Cell Type Preference of a Novel Human Derived Cell-Permeable Peptide dNP2 and TAT in Murine Splenic Immune Cells

Cell-permeable peptides (CPPs) have been widely studied as an attractive drug delivery system to deliver therapeutic macromolecules such as DNA, RNA, and protein into cells. However, its clinical application is still limited and controversial due to the lack of a complete understanding of delivery efficiency in target cells. Previously we identified and characterized the novel and superior CPP, named dNP2, and here we comparatively analyzed intracellular delivery efficiency of dNP2 and TAT in various immune cells of mouse spleen to demonstrate their cell type preference. dNP2- or TAT-conjugated fluorescent proteins were most efficiently taken up by phagocytic cells such as dendritic cells and macrophages while little protein uptake was seen by lymphocytes including T cells, B cells, and NK cells. Interestingly CD8+ lymphoid dendritic cells and CD62LloCD44hi memory like T cell subsets showed significantly better uptake efficiency in vitro and in vivo relative to other dendritic cells or T cells, respectively. In addition, activated macrophages, T cells, and B cells took up the proteins more efficiently relative to when in the resting state. Importantly, only dNP2, not TAT, shows significant intracellular protein delivery efficiency in vivo. Collectively, this study provides important information regarding heterogeneous intracellular delivery efficiency of CPPs such as dNP2 and TAT with cell type preference in the spleen needed for its application in phagocytic cells or activated immune cells.


Introduction
Cell permeable peptides (CPPs), either identified from various native proteins or designed as synthetic peptides, are being utilized to deliver therapeutic macromolecules into cells in vitro and in vivo [1,2]. Since the TAT protein from HIV was found to localize to the nucleus and cytoplasm of cultured cells without the need for transfection reagents [3], the intracellular delivery of proteins by cell permeable peptides has been intensively studied. Arginine or lysine rich cationic peptides such as TAT [4], Antp [5], VP22 [6] and R9 [7] have been used over the last few decades to deliver various macromolecular cargos including siRNA [8,9], oligonucleotides [10], peptides [11], and transcription factors [12,13] to alter cellular behavior and modulate disease pathogenesis as a macromolecular therapeutics. Cationic CPPs bind to negatively charged cell membrane molecules such as proteoglycans on glycoproteins or negatively charged membrane lipid molecules and trigger endocytosis by the cells to uptake the CPPcargo complex [14,15]. Recently, human protein derived CPPs have been identified such as VectoCell [16], Lactoferin [17], Hph-1 [18], Sim-2 [11], LPIN [19], 2IL-1a [20] to minimize possible immunogenicity and toxicity in vivo. However, it has been demonstrated that almost all cells are permeable to CPPs without specificity, which remains one of the major challenges to the use of CPPs as an advantageous drug delivery tool [2]. Although a significant number of clinical trials have been performed using TAT, there are as of yet no available FDA approved CPP-drugs for any human disease. Here, we hypothesized that the understanding of heterogeneous delivery efficiency by a CPP would increase the success of CPP drug development.
Recently, we identified the novel human-derived CPP named dNP2 that was efficiently delivered into primary immune cells including T cells and was even localized to brain tissue through the blood-brain barrier (BBB) in mice [21]. dNP2 contains amphiphilic amino acids with basic sequences and hydrophobic amino acids with helix-loop-helix structure as prediction that seem to enhance cellular uptake by endocytosis. To better use this attractive CPP for valuable drug development, the preferential targeting of dNP2 to specific cells should be investigated. Therefore, the present study was designed to determine the cell type preference of dNP2 and TAT in various immune cells of mouse spleen. Enhanced green fluorescent proteins (EGFPs) fused to CPPs were used to treat splenocytes in vitro and flow cytometric analyses defined the specific cell types targeted by each CPP protein. In addition, intravenous administration of dNP2-EGFP and TAT-EGFP proteins in mice confirmed the in vivo relevance of heterogeneous delivery efficiency among the immune cells. Here we found that phagocytic cells such as DCs and macrophages uptake CPP-proteins more efficiently than lymphocytes. In addition, activated T cells, B cells, and macrophages are better targeted by CPP then these cells in their resting state. In addition, in vivo relevance suggests that there is obvious cell type preference for CPPs with heterogeneous delivery efficiency among various cell population, which should be considered for therapeutic drug design.

Methods
Cell lines and cell culture EL4 (mouse lymphoma cell line) and Jurkat (human lymphoma cell line) cells were purchased from the American Type Culture Collection (ATCC) and cultured using Roswell Park Memorial Institute (RPMI) 1640 media with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin antibiotics. All cells were maintained in a 5% CO 2 incubator at 37stitute (RPMI) 1640 media with 10% fetal bovine seThermo Scientific Hyclone.

Splenocyte isolation
Spleens were removed from 6 to 10-week-old mice and chopped with scissors. The chopped tissue was incubated with 2 mg/ml collagenase D at room temperature for 20 min and red blood cells were depleted by incubating with RBC lysis buffer for 2 min. After centrifugation, the cells were resuspended in 10% FBS and 1% penicillin/streptomycin containing RPMI media and counted.

Flow cytometric analysis of CPP-proteins delivery efficiency of in vitro
To analyze the delivery efficiency of CPP-conjugated proteins or CPP-peptide, 5 μM of each molecule were added to the cells and incubated for 1 h. We used 2.5X10 5 cells/well for Jurkat cells and EL4 cells compared with splenocytes and human PBMCs, and 8X10 6 cells/well for analyzing various splenocyte cell types. Following incubation, cells were washed twice with PBS and trypsinized with 0.25% (for cell line) or 0.05% (for primary cells) trypsin-EDTA at 37°C for 10 min. Cells were washed with PBS and stained with fluorochrome-conjugated cell surface marker specific antibodies and the intracellular delivery efficiency was analyzed using flow cytometry.

Immune cell activation
To activate splenic T cells, anti-mouse CD3 antibody (2 μg/ml, BD Pharmingen) was coated onto 96-well plates (BD Falcon) at 3796-well plateellmousesplenocytes were seeded at a density of 2.5X10 5 cells/well and incubated overnight at 37 r in a 5% CO 2 incubator. To activate B cells, DCs, or macrophages, isolated splenocytes were cultured in 12-well plates at 4X10 6 cells/well with 5 μg/ml (for B cells) or 100 ng/ml (for DCs and macrophages) of Escherichia coli 055:B5 (Sigma-Aldrich) in each well and incubated overnight at 37°C in 5% CO 2 incubator.
Flow cytometric analysis of CPP-protein delivery efficiency in vivo 5 mg of EGFP, CPP-EGFP, or PBS was intravenously injected into 6-week-old C57BL/6 female mice. After 1 h, the mice were sacrificed and spleens were removed for isolating splenocytes from a single cell suspension. The cells were stained with fluorochrome-conjugated surface marker specific antibodies and analyzed via flow cytometry.

Statistical analysis
The data were analyzed using one-way ANOVA, two-way ANOVA with multiple comparison tests or Student's t-test. p values <0.05 were considered significant. Statistical analysis was performed using Prism 6 (Graphad Software, Inc.).

dNP2-and TAT-proteins are delivered more efficiently to phagocytic cells than lymphocytes
To determine cell type heterogeneity of dNP2-protein delivery efficiency in various immune cells of mouse spleen, we treated dNP2-EGFP or TAT-EGFP to splenocytes and intracellular delivery efficiency was analyzed by flow cytometry by the staining of cell type specific markers (Fig 1A-1D). The relative mean fluorescent intensity (Relative MFI) represents fold increase that each MFI is normalized to PBS treated sample (Fig 1B-1D). Statistical significance among the all groups were comparatively analyzed by one-way ANOVA test with Sidak's multiple comparison test (S1 Fig). Both dNP2-EGFP and TAT-EGFP proteins were delivered into CD11c lo CD11b hi F4/80 + macrophages and CD11c + MHCII + DCs, which are phagocytic and antigen presenting cells, with significantly higher efficiency than into lymphocytes. In the case of CD4 T cells, B cells, and NK cells, only dNP2-EGFP showed significant protein delivery efficiency comparing with PBS control while TAT-EGFP did not at 5 μM. This pattern is almost identical at various incubating time points including 10 min, 30 min, 1 h, 2 h or 6 h (S2 Fig) and also consistent in the FACS-sorted each cell population (S3 Fig). Interestingly, even EGFP protein without CPPs was significantly taken up by macrophages as well as DCs although their relative MFI value is much lower than that of dNP2-or TAT-EGFP treated group ( Fig 1B). These results collectively suggest that phagocytic cells such as macrophages and DCs uptake CPP-protein more efficiently than lymphocytes, and CPPs such as dNP2 and TAT could significantly deliver a cargo protein into those cells to modulate innate immunity including antigen presentation, inflammatory response, etc.
CD8 + dendritic cells are the major cell type to uptake dNP2-and TATproteins Splenic DCs can be divided into subsets including lymphoid CD8 + DCs, myeloid CD8 -DCs, and plasmacytoid DCs (pDCs) based on different biological characteristics [22][23][24]. To determine CPP-protein delivery efficiency in splenic DC subsets, splenocytes were treated with red fluorescent dTomato protein fused with dNP2 or TAT (dNP2-dTomato or TAT-dTomato). Intracellular red fluorescence intensity and each DC subset marker were analyzed by flow cytometry. First, we determined the DC subsets as Siglec H + pDC, Siglec H -CD11c +-MHCII + CD8 + lymphoid CD8 + DC, and Siglec H -CD11c + MHCII + CD8myeloid CD8 -DC ( Fig  2A). Both TAT-dTomato and dNP2-dTomato proteins were delivered into the CD8 + DCs with significantly higher efficiency relative to CD8 -DCs. In addition, pDCs did not uptake CPPproteins suggesting that the major cellular subset to uptake CPP-proteins is CD8+ lymphoid DCs (Fig 2B, 2C and 2D). These results suggest that among the DC subsets, CPP-protein uptake efficiency is quite variable even in DC subsets and that lymphoid DCs have higher delivery efficiency than myeloid DCs and pDCs indicating that immunization with CPP-conjugated antigen cargos would be efficient application using CPPs such as dNP2 and TAT. mDCs also showed significantly higher delivery efficiency than pDCs (Fig 2C and 2D) suggesting that targeting mDCs by CPP-mediated cargo delivery would be good approaches such as anti-inflammatory functions.
Activated lymphocytes uptake dNP2-and TAT-proteins more efficiently than resting state lymphocytes Because activated lymphocytes are rapidly proliferating and their membrane proteins are dynamically regulated, we hypothesized that CPP-proteins would be delivered into activated or memory-like lymphocytes with higher efficiency relative to resting state lymphocytes. To examine this, we treated dNP2-dTomato or TAT-dTomato to mouse splenocytes. Then, we at first subdivided CD4 T cells into CD4 + CD62L hi CD44 lo naïve CD4 T cells, CD4 + CD62L lo CD44 hi memory-like CD4 T cells, and CD4 + Foxp3 + regulatory T cells by flow cytometry and determined heterogeneous delivery efficiency by CPP-proteins (Fig 3A). Both dNP2-dTomato and TAT-dTomato protein, but not dTomato without CPPs, delivery efficiencies were significantly higher in CD62L lo CD44 hi memory like activated CD4 T cells than in naïve CD4 T cells and Foxp3 + regulatory T cells, though the efficiency of dNP2 is much higher than TAT (Fig 3B and 3C). Also the delivery efficiency was significantly higher in CD4 + Foxp3 + CD44 high cells than that of CD4 +-Foxp3 + CD44 low cells (Fig 3D and 3E). This result suggests that among the peripheral T cell subsets, CPP-protein delivery efficiency is heterogeneous that memory-like T cells or effector T cells are preferred by CPP-protein delivery than naïve T cells.
Next, we stimulated isolated mouse splenocytes with plate bound anti-CD3 and anti-CD28 antibodies to determine whether activated CD4 and CD8 T cells would uptake CPP-proteins more efficiently relative to unstimulated cells (Fig 4A). Both dNP2-EGFP and TAT-EGFP proteins were delivered into activated CD4 and CD8 T cells with significantly higher efficiency than unstimulated cells (Fig 4C, 4E and 4F). In addition, we stimulated splenocytes with lipopolysaccharide (LPS) to examine whether activated B cells would uptake CPP-proteins ( Fig  4B). From the results, LPS stimulated B cells uptake dNP2-EGFP and TAT-EGFP proteins more efficiently than resting B cells (Fig 4D and 4G). These results collectively suggest that CPPs such as dNP2 and TAT could deliver a cargo protein into activated immune cells with better efficiency than their resting state that it would be valuable to apply on modulating abnormally activated T or B cell responses in allergic or autoimmune diseases.
Activated macrophages, but not activated dendritic cells, uptake dNP2and TAT-proteins more efficiently than resting state We next examined CPP-protein delivery efficiency in activated macrophages and DCs compared to their resting state. Isolated mouse splenocytes were incubated with or without LPS stimulation overnight, and then cells were treated with dNP2-EGFP or TAT-EGFP protein for 1 h. The intracellular fluorescence of CD11c lo CD11b hi F4/80 + macrophages and CD11c +-MHCII + DCs was analyzed by flow cytometry. dNP2-EGFP treated LPS-stimulated macrophages showed significantly higher intracellular fluorescence intensity than unstimulated macrophages and TAT-EGFP treated macrophages also showed a similar delivery pattern even though the delivery efficiency of dNP2 is much higher than TAT (Fig 5A, 5C and 5E). However, CPP-protein delivery efficiency in LPS-stimulated DCs was significantly lower than in resting DCs (Fig 5B, 5D and 5F) possibly due to maturated DCs have less antigen uptake ability then immature state [25]. Interestingly EGFP without CPPs also shows consistently reduced protein uptake efficiency by DCs while no difference was observed in macrophages (Fig 5E and  5F). These results suggest that CPP-protein delivery efficiency is more significant in activated macrophages and immature DCs that it could be valuable to apply it on modulating inflammatory responses mediated by macrophages or on initiating immune responses mediated by DCs.

dNP2-but not TAT-proteins shows consistent cell type preference in various immune cells of mouse spleen in vivo
Finally, we investigated the applicability of in vitro results in vivo by intravenous administration of dNP2-EGFP or TAT-EGFP protein in mice. One hour after the intravenous injection of 5 mg dNP2-EGFP or TAT-EGFP into mice, splenocytes were isolated and various immune cell types were stained with cell type specific marker antibodies ( Fig 6A). As consistent with the in vitro results, F4/80 high CD11b int macrophage cells (blue) are the most major cells to uptake proteins. In addition, delivery efficiency of dNP2-EGFP shows higher delivery efficiency on Ly6C +-CD11b high monocytes than Ly6G + CD11b high neutrophils (Fig 6B and 6C). Among the DC subsets, CD8 + Siglec Hlymphoid DCs (blue) showed higher delivery efficiency than CD8 -Siglec Hmyeloid DCs and Siglec H + pDCs (Fig 6D and 6E). CD62L lo CD44 hi memory-like activated cells (red) in both CD4 and CD8 T cells preferentially transduced by dNP2-EGFP protein relative to CD62L hi CD44 lo naïve CD4 and CD8 T cells (Fig 6F and 6G). However, importantly, TAT-EGFP delivery efficiency was not significant in any cell type population suggesting its limitation for in vivo application. These results suggest that dNP2 efficiently deliver a cargo protein into various immune cells in vivo with heterogeneous delivery efficiency and this should be considered for its application design to modulate cellular responses in vivo.

Discussion
In the present study, we characterized and demonstrated cell type preference of the recently identified human derived BBB-permeable peptide dNP2 [21] and the widely used HIV-derived CPP, TAT [4,26] for the uptake of proteins in various immune cells. Both dNP2 and TAT conjugated fluorescent proteins were more preferentially delivered into phagocytic cells such as macrophages and DCs relative to lymphocytes. In addition, we confirmed that activated immune cells including macrophages, T cells, and B cells could uptake proteins with better efficiency relative to their resting state, collectively demonstrating heterogeneous intracellular delivery efficiency of dNP2 and TAT in vitro and in vivo should be considered for better application of these CPPs in disease modulation.
CPPs are generally regarded as a non-specific macromolecular delivery tool, which has been considered a major disadvantage of CPP usage for therapeutic purposes [27][28][29]. However, previous studies reported that the arginine-based CPP r8 peptide, without any targeting sequences, spontaneously accumulated in xenografted tumors of tumor-bearing nude mice  [30]. Even though the R8 peptide was delivered into normal tissues, it showed a preference for tumor cells. In addition, in the early days of CPPs, Dowdy et al. reported an odd delivery pattern in spleen tissue in vivo where TAT-β-galactosidase was only detected in red pulp while it was clearly not detected in white pulp in mice [4]. These studies suggested that CPPs could be delivered with different efficiencies and specificities depending on target cell type. These previous results suggest the possibility of cell type preference in various types of cells in vivo.
In the present study, we revealed that both dNP2 and TAT have cell type preference in various immune cells. Although the proportion of DCs and macrophages are just 1-2% of total spleen cells, their CPP-protein uptake efficiency is much dominant compare to the lymphocytes including T cells or B cells. This suggests that the phagocytic action or cell's own characteristics are important in effective CPP-protein delivery. High efficiency in phagocytic cells including DCs and macrophages also suggest that the application of CPPs such as dNP2 as an immunization tool to delivery of antigens for vaccine development could be one of the best approaches. Especially, lymphoid DCs were shown to uptake proteins much more efficiently than myeloid DCs and plasmacytoid DCs. Previously, it has been reported that lymphoid DCs have much higher T cell stimulation capacity than myeloid DCs [31]. While, myeloid DCs have been known as having higher migratory capacity compared with lymphoid DCs [32][33][34]. In addition, only lymphoid DC can cross-prime cytotoxic T cells in vivo [35]. Thus, lymphoid DCs which is preferred by CPP-cargo delivery is appropriate target for DC vaccination using cell-penetrating peptide [36]. Similar to the preference and high delivery efficiency of TAT to lymphoid DCs in this study, previous studies showed that TAT-conjugated protein antigen could be a useful DC vaccination methodology for boosting antigen specific CD8 T cell immunity [37,38]. For example, treatment of the TAT conjugated Leishmania antigen TAT-LACK to DCs successfully induced Leishmania specific CD8 T cells in vitro while LACK alone could not [39]. Also, numerous studies reported that DC vaccination is one of the most attractive approaches for cancer immune therapy using specific cancer antigens [40][41][42][43]. TAT conjugated breast cancer specific antigen generated cancer specific CD8 T cells and infiltrated into tumor tissues in vivo [44]. Our study previously showed that dNP2 has higher delivery efficiency relative to TAT suggesting that dNP2 could be more valuable to deliver antigens for the generation of pathogen or cancer specific immunity.
Macrophages are another potent target cell type for modulating the immune system because of their high delivery efficiency with both dNP2 and TAT. Macrophages are one of the most abundant cells existing in every tissue including lymphoid, lung, liver, bone, and brain tissues with phenotypic diversity [45,46]. For this reason, the preference of dNP2-protein delivery to macrophages could be a useful tool for modulating macrophage functions regarding inflammation, disease pathogenesis, or bone remodeling.
Both dNP2 and TAT conjugated protein were taken up by memory like CD4 T cells or activated lymphocytes including CD4 and CD8 T cells or B cells with higher efficiency relative to these cells in their resting state. This preference would have the advantage of modulating TAT-EGFP, dNP2-EGFP, or PBS (5 mg) were injected to 6-week-old C57BL/6 mice and after 1 h the splenocytes were isolated and analyzed by flow cytometry after cell surface marker staining. (B-C) Protein delivery efficiency into Ly6C + CD11b high monocytes (red), F4/80 high CD11b int macrophage cells (blue), and Ly6G + CD11b high neutrophils (orange) was analyzed by flow cytometry. (D-E) Protein delivery efficiency into CD8 -Siglec Hmyeloid DCs (red), CD8 + Siglec Hlymphoid DCs (blue), and Siglec H + pDC (orange) was analyzed by flow cytometry. (F-G) Protein delivery efficiency into CD4 + CD62L high CD44 low , CD4 + CD62L low CD44 high , CD8 + CD62L high CD44 low , or CD8 + CD62L low CD44 high cells was analyzed by flow cytometry. Bar graphs indicate the mean ± s.d. and within groups (EGFP, TAT-EGFP, dNP2-EGFP) the differences of means of each cell types were compared with each other by two-way ANOVA with Tukey's multiple comparison test, * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001, **** indicates p<0.0001 (n = 3).
doi:10.1371/journal.pone.0155689.g006 activated lymphocytes and memory T and B cells by delivering signaling modulatory molecules. Previously we have attempted to deliver cytoplasmic domain of CTLA-4 protein which inhibits T cell receptor signaling pathway in activated T cells modulated abnormal immune responses in allergic or autoimmune disease models [18,21,47]. In addition, CPP-conjugated Foxp3 protein could modulated OVA-induced allergic T cell responses in vivo without affecting the other T cells response to KLH antigen which was still intact [12]. Other previous report showed that CPP-based antisense-oligomer delivery successfully modulated T cell response by altering gene expression [48]. These results suggest CPPs such as dNP2 and TAT could be applied to modulate activated lymphocyte reaction by passive targeting or preference moiety.
In consistency of relative preference on activated T cells, regulatory T cells which is highly expressing CD44, that could be regarded as activated/effector regulatory T cell [49], showed significantly higher CPP-protein uptake efficiency than resting regulatory T cells which have lower level of CD44. Regulatory T cell subsets were reported as CD45RA + Foxp3 low resting regulatory T cells, CD45RA -Foxp3 high activated regulatory T cells and CD45RA -Foxp3 low nonsuppressive T cells [50]. In the case of carcinoma patients, the activated regulatory T cells are increased in the peripheral circulation and antitumor immunity is suppressed [51]. The CPPdrugs could be preferably delivered into activated regulatory T cells for cancer therapy which possibly be a novel approach for cancer-immunotherapy.
Collectively, we demonstrate cell type preference of CPP delivery efficiency using dNP2 and TAT in various immune cells in vitro and in vivo. First, CPP delivery efficiency is the highest into phagocytic cells, especially lymphoid DCs. Second, CPP delivery efficiency is much higher into the activated immune cells than non-activated immune cells. Last, in vivo protein delivery efficiency of TAT is as little as negative control suggesting significant advantage of dNP2 usage in vivo. These findings suggest that heterogeneity of CPP-protein delivery efficiency would be dependent on cells own character such as phagocytic activity, etc. Cell type preference analysis in non-immune cells also should be considered and it would provide important additional information for the accurate application of CPPs in appropriate target cells. 5 μM of EGFP, TAT-EGFP, dNP2-EGFP or PBS were treated to splenocytes from 6-week-old C57BL/6 mice for various reaction times including 10 min, 30 min, 1 h, 2 h or 6 h and protein delivery efficiencies were analyzed by flow cytometry after cell staining with cell type specific markers (CD4 + CD4 T cells, CD19 + B cells, NK1.1 + NK cells, CD11c low CD11b high F4/80 + macrophages, and MHCII + CD11c high DCs). The values were normalized with the mean fluorescence intensity (MFI) of PBS treated samples (relative MFI). The graphs indicate the mean ± s. d. In the case of TAT-EGFP and dNP2-EGFP, macrophages and DCs showed significantly higher delivery efficiency than lymphocytes within every time points. We performed statistical analysis by two-way ANOVA with Tukey's multiple comparisons test and ÃÃÃ indicates p<0.001 and ÃÃÃÃ indicates p<0.0001.