Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Recently, a variety of recombinant bioactive proteins and peptides have been produced in large quantities as candidates for new therapeutic drugs. But these proteins, such as cytokines, often have poor in vivo stability because of proteolysis, short half-lives because of rapid renal excretion, and broad tissue distribution (Donohue and Rosenberg, 1983; Bollon et al., 1988; Tanaka and Tokiwa, 1990). Accordingly, excessive high dosage and frequent administration are necessary for successful clinical effects. This causes disruption of homeostasis, resulting in toxic side effects (Kimura et al., 1987; Rosenberg et al., 1987). In addition, because cytokines exhibit diverse pharmacological actions in various tissues, it is difficult to selectively obtain favorable actions.

Tumor necrosis factor- (TNF-) is a cytokine with a wide variety of biological activities in immune and inflammatory responses. TNF- activates natural killer cells (Ostensen et al., 1987), induces expression of adhesion molecules on endothelial cells (Swerlick et al., 1992), and causes direct cytotoxicity in some tumor cells (Helson et al., 1975). TNF- also causes a hemorrhagic necrosis of implanted murine or human tumor cell lines in mice (Carswell et al., 1975). Clinical applications of TNF- have been attempted as an alternative antitumor agent to commonly used antineoplasmic reagents; however, TNF- clinical applicability is still limited. The administration of TNF- results in chills, tachycardia, hypertension, and chakexia due to its instability and wide nontarget tissue distribution (Tracey et al., 1986; Starnes et al., 1988).

Recently, we reported that chemical modification of bioactive proteins and antimetastatic peptides with water-soluble polymeric modifier, typified by polyethylene glycol (PEG), overcomes these side effects (Tsutsumi et al., 1996a; Kaneda et al., 1998; Maeda et al., 1998). Chemical modification of TNF- with PEG, a water-soluble and nontoxic polymer, effectively improved its resistance to proteinases and plasma half-lives by increasing its stereochemical hindrance and molecular size and resulted in greater therapeutical potency (Tsutsumi et al., 1995, 1996b). We have also shown that certain PEG modification (PEGylation) of TNF- selectively enhances its desirable effects (i.e., therapeutic effects, antitumor effect) and reduces its undesirable effects in vivo. However, we found that the specific activities of PEGylated cytokines decreased with increases in the degree of PEG modification. A decrease in the specific activities of cytokines by polymer conjugation is believed to be caused by modification of the active core or receptor-binding region of the cytokines and by stereochemical hindrance by the modified polymers. To improve the therapeutical potency of PEGylated TNF- (PEG-TNF-) and other polymer-conjugated cytokines, a novel method of polymer conjugation is required.

Because the reaction used for polymer-conjugated proteins (modification of lysine amino groups) is mild for unstable proteins and highly reactive, many polymer-conjugated proteins have randomly modified lysine residues (Delgado et al., 1992). Unfortunately, lysine residues involved in bioactivities can be conjugated because the receptor-binding regions of cytokines are believed to exist on the surface of cytokine. Therefore, conjugation methods that avoid modification to amino groups around the receptor-binding region of cytokines should result in more effective polymer-conjugated cytokine. We attempted to control modification by using the reversible amino-protective reagent dimethylmaleic anhydride (DMMAn). In this study, we assessed the usefulness of a novel polymer conjugation method using DMMAn to create more effective and safer polymer-conjugated cytokines for clinical application.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Methoxypolyethylene glycol succinimidyl succinate (ssPEG; number average molecular weight = 5000) was purchased from Sigma Chemical Co. (St. Louis, MO). DMMAn was purchased from Acros Organics (Springfield, NJ), and fluorescamine (Fluram) was from Fluka (Tokyo, Japan). Human natural TNF- was generously provided by Mochida Pharmaceutical Co., Ltd. (Tokyo, Japan).

Preparation of PEG-TNF- Using DMMAn. TNF- was chemically conjugated with PEG by amide-bond formation between lysine amino groups in TNF- and ssPEG. To reversibly protect some lysine amino groups in TNF-, TNF- in phosphate buffer, pH 8.5, was reacted with a 10 M excess of DMMAn to the lysine amino acids on ice. After a 30-min reaction, a 10 M excess of ssPEG to the lysine amino acids was reacted for 1 h. Then, to regenerate the protected lysine amino groups, the reaction mixture was adjusted to pH 6.0 with 0.1 N HCl and incubated at 37°C for 30 min. PEG-TNF-(+), which was pretreated with DMMAn before PEGylation, was separated by gel filtration HPLC (Superose 12; Amersham Pharmacia Biotech, Uppsala, Sweden) into three fractions of different number average molecular weight values and stored at 80°C until use. Control PEG-TNF- (not treated with DMMAn) was prepared by the usual methods [PEG-TNF-()] (Tsutsumi et al., 1996b). Protein content was measured by HPLC at 280 nm with native TNF- as the standard. The degree of modification was monitored by a fluorimetric method using fluorescamine, which reacts with the remaining amino groups of proteins (Stocks et al., 1986).

Bioassay of PEG-TNF-. The specific activities of PEG-TNF-s were estimated by cytotoxicity assays using LM cells and expressed in terms of the Japan reference unit (JRU) according to the method described by Yamazaki et al. (1986).

Resistancy from Proteases. Native-TNF- and PEG-TNF-s (10 µg/ml) were incubated with trypsin (1:250; Difco Laboratories, Detroit, MI) or Cathepsin G (4000 U/mg; Wako Pure Chemical Co., Ltd., Osaka, Japan) in minimal essential medium at 37°C. The mixture was diluted with minimal essential medium containing 1% FBS to stop proteolysis and used for LM cell cytotoxicity assay.

Antitumor Effects In Vivo. All animal experimental protocols were in accordance with the "Principles of Laboratory Animal Care" (NIH publication 85-23, revised 1985). The antitumor effect of native TNF- and PEG-TNF- was evaluated in mice bearing Meth-A fibrosarcoma. Meth-A cells were implanted intradermally (5 × 10⁵ cells/site) in female BALB/c mice (4 weeks old; Charles River Japan Inc., Yokohama, Japan). On day 7, when the tumor diameter reached 8 mm, cytokine was administered i.v. as a single injection. The antitumor potency of PEG-TNF- was estimated by the area of tumor hemorrhagic necrosis within 24 h after injection. Body weight of mice was measured 24 h after treatment.

Statistical Analysis. Specific activity, tumor necrotic area, and body weight were statistically evaluated with the use of Student's t test.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Characterization of PEG-TNF-(+) Prepared Using DMMAn In Vitro. PEG-TNF-(+) was prepared according to the reactions presented in Scheme 1. To confirm sequential changes in PEG-modified TNF-(+), the remaining free lysine amino groups were measured by fluorimetric analysis (Fig. 1). About 35% of the 18 lysine amino groups in TNF- were protected from PEGylation by DMMAn (Reaction I). PEGylation of residual amino groups was then performed (Reaction II), followed by regeneration to free amino groups in Reaction III. As a control, PEGylated TNF- with nontreated of DMMAn was also synthesized [PEG-TNF-()]. The products, PEG-TNF-(+) and PEG-TNF-(), were purified and separated into three fractions of different molecular sizes (high = H, middle = M, low = L) by gel filtration HPLC, and then specific activities were measured. The specific activities of PEG-TNF-()s and PEG-TNF-(+)s had gradual reduction with increasing degrees of modification (Table 1). The suppressed loss in specific activities or 20% to 40% of improvement in specific activities was observed in each fractions of PEG-TNF-(+)s compared with PEG-TNF-()s. Changes in pH and DMMAn treatment used in this study did not result in the loss of TNF- activity (data not shown).

View larger version (24K): [in this window] [in a new window]   Scheme 1.   Schematic protocol of PEGylation of TFN- using DMMAn. Reaction I, protection of partial amino groups by DMMAn; reaction II, PEGylation to remained lysine amino groups; reaction III, regeneration of amino groups by releasing DMMAn.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Monitoring of the PEGylation of TNF- using DMMAn by the fluorescamine method. Remaining amino groups of native TNF- (), PEG-TNF-() (), and PEG-TNF-(+) () are displayed. Reaction numbers presented below the figure correspond to the numbers in the figure.

View this table: [in this window] [in a new window]   TABLE 1 Improvements in specific activity of PEG-TNF-s using DMMAn

Protective Effects from Various Proteases. Protease resistance of PEG-TNF-s was examined (Fig. 2). The bioactivity of native TNF- rapidly diminished with incubation with trypsin or cathepsin G, which are gastric and lysozymic enzymes, respectively. Both PEG-TNF-() and PEG-TNF-(+) were more resistant to proteolysis than native TNF-, and fractions with higher molecular sizes were more resistant. DMMAn treatment resulted in a small reduction in stability.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Resistance of native TNF- and PEG-TNF-s to proteases. Native TNF- or PEG-TNF-s was incubated with trypsin (left) or cathepsin G (right) at 37°C for indicated times. Reactions were stopped, and remaining bioactivities were measured by bioassays using LM cells. Each value is the mean ± S.D.

Estimation of Antitumor Effects of PEG-TNF-. The in vivo antitumor effects of PEG-TNF-(+)s were compared with those of native TNF- and PEG-TNF-()s. Intravenous injection of TNF- causes hemorrhagic necrosis of Meth-A solid tumors within 24 h (Carswell et al., 1975). Figure 3 shows the necrotic area on intradermally implanted Meth-A fibrosarcoma 24 h after i.v. injection of native or PEG-TNF-s. Hemorrhagic necrosis was caused in mice treated with native TNF- in a dose-dependent manner; an approximately 35% tumor necrotic area was observed at a dose of 30.0 µg/mouse. However, severe body weight loss followed within 24 h. Little or no necrosis was observed at doses of less than 3.0 µg/mouse. Among usual PEG-TNF-()s, the middle fraction of PEG (MPEG)-TNF-() showed the most potent tumor necrotic effect. MPEG-TNF-() at a dose of 1.0 µg/mouse had similar effects to native TNF- at a dose of 16.0 µg/mouse (16-fold increase) The antitumor effects of MPEG-TNF-(+) was 2-fold higher than that of MPEG-TNF-(). Marked improvement in antitumor potency was achieved in all PEG-TNF-s synthesized using the DMMAn method. In addition, no obvious body weight loss was observed in any PEG-TNF--treated mice (Fig. 4).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Tumor necrotic effects of PEG-TNF-s on Meth-A solid tumor in mice. Meth-A-bearing BALB/c mice were given native TNF-, PEG-TNF-(), or PEG-TNF-(+) i.v. The control group was given saline. The area of hemorrhagic necrosis was measured 24 h after injection. Each value is the mean ± S.E. of four animals. *P < .01, statistical significance compared with PEG-TNF-()s. N.D., not detected.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Body weight changes in Meth-A-bearing mice treated with native TNF- or PEG-TNF-s. Meth-A-bearing BALB/c mice were given native TNF-, PEG-TNF-(), or PEG-TNF-(+) i.v. The control group was given saline. Body weight was measured 24 h after i.v. injection. Data were represented as body weight relative to that before injection. Each value is the mean ± S.E. of four animals. *P < .05, **P < .01, statistical significance compared with PEG-TNF-()s.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Frequently used methods of chemical modification with PEG (PEGylation) resulted in random covalent conjugations to lysine residues (Delgado et al., 1992). The attachment of polymers to amino groups in the active core or receptor-binding region of cytokines causes decreases in specific activities; therefore, conjugation methods that avoid modification to amino groups around the receptor-binding region of cytokines should result in more effective polymer-cytokine conjugates for clinical application. The purpose of the present study was to develop a novel method to create polymer-conjugated cytokines using TNF- as a model, with higher efficacies, and to investigate the usefulness of the resultant conjugates in vivo. The protection of partial amino groups in TNF- by DMMAn was confirmed by fluorimetric analysis (Fig. 1). Then, to examine whether the active core of TNF- can be protected from PEGylation using DMMAn, changes in specific activity were examined in vitro. Specific activities improved for all fractions of PEG-TNF-(+) compared with the similar molecular size fractions of PEG-TNF-() tested (Table 1). This result indicates that the use of DMMAn improves cytokine receptor binding. We found similar results for PEGylation of interleukin-6 and granulocyte macrophage colony-stimulating factor (data not shown). It has been demonstrated that the Lys90 of TNF- is involved in receptor binding and is located on the surface of the molecule (van Ostade et al., 1991). These suggest that partial modification of lysine amino groups of TNF- leads to their protection from PEGylation and subsequent inactivation. Although more precise examinations are necessary, this methodology may be applicable for polymer modification of various cytokines with lysine residues in their active cores. Figure 2 shows that PEG-TNF-(+)s have slightly reduced protease resistance compared with PEG-TNF-()s. The degree of modification was similar for each fraction of PEG-TNF-(+) and PEG-TNF-(), suggesting that improvements in specific activity correlate with reductions in stereochemical hindrance at the active core and result in some loss of resistance to proteases.

To examine the influence of improvements in specific activity on antitumor activity in vivo, Meth-A fibrosarcoma-bearing mice were given native and PEG-TNF-s by i.v. injection. TNF- causes tumor hemorrhagic necrosis by specific injury to tumor endothelium. Native TNF- caused hemorrhagic necrosis within 24 h after i.v. injection in Meth-A-bearing mice, and the necrotic area reached about 35% at a dose of 30.0 µg/mouse. However, native TNF- treatment also resulted in dose-dependent body weight loss, limiting the clinical application of TNF-. MPEG-TNF-() was the most potent TNF- not treated with DMMAn, as previously reported (Tsutsumi et al., 1995) (Fig. 3). The administration of 1.0 µg/mouse MPEG-TNF-() had similar effects as 16.0 µg/mouse native TNF-, whereas the antitumor effects of MPEG-TNF-(+) were 2-fold more potent than those of MPEG-TNF-() and 30-fold more potent than those of native TNF-. Significantly, improvements in the specific activity of MPEG-TNF-s were only about 50%, but improvements in antitumor effects were more than 2-fold in vivo. Slight decreases in protease residency of PEG-TNF-(+) compared with PEG-TNF-() did not lead to a severe loss of efficacy in vivo. No obvious side effects were observed in any groups of PEG-TNF--treated mice. Using DMMAn in PEGylation, antitumor effects were improved without increasing unfavorable effects. We previously reported that optimal degree of PEGylation of TNF- (MPEG-TNF-) increases its antitumor effect without any side effects (Tsutsumi et al., 1995). This suppression of side effects was the result of decreases in dose of TNF- and suppression of distribution to side effect-related tissues such as liver. Indeed, PEG-TNF-(+)s had similar molecular sizes as PEG-TNF-()s; therefore, suppression of systemic distribution of PEG-TNF-(+) is presumed. Although further analysis of various side effects is necessary, this may be one of the reasons for the selective improvement of antitumor activity of TNF- by PEGylation using DMMAn. Now, we are engaged in the study of the pharmacokinetics of PEG-TNF-(+) and in the more precise analysis of the mechanism of selective enhancement of antitumor activity by PEG-TNF-s. Even closer control of modification sites may create polymer-conjugated cytokines with higher activity and safety. However, our method is easy and useful for the clinical application of polymer-conjugated cytokines.