Herpes Simplex Virus Type 1–Induced Transcriptional Networks of Corneal Endothelial Cells Indicate Antigen Presentation Function

Abstract

Purpose.: To determine the transcriptional response of cultured human corneal endothelial (HCEn) cells after herpes simplex virus type (HSV-1) infection and to characterize the primary functional elements and antiviral responses.

Methods.: Immortalized HCEn cells were infected with HSV-1, and the global transcriptional profile was determined. The transcriptional networks of HCEn cells were constructed, and the inflammatory network nodes were evaluated for induction of candidate inflammatory mediators by protein array analyses. HSV-1–specific allogeneic T cells isolated from HSV-1–infected donors were co-cultured with HSV-1–pulsed HCEn cells, and T cell activation was assessed for antigen-specific proliferation.

Results.: HSV-1 infection induced a global transcriptional activation with 331 genes significantly up- or downregulated compared with mock-infected HCEn cells (P < 0.01; 4< or 0.25> threshold). Network analysis showed that the HSV-1–induced transcriptome was specifically associated with antigen presentation, interferon-related responses, and cellular development, and was characterized by NF-κB and extracellular signal–regulated kinase signaling pathways. The primary associated function in the transcriptome was antigen presentation. Protein array analysis identified significant elevation of genes related to antigen presentation: IL-6, IP-10, HVEML, and interferon-γ. In addition, inflammatory cytokines including IL-8, MCP-1, TIMP-1, RANTES, I-309, MIF, MCP-2, IL-10, and SDF-1, in descending order, were significantly elevated. Mixed lymphocyte reaction assays showed that HSV-1–pulsed HCEn cells stimulated antigen-specific proliferation of allogeneic T lymphocytes.

Conclusions.: HCEn cells respond to HSV-1 infection by initiating antigen presentation–related inflammatory responses, and they may serve as antigen-presenting cells.

Corneal endotheliitis is a progressive form of corneal endotheliopathy that is characterized by focal, linear, or diffuse corneal edema. It can lead to progressive endothelial cell loss and to endothelial dysfunction. Relevant to this study, an intracameral injection of herpes simplex virus (HSV)-1 can lead to corneal endotheliitis, ¹ and molecular diagnostic methods have shown that HSV contributes to the pathogenesis of corneal endotheliitis. ²  

The most frequent HSV-associated diseases of the cornea are epithelial keratitis and stromal keratitis, although stromal keratitis is known to involve the corneal endothelial cells as well. In contrast, pure endotheliitis without stromal keratitis due to HSV-1 is rare. Generally, detailed evaluations of the endothelial cells after HSV infection cannot be made by slit lamp examination and specular microscopy because of corneal opacification. ³ However, Hillenaar et al. ³ found by in vivo confocal microscopy that 43% of patients with common HSV keratitis had characteristic signs of endotheliitis, including pseudoguttata, enlarged intercellular gaps, infiltration of inflammatory cells into the endothelium, loss of cell boundary, spotlike holes, and endothelial denudation. These alterations of the corneal endothelial cells were shown to be resolved after antiviral and -inflammatory treatment, but the density of the endothelial cells in the affected eye decreased by 10.3%/year. 

Corneal endothelial cells are permissive to HSV infection, as shown in human corneal endothelial (HCEn) cells grown in vitro by Sugioka et al. ⁴ Of note, the HCEn cells had higher susceptibility to HSV-1 and produced more viral particles than the representative permissive CV-1 cell line. So, the question arises as to how HCEn cells resist HSV infection despite their inherent susceptibility to HSV-1 infection. One possible answer to this question is the immune-modulatory properties of HCEn cells. 

Anterior chamber–associated immune deviation (ACAID) is a well-known mechanism of peripheral immune tolerance, ⁵ and HCEn cells appear to be an important player in this process. For example, HCEn cells inhibit the CD3-stimulated proliferation of effector T cells in a cell-contact–dependent manner using programmed cell death 1 ligand 1 (PD-L1). ⁶ The HCEn cells can also convert CD8⁺ T cells into regulatory T cells through membrane-bound TGF-β. ⁷ Thus, HCEn cells have the ability to modulate immune responses; however, it is still not known whether HCEn cells possess antigen-presentation capabilities. 

How corneal epithelial and endothelial cells respond to pathogens is an important unanswered question, as is how they respond globally to pathogens. To try to answer these questions in an earlier study, we used human corneal epithelial cells (HCEps), which are representative cells permissive to HSV-1, to characterize the global transcriptional responses of the HCEp cells to HSV-1 infection. Application of bioinformatic methods showed that HCEp cells responded to HSV-1 infection by initiating mitogen-activated protein kinase-related transcriptional events, and also enhanced the release of IL-6 which induced an array of inflammatory mediators. ⁸  

In the same way, determining how HCEn cells respond to HSV infection may provide important clues about the physiological functions and contribution of HCEn cells. We will show that the global responses of HCEn cells to HSV-1 are markedly different from HCEp cells and are preferentially set to antigen presentation. This antigen-presentation capability was confirmed by their ability to stimulate HSV-1–specific allogeneic T-lymphocyte responses. 

Materials and Methods

Cells

The HCEn cell line was established by transduction with hTERT and the large T gene, as described. ⁹ Retroviral vectors, BABE-hygro-hTERT (for hTERT), and MFG-tsT-IRES-neo (for SV40 large T antigen), were used, as described in detail. ^6,10 The HCEn cells were propagated to confluence on 6- or 96-well plates in DMEM (Dulbecco's modified Eagle's medium; Invitrogen-Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum. 

Virus

Confluent monolayers of Vero cells were infected with HSV-1 (KOS strain). ⁸ To analyze the transcriptome of HSV-1–infected HCEn cells, we used the HCEp transcriptome as a reference, as reported. ⁸ Purified virus stock was prepared as described. ⁸ After 1 hour of adsorption, the medium containing the virus was aspirated, and the monolayers were re-fed with fresh HSV-1–free media. At the maximum cytopathic effect, the media were discarded, and the cells with a small amount of remaining media were frozen, thawed, sonicated, and centrifuged at 3000 rpm for 10 minutes. The supernatant was overlaid onto a sucrose density gradient (10%–60% wt/vol) and centrifuged on a swing rotor (SW28; Beckman, Fullerton, CA) for 1 hour at 11,500 rpm. The resultant visible band at the lower part of the gradient which contained the HSV-1 was washed by centrifugation at 14,000 rpm for 90 minutes and resuspended in a small volume of serum-free DMEM. The sample was then aliquotted and stored at −80°C until use. The infectivity of the virus was determined by plaque titration assay and was typically 1 × 10⁹ plaque forming units (PFU) per milliliter. To infect HCEn cells with HSV-1, the cells were adsorbed with sucrose-density, gradient-purified virus stock for 1 hour and re-fed with fresh medium. 

Microarray Procedures

HSV-infected HCEn cells were transcriptionally analyzed using a whole human genome microarray (Agilent Technologies, Santa Clara, CA) corresponding to 41,000 human genes and transcripts. HCEn cells were infected with HSV-1 at a multiplicity of infection (MOI) of 1. Total RNA was isolated from the HSV-infected HCEn cells 12 hours postinfection (PI; RNeasy Mini Kit; Qiagen, Hilden, Germany), according to the manufacturer's instructions. Mock-infected HCEn cells were used as controls. 

Cyanine-3 labeled cRNA was prepared from 0.25 μg of RNA (One-Color Low RNA Input Linear Amplification PLUS kit; Agilent). Fragmented cRNA was hybridized to the whole human genome oligo microarray (model G4112F, Agilent) using a hybridization kit (Gene Expression Hybridization, G2545A; Agilent) and scanned with a microarray scanner (model G2565BA; Agilent). The acquired data were bioinformatically analyzed (GeneSpring GX 10; Agilent), and the genes differentially up- or downregulated after HSV infection were extracted from the whole genome by using t-test. 

Functional Analysis of Data Set

Functional analysis was used to identify the biological function and/or disease that was most significant to the data set (Ingenuity Pathway analysis 7.0; Ingenuity Systems, Redwood, CA, computer program based on the Ingenuity Pathway Knowledge Base; http://www.ingenuity.com/products/pathways_analysis.html). Genes from the data set that met the cutoff of fourfold difference (P < 0.01) and were associated with biological functions and/or diseases in the Ingenuity Pathway Knowledge Base were selected for the analysis. Fisher's exact test was used to calculate a P value determining the possibility that each biological function and/or disease assigned to that data set was due to chance alone. 

Canonical Pathway Analyses of Data Set

Canonical pathway analyses were used to identify the pathways from the pathways analysis library of canonical pathways that were most significant to the data set. Genes from the data set that met the cutoff of fourfold difference (P < 0.01) and were associated with a canonical pathway in the pathway knowledge base were selected for the analyses. The significance of the association between the data set and the canonical pathway was measured in two ways: (1) a ratio of the number of genes from the data set that map to the pathway divided by the total number of genes that map to the canonical pathway, and (2) the use of Fischer's exact test to calculate a P-value determining the probability that the association between the genes in the dataset and the canonical pathway can be explained by chance alone. 

Network Analysis of the HSV-1–Induced Transcriptome

The set of extracted genes was analyzed for transcriptional networks of molecular events using pathway analysis. The resulting networks were evaluated by the significance scores, which were expressed as the negative logarithm of the P value. The obtained score indicated the likelihood that the assembly of a set of focus genes in a network could be explained by random chance alone. 

Real-time RT-PCR

Total RNA was isolated from the HSV-infected HCEn cells and reverse transcribed using (QuantiTect Reverse Transcription Kit; Qiagen), and the cDNAs were amplified and quantified (LightCycler; Roche, Mannheim, Germany, QuantiTect SYBR Green PCR kit). The sequences of the real-time PCR primer pairs were IFN-a1: forward 5′-GGAGTTTGATGGCAACCAGT-3′ and reverse 5′-CTCTCCTCCTGCATCACACA-3′; and glyceraldehyde-3-phosphate dehydrogenase (GAPDH): forward 5′- AGCCACATCGCTCAGACAC-3′ and reverse 5′- GCCCAATACGACCAAATCC-3′. 

To ensure equal loading and amplification, all products were normalized to GAPDH transcript as an internal control. 

Cytokine Array Analyses

To profile the inflammatory cytokine after HSV infection, supernatants were collected from HCEn cells 12 hours PI and assayed with a cytokine antibody array (Human Body Array; RayBiotech, Norcross, GA). This process determined the level of expression of 80 cytokines. The intensity of the chemiluminescence signals was digitized (LAS-1000plus with MultiGauge software ver. 2.0; Fujifilm, Tokyo, Japan). 

T-Lymphocyte Proliferation Assay

T lymphocytes were prepared from peripheral blood mononuclear cells of human donors with histories of recurrent herpetic lesions by using negative selection with an immune magnetic beads–based isolation kit (IMag; BD Biosciences, Franklin Lakes, NJ). These cells were further negatively selected for CD4⁺ T cells with an isolation kit (IMag). HCEn cells were seeded into 96-well plates, exposed to purified HSV-1 (KOS strain) for 1 hour, and treated with mitomycin C (Sigma-Aldrich, St. Louis, MO) at 4.5 hours PI. The HSV-primed HCEn cells were co-cultured with isolated T cells for 3 days and pulsed with BrdU for 12 hours. The incorporation of BrdU was measured by chemiluminescence-based ELISA (Roche). Interferon-γ levels in the supernatant were measured with an ELISA kit, according to the manufacturer's instructions (eBioscience, San Diego, CA). 

The procedures used conformed to Declaration of Helsinki. Informed consent was obtained from all the participants. 

Statistical Analyses

Data are presented as the mean ± SEM. Statistical analyses were performed using t-tests or ANOVA as appropriate. 

Results

Microarray Analysis of HSV-1–Infected Corneal Endothelial Cells

Viral infection usually induces an interferon response from the host; however, the interferon response is generally silenced by HSV-1 infection by its exploitation. ¹¹ Therefore, we first tested whether the HSV-1 infection induced an interferon response of HCEn cells. Similar to the HCEp response, ⁸ the HCEn cells transcriptionally induced an interferon response that was detected at 12 hours PI and was higher at 24 hours. The expression of IFN-α1 relative to GAPDH was 73.4 ± 19.4 relative copies at an MOI of 1 of HSV-1 and 6.4 ± 0.5 relative copies for a mock infection (P < 0.005). 

To determine the early global responses to HSV-1, we conducted a transcriptional profiling of HSV-1–infected HCEn cells by microarray analysis. We identified 8979 genes that were differentially expressed in HSV-1–infected at 12 hours PI at an MOI of 1 (P < 0.01). To extract sets of virus-responsive genes, we set a threshold of fourfold expression changes. This threshold resulted in the detection of 453 upregulated genes and 8 downregulated genes in the HSV-1–infected HCEn cells (Table 1). Thus, HSV-1 infection of HCEn cells globally activated transcriptional responses. 

Table 1.

 

Significantly Upregulated or Downregulated Genes of Human Corneal Endothelial Cells after HSV-1 Infection

GenBank ID*	Gene Symbol	Change	Regulation
NM_016084	RASD1	239.5	Up
NM_005618	DLL1	158.7	Up
NM_005634	SOX3	118.5	Up
NM_015193	ARC	99.3	Up
NM_001362	DIO3	77.9	Up
NM_002164	IDO1	62.3	Up
AK024457	FLJ00049	62.0	Up
NM_001565	CXCL10	54.1	Up
NM_138800	TRIM43	53.0	Up
ENST00000334770	ENST00000334770	53.0	Up
NM_080657	RSAD2	51.2	Up
NM_005409	CXCL11	50.7	Up
BC141819	BC141819	49.6	Up
NM_020975	RET	48.3	Up
NM_152677	ZSCAN4	47.8	Up
NM_002590	PCDH8	46.7	Up
NM_201589	MAFA	43.8	Up
THC2750782	THC2750782	40.9	Up
NM_002523	NPTX2	38.3	Up
NM_170672	RASGRP3	38.3	Up
NM_005382	NEFM	37.9	Up
NM_020358	TRIM49	37.7	Up
NM_002522	NPTX1	35.7	Up
NM_001005217	FRG2	34.8	Up
NM_052942	GBP5	32.2	Up
NM_003733	OASL	29.8	Up
NM_144614	MBD3L2	29.3	Up
NM_032855	HSH2D	29.1	Up
NM_001012276	PRAMEF8	28.6	Up
NM_004561	OVOL1	27.6	Up
NM_002776	KLK10	27.0	Up
NM_001485	GBX2	26.8	Up
NM_002416	CXCL9	26.6	Up
BC040902	PRAMEF2	26.4	Up
AW105154	AW105154	26.3	Up
NM_006573	TNFSF13B	26.3	Up
ENST00000273083	GRIP2	26.1	Up
NM_016323	HERC5	25.7	Up
XR_016154	LOC642425	25.3	Up
NM_016358	IRX4	24.8	Up
NM_001775	CD38	24.4	Up
NM_002196	INSM1	24.3	Up
NM_014310	RASD2	23.8	Up
NM_003004	SECTM1	23.6	Up
NM_006705	GADD45G	22.5	Up
NM_052941	GBP4	21.9	Up
NM_001040429	PCDH17	21.8	Up
NM_022454	SOX17	21.8	Up
NM_004833	AIM2	21.7	Up
NM_002507	NGFR	21.3	Up
NM_003956	CH25H	20.4	Up
NM_138456	BATF2	19.9	Up
NM_172374	IL4I1	19.6	Up
NM_003810	TNFSF10	19.5	Up
BG547557	BG547557	19.5	Up
NM_004789	LHX2	18.7	Up
NM_001080535	LINCR	18.4	Up
THC2651958	THC2651958	18.2	Up
NM_001712	CEACAM1	18.2	Up
NM_001547	IFIT2	18.0	Up
NM_016135	ETV7	17.9	Up
THC2559380	THC2559380	17.9	Up
NM_000517	HBA2	17.4	Up
NM_004304	ALK	17.0	Up
NM_205848	SYT6	16.9	Up
NM_002985	CCL5	16.9	Up
NM_022147	RTP4	16.8	Up
NM_152611	C20orf75	16.6	Up
NM_014314	DDX58	16.6	Up
NM_182597	FLJ39575	16.4	Up
NM_022168	IFIH1	16.3	Up
NM_017699	SIDT1	16.1	Up
AF007190	AF007190	15.6	Up
NM_002661	PLCG2	15.6	Up
NM_014383	ZBTB32	15.6	Up
NM_005623	CCL8	15.3	Up
NM_017878	HRASLS2	15.2	Up
NM_153456	HS6ST3	15.1	Up
NM_001103	ACTN2	14.8	Up
NM_002201	ISG20	14.7	Up
NM_007365	PADI2	14.3	Up
NM_006877	GMPR	14.3	Up
NM_001008540	CXCR4	14.1	Up
NM_021804	ACE2	14.0	Up
NM_000706	AVPR1A	13.3	Up
NM_002460	IRF4	13.1	Up
BC025340	MGC39372	13.0	Up
NM_033261	IDI2	13.0	Up
NM_006158	NEFL	12.8	Up
NM_002010	FGF9	12.8	Up
NM_001549	IFIT3	12.8	Up
NM_002463	MX2	12.8	Up
AY831680	AY831680	12.6	Up
NM_175887	PRR15	12.5	Up
NM_018295	TMEM140	12.4	Up
BI910665	BI910665	12.4	Up
NM_004976	KCNC1	12.4	Up
NM_001548	IFIT1	12.3	Up
NM_020766	PCDH19	12.3	Up
NM_004848	C1orf38	12.2	Up
NM_203311	CSAG3A	12.0	Up
ENST00000292729	USP41	11.7	Up
BX110856	BX110856	11.6	Up
ENST00000301807	LBA1	11.5	Up
NM_004522	KIF5C	11.5	Up
NM_144583	ATP6V1C2	11.5	Up
NM_017414	USP18	11.4	Up
NM_014398	LAMP3	11.4	Up
NM_031917	ANGPTL6	11.2	Up
NM_002534	OAS1	11.2	Up
NM_018438	FBXO6	11.1	Up
NM_153357	SLC16A11	11.1	Up
NM_003885	CDK5R1	11.1	Up
NM_017654	SAMD9	11.1	Up
AW977362	AW977362	11.1	Up
NM_017805	RASIP1	11.0	Up
THC2657593	THC2657593	10.9	Up
NM_001729	BTC	10.6	Up
NM_005220	DLX3	10.5	Up
NM_001017403	LGR6	10.4	Up
NM_145288	ZNF342	10.4	Up
NM_000076	CDKN1C	10.3	Up
NM_005430	WNT1	9.9	Up
NM_024625	ZC3HAV1	9.9	Up
NM_002699	POU3F1	9.8	Up
ENST00000360954	HS3ST3B1	9.8	Up
NM_017554	PARP14	9.7	Up
AK023743	FLJ31033	9.6	Up
NM_003265	TLR3	9.5	Up
ENST00000302057	IRX2	9.5	Up
NM_016582	SLC15A3	9.5	Up
NM_019891	ERO1LB	9.4	Up
D00044	CCL3	9.4	Up
NM_004256	SLC22A13	9.4	Up
AF085913	AF085913	9.3	Up
NM_015900	PLA1A	9.2	Up
NM_006684	CFHR4	9.2	Up
BC029255	BC029255	9.2	Up
AB002384	C6orf32	9.2	Up
NM_024119	LGP2	9.2	Up
NM_004235	KLF4	9.2	Up
NM_152703	SAMD9L	9.1	Up
NM_004909	CSAG2	9.1	Up
AK074335	NANP	9.1	Up
NM_009587	LGALS9	9.0	Up
NM_014587	SOX8	9.0	Up
NM_004438	EPHA4	9.0	Up
NM_021096	CACNA1I	8.9	Up
NM_024490	ATP10A	8.9	Up
NM_003516	HIST2H2AA3	8.8	Up
NM_206827	RASL11A	8.7	Up
AK002042	BET3L	8.7	Up
NM_152309	PIK3AP1	8.6	Up
NM_005515	HLXB9	8.5	Up
NM_016569	TBX3	8.5	Up
NM_032206	NLRC5	8.5	Up
NM_144602	C16orf78	8.4	Up
NM_000775	CYP2J2	8.3	Up
NM_004165	RRAD	8.3	Up
NM_000246	CIITA	8.2	Up
NM_012193	FZD4	8.2	Up
NM_001080494	C1orf34	8.1	Up
AK125510	C1orf104	8.1	Up
NM_019055	ROBO4	8.0	Up
NM_004031	IRF7	7.9	Up
NM_001243	TNFRSF8	7.9	Up
AK074050	MYO1G	7.8	Up
NM_138433	KLHDC7B	7.8	Up
NM_013435	RAX	7.8	Up
NM_021065	HIST1H2AD	7.8	Up
AK091308	AK091308	7.7	Up
BM928667	BM928667	7.7	Up
NM_006186	NR4A2	7.7	Up
NM_014290	TDRD7	7.7	Up
NM_001007139	IGF2	7.6	Up
NM_138621	BCL2L11	7.6	Up
NM_178445	CCRL1	7.5	Up
NM_004673	ANGPTL1	7.5	Up
XM_211749	LOC285047	7.4	Up
AF086011	AF086011	7.4	Up
NM_021258	IL22RA1	7.4	Up
NM_004378	CRABP1	7.4	Up
BC093991	HSPB9	7.4	Up
AK095727	AK095727	7.3	Up
BC014346	BC014346	7.3	Up
NM_172140	IL29	7.3	Up
NM_000261	MYOC	7.3	Up
NM_002176	IFNB1	7.3	Up
NM_016817	OAS2	7.2	Up
NM_017639	DCHS2	7.2	Up
NM_002448	MSX1	7.2	Up
NM_173544	FAM129C	7.2	Up
NM_003655	CBX4	7.2	Up
ENST00000378953	ENST00000378953	7.1	Up
NM_207339	PAGE2	7.1	Up
NM_006914	RORB	7.1	Up
THC2526402	THC2526402	7.1	Up
NM_003328	TXK	7.0	Up
NM_153606	FAM71A	7.0	Up
NM_130467	PAGE5	7.0	Up
NM_020914	RNF213	7.0	Up
NM_002996	CX3CL1	7.0	Up
NM_198493	ANKRD45	7.0	Up
NM_004380	CREBBP	7.0	Up
NM_172109	KCNQ2	7.0	Up
NM_017523	XAF1	6.9	Up
NM_006144	GZMA	6.8	Up
AK091834	FLJ34515	6.8	Up
THC2474831	THC2474831	6.7	Up
NM_014817	KIAA0644	6.7	Up
NM_000758	CSF2	6.7	Up
NM_005853	IRX5	6.7	Up
NM_000558	HBA1	6.7	Up
NM_007332	TRPA1	6.7	Up
AL834280	AL834280	6.6	Up
AI028577	AI028577	6.6	Up
NM_003959	HIP1R	6.6	Up
NM_006228	PNOC	6.5	Up
NM_153479	CSAG1	6.5	Up
AK025221	LOC441108	6.5	Up
NM_002374	MAP2	6.5	Up
NM_015660	GIMAP2	6.5	Up
NM_020715	PLEKHH1	6.5	Up
ENST00000324559	TMEM16E	6.5	Up
ENST00000332844	ENST00000332844	6.4	Up
NM_000359	TGM1	6.4	Up
NM_172200	IL15RA	6.4	Up
NM_006820	IFI44L	6.4	Up
NM_003141	TRIM21	6.3	Up
NM_003811	TNFSF9	6.3	Up
NM_002147	HOXB5	6.3	Up
NM_007335	DLEC1	6.3	Up
AL117481	DKFZP434B061	6.3	Up
W91942	W91942	6.3	Up
BC073918	BC073918	6.2	Up
NM_000593	TAP1	6.2	Up
BC041467	C17orf67	6.2	Up
NM_006781	C6orf10	6.2	Up
NM_012465	TLL2	6.2	Up
NM_145641	APOL3	6.2	Up
NM_003807	TNFSF14	6.2	Up
NM_032727	INA	6.2	Up
NM_004364	CEBPA	6.1	Up
NM_021052	HIST1H2AE	6.1	Up
AK055279	C8orf53	6.1	Up
ENST00000269499	ZCCHC2	6.1	Up
ENST00000369158	ENST00000369158	6.1	Up
NM_021127	PMAIP1	6.1	Up
NM_005982	SIX1	6.0	Up
AV756170	AV756170	6.0	Up
NM_001040078	LOC654346	6.0	Up
NM_021822	APOBEC3G	5.9	Up
NM_013351	TBX21	5.9	Up
NM_005252	FOS	5.9	Up
DQ786194	DQ786194	5.9	Up
NM_004566	PFKFB3	5.9	Up
NM_006922	SCN3A	5.9	Up
THC2663297	THC2663297	5.9	Up
NM_004496	FOXA1	5.9	Up
NM_000683	ADRA2C	5.8	Up
NM_030641	APOL6	5.8	Up
NM_174896	C1orf162	5.8	Up
NM_012367	OR2B6	5.8	Up
NM_006074	TRIM22	5.8	Up
NM_057157	CYP26A1	5.8	Up
NM_002557	OVGP1	5.8	Up
ENST00000390253	ENST00000390253	5.8	Up
NM_017912	HERC6	5.8	Up
NM_014755	SERTAD2	5.7	Up
ENST00000377186	ENST00000377186	5.7	Up
NM_016642	SPTBN5	5.7	Up
NM_000759	CSF3	5.7	Up
NM_005533	IFI35	5.7	Up
ENST00000382595	FAM90A9	5.7	Up
NM_032265	ZMYND15	5.7	Up
NM_002468	MYD88	5.7	Up
NM_030930	UNC93B1	5.7	Up
NM_005170	ASCL2	5.6	Up
BQ213856	BQ213856	5.6	Up
NM_138819	FAM122C	5.6	Up
NM_006671	SLC1A7	5.6	Up
BC016934	SOD2	5.6	Up
THC2661318	THC2661318	5.6	Up
NM_170699	GPBAR1	5.6	Up
NM_003882	WISP1	5.6	Up
NM_004510	SP110	5.6	Up
NM_004585	RARRES3	5.5	Up
NM_019885	CYP26B1	5.5	Up
NM_005557	KRT16	5.5	Up
NM_006399	BATF	5.5	Up
NM_023940	RASL11B	5.5	Up
ENST00000303310	ENST00000303310	5.4	Up
AK125162	AK125162	5.4	Up
AA455656	AA455656	5.4	Up
NM_152431	PIWIL4	5.4	Up
ENST00000358378	ENST00000358378	5.4	Up
NM_003277	CLDN5	5.4	Up
NM_024783	AGBL2	5.3	Up
NM_002700	POU4F3	5.3	Up
NM_005980	S100P	5.3	Up
NM_031212	SLC25A28	5.3	Up
NM_006620	HBS1L	5.3	Up
NM_021035	ZNFX1	5.3	Up
NM_173042	IL18BP	5.3	Up
NM_033238	PML	5.2	Up
DB518505	DB518505	5.2	Up
NM_017709	FAM46C	5.2	Up
AF305819	AF305819	5.2	Up
NM_000161	GCH1	5.2	Up
NM_025079	ZC3H12A	5.2	Up
NM_019001	XRN1	5.2	Up
NM_020904	PLEKHA4	5.2	Up
NM_032784	RSPO3	5.1	Up
NM_153341	IBRDC3	5.1	Up
NM_153610	CMYA5	5.1	Up
NM_022750	PARP12	5.1	Up
NM_033292	CASP1	5.1	Up
BX109076	BX109076	5.1	Up
AK090515	LOC283663	5.1	Up
BC037791	BC037791	5.0	Up
NM_080552	SLC32A1	5.0	Up
NM_006187	OAS3	5.0	Up
AJ295982	AJ295982	5.0	Up
NM_025179	PLXNA2	5.0	Up
NM_033109	PNPT1	5.0	Up
NM_178140	PDZD2	5.0	Up
NM_005101	ISG15	5.0	Up
NM_017631	FLJ20035	5.0	Up
NM_021105	PLSCR1	5.0	Up
NM_000070	CAPN3	5.0	Up
NM_052886	MAL2	5.0	Up
AK023773	AK023773	4.9	Up
AA573434	AA573434	4.9	Up
NM_033255	EPSTI1	4.9	Up
NM_001080391	SP100	4.9	Up
NM_203393	LOC389458	4.9	Up
NM_014850	SRGAP3	4.9	Up
AK021546	AK021546	4.9	Up
AK056817	FLJ32255	4.9	Up
NM_006576	AVIL	4.9	Up
AK056190	DFNB31	4.9	Up
ENST00000339446	LOC387763	4.9	Up
NM_002135	NR4A1	4.9	Up
NM_004688	NMI	4.9	Up
NM_002198	IRF1	4.9	Up
NM_000901	NR3C2	4.8	Up
AL117235	PTCHD2	4.8	Up
NM_001394	DUSP4	4.8	Up
NM_024956	TMEM62	4.8	Up
NM_003641	IFITM1	4.8	Up
NM_006095	ATP8A1	4.8	Up
NM_014080	DUOX2	4.8	Up
NM_138287	DTX3L	4.8	Up
THC2688196	THC2688196	4.8	Up
BI024548	BI024548	4.8	Up
NM_145019	FAM124A	4.8	Up
NM_006417	IFI44	4.8	Up
AK127223	LOC284296	4.7	Up
NM_006869	CENTA1	4.7	Up
NM_005248	FGR	4.7	Up
NM_002462	MX1	4.7	Up
NM_002286	LAG3	4.7	Up
BC010906	MED9	4.7	Up
NM_152612	CCDC116	4.7	Up
BU561469	BU561469	4.7	Up
NM_203446	SYNJ1	4.7	Up
NM_006424	SLC34A2	4.7	Up
NM_025195	TRIB1	4.7	Up
NM_002218	ITIH4	4.7	Up
NM_030766	BCL2L14	4.7	Up
NM_024778	LONRF3	4.7	Up
NM_021784	FOXA2	4.6	Up
AK027294	AK027294	4.6	Up
NM_175839	SMOX	4.6	Up
NM_006290	TNFAIP3	4.6	Up
BC013171	BC013171	4.6	Up
BC031266	TRIM69	4.6	Up
NM_015907	LAP3	4.6	Up
NM_005419	STAT2	4.6	Up
BC031319	BC031319	4.6	Up
NM_001066	TNFRSF1B	4.6	Up
THC2673062	THC2673062	4.6	Up
THC2513333	THC2513333	4.5	Up
NM_145053	UBQLNL	4.5	Up
NM_018381	FLJ11286	4.5	Up
NM_175065	HIST2H2AB	4.5	Up
NM_000600	IL6	4.5	Up
NM_024913	FLJ21986	4.5	Up
NM_019102	HOXA5	4.5	Up
NM_017539	DNAH3	4.5	Up
AK094730	LOC283454	4.5	Up
NM_173086	KRT6C	4.5	Up
NM_012420	IFIT5	4.5	Up
NM_002053	GBP1	4.5	Up
XR_017251	LOC389386	4.4	Up
BC014971	BC014971	4.4	Up
BC035583	KIAA0999	4.4	Up
NM_080829	C20orf175	4.4	Up
BC008632	C6orf176	4.4	Up
NM_198183	UBE2L6	4.4	Up
NM_139266	STAT1	4.4	Up
NM_014400	LYPD3	4.4	Up
NM_000307	POU3F4	4.4	Up
NM_178516	EXOC3L	4.4	Up
NM_145637	APOL2	4.4	Up
NM_004821	HAND1	4.4	Up
NM_173490	TMEM171	4.4	Up
NM_002089	CXCL2	4.4	Up
NM_138422	LOC113179	4.4	Up
NM_006018	GPR109B	4.3	Up
NM_173198	NR4A3	4.3	Up
ENST00000367675	ENST00000367675	4.3	Up
NM_018964	SLC37A1	4.3	Up
ENST00000319902	KIAA1618	4.3	Up
XR_019109	LOC650517	4.3	Up
NM_001823	CKB	4.3	Up
NM_001451	FOXF1	4.3	Up
NM_000882	IL12A	4.3	Up
NM_194284	CLDN23	4.3	Up
NM_004405	DLX2	4.3	Up
NM_130436	DYRK1A	4.3	Up
XR_015273	LOC728371	4.2	Up
BU681302	PPP2R2D	4.2	Up
NM_031458	PARP9	4.2	Up
NM_003335	UBE1L	4.2	Up
NM_003649	DDO	4.2	Up
NM_004184	WARS	4.2	Up
CD556746	CD556746	4.2	Up
NM_145343	APOL1	4.2	Up
NM_001050	SSTR2	4.2	Up
BX110908	BX110908	4.1	Up
AK128592	DNHD2	4.1	Up
BC024745	BC024745	4.1	Up
ENST00000382591	FAM90A10	4.1	Up
AK226060	BUB3	4.1	Up
NM_001017534	COP1	4.1	Up
NM_014059	C13orf15	4.1	Up
AL049782	CG012	4.1	Up
NR_002139	HCG4	4.1	Up
NM_001003845	SP5	4.1	Up
NM_015564	LRRTM2	4.1	Up
NM_053001	OSR2	4.1	Up
BQ130147	BQ130147	4.0	Up
ENST00000294663	GBP2	4.0	Up
NM_004024	ATF3	4.0	Up
AK092450	AK092450	4.0	Up
BC000772	BC000772	4.0	Up
NM_016585	THEG	4.0	Up
NM_024522	FAM77C	4.0	Up
NM_173649	FLJ40172	4.0	Up
NM_024989	PGAP1	4.0	Up
NM_152542	PPM1K	4.0	Up
AL833749	LOC146439	4.0	Up
NM_014177	C18orf55	5.2	Down
NM_025184	EFHC2	4.9	Down
NM_001629	ALOX5AP	4.5	Down
NM_022469	GREM2	4.4	Down
AK126014	KIAA1211	4.3	Down
NM_000812	GABRB1	4.2	Down
ENST00000379426	ENST00000379426	4.2	Down
NM_014905	GLS	4.0	Down

  P < 0.01.

*   http://www.ncbi.nlm.nih.gov/Genbank/, National Center for Biotechnology Information, Bethesda, MD.

The upregulated genes at the highest ratio were RAS, dexamethasone-induced 1 (RASD1), δ-like 1 (DLL1), SRY-box 3 (SOX3), activity-regulated cytoskeleton-associated protein (ARC), thyroxine deiodinase type III (DIO3), indoleamine 2,3-dioxygenase 1 (IDO1), FLJ00049, and 10 kDa interferon-γ–induced protein (IP-10, CXCL10). 

The downregulated genes at the highest ratio were chromosome 18 open reading frame 55 (C18orf55), EF-hand domain (C-terminal)–containing 2 (EFHC2), and arachidonate 5-lipoxygenase-activating protein (ALOX5AP), which is required for leukotriene synthesis with 5-lipoxygenase. 

Network Analysis of Upregulated or Downregulated Genes in HSV-Infected Human Corneal Endothelial Cells

To obtain a global view of HSV infection–induced phenomena in the HCEn cells, 330 genes were extracted from the 461 genes (fourfold difference, P < 0.01) and were analyzed for signaling interactions using a systems biological approach. 

Using the data set of 330 genes, functional analysis was used to reveal functional associations with the HCEn transcriptome. The highest significant association was detected for antigen presentation function, and as much as 23% of the data set was associated with this function (Table 2). This analysis was followed by determining significant associations with antimicrobial response function and cell-mediated immune response (data not shown). 

Table 2.

 

Molecules Significantly Associated with Antigen Presentation, as Revealed by Functional Analysis

Molecules in Network

P

IL15RA, AIM2, IL6, TBX21, IFIH1, APOL3, CXCL10, SOD2, IFI44L, TNFSF9, CCL8, FGR, TNFSF13B, GZMA, DLL1, CXCL9, TRIM21, HSH2D, ZC3HAV1, LAG3, CSF3, IRF1, APOBEC3G, IL18BP, IRF7, DUOX2, PLCG2, DDX58, RARRES3, IDO1, PNOC, PIK3AP1, STAT2, IL29, TRIM22, CX3CL1, IL12A, IFNB1, CIITA, TNFSF10, TNFAIP3, CCL5, CCL3, LGALS9, SECTM1, TAP1, NGFR, CASP1, CD38, CCRL1, GBP2, TLR3, TNFRSF1B, STAT1, PLSCR1, CXCL11, DHX58, MX2, OAS1, IRF4, MYD88, CXCR4, ALOX5AP, MX1, UNC93B1, IFI44, APOL1, CEACAM1, FOS, ZBTB32, NMI, CXCL2, CSF2, ISG20, BCL2L11, and TNFSF14

5.05 × 10−18–5.34 × 10−4

We next applied canonical pathway analysis to the data set to reveal their relative associations with the pathways. The results shown in Table 3 showed that the HCEn transcriptome is heavily favored toward interferon signaling as the primary pathway. The second association was with the pattern recognition receptor pathway, which would recognize HSV. These associations were consistent with their function as antigen-presenting cells (APCs). 

Table 3.

 

Canonical Pathway Analysis of HSV-1–Induced Transcriptome of Corneal Endothelial Cells

Canonical Pathway	P	Ratio
Interferon signaling	1.11 × 10−11	11/29 (0.379)
Role of patternr ecognition receptors in recognition of bacteria and viruses	1.34 × 10−10	15/88 (0.17)
Activation of IRF by cytosolic pattern recognition receptors	1.2 × 10−7	11/74 (0.149)
IL-15 Production	2.18 × 10−4	5/30 (0.167)
Role of RIG1-like receptors in antiviral innate immunity	4.12 × 10−4	6/52 (0.115)

To obtain a global view of biological interactions in the data set, we applied network analysis using the data base (Pathways Knowledge Base; Ingenuity Systems) of known signaling networks. We successfully generated five major biological networks with their significance scores (P < 10⁻³⁰; Table 4, Fig. 1). 

Table 4.

 

Transcriptional Networks of HSV-1–Infected Corneal Endothelial Cells

Network	Molecules in Network	Score (−log P)	Function
1	AIM2, BATF, DDX58(RIG-1), DHX58, FOXF1, IFI35, IFI44, IFIH1 (MDA5), IFIT1, IFIT2, IFIT3, IFN&α;/β, IL-29, IRF, IRF1, IRF7, ISG15, ISGF3(IRF9), NFkB (complex), Oas, OAS1, OAS2, OAS3, PARP9, RARRES3, REL/RELA/RELB, RNF19B, RSAD2, S100P, SP110, STAT2, Stat1–Stat2, TLR3, TNFSF9 (4–1BB-L), and TRIM69	44	Antigen presentation, antimicrobial response, cell-mediated immune response
2	ADRA2C, ALK, ANGPTL1, ASCL2, BCL2L14, BCR, BTC, CLDN5, CSF3, CX3CL1, CXCR4, DIO3, DUSP4, EPHA4, ERK, Fcer1, Fgf, FGF9, Ige, INSM1, MAP2K1/2, OVGP1, PLC gamma, PLCG2, PPP2R2D, Rap1, RASD1, RASGRP3, RET, Stat1 dimer, SYK/ZAP, SYNJ1, TBX21, TRIB1, and TXK	40	Cellular development, hematological system development and function, hematopoiesis
3	BCL2L11, Caspase, CCL3, CCL5, CCL8, CCRL1, Cytochrome c, FZD4, GLS, GZMA, HBA1, HBA2, HSH2D, IFITM1, IgG, Igm, Ikb, IKK (complex), IL1, IL12 (complex), IL12 (family), IL12A, IRF4, ISG20, Jnk, KRT16, MYD88, NGFR, RRAD, SP100, STAT1, Tnf receptor, TNFRSF1B, TNFSF10, and TNFSF13B	38	Cell-to-cell signaling and interaction, hematological system development and function immune cell trafficking
4	BUB3, CD38, CXCL2, CXCL9, CXCL10, CXCL11, GBP2, HLA-DR, IDO1, IFN Beta, Ifn gamma, IFNB1, IL6, IL15RA, IL18BP, Interferon alpha, IRF3 dimer, MHC CLASS I (family), MX1, MX2, NF-kappaB (family), NfkB-RelA, Nucleotidyltransferase, PMAIP1, PNPT1, Sod, TAP1, Tlr, TNFAIP3, TNFRSF8, TNFSF14(HVEML), TRAF, USP18, WARS, and XAF1	36	Antigen presentation, cell-mediated immune response, humoral immune response
5	ALOX5AP, ATF3, Cbp/p300, CEBPA, CIITA, CREBBP, CYP26A1, FOXA1, FOXA2, GBP1, GCH1, Growth hormone, HISTONE, Histone h3, Histone h4, HOXA5, KLF4, MHC Class I (complex), N-cor, NIACR2, NMI, NR3C2, P38 MAPK, PEPCK, PML, Rar, Rxr, SECTM1, Sox, SOX3, SOX8, SOX17, SWI-SNF, TRIM21, and Vitamin D3-VDR-retinoid X receptor, gamma	30	Cellular growth and proliferation, embryonic development, gene expression

Figure 1.

 

Network analysis of the biological processes underlying the HSV-1 infection–induced responses of HCEn cells. Networks 1 to 5 are summarized as the merged networks. Interactions between the networks are shown as yellow lines.

Network 1 provided the highest significance score (P < 10⁻⁴⁴) and was represented by interferons including IL-29, interferon regulatory transcription factors (IRFs), and interferon-responsive genes including absent in melanoma 2 (AIM2), interferon-induced proteins (IFI), melanoma differentiation associated protein-5 (IFIH1, MDA5), interferon-induced proteins with tetratricopeptide repeats (IFIT), interferon-stimulated protein, 15 kDa (ISG15), 2′-5′-oligoadenylate synthetase (OAS), radical S-adenosyl methionine domain–containing 2 (RSAD2), SP110 nuclear body protein (SP110), and signal transducer and activator of transcription 1 (STAT1) and STAT2. 

Another category in network 1 was the recognition of dsRNA and related molecules. This network included Toll-like receptor (TLR) 3, DEAD box polypeptide 58 (DDX58, RIG-1), IFIH1 (MDA5), RIG-I-like receptor LGP2 (DHX58), and OAS. Of these, TLR3, DDX58 (RIG-1), and MDA5 are representative sensors of dsRNA. In addition, tumor necrosis factor ligand superfamily member 9 (TNFSF9, 4–1BB-L), which are crucial costimulatory molecules for antigen presentation to induce T lymphocyte proliferation, were found to be significantly associated with this network. Thus, network 1 was annotated as antigen presentation, antimicrobial responses, and cell-mediated immune responses. Activation of this network was calculated to be significantly associated with the NF-κB cascade. 

Network 2, with the second highest significance score (P < 10⁻⁴⁰), was annotated as cellular development, hematologic system development and function, and hematopoiesis. This network was summarized to the extracellular signal-regulated kinase (ERK) cascade, and granulocyte colony-stimulating factor (CSF3), CXC chemokines receptor 4 (CXCR4), phospholipase C (PLC) gamma, and spleen tyrosine kinase (SYK)/ζ-associated protein (ZAP) served as crucial nodes. 

Network 3 was annotated as cell-to-cell signaling and interaction, hematologic system development and function, and immune cell trafficking. This network included CCL3 (MIP-1α), CCL5 (RANTES), interleukin (IL)-12 (IL12, IL-12), and tumor necrosis factor (TNF) ligand family molecules, including TNF superfamily, member 13b (TNFSF13B, BAFF), TNF ligand superfamily member 10 (TNFSF10, TRAIL), and TNF receptor superfamily, member 1B (TNFRSF1B, TNFR-2). 

Network 4 was another significant antigen-presentation, function-related network, which was annotated as antigen presentation, cell-mediated immune response, and humoral immune response. This network included antigen-presentation–related genes, Th1-related chemokines and cytokines, and interleukin 6 (IL6, IL-6), which will determine the type of T lymphocyte responses. Essential components of the antigen presentation machinery, including MHC molecules and transporter associated with antigen processing (TAP1), were found in this network. Herpes virus entry mediator-ligand (TNFSF14, HVEML) found in this network is a co-stimulatory factor for T cells to interact with APCs and acts as a receptor for HSV. 

Network 5 was annotated as cellular growth and proliferation, embryonic development, and gene expression. In the context of antigen presentation, Class II Major Histocompatibility Complex and transactivator (CIITA), a master transcriptional regulator essential for class II expression, was located in this network. Network 5 was also characterized by nuclear transcriptional regulation-related genes: cAMP-response element binding protein (Cbp)/p300, CCAAT/enhancer binding protein alpha (CEBPA), CREB binding protein (CREBBP), activating transcription factor 3 (ATF3), histone, and retinoic acid receptor-related genes (cytochrome P450, family 26, subfamily A, polypeptide 1; CYP26A1), nuclear receptor co-repressor 1 (N-cor), promyelocytic leukemia (PML), retinoid acid receptor (Rar), retinoid X receptor (Rxr), and SWI/SNF complex. The retinoid acid receptor family genes are nuclear receptors and act as transcriptional repressors, which are involved in antiproliferative effects of retinoic acid. 

Corneal Endothelial Responses to HSV-1 in Common with Corneal Epithelial Cells

To understand the specific responses of the HCEn cells, we then compared the transcriptome of HCEn cells and the reported transcriptome of HCEp cells after HSV infection (12 hours PI). ⁸ Of the 10 highest induced genes in HCEn cells, RASD1, DLL1, SOX3, ARC, DIO3, FLJ00049, and tripartite motif-containing 43 (TRIM43) were also observed in the transcriptome of HCEp cells. In contrast, IDO1 and IP-10 were observed only in the HCEn transcriptome. Therefore, we reasoned that the networks of the HCEn cells represent general antiviral responses to HSV and corneal endothelium–specific responses. 

To delineate the general responses of HCEn cells to HSV, we constructed shared networks using genes detected in the transcriptomes of both HCEn and HCEp. IPA generated two major biological networks with high significance scores (P < 10⁻³⁰; Table 5). Shared network 1 was annotated as embryonic development, tissue development, and skeletal and muscular system development and function. Shared network 1 was characterized by interferon response, MAPK, and NFκB cascades. IL-12, chemokine (C-X-C motif) ligand 2 (CXCL2), and fibroblast growth factor 9 (FGF9) were identified as shared mediators (Table 5). As a co-stimulatory molecule, TNFSF9 (4–1BB-L) was also observed in this network. Shared network 2 was annotated as cellular development, hematologic system development and function, and hematopoiesis and was characterized by retinoic acid metabolism. 

Table 5.

 

Transcriptional Networks Shared with Corneal Endothelial and Epithelial Cells after HSV-1 Infection

Network	Molecules	Score (−log P)	Functions
1	ALP, CaMKII, CBX4, CDK5R1, CDKN1C, Ck2, CRABP1, Creb, CREBBP, CXCL2, Cyclin A, DLL1, DLX2, FGF9, FOS, FOXF1, GADD45G, Gsk3, IGF2, IL12 (complex), Interferon alpha, KCNC1, LDL, MSX1, MUC5AC, NFkB (complex), P38 MAPK, PDGF BB, RRAD, STAT5a/b, TBX21, Tgf beta, TNFSF9, TRIB1, and WNT1	43	Embryonic development, tissue development, skeletal and muscular system development and function
2	ARC, BST2, CCDC116, CLDN5, CLDN10, ELAVL3, EWSR1, GBX2, GLI1, HTT, IRX4, KCNQ2, LOC387763, MDFI (includes EG:4188), MED9, MIR18A, NEFL, NEFM, NPM1 (includes EG:4869), PAX3, PCDH8, POU5F1, PPARγ ligand-PPARγ-Retinoic acid-RARα, PTCHD2, RARA, RASL11B, retinoic acid, Retinoic acid-RAR, RPS17 (includes EG:20068), SIK3, SOX3, SYNJ1, TBX15, YWHAZ, and ZNF133	33	Cellular development, hematological system development and function, hematopoiesis

Corneal Endothelial Responses to HSV-1 Distinct from Corneal Epithelial Cells

Next, we analyzed how HCEn cells respond to HSV-1 infection. Genes in the transcriptome of HCEn cells that were shared with HCEp with more than fourfold difference compared to the mock-infection control were eliminated. After complementing with statistically significant connecting nodes, the IPA generated four major biological networks with high significance scores (P < 10⁻³⁰; Table 6). 

Table 6.

 

Transcriptional Networks Preferred by Corneal Endothelial Cells after HSV-1 Infection

Network	Molecules in network	Score (−log P)	Functions
1	AIM2, BATF, DDX58, DHX58, IFI35, IFI44, IFIH1, IFIT1, IFIT2, IFIT3, Ifn, IFN TYPE 1, IL29, IL18BP, Interferon-α Induced, IRF, IRF7, ISG15, ISGF3, MX1, MX2, NFkB (complex), OAS1, OAS3, PARP9, RARRES3, REL/RELA/RELB, RNF19B, RSAD2, S100P, STAT2, Stat1–Stat2, TNFRSF8, TRIM69, and UBA7	41	Antimicrobial response, inflammatory response, infection mechanism
2	Akt, ALOX5AP, ANGPTL1, C13ORF15, CD38, CIITA, Collagen Alpha1, CSF2, CYP2J2, FAM65B, GBP1, GLS, GPR109B, Growth hormone, HSH2D, IFITM1, Ikb, IKK (complex), Ikk (family), IL1, Interferon Regulatory Factor, IRF1, IRF4, LDL, NfkB1-RelA, NR4A3, PARP, PARP12, PAR P14, Tnf receptor, TNFAIP3, TNFRSF1B, TNFSF10, TRAF, and WISP1	33	Cell death, cellular growth and proliferation, connective tissue development and function
3	APOBEC3G, APOL2, APOL3 (includes EG:80833), ATF3, CCL3, CCL5, CEBPA, CSF3, CYP26A1, FOXA1, FOXA2, GCH1, HOXA5, IgG, Igm, IL12 (complex), Interferon alpha, IRF3 dimer, ISG20, KLF4, KRT16, Nfat (family), NR4A2, P38 MAPK, PEPCK, PMAIP1, Rxr, SECTM 1, Sod, SOD2, TRIL, TRIM21, UNC93B1, VitaminD3-VDR-RXR, and ZC3HAV1	31	Infection mechanism, infectious disease, embryonic development
4	ACE2, BUB3, CCL8, CCRL1, CHEMOKINE, CXCL9, CXCL10, CXCL11, GBP2, HLA-DR, IDO1, IFN Beta, Ifn gamma, Ifnar, IFNB1, IFNα/β, Iga, IL6, IL23, IL12 (family), IL12A, IRAK, MYD88, NfkB-RelA, Oas, OAS2, Pro-inflammatory Cytokine, SMOX, SP110, Tlr, TLR3, TNFSF14, TNFSF13B, WARS, and XAF1	30	Cell-to-cell signaling and interaction, hematological system development and function, cellular movement

The HCEn-preferred network 1 of highest significance was annotated as antimicrobial responses, inflammatory responses, and infection mechanisms. This network was characterized mainly by interferon responses. In network 2, antigen-presentation–related genes, TNFSF10 (TRAIL), TNFRSF1B (TNFR-2), and CIITA, and granulocyte-macrophage colony stimulating factor (CSF2, GMCSF), were identified. Network 3 was annotated as infection mechanism, infectious disease, embryonic development, and was characterized by antiviral mediators including CCL3, CCL5 (RANTES), IL-12, and interferon α. HCEn-preferred network 4 was annotated as cell-to-cell signaling and interaction, hematologic system development and function, and cellular movement. This network was characterized by antigen presentation and lymphocyte function-determinant–related genes, including IDO1, HLA-DR, TNFSF14 (HVEML), CXCL9, CXCL10 (IP-10), CXCL11, interferons, IL-6, and IL-12. Thus, all four HCEn-preferred networks were found to share or to be involved in the antigen-presentation–related function. 

Production of Inflammatory Cytokines by HSV-1–Infected Corneal Endothelial Cells

We examined whether the observed transcriptional responses may indeed translate into a special profile of secreted proteins. The supernatant collected from HSV-1–infected HCEn cells at 12 hours PI was analyzed for a cytokine secretion profile by using protein array analysis. Significant increases in the secretion of IL-6, IL-8, monocyte chemotactic protein 1 (MCP-1, CCL2), tissue inhibitor of metalloproteinase 1 (TIMP-1), RANTES (CCL5), IP-10, I-309, macrophage migration inhibitory factor (MIF), monocyte chemoattractant protein 2 (MCP-2, CCL8), TNFSF14 (HVEML), IL-10, stromal cell-derived factor 1 (SDF-1), and interferon-γ were found in a descending order (Fig. 2). Antigen-presentation–related genes, including IL-6, IP-10, CCL8, HVEML, and interferon-γ, were confirmed for induction by HSV-1 infection. 

Figure 2.

 

Induction profile of inflammatory cytokines by HSV-1–infected HCEn cells. HCEn cells were adsorbed with HSV-1 at an MOI of 0.1 for 1 hour and refed with DMEM. After 12 hours' incubation, the supernatant of HSV-1–infected HCEn cells was assayed with a cytokine array. A panel of inflammatory cytokines was significantly induced by HSV-1 infection. n = 4 per group. P < 0.05.

Priming of Allogeneic T Lymphocytes by HSV-1–Infected Corneal Endothelial Cells

Finally, we tested whether HCEns may indeed function as APCs. HCEn cells infected with HSV-1 were treated with MMC to suppress DNA synthesis and proliferation, and then co-cultured with allogeneic T cells from donors previously infected with HSV-1. The proliferation of CD4⁺ T cells measured by BrdU uptake was significantly stimulated by HSV-1–primed HCEn cells at an MOI of 5 (Fig. 3A). Allogeneic T cells from healthy donors without a history of HSV-1 infection did not show an appreciable stimulatory effect (data not shown). For the control of allogeneic responses, Vero cells (kidney epithelial cells derived from the African Green monkey) were used as a stimulator. As expected, no significant T-cell proliferation was observed (Fig. 3B) To confirm HSV-1–stimulated allogeneic responses provoke a Th1-type response, we assessed interferon-γ secretion. HSV-1–primed HCEn cells stimulate allogeneic T cells to produce significant amounts of interferon-γ (Fig. 3C). No interferon-γ was produced by T cells cocultured with HSV-1 primed Vero cells. 

Figure 3.

 

HSV-1–specific T-cell proliferation and interferon-γ secretion stimulated by HSV-1–treated HCEn cells. HCEn cells were adsorbed with HSV-1 for 1 hour at the indicated MOI and treated with mitomycin C. HSV-1–specific CD4⁺ T cells were isolated from HSV-1 infected-allogeneic donors, and co-cultured with the HSV-1–adsorbed HCEn cells (responder: stimulator ratio, 4:1) (A) or xenogeneic Vero cells as the control (B). HSV-1–specific T-cell proliferation was examined by BrdU uptake, which was assessed with a chemiluminescence-based, BrdU-specific ELISA. HSV-1–specific interferon secretion from the T cells co-cultured with the HSV-1–primed HCEns or Vero cells was measured with ELISA (C). The HSV-1–primed HCEn cells, but not the Vero cells, significantly stimulated interferon-γ secretion. *P < 0.05.

Discussion

Our results showed that HSV infection affected the expression of numerous genes, and the majority of the mRNAs were transcriptionally activated. Importantly, our bioinformatics analysis of HSV-induced transcriptome of HCEn cells showed that the molecular signature profile of these genes is strongly directed to initiate the acquired immune system as APCs. 

Generally, HSV infection induces global silencing of host-derived transcripts. ^{12 –14} This is mediated by viral proteins or by the immediate early genes including ICP0, ICP27, or ICP34.5. ^12,15,16 Thus, global transcriptional activation after HSV infection, which was also observed in HCEp cells, ^8,17 appears to be an uncommon event. Epithelial cells, including HCEps, are part of the primary defense system that initiates an arsenal of proinflammatory mediators on infection. This may explain the presumed evolutionary requirement to resist transcriptional silencing exploited by HSV. In contrast, corneal endothelial cells are located inside the eye and are not easily accessible to HSV, which is different from corneal epithelial cells. Thus, the resistance of HCEn cells to transcriptional silencing appears to reflect a specific property. 

On infection, HSV hijacks the transcriptional machinery of the host and diverts the canonical NFκB inflammatory signaling cascade for its own replication. ¹⁸ To resist viral replication, HCEn cells initiate an antiviral program with global transcriptional activation, which results in the release of inflammatory cytokines (Fig. 2). These cytokines subsequently prime the acquired cellular immunity to protect the corneal endothelial cells. 

In the HSV-induced transcriptome of HCEn cells, we detected several antigen-presentation–related genes. For example, MHC class II is used for the presentation of exogenous proteins. The expression of MHC class II is regulated by the master transcriptional regulator CIITA (Fig. 1, Tables 4, 6), which is a signature molecule of professional APCs. Moreover, HCEn cells express the co-stimulatory molecules CD80 and CD86 on the cell surface, and they are stimulated to express CD40 after interferon-γ stimulation. ⁶ All these molecules are essential for APCs to provide the appropriate strength of antigen stimuli to recognize T-cell receptors. 

Another important signal for APC function is a maturation stimulus, which is typically mediated by GMCSF (CSF2; Tables 1, 6; Fig. 1). Thus, these observations further support the functional capability of HCEn cells as APCs after HSV-1 infection. 

Recently, an analysis of the transcriptional signature of the genome of dendritic cell (DC) responding to pathogen stimuli has been conducted, and crucial regulatory circuits were found. These circuits comprised 125 transcription factors, chromatin modifiers, and RNA-binding proteins. ¹⁹ The study showed that the responses of dendritic cells to pathogens consisted essentially of inflammatory and antiviral programs. In the inflammatory program, IL-6, IL-12, CXCL2, and IL-1β were representative effector molecules, ¹⁹ and, in our study, these molecules were found in major networks 3 and 4 of HCEn cells (Table 4). In contrast, IP-10 (CXCL10), interferon-stimulated protein, 15 kDa (ISG15), and interferon-induced protein with tetratricopeptide repeats 1 (IFIT1) are other representatives of antiviral programs. ¹⁹ They were identified in networks 1 and 4 of the HCEn cells (Table 4). Interestingly, IP-10 was the eighth highest induced gene in the HCEn transcriptome (Table 1). IP-10 directly inhibits HSV-1 replication. ²⁰  

In dendritic cells, antiviral programs are initiated by viral sensors, including TLRs, melanoma differentiation associated protein-5 (MDA5, IFIH1), and DDX58 (RIG-1), which again are found in network 1 of HCEn cells (Table 4). In contrast, HCEp networks were clearly distinctive in their transcriptional profile, and their identified nodes did not match those of dendritic cells. Thus, the representative transcriptional network nodes of HCEn cells are essentially matched to those for representative effector molecules in dendritic cells. 

In the antiviral program of dendritic cells, signal transducer and activator of transcription 1 (STAT1) and STAT2 regulate components of the antiviral effector molecules. Consistent with this, STAT1 was positioned centrally in the transcriptional network of HCEn cells (Fig. 1). Other representative transcriptional regulators of the antiviral program in dendritic cells were IRF8, IRF9, activating transcription factor 3 (ATF3), ets variant 6 (ETV6), JUN, STAT4, and retinoblastoma-like 1 (RBL1). Of these, the IRFs (network 1) and ATF3 (network 5) were also found in HCEn cell networks. This result is consistent with the functional capability of HCEn cells as APCs, and may also reflect the HCEn cell–specific responses to pathogens. 

Our results showed that HCEns can function as APCs. Generally, HSV-1 is known to block antigen presentation of infected cells. ^15,21 The observed allopriming effect of HCEn cells would be beneficial for the effective eradication of HSV-infected cells. On the other hand, such elimination may lead to endothelial cell loss, which could lead to potentially blinding bullous keratopathy. So, how does the host avoid such a deleterious phenomenon? It has been shown that HCEns can serve as immune regulatory cells that dampen the cytotoxic effects induced by activated T cells. This action may protect the endothelial cells from death while maintaining their priming of immunologic memory responses. For example, HCEn cells impair Th1 CD4⁺ cells by PD-L1, which is strongly expressed on its surface. ⁶ CD8⁺ T cells can also be converted to regulatory T cells by HCEn cells via TGF-β. ⁷ In the HSV-1–induced transcriptome of HCEn cells, the sixth highest induced gene, IDO1, produces an immune regulatory enzyme that induces anergy or regulatory T-cell differentiation. In the HCEn networks, nuclear receptor transrepression pathways appear also to regulate inflammation by N-cor or Rxr, which are representative repressors of inflammatory responsive promoters (Table 4, network 5). ²²  

The most striking difference between the HCEn and HCEp transcriptional networks was the interferon-related response. This result is consistent with their functional ability of antigen presentation. Interferons induce representative antiviral responses and modulate the immune system, and they render neighboring cells resistant to viral infection. In general, interferon responses are commonly observed after viral infection, including human cytomegalovirus. ^23,24 In contrast, interferon responses are generally silenced in cases of HSV infection. ^14,20 This silencing does not occur when viral replication is impaired. ¹⁴ After HSV infection of HCEn cells, we observed an induction of known interferon-inducible antiviral genes, including OAS1/2/3, and myxovirus resistance 1 (MX1)/2 in networks 1 and 4, respectively. OAS activates latent RNase L to induce viral RNA degradation. ²⁵ The MX proteins are dynamin superfamily GTPases that interfere with viral replication. ²⁶ Thus, HCEn cells have a strong propensity for interferon-related antiviral or inflammatory programs to resist HSV-1 infection, despite their susceptibility to infection. 

The HSV-induced host genes of HCEn cells determined by network analysis showed significant association with the Jun N-terminal kinase (JNK), p38, extracellular signal-regulated kinase (ERK), and NF-κB signaling pathways. An association of these cascades with the HCEp network was also observed, ⁸ indicating that they are common signaling cascades after HSV-1 infection. 

After HSV infection, the HCEn cells produce large amounts of IL-6, similar to HCEp cells. ⁸ Network analysis indicated that IL-6 was the most significantly shared effector molecule. It was centrally located in the inflammatory program of the transcriptional network, and activations of NF-κB and JNK were shown to be related to IL-6 induction. ²⁷ In addition, IL-6 is a representative effector molecule downstream of TLR2, TLR3, and TLR9, which sense HSV entry. On infection, IL-6 mediates an acute phase reaction that influences antigen-specific immune responses. ^28,29 Importantly, IL-6 converts T cells into cytotoxic T cells or the Th17 lineage and stimulates B cell differentiation. ³⁰ In herpetic keratitis, IL-6 contributes to the massive neutrophil attraction to the corneal stroma ^{31 –33} and stimulates bystander populations and HCEp cells to induce vascular endothelial growth factor (VEGF). ^{8,32 –34}  

Collectively, our data provide strong evidence that HCEn cells can serve as APCs after HSV-1 infection. Understanding the immune-modulating properties of the corneal endothelium would help develop efficacious strategies to block HSV-1–induced inflammatory responses and endothelial cell loss. 

Footnotes

 Supported by Grand-in-Aid 20592076 and 21592258 for Scientific Research from the Japanese Ministry of Education, Science, and Culture.

Footnotes

 Disclosure: D. Miyazaki, None; T. Haruki, None; S. Takeda, None; S. Sasaki, None; K. Yakura, None; Y. Terasaka, None; N. Komatsu, None; S. Yamagami, None; H. Touge, None; C. Touge, None; Y. Inoue, None

The authors thank Duco Hamasaki for editing the manuscript.