Smad2 and Smad3 (Smad2/3) proteins are key signaling molecules for TGF-β and some related family members regulating the transcription of several hundred genes. TGF-β have key roles in development, tissue homeostasis, and the pathogenesis of many human diseases, including cancer, fibrotic disorders, developmental defects, and neurodegeneration. To study the temporal and spatial patterns of Smad2/3-dependent signaling in normal and pathological conditions in the living organism, we engineered transgenic mice with a Smad-responsive luciferase reporter construct (SBE-luc mice). Using bioluminescent imaging, we assessed Smad2/3 signaling activity noninvasively in living mice. At baseline, this activity was highest in brain, intestine, heart, and skin, and correlated with biochemical measurements of reporter activity. Primary astrocytes cultured from SBE-luc mice showed specific activation of the reporter in response to Smad2/3-activating TGF-β family members. Treatment of mice with the endotoxin LPS resulted in a fast and vigorous, but transient activation of the reporter in the intestine. Although the response was similarly rapid in brain, it remained increased, indicating important but different cellular responses to endotoxin challenge in these organs. Traumatic brain injury with a needle stab resulted in local activation of Smad2/3-dependent genes and a severalfold increase in bioluminescence in living mice. SBE-luc mice can therefore be used to study temporal, tissue-specific activation of Smad2/3-dependent signaling in living mice in normal or pathological conditions as well as for the identification of endogenous or synthetic modulators of this pathway.
The TGF-β superfamily of proteins, including TGF-β, activins, and bone morphogenic proteins (BMPs)4 control cellular processes ranging from patterning and differentiation to proliferation and apoptosis, and have been implicated in tumorigenesis, fibrosis, inflammation, and neurodegeneration (1, 2). The three different isoforms of TGF-β, TGF-β1, -2, and -3 have key roles in injury responses and wound repair in various tissues (3). They activate the TGF-β signaling pathway through a high-affinity transmembrane receptor complex consisting of the TGF-β-type I (activin receptor-like kinase (ALK) 5) and -type II serine/threonine kinase receptor subunits (2, 4). TGF-β binding leads to phosphorylation of ALK5 and recruitment and phosphorylation of receptor-regulated Smad2 or Smad3. Once phosphorylated, these Smads form heterodimeric or heterotrimeric complexes with Smad4 and translocate into the nucleus to regulate gene transcription (4) (see Fig. 1 A). Growth and differentiation factor (GDF) 8/myostatin and GDF-9 can also activate Smad2/3 after binding to ALK5 in combination with activin and BMP-type II receptors, respectively (5, 6). Furthermore, activin and nodal activate Smad2/3 by engaging ALK4 and ALK7 receptors (4, 7, 8). In contrast, BMPs recruit Smad1, Smad5, and Smad8 in combination with Smad4 after binding to BMP-type I and -type II receptors (2, 4). Signaling of TGF-β and related factors is also regulated by cross-talk with other signaling pathways including MAPK, JAK/STAT, and wnt pathways, and TGF-β can activate MAPK and JNK pathways independent of Smads (2, 4).
Global gene expression analysis of cells or tissues has recently generated a more comprehensive understanding of TGF-β and other signaling pathways and gene responses to injury (9, 10). Thus, Smad2/3 may regulate expression of several hundred immediate-early and immediate genes (9). Other studies determined the distribution of mRNAs for TGF-β family members (11). Expression patterns of Smad proteins have also been characterized in mouse embryos (12) or in adult mice (13). Transgenic mice harboring a Smad1/5-responsive reporter gene were used to study signaling by BMPs in embryos (14). Probably the best characterization of spatial and temporal activation of embryonic Smad2/3-dependent TGF-β signaling in vivo comes from an analysis of phosphorylated Smad2 expression in mice (15). Together, these studies underline the importance of TGF-β and related family members in patterning, migration, differentiation, and apoptosis during development and in adulthood. However, they do not directly test whether Smad2/3-dependent signaling results in transcriptional responses and none can be used to follow TGF-β signaling within a single mouse over time. It has also not been investigated how Smad2/3-dependent transcription changes over time in response to injury or endotoxin challenge. Monitoring of signaling pathways in individual animals over time could greatly enhance our understanding of how specific pathways contribute to pathogenic mechanisms and enable analyses of therapeutic interventions that target these pathways. This is particularly relevant in the study of complex diseases where high interindividual variability necessitates the use of large groups of animals and often obscures important molecular interactions or therapeutic effects of drugs. We have engineered transgenic reporter mice that express luciferase in response to activation of Smad2/3. After injecting a luciferase substrate, we can use bioluminescence imaging to obtain and follow optical signatures in a spatial and temporal manner in living mice. We demonstrate here how this novel and versatile tool can be used to detect TGF-β pathway activation after LPS injection and brain injury. The complexity of the TGF-β response is revealed by differing patterns of activation in different organs.
Materials and Methods
Transgenic mice were generated with the SBE-luc plasmid (Fig. 1 B and Ref. 16) on a FVB/N or (SJL/J × C57BL/6J)F1 genetic background using standard procedures. Transgenic lines T9-7, T9-55, T9-90, and T9-98 were selected, bred, and backcrossed for at least three to four generations with C57BL/6J or FVB/N mice. In lines T9-55 and T9-98 between 30 and 45% of the transgene-positive mice did not show an induction of the transgene in response to injury, although primary astrocytes from all transgene-positive mice responded similarly to stimulation by TGF-β1. Such “nonresponder” mice had no basal reporter gene activity in their brains (luciferase activity was similar to nontransgenic mice) and could be identified by a lack of luciferase activity in tail biopsies. In contrast, “responder” mice had luciferase activities that were 20- to 50-fold above background and clearly detectable luciferase activity in tail biopsies. Therefore, all mice used in this study were screened for tail luciferase activity and only responder mice were used. The reason for this lack of reporter gene activity is unclear but was not a result of multiple integration sites of the transgene. It is possible that epigenetic factors contributing to differential DNA methylation may inactivate the reporter in some mice, but this has not been tested so far. All animal handling was performed in accordance with institutional guidelines and approved by the local Institutional Animal Care and Use Committee.
Primary astrocytes were obtained from brains of 1- to 3-day-old postnatal SBE-luc transgenic mice. Ten-day-old primary astrocytes were used for experiments and stimulated with recombinant cytokines (all from R&D Systems) in serum-free medium. Luciferase activity was measured 16 h later using a commercial luciferase assay kit (Promega) and a tube luminometer. F11 reporter cells were generated by stable transfection of a mouse fibroblast cell line obtained from TGF-β1 knockout mice (a kind gift from Dr. J. Munger, New York University, New York, NY) with a Smad-binding element (SBE)-secreted alkaline phosphatase (SEAP) reporter gene (SBE-SEAP, Fig. 1 B) and a CMV-hygromycin-resistant gene. B103 rat neuroblastoma cells have been described previously (17). Cells were cultured in DMEM containing 2 mM glutamine, 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin and transiently transfected with the following plasmids (all 0.3 μg/well except where indicated): CMV-TβRIIΔk, CMV-Flag-Smad2, CMV-Flag-Smad3, CMV-Flag-Smad4 (0.1 μg/well; generously provided by Dr. R. Derynck, University of California, San Francisco, CA), CMV-Flag-Smad1, and CMV-Flag-Smad5 (18) using Superfect (Qiagen). SEAP activity in the culture supernatant was measured using Great EscAPe SEAP Reporter System 2 kits (Clontech Laboratories). Similarly, B103 cells were transfected with various Flag-Smad expression vectors. In addition, these cells were also cotransfected with SBE-luc reporter plasmid (0.1 μg/well) and CMV-TβRIIΔk (0.3 μg/well) to suppress autocrine/paracrine activation of the reporter gene by endogenous TGF-β. All transfections were verified by Western blotting.
Measurement of TGF-β1 and plasminogen activator inhibitor 1 (PAI-1) mRNA
Total RNA was isolated from hemibrains with TRIzol reagent (Molecular Research Center). RNA was analyzed by solution hybridization RNase protection assay with antisense riboprobes complementary to TGF-β1, PAI-1, and actin mRNA as described previously (19).
Brain injury and LPS administration
Acute brain injury was induced in mice temporarily anesthetized with isofluorane (Baxter) by placing two stab wounds (∼2 mm in depth) into the temporal-parietal region of one cortical hemisphere using a 26.5-gauge needle. Unlesioned transgenic mice were included as controls. Acute inflammation was induced by i.p. or i.v. injection of LPS (Escherichia coli serotype 055:B5, 2 mg/kg; Sigma-Aldrich).
Mice were anesthetized with 400 mg/kg chloral hydrate (Sigma-Aldrich) and perfused transcardially with 0.9% saline. Tissues were dissected, weighed, and lysed in 100 to 400 μl of cell culture lysis reagent (Promega). Luciferase activities from tissue homogenates were measured and normalized to weight. To detect bioluminescence in individual isolated organs, mice were anesthetized and luciferin was administered (see below). Five minutes later, mice were sacrificed by cervical dislocation and organs were rapidly dissected, placed in 24-well culture dishes, and imaged exactly 10 min after the initial luciferin administration.
Optical detection of bioluminescence
Bioluminescence was detected with the In Vivo Imaging System (IVIS; Xenogen), consisting of a cooled charge-coupled device camera (Roper camera, 1300 × 1300 pixels, 20 μm2/pixel, cooled to −105°C) mounted on a light-tight specimen chamber, an anesthesia unit, and a Windows computer system. Mice were injected i.p. with 150 mg/kg d-luciferin (Xenogen) 10 min before imaging and anesthetized with isofluorane during imaging. Photons emitted from living mice or dissected organs were acquired as photons per second per square centimeter per steradian (sr) and integrated for 5 min. Images were then converted into pseudo-colored representations of light intensity (blue least intense and red most intense). Gray scale images were obtained before bioluminescence imaging for reference. Gray scale photographic images and bioluminescence color images were superimposed using LIVINGIMAGE version 2.11 software overlay (Xenogen). For photon quantitation, a region of interest was manually selected over the signal intensity. The area of interest was kept constant within experiments and the intensity was converted into photons per second per square millimeter per sr.
Statistical analyses were performed with Prism 4.03 software (GraphPad). Means between two groups were compared with a two-tailed, unpaired Student’s t test; comparisons of means from multiple groups with one control were analyzed with one-way ANOVA and Dunnett’s post hoc test.
Independent lines of reporter mice for TGF-β signaling pathway show strongest basal activity in brain and intestine
To assess the utility of the SBE, we first tested a previously described synthetic reporter gene containing 12 SBE repeats fused to luciferase (SBE-luc; Fig. 1,B and Ref. 16) in transiently transfected B103 neuroblastoma cells. This Smad-binding element (SBE) which consists of one or more CAGA repeats (or GTCT) mediates specific responses to Smad2/3-dependent signaling in an estimated 500 target genes in mammals (4, 9). Smad3/4 or Smad2/3/4 resulted in a close to 100-fold increase in reporter activity but even Smad3 alone or Smad2/4 induced moderate responses (Fig. 1,C). Mouse MFB-F11 fibroblast cells derived from tgfb1−/− mice and stably transfected with the same promotor elements fused to SEAP (Fig. 1,B) were also activated by combinations of Smad2/3/4 but not by Smad1/4/5 (Fig. 1 D), consistent with previous reports (1, 2).
To study activation of the TGF-β signaling pathway in response to physiological and environmental stimuli in the intact organism, we engineered transgenic reporter mice with the same synthetic luciferase reporter gene (SBE-luc) used in cell culture above. Four independent transgenic reporter mouse lines harboring SBE-luc were generated and all responded to brain injury with an increase in luciferase expression (data not shown). Two lines that showed the highest fold induction in response to injury were selected and analyzed in more detail. Tissue analysis of transgene expression at basal level in 2-mo-old SBE-luc mice, in lines T9-55F, T9-7F (Fig. 2 A, on the FVB/N genetic background), and T9-55B (data not shown; same as T9-55F but on C57BL/6 genetic background), demonstrated the highest luciferase activity in the brain, intestine, heart, and skin. Although TGF-β signaling has been described in these organs, these data illustrate for the first time a direct comparison of Smad2/3-dependent gene activity in adult tissues. The prominent reporter activity in the brain is probably most surprising and points to an important role of Smad2/3-dependent signaling in the normal brain.
To confirm the specificity of the reporter gene for the TGF-β signaling pathway in SBE-luc mice, we cultured primary astrocytes from brains of transgenic mice and stimulated them with TGF-β1 or related molecules. TGF-β were the most potent activators of the reporter, whereas activin and nodal required 10-fold higher concentrations to induce similar increases in luciferase expression (Fig. 2 B). BMP-2, brain-derived neurotrophic factor, and epidermal growth factor failed to activate the reporter gene. As a control, the same dose of BMP-2 was able to activate a reporter gene consisting of BMP-responsive elements fused to luciferase (Ref. 20 and data not shown). These results are consistent with the reporter responding primarily to Smad2/3-dependent signaling and preferably to TGF-β.
Tissue-specific activation of Smad2/3-dependent signaling after endotoxin challenge
TGF-β1 has prominent roles in immune function and has both pro- and anti-inflammatory effects (21). To analyze the spatial activation pattern of TGF-β signaling in vivo in living mice after an inflammatory stimulus, we challenged SBE-luc mice with the endotoxin LPS. LPS is a potent inducer of inflammation that stimulates the production of various proinflammatory cytokines and other mediators. It also induces the production of TGF-β1, possibly to balance the action of proinflammatory factors and to induce endotoxin tolerance (22). TGF-β signaling and reporter gene expression were monitored noninvasively in living mice using in vivo bioluminescence imaging, which allows the detection of low-level light production from within living tissues of reporter mice (23, 24, 25). Intraperitoneal injection of luciferin into SBE-luc mice resulted in strong emission of photons from the intestinal region and the brain in living mice (Figs. 3 and 4). Using lower threshold settings, photon emission was also detected in the nose, lower jaw, paws, and tail in unmanipulated and in LPS-treated T9-55 mice (data not shown). Photon emission quantified in rapidly dissected organs using bioluminescence imaging software showed strongest activation again in intestine and brain (Fig. 3, A and B). In addition, photon emission from dissected organs correlated strongly with biochemical luciferase measurements in homogenized tissues across all organs in untreated control mice (r = 0.89, p < 0.0001) as well as in organs from mice 5 h after LPS treatment (r = 0.72, p < 0.0001). LPS treatment resulted in several hundred fold induction of bioluminescence in the intestine (Fig. 3, A, B, and D). Induction was also very prominent in the brain, and to a lesser extent in the heart and spleen (Fig. 3 B).
To identify the origin of intestinal photon emission, mesenteric lymph node, Peyer’s patches, and intestine without Peyer’s patches were dissected from mice treated with LPS, and luciferase activity in the tissue was measured. As shown in Fig. 3,C, Peyer’s patches showed a dramatic increase in luciferase activity 5 h after LPS challenge, while a much smaller, but still significant increase was found in the intestine without Peyer’s patches and no change in the mesenteric lymph node. There was no significant increase in luciferase activity in any of the tissues 48 h after LPS treatment, consistent with optical imaging (Fig. 4 C).
Transient vs chronic activation of Smad2/3 after endotoxin challenge in brain and intestine
To determine the temporal activation pattern of Smad2/3-dependent signaling after endotoxin challenge, photon emission in living SBE-luc mice was quantified at various time points after administration of LPS i.p. or i.v. (Fig. 4). Photon emission was rapidly detected in the intestinal area, peaking around 5 h, and declined almost to baseline levels 12 h after LPS injection (Fig. 4, A and B). To illustrate individual responses, temporal reporter gene activation patterns are depicted for three separate mice (Fig. 4,A). A similar activation profile was obtained when photon emission was quantified over the chest region, presumably reflecting photons emitted from the heart (data not shown). In contrast, bioluminescence over the skull peaked at 5 h but stayed at similar levels for up to 48 h (Fig. 4,C). It is important to note that similar patterns were obtained when LPS was administered i.p. or i.v. (Figs. 3,D and 4, B and C), indicating that the robust activation of Smad2/3 signaling in these regions is independent of the routes of LPS application. These findings could have important implications in understanding how the brain’s response to endotoxins differs from that of peripheral organs and why endotoxin-induced inflammation can persist in the brain for long periods of time.
Rapid activation of TGF-β signaling in response to focal brain injury
To examine activation of the TGF-β signaling pathway to direct tissue injury, brains were lesioned mechanically with a focal penetrating needle stab in the neocortex. Stab lesions resulted in consistent induction of cerebral TGF-β1 mRNA and more pronounced induction of PAI-1 mRNA (Fig. 5, A and B), which has been used frequently as an indicator gene for TGF-β signaling (1, 2). Both genes contain functional SBE elements in their promoters (1, 2). Consistent with this response, stab lesions resulted in significant reporter gene activation that could again be detected noninvasively in living mice (Fig. 5,C). Bioluminescence increased already 2 h after lesioning and peaked at 4 h before the signal slowly decreased to baseline levels at ∼48 h after injury (Fig. 5,D). This rapid activation of the reporter gene is likely a result of activation of latent TGF-β1, which is stored in the tissue in an inactive form but can be mobilized quickly after injury. Biochemical measurements of luciferase activity in postmortem brain tissues at 8 h correlated positively with the bioluminescent measurements obtained in the same mice immediately before they were sacrificed (Fig. 5 E). This demonstrates that noninvasive in vivo imaging can be used to measure Smad2/3-dependent signaling.
The TGF-β signaling pathway fulfills important roles in the normal and pathological function of many cells and tissues (1, 2). Our study describes an in vivo reporter mouse for this pathway and specifically for activators of Smad2/3. We demonstrate that this pathway has a high basal activity in the brain, intestine, heart, and skin and that it is rapidly activated after brain injury and inflammation.
Smad2 and Smad3 mediate transcriptional responses for TGF-β, nodal, activin, and some growth and differentiation factors (2, 4), and the SBE-luc reporter gene used here seems to respond to all of these tested in primary astrocytes derived from the reporter mice (Fig. 2,B). Activation of the reporter gene in vivo at basal state or after injury may therefore reflect signaling by either one or combinations of these growth factors and future studies using targeted inactivation of specific receptors or ligands will have to dissect these responses. At the transcriptional level, the reporter gene may be activated by Smad2, Smad3, or both and this is likely depending on tissue, developmental stage, and type of stimulus. Although Smad2 (in combination with Smad4) induces activation of the SBE reporter plasmids in cell culture only to a fraction of what is seen with Smad3 (Fig. 1), it is clear from recent in vivo studies that many tissues express a splice variant of Smad2 which can fully replace Smad3 function (26). Although Smad2 does not bind to DNA directly, the Smad2 variant lacking exon 3 can bind to the SBE sequence with an affinity similar to Smad3 (27) and activate all target genes essential for rescue of the lethal phenotype observed in mice lacking Smad2 (26). Targeted inactivation of Smad2 or 3 in (adult) SBE-luc mice may help in dissecting which Smad protein is mediating reporter gene activation in response to injury in different tissues.
SBE-luc reporter mice allowed us to image Smad2/3-dependent TGF-β signaling noninvasively in a temporal and spatial manner that reflects biochemical tissue reporter gene activity. Our studies show for the first time activation of the TGF-β signaling pathway in response to endotoxin administration or tissue injury in individual living mice over time and reveal reporter gene activation already 2 h after the challenge. This early induction of the reporter gene could be the result of signaling by pre-made latent TGF-β stored in the extracellular matrix, whereas newly synthesized TGF-β would result in a more long-lasting activation of the reporter.
Unexpectedly, the strongest reporter activity in unmanipulated SBE-luc mice was observed in the brain. Reporter gene activity in the brain may be induced by TGF-β or activin, which are expressed in the normal or injured adult brain (28, 29). This baseline activation of Smad2/3 signaling supports a possibly more important role for this pathway in the brain than previously appreciated. It is consistent with recent studies implicating TGF-β in synaptogenesis and retrograde signaling in Drosophila (30) and data in TGF-β1-deficient mice that identified TGF-β1 as an important factor in maintaining neuronal integrity and survival (31).
Basal reporter gene activity in the intestine, which includes the gut and associated lymphoid and fat tissues, was second to brain among all major organs analyzed. Interestingly, TGF-β signaling has an important role in maintaining homeostasis of intestinal epithelium and mutations in the TGF-β-type II receptor, Smad2, or Smad4 are frequently found in colorectal cancer (32). Unmanipulated mice also showed characteristic emission of photons from the nose, lower jaw, paws, and tail, indicating activation of Smad2/3 signaling in these areas most likely from cells of the skin or cartilage. TGF-β has important functions in the dermis and epidermis and TGF-β signaling is involved in regulating epithelial cell proliferation and differentiation (33).
Administration of the endotoxin LPS results in a prominent activation of proinflammatory pathways. It has also been shown to increase production of TGF-β1, which can stimulate the early phases of inflammation but later counteracts inflammation and induces tolerance against endotoxins (21, 22). LPS induced indeed a prominent activation of the SBE-luc reporter gene in many organs, most prominently in the intestine, brain, heart, and spleen (Fig. 3,C). The effect of peripheral endotoxin challenge on the brain has received increasing attention recently not only for its possibly protective effects on subsequent injury (see below) but also for its ability to cause chronic damage to neurons in the substantia nigra, a process critical to neurodegeneration in Parkinson’s disease (34). In contrast to the intestinal and cardiac response to LPS, reporter activity in the brain did not revert to baseline levels after 24 h but remained at the peak level reached 5 h after the challenge (Fig. 4). Cerebral TGF-β1 along with many other cytokines is known to be increased after LPS challenge (22, 35), although the molecular mechanism has not been elucidated yet. This LPS-induced increase in TGF-β1 may serve to protect neurons against injury as suggested by an elegant study by Boche et al. (35). Thus, LPS treatment reduced neuronal degeneration in response to the neurotoxin kainic acid in mice, and this protection was lost in the presence of Abs that blocked TGF-β signaling (35). LPS administration has also been demonstrated to reduce accumulation of amyloid in transgenic mouse models for Alzheimer’s disease under certain conditions (Ref. 36 and references therein). It is tempting to speculate that increased production of TGF-β1 is responsible, at least in part, for this effect by stimulating microglial phagocytosis of amyloid (37). Similar to LPS, mechanical injury resulted in the rapid activation of Smad2/3-dependent signaling in the brain but the signal decreased almost to baseline 2 days after injury (Fig. 5). It will be interesting to determine how LPS differentially regulates chronic increases in TGF-β signaling in different tissues.
Furthermore, we found quite unexpectedly that LPS administration resulted in the strongest induction of TGF-β signaling in the intestinal area even if LPS was injected i.v. This not only supports the immunoregulatory function of TGF-β in the maintenance of intestinal homeostasis, but also highlights the critical role of the intestine in endotoxemia-related pathological events, such as multiple organ failure syndrome (38). The dramatic increase of TGF-β signaling in the Peyer’s patches suggests that the lymphocytes are the main effector cells in this system and underline their critical role in intestinal function and host defense (39).
In summary, SBE-luc mice offer the possibility to study activation of Smad2/3-dependent TGF-β signaling in normal and pathophysiological conditions in living mice over time. We have used these mice here to demonstrate activation of this signaling pathway in unmanipulated mice, predominantly in intestine and brain. We also showed tissue-specific responses and temporal activation of Smad2/3-dependent signaling in living mice in response to systemic endotoxin challenge or tissue-specific injury. The possibility to monitor and quantify reporter gene activity in individual animals over time is particularly useful for the study of complex diseases in which interindividual variations in disease progression make it often impossible to study disease modifiers or therapeutic strategies. Because reporter mice such as the SBE-luc mice allow us to detect disease progression in each mouse individually, they are ideal tools to study endogenous or therapeutic modifiers of diseases that involve activation of the particular reporter gene. SBE-luc mice will thus allow us to correlate the level and time of reporter gene activation with biological effects of TGF-β such as fibrosis or immunosuppression.
We thank S. Dalai and M. Lin for technical support. We also thank Dr. T. Brunner (University of Bern, Bern, Switzerland) for advice on the dissection and analysis of the intestine and associated lymphoid tissues.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the National Institutes of Health (to T.W-C.; Grant AG20603), the Veterans Administration Geriatric Research Education and Clinical Center, and Mental Illness Research, Education and Clinical Center services (to T.W-C.). A.H.L. was supported by a National Research Service Award.
Abbreviations used in this paper: BMP, bone morphogenic protein; GDF, growth differentiation factor; sr, steradian; SBE, Smad-binding element; PAI-1, plasminogen activator inhibitor 1.