Use of Lentiviral Particles As a Cell Membrane-Based mFasL Delivery System for In Vivo Treatment of Inflammatory Arthritis

During budding, lentiviral particles (LVP) incorporate cell membrane proteins in the viral envelope. We explored the possibility of harnessing this process to generate LVP-expressing membrane proteins of therapeutic interest and studied the potential of these tools to treat different pathologies. Fas-mediated apoptosis is central to the maintenance of T cell homeostasis and prevention of autoimmune processes. We prepared LVP that express murine FasL on their surface. Our data indicate that mFasL-bearing LVP induce caspase 3 and 9 processing, cytochrome C release, and significantly more cell death than control LVP in vitro. This cytotoxicity is blocked by the caspase inhibitor Z-VAD. Analysis of the application of these reagents for the treatment of inflammatory arthritis in vivo suggests that FasL-expressing LVP could be useful for therapy in autoimmune diseases such as rheumatoid arthritis, where there is an excess of Fas-expressing activated T cells in the joint. LVP could be a vehicle not only for mFasL but also for other membrane-bound proteins that maintain their native conformation and might mediate biological activities.

controlled incorporation of specific membrane glycoproteins into LVP could be utilized to generate nanoparticles able to modulate immune system function.
The Fas (CD95)/FasL (CD178) receptor/ligand interaction is a key factor in the induction of programmed cell death (11). Fas/ FasL belong to the tumor necrosis factor (TNF)/TNF receptor (TNFR) family (12) and, like other members of the TNF family, biologically active FasL is a trimer (13). A FasL trimer recruits three Fas molecules, resulting in oligomerization of an intracellular death-inducing signaling complex (14) that through a cascade of several caspases triggers programmed cell death (11). This pathway is active during CTL-mediated killing of virus-infected or cancer cells (15) and in suppression/regulation of activated effector T cells (16,17), among other processes. Fas-mutant (lpr) and FasL-deficient (gld) mice suffer autoimmune pathology due to dysregulated lymphoproliferative processes, indicating the relevance of this mechanism in the regulation of immune system homeostasis (18,19).
Fas expression is observed in a variety of immune and nonimmune cells (19), whereas FasL at the cell membrane (mFasL) is expressed by CD8 + cytotoxic T cells as well as a proportion of activated CD4 + T cells and NK cells. mFasL is processed by metalloproteases to generate soluble FasL (sFasL) that although still able to bind Fas, is much less efficient in killing target cells (20), consistent with the idea that clustering of preassembled Fas trimers is required for optimal apoptosis induction (21).
Here, we generated non-infectious lentiviral expressing murine FasL (FasL-LVP) and tested their efficiency in specific killing of Fas-expressing cells. Exposure of Fas-expressing cells to the FasL-LVP triggered activation of caspase 3 and subsequent apoptosis, as shown by annexin V (AV)-propidium iodide (PI) incorporation and cell cycle analysis, with no crosslinking requirement. FasL-LVP injection in the footpad of mice with established collagen-induced arthritis caused a reduction in paw inflammation, of cells infiltrating the tissue, as well as clear improvement in the clinical score compared to controls. Our in vitro and in vivo results indicate that administration of non-infective, non-replicative LVP expressing FasL could be a novel method for local treatment of inflammatory diseases. This pharmaceutical formulation avoids the requirements for FasL purification that might affect its conformation and/or activity and reduces side effects due to crosslinking requirements with anti-FasL antibodies.

generation of gPi-linked Fasl (Fasl-gPi) constructs
To prepare the construct encoding murine FasL attached to the membrane via a glycosylphosphatidylinositol (GPI) anchor the following oligonucleotides: mFasLFor-AGAGTCGACGCCACCATGCAGCAGCCCATGAATTAC; mFasLRev-GCAAGC TTAGAGCTTATACAAGCCGAAAA AGG; FasLFor-AGAGTCGACGCCACC ATGCTGGGCATC TGGACCCTCCTACCTCTGGTTCTTACGTCTGTTGCTAC ACCCTC; mFasLGPIFor-ACGTCTGTTGCTACACCCTCTG AAAAAAAAGAGC; and mFasLGPIRev-GCAAGCTTGCCA CCAGAGCTTGAACTGAGCTTATACAAGCC GAAAAAGG. DAFGPIFor-GCAAGCTTCCAAATAAAGGAAGTGGAACC; DAFGPIRev-GCATGCGGCCGCTAAGTCAGCAAGCCC ATGGTTAC were used to amplify fragments of mFasL, murine DAF, and a linker sequence from mouse spleen cDNA, and then different combinations of these oligonucleotides were used to amplify overlapping fragments from the first round of PCR to generate a full length mFasL-GPI fragment that was cloned into the pBJ-Neo vector (23).

attachment of lVP to latex Beads
To evaluate FasL-expressing LVP, we used flow cytometry of LVP coupled to latex beads. Aldehyde/sulfate latex beads (4 µm, 4% w/v; A37304, Invitrogen, ThermoFisher Scientific, MA, USA) were sonicated (5 min, RT) and mixed in a 1:3 v/v ratio with LVP (15 min, RT). PBS (1 ml) was added, and the final suspension was incubated (60 min, 4°C) with continuous rocking. Free reactive groups were blocked using glycine (100 µl, 1 M, 30 min, RT). Beads coupled to LVP were washed with PBS/BSA 0.5%. For FACS analysis, the pellet was resuspended in PBS staining buffer and stained with specific antibodies as described above.
induction and assessment of cia Two-month-old DBA/1J mice were immunized intradermally at the tail base with an emulsion of chicken type II collagen (CII) in Freund's complete adjuvant (24). Arthritis was assessed daily by scoring each limb on a 0-4 scale, where 0 = normal, 1 = erythema and mild swelling confined to the tarsals or ankle joint, 2 = erythema and mild swelling extending from the ankle to the tarsals, 3 = erythema and moderate swelling extending from the ankle to metatarsal joints, and 4 = erythema and severe swelling encompass the ankle, foot and digits, or ankylosis of the limb, yielding a maximum score of 16 per mouse. Limbs with a score >2 (Cs0) were inoculated on day 0, 2, 4, 6, and 9 in the footpad with FasL-LVP or LVP (30 µg/ml, 50 µl) and the clinical score re-evaluated (Cst). As control, unaffected limbs were treated in parallel with FasL-LVP or LVP (30 µg/ml, 50 µl).
At the end of the experiment, the score was monitored, and the paws were removed, fixed in 4% formalin, decalcified with 10% EDTA, and paraffin-embedded. Sections (4 µm thick) were stained with hematoxylin and eosin or alkaline phosphatase.

enzyme-linked immunoassay
Microtiter plates were coated with chicken type II collagen (5 µg/ml; 90 min, 37°C). After blockade of remaining proteinbinding sites with 0.5% BSA in PBS, plates were incubated with serial dilutions of serum samples of immunized DBA/1J mice, followed by PO-labeled subclass-specific rabbit anti-mouse antisera. When needed, LVP-and FasL-LVP-coated microtiter plates (1 µg/ml) were incubated with serial dilutions of serum samples of mice injected with LVP in the footpad (time 0 and 14 days after treatment), followed by PO-rabbit anti-mouse IgG (Dako) antisera.

resUlTs lVP express Fasl at the Particle surface
We transiently transfected HEK293T cells with PAX2, pGEM-GFP and FasL-GPI-LVTHM/GFP-murine, or empty LVTHM/ GFP plasmids (control). Transfection efficiency was controlled by assessing GFP expression via fluorescence microscopy, Western blot and flow cytometry, whereas flow cytometry and Western blot analysis using anti-FasL mAb confirmed specific expression of FasL in cells transfected with FasL-GPI-LVTHM/GFP (Figures 1A,B). Flow cytometry analysis of LVP, isolated by centrifugation through a sucrose pellet, using specific mAb showed that FasL could be detected on the surface of LVP obtained from cells transfected with FasL-GPI (FasL-LVP) (Figure 1C, lower panel), but not on LVP obtained from HEK293T cells transfected with a control vector (LVP) (Figure 1C, upper panel). Both types of particles (FasL-LVP and LVP) expressed similar levels of CXCR4, a receptor endogenously expressed by HEK293T cells (Figure 1C). To confirm FasL expression, detergent extracts of FasL-LVP and LVP were evaluated by Western blot using specific anti-FasL mAb ( Figure 1D, upper panel). GFP expression in FasL-LVP and control LVP was analyzed as a loading control and was also used to normalize the number of control and FasL-LVP used in subsequent assays (Figure 1D, lower panel).

mFasl-lVP are Fully Functional In Vitro
Ba/F3 cells constitutively express Fas at the cell surface, which can be upregulated by IL-3 treatment (25) (Figure 2A, left). To test the ability of FasL-LVP to trigger cell apoptosis, Ba/F3 cells (10 6 cells/ml) were plated in RPMI medium supplemented with 5% FCS and IL-3, and exposed to serial dilutions of FasL-LVP, control LVP, or sFasL plus crosslinker as a positive control. Cell cycle analysis by PI incorporation and flow cytometry showed that treatment with FasL-LVP significantly increased cell death compared to control ( Figure 2B).
To test whether FasL-LVP treatment could also induce the death of primary cells, we used thymocytes, as they also express Fas at the cell surface (Figure 2A, right). Thymocytes isolated from 3-month-old C57BL/6 mice were treated with different amounts of FasL-LVP, LVP, or sFasL plus crosslinker for various times (2, 4, 6, 8, or 24 h) and analyzed as above. FasL-LVP significantly increased cell death compared to control LVP, with a maximum effect after 6 h treatment (Figure 2C). Pretreatment of cells with the pan-caspase inhibitor Z-VAD (R&D) blocked FasL-LVP-mediated apoptosis (Figure 2C), suggesting caspase involvement in the cell death process activated by exposure to FasL-LVP.
These results were confirmed by determining AV-PI incorporation in thymocytes, and Ba/F3 cells treated with FasL-LVP or control LVP. As an additional control of specificity, FasL-LVP did not trigger apoptosis of L1.2 cells (Figure 2D). Murine pre-B L1.2 cells do not express Fas at the cell surface, as shown by flow cytometry with anti-mCD95-PE mAb (Figure 2A).
FasL binding to Fas receptor leads to the formation of the DISC (death-inducing signaling complex), autoprocessing of the initiator procaspase 8 and activation of the effector  (c) Immunofluorescence analysis of cytochrome c release using specific antibodies and murine thymocytes treated with FasL-or control LVP for 6 h. One experiment is shown of five performed. procaspase-3, which triggers the extrinsic pathway of apoptosis (26) and/or cytochrome C release from mitochondria and caspase 9 activation (intrinsic pathway). Processing of procaspase 8 was not detected in any of these experiments (not shown); nevertheless, exposure to FasL-LVP specifically induced cleavage of caspase 3 in thymocytes incubated for 6 h with FasL-LVP ( Figure 3A). FasL-LVP-induced cleavage of caspase 3 was blocked when thymocytes were pretreated with the pancaspase inhibitor Z-VAD (Figure 3A), indicating the specificity of the effect. Cleavage of caspase 9 was also observed in these experiments ( Figure 3B). Moreover, rapid, specific release of cytochrome C, indicative of mitochondrial damage, was detected when thymocytes were treated with FasL-LVP, but not LVP controls (Figure 3C). Overall, these data showing involvement of both caspases 3 and 9 and mitochondrial damage in the cell death of thymocytes treated with FasL-LVP suggests that exposure to these particles triggers both the extrinsic and intrinsic pathways of apoptosis.

local administration of Fasl-lVP reduces Paw swelling in a Murine Model of arthritis
The collagen-induced arthritis model in DBA/1J mice was used to assess the potential therapeutic utility of FasL-LVP administration, since Fas expression is enhanced on activated lymphocytes and the immunopathogenesis in this model involves both T-cell and B-cell-specific responses to collagen (27). Arthritis was induced by subcutaneous injection of collagen type 2 into 2-month-old DBA/1J mice. On appearance of the first signs of CIA (score ≥2, day 0), arthritic mice received four injections on days 0, 2, 4, and 6 of either FasL-LVP or LVP (30 µg/ml, 50 µl) in the inflamed paw (left or right limbs, respectively). While disease progressed in LVP-treated mice, a clear reduction in paw swelling (p < 0.01) (Figure 4A) was observed in FasL-LVP-treated animals. This effect was observed as early as 2 days after initiation of treatment and reached maximum at 14 days after the initial inoculation. Treatment with neither FasL-LVP nor LVP promoted any changes in non-inflamed paws. The reduced severity of disease in FasL-LVP-treated mice was not associated with changes in anticollagen type II antibody levels (Figure 4B), suggesting that the effect is confined mainly to the administration site. Histological analysis of paws from both FasL-LVP-and LVP-treated mice showed that the inflammation and pannus parameters differed clearly between these groups. Similarly, the damage observed in cartilage and bone in control mice was absent in those treated with FasL-LVP ( Figure 4C). We nonetheless detected a small but sustained immune response to LVP in treated mice ( Figure 4D). The use of the human HEK293T cell line to prepare the LVP might underlie this side effect.

DiscUssiOn
As they bud from the membrane of the cell, LVP incorporate components of the plasma membrane, including embedded proteins in their lipid environment, into the virus particle (28). Here, we show that this property can be exploited to develop virus-like particles that can subsequently modulate cell activation both in vitro and in vivo.
The Fas molecule belongs to a subset of the TNFR family called death receptors, which can trigger apoptosis by inducing caspase activation. Ligation of Fas triggers apoptosis in activated T cells; this mechanism is involved, for example, in regulation of the clonally expanded T cell population after antigen clearance (11). Fas-mediated apoptosis can also be activated in other, non-immune, cell types including the death of neurons after withdrawal of growth factors, of fibroblasts expressing c-Myc, of thyroid cells and even of endothelial cells (29)(30)(31)(32), which suggests that Fas can be considered a target for therapeutic intervention in several diseases.
We engineered and transiently expressed a chimeric FasL-GPI in HEK293T cells and used them to produce LVP that express FasL, but lack viral envelope proteins. Replacement of C-terminal transmembrane domains by GPI-addition signal peptides allows expression of the proteins on the plasma membrane and provides stable association of the construct with the lipid bilayer (33,34). Plasmids encoding viral envelope proteins, i.e., VSV-G (glycoprotein of the vesicular stomatitis virus), were not used in the viral particle preparation since VSV-G expression facilitates virus dissemination, promoting a range of non-specific changes that could obscure the specific effects of treatment with FasL (35).
In vitro, the virus-like particles induced cell death in Faspositive cells, including the B cell line, BAF/3 and primary thymocytes. The absence of apoptosis in L1.2 cells, which do not express cell surface Fas, and in cells treated with control LVP showed that this effect is specific; it indicated that FasL-LVP-induced cell death is mediated by the FasL expressed on the particles and not by other components expressed at the membrane or by other LVP components. In this context, when VSV-G pseudotyped LVP were used in similar experiments, increased cell death was observed independently of FasL, probably due to the ability of these LVP to infect cells (36).
FADD-mediated activation of the proteolytic activity of caspase 8 is essential for Fas-induced apoptosis in many cell types (37)(38)(39); nonetheless, caspase 8 cleavage was not observed in our experiments. We demonstrated processing of caspase 3 and caspase 9 as well as cytochrome C release from mitochondria, which suggests FasL-mediated activation of both the extrinsic and intrinsic apoptosis pathways. It is possible that in our experimental conditions, modest activation of caspase 8 was undetected; indeed, a Fas signaling pathway involving slowly activated caspase 8 and mitochondrial damage has been reported (40). This pathway involves lower levels of activated caspase 8 which are still sufficient to process the pro-apoptotic protein Bid. Translocation of Bid to the mitochondrial membrane induces oligomerization of Bcl-2, Bak, and Bax, which promote cytochrome release and finally caspase 9 activation (41). Nevertheless, the possibility of FasL-LVP-mediated activation of a caspase 8-independent apoptosis mechanism cannot be excluded. Caspase-independent activation of JNK leading to cell death has been reported (42), as has Fas-mediated cell death that is not blocked by specific caspase 8 inhibitors (43). New experiments are required to clarify this issue. Whatever the mechanism, the data presented here clearly show that LVP displaying FasL can induce apoptosis in Fas-expressing cell lines and primary cells in vitro.
Although apoptotic cells are rarely observed in RA tissues in vivo (44), probably due to intra-and extracellular antiapoptotic processes (45,46), high Fas/FasL levels are present in the synovium of RA patients (44). Since apoptosis induced by anti-Fas antibody or by gene transfer of FasL ameliorates arthritis in the experimental murine model of collagen-induced arthritis, probably via the induction of apoptosis of T cells, macrophages, and synoviocytes (47)(48)(49), we tested the use of our LVP as therapy in this model. Local injection of FasL-LVP into the inflamed footpad specifically reduced paw inflammation. This effect was restricted to the injected paw, as no improvement of inflammation was noted in the contralateral hind paw that received the injection of control LVP. In addition, no effect was observed on anti-collagen type II antibody levels. These data indicate that the injected LVP did not disseminate outside the injection site. Histological analysis reflected a significant reduction of cell infiltration in joints treated with FasL-LVP compared to controls. The simplest interpretation of these experiments is that the FasL-LVP induces death of the activated (Fas-expressing) lymphocytes and macrophages that mediate the autoimmune process. We also detected low anti-LVP antibody levels in the sera of treated mice, which indicated that LVP treatment might activate a slight immune response to the viral particles, which could decrease treatment efficiency. The use of human HEK293T cells to generate the LVP for use in a murine system might underlie this effect. LVP generation in compatible cells should reduce this side effect. LVP have been widely used for human gene therapy for many years, and several strategies have been developed to avoid, suppress, or manipulate the immune response and promote immune tolerance to the LVP (50).
The use of FasL-LVP as a therapeutic strategy has a number of advantages compared to the use of sFasL. It is suggested that nonapoptotic, Fas-mediated signals can promote chronic inflammatory arthritis. For instance, DBA1/lpr mice, which are defective in Fas signaling, are less susceptible to CIA (51). Indeed, sFasL is increased in synoviocytes of RA patients compared to those suffering from OA, and sFasL levels are higher in patients with severe RA compared to those with mild disease, indicating that sFasL may be an exacerbating factor in RA (52). sFasL, cleaved from cell membranes by metalloproteases, contributes to Fas downregulation, thus reducing Fas-mediated apoptosis in RA synovial cells and perpetuating the disease (53). The construct used in the preparation of FasL-LVP lacks the proteolysis site found in membrane FasL, which minimizes the possibility of sFasL production. Another advantage of this construct is that the intrinsic structure of the FasL displayed on the virus-like particle is likely very similar to that on the cell membrane, rendering crosslinking with antibodies unnecessary. The use of monoclonal antibodies to crosslink sFasL is associated with liver toxicity (54), whereas a synthetic human hexameric FasL (APO010, Topotarget) promotes human glioma cell death and prolongs survival in tumor-bearing mice with no need for crosslinkers (55). In vivo, Fas expression is enhanced after lymphocyte activation; we also demonstrated that FasL-LVP, presumably by inducing the death of activated cells, can be used to treat autoimmunity. Our in vitro and in vivo results indicate that the administration of non-infective, non-replicative LVP expressing FasL could be a novel method for local treatment of inflammatory diseases.

eThics sTaTeMenT
All animal experiments were performed in accordance with national and European regulations and were approved by the Animal Experimentation Committees at the CSIC and the Dirección General de Medio Ambiente de la Comunidad de Madrid (PROEX 188/14).