6.2.1.1 Exogenously Added Cholesterol Incorporates into Membranes

Cholesterol-tethering of biomolecules has been used to paint the surface of cells. Exchange between cholesterol derivatives found in lipid suspensions with those in erythrocyte ghost membranes have been detected upon incubation for 24 h [10,11]. In order to further investigate the “spur” red cell phenomenon in which red blood cells from patients with spur cell anemia uptake cholesterol found in the serum, Cooper et al. [12] attempted to artificially reproduce the environment. They found that when phospholipids were sonicated with cholesterol, the cholesterol-to-phospholipid (C/P) molar ratio affected uptake of cholesterol in red cells that were incubated with the mixture. When the C/P ratio was less than 1.0, cholesterol was depleted from the red cells and when the C/P ratio was equal to 1.0, no change in cholesterol content in red cells was seen. However, when the C/P ratio was greater than 1.0, cholesterol was deposited into normal human red cell membranes and the spur cell phenomenon was detected [10–12]. As the C/P ratio was increased, increased incorporation of cholesterol in red cells was also detected in a dose-dependent manner over 24 h [12]. These results show that cholesterol can be exogenously added to incorporate into cell membranes.

6.2.1.2 Cholesterol-Tethering of Biomolecules to Membrane Surfaces

Due to the ease with which exogenously added cholesterol can incorporate into cell membranes, a method of membrane painting to modify the cell surface with biomolecules was established (Figure 6.1A). This method involves the use of cholesterol derivatives, such as N-(3β)-cholesterylglycine, that are tethered to biomolecules [13]. The addition of amino groups on cholesterol derivatives that decrease the cholesterol–phospholipid affinity in the donor membranes leads to enhanced incorporation of the cholesterol derivative in the accepting cell membrane vesicles [14]. Conjugation of a biomolecule, such as biotin or epitopes for anti-HA or anti-FLAG antibodies, to the cholesterol derivatives creates an amphiphilic complex. The amphiphilic nature allows for insertion of the cholesterol derivative of the biomolecule into lipid rafts, whereas the polar segment of the biomolecule remains exposed on the cell surface. Upon incubation of the cholesterol-tethered molecules for 1 h at 37°C in RPMI medium with 10% fetal bovine serum (FBS), stable dose-dependent incorporation is detected within lipid rafts of cellular membranes [13].

6.2.1.3 Application of Cholesterol-Tethering to Create Synthetic Receptors

Modification of cell surfaces by cholesterol-tethering of biomolecules allows for delivery of drugs or antitumor agents into target cells. This delivery system occurs when a cholesterol-tethered receptor bound to its ligand is internalized. Beforehand, a cholesterol derivative is tethered to the biomolecule receptor, which is then allowed to incorporate into lipid rafts of cells for 1 h. When the surrogate receptor (SR)-incorporated cells are incubated with ligand, nearly 100% of the cells bind to the ligand in a matter of 10 min. Upon receptor–ligand binding, internalization of the receptor–ligand complex was initiated, and as the concentration of SR added to the cells increased, the amount of internalization also increased upon ligand addition in a dose-dependent manner [13]. Furthermore, Hussey et al. [15] showed that internalization occurred partially within 10 min of incubation and completely within 4 h of incubation through a clathrin-mediated endocytosis pathway.

The length of the tethering linker between the cholesterol derivative and the biomolecule is imperative [16]. A longer linker allows for full insertion of the cholesterol derivative into the membrane, whereas the shorter linker prevents this. An 11-atom linker resulted in membrane insertion of a cholesterol derivative bound to biotin that was capable of internalizing its ligand, streptavidin, within 10 min [16].

Using this approach of creating synthetic cell surface receptors, Martin and Peterson [13] showed that tethering an HA epitope to a cholesterol derivative allowed for surface expression of the epitope on Jurkat cells. An anti-HA antibody complexed with a protein A-β-galactosidase conjugate bound the surface expressed HA epitope resulting in engulfment of the protein A-β-galactosidase complex. Furthermore, when the substrate to β-galactosidase was added to these Jurkat cells, fluorescence was detected intracellularly within 75% of the cells. These experiments suggest that not only can cholesterol-derived SRs be used to modify the cell surface, but also the added protein remains functionally active even when endocytosed within the cell. Similar binding and internalization studies were conducted with cholesterol-derived SRs bound to either FLAG epitopes or streptavidin ligands, such as biotin [13,16].

Cholesterol-derived SRs can be added to any cells that contain lipid rafts. Many multidrug-resistant tumors have been shown to produce increased levels of lipid raft components [16]. Therefore, the addition of cholesterol-derived SRs to these multidrug-resistant cancer cells via intratumoral injections should readily lead to the uptake and expression of the receptor on the tumor cell surfaces that contain increased lipid raft components. In this manner, many drugs or toxins can be preferentially targeted to these cancer cells as opposed to neighboring healthy cells that have fewer lipid raft components in comparison.

6.2.1.4 Advantages/Disadvantages of Cholesterol-Tethering

Cholesterol-tethering of biomolecules onto cell membrane surfaces provides for a simple method to express proteins on cell surfaces due to the ease with which cholesterol and cholesterol derivatives incorporate into the cell membranes. Since lipid rafts are commonly found on cells, albeit at different levels, incorporation of cholesterol-tethered biomolecules could occur on almost any cell surface. Through the production of SRs, this method can also lead to the engulfment of high molecular weight proteins, such as enzymes like A-β-galactosidase, toxins, or drugs, that remain functionally active within the cells, allowing for cell targeting via this mechanism [13,16].

However, the production of these cholesterol-tethered biomolecules is not an easy task; it involves many chemical additions and the linkage between the biomolecule and cholesterol derivative is vital for incorporation and ligand binding. Moreover, there is a possibility that lipids present in circulation may bind to these administered conjugates and block incorporation onto the defined target cells.

6.2.2.1 Palmitolyation to Create Surrogate Cell Surface Receptors

Originally, artificial receptors were incorporated onto B cells by palmitoylating antibodies. Peacock et al. [20] showed that antibodies can be palmitoylated and subsequently inserted into a cell membrane. These membrane-bound antibodies can act as receptors, and upon binding to target antigen can induce an immune response by delivering signals to the decorated cells. In the study, protein 315, a secreted antibody against dinitrophenyl (DNP) lacking the Fc region, was palmitoylated before being incubated with murine splenocytes at 37°C. The palmitoylated antibody spontaneously incorporated onto splenocyte cell membranes in cell media, specifically via the palmitate groups since nonpalmitoylated proteins did not incorporate. Incorporation of the palmitoylated antibody on cell membranes was concentration and time dependent, and the incorporated palmitoylated antibody was still able to bind the DNP antigen and induce a polyclonal antibody response specific to DNP [20]. However, the degree of palmitoylation was also important for functional antigen binding. Increasing palmitoylation led to decreased ability of the incorporated antibody to bind antigen [20]. Furthermore, through fluorescence photobleaching recovery (FPR), palmitoylated-antibody receptors that incorporated exogenously onto B cells were found to have similar membrane properties as surface immunoglobulin (Ig) expressed naturally on B cells [23]. Therefore, exogenously decorating B cells with palmitoylated antibodies could provide insight into how surface antibodies are able to exhibit an antigen-dependent immune response.

This antibody-anchoring technology was further developed to create “SRs” on the cell surface. The establishment of SRs allows for the creation of “custom-designed killer cells” in which palmitoylated antibodies incorporated onto cell membranes provide specificity for target cells [24]. Either palmitoylated F(ab′)₂ against mouse Ig or intact anti-TNP antibody was incorporated onto the membrane of nylon wool nonadherent spleen cells from mice. The antimouse Ig decorated cells were then incubated with syngeneic B cells stimulated with LPS, while the spleen cells decorated with anti-TNP antibody were incubated with TNP-modified EL4 T-lymphoma cells. Both sets of decorated cells demonstrated the ability to lyse their targets, while their undecorated counterparts did not. Furthermore, neither set of decorated cells exhibited the ability to lyse cells not displaying the antigen specific for their SR, demonstrating the specificity of the SR. Of particular interest, this SR-mediated cellular cytotoxicity (SRMCC) occurred even when the SR did not have an Fc region, suggesting that SRMCC is independent of Fc receptor interactions [25]. Although SRMCC was not detected when NK cells were depleted by anti-asialo GM₁, SRMCC could induce lysis of NK-resistant cells. This suggests that NK cells, while not the sole effector cells, play a key role in SRMCC [24]. Thus, incorporation of palmitoylated SRs onto cells provides a means to engineer killer cell surfaces to express receptors that can specifically target cells expressing particular ligands.

6.2.3.1 Expression of Proteins on the Cell Surface by Protein Transfer of GPI-APs

Purified GPI-APs contain the unique property of being able to reincorporate into cell membranes and still retain their biological activity. In 1984, Medof et al. [32] showed that the purified GPI-AP, DAF, was able to reincorporate into erythrocyte membranes by simply incubating purified DAF with erythrocytes. This reconstitution of DAF prevented complement protein C4b2a activity on the erythrocytes in a DAF dose-dependent manner. Furthermore, in the acquired disorder paroxysmal nocturnal hemoglobinuria (PNH), the absence of DAF in erythrocytes from PNH patients leads the erythrocytes to be susceptible to complement-induced lysis. However, this lysis is not detected when exogenous DAF is reconstituted back into erythrocyte membranes [33]. Additionally, in PNH patients, another complement regulatory protein, homologous restriction factor (HRF), later known as CD59, is also absent from erythrocyte membranes. Incubation of purified radiolabeled HRF with erythrocytes led to insertion of HRF into erythrocyte membranes and the inserted HRF remained biologically functional in preventing complement lysis by C5b-9 in vitro [34–38]. Oligodendrocytes that also lacked CD59 became resistant to complement-induced lysis after exogenously incorporating purified CD59 into their membranes as well [39]. Another complement regulatory protein which behaved similarly to DAF and CD59 is C8-binding protein (C8bp). In a comparable manner, C8bp, which is also not found in erythrocytes obtained from PNH patients, can be inserted into erythrocyte membranes exogenously after purification of the protein and again inhibit complement-induced lysis of erythrocytes [40].

Additionally, adhesion molecules such as lymphocyte function-associated antigen 3 (LFA-3) are also found to be deficient on erythrocytes obtained from PNH patients [41]. LFA-3 is expressed as a GPI-anchored form in erythrocytes, whereas nucleated cells express both GPI-anchored forms and polypeptide anchored forms [42]. Consequently, partial deficiency of LFA-3 was observed in nucleated cells from PNH patients. The interaction between LFA-3 on erythrocytes and its receptor CD2 on T cells induces formation of T cell rosettes [43]; however, erythrocytes from PNH patients did not form rosettes upon incubation with CD2 expressing Jurkat cells. Interestingly, when exogenous purified GPI-anchored LFA-3 was added back to erythrocytes from PNH patients in vitro, rosettes were formed, suggesting that protein transfer of exogenous LFA-3 leads to stable as well as functional LFA-3 expression on the cell surface [41]. Also, exogenous incorporation of purified CD16B (Fcγ receptor III), another naturally occurring GPI-AP, onto both a human T cell leukemia and B lymphoblastoid cell lines in vitro demonstrated that protein transferred CD16B on nucleated cells can still retain its function and mediate endocytosis of its ligand [44]. Another GPI-AP, Thy-1, was able to exogenously incorporate onto murine Thy-1-negative cells in vitro. After this protein transfer, staining of the exogenously incorporated Thy-1 was similar to staining patterns of endogenously expressed Thy-1 and the lateral mobility of both endogenously and exogenously inserted Thy-1 was similar as determined by fluorescence recovery after photobleaching (FRAP) [45]. Furthermore, even nonmammalian GPI-APs, such as the membrane form of variant surface glycoprotein (mfVSG) expressed on the parasite Trypanosoma brucei, can also be purified and exogenously incorporated onto cell surfaces with similar lateral mobility as endogenously expressed mfVSG [46]. However, PI-PLC treatment of purified GPI-APs decreases the insertion levels of these proteins onto cell membranes suggesting that protein transfer is due to the presence of the GPI-anchor [36,41,45]. All of the above examples show that purified GPI-APs can be incorporated via the GPI-anchor into cell membranes by simply exogenously adding the purified GPI-APs and that after incorporation onto cell membranes, the biological function and lateral mobility of the protein remains similar to that of the constitutively expressed protein counterpart.

6.2.3.2 Characteristics of GPI-AP Protein Transfer

Protein transfer of GPI-APs onto cell membranes (Figure 6.1C) occurs in a concentration-dependent manner. As higher concentrations of GPI-APs are added to the membranes, increased incorporation occurs [32,44,47]. However, a saturation point is expected to occur in which all GPI-AP sites on the membrane will be occupied and no more GPI-APs can incorporate by protein transfer without disrupting the integrity of the membranes. GPI-anchor-mediated protein transfer also results in surface expression on 100% of the cells incubated, as shown by flow cytometry [48]. Furthermore, GPI-anchor-mediated protein transfer is time dependent. As the time for protein transfer is increased, increased expression of GPI-APs on membrane surfaces occurs. Incorporation was detected in as little as 20 min postincubation of GPI-APs with membranes and incorporation levels increased with incubation time for up to at least 4 h [32,44,47]. Medof et al. [32] noticed that incorporation of radiolabeled DAF into erythrocyte membranes occurred best at 37°C followed by 20°C with no detectable DAF incorporation occurring at 0°C. Poloso et al. [47] detected a similar phenomenon in which protein transfer occurs best on cell membrane vesicles at 37°C followed by 25°C followed by 4°C. These observations suggest the necessity of membrane mobility for the incorporation of GPI-APs. Plasma membranes are most fluid at 37°C compared to lower temperatures; therefore, the dynamics between GPI-AP micelles and the lipid bilayer occurs more readily at 37°C [49].

Optimum incorporation also occurs without the presence of certain serum proteins. Incorporation of GPI-APs is inhibited by the presence of human serum lipoproteins, such as HDL and LDL, and fatty-acid-binding proteins, such as bovine serum albumin (BSA) and orosomucoid [32,41,44]. High concentrations of ovalbumin on the other hand do not affect incorporation of GPI-APs onto cell membranes [32,44]. Once incorporated onto the cell surface by protein transfer, GPI-APs remain stably integrated into the cell membrane of erythrocytes and cannot be removed by washing even with high salt concentrations [32]. After protein transfer, GPI-APs can however be removed by solubilization with detergents, such as NP-40. However, incorporation onto live or irradiated nucleated cells leads to loss of incorporated proteins under culture conditions [48].

Protein transfer leads to the incorporation of exogenously added GPI-APs into lipid rafts of cell membranes [8,50]. After Triton X-100 lysis of cells that have been protein transferred with GPI-APs, the GPI-APs are detected in detergent insoluble microdomains as are other lipid raft specific proteins [50]. Furthermore, enhanced incorporation of green fluorescent protein (GFP)-labeled GPI-anchors into lipid rafts was also associated with the presence of lipid raft-associated molecules (cholesterol, sphingomyelin, and dipalmitoyl-phosphatidylethanolamine) further proving that protein-transferred GPI-APs associate with lipid raft structures [8]. Upon initial incorporation of GPI-APs onto cell membranes by protein transfer, the GPI-APs are widely distributed throughout the cell membrane as observed by confocal microscopy [51]. However, hours after incorporation, the GPI-APs localize into lipid raft microdomains as detected by confocal microscopy or by observing association of GPI-APs with lipid raft domains after lysis with Triton X-100 [50,51].

Similar to the erythrocyte-incorporated proteins described earlier, these protein-transferred GPI-APs expressed on nucleated cells and cell membranes retain their functional biological activity in comparison to endogenously expressed forms of the proteins [44,47,48,52]. Not only do exogenously incorporated GPI-APs retain the functional extracellular activity of the GPI-AP, but they can also retain functional intracellular activity [50,51]. Exogenously incorporated GPI-APs onto cells led to association of the GPI-AP with active intracellular kinases [50,51]. Upon antibody crosslinking of exogenously incorporated GPI-B7-1, DAF, or CD59, phosphorylation induced by active kinase activity was detected [50,51] as well as an induced transient Ca²⁺ flux signal [51].

It has also been observed that cells genetically transfected to express the GPI-anchored form of the cytokine, GM-CSF, not only expressed functional surface-bound GM-CSF that was susceptible to PI-PLC cleavage, but GM-CSF was also detected in the supernatant in a soluble form. This soluble GM-CSF was released by either shedding or proteolytic cleavage and was not released by secretion [53]. Similarly, IL-2 was detected in the supernatant of cultured cell lines transfected to express GPI-IL-2; however GPI-IL-12 was not secreted when expressed on CHO cells [54,55]. It is not clear as to how only IL-2 was released; however, proteins differ in protease cleavage sites and also may differ in shedding capabilities. Therefore, this dual ability of cytokines to not only be expressed on cell surfaces by a GPI-anchor but also to be shed into the extracellular space could lead to a cytokine gradient allowing for the infiltration and activation of immune cells.

More importantly, after protein transfer of GPI-APs onto the lipid bilayers, the newly incorporated GPI-APs remain functionally active. Although it has been stated that naturally occurring GPI-APs can be reincorporated exogenously by protein transfer and retain their biological functional activity, even recombinant GPI-APs, such as GPI-B7-1 and GPI-IL-12, after protein transfer retain their biological functional activity [47,48,52,56,57]. Protein-transferred GPI-B7-1 and GPI-IL-12 incorporated on tumor cells and tumor membrane vesicles (TMVs) retained their ability to induce T cell proliferation.

6.2.3.3 Advantages and Limitations of GPI-AP Protein Transfer

Modification of membranes and cell surfaces provides a wide range of application possibilities. Protein transfer allows for membrane modification in an easy, quick, and efficient manner compared to the traditional genetic engineering approaches. (1) With protein transfer, purified GPI-APs incorporate onto cell membranes in a matter of minutes with maximal incorporation occurring around 2–4 h [47], whereas genetic engineering approaches take much longer from the order of days to months to establish cultured transfected cell lines. (2) Genetic engineering makes modifying some cell lines extremely difficult especially when primary cell lines, such as bone marrow progenitors and primary cultures, are hard to transfect. Also, protein transfer can be performed on small cell numbers and does not require cell division to occur. (3) Using protein transfer, the amount of incorporation can be finely controlled by adjusting the amount of exogenously added GPI-APs. (4) Finally, multiple GPI-APs can be incorporated onto the same cell membranes by protein transfer, whereas double and triple transfections by genetic engineering are more difficult to attain [58].

However, membrane anchors are known to affect the functionality of external domains of some proteins [59]. It has been noted that PI-PLC treatment of some GPI-APs renders some antibodies nonresponsive to the protein even if the antibody binds the GPI-AP in the presence of the GPI-anchor [60–62]. Therefore, the addition of a GPI-anchor can affect the structural conformation and function of proteins. Also, if important antibody epitopes are found at the C-terminal end of a protein, the addition of the GPI-anchor at the C-terminus may alter the antigenicity of the protein. Similarly, the GPI-anchor attaches the protein close to the membrane with less than 10–14 Å space between the protein and the membrane surface as determined by fluorescence resonance energy transfer (FRET) [62,63]. This close proximity to the membrane surface may alter accessibility of important antibody-binding sites. Therefore, it is imperative that the functionality of each GPI-AP be tested before use. A bottleneck of the protein transfer method using GPI-APs is the purification of GPI-APs. Purification using immunoaffinity chromatography is the best defined manner for obtaining pure fractions of GPI-APs; however, optimization of the process for each individual GPI-AP is time-consuming. Despite these disadvantages, when optimal purification of GPI-APs that retain their biological functionality is attainable, protein-transferred GPI-APs provide an efficient method to controllably modify membrane surfaces.

6.2.3.4 Protein Transfer of Antigen-Presenting Cells with GPI-Anchored MHC Complexes

During viral infections, CTL responses are important to combat infected cells. For a CTL to become activated, the T cell has to be engaged with MHC I that expresses the viral peptide; however, many viruses downmodulate the expression of MHC molecules on the cell surface to escape from the immune system. Therefore, Huang et al. [64] constituted a method to control the amount of MHC–peptide complexes expressed on a cell by protein transfer. Their model system utilized a HLA-A2.1 MHC complexed with a hepatitis B virus (HBV) peptide that induces a dominant T cell response. They constructed a GPI-anchored form of HLA-A2.1 by attaching a GPI-anchored signal sequence to the heavy chain and contransfected Schneider S2 Drosophila melanogaster cells with the beta 2 microglobulin gene. Upon purification of this HLA-A2:GPI, they were able to show that the MHC complex incorporated into HLA-A-negative and HLA-B-reduced cells in a time, temperature, and concentration-dependent manner. They were able to detect incorporation by flow cytometry within 1 min of incubation with cells; however, maximal incorporation occurred after 1 h incubation at 37°C. Additionally, incorporation was inhibited by the presence of 10% FBS, but if FBS was added after protein transfer, then no affect on cell surface expression of the MHC complex was noticed. Peptide was loaded onto the GPI-MHC molecules at 4°C overnight before protein transfer leading to stable expression of the MHC–peptide complex and MHC–peptide specific cell lysis by CTLs after protein transfer. These results showed that protein transfer with GPI-anchored MHC complexes could be used to express desired quantities of specific MHC complexes on cell surfaces. Thus, antigen-presenting cells (APCs) can also be decorated with desired MHC–peptide complexes allowing further studies of APC and T cell interactions to elucidate specific peptide-mediated T cell activation and tolerance.

6.2.3.5 Protein Transfer of GPI-APs in Tumor Immunotherapy

Another application of protein transfer using GPI-APs is in tumor immunotherapy. Cancer cells have the ability to continuously evade the immune system; therefore, finding a way to activate the immune system against these cells is key to combating tumors. One method in which tumor cells evade the immune system is through the downregulation of costimulatory molecules, such as B7-1, on tumor cell surfaces. This downregulation renders tumor-specific T cells anergic and nonresponsive to the tumor. Therefore, tumor cells have been transfected to express B7-1 in order to improve the immunogenicity of cell-based therapeutic vaccines. However, in order to transfect tumor cells with immune stimulatory molecules, the establishment of primary tumor cells lines from patients’ tumor tissue is required. However, tumor cells derived from patients are extremely difficult to culture; therefore, McHugh et al. [52] proposed the use of protein transfer to express GPI-anchored B7-1 on tumor cells or on TMVs. The TMVs that are derived from the homogenization of tumor cells still contain tumor-associated antigens along with a lipid bilayer, thus they can be modified with GPI-APs by protein transfer. It was shown that TMVs derived from EG7 tumors, a murine T cell lymphoma line, could be protein transferred to express GPI-B7-1 on the surface. They found that these protein-transfer-modified TMVs could protect against challenge with EG7 tumor cells. Furthermore, B7-1-modified TMVs led to T cell proliferation and cytolytic T cell activity in vaccinated mice ex vivo against the parental tumor [52].

Additionally, protein transfer was also shown to be conducted on TMVs derived from surgically removed human tumor tissue. These TMVs were stably incorporated with GPI-B7-1 by protein transfer in a concentration-, time-, and temperature-dependent manner. These incorporated TMVs were functionally active in stimulating T cell proliferation [47].

Many cytokine therapies, such as the administration of the cytokine IL-2, have been used as a tumor immunotherapy in order to activate immune cells against the tumor [65]. Although cytokine administration has worked in activating the immune system, systemic toxicity also results which impedes the success of the treatment. Therefore, human TMVs modified by protein transfer to express GPI-cytokines were used as a method of delivery of not only the cytokine in a less toxic manner, but also tumor-associated antigens [47,57]. Using protein transfer, purified GPI-anchored IL-12 was incorporated onto human and murine TMVs from mammary carcinomas, renal cell carcinomas, melanoma, and B cell lymphoma cultured cell lines, and the resulting TMVs expressed functionally active IL-12 that could induce T cell proliferation and IFN-γ secretion. IL-12 also worked well in combination with CD80 or CD40 to enhance T cell proliferation and IFN-γ secretion [57].

Taken together, protein transfer of costimulatory molecules and cytokines that have been converted into GPI-anchored forms has led to a promising approach for inducing antitumor immunity. Because GPI-anchored cytokines are attached to the membrane surface, a slow-release depot of the cytokine is ensured circumventing the systemic toxicity associated with administration of soluble cytokines.