**3. Discussion**

In the present study, we explored changes in miRNA expression profiles following pulmonary ricin intoxication. We detected 21 DE miRNAs (9 upregulated and 12 downregulated) in mouse lung tissues 24 h after ricin exposure. Five miRNAs (miR-223, miR-1224, miR-503, miR-10a, and miR-200c) had similar changes in expression in a validation cohort. Next, we utilized the InnateDB dataset, which has been developed to facilitate system level investigations of the mammalian (human, mouse, and bovine) innate immune response [12], and noted significant enrichment of the DE miRNA in the InnateDB dataset (*p*-value <0.005). This stands in agreemen<sup>t</sup> with previous studies that explored ricin-induced lung injury and innate immune response [11,20]. Associations of miR-223, miR-1224, miR-10a, miR-200c, and miR-503 with immune regulation were previously established. miR-223 was implicated in polymorphonuclear cell development and function [21]. Further, miR-223 may play a crucial role during granulopoiesis [22]; it is upregulated throughout granulocyte differentiation and is the first miRNA that was found to dramatically alter granulocyte fate [23]. miR-223 critically fine-tunes myeloid cell activity and is involved in inflammatory diseases by regulating multiple gene transcripts including E2F1, NOD-like receptor activation, and the NF-κB pathway [24]. Overexpression of miR-223

was demonstrated to dampen acute lung injury [25]. miR-1224 was previously shown to regulate tumor necrosis factorα (TNFα) gene activity and miR-1224 expression was demonstrated to inversely a ffect LPS-induced TNFα mRNA expression [26]. Macrophages infected with mycobacterium demonstrated significantly higher miR-1224 expression, suggesting a potential role of miR-1224 in host responses upon mycobacterium infection [27]. miR-10a negatively regulates Bcl-6 expression in T cells and has an inverse e ffect on germinal center reactions [28]. miR-10a is involved in stabilizing the expression of Foxp3 in regulatory T cells [29] and can suppress proinflammatory monocytic cell activation by inhibiting the activation of the proinflammatory nuclear factor κB pathway [30]. miR-200c directly regulates expression of IL-6, IL-8, and CCL-5 transcripts by binding to their 3'UTRs [31]. It suppresses signaling pathways leading to NF-κB activation after TLR4 ligation; miR-200c mimics have been shown to cause decreased expression of transcripts encoding MyD88 and to induce the expression of inflammatory molecules in response to LPS [32]. Reduced miR-503 expression augments lung fibroblast VEGF production and promotes lung fibrosis [33]. Downregulation of miR-503 in bone marrow-derived mesenchymal stem cells was linked to attenuation of lung injury after infection [34], and miR-503 levels were found to be reduced in acute lung injury [35].

Lee et al. explored miRNA expression in acute lung injury (induced by LPS administration) and demonstrated significant changes in miR-223 and miR-503 expression, as well as in six other miRNAs not shared with our expression profile [17]. These discrepancies may be related to the vast di fferences between the ricin and LPS immune stimulation mechanisms [36]. Vaporidi et al. discovered miRNA expression profiles in ventilator-induced lung injury [18], which included changes in 65 miRNAs, among them miR-223, miR-503, and miR-200c. This abundance of deregulated miRNAs compared to our results could be explained by the pathologic di fferences in mechanic and pathogenic etiology of acute lung injury [37] and supports the notion that changes in miRNA expression in acute lung injury di ffer by their causative agents (also see Table 2).

To functionally validate the results of the miRNA expression analysis and to decipher its biological relevance, we investigated the lung transcriptome of mice exposed to ricin. Enrichment of immune regulation mediators and pathways, such as cellular response to bacterial stimulus, IL-1 and NF-kB signaling, leukocyte migration, and chemokine activity were discovered in biologic and molecular gene ontology analyses. Lung infections and immune related pathologies, such as chronic obstructive pulmonary disease, asthma, and bacterial infections were seen when utilizing the disease ontology database and PubMed indexed articles. Taken together, the miRNA expression and its transcriptome counterpart highlight the alteration in immune response following ricin intoxication.

PBMCs are a promising option for assessment of biologically distinct responses to various pathologies when the a ffected organs cannot be biopsied without further compromising the patient's health. Specifically, genomic analysis of PBMCs has been shown to distinguish between several lung related pathologies, including lung cancer [38], asthma [39], and pneumonia [40]. Here, we observed little resemblance between PBMC miRNA profiles and lung miRNA expression, with only the miR-223 PBMC profile mimicking its lung profile. A possible explanation for this is that most of the miRNAs expressed in ricin-induced lung toxicity originate in lung parenchyma (mostly lung epithelial cells) and the supporting stromal cells (fibroblasts, dust cells) and hence cannot be fully detected in PBMCs. Another potential explanation is that the time points used in this study are not conducive to a correlative observation between lung and PBMC miRNA responses. Future studies would benefit from a time-course assessment of miRNA profiles in response to ricin-induced intoxication.

A major limitation of our study is the usage of a rodent model to study human toxicity. Due to the complexity of running these experiments in humans, we cannot extrapolate our results to such a context. Nonetheless, we feel that our work represents an important attempt to understand the miRNAs and mRNAs modified during ricin intoxication in animals. Given that all five identified miRNAs are conserved in mammals, it is plausible that the outcome may have similarities among organisms of this class. Major strengths of our study include the comprehensive evaluation, identification, and validation of all miRNA and mRNA expression levels, and also the network and regulation analysis.

In conclusion, we describe a unique lung miRNA expression profile of pulmonary ricin intoxication that was validated in two separate in vivo experiments. The DE miRNAs detected are enriched for innate immune response and elucidate the regulatory roles of miRNAs following ricin exposure, as was further supported by transcriptome analysis. Additional studies should be conducted to further characterize the miRNA regulatory networks and to translate these findings into a miRNA-based diagnosis.
