Postgenomics Biomarkers for Rabies—The Next Decade of Proteomics

Abstract

Rabies is one of the oldest diseases known to mankind. The pathogenic mechanisms by which rabies virus infection leads to development of neurological disease and death are still poorly understood. Analysis of rabies-infected proteomes may help identify novel biomarkers for antemortem diagnosis of the disease and target molecules for therapeutic intervention. This article offers a literature synthesis and critique of the differentially expressed proteins that have been previously reported from various in vitro/in vivo model systems and naturally infected clinical specimens. The emerging data collectively indicate that, in addition to the obvious alterations in proteins involved in synapse and neurotransmission, a majority of cytoskeletal proteins are relevant as well, providing evidence of neuronal degeneration. An interesting observation is that certain molecules, such as KPNA4, could be potential diagnostic markers for rabies. Importantly, proteomic studies with body fluids such as cerebrospinal fluid provide newer insights into antemortem diagnosis. In order to develop a complete integrative biology picture, it is essential to analyze the entire CNS (region-wise) and in particular, the brain. We suggest the use of laboratory animal models over cell culture systems using a combinatorial proteomics approach, as the former is a closer match to the actual host response. While most studies have focused on the terminal stages of the disease in mice, a time-series analysis could provide deeper insights for therapy. Postgenomics technologies such as proteomics warrant more extensive applications in rabies and similar diseases impacting public health around the world.

Introduction

Rabies virus (RV) is a neurotropic, negative stranded RNA virus that belongs to the rhabdoviridae family with a wide host range; canines, bats, raccoons, humans, etc. (Richman et al., 1997; Topley et al., 1998). Post exposure vaccination has been 100% effective in preventing rabies, yet the current WHO estimate of the global burden of the disease is over 60,000 human deaths per annum; Asia alone accounts for 30,000 of them (Mani et al., 2013). RV initially infects the peripheral nervous system by binding to specific neuronal receptors and utilizes the axonal retrograde transport to invade the central nervous system (CNS) (Jackson et al., 2002). The clinical presentations of the disease usually include hydrophobia encephalitis, flaccid paralysis, and coma. Postmortem diagnosis using brain tissue is 100% confirmatory, but antemortem clinical diagnosis proves to be a major challenge till date. Unfortunately, for antemortem diagnosis, a negative test does not rule out RV infection and once the clinical signs begin to appear (i.e., after the virus reaches the brain), it invariably proves to be fatal (Gadre et al., 2010; Madhusudana et al., 2008).

Rabies is one of the oldest diseases known to mankind. Despite this, the pathogenic mechanisms by which rabies virus infection leads to development of neurological disease and death are not well understood (Dhingra et al.,2007; Wang et al., 2005). Electroencephalographic (EEG) abnormalities have also been recorded in animal models infected with different strains of RV. These alterations in neurophysiology could be attributed to the malfunction in neurotransmission. Functional genomics and proteomics have previously been used to determine host responses in cells/animals infected with different viruses (Dhingra et al., 2007). While there may be some understanding of the role of viral proteins in rabies pathogenesis, neither the contribution of host factors to rabies virus transcription/replication and axonal/trans-synaptic spread is understood nor are the mechanisms by which these factors determine the outcome of the disease precisely known (Dietzschold et al., 2008; Faber et al., 2004; Prosniak et al., 2001).

The current focus is therefore on identifying potential protein biomarkers in the host's CNS during the disease. A proteomics-based approach would not only provide us with the potential diagnostic and therapeutic targets but also give us a better insight into the pathological mechanisms and the associated host response. It would also help to identify differentially regulated proteins at the transcriptional, translational, as well as post-translational stages. It is a formidable task to identify and isolate a specific biochemical pathological course of the disease as the host would vary, depending on the virus strain (street/attenuated). In the current review, we have attempted to provide a comprehensive overview of the significant contributions in the field of rabies proteomics. The article particularly emphasizes the host proteins found to be differentially expressed when compared with uninfected controls.

Comparative Proteomics Analyses with Naturally Infected Specimens

Proteomic studies on rabies infection conducted to date are based on varied host systems such as mouse and human brain tissue samples, neuronal, and fibroblastic cell lines, which also contribute to the diversity in the type of responses observed. Table 1 provides a summary of the proteome analysis in RV infected different in vitro/in vivo host model systems and clinical samples.

Table 1.

Proteome Analyses of Rabies Virus with Different in vitro/in vivo Host Model Systems and Clinical Specimens

Serial No.	Host model	RV strains used	Technique employed	Number of differential proteins identified	Reference
In vitro systems
1.	Vero cells	Human Louis Pasteur 2061	2-D DIGE+MALDI-TOF-MS/MS	47	(Kluge et al., 2013)
2.	CCL131 cells	CVS, PV	2-DE+MALDI-TOF/TOF-MS	40	(Zandi et al., 2013)
3.	N2a cells	BD06, SRV9, CVS-11	2-DE+MALDI-TOF-MS/MS	35	(Wang et al., 2011)
4.	BHK21-2P cells	CVS	2-DE+ESI-MS/MS	14	(Zandi et al., 2009)
In vivo systems
5.	NMRI mice (lymphocyte)	Street virus, PV	2-DE+MALDI-TOF/TOF-MS	10	(Vaziri et al., 2013)
6.	ICR mice (brain)	B2C, SHBRV	2-D DIGE+MALDI-TOF/MS	33	(Dhingra et al., 2007)

	Source/ specimen	Region analyzed	Technique employed	No of differential proteins identified/reported	Reference
Clinical specimens
7.	Human brain	Right frontal cortex	iTRAQ+LC-MS/MS	94	(Venugopal et al., 2013)
8.	Human brain	Occipital lobe	2-DE+MALDI-TOF-TOF	14	(Farahtaj at al., 2013)
9.	Canine CNS	Hippocampus, brainstem, spinal cord	2-DE+q-TOF MS & MS/MS	71	(Thanomsridetchai et al., 2011)

B2C, attenuated strain derived from CVS-24; BD06, street virus isolated from dog; CVS, Challenge Virus Standard; 2-DE, two-dimensional gel electrophoresis; DIGE, Difference in Gel Electrophoresis; ESI, Electrospray Ionization; iTRAQ, isobaric tags for relative and absolute quantification; MALDI-TOF-MS/MS, Matrix Assisted Laser Desorption Ionization-Time of Flight-Mass Spectrometer; PV, Pasteur Virus; q-TOF, Quadrupole Time of Flight Analyzer; RV, Rabies Virus; SHBRV, Silver Haired Bat Rabies Virus; SRV9, Street Rabies Virus.

Proteome profiles of in vitro/in vivo systems vs. clinical specimens

Rabies virus propagation in animal models is preferred over tissue culture-based systems because unlike most viruses, rabies infection/replication in routine cell culture is not accompanied by any clear definable cytopathic effect, though the virus adapts easily on several cell lines such as neuroblastoma (N2a), Vero, and baby hamster kidney (BHK-21) cells (Meslin et al. 1996). Further, it is simpler to track the course of the disease in laboratory animals based on development of clinical signs. Proteomic analysis of the most common rabies infected in vitro models/in vivo host systems and street virus infected clinical samples revealed a list of differentially expressed proteins when compared to uninfected controls. Based on their functions, all the proteins (listed in Table 2 and Table 3) have been classified into eight broad categories as depicted in Figure 1.

FIG. 1.

Percentage distribution of differentially expressed proteins identified in rabies infected in vitro/in vivo model systems (A) and in naturally infected clinical specimens (B) and classified under eight categories: Synapse, neurotransmission, and neurophysiology; signal transduction; ion homeostasis and apoptosis; stress response proteins; metabolic processes; cytoskeletal elements; heat shock proteins/chaperones; and others.

Table 2.

Differentially Expressed Proteins with Respect to Uninfected Controls Reported in Various in Vitro/in-Vivo Models of Rabies Infection

			Fold change in expression
					SHBRV				B2C
Categor	Protein	Symbol	CVS/CVS-11	BD06	i.c	i.m	DRV	SRV9	i.c	i.m	PV	References
Synapse, neurophysiology and neurotransmission	E3 ubiquitin-protein ligase TRIM9	TRIM 9			3.0 ↓	NK						(Dhingra et al., 2007)
	Alpha-soluble NSF attachment protein	α-SNAP			3.0 ↓	2.0 ↓
	Pallidin (syntaxin 13)	PA, PLDN			2.2 ↓	NK
	Synexin, Annexin A7	SNX, ANXA7			2.0 ↓	2.4 ↓
	Syntaxin 18	STX18			1.8 ↓	NK
	Neurofilament protein NF 66	INA							3.0 ↓	NK
	Alpha-internexin	INA/NEF5							NK	3.6 ↓
	Annexin A1	ANXA1					1.7 ↓				1.9 ↓	(Vaziri et al., 2012)
	Collapsin response mediator protein 2	CRMP-2							3.6 ↑	NK		(Dhingra et al., 2007)
	Dihydropyrimidinase-related protein 1	CRMP-1							NK	2.8 ↑
Signal transduction	Galectin-1	LGALS1	1.75 ↓	2.9 ↓				1.49 ↓				(Wang et al., 2011)
	14-3-3 Protein zeta/delta	YWHAZ	3.23 ↑	1.02 ↑				2.27 ↑
	14-3-3 Protein gamma	YWHAG	3.57 ↓	1.12 ↑				ND
	14-3-3 Protein epsilon	YWHAE	2.5 ↓	1.05 ↑				1.23 ↓
	Nucleoside diphosphate kinase B	NDPK-B	1.69 ↓	4.55 ↑				2.7 ↓
	Alpha-enolase	ENO1	2.78 ↑	1.96 ↑				4.17 ↑
Ion homeostasis and apoptosis	Na⁺/K⁺ ATPase	ATP1A1			3.0 ↑	2.6 ↑						(Dhingra et al., 2007)
	H⁺ ATPase subunit a isoform 1	ATP6V0A1			2.6 ↑	NK
	H⁺ ATPase synthase subunit b	MT-ATP6							2.8 ↓	2.2 ↓
	Villin 1	VIL1							2 ↓	NK
	Calretinin	CALB2							2.2 ↓	NK
	Caspase 1/ Interleukin 1	IL-1			3.2 ↓	NK						(Wang et al., 2011)
	Annexin				NK	2.2 ↓						(Dhingra et al., 2007)
	Apoptotic protease-activating factor	APAF 1			3.0 ↓	NK			2.4 ↑	2.8 ↑
	Ras-GTPase- activating protein	GAPs/ RGS			1.8 ↓	2.2 ↓
Stress response proteins	Peroxiredoxin 6	PRDX6	1.92 ↑	2.9 ↑				2.06 ↑				(Wang et al., 2011)
	Peroxiredoxin 4	PRDX4	5.22 ↑	1.99 ↑				4.01 ↑
	Periredoxin 1	PRDX1					1.9 ↓				2.5 ↓	(Vaziri et al., 2012)
	Thioredoxin		1.1 ↓	1.53 ↑				2.3 ↓
	Thioredoxin-like protein 1	TXNL1	2.9 ↓	1.75 ↓				1.63 ↓
	Thioredoxin domain containing 12 precursor	TXNDC12	3.3 ↓									(Zandi et al., 2013)
	Heat-shock protein 1	Hsp-1	2.3 ↓	1.54 ↑				1.9 ↓				(Wang et al., 2011)
	Stress-induced phosphoprotein 1	STIP1	3.2 ↓	1.37 ↑				2.28 ↑
	Cytochrome b-c1 complex subunit 1	UQCRC1	1.06 ↑	2.08 ↑				1.85 ↑
	Endoplasmic reticulum protein ERp29	ERP29	1.2 ↓	1.8 ↓				1.2 ↓
	Endoplasmic reticulum protein ERp29		1.2 ↓	2.3 ↓				2.02 ↑
	Superoxide dismutase [Mn],mitochondrial precursor	SOD2	7 ↑									(Zandi et al., 2009)
	Ubiquinol-cytochrome-c reductase complex core protein 1 precursor	UQCRC1	1.7 ↑									(Zandi et al., 2013)
Metabolic processes	Heterogeneous nuclear ribonucleoprotein K	hnRNPK	3.15 ↑	2.52 ↑				2.08 ↑				(Wang et al., 2011)
	Heterogenous nuclear ribonucleoprotein L	hnRNPL									4.5 ↓	(Vaziri et al., 2012)
	Splicing factor, arginine/serine-rich 1	SRSF1	1.69 ↓	1.61 ↓				2.32 ↑				(Wang et al., 2011)
	Proteasome subunit alpha type-3	PSMA3	1.59 ↑	0				1.98 ↑
	T-Complex protein 1 subunit epsilon	CCT5	1.72 ↓	4.63 ↑				5.26 ↑
	Inorganic pyrophosphatase	PPA1	1.63 ↓	1.56 ↓				1.96 ↓
	UDP-glucuronosyltransferase 1A1	UGT1A1	1.79 ↑	0				3.41 ↑
	Zinc finger protein 12	Zfp12	1.34 ↑	1.03 ↓				3.23 ↑
	Similar to MusmusculusmCASP		ND	3.36 ↑				2.97 ↑
	Cytokine-induced apoptosis inhibitor 1	CIAPIN1	3.58 ↑	1.96 ↑				4.16 ↑
	SRF accessory protein 1	SAP 1			2.8 ↓	NK						(Dhingra et al., 2007)
	Cell division control protein 42	CDC 42			2.2 ↓	NK
	Proteasome component 9				1.8 ↓	NK
	Trehalase	TREH							2.4 ↓	NK
	Endoplasmic reticulum resident protein	ER60 precursor							3.0 ↑	NK
	Malate dehydrogenase	MDH							2.1 ↑	NK	2 ↓	(Dhingra et al., 2007; Vaziri et al., 2012)
	Cathepsin								2.0 ↑	NK		(Dhingra et al., 2007)
	Elongation factor 2	EF-2							2.5 ↑	3.0 ↑
	Isocitrate dehydrogenase	IDH							NK	3.2 ↑
	Isocitrate dehydrogenase 3 alpha, isoform CRA	IDH3A									2.4 ↓	(Vaziri et al., 2012)
	60S acidic ribosomal Protein P₀	RPLP60	3 ↓									(Zandi et al., 2009)
	SH3 domain binding protein 2	SH3BP2							2.6 ↑	2.8 ↑	1.4 ↓	(Dhingra et al., 2007; Zandi et al., 2009)
	G protein–coupled receptor 44, Prostaglandin D2 receptor 2	PTGDR2							3 ↑	NK		(Dhingra et al., 2007)
Cytoskeletal elements	Vimentin	VIM	1.5 ↓	5.5 ↓				ND				(Zandi et al., 2013)
	Capping protein (Actin filament)muscle Z-line, beta –Homo sapiens (Human)	CP	P									(Zandi et al., 2013)
	Tubulin beta 5	TUB β	3.12 ↑	2.27 ↑				3.14 ↑				(Wang et al., 2011)
	Tubulin alpha 1	Tub α	5.23 ↑	1.27 ↑				4.37 ↑
	Beta-tropomyosin	TPM 2(beta)	2.58 ↑	1.84 ↑				2.61 ↑
	Stathmin	STMN 1	P									(Zandi et al., 2013)
	Cofilin1, non-muscle	CFL 1									1.56 ↓	(Vaziri et al., 2012)
Heat shock proteins/ Chaperons	60 kDa heat shock protein, mitochondrial	HSP60							3.2 ↑	3.6 ↑
	Heat shock 70 kDa protein 1A/1B	HSPA1A/1B							NK	2.1 ↑

B2C, attenuated strain derived from CVS-24; BD06, street virus isolated from dog; CVS, Challenge Virus Standard; DRV, Dog Rabies Virus; i.c., intracerebral route of inoculation; i.m., intramuscular route of inoculation; ND, not done; NK, not known; PV, Pasteur Virus; SHBRV, Silver Haired Bat Rabies Virus; SRV9, Street Rabies Virus; ↑, upregulation; ↓, downregulation; P, present.

Table 3.

Differentially Expressed Proteins with Respect to Uninfected Controls Reported in Various Clinical Specimens

			Human PM brain				Dog PM brain and spinal cord
				Frontal cortex			Hippocampus		Brainstem		Spinal cord
Category	Protein	Symbol	Occipital lobe	P	F	PV	P	F	P	F	P	F	References
Synapse, Neurophysiology and Neurotransmission	Contactin 1	CNTN1		1.0 ↑	2.0 ↑	1.7 ↑							(Venugopal et al., 2013)
	Syntaxin 1A	Stx1a			2.4 ↑	3.9 ↑
	Calcium/calmodulin-dependent protein kinase IIA	CAMK2A		2.1 ↑	1.6 ↑	4.3 ↑
	Adaptor-related protein complex 3, beta subunit	AP3B2			2.8 ↓	2.2 ↓
	Neurofascin precursor	NFASC		3.3 ↓	1.3 ↓	1.1 ↓
	ATPase family AAA domain-containing protein 1	ATAD1		1.6 ↓	2.5 ↓
	Dihydropyrimidinase related protein-2	DRP-2/CRMP-2									NS	↓	(Thanomsridetchai et al., 2011)
	Dihydropyrimidinase related protein-2 isoform 6	DRP-2/CRMP-2					NS	↑	↑	NS	↓	↓
	Dihydropyrimidinase related protein-2 isoform 4	DRP-2/CRMP-2									↓	↓
	Hippocalcin-like protein 4	HPCAL4		2.4 ↑	1.5 ↑	1.7 ↑							(Venugopal et al., 2013)
Signal transduction	Guanine nucleotide binding protein subunit beta 5 (Isoform 1)	GNB5	2.41 ↓										(Farahtaj et al., 2013)
	Alpha-enolase isoform		1.81 ↑
Ion homeostasis and apoptosis	Nitrilase homolog 1 Isoform 2	NIT1	1.68 ↓
	Annexin A2 isoform	ANXA2									NS	↓	(Thanomsridetchai et al., 2011)
	Annexin A2 isoform										↑	NS
	Annexin A6	ANXA6					NS	↑	NS	↓
	Cytochrome P450 2B12	CYPIIB12					↓	NS
	Mitochondrial carrier homolog 2	MTCH2			2.7 ↑								(Venugopal et al., 2013)
	Optic atrophy 1 isoform 2	OPA1		1.5 ↑	5.1 ↑	1.2 ↑
	Astrocyticphoshpoprotein	PEA15		2.1 ↑	1.7 ↑	1.9 ↑
	Catenin beta 1	CTNNB1		1.2 ↑	1.8 ↑	2.03 ↑
	Programmed cell death 6 interacting protein	PDCD6IP		1.6 ↓	2.5 ↓	5.0 ↓
	Protein kinase C, alpha	PRKCA		2.17 ↓	1.07 ↓	2.5 ↓
Stress response proteins	150 kDa oxygen-regulated protein precursor	Orp150									↑	NS	(Thanomsridetchai et al., 2011)
	Glutathione S-transferase Mu 3	GSTM3-3							↑	NS
	Oxygen-regulated protein 1	ORP1									↓	NS
	Peroxiredoxin 1	PRDX1									↓	↓
	Peroxiredoxin 5 precursor, isoform b	PRDX5		1.2 ↓	1.2 ↓	2.5 ↓							(Venugopal et al., 2013)
Metabolic processes	Bifunctional purine biosynthesis protein PURH	ATIC	1.59 ↓										(Farahtaj et al., 2013)
	Glutamine synthetase/ Glutamate-ammonia ligase	GLUL	1.36 ↓	2.66 ↑	4.85 ↑	5.48 ↑							(Farahtaj et al., 2013; Venugopal et al., 2013)
	NADH dehydrogenase [ubiquinone] iron-sulfur protein 8, mitochondrial	NDUFS8	1.74 ↓										(Farahtaj et al., 2013)
	Isoform Short of Proteasome subunit alpha type-1	PSMA1	1.54 ↓
	Fatty acid-binding protein,	FABP3	1.31 ↓
	NADH dehydrogenase, (ubiquinone) Fe-S protein 3	NDUFS3		1.6 ↑	2.1 ↑	2.6 ↑							(Venugopal et al., 2013)
	NADH dehydrogenase (ubiquinone) 1	NDUFAB1		2.0 ↑	2.4 ↑	3.7 ↑
	Tumor protein D52	TPD52		2.8 ↑	1.4 ↑	2.4 ↑
	Aldehyde dehydrogenase 2 family	ALDH2		2.2 ↑	1.9 ↑	2.3 ↑
	Pyruvate dehydrogenase (Lipoamide), beta chain	PDHB		2.0 ↓	2.5 ↓	1.7 ↓
Cytoskeletal elements	Tubulin alpha-1C chain	TUBA1C	1.81 ↓										(Farahtaj et al., 2013)
	Beta-centractin	ACTR1B	1.57 ↓
	Destrin	DSTN	1.58 ↓
	Dynein, light chain, roadblock-type 1	DYNLRB1	1.74 ↓
	Dynamin	DNM					↓	NS					(Thanomsridetchai et al., 2011)
	Fascin 1	FSCN1							↓	NS
	Glial fibrillary acidic protein, astrocyte isoform	GFAP									NS	↑
	Myosin, heavy chain 2, skeletal muscle, adult	MYH2							↓	NS
	Nebulin-related anchoring protein isoform 2	NRAP							↓	↓
	Neurofilament medium polypeptide	NEFM									↓	↑
	Neurofilament, heavy polypeptide	NEFH							↓	↓
	Septin-8	SEPT8									↓	↓
	TUBB2B protein						NS	↓
	Tubulin, alpha-1 isoform 9	TUBA1							↑	NS
	Tubulin, alpha-2 chain isoform 7	TUBA2							↓	NS
	Vinculin (Metavinculin)	VCL									↑	↑
	Xin actin-binding repeat containing 2 isoform 1	XIRP1									↑	NS
	Transgelin 3	TAGLN3		2.3 ↑	1.4 ↓								(Venugopal et al., 2013)
	Proteolipid protein 1 isoform 2	PLP1		1.1 ↓	1.2 ↓	2.7 ↑
	Myelin basic protein	MBP		1.4 ↓	2 ↓	2.1 ↑
Heat shock proteins/chaperons	Heat shock protein Beta-5	Hspβ5									↓	↑	(Thanomsridetchai et al., 2011)
	Heat shock protein Beta-5 (Rosenthal fiber component)	Hspβ5							NS	↑	↑	↑
	DnaJ (Heat shock protein 40)	Hsp 40									↓	↓
	Heat shock cognate 71 kDa protein	HSPA8									↑	NS
	Heat shock protein 90 kDa beta, member 1	Hsp 90							↑	NS	NS	↓
	Heat shock protein beta-1	HSPB1									NS	↑
	Heat shock protein beta-1	HSPB1									↑	NS
	Heat shock protein 70kDa protein 5	HSPA5		2.2 ↑	1.5 ↑	1.2 ↑							(Venugopal et al., 2013)
	Heat shock protein 70kDa protein 9	HSPA9		1.4 ↑	2.1 ↑	1.4 ↑
Others	Macrophage migration inhibitory factor	MIF	1.90 ↓										(Farahtaj et al., 2013)
	Immunoglobulin heavy chain variable region								↑	NS			(Thanomsridetchai et al., 2011)
	Interferon alpha 4	IFNA4					NS	↑
	SARM1 protein						NS	↑
	Karyopherin alpha 4	KPNA4		3.1 ↑		7.3 ↑							(Venugopal et al., 2013)
	Importin 5	IPO5		1.9 ↑	2.9 ↑	1.9 ↑
	Poteasomemacropain subunit alpha 2	PSMA2		2.6 ↑	3.1 ↑	2.5 ↑
	Guanine nucleotide binding protein alpha activating activity polypeptide	GNAO1		1.6 ↑	2.2 ↑	2.3 ↑

F, Furious Rabies, NS, not significant; P, Paralytic Rabies, PM, postmortem; PV, partially vaccinated; ↑, upregulation; ↓, downregulation.

It was observed that the expression of a vast majority of the metabolic proteins (33%) were altered in the cell culture and animal model systems (Dhingra et al., 2007; Kluge et al., 2013; Vaziri et al., 2012; Wang et al., 2011; Zandi et al., 2009; 2013) while analysis of rabid human and dog brain samples revealed changes in the expression of a large number of cytoskeletal associated proteins (27%) (Faber et al., 2004; Thanomsridetchai et al., 2011; Venugopal et al., 2013). This difference may be attributed to the diversity in virus cultivation, cell line, media component, and culture conditions. It has been demonstrated earlier that adaptation and propagation of viruses in different animal cells with different cultivation conditions involve changes in virus replication dynamics and protein expression. Changes in media components and differences in process conditions also have shown to alter expression of proteins, particularly those involved in metabolism and stress response (Kluge et al., 2013). This indicates that the hypotheses put forward based on laboratory model systems may sometimes deviate largely from the actual findings.

We also observed that, of all the differentially expressed proteins reported, only four proteins—CRMP2 (collapsin response mediator protein), anti-oxidative stress protein- PRDX1 (peroxiredoxins), signal transducing protein- α-enolase, and chaperone Hsp70—are common in the two groups [i.e., laboratory host model systems (in vitro/in vivo) and clinical specimens]. Further, the similarity in expression pattern is also noteworthy (Table 4). It is possible that this limited overlap in the protein profile may be due to the variability of different model systems or virus strains and limitations in the individual techniques employed.

Table 4.

Common Proteins Reported Between Laboratory Host Model System and Clinical Specimens

Protein	Laboratory host model system	Expression pattern	Clinical specimens	Expression pattern	References
CRMP2	Mouse brain	↑	Dog brain	↑	Dhingra et al., 2007, Thanomsridetchai et al., 2011
PRDX1	Mouse lymphocytes	↓	Dog spinal cord	↓	Vaziri et al., 2013, Thanomsridetchai et al., 2011
α-Enolase	N2a cells	↑	Human brain (occipital lobe)	↑	Wang et al., 2011, Farahtaj at al., 2013
Hsp70	Mouse brain	↑	Human brain (right frontal cortex)	↑	Dhingra et al., 2007, Venugopal et al., 2013

↑, Upregulation;↓, Downregulation.

Techniques employed to study differential expression

Several gel-based proteomic studies have been carried out either using 2-DE (two- dimensional gel electrophoresis) or DIGE (difference in gel electrophoresis) to study RV host cell interactions. Different mass spectrometric (MS) based approaches following all gel/non-gel based platforms have made it possible to analyze such virus–host interactions successfully. A recent study has employed in comparison an enhanced technique of iTRAQ (isobaric tags for relative and absolute quantification) labeling to quantitatively analyze the proteome of human frontal cortex tissue in paralytic, furious, and partially vaccinated cases. It is obvious from Table 1 that, with high throughput techniques such as iTRAQ coupled with LC-MS/MS analysis, greater number of proteins can be identified. Venugopal and colleagues (2013) have identified 402 proteins, of which 94 were found to be differentially expressed with respect to controls. Nevertheless, 2-DE followed by q-TOF MS & MS/MS has helped identify 71 altered proteins in the hippocampus, brainstem, and spinal cord of street virus-infected CNS of dogs expressing paralytic and furious forms (Thanomsridetchai et al., 2011). Thus, tandem mass spectrometry increases the likelihood of finding proteins that may have not been considered previously. Pre-fractionation to 2-DE using reverse phase high performance liquid chromatography (RP-HPLC) can improve the chances of resolving and thus identifying low abundance proteins (Vaziri et al., 2006). It therefore appears that with each type of technique altogether different sets of proteins are being identified and hence an extremely narrow overlap is observed on comparing the proteins reported from laboratory model systems to those from clinical specimen. A combination of multiple proteomic methods may be utilized to develop an effective output. Complementary information may be obtained on using combinatorial strategies such as 2-DE/DIGE followed by iTRAQ and LC-MS/MS analysis to generate a better picture of the host protein profiles.

Proteomics insights: Neuronal dysfunction versus neuronal degeneration

It has been well established that the virus propagates at the synapse and modulates the synaptic physiology to disseminate across the trans-neuronal network. Alterations in the rabies-infected CNS are largely attributed to neuronal dysfunction. Proteomic experiments by Dhingra et al. (2007) have demonstrated that proteins involved in the formation of SNARE complexes are downregulated. SNARE-associated proteins such as Trim 9, SNAP α, annexin A7, synataxin 13, and syntaxin 18, which are also involved in docking of synaptic vesicle and exocytosis, are also downregulated. This may explain the decreased docking ability and accumulation of the synaptic vesicles at the presynaptic membranes in rabies-infected neurons. Impaired release and binding of serotonin as well as gamma-aminobutyric acid (GABA) in CVS infected primary neuronal cultures has also been demonstrated earlier. In fact, neurotransmitters such as norepinephrine (NE), dopamine (DA), dihydroxyphenylacetic acid (DOPAC), and serotonin were found to be stimulated in the hippocampus of rats infected with CVS in the early stages of infection, but the expression declined on development of clinical signs (Fu et al., 2005). This is further supported by the NMR findings of a rabid human brain that exhibited reduced levels of N-acetyl-aspartate (NAA) and GABA w.r.t. controls (Reinke et al., 2013) indicating that with progression of the disease, neurons are rendered incapable of releasing neurotransmitters at the synaptic junctions (Fu et al., 2005).

In addition to altered neurotransmission, in clinical specimens, a majority of the cytoskeletal proteins such as microtubule-associated proteins such as TUBA1C, β centractin, dynein light chain, dynamin, and tubulin α-2 were observed to be downregulated except GFAP, tubulin α-1 isoform 9, vinculin, and XIRP1. Defects in dynein/dynactin function due to altered expression are thought to affect the retrograde neuronal transport (Farahtaj et al., 2013), while suppression of actin associated proteins such as destrin, nebulin related anchoring protein, and myosin are further suggestive of abnormalities in the actin dynamics regulatory mechanism. The neurofilament heavy polypeptide was reported to be underexpressed in the brainstem of furious and paralytic rabies cases, whereas the medium polypeptide was downregulated in the spinal cord of paralytic rabies but upregulated in furious cases (Venugopal et al., 2013).

However, recent in vitro studies show an overall upregulation in the expression of a large number of cytoskeletal elements such as vimentin, tubulin subunits, and β-tropomyosin during RV infection. This overexpression may result from either an increased mRNA translation or an augmented cleavage of the cytoskeletal elements or both. Transcriptomic studies are required to validate the former, though several reports indicate the possibility of the latter also. RV-induced collapse in vimentin network and therefore cytoskeletal disruption (Zandi et al., 2013) and downregulation of total vimentin in CVS-infected neuroblastoma N2a cells (Wang et al., 2011) has also been demonstrated. While vimentin is thought to modify the intermediate filament network, CP–-a new form of heterodimeric capping protein (CP) involved in myofibrillogenesis actin organization has also been reported (Zandi et al., 2009). These findings are interesting as glycolysis in primary cells is also found to be cytoskeleton dependent (Kluge et al., 2013).

Thus, rabies, though often associated with absence of morphological changes supported by few or no neuronal degeneration (Jackson 2006), data from proteomic analysis indicate otherwise. Structural changes observed in the cerebral cortex of CVS infected mice such as swollen mitochondria, presence of vacuoles in the axons and pre-synaptic nerve endings, and beading of dendrites and axons, only corroborate the proteomic findings further (Scott et al., 2008). Even electron microscopic investigations reveal complete disappearance or reduction in a number of intracellular organelles such as rough endoplasmic reticulum, free ribosomes, and mitochondria in the pathogenic RV infected neurons (Li et al., 2005). Furthermore, demyelination and axonal injury in peripheral nerves, as well as axonal degeneration, have also been observed in paralytic rabies (Sheikh et al., 2005). With years of debate over neuronal dysfunction or degeneration responsible for pathogenesis in RV infection, it is clear that while neuronal dysfunction may be a major reason contributing to it, there is ample evidence supporting existence of the latter as well.

Potential Biomarkers

Rabies expresses itself in two distinct forms, paralytic or the dumb/numb form in which peripheral nerve demyelination or axonal degeneration have been observed, and the classical encephalitic/furious form characterized typically by hydrophobia wherein anterior horn dysfunction has been reported (Gadre et al., 2010). The paralytic form is often difficult to distinguish from and confused with the Guillain-Barre (GB) syndrome (Madhusudana et al., 2008). Discovery of protein biomarker appears to have the potential to answer the challenging question of clinical diagnosis of the disease antemortem. Mass spectrometry in its various guises has played an important role in discovery of biomarkers for many diseases. Venugopal et al. (2013) have reported several novel proteins found to be differentially expressed on analyses of rabid human proteomes, as listed in Table 5.

Table 5.

Probable Biomarker Molecules

Serial No	Protein	Symbol	Description
1.	Proteoplipid Protein 1	PLP1	Involved in myelination
2.	Glutamate ammonia ligase	GLUL	Synthesized by astrocytes
3.	Calcium calmodulin dependent kinase 2 alpha	CAMK2A	Downstream signaling for building of neuromuscular junctions
4.	Optic atrophy 1	OPA1	Dynamin-related GTPases
5.	Hippocalcin-like protein 4	HPCAL4	Calcium-dependent regulation of rhodopsin phosphorylation
6.	Transgelin 3	TAGLN3	CNS development
7.	Adaptor-related protein complex 3, beta 2 subunit	AP3B2	Component of neuron specific AP3 complex
8.	Programmed cell death 6 interacting protein	PDCD6IP	Induces neuronal apoptosis
9.	Limbic system associated membrane protein	LASMP	Expressed by neurons comprising limbic-associated cortical & subcortical regions that function in cognition, emotion, memory, and learning
10.	Karyopherin alpha 4	KPNA4	Cytoplasmic proteins that recognize nuclear localization signals

PLP1 was found to be downregulated (<two-fold) in both furious and paralytic rabies. It has also been implicated in axonal degeneration in patients suffering from chronic multiple sclerosis (MS) (Garbern et al., 2002). Another molecule, glutamate ammonia ligase (GLUL) is one of the important enzymes involved in ROS associated programmed necrosis (Moquin et al., 2010; Yu et al., 2011). Interestingly, while its overexpression (four-fold) has been reported in the later stages of Alzheimer disease (Venugopal et al., 2013), it is reported to be downregulated in the occipital lobe (Farahtaj et al., 2013), emphasizing the need to study the region-wise expression of such molecules in the entire CNS. GLUL is also reported to undergo oxidation in the later stages of the neurodegenerative disease Alzheimer's (Butterfield et al., 2006). CAMK2A, ubiquitously known to be involved in downstream signaling for building of neuromuscular junctions (Wu et al., 2010) and in synaptic plasticity in the hippocampus was found to be upregulated in paralytic (two-fold) and PV (four-fold) groups. CAMK2A aggregates are found to co-localize within the intracytoplasmic Negri bodies within the neurons. Immunohistochemical studies indicate higher expression in vaccinated cases of paralytic rabies in comparison to both nonvaccinated paralytic and furious cases (Venugopal et al., 2013). CAMK2A is also elevated in individuals with bipolar disorders (Beraki et al., 2005).

Other novel upregulated proteins such as OPA 1 were reported (though <two-fold) in paralytic, furious, and partially vaccinated cases, and HPCAL4 only in paralytic rabies (2.4-fold). Interestingly, PLP1, GLUL, CAMK2A, and OPA1 have all been implicated in schizophrenia too (Martins-de-Souza et al., 2010; Qin et al., 2005; Venugopal et al., 2013). AP3B2 gene expression is also downregulated by the human T-lymphotropic virus type-1 (HTLV) (Nair, 2004). Interestingly, while PDCD6IP is downregulated (<three-fold) in paralytic or furious cases, it is found to decrease five-fold in partially vaccinated cases. While its depletion is known to cause accumulation of unusual actin molecules (Akgul et al., 2009), its increased expression has been associated with survival of injured neurons (Rojo et al., 2011). The roles of TAGLN3, PDCD6IP, and LASMP in neuroviral/neurodegenerative diseases are still not explored and further investigation is needed to determine if either/all hold any potential as therapeutic or diagnostic marker.

Another such molecule that may be investigated for its role as a diagnostic marker to identify paralytic rabies is KPNA4. It was found to increase seven-fold in partially vaccinated cases and three-fold in paralytic cases. Karyopherin alpha 4 has been previously reported to be downregulated in tuberculosis meningitis and hepatocellular carcinoma (Kumar et al. 2012; Yu et al. 2007).

Neurons being the primary host for the rabies virus makes it imperative to study differential protein expression in the brain tissue, however from the diagnosis point of view, rabies-infected CSF proteome may provide a better insight. Metabolomic studies reveal that rabies-infected brain and CSF exhibit vast differences in their metabolite profiles. Upregulation was observed for CSF metabolites such as lactate, quinolinate, glutamate, malate, ornithine, and many other lipid metabolites with progress in disease, while others such as acetate, myo-inositol, pyroglutamate, and succinate were nearly absent. Increased lactate levels were also reported in the frontal cortex and white matter, suggesting anaerobic metabolism in rabies. In addition, the CSF metabolite profiles were found to differ in rabies survivors on comparison with those who died of rabies (O'Sullivan et al., 2013; Reinke et al., 2013). The field of proteomics and transcriptomics has not yet been explored for rabies-infected CSF but hold greater prospect in determining potential antemortem diagnostic markers for rabies.

Conclusions and the Way Forward

Based on the type of protein profiles generated on analysis of infection with different virulent strains of RV, including the wt-street strains, we draw the following conclusions:

i. A vast majority of the proteins that exhibited altered expression in the in vitro/in vivo laboratory model systems belonged to the metabolic processes category (33%). This may be attributed to the differences in virus culture conditions including nature of the cell line, media component, etc.

ii. Changes in the expression of a large number of cytoskeletal associated proteins (27%) have been observed in the clinical specimen of rabid human and dog brain samples, indicating that not only neuronal dysfunction but also neuronal degeneration contributes to the disruption of neurotransmission dynamics in RV infection.

iii. The vast difference in the host protein profiles with only four common proteins—CRMP2, PRDX1, α-enolase, and Hsp70—between the two groups, in vitro/in vivo systems and clinical specimens may be due to the diversity of the host and the techniques employed to study them and their limitations. This may be resolved by adopting combinatorial strategies that can generate complementary information to have a better correlation while studying rabies proteomes in model host systems and actual clinical specimens.

iv. The currently available information on rabies proteome from human samples is restricted to only the occipital lobe and frontal cortex. In order to generate comprehensive data, it is imperative to perform proteomic profiling of all the regions of brain tissue. A targeted region-wise analysis of certain proteins such as GLUL that exhibit different expression patterns in different areas of the brain, an amalgamation of 2-DE, iTRAQ, and LC-MS/MS may provide deeper insight into its patho-physiological role in rabies.

v. Proteins such as KPNA4 may prove to be useful diagnostic biomarker that may help distinguish paralytic rabies from GB, but further studies are warranted.

vi. Proteomic profiling of rabies-infected CSF specimen may provide a better insight into protein biomarkers for antemortem diagnosis.

We therefore conclude that currently only partial data from proteomic analysis of rabies-infected CNS or other varied host cell systems are available. In order to generate a holistic picture, it is essential to analyze the entire CNS (region-wise) and in particular, the brain. We further suggest the use of laboratory animal models over cell culture systems, using a combinatorial proteomics approach, as they are a closer match to the host response observed in reality. While most studies have focused on the terminal stages of the disease in mice, a time-series analysis may provide a better insight for therapy. In addition, the proteomics data may be complemented with more information from metabolomic and transcriptomic studies of the same samples. Using CSF as a clinical specimen, especially in antemortem laboratory investigation, is crucial, as potentially identifiable lead biomolecules may be used either as diagnostic markers or more importantly, for therapeutic intervention.

Footnotes

Author Disclosure Statement

The authors declare that no competing financial interests exist.

References

Akgul

, Zigrino

, Frith

, et al. (2009). Proteomic analysis reveals the actin cytoskeleton as cellular target for the human papillomavirus type 8. Virology, 386, 1–5.

Beraki

, Aronsson

, Karlsson

, et al. (2005). Influenza A virus infection causes alterations in expression of synaptic regulatory genes combined with changes in cognitive and emotional behaviors in mice. Mol Psychiatry, 10, 299–308.

Butterfield

, Abdul

, Newman

, et al. (2006). Redox proteomics in some age-related neurodegenerative disorders or models thereof. Neuro Rx, 3, 344–357.

Dhingra

, Li

, Liu

, et al. (2007). Proteomic profiling reveals that rabies virus infection results in differential expression of host proteins involved in ion homeostasis and synaptic physiology in the central nervous system. J Neurovirol, 13, 107–117.

Dietzschold

, Li

, Faber

, et al. (2008). Concepts in the pathogenesis of rabies. Future Virol, 3, 481–490.

Faber

, Pulmanausahakul

, Nagao

, et al. (2004). Identification of viral genomic elements responsible for rabies virus neuroinvasiveness. Proc Natl Acad Sci USA, 101, 16328–16332.

Farahtaj

, Zandi

, Khalaj

, et al. (2013). Proteomics analysis of human brain tissue infected by street rabies virus. Mol Biol Rep, 40, 6443–6450.

, and Jackson

(2005). Neuronal dysfunction and death in rabies virus infection. J Neurovirol, 11, 101–106.

Gadre

, Satishchandra

, Mahadevan

, et al. (2010). Rabies viral encephalitis: clinical determinants in diagnosis with special reference to paralytic form. J Neurol Neurosurg Psychiatry, 81, 812–820.

10.

Garbern

, Yool

, Moore

, et al. (2002). Patients lacking the major CNS myelin protein, proteolipid protein 1, develop length-dependent axonal degeneration in the absence of demyelination and inflammation. Brain, 125, 551–561.

11.

Jackson

. (2006). Rabies: New insights into pathogenesis and treatment. Curr Opin Neurol, 19, 267–270.

12.

Jackson

, and Wunner

. (2002). Rabies, Academic Press, Amsterdam, Boston.

13.

Kluge

, Rourou

, Vester

, et al. (2013). Proteome analysis of virus-host cell interaction: Rabies virus replication in Vero cells in two different media. Appl Microbiol Biotechnol, 97, 5493–5506.

14.

Kumar

, Venugopal

, Mahadevan

, et al. (2012). Quantitative proteomics for identifying biomarkers for tuberculous meningitis. Clin Proteomics, 9, 12.

15.

, Sarmento

, and Fu

(2005). Degeneration of neuronal processes after infection with pathogenic, but not attenuated, rabies viruses. J Virol, 79, 10063–10068.

16.

Madhusudana

, and Sukumaran

. (2008). Antemortem diagnosis and prevention of human rabies. Ann Indian Acad Neurol, 11, 3–12.

17.

Mani

, and Madhusudana

. (2013). Laboratory diagnosis of human rabies: Recent advances. Sci World J, 2013, 569712.

18.

Martins-de-Souza

, Schmitt

, Roder

, et al. (2010). Sex-specific proteome differences in the anterior cingulate cortex of schizophrenia. J Psychiatr Res, 44, 989–991.

19.

Meslin

, Kaplan

, and Koprowski

. (1996). Laboratory Techniques in Rabies, World Health Organization, Geneva.

20.

Moquin

, and Chan

. (2010). The molecular regulation of programmed necrotic cell injury. Trends Biochem Sci3, 5, 434–441.

21.

Nair

. (2004). Studies of retroviral vectors for in utero gene transfer and investigation of calcium-mediated gene regulation by human T-lymphotropic virus type-1. Ohio State University, PhD thesis.

22.

O'Sullivan

, Willoughby

, Mishchuk

, et al. (2013). Metabolomics of cerebrospinal fluid from humans treated for rabies. J Proteome Res, 12, 481–490.

23.

Prosniak

, Hooper

, Dietzschold

, et al. (2001). Effect of rabies virus infection on gene expression in mouse brain. Proc Natl Acad Sci USA, 98, 2758–2763.

24.

Qin

, Gao

, Xing

, et al. (2005). A family-based association study of PLP1 and schizophrenia. Neurosci Lett, 375, 207–210.

25.

Reinke

, Resch

, Maingat

, et al. (2013). Metagenomic and metabolomic characterization of rabies encephalitis: New insights into the treatment of an ancient disease. J Infect Dis, 207, 1451–1456.

26.

Richman

, Whitley

, and Hayden

. (1997). Clinical Virology. Churchill Livingstone, New York.

27.

Rojo

, Prough

, Falduto

, et al. (2011). Influence of stochastic gene expression on the cell survival rheostat after traumatic brain injury. PLoS One, 6, e23111.

28.

Scott

, Rossiter

, Andrew

, et al. (2008). Structural abnormalities in neurons are sufficient to explain the clinical disease and fatal outcome of experimental rabies in yellow fluorescent protein-expressing transgenic mice. J Virol, 82, 513–521.

29.

Sheikh

, Ramos-Alvarez

, Jackson

, et al. (2005). Overlap of pathology in paralytic rabies and axonal Guillain-Barre syndrome. Ann Neurol, 57, 768–772.

30.

Thanomsridetchai

, Singhto

, Tepsumethanon

, et al. (2011). Comprehensive proteome analysis of hippocampus, brainstem, and spinal cord from paralytic and furious dogs naturally infected with rabies. J Proteome Res, 10, 4911–4924.

31.

Topley

WWCT

. Wilson's principles of bacteriology v, immunology. (1998). Topley & Wilson's Microbiology and Microbial Infections. Vol.1, Virology, Arnold, London.

32.

Vaziri

, Rahimpour

, Eslami

, et al. (2006). RP-HPLC prefractionation and its application in expressional proteomics analysis of an in vitro viral infection model. J Sep Sci, 29, 2284–2291.

33.

Vaziri

, Torkashvand

, Eslami

, et al. (2012). Comparative proteomics analysis of mice lymphocytes in early stages of infection by different strains of rabies virus. Indian J Virol, 23, 311–316.

34.

Venugopal

, Ghantasala

, Selvan

, et al. (2013). Quantitative proteomics for identifying biomarkers for rabies. Clin Proteomics, 10: 3.

35.

Wang

, Zhang

, Sun

, et al. (2011). Proteomic profiles of mouse neuro N2a cells infected with variant virulence of rabies viruses. J Microbiol Biotechnol, 21, 366–373.

36.

Wang

, Sarmento

, Wang

, et al. (2005). Attenuated rabies virus activates, while pathogenic rabies virus evades, the host innate immune responses in the central nervous system. J Virol, 79, 12554–12565.

37.

, Xiong

, and Mei

. (2010). To build a synapse: Signaling pathways in neuromuscular junction assembly. Development, 137, 1017–1033.

38.

, Kim

, Park

, et al. (2007). Identification of molecular markers for the oncogenic differentiation of hepatocellular carcinoma. Exp Mol Med, 39, 641–652.

39.

, and Levine

. (2011). Toll-like receptor, RIG-I-like receptors and the NLRP3 inflammasome: Key modulators of innate immune responses to double-stranded RNA viruses. Cytokine Growth Factor Rev, 22, 63–72.

40.

Zandi

, Eslami

, Soheili

, et al. (2009). Proteomics analysis of BHK-21 cells infected with a fixed strain of rabies virus. Proteomics, 9, 2399–2407.

41.

Zandi

, Eslami

, Torkashvand

, et al. (2013). Expression changes of cytoskeletal associated proteins in proteomic profiling of neuroblastoma cells infected with different strains of rabies virus. J Med Virol, 85, 336–347.