**2. Results**

To assess the e ffect of ricin inhalation on miRNA expression, 12 mice were intranasally challenged with a lethal dose of ricin or saline as a negative control. Twenty-four h after exposure, lungs were excised and total RNA was purified for multiplex miRNA profiling using the Nanostring nCounter system. Of the 600 mature mouse miRNAs probed, 182 had expression levels above the negative control probes (Table S1a). A total of 21 miRNAs were found to be deregulated in the ricin group compared to control samples (Table S2), 9 were significantly downregulated, and 12 were significantly upregulated, all with *p*-values of <0.05. Principal component analysis (PCA) using the di fferentially expressed (DE) miRNAs revealed sample clustering according to ricin exposure status (Figure 1). We used qPCR to validate the Nanostring results (Figure S1). To further corroborate our findings, the ricin exposure experiment and lung RNA extraction were repeated with an additional set of mice (*n* = 12). Five miRNAs (miR-223, miR-1224, miR-503, miR-10a, and miR-200c) exhibited statistically significant changes in expression that matched the results of the first Nanostring analysis.

Since ricin has been shown to induce rapid, massive migration of inflammatory cells—predominantly innate immune cells [11]—we utilized the innate DB database [12], which aims to capture improved coverage of the innate immunity interactome. A total of 54 immune miRNAs appear in the database. Of these, three (miR-223, miR-10a, miR-200c) were di fferentially expressed in our cohort. This number was significantly higher than expected by chance (*p*-value <0.005). To strengthen the immune regulation enrichment observation and to delineate its deregulated pathways, total RNA (the same RNA used for the first miRNA multiplex analysis) was sent for RNA-sequencing (RNA-seq). Major di fferences in gene expression were detected between the groups, with significant expression downregulation (*p* < 0.05) of 2823 genes and upregulation of 3147 genes (Table S1b). The ricin and control groups were distinctly discriminated by their gene expression (Figure S2). These vast di fferences in gene expression concur with previous descriptions of transcriptome analyses of models of acute lung injury mice [13]. For ontology analysis of genes that presented with statistically significant di fferences, high variation in expression (absolute fold-change ≥1) revealed enrichment for immune cell responses after exposure to pathogenic agents, including IL-1 and NF-kB pathway activation, leukocyte chemotaxis and migration, and chemokine activation (Figure 2). Utilizing validated miRNA–target interactions databases, we noted that many of the di fferentially expressed RNA-seq genes were directly regulated by the deregulated five miRNAs (Table 1).

**Figure 1.** miRNA expression, ricin versus control treated. (**a**) PCA analysis of miRNA expression. The ricin and control groups are clearly distinguished by their miRNA expression profiles, as demonstrated by unsupervised clustering. (**b**) Expression of the top eight differentially expressed (DE) miRNAs. Each gray dot represents one sample. The red line indicates mean expression. All miRNAs had adjusted *p*-values < 0.05.

**Figure 2.** Gene ontology enrichment analysis of ricin intoxication. Barplot representation of the gene ontology biological function (**a**) and molecular processes (**b**) analysis of ricin induced transcriptome. The analysis comprised 2236 genes that had over 1-fold change expression between the ricin and control groups, with adjusted *p*-value <0.05. Both biological and molecular analyses showed enrichment for immune cell recruitment and cellular response to infection. The X-axis describes the number of genes involved in each process.



Next, we identified the diseases with over-presentation of the same highly changed genes, using disease ontology and MEDLINE/PubMed indexed articles. Enrichment in lung-related diseases of both infectious and immune related origin was detected (Figure 3). With the aim of translating these results to a clinical setting, we examined whether the ricin induced lung miRNA changes could also be detected in peripheral blood mononuclear cells (PBMCs) as a surrogate tissue that can be safely obtained from patients. PBMCs provide a large pool of gene transcripts that have demonstrated the potential to be highly sensitive to the disease microenvironments on a system-wide basis [15,16]. PBMCs were isolated from mouse blood after intranasal ricin or saline exposure. RNA was extracted and expression levels of miR-223, miR-1224, miR-10a, miR-200c, and miR-503 were evaluated using real time PCR. miR-223 was significantly upregulated in the ricin group, similar to its trend of expression in the lungs. However, other miRNAs did not present the same expression pattern (Figure 4).

**Figure 3.** Disease enrichment analysis of ricin intoxication. Dotplot representation of ricin intoxication disease ontology. The analysis comprised 2,236 genes that had over 1-fold change expression between the ricin and control groups, with adjusted *p*-value < 0.05; these genes were evaluated for over-presentation in (**a**) PubMed and MEDLINE databases and (**b**) Disease Ontology. GeneRatio is the number of genes involved in the specific process divided by the total number of genes (2,236). Dot sizes ("count") represent the total number of predicted gene targets of the total number of genes that are known to be involved in the listed processes and dot color indicates statistical significance. Enrichment of lung related diseases of infectious and immune natures can be detected.

**Figure 4.** Relative expressions of miR-223, miR-1224, miR-10a, miR-200c, and miR-503 in peripheral blood mononuclear cells (PBMCs). Expression levels of miR-223, miR-1224, miR-10a, miR-200c, and miR-503 in PBMCs of ricin and control groups. miR-223 was found to be significantly upregulated in the ricin group, similar to its expression pattern in lung tissue. miR-1224, miR-10a, miR-200c, and miR-503 expression was not significantly different among the study groups. Data are represented as mean ± SEM. \* *p* < 0.05.


**Table 2.** Comparison of miRNAs in studies of injured lungs.
