Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of gene function with the aid of antisense technology provides an opportunity to dissect out the biological function among proteins that share significant sequence homology. Recently a locus was identified in which three different genes (BAH, humbug, and junctin) share exons but differ significantly in sequence composition based upon the use of alternative terminal exons (Dinchuk et al., 2000). Although junctin has been assigned a function as an integral member of the ryanodine receptor complex (Jones et al., 1995; Zhang et al., 1997, 2001), the possible functions of BAH and humbug remain unclear. In the mouse, there are three BAH-related transcripts that are 2.8, 4.5, and 6.6 kb. The two larger transcripts encode full-length catalytically active BAH protein and differ only in the use of alternative polyadenylation signals. The shortest transcript, humbug, which is also known as junctate, encodes a highly charged protein with no known catalytic function (Dinchuk et al., 2000). Full-length BAH protein catalyzes the addition of hydroxyl groups to particular Asp or Asn residues within cEGF domains of numerous proteins, including clotting factors, Notch and its ligands, and a variety of other biologically important molecules (Rebay et al., 1991; Stenflo, 1991; Downing et al., 1996; Goruppi et al., 1997). The function(s) of this hydroxylation has not yet been defined, although the fact that some of these motifs are involved in protein-protein interaction has raised the possibility that hydroxylation may modulate certain receptor-ligand interactions (Monkovic et al., 1992; Lavaissiere et al., 1996). Some data suggest that the charged region of humbug may play a role in modulating calcium release (Dinchuk et al., 2000; Treves et al., 2000).

Because overexpression of BAH and humbug has been associated with a variety of human cancers, a possible role for BAH in tumorigenesis has been suggested (Lavaissiere et al., 1996; Ince et al., 2000). There are three BAH-related transcripts in humans that are 5.2, 4.5, and 2.9 kb. Interestingly, humbug (2.9 kb), which lacks catalytic activity, is as abundantly expressed as BAH (5.2 and 4.5 kb) in all the different cell lines examined (vide infra). To investigate possible functions of the -hydroxylase isoforms, antisense ON tools capable of distinguishing between full-length BAH and humbug were developed. Ability of these ONs to inhibit mRNA and protein expression was demonstrated in vitro in the A549 human lung carcinoma line, and several ONs were further characterized through dose-response and time course studies. Application of fluorescein-labeled phosphorothioate ONs to an in vivo xenograft model revealed a surprising inability of these antisense reagents to effectively penetrate the tumor. Modest effects on tumor growth, together with the absence of -hydroxylase down-regulation in the tumor upon antisense treatment are consistent with the poor tumor uptake of ONs observed. Investigation into the possible functions of -hydroxylase was therefore inconclusive. These data highlight the need for strategies that will enhance delivery of phosphorothioate ONs to solid tumors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

ON Synthesis. All ONs used in these studies, with the exception of ON 1001, were 20 nucleotide 2'-deoxyribonucleotide phosphorothioate ONs and were synthesized and purified as described previously (Ho et al., 1996). Sequences of ONs used are as follows: ON 79, GGCATTCTTACGCTGGGCCA; ON 79M, GGAATTGTTAGGCTCGGACA; ON 1302, GGGACATCAGGTAGGCTGGC; ON 1302M, GGGCATGAGGAAGGGTGGC; ON 1001, agcTTTAAGTATCTGGTGGTac; ON 1001M, agcGTTAGTAACTCGTGTac; and ON 962 (an ON sequence unrelated to the -hydroxylase gene), TGACGCAGCGGCACCAGACC. Uppercase bases denote 2'-deoxyribonucleotide residues, and lowercase bases denote 2'-methoxy-ribonucleotide residues. Underlined nucleotides form mismatches with the target RNA. ONs conjugated with fluorescein at the 5' end were synthesized using fluorescein phosphoramidites (Glen Research, Sterling, VA) according to manufacturer's instructions.

Cell Culture. Human A549 lung carcinoma cells were obtained from the American Type Culture Collection (Rockville, MD) and cultured in Dulbecco's modified Eagle's medium containing 1 g/l glucose (Invitrogen, Carlsbad, CA). The medium was supplemented with 10% fetal bovine serum (FBS), 2 M glutamine, 100 units/ml penicillin, and 10 µg/ml streptomycin. Cells were routinely passaged at 85 to 90% confluence in T-225 flasks. The cells were plated onto 60-mm culture dishes and 100-mm culture dishes for Northern or Western blot analysis, respectively.

Treatment of Cells with ONs. A549 cells at 80 to 85% confluence were washed twice with prewarmed Opti-MEM (Invitrogen). Opti-MEM containing lipofectin (Invitrogen; 10 µg/ml for Northern analysis or 12.5 µg/ml for Western analysis) was added to the cells. The ONs were then added from a 10 µM stock (0.4 µM final ON concentration). The cells were incubated for 4 h at 37°C, and the medium containing the lipofectin was aspirated off. The cells were washed twice with Dulbecco's modified Eagle's medium/10% FBS and incubated with Dulbecco's modified Eagle's medium/10% FBS. Cells used for Northern blot analysis were incubated for 16 h, whereas cells used for Western blot analysis were incubated for 72 h or longer, and a second application of ON was applied at the 48-h time point.

Northern Blot Analysis. Total cellular RNA was isolated using the RNeasy Mini kit (QIAGEN, Valencia, CA) according to the manufacturer's protocol. Total RNA (10-15 µg) was poly(A)⁺ selected using the MicroPoly(A) ⁺ Pure kit (Ambion, Austin, TX). The mRNA was precipitated and separated on a 1.2% agarose/formaldehyde gel, transferred to a nylon membrane (Hybond N; Amersham Biosciences, Piscataway, NJ) by capillary action, and cross-linked. The blots were prehybridized for 1 h at 42°C in UltraHyb (Ambion) and simultaneously probed using cDNAs specific for BAH and for glyceraldehyde-3-phosphate dehydrogenase (G3PDH). The probes were generated by random priming and radiolabeled using [-³²P]dCTP according to manufacturer's protocol (DECAprime II; Ambion). G3PDH, BAH, and humbug bands were visualized and quantified using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Potential unequal loading of the lanes was corrected for by normalizing the BAH and humbug bands to the G3PDH present in each lane.

Western Analysis. Cells treated with antisense ONs were harvested and resuspended in 1.1× LDS sample buffer (Invitrogen) to a final concentration of 2 × 10⁶ cells/ml. Protein quantitation was performed using the DC protein assay (Bio-Rad, Hercules, CA). Twenty-five micrograms of protein lysate was fractionated on a 4 to 20% acrylamide Tris glycine gel (Novex, San Diego, CA) for 1 h at 150 V. Protein was transferred onto a nitrocellulose membrane by electrotransfer and blocked in 5% nonfat milk in phosphate-buffered saline/0.5% Tween. Blots were probed with primary antibody (FB-50, 1:5000 dilution; anti-G3PDH, 1:300 dilution; or actin). FB-50 was the kind gift of Dr. Jack R. Wands (Lavaissiere et al., 1996). FB-50 incubations were carried out at 4°C overnight, whereas the G3PDH antibody was incubated for 1 h at room temperature. Secondary antibody detection was performed using the Western Breeze kit (Invitrogen) according to manufacturer's instructions.

Experimental Animals. Female athymic Swiss nu/nu mice at 6 to 8 weeks of age were obtained from Taconic Farms (Germantown, NY) and were free of known pathogens at the time of use. The animals were housed 10 per cage in polycarbonate, filter-capped micro-isolation cages in temperature-controlled rooms maintained in a barrier facility on 12-h light/dark cycles and provided food and water ad libitum. All animal studies were conducted in a facility accredited by The American Association for the Accreditation of Laboratory Animal Care.

A549 Xenograft Study. A549 human lung epithelial cells were implanted s.c. in the inguinal area of the nude mice at 1 × 10⁷ (0.01 ml) per mouse. Tumors were measured using perpendicular diameters, and the volume was calculated using the formula for a prolate ellipsoid as described previously (Wexler et al., 2000). Before the first dose, the animals were randomized so there were no significant differences in tumor volumes among groups (75 ± 5 mm³). Starting on day 21, the ONs 79 and 79M were injected s.c. at 13.5 mg/kg/day, rotating injection sites to prevent scarring. At the termination of the study, tumors were removed and weighed. Tumor tissue was trimmed, connective tissue was removed, and the tumor was cut into four pieces and stored in RNALater (Ambion) at 20°C until processed for Northern blot analysis.

Visualization of Fluorescent ONs. In the first uptake experiment, tumor-bearing mice from the above-described xenograft study were injected with fluorescein-labeled ON 79 after having received 13.5 mg/kg of ON 79 for 14 days. In the second study, tumor-bearing mice were injected with 13.5 mg/kg of either ON 79, ON 1302, or ON 962. All mice were euthanized 4 and 24 h after administration of fluorescein-labeled ON and the tumor, liver, and kidney were collected immediately into 10% neutral buffered formalin. Tissues were processed and paraffin embedded according to the following schedule: 6 h in 10% neutral buffered formalin; 3 h in 80% ethanol; 1 h in 95% ethanol 2×; 1 h in 100% xylene 3×; 1 h in paraffin; and finally, 0.5 h in paraffin (Tissue Processor Tissue-Tek II; Labtek, Tokyo, Japan). Tissue blocks were sectioned at 4 µm (Microtome Cut 4055; Olympus, Lake Success, NY), and sections were placed onto silane-coated glass slides (Polysciences, Warrington, PA) and stored at 80°C. Samples were rehydrated the day of visualization, counterstained, and coverslipped with Vectashield containing propidium iodide (Vector Laboratories, Burlingame, CA). Separate photomicrographs of the tissue for fluorescein and propidium iodide were collected and overlaid to create two color images of the distribution of the ONs within the tissues. Fields were viewed at 10, 40, and 100× in oil, using the AX70 microscope (Olympus), the MCID M5 image collection system (Imaging Research, St. Catherines, Ontario, Canada), and a videocamera (DXC-970MD; Sony, Tokyo, Japan).

Processing of Tumor for Northern Analysis. Tumor sections were removed from buffer and cut into smaller pieces in a sterile Petri dish. The pieces were placed in a 14-ml Falcon tube with buffer RLT (QIAGEN Midi RNA kit) and homogenized with a polytron for 30 to 45 s. The resulting lysate was frozen at 80°C until ready for use. The samples were thawed in a 37°C bath to eliminate any precipitation. Total RNA was isolated using the QIAGEN Midi RNA kit.

Statistics. Statistical differences were determined using the Student's t test. A p < 0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

BAH RNA transcripts obtained from an expression plasmid encoding the full-length coding sequence of the BAH gene were probed with ON libraries to identify sites in the RNA that may be accessible to antisense ON hybridization (Ho et al., 1996, 1998, 1999). Such regions in the BAH RNA were located by using ribonuclease H, which produces endonucleolytic cleavages in the RNA only at sites that are hybridized with DNA (from the ON libraries). Sequencing of the resulting RNA fragments revealed the location of these accessible sites. Antisense ONs, 20 nucleotides in length and consisting of the 2'-deoxyribonucleotide phosphorothioate chemistry, were synthesized to target these sites (Fig. 1A).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   A, location of 20-nucleotide antisense ONs and epitope site for FB-50. ONs are identified using their 5'-nucleotide position (accession no. U03109). Eight ONs, labeled A through H, were directed against the 3'-untranslated region (exon 14a) of the humbug mRNA. B, antisense inhibition of BAH and humbug RNA. The degree of inhibition was determined by Northern analyses and expressed relative to untreated cells. Top, 5.2-kb BAH RNA; middle, 4.5-kb BAH RNA; bottom, 2.9-kb humbug RNA. Results are an average of at least two experiments. Error bars represent standard deviations.

To determine the ability of these ONs to interfere with BAH expression, RNA isolated from ON-treated A549 cells was examined by Northern blot analysis. The degree of reduction of -hydroxylase RNA transcripts was determined relative to RNA obtained from cells incubated in the absence of any ON. Alternate splicing at nucleotide position 1011 results in the 2.9-kb humbug message. Most of the ONs tested produced substantial reductions of BAH RNA (Fig. 1B). Sequences upstream of nucleotide 1011 down-regulated both BAH and humbug expression. However, sequences targeted downstream of nucleotide 1011 were complementary only to the full-length BAH transcript and were ineffective at reducing levels of humbug.

Eight different antisense ONs directed against exon 14a of humbug (Fig. 2A) were chosen based on thermodynamic considerations (Oligo Primer Analysis Software, version 5.0; National Biosciences, Plymouth, MN). Although all eight sequences (ONs A-H) reduced humbug expression (Fig. 2B), these ONs did not exhibit the desired selectivity for humbug but reduced expression of full-length BAH RNA to approximately the same degree as well.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   A, differential splicing of the BAH gene results in two different transcripts. Eight antisense ONs, A through H, were targeted against exon 14a of humbug. The 5'-nucleotide position of the ONs (humbug gene accession no. AF289489) are as follows: 973 (A), 990 (B), 1305 (C), 1845 (D), 1884 (E), 1929 (F), 2266 (G), and 2339 (H). ON 1001 was directed against the exon 13-exon 14a junction of humbug. B, effect of ONs A through H on BAH and humbug RNA expression. Solid, average of the 5.2- and 4.5-kb BAH RNA; waves, humbug RNA. C, dose response of ONs 79, 1302, and 1001 and their mismatch controls on BAH and humbug RNA expression. The degree of inhibition of BAH mRNA expression is an average of the inhibition of the 5.2- and 4.5-kb transcripts. Results in B and C are an average of at least two experiments, with error bars representing standard deviations.

To enable selective targeting of humbug, an ON was designed to hybridize across the exon 13-exon 14a junction that is unique to the humbug transcript (Fig. 2A). Analysis of this antisense sequence revealed that the ON possessed low thermodynamic stability due to a preponderance of A-T base pairs, and this was confirmed by the poor efficacy of the ON when tested in A549 cells. The 2'-methoxy-ribonucleotide chemistry was therefore incorporated to increase the affinity of this sequence for its target RNA. The resulting ON 1001 effectively reduced levels of humbug without producing significant effects on BAH mRNA.

In addition to ON 1001, which is selective for humbug, two other sequences, ON 79 (which targets both BAH and humbug) and ON 1302 (which is selective for BAH RNA), were chosen for further characterization. Mismatch control ONs were synthesized and dose-response experiments were conducted on the antisense and mismatch molecules. Although the antisense ONs produced dose-dependent effects on RNA expression (ON 79 on all three RNA molecules, ON 1302 on BAH RNA, and ON 1001 on humbug RNA), their mismatch control sequences had minimal effects (Fig. 2C).

Examination of protein lysates by Western analysis showed a substantial degree of protein inhibition 72 h after antisense treatment with ONs 79, 1302, and 1001 (Fig. 3A). As in the Northern analyses, target selectivity was observed. ON 79 inhibited expression of BAH and humbug proteins, whereas ON 1302 selectively reduced BAH protein expression. ON 1001 reduced humbug and not BAH protein levels. The mismatch control ONs did not alter BAH and humbug protein expression levels to any appreciable extent. An 8-day study showed that protein expression could be reduced for an extended period by repeated administration of ON 1302 (on days 0 and 2) (Fig. 3B). BAH protein expression recovered upon withdrawal of the antisense ON. The pattern of expression of BAH and humbug was examined in several lung cell lines, including A549 cells (Fig. 4). Northern analysis showed the presence of BAH and humbug mRNA in all cells lines with the exception of the adenocarcinoma H522 cells (Fig. 4A). Interestingly, despite the absence of catalytic activity in humbug, the message and protein of this -hydroxylase isoform are expressed in substantial quantities relative to full-length BAH.

View larger version (41K): [in this window] [in a new window]   Fig. 3.   A, selective inhibition of protein expression by ONs 79, 1302, and 1001 determined 72 h after ON treatment. B, time course of antisense inhibition by ON 1302 and its mismatch control. The ONs were administered on days 0 and 2 for 4 h at a time.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   A, Northern blot analysis of five lung cell lines. A549, lung carcinoma; WI-38, normal lung fibroblast; H522, lung adenocarcinoma (nonsmall cell); H358, lung bronchioalveolar carcinoma (nonsmall cell); and H460, lung carcinoma (large cell). B, Western blot analysis of five lung cell lines.

Nude mice bearing A549-derived tumors were dosed subcutaneously once a day at 13.5 mg/kg with ON 79 and its mismatch control sequence ON 79M. An uptake study designed to determine the efficiency of phosphorothioate ON uptake and distribution in A549 tumors was conducted simultaneously as part of the xenograft experiment. Because fluorescence microscopy is among the simplest and most sensitive of methods for determining ON localization (Butler et al., 1997), uptake was investigated using fluorescein-labeled ON 79. Two of the mice were injected with a single dose of fluorescein-labeled ON 79 after 14 consecutive days of dosing with unlabeled ON 79. H&E staining revealed tumors that were encapsulated, poorly vascularized, and heterogeneous in nature (Fig. 5, A and B). Areas in the periphery of the tumor seemed viable, whereas regions within the tumor interior contained a mixture of necrotic cells and morphologically viable cells. Although fluorescein-labeled ON 79 was consistently observed within the connective tissue and the periphery of the tumor (Fig. 5, C and E), there was poor penetrance of the ON to the interior regions of the tumor (Fig. 5, D and E). However, fluorescent signals that were observed in the tumor interior were in the same planes of focus as the nuclei and the cytoplasm of the tumor cells (Fig. 5E), indicating some intracellular localization.

View larger version (120K): [in this window] [in a new window]   Fig. 5.   Tumor photomicrographs 4 h after injection of fluorescein-labeled ON 79. A, H&E section of tumor, including capsule at 20×. B, composite image (tiled at 10×) of tumor stained with H&E. C and D, fluorescence photomicrographs of tumor at 40×. E and F, same tumor sections at 100× under oil.

Figure 6A shows the tumor volumes of the mice from days 21 to 44. The average tumor volume of the antisense-treated group was significantly less (p < 0.05) than either of the other two control groups on days 39 and 44 of the study. ON 79 produced a 23% reduction in tumor volume. At the conclusion of the experiment, RNA from the tumors was examined by Northern analysis (Fig. 6B). No differences in BAH or humbug RNA levels were observed among the various treatment groups. This finding was confirmed by real-time polymerase chain reaction analysis of the RNA.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   A, effect of ON 79 on the growth of A549 tumors in nude mice. ON administration was initiated on day 21 (n = 8 to 10 mice/group). Error bars represent standard error of the mean. p < 0.05 for the ON 79 group with respect to the vehicle and ON 79M groups. B, effect of ON 79 on tumor -hydroxylase mRNA expression. Solid, vehicle (n = 8); waves, ON 79 (n = 8); and dots, ON 79M (n = 8). RNA levels are determined relative to one of the vehicle-treated mice. Error bars represent standard error of the mean.

In a separate experiment, three different ONs were injected subcutaneously into nude mice to determine whether a sequence-dependent effect on ON uptake exists in tumors. The results obtained do not reveal any appreciable differences in the pattern of ON uptake (Fig. 7, A-D). Consistent with data obtained previously using ON 79, ONs 1302 (Fig. 7C) and 962 (Fig. 7D) did not penetrate into the tumor interior to any significant extent but seemed to be largely associated with the extracellular matrix. This observation is in contrast with Fig. 7, E and F, which shows effective penetration of ON into the liver and kidney. As with the tumor cells, fluorescent ON was found within the hepatocytes (Fig. 7E) and epithelial cells of the kidney (Fig. 7F) (fluorescent signals were in the same planes of focus as the nuclei and the cytoplasm). The distribution pattern of ON in all the tissues examined was stable for at least 5 months.

View larger version (129K): [in this window] [in a new window]   Fig. 7.   Photomicrographs of tumor, liver, and kidney taken from mice 4 or 24 h after injection with fluorescein-labeled ONs. A549 tumor micrograph from an ON 79-injected mouse at 4 h (A), tumor at 24 h after ON 79 injection (B), tumor from an ON 962-injected mouse at 4 h (C), tumor from an ON 1302-injected mouse at 4 h (D), liver from an ON 79-injected mouse at 4 h (E), and kidney from an ON 79-injected mouse at 4 h (F). All micrographs are at 100× under oil.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Overexpression of -hydroxylase protein has been reported in various carcinomas and transformed cell lines (Lavaissiere et al., 1996). In those studies, -hydroxylase was detected using the FB-50 antibody, which we now know recognizes both full-length and truncated forms of the protein (Dinchuk et al., 2000). humbug lacks the catalytic domain of the -hydroxylase protein, but in spite of the absence of enzymatic activity, it is abundantly expressed in various lung cell lines (Fig. 4) and other transformed lines (data not shown). These observations raise the question of the role of BAH and humbug in pathophysiology, particularly in cancer. Therefore, in an attempt to study the biological role of BAH and humbug, an antisense ON approach was used to selectively and differentially down-regulate the expression of the various -hydroxylase isoforms.

Because RNA molecules have extensive secondary and tertiary structure, not all regions of an RNA molecule are equally accessible to hybridization with an antisense ON. The RNA mapping approach facilitated the identification of numerous active sequences both upstream and downstream of the exon 13-exon 14 splice junction of BAH, with ONs targeting regions downstream of the splice junction being selective for only BAH mRNA (Fig. 1B). Antisense ONs directed against the 3'-untranslated region of humbug (exon 14a) unexpectedly reduced both humbug and BAH mRNA expression. In addition, each of these ONs (A-H) reduced BAH and humbug to similar levels (Fig. 2B). These observations suggest the possibility that ONs A through H may be acting on a molecular target common to BAH and humbug, the precursor mRNA of the hydroxylase gene before splicing, and imply that inhibitory activity of the hydroxylase antisense ONs may be occurring in the nucleus. A strategy to target a site that is unique to mature humbug mRNA, the exon 13-exon 14a junction, yielded a sequence that selectively down-regulates humbug mRNA. Western blot analyses with FB-50 show the presence of three protein bands. Western blotting analysis of antisense studies with ONs 79, 1302, and 1001 demonstrates that the higher molecular weight proteins (M_r, ~130,000) are encoded for by the 5.2- and 4.5-kb BAH transcripts, and allows the correlation of the 2.9-kb humbug transcript with the lower molecular weight protein (M_r, ~55,000) [The small differences in relative molecular weights of these hydroxylase proteins, compared with the molecular weights reported previously (Dinchuk et al., 2000), are due to the different gel systems used in these two studies]. The ability of these particular antisense ONs to selectively modulate expression of BAH and humbug at the mRNA and protein level highlights their potential as tools to dissect out the biological function of BAH and humbug.

Although the pharmacokinetics of phosphorothioate ONs in tumors has been reported previously (DeLong et al., 1997), to the best of our knowledge, the spatial localization of these ONs in tumors has not been investigated. Hence, a fluorescence microscopy uptake experiment was conducted concurrently with the antisense tumor xenograft study. The pattern of distribution and persistence of ON in the peripheral organs we examined were consistent with the findings of other investigators (Rifai et al., 1996; Butler et al., 1997). For example, prominent ON labeling was observed in the proximal tubule cells of the kidney, but there was little uptake in the glomerulus and cells in Bowman's capsule. ON was also observed in hepatocytes. The penetrance of fluorescein-labeled ON 79 into the tumor, however, was in stark contrast to its abundant distribution in the liver and kidney. Although a strong presence of ON was observed in the connective tissue capsule and extracellular space in the outer layers of the tumor, regions within the interior of the tumor were sparsely labeled. Hence, although tumors seem to accumulate modest amounts of phosphorothioate ON (2-3% of total injected dose per gram of tumor tissue) (DeLong et al., 1997), in A549 tumors this study shows that ON is largely absent from the tumor cells themselves. Nonspecific binding of phosphorothioate ONs to extracellular matrix proteins such as laminin and fibronectin (Guvakova et al., 1995; Khaled et al., 1996) may account for the retention of ON in the stroma and may have hindered efficient ON penetration from the extracellular matrix into the tumor cells.

In in vivo studies with ON 79, a modest, although significant reduction in tumor volume was observed. However, analysis of the RNA derived from those tumors did not reveal inhibition of BAH or humbug mRNA. Because RNA was processed from opposing quadrants of the tumor mass, the majority of the RNA extracted would have come from regions of the tumor with poor ON accumulation. This may explain the lack of down-regulation of -hydroxylase message. A tumor uptake study with phosphodiester ONs has been reported (Plenat et al., 1995). In that study, 3'-fluorescein or 3'-digoxigenin-conjugated phosphodiester ONs were reported to penetrate the extracellular matrix into the tumor interior, a finding that is contradictory to our results. This difference may be due to the decreased stability of phosphodiester ONs, even when partially protected at the ends (Hoke et al., 1991; Fisher et al., 1993; Sands et al., 1995). The half-lives of various 3'-protected phosphodiester ONs in mouse serum ranged from 10 to 60 min (Sands et al., 1995). In contrast, phosphorothioate ON incubated in mouse serum remained unchanged even after 5 h (Sands et al., 1994). The Plenat study did not provide any direct evidence that the fluorescein or digoxigenin signal observed within the tumors derived from intact, full-length ON.

We explored the possibility that ON penetrance into tumors may be sequence-dependent. Rifai compared the uptake of two ON sequences that differed substantially in their purine content and found differences in the amount of ON taken up by the heart, liver, and skin (Rifai et al., 1996). However, our study comparing three different ON sequences that also differed considerably in their purine content did not reveal a sequence-dependent effect on uptake by A549 tumors.

In light of the poor ON penetrance into A549 tumors, and the absence of down-regulation of -hydroxylase mRNAs, we are uncertain as to whether the modest inhibition of tumor growth observed derived from actual antisense effects. The finding of poor tumor penetrance suggests additional areas of investigation and highlights the importance of demonstrating significant uptake of ON by the target tissue of interest. A possible approach to increasing ON delivery to the tumor may be through an i.v. route of administration. The i.v. route produces 5-fold greater peak plasma levels of ON than subcutaneous dosing (Phillips et al., 1997) and was used in other A549 xenograft studies (Dean et al., 1996; Monia et al., 1996; Monia, 1999). For poorly vascularized tumors such as these A549 tumors, the higher peak plasma levels may help produce more effective ON delivery to the tumor. An alternative approach may involve continuous infusion of the ON through osmotic mini pumps. Because a high concentration of ON seems to be retained by the extracellular matrix of the tumor, strategies that decrease adherence to extracellular matrix proteins may allow better diffusion and penetration of the ON into the tumor cells. The polyanionic nature of phosphorothioate ONs contributes to its affinity for numerous heparin-binding proteins, including proteins of the extracellular matrix (Guvakova et al., 1995; Khaled et al., 1996). Although phosphodiester ONs also suffer from this nonsequence-specific effect, the presence of the sulfur moiety on phosphorothioate ONs enhances such protein-binding substantially (Khaled et al., 1996). A possible solution to this problem may be to replace a portion of the phosphorothioate residues with uncharged, nuclease-resistant nucleotide analogs such as methylphosphonates (Giles et al., 1995) and polyamide nucleic acids (Pooga et al., 1998) or with second generation nucleotide analogs such as methoxyethyl-ribonucleotide residues (Zhang et al., 2000), phosphoramidate nucleotides (Faria et al., 2001), or locked nucleic acids (Wahlestedt et al., 2000).

In summary, antisense ONs capable of selectively down-regulating the expression of closely related isoforms of -hydroxylase have been developed. Studies to investigate phosphorothioate ON uptake in tumors after subcutaneous administration revealed poor penetrance of the ONs in A549 tumors. This finding is consistent with the modest effect on tumor growth observed and the lack of -hydroxylase mRNA down-regulation. Although investigation into the in vivo function(s) of BAH and humbug was inconclusive, nonetheless, the work reported above provides a foundation for additional studies on this important family of proteins.