Phosphatidylserine Asymmetry and Cell Survival

Journal Articles

Ari Helenius

Mercer J, Helenius A. 2010. Apoptotic mimicry: phosphatidylserine-mediated macropinocytosis of vaccinia virus. Ann. NY Acad. Sci. 1209: 49-55.

Mercer J, Knébel S, Schmidt FI, et al. 2010. Vaccinia virus strains use distinct forms of macropinocytosis for host-cell entry. Proc. Natl. Acad. Sci. USA 107: 9346-51.

Mercer J, Helenius A. 2008. Vaccinia virus uses macropinocytosis and apoptotic mimicry to enter host cells. Science 320: 531-5.

Chris Reutelingsperger

Schutters K, Reutelingsperger C. 2010. Phosphatidylserine targeting for diagnosis and treatment of human diseases. Apoptosis 15: 1072-82.

Kenis H, Reutelingsperger C. 2009. Targeting phosphatidylserine in anti-cancer therapy. Curr. Pharm. Des. 15: 2719-23.

Alan Schroit

Mirnikjoo B, Balasubramanian K, Schroit AJ. 2009. Suicidal membrane repair regulates phosphatidylserine externalization during apoptosis. J. Biol. Chem. 284: 22512-6.

Mirnikjoo B, Balasubramanian K, Schroit AJ. 2009. Mobilization of lysosomal calcium regulates the externalization of phosphatidylserine during apoptosis. J. Biol. Chem. 284: 6918-23.

Balasubramanian K, Mirnikjoo B, Schroit AJ. Regulated externalization of phosphatidylserine at the cell surface: implications for apoptosis. J. Biol. Chem. 282: 18357-64.

Luster TA, He J, Huang X, et al. 2006. Plasma protein beta-2-glycoprotein 1 mediates interaction between the anti-tumor monoclonal antibody 3G4 and anionic phospholipids on endothelial cells. J. Biol. Chem. 281: 29863-71.

Balasubramanian K, Schroit AJ. 2003. Aminophospholipid asymmetry: A matter of life and death. Annu. Rev. Physiol. 65:701-34.

Philip Thorpe

DeRose P, Thorpe PE, Gerber DE. 2011. Development of bavituximab, a vascular targeting agent with immune-modulating properties, for lung cancer treatment. Immunotherapy 3: 933-44.

He J, Yin Y, Luster TA, et al. 2009. Antiphosphatidylserine antibody combined with irradiation damages tumor blood vessels and induces tumor immunity in a rat model of glioblastoma. Clin. Cancer. Res. 15:6871-80.

David Ucker

Mitchell JE, Cvetanovic M, Tibrewal N, et al. 2006. The presumptive phosphatidylserine receptor is dispensable for innate anti-inflammatory recognition and clearance of apoptotic cells. J. Biol. Chem. 281: 5718-25.

Ucker DS, Jain MR, Pattabiraman G, et al. 2012. Externalized glycolytic enzymes are novel, conserved, and early biomarkers of apoptosis. J. Biol. Chem. 287: 10325-43.

Sponsors

Presented by

Silver Sponsor

Highlights

• In young, healthy cells almost all PS molecules face the cytoplasm within the two layers of the cell membrane.
• Disruption of PS asymmetry is an important signal in clotting and in clearing aging cells from the bloodstream.
• Red blood cells (RBCs) lose the ability to internalize PS as they age. PS externalization serves as a tolerizing signal that prevents the body from triggering an immune response to these cells when they undergo apoptosis.

Understanding PS asymmetry via red blood cells

Much of our scientific understanding about PS and its asymmetric distribution in the cell has come from studies of red blood cells (RBCs). An average of one trillion of these cells are cleared from the human body each day. An important biological question about RBCs has always been how the body knows to clear these cells without triggering an immune response, said Alan Schroit of UT Southwestern Medical Center. A better understanding of PS asymmetry has helped researchers answer this question.

Because the networks of lipids and proteins within lipid membranes are associated non-covalently, studies to understand their distribution have required careful design. Early studies by Maddy probed the membrane distribution of lipids by incubating molecules that would react with amines on the outside surface of the membranes of RBCs to see which molecules were modified. Molecules facing the cytoplasm of the cell would be protected. Nearly all of the PS and most of the phosphatidylethanolamine (PE) was not modified. Other studies confirmed these initial findings that PS was primarily on the inner leaflet of the plasma membrane. But researchers did not understand how and why cells maintained this asymmetry, what stimuli perturbed it, and what consequences might arise when the distribution was perturbed.

About a decade later, in 1982, researchers showed that stimuli such as calcium ions and thrombin could prompt platelets to scramble their lipids and to distribute them evenly across both membranes. PS on the surfaces of these cells bound to activated Factor V and Factor X, which bound to prothrombin and fibrinogen and converted them to their active forms. Therefore the movement of PS to the surface had functional significance in specific biological processes, such as clotting.

Chemical tools that add a fluorescent or spin-labeled probe to lipid molecules have provided ways to investigate the dynamics of lipid distribution within the cell. The incubation of cells with labeled lipids under various conditions allowed researchers to follow the movement of the labeled lipids from the outside to the inside of the bilayer. Almost all PS will move from the outside to the inside of the cell, along with 25% of the PC and most of the PE, and that movement depends on both ATP and temperature.

Fluorescence- and spin-labeling of lipids allows researchers to probe their distribution in the cell membrane. (Image courtesy of Alan Schroit)

So why would cells invest energy in maintaining this distribution of lipids? Studies of a range of RBCs of different ages showed that young cells readily internalize PS, but that older cells gradually lose this ability. This waning ability to internalize PS also correlates with the phagocytosis, or "devouring," of these cells by macrophages. Imaging studies show that macrophages form rosettes around RBCs that have PS on their membrane's exterior. Electron microscopy shows that the interaction between macrophages and RBCs is precise and does not distort cell shape particularly when compared with binding between antibodies and RBCs.

In other experiments researchers injected labeled RBCs with PS on their surfaces into live mice. Within 30 seconds most of those cells had disappeared from the blood. Within minutes the labeled cells were detected in the liver and spleen of the mice, indicating that they had been cleared from circulation.

ATP-dependent flippase enzymes facilitate the movement of PS from the outer to the inner leaflet of the plasma membrane. ATP binding cassette (ABC) transporters can serve as floppases that move lipid molecules from the inner leaflet to the outer leaflet. Calcium-dependent scramblases completely disrupt the asymmetry of the plasma membrane in coagulation, to clear apoptotic cells and as a signal that the immune system to tolerate these cells (a tolerizing signal) rather than considering them as foreign bodies.

This tolerizing signal for the immune system has a variety of implications in cancer and in infectious disease. Tumor cells express PS on the surface which explains how these cells shut down the body's autoimmune response. Viruses and parasites also take advantage of PS to evade the immune system.

The cast of characters that control PS asymmetry. (Image courtesy of Alan Schroit)

Speaker:
David Ucker, University of Illinois at Chicago

Highlights

• Although the RBC studies have pointed to PS as an important factor for immune tolerance during apoptosis, detailed studies of the steps of early apoptosis ruled out PS as the triggering factor for the dampened immune response.
• Glycolytic enzymes that function within the cytosol can trigger early apoptosis when they are externalized to the outer surface of the cell.

Examining the biology of early apoptosis

In their studies of the events that happen early in apoptosis, David Ucker and his colleagues at the University of Illinois at Chicago wondered whether the movement of PS to the exterior of the membrane might dampen the immune response to these cells. But in a series of experiments, they showed that PS was not the culprit. Instead, these cells shuttle glycolytic enzymes to their surfaces, which triggers immunity.

Consistent with earlier results, Ucker observed that apoptotic cells, unlike viable or necrotic cells, inhibit the release of two pro-inflammatory cytokines, IL-6 and TNF-α. Those cells don't just block the synthesis of those proteins, but they also block the expression of the genes themselves. Therefore the expression of these genes during apoptosis was not a factor in the suppression of the immune response in these studies.

The researchers decided to look at the engulfment of apoptotic cells by macrophages. When they treated apoptotic cells with cytochalasin D, a molecule that blocks the reorganization of the cytoskeleton, necessary for cell–cell recognition, they showed that the apoptotic cells and macrophages need to recognize each other to block the inflammatory response. But the macrophages did not need to engulf the apoptotic cells to block the inflammatory response.

Apoptotic and necrotic cells show distinct signals that affect how these cells are recognized (Image courtesy of David Ucker)

In studies that compared apoptotic and necrotic cells, macrophages recognized these cells by distinctly different mechanisms. When the researchers used propidium iodide, an intercalating and staining agent to examine the overall membrane integrity of these cells types, the membranes in necrotic cells fell apart almost immediately. Viable cells showed little infiltration from the propidium iodide, and early apoptotic cells maintained much of the membrane integrity.

Cells across species show similar non-immunogenic responses to apoptotic cells, but to understand whether PS was involved the researchers examined PLB-985 cells, a cell line that does not externalize PS. Surprisingly, however, those cells showed the same patterns of membrane integrity and non-immunogenic activity as seen in cells that can externalize this lipid. So PS on the outside of the cell could not be causing the suppression in immune response.

Comparable immune suppression occurred in a variety of organisms, including in flies, frogs, mice, and humans, which suggested a common molecular cause found in all these organisms. So Ucker and his colleagues were looking for another factor that might mitigate the immune response.

Two dimensions of immunity

Building on some of the ideas in Schroit's talk, Ucker described the two dimensions of immunity. Classical immunity is the process by which an organism differentiates between itself and an invading, potentially threatening pathogen. A different dimension of immunity comes from the process by which a cell ages from a young viable cell toward apoptosis. The elimination of these older cells occurs through an immune process distinct from that of classical immunity.

Some viral pathogens take advantage of the same immunity tricks that aging, apoptotic cells use. Cells that die in response to an infection from adenovirus produce cytopathic effect (CPE) bodies that mimic the immunosuppression observed in apoptotic cells.

But with evidence that the chemical trigger for this process was not PS, the researchers treated apoptotic cells with proteases, to see if the protective factor might be a protein. Their results showed that apoptotic cells are enriched in proteins in the plasma membrane and are protease sensitive. So Ucker and his colleagues now knew that they were looking for evolutionarily conserved proteins. Because glycolytic enzymes are the largest group of proteins overexpressed on the membranes of apoptotic cells, they were potential candidates. The researchers showed both with proteomic studies and with assays that these glycolytic proteins are enriched on the surfaces of early apoptotic cells and that this externalization does not occur in viable cells. This process happens at the same time that PS externalization typically occurs.

The externalization of glycolytic enzymes accounts for the binding of plasminogen to apoptotic cells. Caspase activity appears to be one of the triggers of this process because blocking caspase activity prevented the externalization of these enzymes.

Speaker:
Ari Helenius, ETH Hönggerberg, Switzerland

Highlights

• Viruses interact with surface proteins to trigger macropinocytosis, and PS in the viral membrane can trigger this process.
• The activation of surface proteins does not necessarily occur via direct interaction between PS in the viral membrane and proteins on the surface of host cells. Adaptor proteins bridge some of these interactions.
• PS on the surface of viruses furnishes them with the immunity conferred to apoptotic cells.

PS in the viral membrane

Viruses enter the cell through endocytosis, a cascade of signals that allow viruses to bind to the surface of the hosts' cells and become internalized by the membrane. Eventually a virus's nucleic acids are transported to the nucleus where they hijack the activity of a host cell. Ari Helenius of ETH in Hönggerberg described recent studies to uncover the role of PS in viral entry.

A variety of attachment factors can concentrate viruses on the surface of any given host cell and interactions with various receptor proteins can facilitate viral entry through a number of mechanisms: endocytosis, conformational changes, or membrane fusion. Endocytosis occurs through a variety of visually distinct mechanisms where the cells form protrusions or indentations that surround the external material. In macropinocytosis a finger-like projection extends from the plasma membrane and engulfs the target viral particle. This mechanism is active in the endocytosis of a variety of viral particles including HIV, Ebola, and vaccinia virus particles.

To study viral entry, researchers can use microscopy and biochemical techniques to view the cells and to track different molecules. Inhibitors and siRNA silencing allow researchers to perturb viral entry to understand the proteins that facilitate these processes, Helenius said.

Identifying the infectome

In their recent studies of viral entry, Helenius and postdoctoral researcher Jason Mercer have focused on vaccinia virus, one of the most complex animal viruses. As the virus infects cells, the cells extend blebs (bulges in the membrane) and filopodia and then retract the blebs as particles are internalized through macropinocytosis. Mercer and Helenius wanted to better understand the genes in the host cells that assist in this process of infection, called Trojan genes. To do so they used high-throughput screening techniques to identify the infectome, the body of genes that facilitate infection, of these viruses.

They initially screened HeLa cells with vaccinia virus particles against a Qiagen library of 7000 druggable genes, and in every case they tested 3 or 4 siRNAs that targeted each gene. The researchers identified an initial set of hits based on whether 2 or 3 of the siRNAs targeting these genes reduced viral infection by 50% or more. They then narrowed this group of primary hits by screening against a secondary library available from Ambion, which produced 300 secondary hits.

Using bioinformatics, the team analyzed the identified genes and found that many are in pathways associated with the epidermal growth factor receptor (EGFR) and with the proteins that facilitate macropinocytosis. The researchers then studied EGFR inhibitors against vaccinia virus in mixtures of cells. The inhibitors blocked entry of and infection by the virus.

The role of PS

The vaccinia virus expresses PS on its surface as a way of evading the classical immune response, and surface expression of PS is required for the virus to enter cells. Treatment with annexin V (a protein that binds PS) interferes with infection of and blebbing on the host cells, but the viral particles still bind to the cell surface. If PS is eliminated from the virus and then added back via new liposomes, viral activity is restored. The virus also regains its infectivity if another negatively charged phospholipid is added.

To identify the surface proteins that recognize PS directly or indirectly, Helenius and Mercer collaborated with Alexander Frei and Bernd Wollscheid, at ETH in Zurich. Frei and Wollscheid had developed ligand-based receptor capturing (LCR), a technique that takes advantage of a crosslinking molecule with three functions—one functional component can conjugate to carbohydrates on the surface of cells, another part of the molecule reacts with protein side chains, and the molecule's biotin moiety aids purification of the "captured" receptor.

Researchers can identify viral receptors using ligand-based receptor capturing. (Image courtesy of Ari Helenius)

Using this technique the researchers could bind cellular receptors, purify the receptors from a digested cell lysis mixture, and identify them from the peptide sequences in the mixture using mass spectrometry. In this way, the team identified seven candidate receptors likely involved in viral entry and used siRNA to monitor perturbations in viral entry in response to the elimination of five of those receptors: Axl, the mannose 6-phosphate receptor, dystroglycan-1, chondroitin sulfate proteoglycan-4, and cadherin 13.

Many of the receptor tyrosine kinases that recognize PS on the surface of viruses act through an additional untethered connector protein. Gas6 may serve as one of these proteins that bridge PS and the identified surface receptors.

Speaker:
Chris Reutelingsperger, Maastricht University, The Netherlands

Highlights

• Proteins that bind to PS, such as annexin A5, provide tools for imaging and the basis for new treatments for disease.
• A fusion protein of annexin A5 with a cytotoxic protein can kill tumor cells.
• A mutation that changes the affinity of annexin A5 binding can prevent the formation of atherosclerotic plaques in rats.

Binding PS

PS on the outer membrane surfaces of viruses or of diseased cells, such as those in tumors, can differentiate those cells from healthy ones. Therefore, proteins that bind to PS could provide tools for both diagnosis and treatment of diseases, says Chris Reutelingsperger of Maastricht University. Expression of PS on the surface of cells serves several purposes in the body, both to cue the removal of aging or dying cells from the system and to activation of various clotting and immune responses.

The proteins annexin A5, synaptotagmin I, and lactadherin all bind to membranes that contain PS. For the first two proteins, binding also requires calcium ions. Reutelingsperger presented his work with annexin A5, a 36 kDa protein with a convex binding face that interacts with phospholipids. Annexin A5 also has a concave binding face.

When annexin A5 is labeled with various types of chemical tags, researchers can use optical imaging, magnetic resonance imaging, nuclear imaging, and ultrasound imaging to highlight the tissues it binds within the body. Researchers have used annexin A5 as a non-invasive imaging agent both in animal models and in human studies to examine diseased tissue from a range of diseases including Crohn's disease, rheumatoid arthritis, atherosclerosis, and cancer.

Tumor imaging with annexin A5

Reutelingsperger focused most of his talk on the application of these techniques to studies of cancer. He demonstrated how his team employed labeled annexin A5 to image tumor cells in the neck of an individual with lymphoma. They could follow the uptake of the annexin A5 into the tumor after treatment with radiation.

Labeled annexin A5 allows researchers to follow radiation treatment in an individual with lymphoma. (Image courtesy of Chris Reutelingsperger)

One challenge in using annexin A5 is the delivery of the protein to the tumor cells. Because of the leaky vasculature in tumors, the uptake of these proteins depends on the fluid pressure. Annexin A5 falls into a size and affinity gap that limits its ability to infiltrate tumor cells efficiently. To avoid this problem the researchers either needed a smaller molecule with higher affinity for PS or a larger molecule with similar affinity. This second option was easier to develop.

To construct a larger, macromolecular variant of the protein, Reutelingsperger and his colleagues built in several sites labeled with biotin that bind tightly to streptavidin, increasing the overall bulk of the annexin A5 complex to 200 kDa. They also designed control structures that included the backbone of the annexin protein but without the calcium binding residues so that they would not bind to PS. But this bigger complex did not give better results in studies with nude mice that had colon cancer tumors. The larger complex had a long half-life, and therefore the researchers could not distinguish a tumor that had been treated from one that hadn't. Though it will be more difficult to construct the smaller annexin-type derivatives with higher affinity, the researchers are currently working on this strategy.

Annexin A5 to treat tumors

The researchers also looked at treatment strategies using these annexin A5-based molecules. They developed a fusion protein that linked annexin A5 with ETA, an exotoxin from Pseudomonas aeruginosa that kills through interactions with a cytosolic protein. The linkage between the two proteins includes a protease cleavage site. Cell culture studies showed that the fusion protein still showed calcium-dependent binding to PS. The fusion protein was taken into the cells and was cleaved, and PS binding enhanced the fusion protein's cytotoxicity.

The annexin A5-ETA fusion protein is cleaved inside cells and kills tumor cells. (Image courtesy of Chris Reutelingsperger)

Reutelingsperger and his colleagues then looked at the effect of the fusion protein on nude mice that had human colon cancer tumors. The treatment appears to be PS dependent, and the inhibition of tumor growth increased when the fusion protein was given with another agent, cyclophosphamide, a chemotherapy drug.

Annexin A5 to treat inflammation

The researchers also investigated looked at another potential therapeutic application of annexin A5. Efferocytosis, the phagocytosis of apoptotic cells, doesn't occur appropriately in inflammatory diseases such as atherosclerosis. To enhance this process, they developed a new annexin A5-based molecule that included an RGD (arginine-glycyl-aspartic acid) motif that enhances the binding of phagocytes.

Although wild-type annexin A5 inhibited efferocytosis as expected, the modified protein enhanced this process, suggesting that the protein was oriented as predicted. In initial in vivo studies, the effect varied depending on the population of the phagocytes, but the team did see enhancement of one inflammatory factor, IL-10, in the presence of the modified annexin protein. These results suggested that this therapeutic strategy could work.

Reutelingsperger and his team then tested these proteins in rats that were fed a Western-type diet. The wild-type annexin A5 protein had no effect on the plaques measured in the carotid arteries of the rats. However, the RGD-annexin A5 inhibited plaque formation in rats. In animals where plaques had already formed, RGD-annexin A5 stabilized plaques those plaques so that they did not continue to grow.

RGD-annexin A5 blocks the formation of atherosclerotic plaques in rats fed a Western diet. (Image courtesy of Chris Reutelingsperger)

Speaker:
Philip Thorpe, UT Southwestern Medical Center

Highlights

• Bavituximab, an antibody drug in development by Peregrine Pharmaceuticals, targets PS in tumor blood vessels and can reactivate immunity to cancer cells.
• This antibody can serve as both an imaging agent and as an adjuvant treatment with other therapies.
• Bavituximab can induce tumor-specific T-cell immunity in some situations.

Targeting tumor vasculature

Taking advantage of the externalization of PS on tumor cells could provide new treatment strategies for cancer treatments. Philip Thorpe of UT Southwestern Medical Center described preclinical and clinical studies of bavituximab, an antibody drug developed by Peregrine Pharmaceuticals that targets PS in the vasculature of tumors.

As tumors develop, they create a stressful, hypoxic cellular environment that results in the formation of reactive oxygen species and the release of inflammatory cytokines. In this environment, the asymmetry of lipid distribution in the membrane of the endothelial cells that line tumor blood vessels breaks down, leading to the externalization of PS. In addition, PS is often exposed on the surface of tumor cells.

As a strategy for targeting PS externalization, Peregrine Pharmaceuticals initially developed a mouse monoclonal antibody that targets PS by binding to two molecules of β₂-glycoprotein-I that in turn bind to PS on the surface of cells. With only a single molecule of the protein, binding to the surface of cells is relatively weak, but the affinity increases by a factor of 30,000 with the addition of a second molecule of β₂-glycoprotein-I. The researchers have constructed a series of antibodies that targets the tumor vasculature. Two are mouse antibodies, 3G4 and 2aG4. Bavituximab is a chimeric antibody with mouse PS binding regions fused to human IgG. The fully human antibody is known as PGN635.

These antibodies localize specifically to the blood vessels in tumors, which makes them useful as imaging agents. Studies with normal tissues and various tumor types reveal that these antibodies show affinity for a range of tumor types but minimal binding to a variety of normal tissues.

Bavituximab as an anti-tumor agent

In rodent studies, antibodies alone inhibit tumor growth by 30%–90%. That variation most likely reflects the differences in PS expression on blood vessels in different tumors, Thorpe said. However, combining the antibody treatment with other treatments, including chemotherapy, radiation, or androgen deprivation, improved these results. In studies of prostate tumors in mice, the combination of the mouse antibody 2aG4 and docetaxel slowed tumor growth by 98% compared with 90% for docetaxel alone and 80% for antibody alone. The PS-targeting antibody in combination with intense radiation can shrink lung cancer tumors that are resistant to radiation alone.

How does the antibody work? As one proposed mechanism of action, these PS-targeting antibodies damage tumor vessels by a process known as antibody-dependent cellular cytotoxicity. As poor blood flow and hypoxia lead cells in the tumor vasculature to expose PS on their surfaces, bavituximab targets those cells and shuts down blood flow, which starves those tumor cells of nutrients.

As a second mechanism, bavituximab can overcome the immunosuppressive processes observed with cells that externalize PS. In tumors the expression of PS on cell surfaces keeps macrophages in their M2 state, an activation state that suppresses inflammation and supports angiogenesis. It also prevents myeloid-derived suppressor cells (MDSCs) from differentiating into macrophages and dendritic cells (DCs). In addition, PS inhibits the maturation of DCs, which prevents the presentation of antigens to T cells. Bavituximab repolarizes macrophages to their M1 state, which allows them to kill cells that express PS, including the tumor vascular endothelium. It also leads to differentiation of the MDSCs and prompts DCs to mature and to present tumor antigens to T cells.

Tumor-associated macrophages have different functions depending on the environment. (Image courtesy of Philip Thorpe)

Tumor-specific immunity and clinical trials

Bavituximab can also prompt the development of tumor-specific immunity. In experiments with rats with treatment-resistant glioma tumors, 15% of the animals treated both with radiation and with antibody therapies did not die from those tumors. When the researchers reintroduced glioma tumor cells into these surviving rats, the rats were immune to the cancer. The team determined that spleen cells from these animals could kill gliomas, which indicated the antibody treatment had induced T cell immunity.

Bavituximab is currently being tested clinically. More than 400 people have been treated with the drug, which is well tolerated and appears not to show combined toxicity when used with chemotherapy, Thorpe said. A variety of single-arm, early phase II studies have been completed as a second line or first line treatment in people with advanced breast cancer and as a front line treatment for non-small cell lung cancers (NSCLC). Randomized phase II studies in NSCLC and pancreatic cancer are also underway.

Why do apoptotic cells have glycolytic enzymes on their surfaces? How is this process regulated? Could these proteins serve as therapeutic targets?

How many other proteins are involved in recognizing PS on the surfaces of viruses and mediating entry into host cells? Can researchers develop new antiviral agents based on an understanding of these processes?

Can annexin A5 and related proteins that bind to PS serve as therapeutics for cancer or inflammatory diseases?

Will bavituximab or related drugs show efficacy in large-scale clinical trials?