An alternative receptor to poly I:C on cell surfaces for interferon induction

Authors

  • Itsuro Yoshida,

    Corresponding author
    • Department of Microbiology and Immunology, Asahikawa Medical University, Midorigaoka Higashi 2-1, Asahikawa, Hokkaido 078-8510, Japan
    Search for more papers by this author
  • Masanobu Azuma

    1. Department of Microbiology and Immunology, Asahikawa Medical University, Midorigaoka Higashi 2-1, Asahikawa, Hokkaido 078-8510, Japan
    Search for more papers by this author

Correspondence

Itsuro Yoshida, Midorigaoka Higashi 2-1, Asahikawa, Hokkaido 078-8510, Japan.

Tel: +81 66 68 2391; fax +81 66 68 2397; email: iyoshida@asahikawa-med.ac.jp, iyoshida@af.wakwak.com

ABSTRACT

Not-self or denatured nucleic acids are recognized by pattern recognition receptors localized mainly in endosomes and cytoplasm, such as Toll-like receptor (TLR) 3, TLR7, TLR9, retinoic acid-inducible gene-I, DNA-dependent activator of IFN-regulatory factors and other receptors. The binding of polyriboinosinic:polyribocytidylic acid (poly I:C), a synthetic dsRNA that robustly induces type I interferon, to a putative cell-surface receptor on a rabbit kidney cell line, RK13, has been analyzed by the authors and RK13 cells found to capture poly I:C in a specific fashion with sufficient affinity. These findings suggest that an alternative receptor to poly I:C participates in the induction of type 1 interferon, which localizes on cell surfaces. Although the nature of this molecule has not yet been identified, accumulating evidence has led the present authors to speculate that there are undefined classes of RNA-recognition molecules on cell surfaces and that these are unlikely to be categorized as previously reported dsRNA receptors. Although many years have passed since this possibility was first reported by the present authors, it remains attractive. In this article, previously reported cell-surface dsRNA receptors are reviewed in comparison with other receptors reported to date that are firmly involved in the innate immune-sensing of nucleic acids.

Abbreviations
AIM2

absent in melanoma 2

DEAE

diethylaminoethyl

DAI

DNA-dependent activator of IFN-regulatory factors

DAMP

damage-associated molecular pattern

IFN

interferon

MDA5

melanoma differentiation-associated antigen 5

PAMP

pathogen-associated molecular pattern

PFU

plaque-forming unit

poly I:C

polyriboinosinic:polyribocytidylic acid

PRR

pattern-recognition receptor

PR-RK

poly I:C resistant RK

RIG-I

retinoic acid-inducible gene I

TLR

Toll-like receptor

Pathogen associated molecular patterns and DAMPs are recognized by PRRs and induce activation of the innate immune system, including production of type I IFN. After TLR3 was identified as one of the poly I:C receptors participating in interferon induction [1], a number of membrane and cytoplasmic PRRs have been reported to be sensors for PAMPs and DAMPs. Table 1 summarizes these PRRs, which recognize nucleic acid molecular patterns [2-8].

Table 1. Pattern-recognition receptors to nucleic acids participating in innate immunity in mammalian system
PRRLigandsLocalization
TLR 3dsRNA, poly I:CEndosome (2,3)
 RNAs from damaged cellsCell surface (3,4)
TLR 7ssRNAEndosome (2)
 Imidazoquinoline 
TLR 8ssRNAEndosome (2)
 Imidazoquinoline 
TLR 9Non-methylated CpGEndosome, ER (2)
 Chromatine DNA 
RIG-IdsRNA (short)Cytoplasm (5)
 5′-triphosphated RNA 
MDA 5dsRNA (long)Cytoplasm (6)
 poly I:C 
DAIdsDNACytoplasm (7)
AIM 2dsDNACytoplasm (8)

All these nucleic acid-recognizing PRRs reside in immune cells such as monocytes and dendritic cells and localize on the endosomal membrane, endoplasmic reticulum or in the cytoplasm. On the other hand, TLR3 reportedly localizes on cell surfaces of fibroblast cells [4]. Other PRRs such as TLR1, TLR2, TLR4, TLR5, TLR6, which recognize bacterial peptidoglycan, lipopolysaccharide and flagellin, localize on cell surfaces. The presence of sensor molecule on cell surfaces facilitates rapid deployment of the host defense system. We have examined the possibility of the occurrence of nucleic acid receptors on epithelial and fibroblast cell surfaces. Our earlier results will be discussed in conjunction with recent findings concerning PRRs for recognition of nucleic acids.

ESTABLISHMENT OF POLY I:C RESISTANT CELL LINE FROM RK13

RK13, an epithelial-like cell line of rabbit kidney cells that is sensitive to the biological effects of poly I:C, produces IFN when treated with poly I:C with and without DEAE-dextran. After mixing of poly I:C and DEAE-dextran, complexes of polyanions and polycations are formed; these complexes are internalized by the cell by endocytosis, a type of transfection of high molecular weight nucleic acids. However, it is difficult to internalize poly I:C itself without DEAE-dextran. RK13 cells appear to produce IFN without internalization of poly I:C, even when treated with poly I:C alone. RK13 cells are also sensitive to the cytotoxic effects of poly I:C. After exposure of RK13 cells to poly I:C, IFN is produced and large numbers of RK13 cells undergo cell death, which is caused by poly I:C with and without DEAE-dextran.

Taking these basic experiments into consideration, we tried to establish a subline of RK13. After repeated treatment of RK13 cells with poly I:C without DEAE-dextran (Fig. 1), we have established a cell line that is completely non-sensitive to both IFN inducibility and cytotoxicity of poly I:C and have named this cell line PR-RK (poly I:C resistant RK) [9]. PR-RK cells produce IFN after treatment with poly I:C with DEAE-dextran, indicating the presence of the intracellular IFN induction system that recognizes internalized poly I:C. We interpreted this finding as meaning that RK13 cells usually possess both cell-surface and cytoplasmic receptors for poly I:C and that PR-RK cells have lost their cell-surface receptors, making them resistant to extrinsic stimulation by poly I:C. However, PR-RK cells still harbor cytoplasmic receptors to poly I:C, which can respond to internalized poly I:C. In fact, binding of 3H labeled poly I:C to PR-RK is reportedly 60–70% of its binding to RK-13 [9]. However, the possibility of RK13 cells having a dual recognition system and lacking the surface one that PR-RK cells have had not been proven at the time of that report[9].

Figure 1.

Establishment of PR-RK cells.

PARTIAL CHARACTERIZATION OF POLY I:C RESISTANT-RK

32P labeled poly I:C was prepared by using T4 polynucleotide kinase and the binding of 32P labeled poly I:C at 37°C by RK-13 and PR-RK compared [10]. Both in RK13 and PR-RK, the binding was saturated within 30 mins of exposure to poly I:C and the binding of poly I:C to PR-RK was almost 60% that of RK-13, which is essentially the same as previous results [9]. Interestingly, the binding of 32P labeled poly A:U, a double-stranded RNA of polyriboadenylic acid and polyribouridylic acid, to PR-RK is essentially the same as its binding to RK-13 [10], suggesting that the cell surface receptor to poly I:C we predicted is sufficiently specific to the chemical structure of nucleic acids. Consistent with these binding characteristics, PR-RK cells are sensitive to the cytotoxic effects of poly A:U [10]. A murine monoclonal antibody (6D1) that inhibits IFN induction by poly I:C in RK 13 [10] was prepared and used to attempt detection of a component of the poly I:C receptor in the solubilized membrane protein fraction of RK-13 cells. A 60 kDa protein band was detected by western blot analysis with monoclonal antibody 6D1 [10], suggesting this protein is one component of the cell surface receptor to poly I:C of RK-13 cells.

To examine the binding of poly I:C on cell surfaces, we have cloned a hyper-sensitive cell line from RK13 and a non-sensitive cell line from PR-RK. Figure 2 shows the sensitivities of those cell clones to the antiviral effects of poly I:C. When the cells were pretreated with poly I:C, more than 80% of a cell clone, 3A2 derived from RK13, survived after infection with 102 PFU/well of challenge virus. On the other hand, almost none of the cell clone 5B11 derived from PR-RK survived at 102 PFU/well of challenge virus with and without pretreatment of poly I:C. Using these two cell clones, 3A2 and 5B11, we compared the binding efficiencies of poly I:C to the cells at 4°C. Figure 3 shows the binding of 32P labeled poly I:C to 3A2 and 5B11 cells. The differences between 3A2 and 5B11 in such binding show a saturated binding curve of ligand to receptor (dotted line in Fig. 3), and Scatchard plot analysis of this saturated binding curve gives a linear line (data not shown). These findings indicate the presence of the specific poly I:C receptor on the cell surface at around 104 sites/cell. The dissociation constant Kd was 4.0 × 10−9 M in these experiments, indicating nM order affinity between receptor and ligand (Table 2).

Figure 2.

Sensitivity of cloned cells to antiviral effect of poly I:C. The upper panel shows RK-13 and its clone 3A2 and the lower panel PR-RK and its clone 5B11. Cells in 96-well microplates were treated with (closed) or without (open) poly I:C alone for 1 hr, then washed and replenished with fresh medium. After 24 hrs cultivation the cells were infected with VSV at the indicated PFU/well. Forty-eight hours later, the cells were stained and their viabilities measured by using an optical microplate-reader. Non-infected cells were defined as 100% survival.

Figure 3.

Direct binding analysis of poly I:C on clone 3A2 and 5B11 cells. 32P labeled poly I:C was added to cells cultured in 24-well plates placed on ice, the cells were incubated at 4°C for 30 mins, then washed, solubilized and the radioactivities of the bound poly I:C to cells counted by using a liquid scintillation counter. Results were normalized by CPM/10 μg protein of cells in each well. Subtraction of the binding to 5B11 cell from that to 3A2 cells produced data for the saturated binding curve (3A2-5B11/dotted line).

Table 2. Summary of Scatchard plot analysis of poly I:C binding on cell surfaces
CellsBinding dataKd (M)Binding site/cell
RK13Fig. 34.0 × 10−91.0 × 104
3A2   
RK13Fig. 46.4 × 10−96.4 × 103
3A2   
HFCFig. 51.1 × 10−84.3 × 103

A general method for analyzing the specific binding of ligand to receptor is to subtract non-specific binding measured in the presence of excess amounts of unlabeled ligand. Figure 4 shows the binding of 32P-poly I:C to 3A2 cells in the absence and presence of excess amount (2 mg/mL) of unlabeled poly I:C. Subtraction of poly I:C binding in the presence of unlabeled poly I:C from binding in the absence of unlabeled poly I:C provided a saturated binding curve. From this, we calculated that the number of receptors for poly I:C is 6.4 × 103 sites/cell and their affinity Kd = 6.4 × 10−9 M by Scatchard plot analysis (Table 2). These results are essentially the same as those shown in Figure 3, indicating that the binding of 32P-poly I:C to 5B11 cells is non-specific; that is, non-specific binding does not participate in the biological effects of poly I:C. Based on these binding analyses, we defined the specific binding of poly I:C to the cell surface receptors of RK13 cells that would sustain induction of IFN, as well as help poly I:C to attain the endosomes where TLR3 are situated.

Figure 4.

Direct binding analysis of poly I:C on clone 3A2 cells. In the presence (+I:CEX) or absence (32P-I:C) of excess amounts of unlabeled poly I:C (2 mg/mL),32P labeled poly I:C was added to cells cultured in 24-well plates placed on ice and the cells incubated at 4°C for 30 mins. Radioactivities of bound poly I:C were counted as described in Figure 3. Results were normalized by CPM/10 μg protein of cells in each well. Subtraction of the binding to 3A2 cells in the presence of unlabeled poly I:C produced data for the saturated binding curve (32P-I:C − I:CEX/dotted line).

The findings described above indicate that the subtraction method for determining non-specific binding by measuring in the presence of excess amount of unlabeled ligand is applicable to the analysis of poly I:C receptors in other cells. We have therefore analyzed the binding of 32P-poly I:C to the cell surfaces of human diploid fibroblast cells. Like RK-13 cells, these cells are also sensitive to IFN inducibility of poly I:C without DEAE-dextran. As shown in Figure 5, the specific binding of 32P-Poly I:C to those cells provides a saturated binding curve from which we calculated that the number of receptors for poly I:C on human diploid fibroblast cell surfaces is 4.3 × 103 sites/cell and the affinity Kd = 1.1 × 10−8 M by Scatchard plot analysis (Table 2).

Figure 5.

Direct binding analysis of poly I:C on human diploid fibroblast cells. In the presence (+I:CEX) or absence (32P-I:C) of excess amount of unlabeled poly I:C (2 mg/mL), 32P labeled poly I:C was added to cells cultured in 24-well plates placed on ice and incubated at 4°C for 30 mins. Radioactivities of bound poly I:C were counted as described in Figure 3. Results were normalized by CPM/10 μg protein of cells in each well. Subtraction of the binding to HFC in the presence of unlabeled Poly I:C produced data for the saturated binding curve (32P-I:C − I:CEX/dotted line).

CONCLUSIONS

The findings described above clearly verify the presence of alternative poly I:C receptors that possibly participate in IFN induction on the cell surfaces of rabbit epithelial and human fibroblast cells. The number of cell surface receptors for poly I:C is 4–6 × 103 sites/cell and their affinity for poly I:C Kd = 6–11 × 10−9 M, as shown in Table 2. In 2002, it was reported that TLR3 localizes on the cell surface of human fibroblast cells and that function-blocking studies using a specific monoclonal antibody against human TLR3 showed that surface-expressed TLR3 participates in induction of type I IFN by poly I:C [4]. In 2009, it was also reported that Scavenger receptor class-A on the surface of human epithelial cells recognizes dsRNA and participates in induction of type I IFN by poly I:C [12]. Our preliminary analysis of membrane proteins solubilized from human fibroblast cells suggested a 60 kDa protein as a candidate for the receptor component that constitutes the cell surface poly I:C receptor we assigned from our experiments [10]. Because human TLR3 is a >100 kDa protein, the poly I:C receptor or its component we discovered has not yet been identified or characterized. Although, according to recent reports, TLR3 is cleaved into short fragments by lysosomal enzymes [13], we did not determine what happened on cell surface receptors at that time. A recent report by Matsumoto and her colleagues provides evidence that, besides scavenger receptor class A, CD14 and DEC205, other cell-surface receptors for dsRNA for capturing dsRNA outside cells must exist and that these facilitate carrying dsRNA over endosomes together with raftlin [14]. The receptor for dsRNA capture has not yet been identified. The nature of our poly I:C receptor would remind us of this fascinating speculation for future studies.

Accumulating observations have led us to conclude that several kinds of sensor molecules recognize not-self or altered nucleic acids on the surfaces of non-immune cells; these may include the alternative receptor to poly I:C we observed in our experiments with rabbit and human cells. Whether or not these cell surface receptors for poly I:C in epithelial and fibroblast cells are competent at inducing type I IFN will be a crucial issue. Linking our results to the current knowledge on the RNA-sensing system will predict the importance of our receptors for promotion of activation of the innate immune system by sensing RNA.

ACKNOWLEDGMENTS

The authors express many thanks to Dr. Tsukasa Seya, Department of Microbiology and Immunology, Hokkaido University School of Medicine, for encouragement to publish this review.

DISCLOSURE

The authors declare no financial or commercial conflict of interest.

Ancillary