Photodynamic inactivation of Leishmania braziliensis doubly sensitized with uroporphyrin and diamino-phthalocyanine activates effector functions of macrophages in vitro

Photodynamic inactivation of Leishmania has been shown to render them non-viable, but retain their immunological activities. Installation of dual photodynamic mechanisms ensures complete inactivation of species in the Leishmania subgenus, raising the prospect of their safe and effective application as whole-cell vaccines against leishmaniasis. Here, we report the successful extension of this approach to L. braziliensis in the Viannia subgenus, viz. genetic engineering of promastigotes for cytosolic accumulation of UV-sensitive uroporphyrin (URO) and their loading with red light excitable phthalocyanines (PC) that was cationized by chemical engineering. The transgenic strategy used previously produced L. braziliensis transfectants, which gave the same phenotype of aminolevulinate (ALA)-inducible uroporphyria as found in Leishmania subgenus, indicative of pre-subgenus evolutionary origin for similar genetic deficiencies in porphyrin/heme biosynthesis. In the present study, 12 independent clones were obtained and were invariably ALA-responsive, albeit to different extent for uroporphyrinogenesis and UV-inactivation. In a separate study, L. braziliensis was also found, like other Leishmania spp., to take up diamino-PC (PC2) for red light inactivation. In vitro interactions of a highly uroporphyrinogenic clone with primary macrophages were examined with the intervention of URO/PC2-medated double-photodynamic inactivation to ascertain its complete loss of viability. Doubly sensitized L. braziliensis transfectants were photo-inactivated before (Strategy #1) or after (Strategy #2) loading of macrophages. In both cases, macrophages were found to take up L. braziliensis and degrade them rapidly in contrast to live Leishmania infection. The effector functions of macrophages became upregulated following their loading with L. braziliensis photodynamically inactivated by both strategies, including CD86 expression, and IL6 and NO production. This was in contrast to the immunosuppressive infection of macrophages with live parasites, marked by IL10 production. The results provide evidence that photodynamically inactivated L. braziliensis are susceptible to the degradative pathway of macrophages with upregulation of immunity relevant cytokine and co-stimulatory markers. The relative merits of the two loading strategies with reference to previous experimental vaccination were discussed in light of the present findings with L. braziliensis.


Aminolevulinate (ALA)-induction of the doubly transfected L. braziliensis for neogenesis of cytosolic uroporphyrin 1 (URO), indicative of ALAD and PBGD expression. L. braziliensis
promastigotes were doubly transfected and cloned, as described (see "Methods"). All 12 independent clones obtained grew stably during repeated cycles of passages under the selective drug pressures. They were fully revivable without losing drug-resistance phenotypes after cryopreservation in liquid nitrogen, further indicative of the transgenic stability. Cells of all 12 clones were found responsive to ALA, producing intense porphyrin-specific fluorescence in the cell pellets after centrifugation and illumination with UV lamp (Fig. 1A, top panel 1-12 plus a positive control C2). No fluorescence was visible in those of the wildtype without transfection (Fig. 1A,C1) or when illuminated with white light (Fig. 1A, bottom panel). Notably, all ALA-exposed transfectant clones remained intact and motile, similar to untreated controls when kept in the dark, as noted by phase contrast microscopy of uroporphyric transfectants. Under this setting of very dim illumination, the promastigotes were seen to remain motile initially for ~ 1 s followed by an abrupt cessation of their flagellar motility, in keeping with the lag time expected from excitation of their cytosolic URO to the production of a sufficient amount of cytotoxic 1 O 2 . Co-sedimentation of the URO fluorescence with cells in pellets after centrifugation is indicative of its cell-association, as noted with similarly treated uroporphyic L. amazonensis (Fig. 1A, top panel, clones 1-12 vs. C2). The faint fluorescence observed in the supernatants was presumably due to efflux of URO from uroporphyric transfectants, as seen previsouly 19 . Notably, WT L. brasiliensis produced no enzymes equivalent to the trangene products, i.e. ALAD and PBGD, as exposure of these cells to ALA produced no porphyrin fluoresence (Fig. 1A, top panel, C1). Fluorescence microscopy of live transfectants showed that porphyrin was diffused cytosolically throughout the cells, including flagella, but also condensed in cytoplasmic vacuoles (Fig. 1B www.nature.com/scientificreports/ The 12 clones obtained were found to vary in their response to ALA for uroporphyrinogenesis in both kinetics and peak levels, as shown by flow cytometric analysis of these cell samples. Changes were noted in the percentage of porphyric cells and their mean fluorescence intensity (MFI) with the progression of time. URO fluorescence of all ALA-exposed samples began to emerge in ~ 24 h at low levels, i.e. 10-20% porphyric cells and up to ~ 200 MFI in URO-positive cells ( Fig. 2A) and peaked in 48 h to variable levels, i.e. 40-70% porphyric cells and 100-700 MFI in URO-positive cells, being lower and the lowest for clones 2-3 and clones 8-9, respectively (Fig. 2B). By 72 h, URO levels decreased in the cells of all 12 clones according to both % porphyric cells and MFI values (Fig. 2C), especially the latter (right panel). This was coincident with the emergence of fluorescence in the supernatants or release of URO, as noted after sedimentation of porphyric cells (Fig. 1A, top panel). Clones # 4-6 and 11-12 were more responsive than clones 2-3 and 8-10 to ALA exposure for porphyrinogenesis.
All 12 porphyric clones were sensitive to inactivation by UV illumination, as shown by the MTT reduction assay for cell viability. Their MTT reduction activities decreased to 20-30% of the respective dark controls in most cases (Fig. 2D, white bar vs. black bar). MTT-reduction activities of clones #4 and #11 were negligible after irradiation relative to the dark controls. Their loss of viability thus appeared far more significant than expected from the URO levels of these two clones, as shown ( Fig. 2A-C, Clones #4 and #11). This incongruity was taken to indicate that cells with URO, but too little for flow cytometric detection remained susceptible to UV-inactivation. Indeed, apparently aporphyric cells were previously observed to lose flagellar motility under the illumination of fluorescent microscopy for porphyrin 19,21 . As expected, the parental WT L. braziliensis were true aporphyric cells, as they remained equally viable with and without UV-exposure, yielding comparable MTT reduction activities (Fig. 2D, WT Lb white bar vs. black bar). These results show that all doubly transfected L. braziliensis clones are susceptible to ALA-induced porphyrinogenesis and photo-inactivation to various extents. Clone #4    To determine light-susceptibility (D), cells were first incubated with ALA (1 mM) in the dark for 48 h and then split into two sets: one set was kept in the dark (black bar, Lb + ALA) and the other set exposed to longwave UV for 30 min (white bar, Lb + ALA + L). Cell viability was assessed by MTT reduction assay. Photo-inactivation of UV-exposed cells for each clone was expressed as % of the corresponding dark control. Negative controls: wild-type L. braziliensis (WT Lb) similarly incubated with ALA followed by exposure (white bar) or no exposure (black bar) to UV. Data are presented as mean ± S.D. from a representative experiment performed in quadruplicate. ***p < 0.001. www.nature.com/scientificreports/ found to contain in each 2-3 WT Lb (gray bars). The values for both parameters were lower when using porphyric Lb (black bars) and the UV-inactivated porphyric Lb (white bars), especially for the #Lb/macrophage, i.e. a reduction to less than half of the values in comparison to the other two samples (Fig. 3A,B, 6 h). At 24 h, while the readings of both parameters decreased for all samples, they remained comparable between WT Lb (gray bars) and porphyric Lb (black bars), and were significantly lower for UV-inactivated porphyric Lb (white vs. gray bars) (Fig. 3A,B, 24 h). The differences between porphyric Lb and UV-inactivated porphyric Lb are illustrated in the representative photomicrographs (Fig. 3C). Exposure of wildtype L. braziliensis to UV did not affect their www.nature.com/scientificreports/ viability (Supplemental Fig. 1), indicating that it did not contribute directly to the reduced uptake of porphyric and UV-inactivated porphyric L. braziliensis by BMDM. The difference in the levels of uptake between the two is proportional to the levels of their light exposure, i.e. incidental stray light in the former case and UV illumination in the latter. Also, the quantitative data presented represent a net sum of both uptake and intracellular degradation in the acidic environment of phagosome-lysosome vacuoles. UV-inactivated porphyric cells were nonviable and thus more rapidly degraded than substantially viable porphyric cells exposed only to incipient light.

Fate of L. braziliensis transfectants in BMDM with the intervention of double photo-inactivation by two different loading strategies.
We first established that L. braziliensis was susceptible to loading with PC2 for inactivation by red light, as found with other Leishmania species 22,23 . By MTT reduction assays, both PC2-loaded WT L. braziliensis and porphyrinogenic transfectants were found to lose their viability in a concentration dependent manner after red-light exposure in contrast to those kept in the dark (Supplemental Fig. 2, white vs. black bars). PC2 was thus used to sensitize uroporphyric clone #4 for double photo-inactivation to examine their interactions with BMDM according to the two loading strategies presented in Fig. 4. Both strategies inactivated Lb effectively, rendering them susceptible to degradative pathways of BMDM, as indicated by a significant reduction of intracellular parasites when assessed ~ 1 day after loading (Fig. 5). Doubly PS-sensitized, but not light-inactivated clone #4 and wildtype gave a comparable rate of infection at ~ 50% ( Fig. 5A gray bar vs black bar), but the average number of parasites per cell was twice fewer in the former than in the latter, i.e. 1.5 and 3 after 24 h ( Fig. 5B gray bar vs black bar). While the doubly sensitized clone #4 Leishmania were not specifically illuminated, they were unavoidably exposed to incipient dimmed room light, accounting for the decreased values shown. The values for both parameters were significantly lowered upon loading BMDM with photodynamically inactivated parasites via both strategies (Fig. 5A,B, orange and red bars).  www.nature.com/scientificreports/ on the viability of the promastigotes (Supplemental Fig. 1) or indirectly via its effect on macrophages (Supplemental Fig. 3), as documented previously 17,23 . A significant increase in superoxide above the levels of all these controls was observed by loading BMDM with photo-inactivated Lb according to both Strategies #1 (orange bar) and #2 (red bar). Taken together, the results indicate that light-inactivation of doubly PS-sensitized Lb already in BMDM via Strategy #2 elevates NO level, and that loading of BMDM with photo-inactivated Lb by either strategy induces the production of superoxide. Loading of BMDM with photodynamically inactivated Lb altered the expression of co-stimulatory molecules CD40 and CD86 to different extent (Fig. 7A,B). Infection of BMDM with live PS-sensitized (black bar) and WT Lb (dark gray bar) elevated the MFI of both CD40 and CD86 molecules above the ground levels of BMDM alone (light gray) and those light-exposed (blank) (Fig. 7A,B). Only CD86 was significantly upregulated above these samples by loading BMDM with photodynamically inactivated Lb via either Strategies #1 (orange bar) or #2 (red bar).

Innate responses of BMDM enhanced by their interactions with doubly photo-inactivated
Loading of BMDM with Lb photo-inactivated by both Strategy #1 (orange bar) and Strategy #2 (red bar) significantly up-regulated IL-6 over their controls, i.e. no treatment (light gray) and infection with live Lb (dark gray and black bar), as true for TNF via Strategy #2 (Fig. 7C,D). Significantly, the reverse was observed for IL-10 ( Fig. 7E), i.e. down-regulation with photo-inactivated Lb (orange and red bar) and up-regulation with live Lb (light gray and dark bar). These results show that loading of BMDM with doubly PS-sensitized and photo-inactivated Lb upregulates the expression of co-stimulatory molecule CD86 and inflammatory cytokines in parallel to their intracellular processing for degradation as seen microscopically.

Discussion
In the present study, Leishmania braziliensis of the Viannia subgenus was successfully photodynamically inactivated by the dual approach of genetic and chemical engineering, which has proven effective to render members of the Leishmania subgenus non-viable for safe and effective use as experimental whole-cell vaccines 12,13 . All transgenic clones produced in the present work were responsive to delta-aminolevulinate (ALA) for uroporphyrinogenesis, i.e. cytosolic de novo synthesis and accumulation of uroporphyrin 1 (URO), making them sensitive to UV inactivation. This outcome of the transfection signifies the genetic defects of L. braziliensis in porphyrin/ heme biosynthesis, exactly as seen previously in the Leishmania subgenus 18,19 , i.e. the absence of the 1st five enzymes from ALA-synthase to uroporphyrinogen decarboxylase in this biosynthetic pathway. The origin of these genetic deficiencies thus appears to predate the split of the two Leishmania subgenera, consistent with the prevailing view for the evolutionary loss of genes coding for heme biosynthetic enzymes in trypanosomatids as a whole 24 . Of significance, not accomplished previously, is the production of as many as12 independently selected clones, which are all ALA-inducible for uroporphyrinogenesis without exception. Thus, the possibility of inadequate or incomplete transfection is effectively excluded from consideration to account for the presence of apparently aporphyric cells with levels of URO undetectable by fluorescent microscopy and flow cytometry. Pleiotropy of ALA transport and/or URO efflux 19,21 was previously proposed to explain heterogeneity of the cell population in the % URO-positive cells and their URO levels observed during uroporphyrinogenesis. Consistent with this proposal is the observation of considerable variations noted among the 12 obtained clones. While the chosen Clone #4 stands out as the most uropohyrinogenic for UV-inactivation, its subjection to an additional round of diamino-phthalocyanine (PC2)-mediated inactivation with red light 22,23 is for prudence to render its cell population totally non-viable with certainty. Photo-excitation of both URO and PC2 is known to produce highly destructive, but extremely short-lived 1 O 2 to initiate rapid oxidative inactivation of Leishmania 21,23,25 . By virtue of their differences in intrinsic photodynamic properties and cellular location, URO and PC2 are synergistic and compensatory to each other for effective photo-inactivation of Leishmania spp. 20 , including L. braziliensis examined in the present study. The interactions of doubly photo-inactivated L. braziliensis with bone marrow-derived macrophages (BMDM) by both strategies described result in immunity-favorable outcome. In Strategy #1, loading of BMDM with nonviable clone #4 cells, which were doubly photodynamically pe-inactivated before use, spares their exposure to potentially detrimental PDT and immunosuppression known to result from infection with live promastigotes 26 . The reduced loading of BMDM with such inactivated Leishmania seen is fully expected, as they are non-viable www.nature.com/scientificreports/ and thus release no secretory products previously shown to facilitate infection of macrophages, e.g. nucleoside diphosphate kinase 27 . Strategy #2 is a trade-off between effective loadings of BMDM with URO/PC2-sensitized, but still viable Leishmania versus the risk of PDT collateral consequences from their exposure to light targeting Leishmania therein for inactivation. Irrespective of the loading strategies used, the inactivated Clone #4 cells disintegrate rapidly versus live parasites intracellularly, as indicated by comparing the loss of their structural integrity with time in BMDM. Both strategies thus produce a similarly favorable outcome in the context of loading antigen-presenting cells (APC) with whole-cell vaccines. While processing of these "vaccines" for antigen presentation to activate T cells awaits further investigation, such outcome is extrapolatable from previous findings with ovalbumin delivered to dendritic cells by photodynamically inactivated Leishmania 23 . The response of BMDM to loading with doubly photodynamically inactivated L. braziliensis via both strategies is, for the most part, in favor of immunity (summarized in Supplemental Fig. 4). Overall, the results are consistent with and in support of the previous findings of microarray analysis in that J774 macrophages functionally returned to normalcy when their intracellular Leishmania were eliminated by single photodynamic inactivation via Strategy #2 17 . Here, degradation of the parasite loads in BMDM is accompanied by maintenance of effector functions, in contrast to immunosuppression known to occur in infection with live Leishmania 26 . Most evident for this is the reversal of cytokine profiles, i.e. the reduction of IL-10 and concomitant elevation of IL6 and TNF, the latter being pronounced in Strategy #2. These alterations are consistent with enhanced processing of antigens by the macrophages for presentation, since IL-10 is associated with non-degradative pathway favoring the survival of Leishmania in infected macrophages and, since TNF together with IFN-γ is known to activate macrophages with increased degradative microbicidal activities 28 . Also evident from both Strategies for photodynamic inactivation of L. braziliensis in BMDM is the recovery of their respiratory burst, which is countered by intricate mechanisms reported for live Leishmania , e.g. release of arginase to reduce nitric oxide (NO) production by substrate deprivation, up-regulation of superoxide dismutase (SOD) to reduce superoxide levels 29 and unknown factor(s) to reduce the production of ROS by inhibiting membrane NADPH oxidase. Increased production of NO and superoxide is noted to vary with the two Strategies for photodynamic inactivation of Leishmania in BMDM. The elevation of NO over the control is most impressive via Strategy #2, but not Strategy #1, accountable perhaps by the difference in their loading efficiency. Superoxide is elevated similarly under both strategies as well-above the un-infected, but also, to a lesser extent, over the other control groups, i.e. live parasite infection and light exposure alone. The elevation of superoxide in BMDM to some extent in these heterogeneous control groups suggests that the production of this ROS may be subjected to the regulations by several different signal pathways. Redundant anti-oxidant systems also exist in both macrophages and Leishmania, presenting multi-layered complexity of the host-parasite interactions in the disposal of ROS, like superoxide. This is in contrast to 1 O 2 , which neither macrophages nor Leishmania is known to produce and thus has evolved no specific mechanism for its detoxification. The examination of co-stimulating molecules is on account of Leishmania strategies, which have been reported to manipulate the co-stimulatory pathways of macrophages to their advantages [30][31][32] . Loading of macrophages with photodynamically inactivated L. braziliensis by both Strategies offers some evidence for upregulation of CD86 and, to a lesser extent, CD40, even though infection of BMDM with live parasites also elevate them over un-infected and light-alone controls . Up-regulation of these co-stimulatory molecules has the potential to facilitate activation of T cells via binding to their CD28 and CD40L, respectively, as previously demonstrated for dendritic cells in self-resolving L. braziliensis infection 33 .
The results obtained shed additional lights on the relative merits of the two strategies for vaccination, both using whole Leishmaina cells to retain their entire antigenic repertoire for presentation to the immune system. Strategy #1 is favorably disposed with significant advantages of preparatory simplicity relatively and safety assurance by using pre-inactivated non-viable Leishmania. These are the overriding issues for consideration of any vaccine development. Doubly photodynamic inactivation of Leishmania indeed renders them non-viable 20 , but preserves their antigenicity and adjuvanicity. The efficacy of Strategy #1 was thus demonstrated in experimental vaccination against cutaneous leishmaniasis 13 . As suggested by the in vitro work presented here with L. braziliensis, Strategy #1 may be further improved to enhance its efficacy by increasing doses for loading APC to match that of Strategy #2. While the latter was previously shown as effective in prophylactic vaccination against experimental visceral leishmaniasis 12 , it is burdened by significant disadvantages, i.e. the cumbersome in situ light illumination after immunization with PS-sensitized, but live Leishmania and the potential risk of infection due to possibility of their incomplete inactivation. In either case, photodynamically inactivated Leishmaina are not lesion-derived live parasites used for leishmanization 34 and are different from other whole-cell live preparations, such as Leishmania attenuated spontaneously or crippled by other chemical, physical or genetic means [8][9][10][11] . Clearly, the use of doubly PS-sensitized and photo-inactivated L. braziliensis for immunization ameliorates the risk of their potential persistence, reversion to virulence and disease reactivation associated with use of attenuated live mutant parasites, thereby providing novel prophylaxis alternatives against leishmaniasis caused by this species.

Methods
Ethics statements. Female BALB/c mice, 6-8 weeks of age, were obtained from the IGM-FIOCRUZ animal facility where they were maintained under pathogen-free conditions. All animal experimentation was conducted in accordance with the Guidelines for Animal Experimentation established by the Brazilian Council for Animal Experimentation Control (CONCEA). All procedures involving animals were approved by the local Institutional Review Board for Animal Care and Experimentation (CEUA-IGM-FIOCRUZ-003/2014.

Construction of expression vectors.
Two different plasmids were used for double transfection of L. braziliensis promastigotes to express mammalian cDNA encoding ALAD and PBGD 19 , i.e. pX-NEO-alad available previously (Supplemental Fig. 5A) and the newly constructed pXG-HYG-pbgd. For the latter, p6.5-pbgd 19 was first BamHI-digested and the pbgd coding sequence released was gel-purified for cloning at the BamHI sites of the pXG-HYG expression vector (Supplemental Fig. 5B). Competent E. coli (DH5α) cells were transformed with the constructs and plasmids from several colonies were restriction-mapped to select for clones with correctly oriented insert.
Transfection of L. braziliensis and selection of transfectants. Promastigotes of L. braziliensis were subjected to two consecutive rounds of electroporation to obtain double transfectants expressing both ALAD and PBGD. The first round used pX-NEO-alad and the resulting single transfectants were subjected to a second round of electroporation with the newly constructed pXG-HYG-pbgd, thereby producing L. braziliensis double transfectants expressing both ALAD and PBGD.
For electroporation, promastigotes grown to mid-log phase were harvested by centrifugation for resuspension to a cell density of 10 8 cells/ml in Tb-BSF electroporation buffer 37 . After electroporation, transfectants were selected on 96-well tissue culture plates using the limiting dilution method in M199 medium supplemented with appropriate antibiotics. For the first round of transfection, mid-log cell suspension (0.4 ml) was mixed with pX-NEO-alad (5 μg) in a pre-chilled 2 mm electroporation cuvette (SIGMA) for electroporation using a Bio-Rad Gene Pulser 38 . A clone well adapted to G418 selection was expanded for a second round of transfection with pXG-HYG-pbgd by electroporation, as described above. Doubly transfected cells were selected for resistance to both hygromycin (HYG) and G418 and cloned in 96-well tissue culture plates with culture medium supplemented with G418 and HYG at 50 μg/ml each. The cultures were monitored for 2-4 weeks. Transfectants emerging in wells containing cells at the highest dilutions were further expanded in selective medium with increasing drug pressures, i.e. G418 and HYG at 100 ug/ml each. Double transfectant clones grew stably in Schneider's medium under the selective pressure of both drugs at the concentration, as indicated.
Aminolevulinate-induced porphyrinogenesis of genetically complemented L. braziliensis clones. Transfectants from 12 independent clones were each grown to stationary phase. These cells were harvested, washed and resuspended to 5 × 10 7 cells/ml in RPMI 1640 (Invitrogen)/0.01% bovine serum albumin (RPMI-BSA). The cells were exposed at 26 °C for up to 48 h in the dark to 1 mM ALA (SIGMA). ALA-exposed and control cells were washed and examined by phase contrast for their integrity. Porphyrin fluorescence of these cells was assessed in three different ways. Aliquots of cell suspension were sedimented for illumination in the dark with an UV-lamp (λmax = 366 nm) to check for porphyrin fluorescence in the cell pellets and in the supernatants for porphyrin release. Microscopically, live cell samples were examined for porphyrin by excitation with Krypton/argon-mixed gas laser at 488 nm and filtered for emission at 605 nm in an Olympus FluoView confocal microscope. ALA-exposed cells of all 12 clones and controls were sampled at different times to assess the kinetics of their porphyrinogenesis by flow cytometry (BD-LSR Fortessa) at an excitation wavelength of 405 nm and emission wavelength of 610 nm. Data from 30,000 events were acquired to determine the % of porphyrin-positive population for each clone. Data were analyzed using FlowJo (Version 10.2). Parental wildtype L. braziliensis and L. amazonensis porphyrinogenic mutants 19 exposed to ALA were included as negative control and positive control, respectively.
Photo-inactivation of porphyric L. braziliensis transfectants and MTT reduction assay to assess their viability. After ALA treatment in the dark, the 12 clones of double transfectants were each suspended to 5 × 10 7 cells/ml in RPMI-BSA and placed at 500 µl/well in 24-well culture plates. Two sets were prepared for each clone. One set was exposed with lid off to longwave UV (λmax = 366 nm) from the top for 30 min. The second set was kept under identical conditions, but foil-wrapped to keep cells in the dark. All samples were subsequently incubated overnight and assessed for cell viability by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) reduction assay (SIGMA) according to the manufacturer's protocol. Parental wild-type cells of L. braziliensis were included as additional controls.
In vitro uptake of photo-inactivated porphyric L. braziliensis and controls by bone marrow-derived macrophages (BMDM). The primary macrophages obtained from mouse bone marrow were suspended in culture medium for seeding at 3 × 10 5 cells/500 μl/coverslip per well in 24-well plates. Monolayers of adherent macrophages formed on the coverslips were washed to remove non-adherent cells and exposed to experimental or control L. braziliensis promastigotes at 10:1 parasite/host cell ratio (clone #4), i.e. Quantitative analysis of nitric oxide, superoxide and cytokines in culture supernatants of BMDM loaded with photo-inactivated transfectants. BMDM were primed for 24 h with LPS (40 ng/ mL) and IFN-γ (10 ng/mL) (all from Invitrogen) or with 0.5 mM hydroxylamine hydrochloride (Acros Organics) for determining NO and superoxide, respectively. BMDM were then loaded with transfectants according to strategies #1 and #2 (Fig. 4). Cell culture supernatants were collected after 18 h for NO and superoxide quantification using Griess reagent (20). Cytokine levels in the culture supernatants collected were determined by ELISA, following manufacturer's instructions (eBioscience).