Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Phosphorothioate oligonucleotides are designed to hybridize to a specific mRNA, thus inhibiting the expression of a specific target gene (Crooke, 1992). Because of Watson and Crick base paring rules, it is possible to make an oligonucleotide that selectively binds to a specific mRNA effectively inhibiting protein expression and, thus, enabling the design of oligonucleotide agents with specific therapeutic activities. Presently, phosphorothioate oligonucleotides are being developed as therapeutic agents in human immunodeficiency virus, cytomegalovirus, cancer, arthritis, inflammatory bowel disease and organ transplantation (Crooke, 1995; Wagner, 1994). The use of these agents will be dependent on both the therapeutic activity and their toxicity profile. In this report we document the dose-response relationship and the biochemical mechanism of a phosphorothioate oligonucleotide-mediated toxicity.

The subchronic toxicity of phosphorothioate oligonucleotides has been investigated in rodents and primates (Srinivasan and Iversen, 1995), and a pattern of species-specific responses has been characterized that is consistently observed with most compounds in this class independent of sequence. Several laboratories have documented a spectrum of changes in mice and rats that can be collectively characterized as immune stimulation after systemic treatment with phosphorothioate oligonucleotides for several days or longer (Henry et al., 1996b; Sarmiento et al., 1994). Although there are differences in the quantitative dose-response relationship for immune stimulation that appear to be sequence-related, the qualitative similarities in the effects of various phosphorothioate oligonucleotides in rodents are suggestive of some common toxicologic properties for this class of molecules (Henry et al., 1997).

A somewhat different spectrum of toxicities is induced by phosphorothioate oligonucleotides in nonhuman primates. Clinical observations in monkeys (i.e., macaques) given large systemic doses include transient lethargy, periocular edema, susceptibility to bruising and acute mortality (Cornish et al., 1993; Galbraith et al., Henry et al., 1996a). Further investigation in instrumented anesthetized animals revealed pronounced hemodynamic alterations, including central hypotension and reductions in cardiac output that were associated with activation of complement (Galbraith et al., 1994). Although other studies have described complement activation associated with administration of phosphorothioate oligonucleotides (Galbraith et al., 1994), this study fully characterizes the dose-response relationship and the pattern of complement split product formation. The objective of this study was to further characterize the relationship between plasma concentration of phosphorothioate oligonucleotides and complement activation, to establish safe plasma levels of oligonucleotide and to define a potential mechanism of complement activation. The data obtained, particularly the existence of a threshold concentration for complement activation, are highly relevant to the continued safe use of phosphorotioate oligonucleotides.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Test material. ISIS 2302 (M_r = 6781) is a 20-base phosphorothioate oligonucleotide (5-GCCCAAGCTGGCATCCGTCA-3) with a sequence that is complementary to a 3-untranslated region of the human intracellular adhesion molecule-1 mRNA. ISIS 2302 was synthesized by Isis Pharmaceuticals (Carlsbad, CA) (purity >92% full-length oligomer). ISIS 2302 was formulated in phosphate-buffered saline (pH 7.4) at a concentration of 4 mg/ml.

Animals. Fifteen female cynomolgus monkeys (Worldwide Primates, Inc., Miami, FL) were randomly chosen and assigned to dose groups based upon body weight. Body weight ranged from 2.5 to 3.5 kg, but the exact age was unknown. Monkeys were housed in elevated metal grid cages and were maintained in an environmentally controlled room (12-hr light/dark cycle) with ad libitum access to water. The amount of food presented each day was approximately 4% of the mean body weight for all animals in the room. All animal husbandry procedures were in full compliance with AAALAC guidelines.

Study design. Monkeys received single doses of 2, 3.3, 6.6, 10 or 20 mg/kg of ISIS 2302 or a vehicle control solution by i.v. infusion for periods ranging from 2 to 120 min (N = 3-4 per group). Some monkeys were used repeatedly (up to 5 separate treatment regimens) for evaluation of acute toxicity after a 2-week wash out period. Monkeys were observed continuously for the first several hours after treatment to document acute clinical signs of toxicity. Blood samples were obtained at various intervals throughout the dosing or monitoring periods in order to evaluate plasma concentrations of ISIS 2302 and potential complement activation or alterations in hematology parameters.

Hemodynamic evaluation. Alterations in mean arterial pressure, heart rate and cardiac output were assessed after a 10-min infusion of 20 mg/kg of ISIS 2302 in ketamine-anesthetized monkeys. A catheter was placed in the femoral artery for recording central pressure and heart rate. A Swan Ganz catheter was placed in the right atrium for determination of cardiac output by the thermal dilution method. ECG recordings were also obtained during this period. Continuous on-line monitoring was performed during the infusion and for 1 to 2 hr after using a MACLAB (FID Instruments, Castle Hill, Australia) computerized multichannel chart recording system.

Analysis of complement activation. The level of complement split products Bb, C3a_{des arg,} C4a_{des arg} and C5a_{des arg} was determined in EDTA plasma samples using commercially available radioimmunoassay or enzyme-linked immunosorbent assay kits. C3a_{des arg}, C4a_{des arg} and C5a_{des arg} were measured using a radioimmunoassay kit from Amersham Life Sciences (Amersham, Little Chalfont, Buckinghamshire, England). The Bb fragment of Factor B was determined using an enzyme immunoassay from Quidel (San Diego, CA).

Factor H binding assay. The relative affinity of Factor H for heparin [Sigma Chemical Co. (St. Louis, MO); M_r = 16,000-17,000] and ISIS 2302 was estimated in a competitive binding assay as described previously (Pangburn et al., 1991). Radiolabeled Factor H was bound to a heparin-sepharose solid-phase column. Solutions containing increasing concentrations of heparin or ISIS 2302 were added to the column. Column eluates were analyzed to determine the amount of Factor H eluted.

ISIS 2302 analysis. Plasma concentrations of intact ISIS 2302 were determined from EDTA plasma using capillary gel electrophoresis and UV detection (Leeds et al., 1996). Separation and resolution of intact ISIS 2302 and its metabolites were obtained using polyacrylamide gel-filled columns under denaturing conditions (7 M urea). Quantitation was based on the UV absorbance at 260 nm of ISIS 2302 peaks relative to an internal standard.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Complement activation and physiological response. Intravenous infusion of 20 mg/kg of ISIS 2302 over 10 min results in strong activation of the complement cascade as indicated by a reduction in CH50 activity and an increase in the biologically active split products C3a and C5a over time (fig. 1, A and B) (Hugli and Muller-Eberhard, 1978). The 60% decrease in total hemolytic complement activity, CH50, suggests extensive complement activation. In addition to the appearance of the anaphylatoxins C3a and C5a, there was a marked increase in Bb, but no significant change in the marker for the classical pathway, C4a (fig. 1A). C3a and C5a are common to both classical and the alternative pathway activation, whereas the complement split product Bb is formed only as a result of the activation of the alternative pathway. Thus, the increases in Bb and unchanged levels of C4a indicate that only the alternative pathway is activated in cynomolgus monkeys by treatment with ISIS 2302. 

View larger version (20K): [in this window] [in a new window]   Fig. 1.   A representative pattern of complement split products generated by a 10-min i.v. infusion of 20 mg/kg of ISIS 2302 in a cynomolgus monkey. A: Plasma concentrations of C5a, C3a and Bb were increased; however, no increase in C4a split product was evident relative to plasma concentrations before treatment. Peak concentrations of complement split product were C3a = 199 ng/ml; C5a = 74.9 ng/ml; Bb = 4.03 µml; C4a = 225 ng/ml. B: The extent of complement activation after this dose resulted in a 60% reduction in the total hemolytic complement activity. The temporal relationship between dosing and complement activation is very consistent among animals and is correlated with peak plasma concentrations of oligonucleotide. Although the pattern of split products is similar for all animals, the absolute amount of split products generated will vary depending on individual animal sensitivity to complement activation.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   A representative pattern of hematology changes generated by a 10-minute i.v. infusion of 20 mg/kg of ISIS 2302 in a cynomolgus monkey. Transient neutropenia followed by a rebound neutrophilia reflect complement activation and the production of C5a. Neutrophilia at later time points is associated with increases in circulating nonsegmented neutrophils (insert). Data represented in this figure were obtained from the same animal as depicted in figure 1. Neutropenia is consistently associated with increased levels of C5a.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   A representative pattern of hemodynamic changes after a 10-min i.v. infusion of 20 mg/kg of ISIS 2302 in a ketamine-anesthetized cynomolgus monkey. Blood pressure increased slightly and decreased in parallel with heart rate near the end of the infusion period. Data represented in this figure were obtained from the same animal as depicted in figure 1. The temporal relationship between dosing and hemodynamic effects is again consistent among animals. The presence and severity of hemodynamic effects, however, vary greatly among animals and are roughly correlated with the extent of C5a increase. Many animals experienced no hemodynamic alterations.

Dose-response relationship. To assess the role of peak plasma concentrations in the activation of complement, a series of infusions of varying length were performed. Keeping the dose constant and increasing the duration of infusion from 10 to 60 min resulted in a marked decrease in the amount of Bb produced (fig. 4). The peak concentration of Bb after a 10-min infusion of 10 mg/kg was 13 µg/ml (fig. 4). By comparison, 30- and 60-min infusions of 10 mg/kg yielded lower peak Bb levels of 3.25 and 1.5 µg/ml, respectively. Decreasing the dose in a 60-min infusion from 10 to 6.6 mg/kg also resulted in a further decrease in Bb split product formation, indicating the dose-dependent nature of this response. Plasma concentrations of Bb were proportional to plasma ISIS 2302 concentrations at the end of infusion, with higher plasma concentrations resulting in greater complement activation. Doses of 6.6 mg/kg infused over 60 min were considered to be noncomplement activating because Bb levels did not exceed the normal range of variability (fig. 4).

View larger version (28K): [in this window] [in a new window]   Fig. 4.   The amount of Bb produced after i.v. infusion of ISIS 2302 was dependent on dose and rate of infusion. Peak levels of Bb decreased as the dose was lowered or the duration of infusion was extended. There were a number of dose regimens that did not result in complement activation. The values represent mean values of at least 3 animals ± S.E. (The dashed line represents the upper limit in the normal range of variability.)

View larger version (13K): [in this window] [in a new window]   Fig. 5.   An ISIS 2302 plasma concentration threshold for complement activation. Concentrations of ISIS 2302 are plotted against the corresponding concentration of Bb produced during i.v. infusion. This figure represents data obtained during the course of infusion from over 50 animals, and ranging from 2 to 20 mg/kg infused from 10 to 120 min. (The dashed line represents the normal range of variability in plasma Bb in cynomolgus monkeys.)

Effects of ISIS 2302 on the APC. Factor H is a plasma protein that plays a key role in regulating amplification of the alternative pathway and has previously been identified as a DNA-binding protein (Gardner et al., 1980b). Monkeys treated with 10 mg/kg of ISIS 2302 infused over 10, 30 or 120 min had a dose-dependent and infusion time-dependent decrease in plasma Factor H levels that were closely correlated with increasing plasma concentrations of ISIS 2302 throughout the infusions (fig. 6). In contrast to the pattern of a threshold plasma concentration effect on complement activation and Bb production, a linear concentration response was observed on plasma Factor H concentrations. Small decreases in Factor H levels were also observed after doses that did not result in complement activation. Decreases in plasma Factor H were dependent upon both dose and plasma concentrations of ISIS 2302. Furthermore, the decrease in Factor H was proportional to the levels of Bb produced during complement activation. Plasma Factor H concentrations stabilized and began to return toward base-line levels after the end of infusion (fig. 6).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Intravenous infusion of ISIS 2302 causes a decrease in the circulating plasma concentrations of Factor H. A: Decreases in concentrations of Factor H were dependent on infusion time and were lowest at the end of infusion. B: Peak plasma concentrations of ISIS 2302 occur at the end of infusion and correspond to the maximal decrease in Factor H. The values represent mean values of at least 3 animals ± S.E.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   In a competition-binding assay, ISIS 2302 was able to elute Factor H from a solid support. Increasing concentrations of heparin and ISIS 2302 resulted in a concentration-dependent elution of radiolabeled Factor H from a heparin-sepharose column. The concentration required to elute 50% of Factor H was 20 µg/ml for ISIS 2302 and 4.5 µg/ml for heparin (2.9 and 0.27 µm, respectively).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Intravenous administration of high doses of ISIS 2302, as well as other phosphorothioate oligonucleotides, results in complement activation in monkeys (Galbraith et al., 1994). In this study we further characterized the profile of complement split products formed in ISIS 2302-treated animals and determined that the alternative pathway was selectively activated, as indicated by increases in Bb concentrations and the lack of any increase in the classical pathway marker, C4a. Close examination of the temporal relationship between the production of the anaphylatoxins, C3a and C5a, and other acute alterations suggests that changes in hematologic and hemodynamic parameters observed in monkeys after phosphorothioate oligonucleotide administration may be secondary to complement activation. Some of these alterations, particularly the fluctuations in neutrophil counts described above, are characteristic of exposure to C5a. C5a has been shown to bind to specific receptors of neutrophils triggering a series of intracellular events resulting in increased expression of surface adhesion molecules and a loss of deformability (Frank and Fries, 1991; Kajita and Hugli, 1990). This activation occurs rapidly and leads to sequestration of neutrophils in capillary beds (Kajita and Hugli, 1990). As C5a is cleared from plasma, neutrophils regained their deformability and return to circulation. In addition, the chemotactic influence of C5a causes recruitment of immature neutrophils (nonsegmented neutrophils) from bone marrow, resulting in rebound neutrophilia (Ember et al., 1992). This pattern of changes in circulating neutrophil counts was observed in monkeys treated with ISIS 2302 and has been reported for monkeys treated with other phosphorothioate oligonucleotides (Galbraith et al., 1994). Shortly after oligonucleotide administration, a pronounced drop in circulating neutrophils occurred that was closely associated in time with the formation of C5a. Subsequently, there was a rebound neutrophilia that largely consisted of immature (presumably bone marrow-derived) neutrophils. These changes in neutrophil counts were not observed in the absence of complement activation. Repeated administration of ISIS 2302 either daily or every other day neither enhanced nor diminished the activation of complement and there was no evidence of cumulative effects after a series of doses (data not shown).

In addition to their effects on neutrophils, the anaphylatoxins (C3a and C5a) have been reported to activate other cell types including mast cells and basophils (Hugli and Muller-Eberhard, 1978), which can lead to release of vasoactive autocoids (e.g., histamine, prostaglandins, leukotrienes, etc.). These, or other, vasoactive agents may also be released from endothelial cells subsequent to their interaction with activated neutrophils. Such mediators may cause changes in vascular permeability and tone, and could be responsible for the hypotension that occurs in ISIS 2302-treated monkeys after doses that activate the complement system. It is also conceivable that other acute clinical observations associated with ISIS 2302 administration (i.e., transient lethargy and periocular swelling) were secondary manifestations of complement activation. Activation of the complement system has been reported to occur in a number of diverse but commonly used clinical procedures including extracorporeal blood processing (e.g., hemodialysis), drug treatment (e.g., sulfonamides, corticosteroids, dextrans, heparin and protamine sulfate), high-dose treatment with immunoglobulins (e.g., i.v. immunoglobulin treatment) or treatment with certain recombinant proteins (e.g., interleukin-2 and tissue plasminogen activator) (Svehag, 1991).

Plasma concentrations of ISIS 2302 can be used to predict the extent of complement activation based on the relationship defined in these studies. We have identified a threshold concentration of 50 µg/ml for ISIS 2302, below which no complement activation was detected. At higher plasma levels of ISIS 2302, the relative amounts of split products formed were generally related to peak plasma concentration and, to some extent, to the duration of time that the plasma levels remained elevated. Decreasing peak plasma levels of ISIS 2302 by lowering the dose or prolonging the duration of i.v. infusion while holding the dose constant resulted in lower amounts of complement split products generated. The incidence and severity of hematologic and hemodynamic disturbances in the monkeys appeared to be closely associated with the amount of complement split products generated, especially C5a, which is consistent with dose-dependent effects of C5a on neutrophil counts and physiological responses in vivo and neutrophil morphology in vitro (Ehrengruber et al., 1994). In our studies, hemodynamic effects in monkeys were observed at plasma ISIS 2302 concentrations of 90 µg/ml or higher, whereas the threshold for complement activation was 50 µg/ml. At concentrations of ISIS 2302 in the plasma above 90 µg/ml, Bb levels were more substantially elevated compared to plasma ISIS 2302 concentrations between 50 and 90 µg/ml, and thus would correlate with increased anaphylatoxin levels. Anaphylatoxins are the known biological mediators of complement activation and the markedly elevated levels observed with some of the dose regimens in this study could explain the changes in hemodynamic parameters.

That the alternative pathway was activated in the absence of classical pathway activation has important implications on the potential mechanism of activation. The APC is involved in recognition and destruction of foreign organisms (Tomlinson, 1993). Generally, this pathway is thought to require some type of activating surface to initiate the cascade, such as virus-infected cells, bacteria or parasites (Pangburn and Muller-Eberhard, 1984). The alternative pathway is considered to be constitutively active and amplification of the pathway under normal conditions is suppressed by the combined action of a number of negative regulatory components including Factor H and Factor I. Thus, Factor H and Factor I are important regulatory components of the alternative pathway. Factor H binds to the BbC3b complex and displaces the Bb, thus providing a binding site for Factor I that cleaves C3b to iC3b, therefore preventing further interaction with Factor B, which would lead to the formation of the C3b convertase and ultimately to full amplification of the pathway (Pangburn and Muller-Eberhard, 1984). Under conditions of activation (e.g., exposure to bacteria) the affinity of Factor H and Factor I for C3b bound to the activating surface is dramatically reduced by the binding of properdin, thus eliminating down-regulation of the pathway and allowing amplification of the complement cascade subsequent to formation of the C3bBb convertase. If oligonucleotide treatment alters the activity or concentration of these regulatory proteins, this could be the stimuli for alternative pathway activation (Meri and Pangburn, 1990).

The APC is a relatively simple system composed of only six plasma glycoproteins (C3 and Factors B, D, H, I and P) (Schreiber et al., 1978). Two of the components of the alternative pathway, Factor H and Factor B, have been previously identified as DNA-binding proteins (Gardner et al., 1980a,b). A specific interaction between Factor H and ISIS 2302 was demonstrated herein by the ability of the oligonucleotide to compete for a binding site on Factor H and to elute the radiolabeled protein from a heparin-sepharose column.

Intravenous administration of ISIS 2302 to monkeys resulted in decreases in plasma Factor H levels. Elimination of Factor H from plasma could theoretically lead to disregulation of the APC, because Factor H plays an important role in down-regulating the alternative pathway activity (Pangburn and Muller-Eberhard, 1984). In studies where the alternative pathway has been reconstituted from purified components, it was demonstrated that regulatory function was compromised when Factor H concentrations were reduced below 50% of base-line levels (Schreiber et al., 1978). Therefore, decreases in plasma concentrations of Factor H after high doses of ISIS 2302 likely resulted in the loss of alternative pathway regulation. This relationship between Factor H levels and alternative pathway activation may explain the apparent threshold for complement activation in that regulation of the pathway is only compromised when enough Factor H has been removed from circulation.

Activation of the alternative pathway through loss or removal of Factor H from the plasma has been demonstrated in vitro. Dextran-sulfate linked to insoluble sepharose particles was shown to sequester Factor H from the fluid phase, resulting in alternative pathway activation in vitro (Bitter-Suermann et al., 1981). In addition, immunoprecipitation of Factor H from normal serum using anti-Factor H antibodies resulted in rapid activation of the alternative pathway (Boackle et al., 1983). Although it appears that depletion of Factor H in plasma resulted in complement activation, it is not clear from these experiments how ISIS 2302 treatment in monkeys leads to the reduction of Factor H levels.

Complement activation by a phosphorothioate oligodeoxynucleotide is not unique to the ISIS 2302 sequence or length (20 nucleotides) and has been observed for a number of oligonucleotides (S. P. Henry, unpublished data), including the human immunodeficiency virus-rev antisense compound mentioned previously (25 nucleotides) (Galbraith et al., 1994). Thus, complement activation appears to be largely independent of sequence and more attributable to exposure to the chemical class. Due to the nonspecific nature of this interaction, it is likely that the larger chain-shortened metabolites (possibly down to 16 nucleotide molecules) may also contribute to complement activation and, therefore, should also be taken into consideration when defining exposure.

Results of these studies indicate that complement activation resulting from i.v. infusion of phosphorothioate oligonucleotides is related to peak plasma concentrations of oligonucleotide, and that there is a threshold concentration for activation. Because several of the acute toxicities associated with phosphorothioate treatment, including hemodynamic alterations and neutrophil fluctuations, appear to be secondary to complement activation, this threshold for activation is important for continued safe dosing of these compounds. The cause of alternative pathway activation appears to be due to reduced levels of plasma Factor H, thus allowing unregulated amplification. The inability to establish an in vitro assay for complement activation has prevented us from determining if this effect of phosphorothioate oligonucleotides in monkeys is relevant to humans. However, complement activation is not expected to occur in humans, as the highest doses of ISIS 2302 that are currently being used in the clinic produce peak plasma concentrations of oligonucleotide that are 4- to 5-fold lower than the threshold for activation (J. Glover, submitted). Identification of the molecular mechanism for complement activation will facilitate the design of future generations of antisense oligonucleotides that have decreased potential for these acute toxicities.