Introduction

Top Abstract Introduction Methods Results Discussion References

Antisense strategy aims at specific inhibition of target proteins. Specificity is achieved by hybridization of the antisense sequence to the complementary sense sequence of the target mRNA. This mechanism also is used physiologically to regulate gene expression: Naturally occurring antisense transcripts have been described for eucaryotic cells (Farrell and Lukens, 1995; Kimelman and Kirschner, 1989). Imitating this regulatory mechanism, antisense transcripts originating from artificially introduced DNA have been applied successfully in studies of gene function (Zhang et al., 1996) and in gene therapy protocols (Redekop and Naus, 1995; Vandendriessche et al., 1995). The application of exogenous antisense oligonucleotides is an alternative antisense strategy. It has been proven that antisense oligonucleotides microinjected into the cytoplasm of cells inhibit their target protein in a specific and efficient manner (Wagner, 1995). However, whereas endogenous antisense transcripts are formed in the nucleus, their proposed site of action, penetration of exogenously administered antisense oligonucleotides through cellular membranes, remains a major obstacle for applying antisense strategies.

The mechanisms for cellular uptake of exogenously administered oligonucleotides are fluid phase endocytosis, relevant for high oligonucleotide concentration greater than 1 µM, and endocytosis mediated by binding of oligonucleotides to cell surface proteins, postulated for lower oligonucleotide concentrations (Beltinger et al., 1995; Benimetskaya et al., 1997). In general, primary cells incorporate oligonucleotides less efficiently than cell lines (Marti et al., 1992). So far, most studies have examined uptake mechanisms and intracellular trafficking of oligonucleotides in cell lines (Iversen et al., 1992b; Stein et al., 1993; Temsamani et al., 1994; Tonkinson et al., 1994; Vlassov et al., 1994). Despite its implication for in vivo studies, information about spontaneous oligonucleotide uptake in primary monocytes is limited.

Several aspects underscore the significance of oligonucleotide uptake in leukocytes: 1) circulating leukocytes are exposed directly to systemically administered oligonucleotides; 2) oligonucleotides elicit inflammatory responses (Hartmann et al., 1996b; Krieg et al., 1995; Liang et al., 1996; Webb et al., 1997); 3) leukocytes form target cells in clinical studies testing antisense oligonucleotides as therapeutics for inflammatory disease including rheumatoid arthritis and Crohn's disease (Bradbury, 1997), for chronic myelogenous leukemia and myelodysplastic syndrome (Bayever et al., 1993; Nichols, 1995; Skorski et al., 1994) and for human immunodeficiency virus infection (Zhang et al., 1995).

Cationic lipid-mediated transfer of DNA into cells is a well-documented approach of gene therapy protocols in clinical studies (Caplen et al., 1995; Nabel et al., 1993). For antisense oligonucleotides, a specific class of DNA, cationic lipids effectively enhance cellular uptake. Complexes of oligonucleotides and the cationic lipid lipofectin bind to and penetrate cellular membranes. Oligonucleotides subsequently are released into the cytoplasm (Bennett et al., 1992). Antisense effects cannot be achieved in intact human leukemia cells unless cytoplasmic delivery of oligonucleotides is facilitated (Giles et al., 1995). We previously showed that specific antisense inhibition of TNF- synthesis depends on enhancement of oligonucleotide uptake by cationic lipids (Hartmann et al., 1996a). Antisense oligonucleotides form one of several therapeutic strategies to suppress TNF synthesis (reviewed in Eigler et al., 1997).

Understanding uptake mechanisms and intracellular trafficking of oligodeoxynucleotides in leukocytes forms an important basis for the application of these compounds in experimental and therapeutic settings. We address this issue by characterizing spontaneous and cationic lipid-mediated uptake of oligonucleotides in human monocytes and lymphocytes.

	Methods

Top Abstract Introduction Methods Results Discussion References

Preparation of cells. Human PBMC were isolated from peripheral blood of healthy fasting volunteers by Ficoll-Hypaque gradient centrifugation (Bøyum, 1968) as described (Endres et al., 1991). As a modification of the protocol, centrifugation was performed in tubes containing a horizontal porous filter disc over the Ficoll layer (Leucosep tubes, Greiner, Frickenhausen, Germany) to facilitate layering of blood. Cells were suspended in CO₂-equilibrated RPMI 1640 culture medium (Biochrom, Berlin, Germany) supplemented with 2 mM L-glutamine and 10 mM HEPES (all from Sigma, Munich, Germany). All compounds purchased had been endotoxin-tested. Viability was determined before and after incubation with oligonucleotides by trypan blue exclusion (conventional microscopy) or by propidium iodide exclusion (flow cytometric analysis). In all experiments, 96 to 99% of cells were viable.

Oligonucleotide uptake. Oligonucleotides were synthesized by Eurogentec (Seraing, Belgium). The 18-mer oligonucleotide (molecular weight, 5730 g/mol, 17 negative charges per molecule) used in the experiments presented is complementary to the translation initiation site of TNF mRNA (5' CAT GCT TTC AGT GCT CAT 3'; Hartmann et al., 1996a). In initial experiments identical results on cellular uptake were obtained with a 10-mismatch control-oligonucleotide (5' CTA GGT TTG TCA CCT CTA 3'). Oligonucleotides were completely phosphorothioate-modified and were used conjugated either to FITC (for flow cytometry) or to rhodamine (for confocal microscopy) at the 5'-end. The cationic lipid lipofectin consists of equal parts of DOTMA (monovalent cationic lipid) and DOPE (not charged), and was purchased from Gibco BRL (Eggenstein, Germany). The stock solution of 1 mg/ml lipofectin solution contains 0.5 mg/ml DOTMA (0.75 mM). Each DOTMA molecule contributes 1 positive charge equivalent. The ratio of positive to negative charges was calculated as follows: adding 25 µg/ml lipofectin (18.8 µM, 1 positive charge per molecule) and 1 µM oligonucleotide (17 negative charges per molecule), the ratio is 18.8/17 = 1.1. Oligonucleotides and lipofectin were prepared in a 10-fold concentrated solution in distilled H₂O and were allowed to form oligonucleotide-lipid complexes for 30 min at room temperature. PBMC in 100 µl supplemented CO₂-equilibrated RPMI 1640 culture medium (final density, 5 × 10⁶/ml) were aliquoted in 6-ml polypropylene tubes (Becton Dickinson, Lincoln Park, Sunnyvale, CA). Supplemented culture medium (100 µl) containing FITC-conjugated oligonucleotides alone or in complex with lipofectin were added for the indicated final concentrations. As a control, PBMC were incubated with unconjugated fluorescein together with lipofectin. During the incubation time (5 min to 4 h) tubes were protected from light and were rotated in a hybridization oven at 37°C to minimize adherence of cells. At the indicated time points oligonucleotide uptake was stopped by adding 4 ml ice-cold 1% human serum albumin (Sigma, St. Louis, MO) in PBS. Cells were pelleted at 400 × g, 4°C for 5 min. Washing with 1% human serum albumin in PBS was repeated three times. For kinetic experiments cells were fixed in 500 µl 4% paraformaldehyde in PBS for 10 min at 4°C and were washed once to remove fixative.

Surface antigen staining. To identify subpopulations of PBMC phycoerythrin-labeled monoclonal antibodies were used. Mouse antibodies against human CD3 (T lymphocytes), CD19 (B lymphocytes), CD16 (natural killer cells) and CD14 (monocytes) as well as IgG1 (isotype control) were purchased from Immunotech (Marseille, France). After the final washing step cells were resuspended in 100 µl PBS containing 20 µl antibody solution and were incubated for 15 min at 4°C in the dark.

Flow cytometry. Flow cytometric data on at least 15,000 cells per sample were acquired on an Epics Profile ll (Coulter, Hialeah, FL) equipped with a 15 mWatt argon ion laser (488 nm emission wavelength) and bandpass filters at 525 nm (for FITC), 575 nm (for phycoerythrin) and 650 nm (for propidium iodide). Alignment fluorospheres (Immuno Check, Coulter Corp. Miami, FL) were used for calibration before each measurement. In two-color flow cytometric analyses, spectral overlap was corrected by compensation. Fluorescence detector settings were adjusted, so that stained cells were on scale for each parameter. Data were analyzed with Epics Elite software (Coulter). In the forward/sideward scatter histogram gates were set to include the lymphocyte and the monocyte population. The monocyte gate contained 90% to 95% CD14-positive cells, the lymphocyte gate contained less than 1% CD14-positive cells. FITC-fluorescence intensity was analyzed for cells presenting higher fluorescence than background. The background was defined as the fluorescence of cells incubated with unconjugated fluorescein alone or together with lipofectin, respectively. Nonviable cells were identified by propidium iodide staining (10 µg/ml). Phycoerythrin-conjugated isotype control antibody was used to define nonspecific background of phycoerythrin fluorescence.

Fluorescence microscopy. Each sample of PBMC analyzed by flow cytometry also was examined by fluorescence microscopy. For conventional fluorescence microscopy (Diaphot TMD, Nikon) cells were placed on a chambered coverglass (Nunc Inc., Naperville, IL). For confocal microscopy, PBMC were fixed in 500 µl 1% paraformaldehyde/0.1% glutaraldehyde in PBS for 30 min at room temperature and were placed on a microscopic slide. After drying, slides were mounted with mounting medium (Permafluor, Immunotech). Successive confocal slices (0.5 µm) of cells were scanned by a confocal laser microscope (LSM 400, Zeiss, Oberkochen, Germany). FITC was excited by a 488 nm argon laser, and rhodamine and phycoerythrin were excited by a 543 nm HeNe laser. To obtain two-color images, one section of a double-stained cell (FITC for oligonucleotide uptake and phycoerythrin for surface antigen staining) was scanned twice and analyzed for each fluorescence signal separately. The transmission light image and the two color-coded fluorescence images were superimposed by image processing.

	Results

Top Abstract Introduction Methods Results Discussion References

Spontaneous uptake of phosphorothioate oligonucleotides into monocytes and lymphocyte subpopulations. Freshly isolated human PBMC were incubated with FITC-labeled oligonucleotides at 37°C for 2 h. Uptake of oligonucleotides on a per cell basis was quantified as fluorescence intensity per cell. Percentage of positive cells was determined as proportion of cells with fluorescence intensity higher than 99% of cells of the control sample (cells incubated with unconjugated fluorescein alone). Cellular uptake of phosphorothioate oligonucleotides differed markedly among subtypes of PBMC and was compared determining the percentage of positive cells for each cell type (fig. 1, grey bars). In the lymphocyte population only B cells showed marked oligonucleotide incorporation as reflected by a high percentage of positive cells (92%, mean of three experiments). In contrast, only a low proportion of T cells and NK cells were oligonucleotide positive (2% and 12%, respectively). Almost all monocytes (96%) incorporated oligonucleotides. Comparing the mean fluorescence intensity per cell of all positive cells, oligonucleotide uptake into monocytes (24 units/cell) was 4-fold higher than uptake into B cells (6 units/cell; fig. 1, black bars).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Oligonucleotide uptake into mononuclear cell subtypes in the absence of cationic lipid. Human mononuclear cells were incubated with 1 µM FITC-labeled oligonucleotides for 2 h. After extensive washing, cells were stained with phycoerythrin-conjugated antibodies against subtype-specific surface antigens as described under "Methods." The percentages of fluorescein-positive cells (grey bars) are compared for different lymphocyte subpopulations and monocytes. For B cells and monocytes, in which the proportion of positive cells approximates 100%, mean fluorescence intensity (black bars) reflecting oligonucleotide uptake per cell is indicated. Results are shown as means (± S.E.M.) of three independent experiments.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Time course of oligonucleotide uptake. Mononuclear cells were incubated with 1 µM FITC-labeled oligonucleotides. Incubation was stopped by washing and fixation at different time points. The percentage of lymphocytes (morphological gating) showing higher fluorescence than background is depicted on the y-axis. Results are given as means (± S.E.M.) of three independent experiments.

The molar ratio of lipofectin to oligonucleotide defines optimal cellular uptake. The cationic lipid lipofectin was used to enhance oligonucleotide uptake by forming complexes with the polyanionic oligonucleotide molecules. The concentration of lipofectin was varied to investigate the effects of the ratio of positive-to-negative charges (±-charge ratio) on oligonucleotide uptake. PBMC were cultured with FITC-labeled oligonucleotides (1 µM) alone or in complex with different concentrations of lipofectin for 2 h at 37°C. Flow cytometric analysis of monocytes, gated by size and granularity, revealed that negatively charged lipofectin-oligonucleotide complexes (15 µg/ml lipofectin; ±-charge ratio, 0.7) did not significantly enhance oligonucleotide uptake in monocytes compared with uptake without lipofectin (fig. 3). At an almost balanced charge ratio of 1.1 (25 µg/ml lipofectin) cellular uptake of oligonucleotides was highest. More cationic groups in the lipofectin-oligonucleotide complexes (35 µg/ml lipofectin; ±-charge ratio, 1.5) did not further increase uptake compared with the balanced ratio. Lipofectin, at the optimal concentration of 25 µg/ml, enhanced oligonucleotide uptake into monocytes maximally 10-fold as compared with spontaneous uptake. It led to a more heterogeneous oligonucleotide uptake of the cells as indicated by fluorescence intensity values ranging across three log scales of magnitude.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Facilitating effect of lipofectin on oligonucleotide uptake in monocytes. Mononuclear cells were incubated for 2 h with 1 µM FITC-labeled oligonucleotide alone or together with 15, 25 and 35 µg/ml lipofectin, corresponding to a ratio of positive to negative charges of 0.7, 1.1 and 1.5. In the forward/sideward scatter histogram of the flow cytometer a gate was set to include mostly monocytes. The corresponding fluorescence intensity histograms are shown for each lipofectin concentration. FITC-fluorescence intensity is depicted on the x-axis (logarithmic scale), and the number of cells is counted on the y-axis. Mean fluorescence intensity (MFI) is indicated for each lipofectin concentration in the diagram. Results of one representative experiment are shown.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Optimal molar ratio of lipofectin to oligonucleotide for oligonucleotide uptake. Mononuclear cells were incubated with different concentrations of FITC-labeled oligonucleotide and lipofectin to determine the optimal molar ratio of lipofectin to oligonucleotide. The ratio of positive charges, contributed by lipofectin, to negative charges, contributed by oligonucleotide, is indicated on the x-axis. Oligonucleotide uptake was examined for three different concentrations of oligonucleotides (0.125, 0.5 and 1 µM). For each concentration of oligonucleotide the concentration of lipofectin was adjusted to the indicated charge ratio. For example, at 1 µM oligonucleotide, 25 µg/ml lipofectin was added to obtain a ratio of 1.1 (18.8 µM cationic to 17 µM anionic groups). Uptake of FITC-labeled oligonucleotides is represented by mean fluorescence intensity of monocytes (A) or by percentage of FITC-positive lymphocytes (B). Results are shown as means of two independent experiments.

Lipofectin-mediated versus spontaneous uptake into subtypes of PBMC. Lipofectin-mediated delivery of oligonucleotides was compared with oligonucleotide uptake without lipofectin in monocytes and lymphocyte subpopulations. After 2 h of incubation with 1 µM oligonucleotide alone or complexed to 25 µg/ml lipofectin, monocytes, B lymphocytes, T lymphocytes and NK cells were identified by staining with phycoerythrin-conjugated antibodies. For the two cell types in which the proportion of FITC-positive cells is almost 100% (monocytes and B lymphocytes), the mean fluorescence intensity per cell in these positive cells was quantified (fig. 5, top panels). The bar charts on the left show the lipofectin-mediated increase in this mean fluorescence intensity as means of three independent experiments. The scatter diagram on the right illustrates the actual read-out of one of these experiments. Although the percentage of monocytes and positive B cells remains unchanged in the presence of lipofectin (close to 100%), oligonucleotide incorporation reflected by mean fluorescence intensity was markedly enhanced. Mean fluorescence intensity increased 4-fold, both in monocytes (from 20 to 80 units/cell) and in B cells (from 7 to 30 units/cell). For the two cell types with a low proportion of FITC-positive cells (T lymphocytes and NK cells) the effect of lipofectin on this proportion of positive cells itself was studied (fig. 5, lower panels). In the presence of lipofectin, the proportion of oligonucleotide-positive T cells increased from 2 to 24% (12-fold). The effect on NK cells was less pronounced (13-23%; 1.8-fold).

View larger version (37K): [in this window] [in a new window]   Fig. 5.   Spontaneous versus optimized lipofectin-mediated uptake in mononuclear cell subpopulations. Mononuclear cells were incubated with 1 µM FITC-oligonucleotide alone (grey bars) or together with 25 µg/ml lipofectin (black bars) for 2 h. On the left, mean fluorescence intensity (MFI) for monocytes and B lymphocytes (upper two panels) and percentage of FITC-positive cells for T lymphocytes and NK cells (lower two panels) are shown as means (± S.E.M.) of three independent experiments (bar charts, note different scaling). On the right, the scatter diagrams of one representative experiment comparing spontaneous and lipofectin-mediated uptake are shown. FITC-fluorescence intensity (uptake of oligonucleotides) is depicted on the x-axis. Phycoerythrin-fluorescence intensity (subtype-specific staining) is depicted on the y-axis. A phycoerythrin-conjugated isotype control antibody was used to set the analysis gates (not shown).

Lipofectin changes the intracellular distribution of oligonucleotides. After incubation of mononuclear cells with oligonucleotides or lipofectin-oligonucleotide complexes, fluorescence microscopy was performed in samples prepared in conditions identical with those for flow cytometric analysis. Oligonucleotides were either FITC- or rhodamine-labeled to demonstrate independence of the fluorochrome. Because of fading of FITC, rhodamine-labeled oligonucleotides were applied for photography. Two-dimensional cellular oligonucleotide distribution was analyzed by conventional fluorescence microscopy. Phase-contrast light microscopy was used to ensure the morphologic integrity of cells and to distinguish the monocytes from lymphocytes by size and granularity. Successive optical slices from the top to the bottom of cells were visualized by confocal laser microscopy.

View larger version (62K): [in this window] [in a new window]   Fig. 6.   Time-dependent intracellular accumulation of oligonuceotides. Mononuclear cells were incubated with 1 µM rhodamine-labeled oligonucleotides and examined by conventional fluorescence microscopy at different time points. The upper panel shows fluorescence images obtained after 5, 30 and 60 min of incubation (from left to right). The lower panel shows corresponding cuttings of the light microscopic images (one monocyte and one lymphocyte indicated by M and L, respectively). Different cell density is caused by different dilution of the cell suspension.

View larger version (74K): [in this window] [in a new window]   Fig. 7.   Time-dependent intracellular oligonucleotide distribution in monocytes. Mononuclear cells were incubated with 1 µM rhodamine-labeled oligonucleotide alone (upper panel) or in complex with 25 µg/ml lipofectin (lower panel). Incubation was stopped by washing after 15 min (left panel) or 120 min (right panel). After fixation representative cells were scanned by confocal laser microscopy. The color-coded transmission light image (dark) and the rhodamine fluorescence image (bright) are superimposed in this representation.

View larger version (55K): [in this window] [in a new window]   Fig. 8.   Intracellular oligonucleotide distribution in monocytes and B lymphocytes identified by cell surface antigens. Mononuclear cells were incubated with 1 µM FITC-labeled oligonucleotide and 25 µg/ml lipofectin for 2 h and were then stained with phycoerythrin (PE)-conjugated antibodies to CD 14 (monocytes) and CD19 (B lymphocytes). After fixation cells were analyzed for FITC- and phycoerythrin-fluorescence by dual laser confocal microscopy. Intracellular localization of oligonucleotides is shown for a CD14-positive monocyte (upper panel) and for a CD19-positive B lymphocyte (lower panel). The overlay image on the left shows the FITC-fluorescence image (oligonucleotide uptake, green) and the phycoerythrin-fluorescence image (surface antigen staining, red) superimposed on each other.

	Discussion

Top Abstract Introduction Methods Results Discussion References

In the present study we have characterized uptake and intracellular distribution of phosphorothioate oligonucleotides in subpopulations of freshly isolated human mononuclear cells. Monocytes spontaneously incorporated approximately four times more FITC-conjugated oligonucleotides than B lymphocytes. Uptake of oligonucleotides in T and NK lymphocytes was low and close to the detection limit of flow cytometry. Complex formation of oligonucleotides with the cationic lipid lipofectin strongly enhanced uptake in all cell types (5- to 10-fold in monocytes). This effect depended on the ratio of positive (cationic lipid) to negative (oligonucleotide) charge equivalents (optimal ratio, 1.1) of the lipid-oligonucleotide complex. Lipofectin not only quantitatively enhanced oligonucleotide uptake into monocytes and lymphocytes but also changed the intracellular distribution. In the absence of lipofectin, oligonucleotides accumulated in cytoplasmic vesicles of monocytes and were not detectable in lymphocytes by fluorescence microscopy. In contrast, application of lipofectin was followed by nuclear localization of FITC-oligonucleotides in monocytes and B lymphocytes. In T lymphocytes and NK cells fluorescence staining was weak even in the presence of lipofectin and could not be assigned to subcellular structures.

We found that the combination of flow cytometry and confocal microscopy is a suitable method for studying oligonucleotide uptake into different cell types. Fluorescence staining, as detected by flow cytometry and microscopy, can be ascribed to intact 5'-labeled oligonucleotides. Neither oligonucleotide degradation products nor free fluorescein have been found after 4 h of incubation in primary murine lymphocytes (Zhao et al., 1994). Accordingly, we found no cell-associated fluorescence when cells were incubated with free fluorescein alone or in combination with lipofectin. Furthermore, in the human monocytic cell line HL-60, detection of incorporated oligonucleotides extracted by the slot-blot technique correlated well with flow cytometric analysis of FITC-labeled oligonucleotides (Zhao et al., 1996).

Conventional fluorescence microscopy gave a rough estimate of the localization of cell-associated fluorescence. Differentiation of intra- and extracellularly located signals was achieved by confocal laser microscopy, which confirmed that FITC-oligonucleotides were localized inside the cells. Thus, higher fluorescence intensity observed after incubation in the presence of lipofectin was not caused by surface-bound fluorescence. An influence of fixation on oligonucleotide distribution was excluded by comparing the fluorescence image of fixed and unfixed cells.

Lipofectin was effective only in a narrow range of concentrations. As postulated in another study, the ratio between oligonucleotide and lipofectin concentration seems to be critical (Felgner et al., 1993). Lipofectin is a lipid mixture consisting of equal parts of DOTMA (monovalent cationic lipid) and DOPE (not charged). Penetration of cellular membranes has been proposed to be optimal at a neutral or slightly positive net charge of the lipid-oligonucleotide complexes. The ratio of positive to negative molar charge equivalents for optimal oligonucleotide uptake was 1.1 (18.8 µM cationic to 17 µM anionic groups). This confirmed the optimal experimental condition which we described previously for antisense-mediated TNF inhibition (0.5 µM oligonucleotide, 12.5 µg/ml lipofectin) (Hartmann et al., 1996a). For different cell lines optimal transfection was described for a ratio of cationic lipid to anionic DNA molar charge equivalents between 0.5 and 2 (Felgner et al., 1993). In previous studies we demonstrated that suppression of TNF synthesis by antisense-oligonucleotides was successful only in the presence of lipofectin, whereas phosphorothioate-oligonucleotides without the use of lipofectin even enhanced TNF production (Hartmann et al., 1996a, b).

Cell toxicity is a potential disadvantage of cationic lipids. For permanent cell lines it has been recommended not to exceed a concentration of 20 µg/ml lipofectin in the incubation medium (Felgner et al., 1993). We determined viability of PBMC (cell density, 2.5 × 10⁶/ml) by propidium iodide staining after 2 h of incubation with different concentrations of lipofectin and oligonucleotides. Lipofectin alone was quite toxic for human PBMC at higher concentrations (25 and 35 µg/ml), whereas lipofectin applied in complex with oligonucleotide showed low toxicity, even less than observed with the high concentration of 1 µM oligonucleotide alone. One might speculate that toxicity of oligonucleotides and lipofectin is reduced by formation of complexes with nearly neutral net charge.

In the present study, the application of cationic lipid resulted in nuclear enrichment of oligonucleotides in monocytes and B lymphocytes. We suggest that cationic lipids not only enhance the overall uptake but, perhaps more notably, also improve intracellular availability of oligonucleotides in monocytes and B lymphocytes. Improvement of intracellular availability with cationic lipids has been reported in studies with permanent cell lines. In these studies, fluorescence-labeled oligonucleotides applied in complex with cationic lipids could be detected rapidly in the cytoplasm and in the nucleus. Rapid accumulation of oligonucleotides in the nucleus also occurs when oligonucleotides are injected directly into the cytoplasm of cells (Fisher et al., 1993). Oligonucleotides incorporated by cell lines in the absence of a lipid carrier were trapped in endosomal structures, which thus prevented subsequent enrichment in the nucleus (Bennett et al., 1992). It has been concluded from these studies that oligonucleotides, which are available outside of endosomal structures in the cytoplasmic compartment, can accumulate in the nucleus; and that this cytoplasmic availability forms a prerequisite for antisense oligonucleotides to affect gene expression. Our results confirm that this lipofectin-dependent compartmentalization of oligonucleotides also takes place in primary human monocytes and B lymphocytes.

Large differences between cell types have been reported regarding oligonucleotide incorporation. Primary keratinocytes, for example, displayed nuclear accumulation of phosphorothioate oligonucleotides and specific inhibition of intercellular adhesion molecule-1 expression in the absence of lipofectin. In the same study, however, endothelial cells, smooth muscle cells and fibroblasts did not show oligonucleotide accumulation in nuclei without lipofectin (Nestle et al., 1994).

In contrast to cell lines, freshly isolated human PBMC represent a more heterogenous cell population regarding the state of activation and differentiation. In addition, some variation in cellular oligonucleotide uptake is observed in cells from different blood donors. To date only one publication describes spontaneous uptake of oligonucleotides in primary monocytes (Pirruccello et al., 1994). Despite a 4-fold greater oligonucleotide concentration (4 µM) than in our study, the percentage of oligonucleotide-positive monocytes (75-85%) and B lymphocytes (42-61%) was smaller. Differences in the study design could contribute to this. First, Pirruccello et al. (1994) incubated cells with oligonucleotides in PBS, which contained no calcium. We found in separate experiments that spontaneous oligonucleotide uptake in monocytes and B lymphocytes depends on the presence of extracellular calcium (Hartmann et al., 1997; in press). Furthermore, the oligonucleotide tested by Pirruccello was longer (27 nucleotides compared with the 18-mer in our study). In the study of Pirruccello et al.(1994), no information is given regarding intracellular distribution of oligonucleotides.

Three other studies have described uptake of oligonucleotides in human PBMC. One stated that in preliminary experiments incorporation of oligonucleotides in PBMC was one tenth compared to H9 cells (a human T-lymphocyte cell line). No specification for mononuclear cell subpopulations was performed (Marti et al., 1992). Another group reported the opposite, namely that uptake of oligonucleotides in PBMC is more than two times higher than uptake in H9 cells (Iversen et al., 1992b). However, uptake in mononuclear cell subpopulations has not been examined in this study. A recent report focused on oligonucleotide uptake in leukemic cells (Zhao et al., 1996). PBMC, bone marrow cells and leukemic cells were compared. The findings of this study regarding differences in oligonucleotide incorporation of B and T lymphocytes are confirmed by our results. Intracellular distribution of oligonucleotides as well as uptake in monocytes and natural killer cells were not examined.

Several studies have investigated oligonucleotide incorporation into murine lymphoid cells (Iversen et al., 1992a; Krieg et al., 1991; Zhao et al., 1993, 1994). Higher uptake in B lymphocytes than in T lymphocytes has been reported (Iversen et al., 1992a; Zhao et al., 1993, 1996). Oligonucleotide uptake was heterogeneous and depended on cell differentiation. Intracellular fluorescence was found in a speckled pattern within the cytoplasm sparing the nucleus. These studies did not apply cationic lipids. The mechanisms responsible for oligonucleotide uptake have been studied mainly in hematopoietic cell lines (Beltinger et al., 1995; Stein et al., 1993; Temsamani et al., 1994; Tonkinson et al., 1994).

To our knowledge, the present study is the first to examine the application of cationic lipids and oligonucleotides in primary monocytes and lymphocytes. Despite disadvantages of cationic lipids when applied systemically, locally administered cationic lipids have a high potential to improve efficacy of antisense oligonucleotides. Gene transfer efficacy of locally administered cationic lipids was demonstrated in a controlled trial in patients with cystic fibrosis (Caplen et al., 1995). In analogy, the application of antisense oligonucleotides combined with cationic lipids could be feasible in a local compartment such as joints or sections of the gastrointestinal tract. Local administration of antisense oligonucleotides is a promising strategy for specific suppression of target proteins in vivo. Recently, Neurath et al. (1996) reported that intraluminal application of antisense oligonucleotides against the transcription factor NF-B abolished established experimental colitis in mice. Clinical trials testing local administration of antisense oligonucleotides against cytomegalovirus retinitis in patients suffering from AIDS have reached study phase III (Cohen, 1995; Hawkins, 1995).

Together with our previous results on oligonucleotide-mediated suppression of TNF synthesis, the present study provides the information necessary for antisense-mediated inhibition of protein synthesis in human monocytes. This study demonstrates that the beneficial effect of cationic lipids for antisense-mediated inhibition of protein synthesis in monocytes is based on the direct release of oligonucleotides into the cytoplasm and the nucleus, bypassing the endosomal compartment. Further studies must examine the therapeutic potential of local application of antisense oligonucleotides together with cationic lipids directed against inflammatory mediators such as TNF in vivo.