Correction: Post-ischemic modification of neurogenesis and oligodendrogenesis in rodent models

Translating translational research: mouse models of human disease

The laboratory mouse is the foremost model organism for interrogating the genetic and molecular basis of human cancer and is a powerful platform for identifying therapeutically effective targets for prevention and treatment of cancer. Research using genetically engineered mouse models (GEMMs) have led to important advances in our understanding of the genetic basis of cancer susceptibility, the function of tumor suppressors and oncogenes, and therapy responses in preclinical and co-clinical studies. Patient Derived Xenograft (PDX) models are an increasingly important model system for in vivo studies of human cancer. These models are created by implanting patient tumors into immunodeficient or humanized mouse hosts and are a powerful translational research platform for preclinical and co-clinical studies. The number of GEMM and PDX mouse models increases significantly every year and the diverse cancer-related data about human cancer models tend to be distributed in ways that makes it difficult for researchers to integrate and interpret the information to find the most relevant model for their research. The Mouse Tumor Biology database (http://tumor.informatics.jax.org) is an expertly curated resource for information and data about genetically defined mouse strains and PDX models of human cancer. MTB provides query tools to enable integrated searches and visualization of these varied data, thus facilitating the assessment of novel mouse models of human cancer and potential preventative and therapeutic treatments. Enforcement of controlled vocabularies and standard gene, allele and strain nomenclature within MTB facilitates precise and comprehensive queries of MTB for pertinent mouse models. MTB contains data from spontaneous or endogenously induced tumors from genetically defined mice including tumor classification, incidence, Quantitative Trait Loci, pathology reports, images and genetic changes in the tumor (somatic) and background strain (germline) genomes. The PDX resource enables queries based on tumor type, cancer diagnosis and genomic properties of the engrafted tumors. Information in MTB is obtained from curation of peer-reviewed scientific publications and direct data submissions from individual investigators and large-scale programs. New features in MTB include the Faceted Tumor Search Form and a Reported Mouse Models table linking the most common fatal human cancers to reported equivalent mouse models. MTB contains over 77,000 Tumor Frequencies and over 2,200 Pathology Reports with over 6,600 images from over 4,200 references. MTB provides access to detailed clinical, pathological, expression and genomics data from over 400 PDX models. Information in MTB is integrated with cancer models data from other bioinformatics resources including PathBase, the Gene Expression Omnibus and ArrayExpress. MTB is supported by NCI grant CA089713. Citation Format: Dale A. Begley, Debra M. Krupke, Steven B. Neuhauser, Joel E. Richardson, John P. Sundberg, Janan T. Eppig, Carol J. Bult. Identifying therapeutically relevant mouse and patient-derived xenograft (PDX) models of human cancer using the mouse tumor biology database (MTB) data resource [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2804. doi:10.1158/1538-7445.AM2017-2804

Comparative transcriptomics of whole blood can be used to evaluate the systemic host response and its concordance between human and mouse malaria and aid the selection of appropriate models for translational malaria research.

The National Heart, Lung, and Blood Institute “Symposium on Phenotyping: Mouse Cardiovascular Function and Development” was held on October 10–11, 2002, at the Natcher Conference Center on the NIH campus in Bethesda, MD. In laying out the program for the meeting, we sought to bring together

Age-dependent behavioral and biochemical characterization of single APP knock-in mouse (APPNL-G-F/NL-G-F) model of Alzheimer's disease.

Animal models have been extensively used to test promising new agents for the treatment and prevention of cancer. Many different animal models are available for preclinical testing, and the choice of the specific model to use is often a critical step in successful drug development. Many different mouse models of human breast cancer have been developed that have been used to test promising anti-cancer drugs. These include mouse models that spontaneously develop mammary tumors, carcinogen-treated mouse models, xenograft models, transgenic mice and gene knockout mice that develop mammary tumors. The particular strengths and weakness of these models for testing therapeutic agents will be reviewed. The most widely used preclinical models for testing agents for the treatment of cancer are xenograft models. Xenograft studies using human cancer cell lines are easily conducted, are relatively rapid and, importantly, are recognized by the Food and Drug Administration as providing evidence of preclinical anti-tumor activity against human cancer. However, certain studies cannot be done using xenografts of human cancer cell lines. These include testing of immune-based therapies or testing of cancer preventive agents. For such studies, other models are needed. Testing of novel vaccines against cancer requires an immunocompetent host, and thus may require vaccination against murine tumors or studies to be done in 'humanized' mice. Studies of cancer preventive agents requires xenografts of human normal breast tissue or carcinoma-in-situ lesions, or alternatively mouse or rat models that develop tumors, either spontaneously or after carcinogen treatment. One of the most commonly used models for testing chemopreventive agents has been the carcinogen-treated rat model. This model has been successfully used to demonstrate the chemopreventive activity of many agents, including selective estrogen receptor modulators (SERMs) such as tamoxifen, raloxifene, and idoxifene, and retinoid compounds, and is particularly useful for testing agents for the prevention or treatment of estrogen receptor-positive mammary cancer. More recently, transgenic mouse models have been used to study the activity of chemopreventive agents, particularly for the prevention of estrogen receptor-negative breast cancer. We have used two such transgenic mice (C3(1)-SV40 T antigen, and mouse mammary tumor virus [MMTV]-erbB2 mice) to investigate the preventive activity of receptor-selective retinoid compounds. Both of these transgenic mouse lines develop premalignant lesions that then evolve into invasive mammary tumors that eventually metastasize. We have found that 9-cis-retinoic acid, which binds both retinoic acid receptors and retinoid X receptors (RXRs), and the RXR-selective retinoid, LGD1069 (bexarotene, Targretin), suppresses the development of non-invasive and invasive mammary tumors in both C3(1) SV40 T antigen mice and MMTV-erbB2 mice. These retinoids interfere with tumorigenesis by suppressing proliferation of normal and premalignant mammary epithelial cells, ultimately suppressing the development of invasive cancer. Based on these results in these mouse models, we initiated a human clinical trial using retinoids for the prevention of human breast cancer, which is now ongoing. These results demonstrate the utility of genetically engineered mouse models for the testing of molecularly targeted agents for the prevention of breast cancer.

Disease-associated oligodendrocyte signatures in neurodegenerative disease: the known and unknown.

Over the past three decades, immunodeficient mouse models carrying human immune cells, with or without human lymphoid tissues, termed humanized immune system (HIS) rodent models, have been developed to recapitulate the human immune system and associated immune responses. HIS mouse models have successfully modeled many human-restricted viral infections, including those caused by human cytomegalovirus (HCMV) and human immunodeficiency virus (HIV). HIS mouse models have also been used to model human cancer immunobiology, which exhibits differences from murine cancers in traditional mouse models. Variants of HIS mouse models that carry human liver cells, lung tissue, skin tissue, or human patient-derived tumor xenografts and human hematopoietic stem cells-derived-human immune cells with or without lymphoid tissue xenografts have been developed to probe human immune responses to infections and human tumors. HCMV-based vaccines are human-restricted, which poses limitations for mechanistic and efficacy studies using traditional animal models. The HCMV-based vaccine approach is a promising vaccine strategy as it induces robust effector memory T cell responses that may be critical in preventing and rapidly controlling persistent viral infections and cancers. Here, we review novel HIS mouse models with robust human immune cell development and primary and secondary lymphoid tissues that could address many of the limitations of HIS mice in their use as animal models for HCMV-based vaccine research. We also reviewed novel HIS rat models, which could allow long-term (greater than one year) vaccinology studies and better recapitulate human pathophysiology. Translating laboratory research findings to clinical application is a significant bottleneck in vaccine development; HIS rodents and related variants that more accurately model human immunology and diseases could increase the translatability of research findings.

Pharmacotherapy with two antiepileptic drugs in combination is usually prescribed to epilepsy patients with refractory seizures. The choice of antiepileptic drugs in combination should be based on synergistic cooperation of the drugs with respect to suppression of seizures. The selection of synergistic interactions between antiepileptic drugs is challenging issue for physicians, especially, if 25 antiepileptic drugs are currently available and approved to treat epilepsy patients. The aim of this study was to determine all possible interactions among 5 second-generation antiepileptic drugs (gabapentin (GBP), lacosamide (LCM), levetiracetam (LEV), pregabalin (PGB) and retigabine (RTG)) in the 6-Hz corneal stimulation-induced seizure model in adult male albino Swiss mice. The anticonvulsant effects of 10 various two-drug combinations of antiepileptic drugs were evaluated with type I isobolographic analysis associated with graphical presentation of polygonogram to visualize the types of interactions. Isobolographic analysis revealed that 7 two-drug combinations of LEV+RTG, LEV+LCM, GBP+RTG, PGB+LEV, GBP+LEV, PGB+RTG, PGB+LCM were synergistic in the 6-Hz corneal stimulation-induced seizure model in mice. The additive interaction was observed for the combinations of GBP+LCM, GBP+PGB, and RTG+LCM in this seizure model in mice. The most beneficial combination, offering the highest level of synergistic suppression of seizures in mice was that of LEV+RTG, whereas the most additive combination that protected the animals from seizures was that reporting additivity for RTG+LCM. The strength of interaction for two-drug combinations can be arranged from the synergistic to the additive, as follows: LEV+RTG > LEV+LCM > GBP+RTG > PGB+LEV > GBP+LEV > PGB+RTG > PGB+LCM > GBP+LCM > GBP+PGB > RTG+LCM.

I would like to thank Vijg and Hasty for their exhaustive commentary on our manuscript that was published recently in Aging Cell (Gentry & Venkatachalam, 2005). In this short article we showed that there are at least 23 genes deleted in the p53+/m mouse model and proposed an alternative hypothesis that questioned the role of p53 in the observed phenotypes (Tyner et al., 2002). As one of the primary authors in both of the above-mentioned articles, it is my responsibility to respond to the comments put forward by Vijg and Hasty that appear in this issue of Aging Cell. At the outset the authors say that our data and interpretation ‘confuse rather than clarify the issue of p53 as a putative pro-aging factor’. This gives an impression that the commentators assume that the ‘putative’ role of p53 in aging is an established fact and that our data are simply a distraction. The authors also cite three perceived problems with our hypothesis related to the comparisons of the p53+/m mouse model with two other mouse models and the lack of evidence for the role of 23 genes on the observed phenotypes. Here I have tried to separate the facts from fiction. Vijg and Hasty argue that it is ‘very likely that the ‘m’ allele produces a truncated form of p53’ in the p53+/m mouse. This speculation is not backed by in vivo evidence that shows the expression of the truncated protein product in p53+/m tissues. The data presented in the p53+/m manuscript are an in vitro transcription–translation experiment that showed the potential for the cloned ‘m’ allele to code for a protein product. The assumption regarding the in vivo existence of the m-protein has two important implications. First, the stability of the m-protein should not be affected due to the absence of the N-terminus that is involved in MDM2 binding that in turn leads to p53 degradation. However, as pointed out earlier, we were unable to detect this m-protein in vivo. Second, in the absence of MDM2-mediated protein degradation, the m-protein is most probably recognized as a misfolded protein and degraded rapidly by the ubiquitin system that might explain our inability to detect the truncated protein product. Even if one considers the remote possibility of the existence of the m-protein, its effects on full-length p53 function are unknown and purely imaginary. In addition the evidence Vijg and Hasty provide for the functional role of the m-protein (when the existence of the protein product is debatable) is just a comparison of the tumor incidence between the p53+/m mouse model and the p53+/null model. There is a fundamental flaw to this argument that makes a big assumption (in the absence of any experimental evidence) that the ‘m’ allele, along with the wt p53 allele, is responsible for the cancer resistance phenotype. The extrapolation of phenotypic data that discounts the heterozygosity of 23 genes in the p53+/m model while comparing it with the p53+/null model is weak, and has no bearing on the hypothesized role of the m-protein. It is also surprising that Vijg and Hasty fail to compare the longevity data for the p53+/m and the p53+/null mice that paradoxically show an increase in the lifespan of p53+/m mice despite their ‘aging-like’ symptoms. The mean lifespan of p53+/m mice is 96 weeks vs. the 71.5 weeks in p53+/null mice (Venkatachalam et al., 2001; Tyner et al., 2002). The dilemma for an impartial audience looking at the longevity data of the wt, p53+/m and p53+/null mice would be the choice of comparisons between the p53+/m model with that of the p53+/null model or the wild-type controls. In other words, if one compares the longevity of the p53+/null mice to that of the p53+/m mice, the data would indicate that the cancer resistance actually leads to an increased lifespan in the p53+/m mice. This caveat further strengthens the possibility that the diverse physiological effects of the haploinsufficiency of 23 genes may be the cause of the paradoxical tumor resistance combined with the appearance of ‘aging-like’ symptoms (that could also be ‘illnesses’) attributable to the large genetic deletion. Vijg and Hasty proceed to compare the phenotypes of the p53null/null model with the p53null/m model and say that the tumor phenotypes are exactly the same. The tumor suppressor functions of p53 are well proven and the complete loss of p53 (nullizygosity) leads to a tremendous increase in tumor susceptibility in mice (Donehower et al., 1992; Jacks et al., 1994; Purdie et al., 1994). Furthermore, the appearance of tumors in the p53null/m animals that occurs as early as 13 weeks of age simply precludes the analysis of aging-like or potential cancer resistance phenotypes in the p53m/null mice. This problem of studying the physiological functions of other genes on a p53null/null background is not new and has been shown to be true for the compound mutants (ku80–/–p53–/–) described earlier by Hasty and colleagues (Lim et al., 2000). It is also important to note that the data presented in our earlier paper indicate a slight delay in the cancer incidence of the p53null/m animals (Fig. 2b in Tyner et al., 2002). This observable delay could very well be due to the absence of 23 genes that function in a variety of pathways including lipid metabolism (lipoxygenase cluster), chromatin remodeling (Chd3) and cell cycle checkpoints (Aurora kinase B). It is important to note that this delay in cancer incidence occurs despite the overwhelming effects of p53 deficiency induced tumorigenesis. In the absence of this over-riding effect of p53 nullizygosity (and the rapid tumor induction), the haploinsufficiency of putative candidates within the deletion may confer an enhanced cancer resistance that becomes more pronounced and measurable in the p53+/m model. Vijg and Hasty also compare the basal levels and the regulation of p53 between the p53+/m and ‘super p53’ mouse model generated by Serrano and colleagues (Garcia-Cao et al., 2002). They point out that the basal levels of p53 are not affected and that the p53 protein is under normal regulatory control in the ‘super-p53’ model. This statement seems to imply that the full-length p53 protein levels are either abnormally high or unregulated in the p53+/m mice. A careful examination of our earlier paper will indicate that the basal levels of full-length p53 in mouse kidneys are actually similar between p53+/null and p53+/m mice while being higher in p53+/+mice (compare control lanes in Fig. 3b in Tyner et al., 2002). Furthermore the p53 protein is still under regulatory control in the p53+/m model as the full-length protein is inducible after DNA damage. However, the p53+/m mice seem to have a higher (∼ threefold) and more prolonged p53 induction than the p53+/null mice. This enhanced p53 protein induction begins to decrease over time indicating that the full-length protein is under regulatory control (Fig. 3b in Tyner et al., 2002). As another measure of p53 activity we also determined the induction of the cell cycle inhibitor p21 (a transcriptional target of activated p53) in wt, p53+/null and p53+/m mouse embryonic fibroblasts (Fig. 3c in Tyner et al., 2002). Interestingly, the p21 response in γ-irradiated mouse embryonic fibroblasts is more complex since the basal levels of p21 are highest in the wild-type cell lines, followed by the p53+/m and the p53+/null MEFs. Once again the initial induction of p21 mRNA levels in the p53+/m MEFs begins to decrease after 4 h indicating the presence of p53 regulation in the p53+/m mice. The constitutive overexpression or the misregulation of the p53 protein in the p53+/m model that Vijg and Hasty allude to is just a hypothetical possibility that awaits experimental evidence. They also compare the p53+/m model to the p44 transgenic mouse model described recently (Maier et al., 2004). Comparing these transgenic mice is like comparing apples to oranges for various reasons cited in Table 1. Even if one were able to accept their speculation about differences in the expression levels of the two truncated protein products (despite our inability to detect the m-protein in various p53+/m tissues) it is important to note that the p44 protein product retains an intact DNA binding domain whereas the m-protein lacks most of it. Furthermore, the time line of the phenotypes exhibited by the two mouse models is vastly different. In fact the ‘aging-like’ phenotypes in the p53+/m mouse model begin to appear only after all the p44 transgenic mice are dead and gone. The comparison of the tumor resistance between the two mouse models is difficult for the following reasons: (i) the vast differences in the median lifespan of the two models (96 weeks in p53+/m mice vs. 32 weeks in p44 transgenic mice), (ii) the absence of quantitative data on the tumor resistance of p44 mice, and (iii) the premature deaths that start as early as 5 weeks in the transgenic lines. Furthermore, it is important to note that the p44 transgenic mice do harbor two copies of the p53 gene, a fact that makes the comparisons more problematic. The numerous differences between the two mouse models are tabulated. In the next section, Vijg and Hasty compare the mutant mouse models for BRCA1 and Ku80 proteins (Vogel et al., 1999; Cao et al., 2003), Terc–/– compound mutants (Wong et al., 2003; Chang et al., 2004) and a future study (Vijg & Hasty, 2006). They also misinterpret the meaning of our sentence about the problems with comparing such models with the p53+/m mouse model. The space constraints of modern-day scientific publishing put limitations on discussion in our original manuscript and this could have contributed to such a misunderstanding. The message we wanted to convey is once again related to the ‘24-gene deletion’ problem. How can one compare compound double mutants to a mouse model that has 24 genes deleted? Furthermore, at the mechanistic level, the chronic activation of p53 shown for the BRCA1 mutants and proposed for other models is not true for the p53+/m mouse model (Fig. 1). Vijg and Hasty refer to the alleviation of premature aging symptoms in the compound mutants at the level of replicative senescence. The extrapolation of replicative senescence phenotypes (in vitro) at the cellular level to in vivo aging is also debatable and the problems with such comparisons have been raised earlier by others (Rubin, 2002; Wright & Shay, 2002). Genotype-phenotype comparisons of the p53+/m mouse model with other mutant mouse models that show chronic activation of p53. Besides this immediate concern that such comparisons have no mechanistic basis (at the level of p53 activation), their argument brings up a more important issue that relates to the utility of such mutant models in the study of aging and the liberal use of the term ‘aging’ while describing such mutants. Are the ‘aging-like’ phenotypes exhibited by such mutant mouse models authentic markers of physiological aging? Are we going to focus on every sick and hunchbacked mouse model and consider it a model for mammalian aging? It is very likely that the phenotypes exhibited by such models may simply be a manifestation of the chronic activation of p53 (or cell cycle inhibitory proteins) and their effects on specific tissues that have a higher turnover (e.g. osteoblasts in bone remodeling). It is also important to note that the process of aging is much more complex than the ‘across-the-board genetic tweaking’ that occurs in genetically modified mice. As alerted earlier by experts in the field, it is time the scientific community takes a cautionary note about the wholesale categorization of such mutants as ‘aging’ models (Lombard et al., 2005; Miller, 2004). In the final part of the discourse, Vijg and Hasty have a legitimate question about the evidence for the role of the 23 genes in the phenotypes exhibited by the p53+/m mouse model. The effect of 23 genes on the phenotypes exhibited by the p53+/m mouse model and relating the phenotypes to specific genotypes will require time and experimentation. Furthermore, the complexity of the deleted region indicates that a single candidate may not be responsible for all the phenotypes. At this juncture, I would like to reiterate that the purpose of our publication was to further clarify the genetic deletion and formulate an alternate hypothesis that draws upon the known functions of the candidate proteins present in the deletion. In addition, the caveat of cancer resistance combined with lifespan reduction presented by the p53+/m mouse model (if one is to compare the longevity data of the p53+/m mice only to the wild-type controls and not that of the p53+/null animals) and the weak molecular data that accompanied our original unproven hypothesis provided the impetus for our attempts to characterize the genetic deletion. Given this lack of molecular or biochemical evidence for the role of any of the deleted genes (including p53) for the observed phenotypes, it is prudent that we do not come to premature conclusions. Vijg and Hasty also present a model (lacking experimental evidence) based on the hypothetical role of the m-protein in activation of p53. The constitutive activation of p53 that the commentators refer in the p53+/m mouse is still a hypothetical possibility. The model also draws partial and inappropriate comparisons with mouse models that are genetically and mechanistically different. The importance of the p53 tumor suppressor in a variety of cellular responses to DNA damage cannot be overstated. However extrapolating the role of p53 in organismal aging and basing it on hypothetical models that lack experimental evidence is risky. Questioning the role of an important tumor suppressor protein in aging and not discounting the possible role of 23 other proteins may not be popular, but this author believes that scientific research is not a game of poker but an endeavor that derives its strength from sound experimentation and an unrelenting quest for truth.

Model organism databases are in jeopardy.

The first Mouse Models for Human Cancer Consortium (MMHCC) meeting for human nervous system cancer convened in New York City in November of 2000, with the proceedings of that meeting subsequently reported in Oncogene (1).The emphasis of the first meeting was on the comparative pathology of mouse models and corresponding human tumors. Recommendations from the meeting included the need to recapitulate human CNS tumor gene alterations in mouse model tumors, a need for increased emphasis on the molecular characterization of mouse model tumors for establishing consistency with corresponding human tumors, and the need to utilize mouse models in the preclinical evaluation of new therapies for treating CNS cancer. As indicated in the report ofthe 5th MMHCC meeting for nervous system cancer (Montreal, November 2010) in the current issue of Neuro-Oncology (2), these recommendations have proven influential, as most mouse models for CNS cancer are now 1) based on gene alterations suspected ofdriving corresponding human tumor development; 2) routinely subjected to extensive molecular characterization; and 3) are experiencingincreased use for therapeutic hypothesis testing.In addition, the sophistication of current models has increased substantially, with many utilizing information regarding potential cell of origin to regulate the expression or inactivation of genes in precursor cells thought to give rise to corresponding human tumors. Perhaps unforeseen at the 2000 MMHCC meeting was the revitalization of xenograft models for studying CNS cancer, which has been driven by the growing appreciation that histopathologically defined classes of CNS cancer, such as glioblastoma, consist of multiple distinct tumor subtypes, many of which can be propagated as xenografts, and the understanding that individual tumors are composed of distinct subpopulations of cells with distinct biologic properties.As a result of such understanding, human tumor xenografts, established either by direct transplantation from surgical specimens or by transplantation from primary cultures of surgical specimens, are viewed as model systems in their own right. Currently available genetically engineered mouse(GEM) models and human tumor xenografts now allow for detailed investigation of tumor initiation, progression, and response to therapy and have provided investigators with a resource armamentarium for the in-depthstudy of the molecular, cellular, and tumor biology of CNS cancer.It will be of interest to follow the progress during the next 10 years, as GEM modelers will undoubtedly utilize the rapidly growing base of information regarding developmental regulation of gene expression to further refine models for appropriate temporal, spatial, and cellular recapitulation of corresponding human tumor development.

Pathogenicity of swine influenza viruses possessing an avian or swine-origin PB2 polymerase gene evaluated in mouse and pig models

Progress has been made in the field of neural interfacing using both mouse and rat models, yet standardization of these models’ interchangeability has yet to be established. The mouse model allows for transgenic, optogenetic, and advanced imaging modalities which can be used to examine the biological impact and failure mechanisms associated with the neural implant itself. The ability to directly compare electrophysiological data between mouse and rat models is crucial for the development and assessment of neural interfaces. The most obvious difference in the two rodent models is size, which raises concern for the role of device-induced tissue strain. Strain exerted on brain tissue by implanted microelectrode arrays is hypothesized to affect long-term recording performance. Therefore, understanding any potential differences in tissue strain caused by differences in the implant to tissue size ratio is crucial for validating the interchangeability of rat and mouse models. Hence, this study is aimed at investigating the electrophysiological variances and predictive device-induced tissue strain. Rat and mouse electrophysiological recordings were collected from implanted animals for eight weeks. A finite element model was utilized to assess the tissue strain from implanted intracortical microelectrodes, taking into account the differences in the depth within the cortex, implantation depth, and electrode geometry between the two models. The rat model demonstrated a larger percentage of channels recording single unit activity and number of units recorded per channel at acute but not chronic time points, relative to the mouse model Additionally, the finite element models also revealed no predictive differences in tissue strain between the two rodent models. Collectively our results show that these two models are comparable after taking into consideration some recommendations to maintain uniform conditions for future studies where direct comparisons of electrophysiological and tissue strain data between the two animal models will be required.

The anti-inflammatory effect of miR-155 was closely linked to transforming growth factor-β-activated kinase-1-binding protein 2 (TAB2) and autophagy. This study investigated the role of miR-155 in attenuation of septic lung injury through TAB2 and autophagy in mouse model and in vitro. Patients who underwent fiberoptic bronchoscope examination with or without septic lung injury were recruited for the collection of bronchoalveolar lavage fluid (BALF) samples. Mouse model of septic lung injury was established by cecal ligation puncture, while alveolar macrophage cell line was treated with lipopolysaccharide (LPS). Agomir miR-155 transfection into the mouse airways was performed to induce miR-155 expression, while miR-155 mimic, miR-155 inhibitor, or TAB2-siRNA was transfected into NR8383 macrophages. Mouse BALF and cell cytokine levels, lung tissue pathology and wet/dry ratio, numbers of autophagy bodies, miR-155, gene and protein expressions were also examined accordingly. Expression of miR-155 was increased in the BALF of septic lung injury patients, in mouse model and NR8383 macrophages after LPS treatment. Increased numbers of autophagy bodies were also observed in mouse and macrophage models. MiR-155-transfected mice showed alleviation of inflammation, lower water content in lung tissues, increased number of autophagy bodies, increased expression of microtubule-associated protein 1 light chain 3 (LC3 II/I), reduced expressions of cysteinyl aspartate-specific protease-1 (Caspase-1), and TAB2, and decreased cytokines levels. Similar results were obtained in macrophages after LPS treatment. Cells transfected with miR-155 inhibitor showed increased expression of TAB2 and Caspase-1, fewer autophagy bodies, lower LC3 II/I expression, and higher cytokine levels. The current study observed a higher level of miR-155 in the BALF from sepsis patients with acute respiratory distress syndrome and demonstrated that miR-155 alleviated inflammation in septic lung injury in mouse and cell models by inducing autophagy via inhibition of TAB2.