Wells' Lab - Summary of Research Programs

Alan Wells, MD, DMS

The Wells Laboratory research program, in close collaboration with its research partners, aims to understand cell migration in terms of how motility processes are regulated, and understand how this regulation of migration plays a role in physiologic and pathologic situations. We are integrating the knowledge gained from our biochemical and biophysical mechanistic studies into our investigations concerning conditions of dysregulated (tumor invasion) and orchestrated (wound healing and organogenesis) cell motility. As part of understanding the motility response, we are investigating both how this particular integrated cell response is selected from among others and the metabolic consequences of motility. This integrative approach provides reinforcing insights and novel avenues for exploration into the basic signaling pathways as well as functioning of whole organism. As a model system, we explore motility signaling from the epidermal growth factor receptor (EGFR) in adherent cells. EGFR plays a central role in the functioning in a wide variety of both stromal and epithelial tissues, and is the prototype for other receptors with intrinsic tyrosine kinase activity. Thus, these studies should have widespread implications.

Cell Biology - Biochemistry and Biophysics

Initially, we sought to define the signaling pathway from the EGFR to the cell motility machinery. Philip Chen, defined the need for receptor tyrosine kinase activity and activation of phospholipase Cg for EGFR-mediated cell motility. We then demonstrated that the hydrolysis of phospho-inositide bisphosphate mobilizes actin-modifying proteins, using gelsolin as a marker for this group, and that this event directly remodels the actin cytoskeleton. Jeff Chou has found differential effects of PLC activity on the actin-modifying proteins gelsolin, profilin and cofilin.  Gelsolin, while relatively deficient in the extending lamellipod is incorporated into the remodeled cytoskeleton, whereas profilin is enriched in this motility cellular organ.  Furthermore, inhibition of PLCg signaling abrogated EGF-induced but not basal haptokinetic motility, suggesting that this pathway is preferential to growth factor-induced motility. These studies defined the first direct path from any cell surface receptor to alteration of the actin cytoskeleton.

We have extended our biochemical and cell biologic findings to probe the biophysical underpinnings of induced cell motility. In collaboration with Doug Lauffenburger and Linda Griffith and their laboratories (MIT), the actual motile events (lamellipodial extension, cell-substratum adhesion flux, or contractility) are being aligned with specific signaling pathways. Marti Ware (at MIT) found that PLCg-mediated pathway increases cell motility by modulating cytoskeletal reorganization required for asymmetric cell locomotion. In the course of these studies, evocative findings demonstrated that EGFR activation speeds movement but shortens persistence. We will assess the implications of cell speed being intrinsic upon EGFR activation but cell direction (as defined by persistence) being extrinsic in both physiologic (wound healing) and pathologic (tumor invasiveness) conditions.

Other pathways and intracellular signaling events also are required for EGFR-mediated cell motility. Heng Xie and Philip Chang examined alterations in focal adhesion strength and constitution in collaboration with Joanne Murphy-Ullrich (UAB). We have determined that EGFR-mediated focal adhesion disassembly occurs through an erk MAP kinase-dependent mechanism. Inhibition of this signaling pathway prevents induced cell motility.  Futher downstream, Angela Glading has shown that calpain-mediated proteolysis is actively involved in EGFR-mediated motility. EGF activates calpain with focal adhesion components demonstrating proteolytic cleavage. Unexpectedly, this is linked to the MAP kinase-associated de-adhesion likely through a direct phosphorylation/activation mechanism.  this entire enzymatic cascade occurs at the inner face of the plasma membrane, localizing calpain activity in the region of cytoskeletal linkages to the substratum.  These two events demonstrate that EGFR induces motility by modulating at least some of the pathways utilized by haptokinetic signals.

Marti Ware and Gargi Maheshwari (at MIT), in collaboration with Philip Chang , determined the biophysical consequences of EGFR-mediated motility on cell-substratum adhesion. EGF causes an acute decrease in cell adhesiveness to substratum, that reverses within 2 to 4 hours. Interestingly, for the first hour or so after EGF stimulation cell movement is reduced from basal levels, despite vigorous membrane ruffling. Both the detachment and subsequent re-attachment are necessary for cell motility, as cells on poorly adhesive surfaces do not regain adhesion and fail to locomote, while cells on tightly adhesive surfaces fail to reduce their interactions with substratum sufficiently to allow movement.  Findings by Angela Glading and Hidenori Shiraha show that calpain activation is required for the de-adhesion, though the final linkages and end targets are still undefined.

Fred Allen (at MIT), in collaboration with Eliot Elson's laboratory (Washington University), is probing EGFR-mediated cell contraction. They have found that EGFR signaling promotes cell contraction in a complex manner that could reflect two distinct biochemical pathways working on different timescales. Using signaling-restricted mutants and specific pharmacological and molecular inhibitors, these pathways are being parsed. Furthermore, we are pursuing the possibility that select chemokines switch EGFR signaling from a locomotive response to a contractile response similar to what is seen when preventing calpain-mediated de-adhesion.

These studies have defined the biophysical consequences of two motility-related signaling pathways - that PLCg activation leads to membrane protrusion and MAP kinase activation allows detachment from substratum. Current studies aim to fully describe the individual biophysical aspects of cell motility in biochemical terms.

Pathology - Tumor Progression

Upregulated EGFR signaling has been correlated with tumor progression to the invasive and metastatic state. As prostate cancers present a TGFa-EGFR autocrine signaling loop, we asked if, and how this signaling contributes to prostate cancer invasiveness. To address this question, we generated a series of human prostate carcinoma cells (DU-145) genetically-engineered to overexpress wild-type EGFR or a fully-mitogenic but non-motogenic truncated EGFR. Tim Turner (now at Tuskegee University) demonstrated that in vitro invasiveness of these cells was dependent on EGFR signaling, requiring domains in the carboxy-terminal tail of the receptor. A similar requirement for carboxy-terminal EGFR signaling has been observed for colorectal carcinoma metastasis in vivo by our collaborator Robert Radinsky (MD Anderson/Amgen). Furthermore, proteolytic activity was seen not to be rate-limiting for invasiveness of our prostate tumor cells. Tim Turner continued this work to demonstrate that a similar pattern of invasiveness was seen in mouse xenograft models with the cells inoculated into the anterior prostate or peritoneal cavity. To probe the role of the EGFR-PLCg motility pathway, we treated these mice with a pharmacologic inhibitor of PLC; this treatment severely abrogated tumor invasiveness. Tumor invasiveness also can be inhibited by expression of a dominant-negative PLCg fragment. The inhibition of invasiveness was independent of effects on cell proliferation. We have found a similar requirement for PLC activation for glioblastoma cell invasiveness in vitro in collaboration with Paul Penar (Univ Vermont). Thus, for the first time, we described a signaling element required specifically for tumor cell invasion.

Jareer Kassis has found that PLCg-dependent invasiveness is not unique to these cells or even prostate cancer. In collaboration with Norman Greenberg (Baylor), he showed that PLC activation is required for invasiveness of SV40 T-antigen-driven prostate tumors. To extend the generalizability of growth factor receptor-mediated PLCg activity to other tumors, Jareer Kassis determined that abrogation of PLC signaling limits in vitro invasiveness of breast and bladder carcinoma cell lines. In two EGFR-overexpressing cell lines examined, this was dependent on autocrine EGFR-stimulation. Thus, PLCg appears to be required for invasiveness for a broad range of tumors and therefore can be targeted for therapeutic affect.  For final proof, we have generated a transgenic mouse to express a tetracycline-regulatable dominant-negative PLCg fragment in only prostate or breast tissue.  Jareer Kassis and James Solava are mating these mice to models of invasive breast and prostate carcinoma to determine whether tumor progression can be blocked by inhibiting motility.

We also have employed our xenograft model of aggressive prostate cancer to test novel ways of treating progressive cancers. In one collaboration with Andrew Kraft (Univ Colo), Tim Turner and Jareer Kassis demonstrated the therapeutic efficacy of a natural marine product, dolastatin, on prostate cancer. This agent targets cell proliferation and viability, and thus may serve as combination treatment along with PLC activation inhibitors. A second therapeutic intervention against prostate cancer is being investigated by Jose Souto and James Solava in collaboration with Tim Turner's lab at Tuskegee. They investigated how LHRH agonists directly inhibit prostate cancer cell growth. Initial findings suggest that it occurs via downregulation of autocrine EGFR signaling that is normally required for prostate cell proliferation and survival. The biochemical linkages between the G protein-linked LHRH receptor and EGFR signaling involve PKC negative transmodulation.  These collaborative projects complement our approach to prostate cancer progression and its amelioration.

Physiology - Wound Healing & Organ Morphogenesis

EGFR-mediated motility signaling in fibroblasts is negatively affected by inhibitory signals during wound repair and by aging. In a mechanistic study, Hidenori Shiraha is defining the intracellular crosstalk between the pro-motogenic EGFR signaling pathways and the counter signals from the anti-inflammatory chemokine IP-10. IP-10 selectively impairs motility but not proliferation responses, and that this occurs downstream of PLCg and MAP kinase signaling by affecting endpoint motility events. Excitingly, with Angela Glading, he found that EGFR-mediated activation of calpain and subsequent cell-substratum de-adhesion is inhibited by IP-10 signaling. Thus, calpain serves as a central control point for deciding the cell response to potentially pleiotropic EGFR signaling.  Fred Allen (MIT/Drexel) has shown that by inhibiting calpain we shift the fibroblast response to EGF to one of matrix contractility. Latha Satish is pursing the hypothesis that the ELR-negative CXC chemokines, IP-10 and IP-9, that are expressed late during wound healing synchronize the epithelial and dermal compartments during repair.  Initial findings suggest that IP-9 is expressed only in differentiated basal keratinocytes at the time that re-epithelization is complete.  This chemokine limits fibroblast motility suggesting that it may serve to switch the repair response to the resolution phase.

Hidenori Shiraha also found that during aging, both motility and mitogenesis are decreased, with the EGF response being lost in near senescent dermal fibroblasts. This is a consequence of reduced levels of EGFR due to markedly reduced transcription, expression of 'young' levels of EGFR restores the EGF responsiveness. The basis of the reduced transcription rate of the EGFR gene noted during cell aging is currently being investigated.

We have reported that differential trafficking of EGFR, as dictated by specific ligands, alters the cell response to receptor activation Signaling and metabolic consequences. Therefore, we predict that extracellular matrix interactions with ligands may alter EGFR trafficking and resultant cell responses. Scott Swindle and Kien Tran, in collaboration with Linda Griffith's laboratory (MIT) is investigating the consequences of coordinated and linked EGFR and integrin signaling, as well as seeking low affinity ligands for EGFR. Most intriguingly, they have found that some of the EGF-like repeats in matrix components may bind and activate EGFR directly but at affinities 1000-fold lower than classic ligands.  The implications of matrix-tethered ligands for cell responses during the dynamic and changing nature of matrix maturation are immense.  The coordinate activation of EGFR and integrins is expected to alter the balance of signaling to favor a particular cell response. Gargi Maheshwari and Gillian Brown (both at MIT) have found that the responsiveness to EGF is dictated by the density of integrin ligands. This project will address the question of how multiple external signals are integrated for a resultant cell response. We propose that the knowledge gained will provide for engineered biologic surfaces of defined properties.

Organ morphogenesis also requires orchestration of both cell motility and proliferation in response to external signals. Utilizing our transgenic mice which express the dominant-negative PLCz fragment in a tissue-specific, regulatable manner, Hyung Kim and James Solava are investigating the role of PLCg-mediated motility in prostate and mammary gland development. As these organs develop to a significant extent post-natally and are dispensable to survival, they provide a rare opportunity to isolate specific stages in organogenesis. Branching and lobule numbers are reduced when the motility-associated dominant-negative PLCz is expressed in the epithelial cells of these tissues. Thus, epithelial motility is rate-limiting for organogenesis in these tissues

Signaling and Metabolic Consequences

Growth factors signal both proliferation and motility in a highly orchestrated fashion during wound healing and in a dysregulated manner during tumor progression. EGFR ligands abound during wound healing, with the receptor present on both dermal fibroblasts and keratinocytes. A critical question is how the cell decides upon the proper response. My laboratory is pursuing both extracellular signals that modulate which cell response predominates and intracellular crosstalk which affects these responses. Philip Chen demonstrated that activation of the motility-signaling PLCg pathway with subsequent activation of PKC results in preferential feedback attenuation of mitogenic signaling. These data suggest that strong, uninterrupted EGFR signaling favors mitogenesis, while discontinuous signaling is more compatible with motility. This finding provides a mechanistic basis for changing the balance among a myriad of intracellular signaling pathways and altering the resultant cellular response.

It is also possible that the physiochemical properties of ligand-receptor interactions elicit differential cell responses. In collaboration with Doug Lauffenburger and Cartekiya Reddy (now at Bayer Corporation) we found that ligand binding and receptor recycling properties dictate the strength and persistence of signaling from the EGFR. For sustained signaling, dissociative ligands are favored in receptor-limited situations, and non-dissociative ligands when ligand is limited. This explains, in part, why there are multiple ligands for a single receptor, EGFR, and the tissue and pathologic state distribution of the different EGFR ligands. Jason Haugh (at MIT ) is extending these studies to quantitate the signaling balance among different downstream pathways. He found that receptor-ligand internalization and avidity affect the balance of intracellular signaling. Interestingly, though EGFR and PLCg continue to be phosphorylated during endosomal trafficking, PIP2 hydrolysis is noted only from plasma-membrane assoicated EGFR. On the other hand, ras is activated at least as robustly by internalized EGFR as by plasma membrane-associated receptors. But this pathway spatially segregates downstream.  Angela Glading has shown that while erk MAP kinases are activated from both the plasma and intracellular membranes only the plasma membrane-associated erk leads to M-calpain activation.  This suggests a principal of how the integrated biological responses of motility and mitogenesis form the basis of his quantitative approach to the intracellular signaling network triggered by EGFR.

Integrated cell responses of motility and proliferation are energy-requiring. To determine the metabolic flux of glucose, we collaborated with Rob Hardy (UAB) to investigate the consequences of EGFR activation of glucose uptake and disposal. We have expressed signaling-restricted EGFR constructs in adipocytes to serve as analogs for the closely related insulin receptor. In this manner, we can dissect the intracellular signaling pathways which accomplish the glucose disposal and define the defects at play in NIDDM (type II diabetes). This latter problem has been hampered by the presence of high levels of endogenous insulin receptors in appropriate cells, adipocytes and skeletal muscle, which is circumvented by utilizing exogenously-expressed EGFR in 3T3-L1 adipocytes.

In these transgenic adipocytes, EGF induces glucose storage and GLUT4-mediated glucose uptake, similarly to insulin. Induced glucose uptake requires receptor kinase activity and signaling motifs in the EGFR carboxy-terminal tail. Using a candidate effector approach, Mark Van Epps-Fung has shown that PLC activity is involved in EGFR- and insulin receptor-mediated glucose uptake. Mark Van Epps-Fung has also demonstrated that glucose storage as glycogen and lipid is augmented by EGFR signaling similarly to insulin, but that carboxy-terminal signaling motifs are not required. We are investigating the molecular mechanisms of this pathway. EGFR signaling may promote glucose homeostasis by independent pathways from those utilized by the insulin receptor. Thus, alternative pathways may be utilized to bypass insulin resistance in NIDDM. As an example, EGF does not induce IRS-1 phosphorylation. Current investigations are aimed at other mechanisms and models of insulin resistance.