a series of fresh postmortemhuman TG specimens enabled us to combine ex vivo and in situanalyses for the phenotypic and functional characterization of hu-man TG-resident SGC (TG-SGC).

Materials and MethodsClinical specimens

Heparinized peripheral blood and TG specimens, i.e., left and right TG,were obtained from 40 subjects (median age 79 years, range 41–94years) at autopsy with a mean postmortem interval of 6 h (range 2.5–15.5 h). The TG tissue panel consisted of 34 donors with a CNS disease(mainly Alzheimer’s disease and Parkinson’s disease) and six donorswithout evidence of CNS disease. The cause of death was not related to�-herpesvirus infections. No significant differences in the immunolog-ical parameters analyzed were detected between donors with or withouta history of CNS disease (data not shown). Specimens were either snap-frozen (n � 23) or transferred to tubes (n � 17) containing culturemedium consisting of RPMI 1640 (Lonza) supplemented with heat-inactivated 10% FBS (Greiner) and antibiotics. Written informed con-sent from the donor or next of kin was obtained. The local ethicalcommittees approved the study, which was conducted according to thetenets of the Declaration of Helsinki.

Generation of TG single cell suspensions

Generation of single cell suspensions from human TG was performed es-sentially as previously described (12). In brief, the TG were fragmentedand subsequently treated with Liberase Blendzyme 3 (0.2 U/ml, Roche) at37°C for 1 h. Dispersed cells were filtered through a 70-�m pore size cell

*Department of Virology, †Department of Immunology, ‡MS Center ErasMS, and§Pulmonary Medicine, Erasmus Medical Center, Rotterdam, The Netherlands

Received for publication May 26, 2009. Accepted for publication June 9, 2009.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This study was supported in part by the International Consortium on Anti-Virals (toM.v.V.) and the Dutch MS Research Foundation (to J.D.L.).2 Address correspondence and reprint requests to Dr. Georges M.G.M. Verjans, De-partment of Virology, Room Ee1720a, Erasmus Medical Center, s-Gravendijkwal230, 3015 CE Rotterdam, the Netherlands. E-mail address: g.verjans@erasmusmc.nl3 Abbreviations used in this paper: SGC, satellite glial cell; DC, dendritic cell; TG,trigeminal ganglia; PD, programmed death; PD-L1, PD ligand l.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0900890

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

strainer (BD Biosciences), and the flow-through was collected in PBScontaining 1% FBS. From the same donor, PBMC were isolated fromheparinized peripheral blood (�4 ml per donor) by density gradientcentrifugation on Ficoll-Hypaque (12). Donor PBMC and TG single cellsuspensions were directly used for phenotypic and functional analyses.

Flow cytometry

Donor-matched PBMC and TG cells were subjected to multicolor flow cyto-metric analyses using the following fluorochrome-conjugated mAbs: CD3-allophycocyanin (UCHT1; DakoCytomation), CD11b-PE (Bear-1; BeckmanCoulter), CD11c-allophycocyanin (S-HCL3; BD Biosciences), CD14-FITC(TUK4; DakoCytomation), CD40-FITC (5C3; BD Biosciences), CD45-PerCP(2D1; BD Biosciences), CD54-FITC (6.5B5; DakoCytomation), CD68-PE(Y1/82A; BD Biosciences), CD80-FITC (MAB104; Beckman Coulter),CD83-allophycocyanin (HB15e; BD Biosciences), CD86-PE (FUN-1;BD Biosciences), HLA-DR-PerCP (L243; BD Biosciences), CD94-FITC (DX22; eBioscience), NKG2A-allophycocyanin (131411; eBio-science), programmed death (PD)-1-PE (MIH4; eBioscience), and PDligand 1 (PD-L1)-PE (MIH1; eBioscience). Cells were labeled accord-ing to the manufacturers’ instructions and appropriate isotype- and flu-orochrome-matched unrelated mAbs were included as negative con-trols. Cells and data were analyzed on a BD FACSCalibur flowcytometer and BD CellQuest Pro software (BD Biosciences).

In situ analyses

Snap-frozen TG were embedded in Tissue Tek OCT compound (Sakura)and cut into 6-�m sections on a Leica CM 3050S cryostat. Sections werefixed in acetone for 10 min and incubated with the following unconjugatedmAbs according to the manufacturer’s instructions: CD11b (ICRF44; BDBiosciences), CD11c (B-ly6; BD Biosciences), CD14 (TUK4; DakoCyto-mation), CD16 (3G8; BD Biosciences), CD40 (5D12; Pangenetics), CD45(2B11�PD7/26; DakoCytomation), CD54 (LB-2; BD Biosciences), CD64(32.2; DakoCytomation), CD68 (KP1; DakoCytomation), CD80 (M24; In-nogenetics), CD83 (Hb15a; Beckman Coulter), CD86 (1G10; Pangenetics),CD94 (HP-3B1; Immunotech), HLA-E (4D12) by gift from D. E. Geraghty(Fred Hutchinson Cancer Research Center, Seattle, WA), PD-1 (MIH4;eBioscience), and PD-L1 (MIH1; eBioscience). Primary mAb were visualizedusing the avidin-biotin system (DakoCytomation) and AEC (3-amino-9-eth-ylcarbazole; Sigma-Aldrich) as substrate, and sections were counterstainedwith hematoxylin (Sigma-Aldrich), examined under a Zeiss Axioskop, andphotographed using a Nikon DC-U1 camera. For each donor and each marker,three sections and three fields per section were analyzed. Human tonsil sec-tions were used as positive control tissue, and appropriate isotype and conju-gate-negative control stainings were included.

For double stainings, sections were fixed in acetone, and endogenousperoxidase activity and endogenous biotin were blocked before incubationwith the first primary Ab CD14 (TUK4) or CD45 (2B11�PD7/26). Thefirst mAb was detected using an avidin-biotin-HRP system (Biogenex).Before substrate incubation, sections were incubated with normal mouseserum (10%) and a CD11c-PE mAb (B-ly6), which was visualized using ananti-PE secondary Ab (AbD Serotec) and an alkaline phosphatase-conju-gated tertiary Ab (Sigma-Aldrich). Slides were first developed with Fastblue substrate, followed by incubation with AEC substrate solution.

Enrichment of peripheral blood- and TG-derived cellpopulations

Monocytes and TG-SGC were isolated using anti-CD14 microbeads and aMACS magnetic separator (Miltenyi Biotec) according to the manufactur-er’s instructions. T cells were isolated from PBMC of healthy blood donorsusing anti-CD3 microbeads (Miltenyi Biotec). Flow cytometry confirmedthat the enriched cell fractions contained �85% CD14� cells and �95%CD3� cells, respectively (data not shown).

Phagocytosis assay

TG single cell suspensions were incubated with fluorescein-labeledEscherichia coli K-12 strain bioparticles (Invitrogen) in a cell-to-par-ticle ratio of 1:100 according to the manufacturer’s instructions. Afterincubation at 37°C for 2 h, cells were washed extensively and subjectedto flow cytometry or used for immunocytological analyses. For the lat-ter procedure, E. coli-treated TG-SGC were enriched using anti-CD14beads, spun down onto glass slides, fixed with 4% paraformaldehydeand stained with Alexa Fluor 610-PE-conjugated anti-CD68 mAb (KP1;DakoCytomation). Cytospins were mounted in ProLong Gold anti-fadereagent with DAPI (4�,6-diamidino-2-phenylindole; Invitrogen) and an-alyzed on a confocal laser-scanning microscope (LSM510 Meta; Zeiss).Pictures were made using multitrack recording with a 405 nm diode,

488 nm argon, and 561 nm diode laser to detect DAPI, fluorescein, andAlexa Fluor 610-PE, respectively.

Allogeneic MLR assay

CD14-enriched TG-SGC, peripheral blood-derived monocytes and maturedendritic cells (DC) were used as stimulator cells in allogeneic MLR as-says. Due to the low number of monocytes recovered from TG donors’PBMC, mature DC were generated from peripheral blood samples ofhealthy blood donors (n � 2). To obtain mature DC, CD14-enriched pe-ripheral blood-derived monocytes were cultured with IL-4 and GM-CSFfor 6 days to generate immature monocyte-derived DC, and subsequentlymatured with a cytokine mixture as previously described (13, 14). Themature DC phenotype, characterized by high CD80, CD83, and CD86 ex-pression (15), was confirmed by flow cytometry (data not shown). Theeffector cells, i.e., allogeneic peripheral blood T cells, were labeled withCFSE (Invitrogen) at a final concentration of 0.5 �M. The stimulator cellswere cocultured with effector cells at a ratio of 1:10 at 37°C. At day 7, cellswere harvested for flow cytometric analyses. Cells were stained with CD3-allophycocyanin (UCHT1; DakoCytomation) to discriminate between Tcells and stimulator cells.

ResultsHuman TG-SGC express typical macrophage markers

We have previously shown that TG-SGC uniformly express MHCclass II, suggesting that they have a role as APC (12). Tissue-resident APC, including macrophages and DC, express the com-mon leukocyte marker CD45 enabling their distinction from stro-mal cells like fibroblasts. Paired TG-derived cells and PBMCsamples were assayed for CD45 expression. In contrast to PBMC,the TG-derived CD45� cell pool included two distinct cell popu-lations: CD45high and CD45low cells (Fig. 1A). Whereas theCD45high cells consisted mainly of T cells (data not shown), allCD45low cells expressed the monocyte/macrophage marker CD14(Fig. 1B and Table I). In situ analyses showed that CD14 wasexpressed by TG-SGC (Fig. 1B). As hinted upon by a previousreport (8), the macrophage-specific marker CD68 was expressedintracellularly (Fig. 1C and Table I), but not at the cell surface ofTG-SGC (data not shown). Additionally, TG-SGC selectively ex-pressed Ag uptake receptors like CD11b and CD11c (Fig. 1, D andE, and Table I), as well as CD16 and CD64 (data not shown). Insitu double stainings confirmed the flow cytometry data and dem-onstrated coexpression of CD14, CD45 and CD11c on TG-SGC(Fig. 2).

Human TG-SGC have an immature myeloid DC phenotype

The complement receptor CD11c is commonly used as a marker todiscriminate between myeloid (DC; CD11c�) and plasmacytoid(CD11c�) DC (16). Maturation of myeloid DC is characterized byup-regulation or induction of surface markers like MHC class IIand the costimulatory molecules CD80, CD83, and CD86 essentialfor T cell interaction and stimulation (15). Surface expression ofCD83 is considered characteristic for functionally mature DC (17).

The expression of CD11c and MHC class II on TG-SGCprompted us to determine the expression of additional DC markers.Whereas the TG-SGC expressed both CD80 and CD86 (Fig. 1, Fand G and Table I), the mature DC marker CD83 could not bedetected (Fig. 1H and Table I). Furthermore, TG-SGC coexpressedMHC class II and the T cell adhesion molecule CD54 (Fig. 1I).Except for CD40, all markers determined were expressed uni-formly on all TG-SGC. Whereas all TG-SGC of the TG donors(n � 4) analyzed were CD40� by flow cytometry, in situ analysesrevealed interdonor variation of CD40 expression on TG-SGC.Two of six TG donors analyzed showed weak but positive CD40staining on TG-SGC, which was occasionally associated with Tcell clusters (Fig. 1J). The discrepancies observed could be due tothe use of two different anti-CD40 mAbs in the separate assays,and, in case of the differential CD40 expression observed in the in

2457The Journal of Immunology

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

situ analyses, may be attributed to unknown TG donor-specific char-acteristics. Similar to the other markers analyzed, CD40 expressiondid not correlate with the presence of interacting T cells or latent�-herpesvirus infection (data not shown). Table II presents a compar-ative overview of the phenotype of human TG-SGC.

Human TG-SGC phagocytose bacterial particles

A critical role of macrophages is to phagocytose cellular debrisand pathogens. Because the TG-SGC have a macrophage pheno-type, we determined their capability to phagocytose bacterial par-ticles. Whole TG cell suspensions were incubated with fluoresce-in-conjugated E. coli after which the phagocytic cell type wasidentified by flow cytometry. Bacteria were predominantly asso-ciated with the CD45low TG cells, identified in the experiment asTG-SGC (Fig. 3A). Because this assay does not discriminate be-tween membrane bound and internalized bacteria, the E. coli-treated TG-SGC were isolated using anti-CD14 magnetic beadsand subsequently subjected to immunocytology. Confocal laserscanning microscopy revealed that the bacteria colocalized withthe late endosome marker CD68 (Fig. 3A), demonstrating that TG-SGC have actively phagocytosed the bacteria.

Human TG-SGC are unable to induce an allogeneic MLR

Although immature myeloid DC primarily function as phagocytes,DC maturation is associated with up-regulation of costimulatoryand MHC molecules, secretion of cytokines, down-regulationof phagocytic capacity, and increased ability to induce T cellresponses (15). It is well established that mature DC are potentstimulators of an allogeneic MLR, a characteristic that distin-guishes them from other APC (18). Because TG-SGC expresseda myeloid DC phenotype, they were used as stimulator cells inallogeneic MLR assays. From the same donor, peripheral blood-derived CD14� monocytes and CD14� TG-SGC were cocul-tured with CFSE-labeled allogeneic T cells. In contrast to ma-ture monocyte-derived DC, both monocytes and TG-SGC wereunable to induce T cell proliferation (Fig. 3B), indicating thathuman TG-SGC resemble immature myeloid DC both pheno-typically and functionally.

Human TG-infiltrating T cells express T cell inhibitorymolecules and TG-SGC the respective ligands

Although neuron-interacting CD8� T cells express cytolytic mol-ecules, like perforin and granzyme B, neuronal damage is not ob-served in type-1 HSV latently infected TG, suggesting that thecytolytic activity of the CD8� T cells is inhibited (7–12, 19). Re-cently, Suvas et al. (19) have shown that the NK inhibitory mol-ecule complex CD94/NKG2A prevents CD8� T cell-mediated TG

FIGURE 1. Human TG-resident SGC express macrophage- and DC-specific markers. A, Dot plots of paired TG cells (left) and PBMC (right)stained for CD45 to demonstrate that human TG harbor a unique cell pop-ulation expressing CD45 at low levels (gate R2). CD45low and CD45high

cells are arbitrarily green and blue in all dot plots, respectively. Subsequentpanels show representative ex vivo flow cytometric analysis (left) (n � 14donors) and in situ analysis (right) (n � 6 donors) of the expression ofCD14 (B), CD68 (C), CD11b (D), CD11c (E), CD80 (F), CD86 (G), CD83(H), CD54 (I), and CD40 (J) detected. The number for each quadrant in dotplot represents the percentage of cells expressing the indicated markerdefined on matched isotype control mAb stainings. Sections were devel-oped with AEC (bright red precipitate) and counterstained with hematox-ylin (blue nuclei). Original magnifications are �200 (B–E, and J) and�400 (F–I).

FIGURE 2. Human TG-resident SGC express APC markers. HumanTG single cell suspensions, and frozen TG biopsy specimens, were ana-lyzed for the markers CD45 and CD11c (A), and CD14 and CD11c (B)in cytometry analysis (n � 14 donors) and double-color in situ analysis(n � 6 donors) on consecutive sections, respectively. CD45low andCD45high cells are arbitrarily green and blue in all dot plots, respec-tively. The number for each quadrant in the dot plot represents cellsexpressing the indicated marker. Slides were developed with AEC andFast blue resulting in red and blue staining patterns, respectively. Adouble positive cell, stained purple, is enlarged for experiment (farright). Original magnification is �200.

Table I. Marker expression on CD45low human TG-SGC

MarkerPercentage of Positive

TG-SGC � SDa No. of Donors

CD14 95.9 � 4.7 9CD68 97.3 � 1.1 2CD11b 92.3 � 6.6 2CD11c 88.5 � 8.3 8CD80 82.3 � 18.5 5CD86 94.8 � 5.6 6CD83 6.3 � 4.1 3CD40 91.9 � 7.2 4

a Data represent the average of TG-SGC that express the indicated marker deter-mined by flow cytometry.

2458 HUMAN SGC REPRESENT TISSUE-RESIDENT APC

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

neuron destruction in mice. Whereas the majority of the TG-infil-trating HSV-specific CD8� T cells expressed CD94/NKG2A, bothneurons and CD11b� cells expressed the cognate ligand Qa-1b(19). Analogous to the mouse, human TG-infiltrating T cells co-expressed CD94 and NKG2A (Fig. 4A). Moreover, the frequencyof CD94/NKG2A� T cells in TG (mean frequency 13 � 4%) washigher compared with peripheral blood (mean frequency 3 � 1%),suggesting selective infiltration or differentiation of T cells to ex-press CD94/NKG2A locally. The cognate receptor HLA-E (20)was expressed throughout the TG tissue, including TG-SGC, and

CD94 expression colocalized with CD3 within neuron-interactingT cell clusters (Fig. 4A).

In addition to NK inhibitory molecules, several studies haveindicated that the molecule PD-1 and its ligand PD-L1 negativelyregulate T cell effector functions (21–25). Both CD4� and CD8�

TG-infiltrating T cells expressed PD-1, but percentages and ex-pression levels did not differ between donor-matched TG-derivedT cells (mean 29 � 7%) and peripheral blood T cells (mean 35 �12%) (Fig. 4B). However, in situ analyses revealed that neuron-interacting T cell clusters tended to have a higher PD-1 expression,compared with scattered T cells (Fig. 4B). Notably, PD-L1 expres-sion was confined to TG-SGC and appeared to be higher on TG-SGC in proximity to the T cell clusters (Fig. 4C).

DiscussionFor decades, SGC have been regarded as nursing cells providingphysical support to neuron somata in sensory ganglia. The currentstudy demonstrates that human TG-SGC have phenotypic and

FIGURE 3. Human TG-resident SGC share functional characteristicswith macrophages and immature myeloid DC. A, Human TG-SGC wereincubated with fluorescein-conjugated bacteria to determine their phago-cytic function by flow cytometry (left) and confocal laser scanning micros-copy (right). CD45low and CD45high cells are arbitrarily green and blue inthe dot plot, respectively (left). Cytospins of CD14-enriched TG-SGCtreated with fluorescein-conjugated bacteria (bacteria in green) werestained for CD68 (late endosomes in red) and DAPI (cellular nuclei inblue) and examined by confocal laser scanning microscopy (right). B, Dotplots of a representative allogeneic MLR using mature monocyte-derivedDC generated from peripheral blood-derived monocytes of a healthy blooddonor (DC, left), and CD14-enriched peripheral blood monocytes (PBCD14�, middle), and CD14-enriched TG-SGC (TG CD14�, right) recov-ered from the same TG donor, hereby used as stimulator cells in combi-nation with CFSE-labeled allogeneic T cells. The percentage indicates thefrequency of T cells that proliferated upon incubation at 37°C for 7 days.Results are representative of two experiments performed on two TGdonors.

FIGURE 4. Human TG-infiltrating T cells express inhibitory molecules.A, Dot plot of ex vivo flow cytometry (left) on CD94 and NKG2A expres-sion on gated T cells, and in situ analyses of CD3, CD94, and HLA-E onconsecutive sections. B and C, Dot plots of ex vivo flow cytometric (top)and in situ analyses (bottom) of CD3 and PD-1 (B), and CD3 and PD-L1(C) on consecutive sections. The number for each quadrant in dot plotrepresents the percentage of cells expressing the indicated marker definedon matched isotype control mAb stainings. Dot plots in A and B are gatedon CD3� cells. Representative data from six TG donors are shown. Sec-tions were developed with AEC (bright red precipitate) and counterstainedwith hematoxylin (blue nuclei). Original magnification is �200.

Table II. Comparison of phenotype and functional characteristics of TG-SGC to other human APCa

Macrophageb Immature DCb Mature DCb CNS Microgliab TG-SGC

PhenotypeCD14 and CD68 � � � �c �CD16 and CD64 � � � � �CD11b and CD11c � � � � �MHC class II � � �� � �CD45 High High High Low LowCD40 and CD54 � � �� � �CD80 and CD86 � � �� � �CD83 � � � �c �

FunctionPhagocytosis � � � � �Allogeneic MLR � � � �c �

a Results indicate the presence (�), intensity (��; high and low), or absence (�) of the markers or functional characteristicsindicated.

b Data previously described (14, 24, 28).c Upon stimulation with LPS, microglia express CD14 and CD83, and are able to induce an allogeneic MLR (24).

2459The Journal of Immunology

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

functional APC properties. Two main findings are reported. First,human TG-SGC have a unique leukocyte phenotype, with featuresof both macrophages and immature myeloid DC. Second, TG-in-filtrating T cells expressed the T cell inhibitory molecules CD94/NKG2A and PD-1, and the interacting TG-SGC expressed the cog-nate ligands HLA-E and PD-L1, respectively.

Current knowledge on CNS-resident glial cells advocate theirrole as critical participants in the healthy and diseased brain bymaintaining axonal integrity and myelination, providing nutrients,controlling synapse formation and function, and immune regula-tion (26–30). Whereas macroglia, like astrocytes and oligodendro-cytes, are derived from the neuroectoderm (31), microglia expressseveral leukocyte cell markers implicating their origin from my-eloid progenitor cells (26, 32). Microglia are the main CNS-resi-dent APC that constantly sense and sample the brain environmentand coordinate immune responses in response to danger signals(26–28). They resemble macrophages and immature myeloid DCand have been implicated in neurodegenerative disorders likemultiple sclerosis (27, 29, 30). Both human and rodent micro-glia express low levels of the membrane molecule CD45, amarker commonly used to distinguish microglia (CD45low)from stromal cells and macroglia (both CD45�) and infiltratinglymphocytes (CD45high) (33, 34).

In contrast to CNS glial cells, the immune function of peripheralnervous system resident SGC is poorly defined. Our data demon-strate that human TG-SGC closely resemble CNS microglia bothphenotypically and functionally (Table II). Microglia and TG-SGCare CD45low, and express similar macrophage- and DC-associatedmarkers and T cell costimulatory molecules (Figs. 1 and 2, andTable II) (26). Furthermore, both cell types actively phagocytosebioparticles and are unable to induce primary T cell responses (Fig.3) (35, 36). Hitherto, peripheral nervous system-resident SGC havebeen considered to be neuroectoderm-derived (2, 3, 37). The cur-rent study challenges this concept, suggesting that human TG-SGCarise from myeloid progenitors analogous to microglia (26, 32).

Recent data obtained by the Carbone group (38) support thishypothesis. The authors studied the local effector cells involved inmaintaining virus-specific CD8� T cell responses that controlHSV-1 latency in sensory ganglia of experimentally infected mice.It was shown that CD8� T cell homeostasis was depending on atripartite interaction that includes infiltrating CD4� T cells andrecruited DC. The effector DC originated from circulating mono-cytes and expressed high levels of CD11b, CD11c, MHC class II,and F4/80. In situ analyses showed that the CD11c� DC wereoccasionally found in close proximity to CD8� T cells, but morestrikingly they appeared to surround the neuronal somata (38). Thecomparable phenotype and anatomic localization of murine sen-sory ganglia-resident DC and human TG-SGC suggest that theyrepresent the same cell type. This local APC may present the cog-nate HSV-1 Ags to infiltrating virus-specific CD8� T cells. Studiesin mice support this notion (39, 40). Alternatively, HSV-1-specificCD8� T cells may penetrate the SGC sheet to interact directly withthe latently infected neurons (7). Because neurons do not expressMHC class II, infiltrating virus-specific CD4� T cells most likelyinteract with TG-SGC.

In contrast to a previous study on human TG, the majority of theinvestigated markers analyzed in this study were uniformly ex-pressed by TG-SGC of the TG donors studied. This finding wasirrespective of the HSV status of the donor and varicella zostervirus serostatus, and the presence of infiltrated T cells (data notshown). A major difference between the preceding and presentstudy is the median age of the TG donors analyzed: 29 vs 79 years,respectively (8). Animal studies have demonstrated that aging in-duces the transition of naive microglia into an activated state, char-

acterized by up-regulation of MHC class II and CD68 (41, 42).Consequently, the discrepancy between both studies may in part beattributed to the relatively high age of the TG donors analyzed inthis study.

It is generally established that TG-infiltrating CD8� T cells in-hibit HSV-1 reactivation by means of IFN-� and cytolytic effectormolecules (7, 8, 12, 43, 44). Nevertheless, the latently infectedneurons encountered are not damaged, suggesting that cytolytic Tcell effector functions are inhibited (8, 12, 43). The expression ofCD94/NKG2A on human TG-infiltrating T cells is consistent witha previous study on mouse TG, demonstrating that blocking theCD94-NKG2A/Qa-1b interaction in ex vivo TG cultures resultedin neuronal cell lysis (19). CD94 expression in human TG wasselectively expressed by T cells interacting with neuronal somata,suggesting an analogous role of the CD94-NKG2A/HLA-E inter-action in human latently infected TG. Notably, Qa-1b was ex-pressed by neurons, but also CD11b� cells in mouse TG (19). TheCD11b� Qa-1b� cells may represent the effector DC that are func-tionally involved in controlling local T cell responses in HSV-1latently infected mouse sensory ganglia (38).

In addition to CD94/NKG2A, the data on human TG suggest theinvolvement of the T cell inhibitory molecule PD-1. Human TG-infiltrating T cells and TG-SGC expressed PD-1 and PD-L1, re-spectively. Notably, the expression of both markers appeared to behigher within neuron-interacting T cell clusters. IFN stimulationup-regulates PD-1 and PD-L1 expression on receptive cells (22,45). Consequently, the differential PD-1 and PD-L1 expressionobserved may be attributed to IFN-� secreted by activated T cellsrecognizing the latent virus. Functional studies are mandatory toinvestigate the role of both the HLA-E/CD94-NKG2A and PD-1/PD-L1 pathway to inhibit cytolytic T cell effector function in hu-man HSV-1 latently infected TG. Moreover, elucidation of the Tcell inhibitory mechanisms used in the peripheral nervous systemmay provide tools for the development of future therapeutic inter-vention strategies to counteract undue cell damage associated withT cell-mediated chronic diseases.

In conclusion, the data presented in this study show that humanTG-resident SGC have a unique leukocyte phenotype, sharingproperties with macrophages and immature myeloid DC. We hy-pothesize that TG-SGC are tissue-resident APC involved in sens-ing the local environment and the control of local T cell responsesto protect the irreplaceable neuronal somata in TG.

AcknowledgmentsWe thank the Netherlands Brain Bank team for their efforts to provide thehuman TG specimens and are indebted to the donors who agreed to provideTG specimens for research purposes. We also thank D. E. Geraghty (FredHutchinson Cancer Research Center, Seattle, WA) for providing the anti-HLA-E mAb 4D12. The authors acknowledge discussions within the Eu-ropean Cooperation in Science and Technology (COST) Action BM0603Inflammation in Brain Disease Neurinfnet, and networking supportfrom COST.

DisclosuresThe authors have no financial conflict of interest.

References1. Bechmann, I., I. Galea, and V. H. Perry. 2007. What is the blood-brain barrier

(not)? Trends Immunol. 28: 5–11.2. Hanani, M. 2005. Satellite glial cells in sensory ganglia: from form to function.

Brain Res. Rev. 48: 457–476.3. Pannese, E. 1981. The satellite cells of the sensory ganglia. Adv. Anat. Embryol.

Cell Biol. 65: 1–111.4. Elson, K., A. Simmons, and P. Speck. 2004. Satellite cell proliferation in murine

sensory ganglia in response to scarification of the skin. Glia 45: 105–109.5. Lee, S. E., H. Shen, G. Taglialatela, J. M. Chung, and K. Chung. 1998. Expres-

sion of nerve growth factor in the dorsal root ganglion after peripheral nerveinjury. Brain Res. 796: 99–106.

2460 HUMAN SGC REPRESENT TISSUE-RESIDENT APC

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

6. Roizman, B., D. Knipe, and R. J. Whitley. Fields’ Virology, 5th ed. D. Knipe,and P. Howley, eds. Philadelphia, Pa. Lippincott Williams & Wilkins; 2007:2501–2601.

7. Khanna, K. M., R. H. Bonneau, P. R. Kinchington, and R. L. Hendricks. 2003.Herpes simplex virus-specific memory CD8� T cells are selectively activated andretained in latently infected sensory ganglia. Immunity 18: 593–603.

8. Theil, D., T. Derfuss, I. Paripovic, S. Herberger, E. Meinl, O. Schueler,M. Strupp, V. Arbusow, and T. Brandt. 2003. Latent herpesvirus infection inhuman trigeminal ganglia causes chronic immune response. Am. J. Pathol. 163:2179–2184.

9. van Lint, A. L., L. Kleinert, S. R. Clarke, A. Stock, W. R. Heath, andF. R. Carbone. 2005. Latent infection with herpes simplex virus is associated withongoing CD8� T-cell stimulation by parenchymal cells within sensory ganglia.J. Virol. 79: 14843–14851.

10. Khanna, K. M., A. J. Lepisto, V. Decman, and R. L. Hendricks. 2004. Immunecontrol of herpes simplex virus during latency. Curr. Opin. Immunol. 16:463–469.

11. Liu, T., K. M. Khanna, X. Chen, D. J. Fink, and R. L. Hendricks. 2000. CD8�

T cells can block herpes simplex virus type 1 (HSV-1) reactivation from latencyin sensory neurons. J. Exp. Med. 191: 1459–1466.

12. Verjans, G. M., R. Q. Hintzen, J. M. van Dun, A. Poot, J. C. Milikan,J. D. Laman, A. W. Langerak, P. R. Kinchington, and A. D. Osterhaus. 2007.Selective retention of herpes simplex virus-specific T cells in latently infectedhuman trigeminal ganglia. Proc. Natl. Acad. Sci. USA 104: 3496–3501.

13. Sallusto, F., and A. Lanzavecchia. 1994. Efficient presentation of soluble antigenby cultured human dendritic cells is maintained by granulocyte/macrophage col-ony-stimulating factor plus interleukin 4 and downregulated by tumor necrosisfactor �. J. Exp. Med. 179: 1109–1118.

14. Allard, S. D., K. Pletinckx, K. Breckpot, C. Heirman, A. Bonehill, A. Michiels,C. A. van Baalen, R. A. Gruters, A. D. Osterhaus, P. Lacor, et al. 2008. Func-tional T-cell responses generated by dendritic cells expressing the early HIV-1proteins Tat, Rev and Nef. Vaccine 26: 3735–3741.

15. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control ofimmunity. Nature 392: 245–252.

16. Colonna, M., G. Trinchieri, and Y. J. Liu. 2004. Plasmacytoid dendritic cells inimmunity. Nat. Immunol. 5: 1219–1226.

17. Prechtel, A. T., and A. Steinkasserer. 2007. CD83: an update on functions andprospects of the maturation marker of dendritic cells. Arch. Dermatol. Res. 299:59–69.

18. Crow, M. K., and H. G. Kunkel. 1982. Human dendritic cells: major stimulatorsof the autologous and allogeneic mixed leucocyte reactions. Clin. Exp. Immunol.49: 338–346.

19. Suvas, S., A. K. Azkur, and B. T. Rouse. 2006. Qa-1b and CD94-NKG2a inter-action regulate cytolytic activity of herpes simplex virus-specific memory CD8�

T cells in the latently infected trigeminal ganglia. J. Immunol. 176: 1703–1711.20. Kaiser, B. K., F. Barahmand-Pour, W. Paulsene, S. Medley, D. E. Geraghty, and

R. K. Strong. 2005. Interactions between NKG2x immunoreceptors and HLA-Eligands display overlapping affinities and thermodynamics. J. Immunol. 174:2878–2884.

21. Blank, C., and A. Mackensen. 2007. Contribution of the PD-L1/PD-1 pathway toT-cell exhaustion: an update on implications for chronic infections and tumorevasion. Cancer Immunol. Immunother. 56: 739–745.

22. Sharpe, A. H., E. J. Wherry, R. Ahmed, and G. J. Freeman. 2007. The functionof programmed cell death 1 and its ligands in regulating autoimmunity and in-fection. Nat. Immunol. 8: 239–245.

23. Barber, D. L., E. J. Wherry, D. Masopust, B. Zhu, J. P. Allison, A. H. Sharpe,G. J. Freeman, and R. Ahmed. 2006. Restoring function in exhausted CD8 T cellsduring chronic viral infection. Nature 439: 682–687.

24. Maier, H., M. Isogawa, G. J. Freeman, and F. V. Chisari. 2007. PD-1:PD-L1interactions contribute to the functional suppression of virus-specific CD8� Tlymphocytes in the liver. J. Immunol. 178: 2714–2720.

25. Jurado, J. O., I. B. Alvarez, V. Pasquinelli, G. J. Martinez, M. F. Quiroga,E. Abbate, R. M. Musella, H. E. Chuluyan, and V. E. Garcia. 2008. Programmeddeath (PD)-1:PD-ligand 1/PD-ligand 2 pathway inhibits T cell effector functionsduring human tuberculosis. J. Immunol. 181: 116–125.

26. Guillemin, G. J., and B. J. Brew. 2004. Microglia, macrophages, perivascularmacrophages, and pericytes: a review of function and identification. J. LeukocyteBiol. 75: 388–397.

27. Hanisch, U. K., and H. Kettenmann. 2007. Microglia: active sensor and versatileeffector cells in the normal and pathologic brain. Nat. Neurosci. 10: 1387–1394.

28. Nimmerjahn, A., F. Kirchhoff, and F. Helmchen. 2005. Resting microglial cellsare highly dynamic surveillants of brain parenchyma in vivo. Science 308:1314–1318.

29. Streit, W. J. 2004. Microglia and Alzheimer’s disease pathogenesis. J. Neurosci.Res. 77: 1–8.

30. Ulvestad, E., K. Williams, R. Bjerkvig, K. Tiekotter, J. Antel, and R. Matre.1994. Human microglial cells have phenotypic and functional characteristics incommon with both macrophages and dendritic antigen-presenting cells. J. Leu-kocyte Biol. 56: 732–740.

31. Dong, Y., and E. N. Benveniste. 2001. Immune function of astrocytes. Glia 36:180–190.

32. Chan, W. Y., S. Kohsaka, and P. Rezaie. 2007. The origin and cell lineage ofmicroglia: new concepts. Brain. Res. Rev. 53: 344–354.

33. Becher, B., and J. P. Antel. 1996. Comparison of phenotypic and functionalproperties of immediately ex vivo and cultured human adult microglia. Glia 18:1–10.

34. Greter, M., F. L. Heppner, M. P. Lemos, B. M. Odermatt, N. Goebels, T. Laufer,R. J. Noelle, and B. Becher. 2005. Dendritic cells permit immune invasion of theCNS in an animal model of multiple sclerosis. Nat. Med. 11: 328–334.

35. Carson, M. J., C. R. Reilly, J. G. Sutcliffe, and D. Lo. 1998. Mature microgliaresemble immature antigen-presenting cells. Glia 22: 72–85.

36. Davoust, N., C. Vuaillat, G. Androdias, and S. Nataf. 2008. From bone marrowto microglia: barriers and avenues. Trends Immunol. 29: 227–234.

37. Li, H. Y., E. H. Say, and X. F. Zhou. 2007. Isolation and characterization ofneural crest progenitors from adult dorsal root ganglia. Stem. Cells 25:2053–2065.

38. Wakim, L. M., J. Waithman, N. van Rooijen, W. R. Heath, and F. R. Carbone.2008. Dendritic cell-induced memory T cell activation in nonlymphoid tissues.Science 319: 198–202.

39. Pereira, R. A., D. C. Tscharke, and A. Simmons. 1994. Upregulation of class Imajor histocompatibility complex gene expression in primary sensory neurons,satellite cells, and Schwann cells of mice in response to acute but not latentherpes simplex virus infection in vivo. J. Exp. Med. 180: 841–850.

40. Liu, T., Q. Tang, and R. L. Hendricks. 1996. Inflammatory infiltration of thetrigeminal ganglion after herpes simplex virus type 1 corneal infection. J. Virol.70: 264–271.

41. Dilger, R. N., and R. W. Johnson. 2008. Aging, microglial cell priming, and thediscordant central inflammatory response to signals from the peripheral immunesystem. J. Leukocyte Biol. 84: 932–939.

42. Perry, V. H., M. K. Matyszak, and S. Fearn. 1993. Altered antigen expression ofmicroglia in the aged rodent CNS. Glia 7: 60–67.

43. Knickelbein, J. E., K. M. Khanna, M. B. Yee, C. J. Baty, P. R. Kinchington, andR. L. Hendricks. 2008. Noncytotoxic lytic granule-mediated CD8� T cell inhi-bition of HSV-1 reactivation from neuronal latency. Science 322: 268–271.

44. Liu, T., K. M. Khanna, B. N. Carriere, and R. L. Hendricks. 2001. Gammainterferon can prevent herpes simplex virus type 1 reactivation from latency insensory neurons. J. Virol. 75: 11178–11184.

45. Muhlbauer, M., M. Fleck, C. Schutz, T. Weiss, M. Froh, C. Blank,J. Scholmerich, and C. Hellerbrand. 2006. PD-L1 is induced in hepatocytes byviral infection and by interferon-� and -� and mediates T cell apoptosis. J. Hepa-tol. 45: 520–528.

2461The Journal of Immunology

by guest on April 6, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from