Abstract
We attempted the development of a novel polymer conjugation to further improve the therapeutic potency of antitumor cytokines compared with PEGylation for clinical application. Compared with native tumor necrosis factor (TNF)-α in vitro, specific bioactivities of polyvinyl-pyrrolidone (PVP)-modified TNF-αs(PVP-TNF-αs) were decreased by increasing the degree of PVP attachment. PVP-TNF-α fraction 3, Mr101,000, had the most effective antitumor activity of the various PVP-TNF-αs in vivo.
PVP-TNF-α fraction 3 had >200-fold higher antitumor effect than native TNF-α, and the antitumor activity of PVP-TNF-α fraction 3 was >2-fold higher than that of MPEG-TNF-α(Mr 108,000), which had the highest antitumor activity among the polyethylene glycol (PEG)-conjugated TNF-αs. Additionally, a high dose of native TNF-α induced toxic side effects such as body weight reduction, piloerection, and tissue inflammation, whereas no side effects were observed after i.v. administration of PVP-TNF-α fraction 3. The plasma half-life of PVP-TNF-α fraction 3 (360 min) was about 80- and 3-fold longer than those of native TNF-α (4.6 min) and MPEG-TNF-α (122 min),respectively. The mechanism of increased antitumor effect in vivo caused the prolongation of plasma half-life and increase in stability. These results suggested that PVP is a useful polymeric modifier for bioconjugation of TNF-α to increase its antitumor potency, and multifunctionally bioconjugated TNF-α may be a potentiated antitumor agent for clinical use.
INTRODUCTION
TNF3-α was originally identified as an endotoxin-induced factor able to induce hemorrhagic necrosis of tumors in mice (1). The antitumor effect of TNF-α is known to result not only from its direct cytotoxicity against tumor cells but also from activation of antitumor effector immune cells in the blood, such as macrophages, cytotoxic lymphocytes, and neutrophils (2, 3), and furthermore from specific damage to tumor blood vessels (4, 5, 6). Clinical applications of TNF-α have been attempted as novel antineoplastic agents to tumors instead of traditional antitumor drugs (7, 8). However, because of its very low stability and pleiotropic action in vivo, attempts to use TNF-α as a systemic anticancer agent in humans failed because of the appearance of severe systemic side effects, such as a fever and decreased blood pressure the same as an endotoxin-like shock, before therapeutic doses could be reached (9, 10). Accordingly, clinical applications of TNF-α for cancer therapy are still limited despite high expectations(11, 12).
In general, it is difficult to use bioactive proteins, such as cytokines, for therapeutic use because of poor stability and short half-lives. These proteins are rapidly cleared from blood by the liver,kidney, and others organs, and the rate of clearance depends on the size of molecule and degree of proteolysis. Plasma proteases cause degradation and rapid loss of biological activity; therefore, achieving a clinical effect is still a problem. To overcome these problems,soluble-polymer technology (i.e., bioconjugation) attempts to bind polymers for protection against degradation by the host(13). Bioconjugation, forming a stable linkage between the protein and polymer, can prolong the half-life and preserve biological activity. Therefore, therapeutically useful proteins would be modified with various polymers to improve their pharmacokinetic properties and may acquire many advantages after polymer conjugation(14).
Generally, the more polymer chains attached per protein molecule,the greater the extension of half-life, but specific activity may be proportionally reduced (15). The increased plasma half-lives of proteins are attributable to several mechanisms,increased size of protein conjugates surpasses the limit for glomerular filtration (Mr 70,000) and decreased proteolysis because of polymer steric hindrance. Previously, we conjugated TNF-α with PEG to increase stability in vivoand selectively obtain antitumor activity for clinical use(16, 17, 18). Optimal PEGylated TNF-α effectively increased the protease resistance and circulation time by sterically blocking proteolysis and decreasing clearance via glomerular filtration(19). These effects were attributed to increases in steric hindrance and in molecular weight of bioconjugated proteins, both of which were attributable to attachment of polymers. As a result,optimally bioconjugated cytokines showed much better therapeutic effects than native (unconjugated) cytokines. However, to design conjugated cytokines with greater efficacy and safety, i.e.,to selectively enhance desirable therapeutic activities of cytokines without increasing their side effects, it is necessary to closely regulate their behavioral characteristics in vivo, taking into account their mechanism of action.
Thus, it is necessary to develop polymeric modifiers with useful DDS functions, which can regulate their behavioral characteristics in vivo, such as targeting capability and controlled release. The molecular structure of PEG as a general polymeric modifier does not readily allow the addition of such new functions, and other polymeric modifiers to which these new functions can be added are required. Recently, we evaluated the behavior of various water-soluble polymers, and PVP, to which useful functional groups can be introduced by radical copolymerization, was found to be retained in the blood better than PEG.4PVP, which was widely used as a plasma expander until clinical studies revealed a lack of efficacy, is a water-soluble polymer available in molecular weights ranging from a few thousand to several hundred thousand. Therefore, PVP may be a better polymeric modifier to increase the half life of cytokines by conjugation.
In this study, we conjugated TNF-α with PVP (PVP-TNF-α) to increase its half-life and selectively increase its therapeutic potency as a systemic antitumor agent. The results of this study will provide fundamental information enabling us to design useful bioactive protein derivatives showing selective localization in the vascular space against tumors and various cardiovascular diseases and to develop better polymeric modifiers with DDS functions to enhance the therapeutic activity and safety of conjugated bioactive proteins.
MATERIALS AND METHODS
Materials.
Natural human TNF-α (1.78 × 106JRU/mg TNF-α in 0.2 m phosphate buffer, pH 7.2) was kindly supplied by Hayashibara Biological Laboratories (Okayama,Japan). N-Vinyl-2-pyrrolidone was purchased from Wako(Osaka, Japan), 4–4′-azobis-4-cyanovaleric acid was from Aldrich(Milwaukee, WI), and 3-mercaptopropionic acid was from Dojindo Laboratories (Kumamoto, Japan). Other reagents and solvents were of analytical grade.
Synthesis and Activation of PVP.
The terminal COOH-bearing PVP was synthesized from N-vinyl-2-pyrrolidone (27 mmol) by radical polymerization in dimethyl formamide (7 ml) with the aid of 4,4′-azobis-(4-cyanovaleric acid) (1.2 mmol) as a radical initiator, and 3-mercaptopropionic acid(2.7 mmol) as a chain transfer agent at 60°C for 6 h. Resultant PVPs with an average molecular weight of Mr 6,000 [polydispersity(Mr/average Mr), 1.14] were separated and purified by high-performance liquid chromatography (GF-HPLC; TSK Gel G4000PwxL; Toso Co., Ltd., Tokyo, Japan). The terminal COOH group of synthetic PVP was activated by the N-hydroxysuccinimide/dicyclohexyl carbodiimide method.
Conjugation of TNF-α with Activated PVP.
TNF-α was reacted with a 60-fold molar excess of activated PVP at room temperature for 10 min, and then ε-amino caploic acid (5-fold molar excess against activated PVP) was added to stop the reaction. The resultant PVP-TNF-α was separated into five fractions of various molecular weights by gel filtration chromatography (TSK Gel G3000SwxL;Tosoh Co., Ltd.). The average Mr of PVP-TNF-αs were estimated by gel filtration chromatography analysis,and degree of PVP-attachment to TNF-α was calculated from the average Mr (protein standard; Gel Filtration LMW Calibration kit; Amersham Pharmacia Biotech, Buckinghamshire,United Kingdom). The specific bioactivities in vitro of PVP-TNF-αs were measured by L-M cell cytotoxicity assay, according to the method described by Yamazaki et al. (20)and were expressed in terms of JRU defined previously(20).
Screening and Evaluation of Antitumor Effects in Vivo.
S-180 cells (5 × 105 cells) were implanted intradermally in the abdomen of male ddY mice (5 weeks of age; SLC, Hamamatsu, Japan). After 7 days (tumor size, 8–9 mm in diameter), native TNF-α and PVP-TNF-αs were administered in a single i.v. injection. Antitumor effects of PVP-TNF-αs in vivo were screened by determining mean tumor volume calculated from the formula described by Haranaka et al.(21). The in vivo antitumor potencies of native TNF-α, MPEG-TNF-α, in which 56% of the TNF-α lysine amino groups were conjugated with PEG, and PVP-TNF-α fraction 3 in which 40% of TNF-α lysine amino groups were conjugated with PVP, were evaluated as follows. Meth-A fibrosarcoma cells (4 × 105 cells) were inoculated intradermally into the abdomen of 5-week-old female BALB/c mice (SLC). Seven days later (tumor size, 7–8 mm in diameter), native TNF-α, MPEG-TNF-α, and PVP-TNF-α fraction 3 were i.v. administered by a single injection. Drug efficacy against Meth-A tumors was expressed by a score of tumor hemorrhagic necrosis according to the method of Carswell et al. (1). Briefly, the maximal necrotic response(score 3) indicates that ≥50% of the tumor mass is necrotic, the moderate response (score 2) indicates 25–50% necrotic, the minimal response (score 1) indicates <25% necrotic, and no response (score 0)indicates no visible necrosis.
Assessment of Plasma Clearance.
Native TNF-α, MPEG-TNF-α, and PVP-TNF-αs were radiolabeled with 125I by the lactoperoxidase method, yielding 125I-native TNF-α, 125I-MPEG-TNF-α, and 125I-PVP-TNF-αs with specific activities of 4.44 μCi/mg protein. The biological activities of 125I-radiolabeled TNF-α and its derivatives were indistinguishable from those of nonradiolabeled TNF-α and its derivatives (data not shown). Their pharmacokinetic profiles in blood circulation after i.v. injection into female BALB/c mice (5 weeks of age) were studied at a dose of 31.6 ng of protein/mouse. Blood was collected from the tail vein at various time points, and radioactivity was measured in each sample.
Statistical Analysis.
The hemorrhagic necrosis scores and tumor volume were statistically evaluated by the Student t test.
RESULTS
Preparation of PVP-TNF-αs.
Natural human TNF-α was conjugated with activated PVP (average Mr 6000; Mr/average Mr, 1.14) via amide bonds between amino groups of TNF-α and N-hydroxyscuccinimide groups of PVP at the end of the main chain. The rate of PVP attachment to TNF-αincreased with increasing reaction time and molar mass of PVP to TNF-α (data not shown). The resulting TNF-α-conjugated PVPs were purified from native TNF-α and separated into five fractions of various molecular sizes by gel filtration-high performance liquid chromatography (protein standard). Table 1 shows the average Mr, degree of PVP modification of these separated PVP-TNF-αs, and their activities compared with that of native TNF-α. The activities of PVP-TNF-αs decreased with increasing of molecular weight and degree of PVP modification (PVP attachment to TNF-α). This result was also observed when TNF-α was conjugated with PEG (Fig. 1), and this profile of changes in the bioactivity of PVP-TNF-αs was similar to that observed with modification of TNF-α with PEG (average Mr, 5000; Mr/average Mr, 1.32) reported previously(17).
Differing Antitumor Activities of Various PVP-TNF-αs in Vivo.
The antitumor effects of PVP-TNF-αs on S-180 solid tumors were compared with those of native TNF-α and MPEG-TNF-α by single i.v. injection (Fig. 2). The antitumor effects were evaluated by a score of hemorrhagic necrosis 24 h after sample administration. Native TNF-αinhibited tumor growth dose dependently, but marked side effects(e.g., sudden death, transient decrease in body weight, and others) were observed in all mice administered native TNF-α at a dose of 10,000 JRU. Therefore, 10,000 JRU of TNF-α was found to be the maximal applicable dose. All PVP-TNF-αs were injected i.v. at a dose of 1,000 JRU. PVP-TNF-α fraction 1 (average Mr, 134,000) and fraction 5 (average Mr, 74,000) slightly observed tumor necrosis. PVP-TNF-α fraction 2 (average Mr, 117,000) and fraction 4 (average Mr, 84,000) had showed antitumor effects comparable with that of native TNF-α at a dose of 10,000 JRU. The tumor growth- inhibitory effect of PVP-TNF-α fraction 3 (average Mr, 101,000), in which 40% of the total lysine amino groups of TNF-α were coupled with PVP, was markedly higher than that of native TNF-α at a dose of 10,000 JRU/mouse and induced complete regression in two of seven mice (data not shown). No side effect was observed, such as decrease in body weight, in any mice administered PVP-TNF-αs.
Antitumor Effect of PVP-TNF-α Fraction 3 in Vivo.
PVP-TNF-α fraction 3 showed the best characteristics among the PVP-TNF-αs examined. To clarify the usefulness of PVP as a polymeric modifier and PVP-TNF-α fraction 3 as a systemic antitumor agent, we compared the antitumor potency of PVP-TNF-α fraction 3 to those of native TNF-α and MPEG-TNF-α with scheduled i.v. injections on Meth-A solid tumors. Control mice (saline or PVP alone) showed no antitumor effect and hemorrhagic necrosis, whereas native TNF-αshowed suppression of tumor growth in a dose-dependent manner (Fig. 3). However, three of eight mice administered native TNF-α at a dose of 10,000 JRU died within 24 h, and the remaining mice developed piloerection and tissue inflammation (e.g., erythema) and showed a decrease in body weight (data not shown). This dose of native TNF-α completely inhibited tumor growth up to day 28 after tumor inoculation, but tumor growth was observed in one of eight mice after that. PVP-TNF-α fraction 3 and MPEG-TNF-α had markedly increased their antitumor potencies compared with native TNF-α. As shown in Table 2, PVP-TNF-α fraction 3 and MPEG-TNF-α at a dose of 200 JRU showed the maximal antitumor effects without any toxic side effects (such as sudden death and others) and had antitumor effects superior to that of native TNF-α at a dose of 10,000 JRU. On the other hand, only 50 JRU of PVP-TNF-α fraction 3 was needed to exhibit a marked antitumor potency, and tumor growth was completely inhibited for the observation period, as in 10,000 JRU native TNF-α and 100 JRU MPEG-TNF-α. These results indicated that PVP-TNF-α fraction 3 was approximately 200-and 2-fold more potent an antitumor agent than native TNF-α and MPEG-TNF-α, respectively.
Pharmacokinetics of PVP-TNF-αs.
The pharmacokinetics of native TNF-α, MPEG-TNF-α, and PVP-TNF-αs were examined (Fig. 4). Native TNF-α rapidly disappeared from the circulation, and its plasma half-life was only 3.0 min. This half-life corresponds with that reported previously. This rapid clearance of native TNF-α was found to be attributable to proteolysis, renal excretion, and broad distribution to various tissues (data not shown). In contrast, the plasma clearance of PVP-TNF-α fraction 3 was markedly decreased relative to those of native and MPEG-TNF-α, and its plasma half-life(365 min) was about 80- and 3-fold longer than those of native TNF-αand MPEG-TNF-α (123 min), respectively.
DISCUSSION
The distribution of TNF-α from blood into the adversely affected tissue, such as the liver, causes toxic side effects (8, 22). The improved retention of TNF-α in the vascular space and the resultant decrease in transfer of TNF-α to these tissues is expected to reduce the side effects of TNF-α therapy. Thus, the improvement in circulation time may selectively enhance its antitumor action without increasing its side effects, resulting in augmentation of its bioavailability. To assess the usefulness of PVP as a polymeric modifier for protein conjugation, we evaluated the therapeutically potency of PVP-TNF-αs compared with that of PEG conjugate (PEG-TNF-α) in vivo. The results, shown in Table 1, indicated that the bioactivity in vitro was decreased with increasing the degree of modification. These findings suggested two hypotheses: (a) PVP chain TNF-α sterically inhibited TNF-receptor binding of PVP- TNF-αs; and (b) lysine amino residues of TNF-α play an important role in its bioactivity. We assumed that decreased activity in vitro was caused by steric hindrance, inhibiting the binding with TNF receptor. In fact,lysine 11 of TNF-α is known to fulfill a structural role(23). PVP was reacted with ε-amino groups; therefore, if this lysine 11 has been conjugated with the polymer, the bioactivity of TNF-α was almost diminished or vanished. Although this lysine has not been conjugated, if the polymers conjugate in the neighborhood of lysine 11, it is assumed the activity of bioconjugated TNF-α is reduced because of the steric hindrance.
In Fig. 2, the antitumor activities in vivo were evaluated by hemorrhagic necrosis between fractionated PVP-TNF-αs. The results showed that PVP-TNF-α fraction 3, Mr 101,000, had the highest antitumor activity in vivo. MPEG-TNF-α, which showed the highest antitumor activity of the PEG-bioconjugated TNF-α, has a Mr 108,000. This finding suggests that the optimal molecular weight of bioconjugated TNF-α is Mr ∼100,000 to increase the bioactivity in vivo. However, the optimal molecular weight was limited in this case to that of conjugated TNF-α with PEG or PVP. Our previous study showed that DIVEMA-TNF-α (24, 25),which has a Mr 63,000, showed the most marked antitumor activity. DIVEMA, which is known as a biological response modifier (26, 27), has many reactive anhydride residues that form amide bonds with TNF-α. PEG or PVP are reacted only at the end point of the main chain. It is assumed that the optimal bioconjugated conditions would exist because of the selected polymeric modifiers.
In Fig. 3, to investigate the antitumor effect of PVP-TNF-α fraction 3, which showed the highest effects in more detail, the comparison between native- and MPEG-TNF-αs was determined by scheduled administration. The results showed that the antitumor effect was almost same for doses of 10,000 JRU native TNF-α and dosage of 50 JRU of PVP-TNF-α fraction 3; thus, the antitumor effect increased 200-fold over the native TNF-α. Complete regression was observed in only half the mice by administration of native TNF-α at 10,000 JRU, whereas administration of PVP-TNF-α fraction 3 only 200 JRU inhibited tumor growth completely. In the case of native TNF-α administration, the maximal dose was most effective but caused side effects(e.g., sudden death and others.), whereas the minimal dose was not effective for tumor growth inhibition. However,this problem did not arise with PVP-TNF-αs; therefore, we are confident that PVP-TNF-αs, especially PVP-TNF-α fraction 3, will be beneficial for cancer therapy. The reduction in administered dose was possible by increasing the stability of TNF-α in vivo and eliminating the high initial dose that caused side effects. This increase in circulation time of PVP-TNF-α fraction 3 was found to be caused by an increase in stability attributable to the shielding of proteolytic cleavage sites of TNF-α by the PVP chain and reduction of the renal clearance attributable to increased molecular size through attached PVP. PVP has no toxicity when given i.v. administration at this dose and has not shown antitumor effects on tumor-bearing mice. Additionally, after i.v. administration of PVP, experience confirms that polymers with a Mr <20,000 are completely eliminated through the kidneys (28). Therefore, it is assumed that the increasing antitumor effect of PVP-TNF-α fraction 3 is not caused by direct action of PVP. PVP-TNF-α fraction 3 was mainly localized in the vascular space. In general, it has been well known that macromolecules are accumulated and retained in the tumor tissue effectively. This phenomenon is termed the“enhanced permeability and retention” effect (29), and many macromolecular anticancer agents, such as synthetic polymer-conjugating drugs and polymeric micelle-containing drugs and others, have been reported (30, 31). We have reported previously that the tumor distribution of MPEG-TNF-α was markedly enhanced compared with native TNF-α and gradually increased over time(19). About 9-fold more MPEG-TNF-α was distributed to the tumor than native TNF-α. Thus, we come to the conclusion that the marked increase in the antitumor potency of MPEG-TNF-α resulted from the enhanced tumor accumulation. PVP-TNF-α fraction 3(Mr 101,000) has the same molecular size as MPEG-TNF-α (Mr 108,000). Thus, it is assumed that the concentration PVP-TNF-α fraction 3 would be higher than that of native TNF-α in tumor site.
As described in the introduction, antitumor effects of TNF-α were attributable not only to direct cytotoxicity against tumor cells but also to specific injury of the tumor vascular and effective activation of antitumor immune cells. TNF-α selectively enhanced the vascular permeability of tumor vessels. The enhancement of TNF-α half-time may lead to a decrease in its distribution to the liver and spleen, which are the major sites of side effects, and would selectively increase its antitumor effects. Therefore, in Fig. 4 we are confident that the increased antitumor potency of PVP-TNF-α fraction 3 relative to native TNF-α and MPEG-TNF-α may be attributed to increased half-life. It is important to determine why the circulation time of PVP-TNF-α fraction 3 was much longer than that of MPEG-TNF-α,although their molecular sizes were almost the same. We found previously that the mean residence time in blood of PVP6,000 is about seven times longer than that of PEG5,000. Thus, we assumed that the behavior of these conjugated TNF-αs in vivo was affected by the behavioral characteristics of attached polymeric modifiers.
In this study, we showed that the conjugation of TNF-α to PVP selectively increases its antitumor effects, and PVP-TNF-α fraction 3 may be useful as a more potent antitumor therapeutic agent than PEGylated TNF-α. In addition, we demonstrated that PVP is a more useful polymeric modifier than PEG. These findings indicated that the conjugation of TNF-α with PVP effectively increased its antitumor potency without adverse effects. We are currently synthesizing various PVP derivatives by radical copolymerization. PEG has been used as a polymeric modifier, but it is difficult to introduce functions, such as targeting capability and controlled release, into PEG. Therefore, we examined other candidates as polymeric modifiers for introduction of these new functions. PVP can be easily introduced various reactive residues by radical copolymerization. Furthermore, PVP, used as a suspending agent, binder disintegrant, and tablet lubricant for preparation of various medicines (32), is a highly biocompatible, amphiphatic, nontoxic, and nonimmunogenic polymer similar to PEG. In addition, we found that the plasma half-life of PVP itself was longer than that of PEG after i.v. injection.
In the process of hemorrhagic necrosis in tumor vessels (33, 34), the vascular permeability is selectively increased(35, 36), promoting transport from blood to tumor tissue(37). We are attempting to introduce functions into PVP such as immunopotential action of polydivinylether-maleic anhydride(DIVEMA) and targeting capability to specific tissues for obtaining useful TNF-α derivatives. We developed a novel polymeric modifier that can be used to introduce some useful DDS functions for controlling the behavior of conjugated bioactive proteins in vivo to further increase their therapeutic activities and safety by isolation of desirable activities for clinical use.
Antitumor effects of PVP-TNF-αs on S-180 tumor-bearing mice. Seven days after tumor inoculation, native TNF-α, MPEG-TNF-α,or PVP-TNF-α was given in a single i.v. injection. Mice were used in groups of seven. Each value is a mean; bars, SE. Significance compared with PVP control. ∗, P < 0.05;∗∗∗ P < 0.01; ∗, not detected.
Antitumor effects of PVP-TNF-αs on S-180 tumor-bearing mice. Seven days after tumor inoculation, native TNF-α, MPEG-TNF-α,or PVP-TNF-α was given in a single i.v. injection. Mice were used in groups of seven. Each value is a mean; bars, SE. Significance compared with PVP control. ∗, P < 0.05;∗∗∗ P < 0.01; ∗, not detected.
Antitumor effect of native TNF-α (A) and PVP-modified TNF-α (B) by i.v. injection twice a week on necrotic score after Meth-A fibrosarcoma inoculation. Mice were used in groups of five. Each value is a mean; bars, SE. Significance compared with control. ∗, P < 0.01. N. D., not detected.
Antitumor effect of native TNF-α (A) and PVP-modified TNF-α (B) by i.v. injection twice a week on necrotic score after Meth-A fibrosarcoma inoculation. Mice were used in groups of five. Each value is a mean; bars, SE. Significance compared with control. ∗, P < 0.01. N. D., not detected.
Changes in TNF-α concentration in serum after administration of native TNF-α and bioconjugated TNF-α. After i.v. administration of [125I]-native TNF-α and[125I]-bioconjugated TNF-α to BALB/c mice, blood was collected from tail vein at various times, and radioactivity was measured. Mice were used in groups of four. Each value is a mean; bars, SD.
Changes in TNF-α concentration in serum after administration of native TNF-α and bioconjugated TNF-α. After i.v. administration of [125I]-native TNF-α and[125I]-bioconjugated TNF-α to BALB/c mice, blood was collected from tail vein at various times, and radioactivity was measured. Mice were used in groups of four. Each value is a mean; bars, SD.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Supported in part by a grant from the Uehara Memorial Foundation, in part by Grants-in-Aid for Cancer Research and for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan on Priority Areas (Cancer/Biotarget), for Scientific Research and for Encouragement of Young Scientists, and in part by Health Sciences Research Grants for Research on Health Sciences focusing on Drug Innovation from the Ministry of Health and Welfare. This study supported in part by Cancer Research Fellowships of the Japan Society for the promotion of Science for Young Scientists.
The abbreviations used are: TNF, tumor necrosis factor; PVP, polyvinylpyrrolidone; PEG, polyethylene glycol; MPEG,middle fraction of PEG; DIVEMA, divinyl ether and maleic anhydride copolymer; JRU, Japan reference unit(s); IL, interleukin; DDS, drug delivery system.
Unpublished data.
Preparation of PVP-TNF-αs
Fraction no. . | Average molecular weighta . | Degree of modification (%)b . | Specific activity (× 105 JRU/mg TNF)c . | Remaining activity (%) . | Yield (%) . |
---|---|---|---|---|---|
1 | 134,000 | 70.2 | 2.40 ± 0.59 | 13.5 | 7.7 |
2 | 117,000 | 55.0 | 7.35 ± 1.12 | 41.3 | 13.5 |
3 | 101,000 | 39.9 | 9.35 ± 0.07 | 52.5 | 16.3 |
4 | 84,000 | 24.7 | 11.1 ± 0.2 | 62.5 | 13.2 |
5 | 74,000 | 9.6 | 15.3 ± 0.1 | 86.2 | 7.9 |
(Native) | 58,000 | 0 | 17.8 ± 0.5 | 100.0 | 41.4 |
MPEG-TNF-α | 108,000 | 56.0 | 11.4 ± 2.1 | 52.3 |
Fraction no. . | Average molecular weighta . | Degree of modification (%)b . | Specific activity (× 105 JRU/mg TNF)c . | Remaining activity (%) . | Yield (%) . |
---|---|---|---|---|---|
1 | 134,000 | 70.2 | 2.40 ± 0.59 | 13.5 | 7.7 |
2 | 117,000 | 55.0 | 7.35 ± 1.12 | 41.3 | 13.5 |
3 | 101,000 | 39.9 | 9.35 ± 0.07 | 52.5 | 16.3 |
4 | 84,000 | 24.7 | 11.1 ± 0.2 | 62.5 | 13.2 |
5 | 74,000 | 9.6 | 15.3 ± 0.1 | 86.2 | 7.9 |
(Native) | 58,000 | 0 | 17.8 ± 0.5 | 100.0 | 41.4 |
MPEG-TNF-α | 108,000 | 56.0 | 11.4 ± 2.1 | 52.3 |
The molecular size was determined by gel filtration chromatography (protein standard).
Calculated from molecular size.
The specific activities of native TNF-α and PVP-TNF-αs were measured by growth inhibition L-M cytotoxic assay.
Antitumor effect of PVP-TNF-α by scheduled administrationa on survival days after Meth-A tumor inoculation
Run . | Single i.v. injection dose (JRU/mouse) . | Survival timeb (Days) . | Complete regressionc . |
---|---|---|---|
Saline | 0 | 31 ± 1.0 (27, 28, 29, 29, 31, 32, 33, 33, 37) | 0 /9 |
PEG | 0 | 31 ± 1.2 (28, 28, 28, 32, 33, 34, 35) | 0 /7 |
PVP | 0 | 32 ± 1.0 (28, 29, 29, 31, 31, 31, 31, 32, 35, 36, 38) | 0 /10 |
Native TNF-α | 10,000 | >48 ± 15.9 ( 7, 7, 7, 16, 50, >100, >100, >100) | 3 /8 |
5,000 | 46 ± 1.1 (43, 43, 44, 48, 48, 48, 50) | 0 /7 | |
2,000 | 43 ± 1.0 (40, 40, 42, 43, 43, 44, 48) | 0 /8 | |
MPEG-TNF-α | 500 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 |
200 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 | |
100 | >72 ± 10.0d (46, 48, 52, 59, >100, >100, >100) | 3 /7 | |
PVP-TNF-αFr.3 | 500 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 |
200 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 | |
100 | >88 ± 5.5d (58, 60, 74, 91, >100, >100, >100, >100, >100, >100,) | 6 /10 | |
50 | >87 ± 6.1d (40, 74, 77, 89, 91, >100, >100, >100, >100, >100,) | 5 /10 |
Run . | Single i.v. injection dose (JRU/mouse) . | Survival timeb (Days) . | Complete regressionc . |
---|---|---|---|
Saline | 0 | 31 ± 1.0 (27, 28, 29, 29, 31, 32, 33, 33, 37) | 0 /9 |
PEG | 0 | 31 ± 1.2 (28, 28, 28, 32, 33, 34, 35) | 0 /7 |
PVP | 0 | 32 ± 1.0 (28, 29, 29, 31, 31, 31, 31, 32, 35, 36, 38) | 0 /10 |
Native TNF-α | 10,000 | >48 ± 15.9 ( 7, 7, 7, 16, 50, >100, >100, >100) | 3 /8 |
5,000 | 46 ± 1.1 (43, 43, 44, 48, 48, 48, 50) | 0 /7 | |
2,000 | 43 ± 1.0 (40, 40, 42, 43, 43, 44, 48) | 0 /8 | |
MPEG-TNF-α | 500 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 |
200 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 | |
100 | >72 ± 10.0d (46, 48, 52, 59, >100, >100, >100) | 3 /7 | |
PVP-TNF-αFr.3 | 500 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 |
200 | >100 ± 0d (>100, >100, >100, >100, >100, >100, >100, >100, >100, >100,) | 10 /10 | |
100 | >88 ± 5.5d (58, 60, 74, 91, >100, >100, >100, >100, >100, >100,) | 6 /10 | |
50 | >87 ± 6.1d (40, 74, 77, 89, 91, >100, >100, >100, >100, >100,) | 5 /10 |
Samples were administered on days 7, 10, 14, and 17.
Days after tumor inoculation(mean ± SE).
Complete regression was defined when tumor was not regrown for >100 days.
Statistical significance compared with saline control: P < 0.01.