Table 1.
ConA-induced dry eye disease in rabbits.
Fig 1.
DED histopathology of the rabbit.
Cross tissue sections of superior bulbar conjunctiva on day 18 from NZW rabbits (H&E stain). Representative images from naïve (A) and rabbis with DED (B), with the latter showing disrupted architecture (increased stratification, goblet cell loss, and chronic inflammatory infiltrate) not present in the normal tissue. Scale bar = 100 μm.
Fig 2.
Differential gene expression in the conjunctiva of rabbits with DED.
Transcriptome profiling of conjunctival tissue from rabbits with DED and normal controls was determined using microarrays as in Methods. The numbers of upregulated (red arrows) and downregulated (blue arrows) genes are shown; they are based on their annotation or human homology.
Fig 3.
Cluster analysis of gene expression in normal and DED rabbits.
A. Heatmap of 831 differentially expressed gene transcripts from Control (left) and DED (right) rabbits (n = 3 each). Red and blue colors denote upregulation and downregulation, respectively. B: Scatterplot representing the data distribution between DED and normal rabbits. C: The volcano plot depicts the statistical significance (p value) versus fold change in geneexpression. Two subclusters of the genes overex-pressed (ORM1) or downregulated (PIP) in DED rabbits compared to normal contols (enlargement to right of A).
Fig 4.
Clusters of immune response and cell cycle regulation in DED.
Of the genes upregulated (FC >1.5, p<0.05) in rabbit DED, 55% were in two clusters, A (red balls), corresponding to the innate and adaptive immune response, and B, corresponding to the cell cycle regulating gene network (blue balls). The analysis was performed as in Methods.
Table 2.
Selected differentially expressed genes based on fold change.
Table 3.
The 25 most involved pathways in rabbit DED.
Table 4.
Putative rabbit DED transcript markers not previously correlated with the disease.
Fig 5.
RT-PCR analysis of immune response genes.
A. Correlation of RNA expression levels of inflammatory mediators in DED rabbits obtained with both: RT-PCR and microarray methods. (PCC R = 0.622, n = 84, p-value < 0.001.) B. The relative expression (RT-PCR) of 18 selected genes representing innate, transition (APC, antigen presenting cells) and adaptive immunity. Each bar represents mean ± SD of fold change (n = 8, *p < 0.001) in DED compared to normal mRNA levels, normalized to the 5 host genes as in Methods.
Fig 6.
Correlation of gene expression between Sjogren syndrome and rabbits with DED.
A. Gene expression levels (fold change, FC) in rabbits with DED were correlated with that in DED patients with Sjogren’s syndrome (FC≥1.3 up- or down-regulated, p-value <0.001, n = 18). PCC R = 0.756, p<0.001. B. Similar correlation between rabbit transcripts and human DED protein markers included in this analysis (FC≥1.3, p-value<0.001, n = 9) was found (R = 0.662, p<0.05), as in Methods.
Fig 7.
Correlation of gene expression between DED patients without Sjogren syndrome and rabbits with DED.
A. Gene expression in rabbits with DED was compared to that in patients with non-Sjogren’s syndrome DED. There was a significant correlation between the 32 most changed rabbit and the corresponding human genes included in this analysis (FC≥1.3 up- or down-regulated, PCC R = 0.545; p-value <0.001; n = 32). B: The expression of 475 of 910 human genes included in this analysis (that were annotated in rabbit transcriptome arrays), was similar in both species (PCC R = 0.917, p<0.0001), as described in Methods. The most correlated genes were cut-off by using the inflection point, determined by the function (small graph on the left) of similarly expressed genes (in both species) and calculated PCC Rs (S1 Table).
Fig 8.
Gene expression in patients with DED not related to Sjogren syndrome.
A. Venn Diagram depicts the overlap between rabbit and human non-Sjogren’s DED microarray data. Of the 910 overlapping and annotated genes, 217 were similarly upregulated and 283 downregulated in both species, the remaining 410 were changed in opposite directions. The distribution of these genes based on the threshold of statistical significance (TAC, as in Methods) is also shown. B. 28 genes were subtracted from the p<0.001 group, based on their FC values and listed in descending order according to the FC in rabbit gene expression. The String database (DB) revealed a cluster among them related to the immune response (upper part), whereas the Reactome DB system identified several pathways, with the top one being the immune system, common to human and rabbit transcriptomes.
Fig 9.
Transcriptome changes during the vicious cycle of DED.
The four components of the vicious cycle of the pathophysiology of DED are shown. Representative changes in gene expression observed in rabbits with DED are linked to each stage of the cycle. Changes in genes involved in transcriptional control (depicted at the center of the cycle) likely affect all stages of DED evolution.