Developing Crowdsourced Training Data Sets for Pharmacovigilance Intelligent Automation.

INTRODUCTION: Machine learning offers an alluring solution to developing automated approaches to the increasing individual case safety report burden being placed upon pharmacovigilance. Leveraging crowdsourcing to annotate unstructured data may provide accurate, efficient, and contemporaneous training data sets in support of machine learning. OBJECTIVE: The objective of this study was to evaluate whether crowdsourcing can be used to accurately and efficiently develop training data sets in support of pharmacovigilance automation. MATERIALS AND METHODS: Pharmacovigilance experts created a reference dataset by reviewing 15,490 de-identified social media posts of narratives pertaining to 15 drugs and 22 medically relevant topics. A random sampling of posts from the reference dataset was published on Amazon Turk and its users (Turkers) were asked a series of questions about those same medical concepts. Accuracy, price elasticity, and time efficiency were evaluated. RESULTS: Accuracy of crowdsourced curation exceeded 90% when compared to the reference dataset and was completed in about 5% of the time. There was an increase in time efficiency with higher pay, but there was no significant difference in accuracy. Additionally, having a social media post reviewed by more than one Turker (using a voting system) did not offer significant improvements in terms of accuracy. CONCLUSIONS: Crowdsourcing is an accurate and efficient method that can be used to develop training data sets in support of pharmacovigilance automation. More research is needed to better understand the breadth and depth of possible uses as well as strengths, limitations, and generalizability of results.

Exercise preconditioning improves traumatic brain injury outcomes.

PURPOSE: To determine whether 6 weeks of exercise performed prior to traumatic brain injury (TBI) could improve post-TBI behavioral outcomes in mice, and if exercise increases neuroprotective molecules (vascular endothelial growth factor-A [VEGF-A], erythropoietin [EPO], and heme oxygenase-1 [HO-1]) in brain regions responsible for movement (sensorimotor cortex) and memory (hippocampus). METHODS: 120 mice were randomly assigned to one of four groups: (1) no exercise+no TBI (NOEX-NOTBI [n=30]), (2) no exercise+TBI (NOEX-TBI [n=30]), (3) exercise+no TBI (EX-NOTBI [n=30]), and (4) exercise+TBI (EX-TBI [n=30]). The gridwalk task and radial arm water maze were used to evaluate sensorimotor and cognitive function, respectively. Quantitative real time polymerase chain reaction and immunostaining were performed to investigate VEGF-A, EPO, and HO-1 mRNA and protein expression in the right cerebral cortex and ipsilateral hippocampus. RESULTS: EX-TBI mice displayed reduced post-TBI sensorimotor and cognitive deficits when compared to NOEX-TBI mice. EX-NOTBI and EX-TBI mice showed elevated VEGF-A and EPO mRNA in the cortex and hippocampus, and increased VEGF-A and EPO staining of sensorimotor cortex neurons 1 day post-TBI and/or post-exercise. EX-TBI mice also exhibited increased VEGF-A staining of hippocampal neurons 1 day post-TBI/post-exercise. NOEX-TBI mice demonstrated increased HO-1 mRNA in the cortex (3 days post-TBI) and hippocampus (3 and 7 days post-TBI), but HO-1 was not increased in mice that exercised. CONCLUSIONS: Improved TBI outcomes following exercise preconditioning are associated with increased expression of specific neuroprotective genes and proteins (VEGF-A and EPO, but not HO-1) in the brain.

Behavioral effects and mechanisms of migraine pathogenesis following estradiol exposure in a multibehavioral model of migraine in rat.

Migraine is one of the most common neurological disorders, leading to more than 1% of total disability reported and over 68 million visits to emergency rooms or physician's offices each year in the United States. Three times as many women as men have migraine, and while the mechanism behind this is not well understood, 17ß-estradiol (estradiol) has been implicated to play a role. Studies have demonstrated that exposure to estrogen can lead to activation of inflammatory pathways, changes in sodium gated channel activity, as well as enhanced vasodilation and allodynia. Estradiol receptors are found in trigeminal nociceptors, which are involved in signaling during a migraine attack. The purpose of this study was to investigate the role of estradiol in migraine pathogenesis utilizing a multibehavioral model of migraine in rat. Animals were surgically implanted with a cannula system to induce migraine and behavior was assessed following exposure to a proestrus level of estradiol for total locomotor activity, light and noise sensitivity, evoked grooming patterns, and enhanced acoustic startle response. Results demonstrated decreased locomotor activity, increased light and noise sensitivity, altered facial grooming indicative of allodynia and enhanced acoustic startle. Further examination of tissue samples revealed increased expression of genes associated with inflammation and vasodilation. Overall, this study demonstrates exacerbation of migraine-like behaviors following exposure to estradiol and helps further explain the underlying mechanisms behind sex differences found in this common neurological disorder.

Differential response of miRNA-21 and its targets after traumatic brain injury in aging mice.

The present study investigated the possible role of miR-21, a miRNA that has known prosurvival function, in poor outcomes in the elderly following traumatic brain injury compared to adults. Controlled cortical impact injury was induced in adult (5-6 months) and aged (22-24 months) C57/BL6 mice. miR-21 and four of its targets (PDCD4, TIMP3, RECK, PTEN) were analyzed at 1, 3, 7 days post injury in samples of injured cortex using real-time PCR analysis. Basal miR-21 expression was higher in the aged brain than in the adult brain. In the adult brain, miR-21 expression increased in response to injury, with the maximum increase 24 hours after injury followed by a gradual decrease, returning to baseline 7 days post-injury. In contrast, in aged mice, miR21 showed no injury response, and expression of miR-21 target genes (PTEN, PDCD4, RECK, TIMP3) was up-regulated at all post injury time points, with a maximal increase at 24 hours post injury. Based on these results, we conclude that the diminished miR21 injury response in the aged brain leads to up-regulation of its targets, with the potential to contribute to the poor prognosis following TBI in aging brain. Therefore, strategies aimed at up-regulation of miR-21 and/or down regulation of its targets might be useful in improving outcomes in the elderly following TBI.

Neuroglobin overexpression improves sensorimotor outcomes in a mouse model of traumatic brain injury.

There is a significant need for novel treatments that will improve traumatic brain injury (TBI) outcomes. One potential neuroprotective mechanism is to increase oxygen binding proteins such as neuroglobin. Neuroglobin has a high affinity for oxygen, is an effective free radical scavenger, and is neuroprotective within the brain following hypoxia and ischemia. The purpose of this study was to determine whether neuroglobin overexpression improves sensorimotor outcomes following TBI in transgenic neuroglobin overexpressing (NGB) mice. Additional study aims were to determine if and when an endogenous neuroglobin response occurred following TBI in wild-type (WT) mice, and in what brain regions and cell types the response occurred. Controlled cortical impact (CCI) was performed in adult (5 month) C57/BL6 WT mice, and NGB mice constitutively overexpressing neuroglobin via the chicken beta actin promoter coupled with the cytomegalovirus distal enhancer. The gridwalk task was used for sensorimotor testing of both WT and NGB mice, prior to injury, and at 2, 3, and 7 days post-TBI. NGB mice displayed significant reductions in the average number of foot faults per minute walking at 2, 3, and 7 days post-TBI when compared to WT mice at each time point. Neuroglobin mRNA expression was assessed in the injured cortex of WT mice prior to injury, and at 1, 3, 7, and 14 days post-TBI using quantitative real time polymerase chain reaction (qRT-PCR). Neuroglobin mRNA was significantly increased at 7 days post-TBI. Immunostaining showed neuroglobin primarily localized to neurons and glial cells in the injured cortex and ipsilateral hippocampus of WT mice, while neuroglobin was present in all brain regions of NGB mice at 7 days post-TBI. These results showed that overexpression of neuroglobin reduced sensorimotor deficits following TBI, and that an endogenous increase in neuroglobin expression occurs during the subacute period. Increasing neuroglobin expression through novel therapeutic interventions during the acute period after TBI may improve recovery.

Exposure to bisphenol A exacerbates migraine-like behaviors in a multibehavior model of rat migraine.

Migraine is a common and debilitating neurological disorder suffered worldwide. Women experience this condition 3 times more frequently than men, with estrogen strongly implicated to play a role. Bisphenol A (BPA), a highly prevalent xenoestrogen, is known to have estrogenic activity and may have an effect in migraine onset, intensity, and duration through estrogen receptor signaling. It was hypothesized that BPA exposure exacerbates migraine symptoms through estrogen signaling and downstream activation of nociception related pathways. Utilizing a multibehavior model of migraine in ovariectomized female rats, changes in locomotion, light and sound sensitivity, grooming, and acoustic startle were examined. Furthermore, changes in the expression of genes related to estrogen (ERα, GPR30), and nociception (extracellular signal regulated kinase, ERK, sodium gated channel, Nav1.8, and fatty acid amide hydrolase, FAAH) were studied following behavioral experiments. The following results were obtained: BPA treatment significantly exacerbated migraine-like behaviors in rats. Rats exposed to BPA demonstrated decreased locomotion, exacerbated light and sound aversion, altered grooming habits, and enhanced startle reflexes. Furthermore, BPA exposure increased mRNA expression of estrogen receptors, total ERK mRNA and ERK activation, as well as Nav1.8, and FAAH mRNA, indicative of altered estrogen signaling and altered nociception. These results show that BPA, an environmentally pervasive xenoestrogen, exacerbates migraine-like behavior in a rat model and alters expression of estrogen and nociception-related genes.

Effects of aging on blood brain barrier and matrix metalloproteases following controlled cortical impact in mice.

Aging alters the ability of the brain to respond to injury. One of the major differences between the adult and aged brain is that comparable injuries lead to greater blood brain barrier disruption in the aged brain. The goals of these studies were to quantify the effects of age on BBB permeability using high field strength MRI T1 mapping and to determine whether activation of matrix metalloproteases, their inhibitors, or expression of blood brain barrier structural proteins, occludin, zonnula occludins-1 (ZO-1) and claudin-5 were altered following injury to the aged C57/BL6 mouse brain. T1 mapping studies revealed greater blood brain barrier permeability in the aged (21-24 months old) brain than in the adult (4-6 months old) following controlled cortical impact. The increased blood brain barrier permeability in the pericontusional region was confirmed with IgG immunohistochemistry. MMP-9 activity was increased following controlled cortical impact in the aged brain, and this was accompanied by increased MMP-9 gene expression. MMP-2 activity was higher in the uninjured aged brain than in the adult brain. Occludin and ZO-1 mRNA levels were unchanged following injury in either age group, but claudin-5 mRNA levels were lower in the aged than the adult brain following injury. These results demonstrate quantitative increases in blood brain barrier permeability in the aged brain following injury that are accompanied by increased MMP-9 activation and decreased blood brain barrier repair responses.

Sex differences in behavior and expression of CGRP-related genes in a rodent model of chronic migraine.

OBJECTIVE: The objectives of this study were to develop a preclinical rodent model that produces migraine-like behaviors based on International Headache Society diagnostic criteria, to determine whether sex differences are present, and to determine whether expression of calcitonin gene-related peptide (CGRP) and the genes encoding its receptor in trigeminal ganglion or medulla correlates with those behaviors. BACKGROUND: Few animal studies of migraine have tested behaviors associated with migraine diagnostic criteria. In this study, changes in activity and in mechanical sensitivity of facial regions following application of inflammatory soup (IS) or vehicle (phosphate-buffered saline [PBS]) to the dura were measured to model changes in routine activity and allodynia. CGRP, an important mediator of migraine pathogenesis, and the 3 components of its receptor, calcitonin-like receptor (CLR), receptor activity-modifying protein 1 (RAMP1), and receptor component protein (RCP) mRNAs were quantified in the trigeminal ganglion and medulla to identify baseline sex differences and changes associated with application of IS or PBS to the dura. METHODS: Male and female Sprague-Dawley rats were implanted with a dural cannula. Groups of rats were treated with 10 or 20 µL volumes of IS or PBS. Baseline behavioral testing was conducted prior to surgery and again at 7 days postsurgery, and dural application of IS or PBS was performed repeatedly for a total of 8 applications. Locomotor activity was assessed using force plate actimetry during and following application to provide information on distance traveled, bouts of low mobility, spatial confinement, and focused energy. Periorbital and perimasseter sensory testing was performed 20 minutes post-application to measure allodynia. The rats were sacrificed 30 minutes following the final dural treatment, tissue was dissected and total RNAs were isolated from ipsilateral trigeminal ganglia and ipsilateral medulla. Quantitative real-time polymerase chain reactions were used to measure the expression of amplified constructs using gene-specific primers for CGRP, RAMP1, CLR, and RCP. RESULTS: Both males and females showed behavioral effects of IS application, but there were pronounced sex differences. Females showed effects at the lower dose, and activity changes were present for a longer duration, but males required fewer applications of IS to exhibit behavioral changes. Females showed increased withdrawal responses for periorbital and perimasseter mechanical testing (10 µL IS groups), and males showed increased perimasseter withdrawal responses (20 µL IS group). In the trigeminal ganglion, there were no baseline sex differences in CGRP-encoding mRNA, but females had lower baseline expression of RAMP1, CLR, and RCP-encoding mRNAs. In the medulla, females had higher baseline levels of CGRP-encoding mRNAs and lower baseline levels of RAMP1, CLR, and RCP-encoding mRNAs than males. Both IS and PBS increased expression of mRNAs encoding CGRP, RAMP1, RCP, and CLR in the trigeminal ganglion in males, but in females, only CLR and RCP were increased. In the medulla both IS and PBS increased expression of CGRP, CLR in males and CLR and RCP in females. Thus, expression of CGRP-related genes did not mirror the behavioral differences between IS and PBS groups. Instead, CGRP-related genes were upregulated by both IS and PBS applications. CONCLUSIONS: This study demonstrates significant changes in locomotor activity and facial allodynia associated with application of IS to the dura as well as significant sex differences, demonstrating that International Headache Society diagnostic criteria can be used to design a rodent behavioral model of migraine. In addition, there were prominent baseline sex differences in expression of CGRP and its receptor in both the trigeminal ganglion and medulla, but the majority of changes in expression of CGRP and its receptor were present in both the IS and PBS treated rats. This suggests that the CGRP pathway responds to changes in intracranial pressure or meningeal stretch, while migraine-like behaviors occur after meningeal inflammation.

Upregulation of inflammatory mediators in a model of chronic pain after spinal cord injury.

Chronic neuropathic pain is a disabling condition observed in large number of individuals following spinal cord injury (SCI). Recent progress points to an important role of neuroinflammation in the pathogenesis of central neuropathic pain. The focus of the present study is to investigate the role of proinflammatory molecules IL-1ß, TNF-α, MCP-1, MMP-9 and TIMP-1 in chronic neuropathic pain in a rodent model of SCI. Rats were subjected to spinal cord contusion using a controlled linear motor device with an injury epicenter at T10. The SCI rats had severe impairment in locomotor function at 7 days post-injury as assessed by the BBB score. The locomotor scores showed significant improvement starting at day 14 and thereafter showed no further improvement. The Hargreaves' test was used to assess thermal hyperalgesia for hindpaw, forepaw and tail. A significant reduction in withdrawal latency was observed for forepaw and tail of SCI rats at days 21 and 28, indicating the appearance of thermal hyperalgesia. Changes in expression of mRNAs for IL-1ß, TNF-α, MCP-1, MMP-9 and TIMP-1 were assessed using real-time polymerase chain reaction in spinal cord including the injury epicenter along with regions above and below the level of lesion at day 28 post-injury. A significant increase was observed in the expression of MCP-1, TNF-α, TIMP-1 and IL-1ß in the injury epicenter, whereas only TIMP-1 was upregulated in the area below the injury epicenter. The results of the study suggest that prolonged upregulation of inflammatory mediators might be involved in chronic neuropathic pain in SCI, and that TIMP-1 may play a role in maintenance of chronic below level pain.

Impact of the SpeB protease on binding of the complement regulatory proteins factor H and factor H-like protein 1 by Streptococcus pyogenes.

Microbial pathogens often exploit human complement regulatory proteins such as factor H (FH) and factor H-like protein 1 (FHL-1) for immune evasion. Fba is an FH and FHL-1 binding protein expressed on the surface of the human pathogenic bacterium Streptococcus pyogenes, a common agent of pharyngeal, skin, and soft-tissue infections. Fba has been shown to contribute to phagocytosis resistance, intracellular invasion, and virulence in mice. Here, we look at the role of Fba in recruitment of FH and FHL-1 by five serotype M1 isolates of streptococci. Inactivation of fba greatly inhibited binding of FH and FHL-1 by all isolates, indicating that Fba is a major FH and FHL-1 binding factor of serotype M1 streptococci. For three isolates, FH binding was significantly reduced in stationary-phase cultures and correlated with high levels of protease activity and SpeB (an extracellular cysteine protease) protein in culture supernatants. Analysis of a speB mutant confirmed that SpeB accounts for the loss of Fba from the cell surface, suggesting that the protease may modulate FH and FHL-1 recruitment during infection. Comparisons of fba DNA sequences revealed that the FH and FHL-1 binding site in Fba is conserved among the M1 isolates. Although the ligand binding site is not strictly conserved in Fba from a serotype M49 isolate, the M49 Fba protein was found to bind both FH and FHL-1. Collectively, these data indicate that binding of FH and FHL-1 is a conserved function of Fba while modulation of Fba function by SpeB is variable.

Acquisition of regulators of complement activation by Streptococcus pyogenes serotype M1.

Opsonization of bacteria by complement proteins is an important component of the immune response. The pathogenic bacterium Streptococcus pyogenes has evolved multiple mechanisms for the evasion of complement-mediated opsonization. One mechanism involves the binding of human regulators of complement activation such as factor H (FH) and FH-like protein 1 (FHL-1). Acquisition of these regulatory proteins can limit deposition of the opsonin C3b on bacteria, thus decreasing the pathogen's susceptibility to phagocytosis. Binding of complement regulatory proteins by S. pyogenes has previously been attributed to the streptococcal M and M-like proteins. Here, we report that the S. pyogenes cell surface protein Fba can mediate binding of FH and FHL-1. We constructed mutant derivatives of S. pyogenes that lack Fba, M1 protein, or both proteins and assayed the strains for FH binding, susceptibility to phagocytosis, and C3 deposition. Fba expression was found to be sufficient for binding of purified FH as well as for binding of FH and FHL-1 from human plasma. Plasma adsorption experiments also revealed that M1(+) Fba(+) streptococci preferentially bind FHL-1, whereas M1(-) Fba(+) streptococci have similar affinities for FH and FHL-1. Fba was found to contribute to the survival of streptococci incubated with human blood and to inhibit C3 deposition on bacterial cells. Streptococci harvested from log-phase cultures readily bound FH, but binding was greatly reduced for bacteria obtained from stationary-phase cultures. Bacteria cultured in the presence of the protease inhibitor E64 maintained FH binding activity in stationary phase, suggesting that Fba is removed from the cell surface via proteolysis. Western analyses confirmed that E64 stabilizes cell surface expression of Fba. These data indicate that Fba is an antiopsonic, antiphagocytic protein that may be regulated by cell surface proteolysis.