The effect of bovine viral diarrhea virus (BVDV) strains on bovine monocyte-derived dendritic cells (Mo-DC) phenotype and capacity to produce BVDV

Background Dendritic cells (DC) are important antigen presentation cells that monitor, process, and present antigen to T cells. Viruses that infect DC can have a devastating impact on the immune system. In this study, the ability of bovine viral diarrhea virus (BVDV) to replicate and produce infectious virus in monocyte-derived dendritic cells (Mo-DC) and monocytes was studied. The study also examined the effect of BVDV infection on Mo-DC expression of cell surface markers, including MHCI, MHCII, and CD86, which are critical for DC function in immune response. Methods Peripheral blood mononuclear cells (PBMCs) were isolated from bovine blood through gradient centrifugation. The adherent monocytes were isolated from PBMCs and differentiated into Mo-DC using bovine recombinant interleukin-4 (IL-4) and granulocyte-macrophage colony-stimulating factor (GMCSF). To determine the effect of BVDV on Mo-DC, four strains of BVDV were used including the severe acute non-cytopathic (ncp) BVDV2a-1373; moderate acute ncp BVDV2a 28508-5; and a homologous virus pair [i.e., cytopathic (cp) BVDV1b TGAC and ncp BVDV1b TGAN]. The Cooper strain of bovine herpesvirus 1 (BHV1) was used as the control virus. Mo-DC were infected with one of the BVDV strains or BHV-1 and were subsequently examined for virus replication, virus production, and the effect on MHCI, MHCII, and CD86 expression. Results The ability of monocytes to produce infectious virus reduced as monocytes differentiated to Mo-DC, and was completely lost at 120 hours of maturation. Interestingly, viral RNA increased throughout the course of infection in Mo-DC, and the viral non-structural (NS5A) and envelope (E2) proteins were expressed. The ncp strains of BVDV down-regulated while cp strain up-regulated the expression of the MHCI, MHCII, and CD86 on Mo-DC. Conclusions The study revealed that the ability of Mo-DC to produce infectious virus was reduced with its differentiation from monocytes to Mo-DC. The inability to produce infectious virus may be due to a hindrance of virus packaging or release mechanisms. Additionally, the study demonstrated that ncp BVDV down-regulated and cp BVDV up-regulated the expression of Mo-DC cell surface markers MHCI, MHCII, and CD86, which are important in the mounting of immune responses.


Background
Bovine viral diarrhea virus (BVDV) is an important disease of the cattle industry in the USA and worldwide. The virus belongs to the family Flaviviridae and is a single-stranded, positive-sense RNA virus with a genome of approximately 12.5 kb [1]. BVDV can further be classified on the basis of genotype and biotypes. Specifically, genotypes are divided into Type 1 (BVDV1) or Type 2 (BVDV2), and are distinguishable using monoclonal antibodies [2], while BVDV biotypes are classified as either cytopathic (cp) or non-cytopathic (ncp), based on the effect the virus has on cell culture [3]. Likewise, BVDV infection in cattle has multiple variations. Infection of the bovine fetus with ncp BVDV during the first trimester often results in persistently infected (PI) calves that are immunotolerant to BVDV, and thus remain a source of infection to other animals. Additionally, superinfection of PI animals with an antigenically homologous cp strain of BVDV typically results in fatal mucosal disease [4]. In the acutely infected animal, initial infection and replication of BVDV occurs in the oronasal mucosa and oropharyngeal lymphoid tissue [5], and the subsequent systemic spread occurs through lymphatic and blood circulation systems [6]. The BVDV virus can be detected in the blood approximately 4 to 8 days after initial exposure, and the buffy coat sample comprises the white blood cells (WBCs) present in peripheral blood, typically contain more virus than serum [7]. The buffy coat contains various antigen presenting cells including monocytes and dendritic cells. The DC actively provides surveillance for antigen and presents it to T cells after processing. As such, the infected DC may play an important role in BVDV dissemination in body. The infected DC may have altered surface marker expression that interferes with mounting an effective immune response. In this study, Mo-DC were used as an in vitro model of DC. The ability of BVDV to replicate and produce infectious virus in monocytes and Mo-DC was investigated along with the effect of BVDV infection on MHCI, MHCII, or CD86 expression. Four strains of BVDV were used in this study including the severe acute ncp BVDV2a 1373 strain, the moderate acute ncp BVDV2a 28508-5 strain, and a virus pair (cp BVDV1b TGAC and ncp BVDV1b TGAN) recovered from an animal that died of mucosal disease.

BVDV do not affect the Mo-DC viability
To determine the effect of BVDV infection on Mo-DC viability, Mo-DC were infected with cp BVDV1b TGAC, ncp BVDV1b TGAN, ncp BVDV2b 1373, or ncp BVDV2b 28508-5 strains of BVDV at a multiplicity of infection (MOI) of 6. Infected and mock-infected control Mo-DC were collected at 1, 6, 12, 24, 48, and 72 hour(s) post-infection. The viability of Mo-DC was determined via trypan blue exclusion assay, as described by Strober [8]. Results revealed that Mo-DC viability was not altered significantly (p > 0.05) during the course of infection, regardless of BVDV strain ( Figure 4).

Mo-DC lose the ability to produce infectious BVDV during differentiation
Cells were collected during the intermediate stages of differentiation from monocytes to Mo-DC (i.e., at 48, 72, 96, and 120 hours) and infected with ncp BVDV2a 1373 at an MOI of 6. The cells were then collected at 24 or 48 hours post-infection and analyzed for virus titer in cell lysate, as described by Reed and Muench [9]. Results indicated that the ability of Mo-DC to produce virus is lost during differentiation, and production of infectious virus was completely halted by 120 hours of differentiation. At 48 hours of differentiation from monocyte to Mo-DC, cells produced 3.69 ± 0.00 log10/ml and 2.69 ± 0.00 log10/ml infectious BVDV at 24 and 48 hours post-infection, respectively; whereas, at 72 hours of differentiation, the intermediate monocyte to Mo-DC produced 2.19 ± 0.71 log10/ml and 1.68 ± 0.00 log10/ml BVDV at 24 and 48 hours post-infection, respectively. At 96 hours of differentiation, BVDV titers from Mo-DC were comparatively less (i.e., 1.69 ± 0.00 log10/ml and 0.85 ± 1.20 log10/ ml at 24 and 48 hours post-infection, respectively), and no virus was produced by Mo-DC at 120 hours of differentiation at 24 and 48 hours post-infection ( Figure 7).

BVDV replicates viral RNA in Mo-DC
Mo-DC were not able to produce infectious BVDV; therefore, the ability of Mo-DC to produce BVDV RNA was examined. Viral RNA was extracted from BVDVinfected Mo-DC at 0, 1, 6, 12, 24, 48, 72, 96, 144, 168, and 192 hours post-infection. Viral RNA was quantified at each time point using quantitative real-time polymerase chain reaction (qRT-PCR) [10]. Testing indicated that viral RNA replicated in Mo-DC. Interestingly, the kinetics of viral RNA production differed between viral strains. Specifically, replication of viral RNA for the ncp BVDV2a 1373 strain began, peaked, and started to decline at 24, 144, and 192 hours post-infection, respectively ( Figure 8A); replication of cp BVDV2a 28508-5 viral RNA began at 6 hours, began to decline at 96 hours, and was no longer producing viral RNA at 144 hours post-infection ( Figure 8B); replication of the cp BVDV1b TGAC viral RNA began, peaked, and started to decline at 1, 144, and 168 hour(s) post-infection, respectively ( Figure 8C); and production of the ncp BVDV1b TGAN viral RNA began replicating at 12 hours, peaked at 72 hours, and began to decline at 144 hours post-infection.
Comparisons were made between replication of BVDV RNA in monocytes and Mo-DC. Freshly collected monocytes were infected with ncp BVDV2a 1373 at an MOI of 6. The viral RNA was extracted from infected monocytes at 0, 1, 6, 12, 24, 48, 72, 96, 144, 168, and 192 hour(s) post-infection and was quantified at each time point using qRT-PCR. Results showed that BVDV viral RNA started to replicate in monocytes as early as 1 hour post-infection and had ceased at 168 hours post-infection; whereas replication of viral RNA in Mo-DC began at a later time point and had a longer duration (i.e., 24 through 168 hours post-infection) ( Figure 9).
Mo-DC produced the BVDV non-structural (NS5A) and envelope (E2) viral proteins Viral RNA replicated in Mo-DC, but infectious virus was not produced; therefore, to determine whether or not the RNA was translated into viral proteins, a Western Blot was performed. The test was completed using lysate harvested at 72 hours from Mo-DC or MDBK cells infected with ncp BVDV2a 1373 at an MOI of 6. The ncp BVDV2a 1373 infected MDBK cells were used as positive control while mock infected Mo-DC or MDBK cells were used as negative control. The results showed that Mo-DC lysate contained NS5A ( Figure 10A) or E2 ( Figure 10B), indicating that the infected Mo-DC produced viral proteins but did not release infectious virus.
The cp biotype up-regulated and ncp biotypes downregulated MHCI, MHCII, and CD86 expression on Mo-DC Following infection or mock-infection (i.e., control cells) with the four strains of BVDV, the Mo-DC were stained for MHCI, MHCII, or CD86 and fixed at 0, 1, 6, 12, 24, 48, and 72 hour(s) to determine the effects of infection on cell surface expression. The mean fluorescent intensity (MFI) in mock-infected Mo-DC at 0 hours was utilized as 100%, the mock-infected Mo-DC at each time point were used as control, and the percent change at each time was calculated.
Expression of MHCII increased during the course of infection in Mo-DC infected with cp BVDV1b TGAC. MHCII expression increased in cp BVDV1b TGAC infected Mo-DC from 100% at 0 hour to 156.83 ± 54.48% (i.e., roughly 56%) at 72 hours post-infection The MHCII expression on Mo-DC following 6 hours infection with cp BVDV1b TGAC was significantly higher than its time point mock-infected control (P > 0.05) (Figure11B).

Discussion
This study revealed that Mo-DC and its progenitor cells (i.e., monocytes) could become infected with the BVDV virus; and while monocytes subsequently produced infectious virus, Mo-DC did not. Furthermore, virus production decreased with monocyte differentiation to Mo-DC and had completely ceased by120 hours of differentiation; thus, monocytes lost the ability to produce infectious virus with differentiation to Mo-DC. It was also determined that Mo-DC supported BVDV viral RNA replication but; the kinetics of viral RNA production was different between different viral strains. Additionally, viral proteins were translated in BVDV infected Mo-DC; thus, the accumulation of BVDV viral RNA and its subsequent translation into viral proteins suggest that failure to produce infectious virus may be the result of a hindrance in viral assembly, packaging, or release in Mo-DC. Further studies using electron microscopy may confirm the actual level of hindrance in infectious virus production.
A previous study that utilized a homologous pair of BVDV ncp and cp viruses (i.e., Pe515ncp and Pe515cp) revealed that Mo-DC were able to replicate and produce infectious BVDV [11]. The difference in the ability of the Mo-DC to produce infectious virus between that study and the present study may be due to the use of different monocyte isolation and Mo-DC culturing methods, as the previous study utilized CD14 + Mo-DC, while the current study examined Mo-DC that were determined to be non-adherent CD14 − cells. CD14 is a cell surface marker specifically for monocytes and macrophages. It acts as a co-receptor for bacterial lipopolysaccharide (LPS) with toll-like receptor-4 (TLR 4) [12]. CD14 expression is reduced during differentiation of monocytes to Mo-DC [13,14]. Likewise, in another study, bovine CD11b + Mo-DC also produced infectious BVDV; however, that study did not describe the culturing method, phenotype, or morphology of the Mo-DC used [15]. CD11b is an adhesion molecule and complement receptor that is expressed on neutrophils, monocytes, macrophages and Mo-DCs [16,17]. The above cell surface molecules in combination with DEC205 are routinely used to characterize the Mo-DC. DEC205 is a specific marker of Mo-DC [18].
A number of studies have been performed using Mo-DC and hepatitis C virus (HCV; another member of the Flaviviridae family). Similarly, both BVDV and HCV have a similar structural organization, can cause chronic infections in their respective hosts, utilize the lowdensity lipoprotein (LDL) receptor to enter the host cells, and use a functionally similar internal ribosome entry site (IRES) for translation. In addition, both viruses use NS5B RNA-dependent RNA polymerase and have a similar mechanism for virion maturation, assembly, and release [19]. In one study, the J6/JFH strain (a chimeric HCV of genotype 2a) did not replicate in B or T lymphocytes, monocytes, macrophages, or dendritic cells, while it did replicate and produce infectious virus in Huh-7.5 cells [20]; whereas, another study found that Mo-DC infected with the JFH1 strain of HCV with an MOI of 1did not support HCV RNA replication or antigen production [21]. In contrast to the present study, the mechanism responsible for the inability to produce infectious HCV was different, as BVDV viral RNA was replicated and translated into viral protein. This variation in HCV and BVDV replication may be due to their replication requirements, as HCV tropism is restricted   to hepatocytes of humans and chimpanzees [22], while BVDV can replicate in epithelial and non-epithelial cells [23].
A trial utilizing influenza virus in mouse bone marrowderived DC resulted in abortive replication. However, viral genome transcription and replication occurred for each gene segment along with the translation of viral hemagglutinin and nucleoprotein. Further the study using electron microscopy examination revealed that the resulting failure was associated with defective viral release, possibly due to insufficient synthesis or stability of one or more viral proteins required for release of the virus [24]. Similarly, the current study showed the abortive replication of BVDV in Mo-DC. The exact mechanism of abortive BVDV replication in Mo-DC is not known. However future ultrastructural examination of BVDV infected Mo-DC using electron microscopy could be intriguing to understand the stage at which production of infectious virus is blocked (i.e. assembly, egress etc).
The current study showed that Mo-DC precursor monocytes produced infectious BVDV while Mo-DC failed to do so. It would be interesting to know the dynamics of Mo-DC differentiation/maturation and viral production. Previous studies have demonstrated that Mo-DC differentiated with GMCSF and IL-4 were immature and could be matured by proinflammatory cytokines, CD40L, IFN-γ or LPS [13,25]. There has been little work done on effect of BVDV strains on Mo-DC maturation. While it is well established that cp BVDV induced more IFN-γ than ncp BVDV in vitro [26,27], it is likely that cp BVDV strains induce Mo-DC maturation more than ncp BVDV strains in vitro. To confirm this hypothesis, a future study would be helpful using different BVDV strains along with lipopolysaccharide (LPS) as maturation differentiation control.
One of the factors that may affect the production of virus in Mo-DC is interaction with other cell populations. In this in vitro study, isolated cell populations were used; whereas, testing in the in vivo environment may result in different outcomes. Interaction of Mo-DC with T cells in vivo may trigger the production of infectious virus. In one study, the interaction of antigen presenting cells (APC) with T cells enhanced HCV infection, as the interaction of APC with T cells led to activation of T cells [27]. Activated T cells produce interleukin-2 (IL-2), a substance that enhances CD81 expression in thymocytes [28]. Expression of CD81 on thymocytes facilitates the entry of HCV into cells, and thus, its replication [29]. Similarly, co-stimulation through interaction of CD28 and B7 increased HIV type 1 replication [30]. Therefore, it stands to reason that other cell populations could be affected. For example, the macrophage may become infected by phagocytizing the apoptotic Mo-DC containing viral RNA and may then begin producing infectious BVDV. As these probabilities are theoretical, exploration using T cell activation assay or in vivo trials would be beneficial.
The cell susceptibility can greatly be influenced with changes in its surface receptor expression, particularly receptors that mediate virus entry. BVDV infects cells via Receptor-mediated endocytosis using the LDL receptor [19]. Given that oxidized LDL has been demonstrated to promote differentiation of dendritic cells from monocytes in other species [31], it would be interesting to explore the dynamics of LDL receptor expression during monocyte to Mo-DC differentiation with respect to cell infectivity and virus production. These future studies could also explore the effect of different BVDV on LDL expression for better understanding of the BVDV pathogenesis.
In the current study, infection with BVDV affected cell surface marker expression on Mo-DC. The cp BVDV1b-TGAC strain enhanced MHCI, MHCII, and CD86 expression, while ncp strains of BVDV (i.e., ncp BVDV2a 1373, ncp BVDV2a 28508-5, and ncp BVDV1b TGAN) reduced the MHCI, MHCII, and CD86 expression during the course of infection. In another study that used bovine monocyte-derived macrophages (MDM), MHCI was up-regulated following infection with cp BVDV and was down-regulated in cells infected with a ncp strain of BVDV, while both cp and ncp strains down-regulated the MHCII expression in MDM [32]. The difference in MHCII results between that study and the present study may be due to a difference in the cell types chosen for analysis, as the Mo-DC reported here were non-adherent CD14 − cells, while the MDM were adherent CD14 + cells. A study using differential detergent fractionation (DDF) analysis of bovine monocytes showed that 53 bovine proteins involved in the immune function of professional APC were altered following BVDV infection. These altered molecules include adhesion 151 molecules, toll-like receptors (TLR 1, 6, and 8), antigen uptake, MHC I and II, cytokines, and growth factor synthesis molecules [33]. The above study supported the current finding that showed (See figure on previous page.) Figure 11 Effect of BVDV on Mo-DC cell surface marker expression. Differentiated Mo-DC were infected with either of four strains of BVDV, including ncp BVDV2a 1373, ncp BVDV2a 28508-5, or a homologous pair of ncp or cp BVDV1b viruses (ncp BVDV1b TGAN or cp BVDV1b TGAC) with an 6 MOI. The cells were stained for A) anti-MHCI; B) anti-MHCII or C) anti-CD86 antibody at 1, 6, 12, 24, 48, and 72 hour(s) post-infection. The mock-infected Mo-DC) at each time point was used as control (white legend). The mean fluorescent intensity (MFI) of expression was analyzed using flow cytometry. The each experiment was repeated three times. Asterisks (*) indicate significant difference from controls (P < 0.05).
the alteration of MHCI, MHCII and CD86 expression in BVDV infected Mo-DC. The above study also showed the alteration of cell adhesion molecules and capability of antigen uptake by APC, that may affect the Mo-DC infectivity and ultimately virus producing capability that need to be explored.
The changes noted in surface marker expression on Mo-DC could be due to differences in interferon expression between cp and ncp biotypes of BVDV. Type 1 IFN plays an important role in up-regulation of MHCI expression [34], thus, reduced type 1 IFN production in Mo-DC infected with ncp BVDV may explain the down-regulation of MHCI and MHCII. Similarly, up-regulation of MHCI and MHCII in Mo-DC infected with cp BVDV may be due to an increase in type 1 IFN [26,35]. It would be interesting to validate the effect of Type 1 IFN on cell surface marker expression by using poly I:C treated cells as control. The previous study showed that ncp BVDV inhibited of IFN induction was a result of blocking the IRF-3 pathway [27]. It may be possible that BVDV affects other pathways such as the melanoma differentiation-associated gene 5 (MDA5) signaling pathway, which will need to be explored for better understanding of BVDV pathogenesis. Alternatively, infection may have a direct effect on MHC folding and assembly, and consequently, expression. For example, infection with HCV changed MHCI expression in infected cultured cancerous cells lines (i.e., human hepatoma and mouse lymphoma cells) via disruption in MHCI protein folding and assembly [36][37][38]. This study found that HCV replicons induced endoplasmic reticulum (ER) stress that resulted from a decline in protein glycosylation. The decrease in protein glycosylation disrupted protein folding and prevented the assembly of MHCI molecules. Reduced assembly of MHCI molecules ultimately resulted in MHCI down-regulation. Similarly, Human Immunodeficiency Virus (HIV) was shown to reduce MHCI and CD86 expression via the Nef protein in a human embryonic kidney cell line (293 T), the human monocytic U937 cell line, as well as in mouse macrophages and dendritic cells. The Nef protein binds and endocytoses MHCI molecules by the ARF6 pathway, thus resulting in reduced MHCI expression [39,40]. The mechanism by which BVDV infection alters cell surface marker expression needs to be explored, as this information would provide a better understanding of immunosuppression caused by BVDV.

Conclusions
Findings from the current study demonstrated that Mo-DC failed to produce infectious BVDV. The ability to produce infectious BVDV was reduced with differentiation of monocyte to Mo-DC and completely stopped at 120 hours of differentiation. However, BVDV viral RNA replicated and was translated into viral proteins in Mo-DC, indicating that inability to produce infectious virus may be due to assembly or release of virus. Infection of Mo-DC with the cp strain of BVDV up-regulated expression, while infection with ncp strains of BVDV down-regulated expression of the cell surface markers MHCI, MHCII, and CD86. The up-regulation of cell surface marker expression that results from cp BVDV infection may explain why a better immune response is induced when modified live virus vaccines that contain cp strains of BVDV are used. On the other hand, the down-regulation of cell surface marker expression by ncp strains of BVDV may be an important function of immunosuppression and subsequent development of persistent infection.

Animals
Brown Swiss calves (8-12 months of age) were used in this study. All animals were deemed healthy prior to use and were housed at the Dairy Farm at South Dakota State University, Brookings, SD, U.S.A. Animal handling and blood collection were approved (approvals # 09-039E and 12-052A) and performed according to the guidelines of the SDSU Institutional Animal Care and Use Committee.

Virus
The Cooper strain of bovine herpesvirus1 (BHV1) and four strains of BVDV [i.e., a homologous pair of ncp and cp viruses (ncp BVDV1b TGAN and cpBVDV1b TGAC) recovered from an animal that died of mucosal disease; the severe acute ncp BVDV2a 1373); and the moderate acute ncp BVDV2a 28508-5 strain] were used in this study. All the BVDV virus stock used in the study was kindly provided by Dr. Julia F Ridpath (Ruminant Diseases and Immunology Research Unit, National Animal Disease Center, USDA, Ames, IA, USA).

Virus titration
Freshly collected monocytes, fully differentiated Mo-DC, and MDBK cells were infected with ncp BVDV2a 1373, ncp BVDV2a 28508-5, cp BVDV1b TGAC, or ncp BVDV1b TGAN at an MOI of 6. The cell lysate or supernatant were collected at 0, 1, 6, 12, 24, 48, 72, and 96 hour(s) post-infection. Virus titers in cell lysate or supernatant at each time point were determined as described by Reed and Muench [9]. Additionally, cells were infected with BHV1 at the same MOI and were used as positive control for virus replication at the same time points.
To determine at which time point of differentiation Mo-DC lose infectious virus production capacity, the monocytes, cells of intermediate stages at 48, 72, 96 and 120 hours of differentiation were infected with ncp BVDV2a 1373 at an MOI of 6. The cells were collected at 24 and 48 hours post-infection and analyzed for virus titer in cell lysate.
To compare the viral replication dynamics in monocytes, fully differentiated Mo-DC and MDBK cells, the cells were infected with ncp BVDV2a1373. Viral RNA was isolated and quantified using qRT-PCR, as described above. The RNA extracted from non-infected cells was used as internal control at each reaction.

Western blot
To determine whether BVDV viral RNA was translated into viral proteins in Mo-DC, a Western blot was performed targeting one structural protein (envelope protein -2: E2) that play a major role in virus attachment and entry to host cell [41] and non-structural protein 5A (NS5A), that interacts with the host cellular protein and inhibits NF-B activation [42]. Western blot was performed as described by Devireddy and Jones [43], with some modifications. The MDBK cells or Mo-DC were infected with ncp BVDV2a 1373 at an MOI of 6for 72 hours, washed with PBS, and then lysed with 200 μl of RIPA buffer containing the complete ULTRA tablet containing protease inhibitor (Roch Diagnostic, GmbH, Sandhofen, Germany). The mock infected MDBK cells or Mo-DC were used as control. The cell lysate was centrifuged at 13,000 rpm for 5 minutes and the supernatant was collected. Forty μl of cell lysate was loaded in 12.5% sodium dodecyl sulfate -polyacrylamide gel electrophoresis (SDS-PAGE) and the dual color protein ladder (10-250kD; Bio-Rad, Hercules, CA, USA) was used as a standard. The proteins were transferred onto a nitrocellulose membrane (Whatman GmbH, Dassel, Germany), blocked with5% skim milk in PBS for 30 minutes at room temperature, and then incubated at 4°C overnight with either anti-BVDV NS5A rabbit polyclonal antibody (produced in our laboratory using a recombinant BVDV NS5A as an immunogen) or anti-BVDV E2 mouse monoclonal antibody (15-C5, IDEXX Laboratories, Inc., Westbrook, Maine, USA) at a dilution of 1:1000 in PBS containing 1% skim milk. Following three washes with PBS containing 0.5% Tween 20, the membrane was incubated overnight at 4°C with for BVDV E2 or BVDV NS5A with diluted (1:3000) PBS containing 1% skim milk and either goat anti-mouse (IRDye 800CW, LI-COR Biosciences, Lincoln, NE, USA) or goat anti-rabbit (IRDye 800CW, LI-COR Biosciences, Lincoln, NE, USA), respectively. The specific protein band was visualized by using the Odyssey Imaging system and software (LI-COR Biosciences, Lincoln, NE, USA).