If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Otorhinolaryngology & Clinical Allergy Center, The First Affiliated Hospital, Nanjing Medical University, Nanjing, ChinaInternational Centre for Allergy Research, Nanjing Medical University, Nanjing, China
Corresponding author. Department of Otorhinolaryngology & Clinical Allergy Center, The First Affiliated Hospital, Nanjing Medical University, 300 Guangzhou Road, Nanjing, 210029, China
Affiliations
Department of Otorhinolaryngology & Clinical Allergy Center, The First Affiliated Hospital, Nanjing Medical University, Nanjing, ChinaInternational Centre for Allergy Research, Nanjing Medical University, Nanjing, China
Circular RNAs (circRNAs) are involved in inflammation; however, their role in allergic rhinitis (AR) remains unclear. In this study, we analyzed circRNA expression and identified a circRNA-miRNA-mRNA network through which circRNAs regulate AR pathogenesis.
Methods
We analyzed circRNA, miRNA, and mRNA expression profiles in the nasal mucosa by high-throughput sequencing (HTS), using a fold-change >1.5 and p-value < 0.05 to pinpoint significantly differentially expressed (DE) circRNAs, miRNAs, and mRNAs in AR. A DEcircRNA-DEmiRNA-DEmRNA crosstalk network was then constructed using bioinformatics and statistical analysis. Gene ontology and Kyoto encyclopedia of genes and genomes pathway analyses were performed to identify the biological terms enriched in the network; whereas RT-PCR and Sanger sequencing were used to confirm the circRNAs.
Results
A total of 264 DEcircRNAs were identified by HTS, including 120 upregulated and 144 downregulated in AR compared to controls. A DEcircRNA-DEmiRNA-DEmRNA crosstalk network was constructed with 17 miRNAs, 11 circRNAs, 29 mRNAs, and 64 interaction pairs. These genes were involved in the Wnt signaling pathway, TNF biosynthesis, inflammatory responses, the PI3K-Akt signaling pathway, and Toll-like receptors. Of the 11 DEcircRNAs, hsa_circ_0008668 and circTRIQK were upregulated, whereas hsa_circ_0029853 and circRNA_01002 were downregulated in AR tissues. Sanger sequencing confirmed the back-splicing junctions of these circRNAs.
Conclusions
We constructed a novel DEcircRNA-DEmiRNA-DEmRNA network for AR that provides a basis for future studies to investigate its underlying molecular mechanisms.
Allergic rhinitis (AR) is a chronic multi-factorial inflammatory disease, with a prevalence rate of 17.6% in adults that is expected to increase yearly in China.
have reported that circANKRD36 expression positively correlates with the level of inflammatory factor IL-6 and may be involved in type 2 diabetes mellitus and inflammation-related pathways. Moreover, Liu et al
have found circHIPK3 is highly expressed in AR mucosa, and that circHIPK3 can upregulate Th2 cell-specific transcript factor GATA-3 via modulating miR-495. Wang et al
For instance, circRNAs have been shown to regulate mRNAs by sponging miRNA, ie, competitively binding to miRNAs via common miRNA response elements (MREs).
have constructed a network of differentially expressed (DE) circRNAs, miRNAs, and mRNAs, with gene enrichment analysis revealing that they regulate cell adhesion, cell activation, and immune responses in atherosclerosis. CircRNA-related ceRNA networks have also been established for illustrating the pathogenesis of various diseases, including liver cancer,
However, no such networks have been elucidated for AR.
Therefore, in this study, we described circRNA, miRNA, and mRNA expression profiles in the nasal mucosa from patients, with and without AR, using high-throughput sequencing (HTS), and constructed a DEcircRNA-DEmiRNA-DEmRNA crosstalk network to illustrate the potential molecular mechanisms of AR progression.
Methods
Clinical data and nasal mucosa tissues
The nasal mucosa tissues were collected from a cohort of 20 adult subjects (10 with perennial AR and 10 non-allergic controls) aged 18–53 years who received endoscopic septoplasty from August 2019 to January 2020. Tissue samples were immediately frozen and stored in liquid nitrogen, and then subjected to transcriptomic profiling analysis. AR was diagnosed according to the clinical guidelines.
Serum total IgE and allergen-specific IgE were measured by the Automated Immuno-Strip Analysis System (LG Life Sciences Ltd., Korea) in all the subjects. Patients with perennial AR enrolled were sensitized mainly by house dust mites, including Dermatophagoides pteronyssinus and/or Dermatophagoides farinae (Table S1). The controls had negative IgE test results in sera. All subjects had ceased treatment with glucocorticoids, antihistamines, antileukotrienes, and other immunomodulating drugs for at least 4 weeks prior to surgery. Subjects who had an immune deficiency, upper respiratory tract infection, chronic rhinosinusitis, nasal polyps, sinonasal neoplasms, asthma and other allergic diseases, and a history of nasal surgery or immunotherapy were excluded.
Total RNA isolation, library construction, and sequencing
Total RNA was extracted from nasal mucosa samples using TRIzol reagent (Life Technologies, CA, USA) and its quantity and quality were measured using a NanoDrop ND-1000 (Thermo Scientific, DE, USA). RNA integrity was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA) and RNAs with an integrity value of ≥7 were retained.
Six randomly selected samples (three AR and three controls) from the 20 subjects were used to construct circRNA, miRNA, and mRNA libraries, and the 6 samples and the remaining 14 samples were sent to reverse transcription-polymerase chain reaction (RT-PCR). Approximately 1 μg of RNA was extracted from each sample, and ribosomal and linear RNAs were depleted from the samples using Ribo-zero (Illumina, CA, USA) and RNase R (Epicenter, CA, USA), respectively, according to the manufacturers' instructions. CircRNA libraries were constructed using a TruSeq Stranded Total RNA HT/LT Sample Prep Kit (Illumina) according to the manufacturer's instructions. Library quality was assessed using an Agilent 2100 system (Agilent Technologies) and the libraries were sequenced using an Illumina sequencing platform (HiSeqTM 2500).
RNA (1 μg) from each of the 6 samples was used as input material for small RNA library generated using TreSeq Small RNA Sample Prep Kits (Illumina), according to the manufacturer's instructions. Library quality was assessed using an Agilent 2100 system and libraries were sequenced using an Illumina sequencing platform (HiSeqTM 2500). All sequencing procedures and analyses were performed by OEbiotech (Shanghai, China).
Identification of miRNA-target interactions
MiRNA-mRNA and miRNA-circRNA pairs were selected using miRanda, which predicts miRNA targets based on sequence complementarity, conserved target sites, and free formation energy. Next, we filtered out miRNAs with crosstalk with circRNAs and mRNAs and used the 3 types of RNAs to produce a circRNA-miRNA-mRNA network. To identify circRNA-mRNA pairs, we used a ceRNA score that evaluates the significance (p-value) of miRNAs shared by each circRNA and mRNA as the ratio of common miRNA MRE to circRNA-miRNA binding MRE, as follows:
where M1 represents the quantity of all miRNAs, M2 represents the quantity of miRNAs regulating mRNAs, M3 represents the quantity of miRNAs with regulatory effects on the circRNA, and M4 represents the quantity of common miRNAs. p-values < 0.05 were considered statistically significant.
Bioinformatics and data analysis
Gene ontology (GO) analysis was used to annotate gene function. Enrichment scores were calculated as the negative logarithm of the p-value and used to determine the statistical significance of GO term clusters targeted by DE genes. Kyoto encyclopedia of genes and genomes (KEGG) analysis was used to determine the biological pathways involving genes in the network, as quantified by the enrichment score.
Sanger sequencing and RT-PCR
Sanger sequencing was performed by Tsingke Biotechnology (Nanjing, China) to confirm the back-splicing junctions of each circRNA. Sanger sequencing data that matched the HTS results and displayed a unimodal dissolution curve were retained. To ensure that each PCR amplification fragment contained junction sites, the designed circRNA primers were divergent, instead of the frequently used convergent primers (Table S2). The experiment was repeated 3 times for each sample from the 20 subjects (10 with perennial AR and 10 controls). The RT-PCR dissolution curve was unimodal, proving that each pair of specific reverse primers had good repeatability and a single product.
Statistical analysis
All statistical analyses were carried out using SPSS 20.0 (IBM SPSS, Chicago, IL, USA). Between-group differences were analyzed using Student's t-tests. CircRNA expression was displayed as a fold-change calculated using the 2-△△ct method on the RT-PCR data. Data from at least 3 independent experiments were presented as the mean ± standard deviation (SD). p-values < 0.05 were considered statistically significant.
Results
CircRNA expression and characteristics in the nasal mucosa
Three pairs of nasal mucosa tissues from 10 patients with AR and 10 controls were randomly selected for HTS analysis. The clinical characteristics of the subjects are shown in Table S1. All normalized sample datasets were found to display similar circRNA distribution densities (Fig. 1A). CIRI software
was used to identify and annotate 30,936 circRNAs from the 2 groups, including 12 376 known and 18 560 novel (Fig. 1B), and widely distributed circRNA transcripts were found on all chromosomes (Fig. 1C). CircRNA transcript expression was quantified using the spliced reads per million (RPM) distribution density, revealing that almost all circRNA transcripts were expressed at low levels (Fig. 1D).
Fig. 1CircRNAs in the nasal mucosa. (A) Box plots showing that circRNA distribution intensities from all normalized datasets were similar between the six samples. (B) A total of 30,936 circRNAs were identified in the nasal mucosa, of which 12,376 were known and 18,560 were novel, by comparing the predicted results with the circBase database (http://www.circbase.org/). (C) Chromosomal circRNA expression distribution. (D) Spliced reads per million (RPM) circRNA distribution. circRNA, circular RNA
Among the DEcircRNAs, 1119 were antisense and 29 335 were sense, including 28 694 exonic, 641 intronic, and 1119 intergenic circRNAs. In particular, 93% of the circRNAs were sense exons and most circRNA transcripts carried 1–7 exons, with a length of 200–400bp or >2000bp. Hierarchical clustering revealed DE circRNAs between samples (Fig. 2A) and HTS analysis identified 264 DEcircRNAs, including 120 upregulated and 144 downregulated in AR. Based on the number of possible miRNA-binding sites, structure, energy score and p-value, the top 300 relevant circRNA-miRNA pairs were screened and their network illustrated using Cytoscape software (Fig. 3).
Fig. 2CircRNA, miRNA and mRNA expression. Clustered heat map showing all differentially expressed circRNAs (A), miRNAs (B) and mRNAs (C), respectively (fold-change > 1.5, p < 0.05). Red and blue represent high and low relative expression, respectively. circRNA, circular RNA; miRNA, microRNA
Fig. 3Top 300 predicted DEcircRNA-miRNA relationships. Some circRNAs were annotated using circBase. Novel circRNAs were described in terms of their genome location. Green and red nodes represent circRNAs and miRNAs, respectively. circRNA, circular RNA; miRNA, microRNA; DE, differentially expressed
MiRNA and mRNA expression profiles in the nasal mucosa
Next, we used HTS to analyze the miRNA and mRNA expression profiles in the nasal mucosa tissues from the AR patients and controls. Of the 1784 miRNAs available, 13 were upregulated and 15 were downregulated. On the other hand, of the 2030 mRNAs, 93 were upregulated and 79 were downregulated. Hierarchical clustering revealed that these DEmiRNAs (Fig. 2B) and DEmRNAs (Fig. 2C) were distinct between samples, confirming that miRNAs and mRNAs were DE in patients with AR, compared to controls.
DEcircRNA-DEmiRNA, DEmiRNA-DEmRNA, and DEcircRNA-DEmRNA interactions
DEmiRNA-DEcircRNA and DEmiRNA-DEmRNA interactions were detected using miRanda software, which defined 911 and 379 pairs, respectively. Next, we calculated the Pearson correlation coefficient (PCC) of DEmiRNA-DEcircRNA and DEmiRNA-DEmRNA pairs according to their expression values, with p-values < 0.05 and an absolute PCC of ≥ 0.7 (PCC ≥ 0.7 indicated positive correlation and PCC ≤ −0.7 indicated negative correlation). A total of 941 DEmiRNA-DEcircRNA and 865 DEmiRNA-DEmRNA pairs with negative expression correlation were screened out, while 116 DEmiRNA-DEcircRNA and 379 DEmiRNA-DEmRNA pairs were screened by combining miRanda and correlation analyses. Since circRNAs act as ceRNAs form RNAs, we constructed a DEcircRNA-DEmRNA network (Fig. 4) based on 11 and 29 positively correlated circRNAs and mRNAs, respectively.
Fig. 4Depiction of the DEcircRNA-DEmRNA network. The network includes 11 circRNAs (green nodes) and 29 mRNAs (yellow nodes). CircRNAs were annotated using circBase. Novel circRNAs were described in terms of their genome location. circRNA, circular RNA; DE, differentially expressed
DEcircRNA-DEmiRNA-DEmRNA network and functional annotation
Next, we identified miRNAs interacting with both circRNAs and mRNAs to construct a DEcircRNA-DEmiRNA-DEmRNA network, which included 17 miRNAs, 11 circRNAs (Table 1), 29 mRNAs, and 64 interaction pairs (Fig. 5).
Table 1Details of the 11 circRNAs in the DEcircRNA-DEmiRNA-DEmRNA network in AR
Fig. 5Depiction of the DEcircRNA-DEmiRNA-DEmRNA crosstalk network. The network includes 17 miRNAs (red nodes), 11 circRNAs (green nodes), and 29 mRNAs (yellow nodes). CircRNAs and miRNAs were annotated using circBase and miRbase, respectively. Novel circRNAs and miRNAs were described in terms of their genome location. circRNA, circular RNA; miRNA, microRNA; DE, differentially expressed
To determine the biological roles of the DEcircRNAs, we performed GO and KEGG analyses on the DEmRNAs in the DEcircRNA-DEmiRNA-DEmRNA crosstalk network. These genes converged in 159 terms, including 172 biological processes, 47 cell components, and 46 molecular functional components. The GO terms mainly involved the Wnt signaling pathway, lung epithelial cell development, and TNF biosynthesis (biological processes), endoplasmic reticulum lumen, flotillin complex, and intercellular canaliculus (cell components), fibronectin binding, collagen binding, and calcium: sodium antiporter activity (molecular functions; Fig. 6A). In addition, these genes were enriched in the inflammatory response and various signaling pathways (such as cell surface receptor signaling pathways), indicating that circRNAs may participate in multiple pathways related to AR pathogenesis. KEGG pathways analysis revealed 24 significantly involved pathways, of which local adhesion, ECM receptor interaction, protein degradation and absorption, and PI3K-Akt signaling pathways were the most enriched (Fig. 6B). However, other pathways were related to AR, such as the Toll-like receptor (TLR) signaling pathway, suggesting that circRNAs may act as ceRNAs in AR to regulate mRNAs by competitively binding to miRNAs.
Fig. 6Bioinformatics analysis of the DEcircRNA-DEmiRNA-DEmRNA crosstalk network. (A) GO enrichment for biological processes, cell components, and molecular functions, respectively. (B) KEGG pathway analysis of the involvement of linear transcripts in different biological pathways. Bar plot depicting the enrichment score of significantly enriched pathways. circRNA, circular RNA; miRNA, microRNA; DE, differentially expressed; GO, gene ontology; KEGG, Kyoto encyclopedia of genes and genomes
Sanger sequencing verification and RT-PCR quantification of circRNAs
The DEcircRNA-DEmiRNA-DEmRNA crosstalk network included 11 circRNAs (Table 1), of which two were known (hsa_circ_0008668 and hsa_circ_0029853) and nine were novel. To verify these circRNAs, hsa_circ_0008668, hsa_circ_0029853, circTRIQK (a sense-overlapping circRNA), and circRNA_01002 (an intergenic circRNA) were subjected to RT-PCR amplification and Sanger sequencing to confirm their back-splicing sites (Fig. 7A–D). The RT-PCR dissolution curve had a single peak, suggesting that the primers had good repeatability and yielded a single product type. In addition, RT-PCR revealed that hsa_circ_0008668 (Fig. 7E) and circTRIQK (Fig. 7G) were upregulated in AR, whereas hsa_circ_0029853 (Fig. 7F) and circRNA_01002 (Fig. 7H) were downregulated. These findings were consistent with the HTS data and suggested that the HTS results are highly reliable.
Fig. 7Sanger sequencing and RT-PCR of circRNAs. (A–D) Back-splice junction sequences for hsa_circ_0008668, hsa_circ_0029853, circTRIQK, and circRNA_01002 were confirmed by Sanger sequencing, respectively. Arrows represent divergent primers binding to the genomic region of circRNAs. (E–H) RT-PCR results for circRNAs in the nasal mucosa of ten patients with AR and ten control subjects. Hsa_circ_0008668 (E) and circTRIQK (G) were upregulated in AR. whereas hsa_circ_0029853 (F) and circRNA_01002 (H) were downregulated, consistent with the HTS data. Data represent the mean ± SD of at least three independent experiments.∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. circRNA, circular RNA
Recent epigenetic research has yielded considerable advances in our understanding of AR; however, these studies focused on protein-coding genes, miRNAs, and lncRNAs. Research on circRNAs has shown that they can act as new diagnostic markers for cancers.
the majority of which harbor highly conserved sequences and act in multiple cellular processes via post-transcriptional regulation. Thus, disturbed miRNA activity can be partly explained by circRNA involvement.
With the help of MREs, circRNAs, miRNAs and mRNAs participate in large-scale ceRNA crosstalk regulating post-transcriptional genes in various manners. Previous studies have found that Ddx17, circHIPK3 and circARRDC3 can function as ceRNAs to sponge miRNAs in AR.
However, each of these studies has just proved one circRNA related to immunity regulation. To the best of our knowledge, our study is the first to investigate circRNA expression profiles and ceRNA network in AR.
We predicted DEmiRNA-DEcircRNA interactions using miRanda and screened 116 pairs with negative expression relationships, including 17 miRNAs (such as hsa-miR-98-5p and hsa-miR-455-5p), while HTS revealed that hsa-miR-98-5p was significantly upregulated. Takuse et al
have found that hsa-miR-98-5p negatively correlates with peripheral IL-10+ cell count and IL-10. Moreover, an increasing number of studies have demonstrated imbalance between CD4+ T cells and B cells in peripheral blood plays a role in AR.
Hence, hsa-miR-98-5p may affect allergen tolerance in AR by negatively regulating Tregs, Bregs, and IL-10 secretion. Our DEmiRNA-DEcircRNA network revealed that hsa-miR-98-5p competed with several circRNAs, such as hsa_circ_0029853, suggesting that circRNAs may interact with hsa-miR-98-5p during AR immune tolerance.
Similarly, we identified 379 DEmiRNA-DEmRNA pairs and many DE genes (such as DOCK8 and MUC19) that were significantly associated with AR. In particular, DOCK8 expression was decreased in the nasal mucosa of patients with AR. DOCK8 deficiency plays a key role in allergic inflammation, susceptibility to infection, autoimmunity, and malignant tumors
have found that DOCK8 is related to Treg cell homeostasis and function and enhances peripheral B cell tolerance. In addition, DOCK8 modulates MyD88-dependent TLR signal transduction and activates the STAT3 transcription factor.
By examining DEmiRNA-DEmRNA relationships, we found that DOCK8 could bind to hsa-miR-548aq-3p, hsa-miR-582-5p and hsa-miR-96-5p, and that its expression correlated negatively with that of these miRNAs. We also found that MUC19, SLC8A3
in the ceRNA network were associated with AR, while PTPN22 and FCRL5 expression was downregulated in the nasal mucosa. Since single nucleotide polymorphisms in PTPN22
is associated with asthma and AR in Chinese children, future studies should focus on the mechanisms via which DE genes regulate AR.
Finally, we constructed a novel DEcircRNA-DEmiRNA-DEmRNA network for AR consisting of 64 ceRNA pairs, including 11 circRNAs bound to 17 miRNAs, such as hsa-miR-98-5p, to regulate the expression of 29 genes. Previous studies have found that AR is regulated by many signaling pathways, including the PI3K-Akt
signaling pathways. Consistent with these findings, we found that the 29 AR-related genes were involved in various important biological functions and pathways, including the PI3K-Akt and TLR signaling pathways, suggesting that these pathways play key roles in AR development.
The DEcircRNA-DEmiRNA-DEmRNA network also revealed that hsa_circ_0029853 could competitively bind to hsa-miR-98-5p, hsa-miR-140-5p, and hsa-miR-3065-5p to regulate ERRFI1 and NRCAM (Fig. 5), whose expressions were significantly downregulated in the nasal mucosa of patients with AR. ERRFI1 is also downregulated in human breast, skin, pancreatic, and ovarian cancers.
have found that ERFFI1 can significantly inhibit LPS-induced TNF-α, IL-1β, and IL-6 mRNA and protein expression. Moreover, our previous studies revealed that these pro-inflammatory cytokines were upregulated in the nasal mucosa of patients with AR,
suggesting that their overexpression may be related to ERRFI1 downregulation in these patients. In addition, ERRFI1 has been implicated in the epidermal growth factor receptor (EGFR) signaling pathway
Src-dependent EGFR transactivation regulates lung inflammation via downstream signaling involving ERK1/2, PI3Kδ/Akt and NFκB induction in a murine asthma model.
Moreover, intranasal administration of the EGFR inhibitor AG1478 has been found to significantly suppress GM-CSF, eotaxin-1, and RANTES secretion by nasal mucosal cells.
it may participate in AR pathogenesis by regulating the EGFR signaling pathway in the nasal mucosa. Additionally, ERFFI1 is regulated by multiple miRNAs, including miR-126,
; therefore, hsa_circ_0029853 may sponge hsa-miR-98-5p, hsa-miR-140-5p, and hsa-miR-3065-5p to inhibit ERFFI1 expression and thereby promote AR progression.
There are some limitations in the present study. First, we did not perform a functional analysis for the miRNAs (such as hsa-miR-98-5p, hsa-miR-455-5p) and genes (such as DOCK8, ERRFI1), especially their ceRNA network, to determine their roles in AR pathogenesis. The second is that we did not explore the mechanism of circRNAs in sponging miRNAs to regulating mRNAs, or other functions of circRNAs, such as acting as coding protein.
In conclusion, circRNA expression profiles and the ceRNA network for AR could provide a basis for future studies on the underlying molecular mechanisms of AR.
Abbreviations
AR, allergic rhinitis; miRNA, microRNA; lncRNA, long non-coding RNA; circRNA, circular RNA; ceRNA, competing endogenous RNA; MREs, miRNA response elements; DE, differentially expressed; HTS, high-throughput sequencing; GO, gene ontology; KEGG, Kyoto encyclopedia of genes and genomes; RT-PCR, reverse transcription-polymerase chain reaction; SD, standard deviation; RPM, reads per million; PCC, Pearson correlation coefficient; Treg, regulatory T cells; Breg, regulatory B cells; TLR, Toll-like receptor; EGFR, epidermal growth factor receptor.
Ethics approval and consent to participate
This study was approved by the Ethics Committee of the First Affiliated Hospital of Nanjing Medical University (Approval No. 2019-SR-023), and all the participants have signed written informed consent.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Consent for publication
All authors have approved the manuscript for submission.
Author contributions
C-Y Qiu and X-Y Cui performed data analysis and wrote the draft manuscript. M-P Lu, M Yin, W-Y Xu, and X-J Zhu collected samples and performed specimen quality screening. Q Yang performed RT-PCR. L Cheng designed the study and finalized the manuscript. All authors approved the final version of the manuscript.
Funding
This study was supported by a grant (No. 81970862) from the National Natural Science Foundation of China.
Potential competing interests
The authors report no competing interests.
Confirmation of unpublished work
A part of data and materials in this manuscript has been posted to the AUTHOREA preprint server at https://authorea.com/doi/full/10.22541/au.160193459.95947821/v1. The authors confirm that the final version of the manuscript is original, has not been published before, and is not currently being considered for publication elsewhere.
Acknowledgements
This study was supported by a grant (No. 81970862) from the National Natural Science Foundation of China. We thank Associate Professor Yong-Ke Cao at the College of Foreign Languages of Nanjing Medical University for professional English-language proofreading of the manuscript.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Src-dependent EGFR transactivation regulates lung inflammation via downstream signaling involving ERK1/2, PI3Kδ/Akt and NFκB induction in a murine asthma model.
00042-9&id=gr1.jpg){kind=link}
00042-9&id=gr2.jpg){kind=link}
00042-9&id=gr3.jpg){kind=link}
00042-9&id=gr4.jpg){kind=link}
00042-9&id=gr5.jpg){kind=link}
00042-9&id=gr6.jpg){kind=link}
00042-9&id=gr7.jpg){kind=link}