- Email: [email protected]
Therapeutic benefits of oxygen radical scavenger treatments remain unproven
lournal of FreeRadicals in Biology & Medicine, Vol. I, pp. 173-177, 1985 Printed in the USA. All rightsreserved.
0748-5514/85 $3.00+ .00 PergamonPressLtd.
Hypothesis Paper THERAPEUTIC
BENEFITS OF OXYGEN RADICAL TREATMENTS REMAIN UNPROVEN
ROBERT A.
SCAVENGER
GREENWALD
Division of Rheumatology, Long Island Jewish Medical Center, New Hyde Park, NY 11042, USA
(Received 25 April 1985; Accepted 20 May 1985)
Abstract--Pharmaceutical firms and practitioners are rushing to test the medical benefits of oxy radical scavengers in a multitude of clinical situations despite the fact that convincing evidence of oxygen radical tissue damage in vivo is lacking and that properly controlled trials have been few and far between. Analysis of the therapeutic literature reveals disturbing discrepancies: unconvincing animal data, disparities between pharmacologic and enzymatic activity, and prolonged clinical improvements reported in situations where none should be expected. The proper control, inactivated scavenging enzyme, has never been used clinically or in animal models. The results of clinical trials with scavengers have a wider interpretation, since benefits are extrapolated to imply pathophysiologic mechanisms. It is especially important, therefore, no matter how hard we would like to believe that oxygen radical scavenging will be a therapeutic breakthrough, that we insist upon tightly designed clinical trials. Keywords--Arthritis, Scavengers, Superoxide dismutase, Catalase
gotten into the act as well, offering SOD tablets via health food stores, despite the obvious irrationality of oral enzyme therapy. The major areas of medical interest are arthritis, pulmonary disease, neuromuscular disorders, canine cataract and dermatoses, amelioration of radiation and/or anti-neoplastic drug toxicity, and, most excitingly in the past year, inhibition of ischemic tissue damage. Many of the disorders for which oxygen radical scavenging therapy (which I will abbreviate "oxyR x " ) is being advocated are disease processes of unknown etiology or mechanism for which conventional therapy leaves much to be desired. The desire to try new agents in hopes of a breakthrough is a natural motivating force behind medical discovery. It is my thesis that our excitement about the prospects for oxy-Rx in clinical situations cannot be used as an argument to abandon the principles of good experimental design. We must insist on controlled trials wherein attention is paid not only to the patient population, concomitant therapy, control regimen, and parameters of assessing response, but also to the nature of the agent used as the control. A great deal of the data in support of oxy-Rx is anecdotal, defective or missing totally. In lieu of the informal propagation of uncontrolled information which now permeates the literature, we must insist on solid data. To illustrate these points, I will
INTRODUCTION
At the time of this writing, there are at least four U.S. pharmaceutical firms working feverishly on the development of superoxide dismutase preparations for medical use. Their products include bovine hepatic enzyme, bovine enzyme conjugated to polyethylene glycol, and human enzyme produced by cloning techniques involving either yeast or bacteria. Similar development of catalase products is also underway, as is testing of low molecular weight compounds for anti-radical activity. It is now becoming standard in the pharmaceutical industry to screen new compounds with apparent anti-inflammatory activity for action against oxygen derived free radicals. ~ s t in chelation therapy is also resurgent, in view of elm increasingly prominent role of ferrous ion as a mediator of tissue damage. Even the proponents of hyperbaric oxygen therapy, a long standing fringe form of medical treatment for neurologic and chronic degenerative diseases, are gaining some respectability by citing the data that oxygen exposure enhances the levels of endogenous scavenging enzymes. In Europe, superoxide dismutase (SOD) preparations are being heavily promoted for treatment of a variety of rheumtic afflictions (Fig. 1). The quacks have 173
174
R.A. GREENWALD
Fig. 1. Advertisementfrom an Italian pharmaceuticalmanufacturer offering SOD for injectionas treatmentof arthritis, bursitis, tendonitis, etc. concentrate mostly on the arthritis literature, as this is my field of expertise. MECHANISTIC BACKGROUND
Inflammatory arthritis seems like the perfect place to look for medical benefits of scavengers. Joint inflammation can be induced by immune complexes, bacterial products, and crystals, all of which are stimuli of oxy radical generation by cells. Polymorphonuclear leukocytes (neutrophils) and monocyte/macrophages are cell types prominent in joint inflammation which can readily be induced to generate oxygen radicals. Hyaluronic acid, the major macromolecular constituent of synovial fluid, is a marvelous substrate for oxy radical degradation; proteoglycans and collagen are also susceptible to radical damage, albeit to a lesser extent. Joint inflammation is usually characterized by an influx of inflammatory cells, and oxy radicals have been implicated in a chemotactic mechanism (although some of this work has proved to be non-reproducible in my own and others' laboratories). Both free and bound iron
are readily available in joint fluid. Oxygen radical products can inactivate protease inhibitors or activate collagenase, thereby augmenting tissue damage. In view of the fact that intravenously injected SOD, unless conjugated or otherwise modified, has an extremely short circulating half life, the inflamed joint, a contained space, would seem to be the perfect place to test oxyRx. A perfect model of human inflammatory arthritis d o e s not exist. Anti-inflammatory drugs are usually first tested in non-specific systems such as carrageenan paw edema, sponge implant, etc. If there seems to be some therapeutic efficacy, and there are several reports of such for SOD, the next step is usually adjuvant arthritis. In susceptible rats, an injection of adjuvant (homogenised cell wall material from mycobacterium species in mineral oil) induces an inflammatory joint disease characterized by swelling, pain, erythema, systemic illness, radiologic, and histologic evidence of tissue destruction. A similar model, now in widespread use, involves injection of Type II collagen. Adjuvant arthritis provides us with our first example of the "missing data" phenomenon which permeates the clinical literature on oxy-Rx. The first mention of SOD used therapeutically as medication is a six sentence abstract that appeared in 1968. The preliminary anture of this report is exemplified by the statement that " . . . Ontosein (orgotein, SOD) is not an enzyme. ''1 Six years later, in 1974, we find a report of a clinical arthritis trial of SOD from Denmark 2 introduced by the statement that SOD has been proven effective in animal models including adjuvant arthritis; the reference cited is the 1968 abstract. If detailed data on adjuvant arthritis was published between 1968 and 1974, I have been unable to find it, although there were reports of several veterinary uses of the agent. Actually, the effectiveness of SOD for adjuvant arthritis can be questioned. The major work cited is that of Borrelli et. al. 3 from 1979, in which orgotein, 1.255.0 mg/kg was injected IP from days 7 through 21 following adjuvant, and comparison was made to phenylbutazone, 25 or 50 mg/kg. The authors reported a dose-related improvement in various parameters (paw swelling, performance time on a ,Jtating bar, sedimentation rate, and plasma fibrinogen levels). The dosage levels used in this report are best characterized as industrial strength, and the rats used, Sprague-Dawleys, are notoriously resistant to adjuvant arthritis. In another report, it was stated that the "effective dose" of SOD for this model was 3 mg/kg, substantially higher than that needed for any other conventional animal test system. Two reports from Germany4'5 revealed that while various scavengers partially ameliorated the initial transient swelling which accompanied adjuvant arthritis,
Oxygen scavengertherapy none of the substances tested (SOD, catalase, GSH, penicillamine, and others) inhibited the arthritic aspect of the disease. Most importantly, heat inactivated enzymes were j u s t as effective as native material. In our own laboratory (in experiments to be reported elsewhere in detail), neither SOD nor catalase, used as the polyethylene glycol conjugate in large doses and administered frequently by the intraperitoneal route, had any beneficial effect on adjuvant arthritis. It should be borne in mind that during the development of a conventional anti-arthritic drug, such as a propionate or oxicam, or any agent designed as a prostaglandin synthetase inhibitor, a failure to ameliorate adjuvant or collagen arthritis would likely be grounds for abandonment of the compound as a systemic agent. Two studies have tested the effect of local SOD on experimentally induced osteoarthritis models in rabbits. Such lesions can be produced by meniscectomy 6 or by immobilization 7. In the former study, both SOD and the saline-sucrose vehicle used for its delivery caused adverse metabolic consequences on the tissue and failed to ameliorate the disease. In the latter study, the SOD caused worsening of the histologic and inflammatory parameters. Except for uncontrolled trials of SOD administration to race horses, reported exclusively in the veterinary literature, animal model substantiation of SOD efficacy cannot be adduced. PHARMACOLOGY
A comprehensive review of the pharmacology of SOD is be~cond the scope of this article. The native enzyme obviously cannot be used for oral administration; someone has even gone to the trouble of proving that feeding SOD to mice did not raise tissue levels 8. It is therefore destined for either intravenous or local injection. In a remarkable article, Baret et. al.9 reported the fate of injected human CuZn-SOD and Mn-SOD in rats. They noted that CuZn-SOD disappeared rapidly, as had been well known, but that its injection seemed to be associated with a prolonged anti-inflammatory effect even though the levels attained by injection were lower than those of the animal's endogenous enzyme. Mn-SOD, on the other hand, persisted in the circulation at appreciable levels much longer, but had absolutely no pharmacologic activity; this was true despite the fact that both preparations had the same enzymatic specific activity. Disparities like this have serious implications in evaluation of the therapeutic efficacy of scavengers. To prolong the circulating life of SOD, the enzyme is often conjugated to either polyethylene glycoP °'~j or albumin 12. Such maneuvers generally prolong the enzyme's half-life while preserving most of its enzymatic and pharmacologic properties. Again, an important dis-
175
parity must be noted. Cleland et. al. ]3 studied albuminSOD conjugates produced with gluteraldehyde and noted that the product had a prolonged circulating half-life and effective anti-inflammatory activity. However, so did the control preparation, namely albumin-albumin conjugates. Despite the total lack of enzymatic activity in the control material, inhibition of rat paw edema was still detected. One thing does stand out about the pharmacology of SOD: it does appear to be without major toxicity. Both in animal and human studies, adverse effects are few and far between. Most of the observed toxicity has been attributed to the vehicle in which it is given, a problem which should be readily soluble by the pharmaceutical industry. The development of human enzyme should obviate many potential adverse reactions, such as the possibility of immune response, although SOD does appear to be a very weak immunogen. HUMAN STUDIES Studies of new agents for arthritis treatment are frought with difficulty. There is a substantial placebo response rate, generally in excess of 40%, even in the most conscientiously designed trials involving frequent clinical and/or laboratory evaluations. An effective drug most provide at least a 65% response rate for its benefit to be detectable over background unless massive numbers of patients are involved. The heterogeneity of arthritic disorders makes such response rates difficult to attain. Actually, the first human trial of a scavenger for arthritis was done with catalase rather than SOD i4. Equine liver catalase was injected in courses of daily or every other day IM injections for a total of 18 doses. Most of the patients had cervical osteoarthritis, with symptoms such as paresthesias, vertigo, or headache. No criteria of disease severity, radiologic corroboration, or definition of improvement were specified. A "global result" such as good, unchanged, or worse was recorded. The authors rated 69% of the patients as better. No control was involved. There is, of course, no conceivable a priori reason why catalase should have any beneficial effect on nerve entrapment syndromes arising from mechanical impingement secondary to degenerative, non-inflammatory changes of the spine. This is the archtypical example of the uncontrolled, haphazard trial of a putative anti-arthritic agent which makes interpretation of subsequent studies so difficult. The data improved only slightly in 1974 with publication of a Danish trial involving 22 patients with osteoarthritis of the knee:. Criteria for diagnosis were not stated, and the treatment regimens were highly variable, ranging from single injections of 2 mg to multiple
176
R.A. GREENWALD
injections totalling 30 mg. No placebo or control treatment of the contralateral or affected side was used. Interestingly, the authors describe the study as "blind," by which they mean that the assessments were made without reference to the prior visit's results. In conventional trials, " b l i n d " means that the assessor does not know if the treatment has been placebo or active agent. By a variety of parameters, the authors reported prolonged improvements lasting up to 90 days after a single injection. Eight of the patients underwent injections ostensibly into their hip joints, a form of treatment not commonly employed in the U.S. On the bright side, no adverse reactions were noted. The scientific rationale for use of SOD in rheumatoid arthritis (RA) is much stronger than for osteoarthritis (OA) since the inflammatory disorder involves processes such as chemotaxis, protease activation, and matrix dissolution, phenomena which can be linked in vitro to oxygen radicals. The first reported series of SOD treatments for RA, from France in 1980 ~5, was a negative report. In 1980, the first double blind results appeared 16. In a 16 week trial patients received SOD or placebo (identity not mentioned) by intramuscular injection for 12 weeks, followed by 4 weeks of additional observation. The treatment program was reported as having produced improvement in sixparameters (pain, use of analgesics, morning stiffness, knee circumference, grip strength, and global evaluation). Of note, however, is the fact that n o n e of these parameters improved on placebo, i.e. there was no evidence of a placebo effect. Other parameters were assessed, but the results were not reported; notably lacking was data on sedimentation rate, hemoglobin, number of painful joints, walking time, and grip strength. Curiously, the author stated that "orgotein change(s) all abnormal disease parameters in the direction of clinical improvement" but she reported only those in which improvement was noted; the others were not tabulated. It is most extraordinary to find an agent which universally improves every parameter of a complicted disease like RA. Nevertheless, the results suggested that the SOD agent was capable of short term improvement in RA. Three subsequent trials of SOD injected into joints have emanated from Germany ~7-19, two for RA and one for OA. In the first report ~7, the control group received the rather bizarre modality of intra-articular aspirin therapy, a procedure unknown to this writer (and 9 of the 15 had severe local painful reactions). Patients were treated with 6 weeks of either SOD or aspirin injections into their knees, and the efficacy of treatment was indistinguishable from the control except at a final evaluation 6 weeks after therapy had stopped, when some benefit was attributed to the SOD. In contrast, rather profound biochemical changes were reported to have
occurred throughout the entire 12 week period, including decreases in synovial fluid IgM, rheumatoid factor, PGE2 and LDH levels. Such data is uninterpretable in view of the bizarre control, disparity between biochemical and clinical changes, and the lack of improvement during treatment. The same authors followed up by comparing SOD to steroid injections, a more conventional modality 18. Once again, treatment was given for 6 weeks and no improvement was noted save at 12 or 24 weeks after injections had ceased, even though various biochemical parameters showed profound changes much earlier. Finally, in a similarly designed trial involving OA patients, 16 mg of orgotein injected four times over 6 weeks was reported to yield sustained improvement, while a dose of 8 mg or a steroid control showed deterioration 19. Although none of these studies can be used in conclusive support of the effectiveness of local SOD therapy, they do suggest that any trials contemplated in the future be designed with a long term followup component. THE PROPER CONTROL
We have no direct evidence that oxygen radicals are indeed involved in inflammatory processes. There exists no probe which can be introduced into an animal (or human) system, either locally or systemically, where it will undergo a specific, irreversible change that can be unequivocally attributed to an oxy radical product, and which can be quantitated upon recovery of the probe. Development of such a probe, which must also be nontoxic, and non-metabolized, would be a gre~at step forward in tracking down the elusive free radicals in clinical medicine. Our evidence is now only indirect: we extrapolate from in vitro studies of cells, we have circumstantial evidence based on effects which appear to be abolished by scavengers, and we invoke indirect data such as detection of products of free radical action, notably lipid peroxidation reactions, isolated from inflammatory sites. Therapeutic trials therefore have a much wider implication than that of solely providing hints about treatment. The results are invoked mechanistically to attempt to implicate oxy radicals in pathophysiology; if something gets better after SOD or catalase is given, a radical mechanism is surmised. In view of the findings cited above about disparities between enzymatic and anti-inflammatory activity, and the clinical discrepancies between biochemical changes and the time course of improvement, such extrapolations to mechanism appear fraught with danger. If SOD improves things by scavenging superoxide and thereby preventing either direct effects of that radical or secondary generation of other radicals, then it
Oxygen scavenger therapy
seems to me that the proper control is inactive enzyme. Since the apoenzyme reconstitutes itself quickly after injection into a mammalian organism, even exhaustive dialysis of CuZn-SOD against a chelator will not create a proper control. However, SOD can be permanently inactivated by the simple expedient of incubation at pH 9.5 in 10 mM hydrogen peroxide for a short period of time. The protein will be unchanged in primary structure or metal content, but it will lose its dismutase activity. This seems to me to be the proper control for studies of the putative therapeutic effect of SOD; until a properly designed trial has been conducted using such a control, all claims for mechanistic treatment with oxyRx are subject to dispute.
8. 9.
10.
11.
12.
REFERENCES
13. 1. W. Huber, T. L. Schulte, S. Carson, R. E. Goldhamer, and E. E. Vogin. Some chemical and pharmacologic properties of a novel anti-inflammatory protein. Toxicol. Appl. Pharmacol. 12:308 (1968). 2. K. Lund-Olesem amd K. B. Menander. Orgotein: A new antiinflammatory metalloprotein drug. Preliminary evaluation of clinical efficacy and safety in degenerative joint disease. Curr. Ther. Res. 16:706-717 (1974). 3. E Borrelli, C. Serafini, and G. Mattalia. Additional pharmacological aspects of orgotein, a metalloprotein with superoxide dismutase activity. ArzneimitteI-Forsch. 29:781-785 (1979). 4. R. Hirschelmann and H. Bekemeier. Effects of catalase, peroxidase, superoxide dismutase and 10 scavengers of oxygen radicals in carrageenin edema and in adjuvant arthritis of rats. Experientia 37:1313-1314 ( 1981 ). 5. Hirschelmann, R., and H. Bekemeier. Influence of treatment with catalase inhibitors on carrageenin paw edema and adjuvant arthritis in rats. Agents Actions 15:52 (1984). 6. I. A. Rosner, V. A. Goldberg, L. Getzy, and R. M. Moskowitz. A trial of intraarticular orgotein, a superoxide dismutase, in experimentally induced OA. J. Rheumatol. 7:24-29 (1980). 7. J. Gialamas, H. Hoger, G. Partsch, J. Neumulelr, and D. Adamiker. Histologische Untersuchungen uber die Wirkung der Su-
14. 15.
16.
17. 18.
19.
177
peroxid-Dismutase (SOD) an einem Arthrosemodell beim Kaninchen. Zeit fur R heumatologie 43:142-147 (1984). S. Cherr, C. L. Keen, B. Lonnerdal, and L. S. Hurley. Dietary superoxide dismutase does not affect tissue levels. Amer. J. Clin. Nutr. 37:5-7 (1983). A. Baret, G. Jadot, and A. M. Michelson. Pharmacokinetic and anti-inflammatory properties in the rat of superoxide dismutases (Cu SODs and Mn SOD) from various species. Biochem. Pharmac. 33:2755-2760 (1984). F. M. Veronese, E. Boccu, O., Schiavon, et. al. Anti-inflammatory and pharmacokinetic properties of superoxide dismutase derivatized with polyethylene glycol via active esters. J. Pharm. Pharmac. 35:757-758 (1983). P. S. Pyatak, A. Abuchowski, and E F. Davis. Preparation of a polyethylene glycol: superoxide dismutase adduct, and an examination of its blood circulating life and anti-inflammatory activity. Res. Commun. Chem. Pathol. Pharmac. 29:113-127 (1980). K. Wong, L. G. Cleland, and M J. Poznansky. Enhanced antiinflammatory effect and reduced immunogenicity of bovine liver superoxide dismutase by conjugation with homologous albumin. Agents Actions 10:231-239 (1980). L. Cleland, J. Bielicki, B. Vernon-Roberts, and W. H. Betts. Superoxide dismutase (SOD) and albumin conjugates with delayed clearance from plasma and body cavities. Is SOD antiinflammatory? In: Oxy Radicals and their Scavenger Systems, Volume II (R. A. Greenwald an G. Cohen, eds.), pp. 268-273 Elsevier, Amsterdam (1983). R. Riu, R. Le Den, and C. Le Mouel. Interet de L'Equilase en O.R.L. Medicine Interne 6:811-815 (1971). J. P. Camus, I. Emerit, A. M. Michelson, A. Prier, A. C. Koeger, and C. Merlet. Superoxyde dismutase et polyarthrite rhumatoide. Revue Rhum. 47:489-492 (1980). K. Menander-Huber, K. Double blind controlled trials in man with bovine copper zinc superoxide dismutase (orgotein). In: Biological and Clinical Aspects of Superoxide and Superoxide Dismutase (W. H. Bannister and J. V. Bannister, eds.) pp. 408423 Elsevier, Amsterdam (1980). K. M. Goebel, U. Storck, and F. Neurath. Intrasynovial orgotein therapy in rheumatoid arthritis. Lancet 1:1015-1017 (1981). K. Goebeland U. Storck. Effect ofintra-articular orgotein versus a corticosteroid on rheumatoid arthritis of the knees. Am. J. Med. 74:124-128 (1983). W. Gammer and L. G. Broback. Clinical comparison of orgotein and methylpredisolone acetate in the treatment of osteoarthritis of the knee joint. Scand. J. Rheumatol. 13:108-112 (1984).
0748-5514/85 $3.00+ .00 PergamonPressLtd.
Hypothesis Paper THERAPEUTIC
BENEFITS OF OXYGEN RADICAL TREATMENTS REMAIN UNPROVEN
ROBERT A.
SCAVENGER
GREENWALD
Division of Rheumatology, Long Island Jewish Medical Center, New Hyde Park, NY 11042, USA
(Received 25 April 1985; Accepted 20 May 1985)
Abstract--Pharmaceutical firms and practitioners are rushing to test the medical benefits of oxy radical scavengers in a multitude of clinical situations despite the fact that convincing evidence of oxygen radical tissue damage in vivo is lacking and that properly controlled trials have been few and far between. Analysis of the therapeutic literature reveals disturbing discrepancies: unconvincing animal data, disparities between pharmacologic and enzymatic activity, and prolonged clinical improvements reported in situations where none should be expected. The proper control, inactivated scavenging enzyme, has never been used clinically or in animal models. The results of clinical trials with scavengers have a wider interpretation, since benefits are extrapolated to imply pathophysiologic mechanisms. It is especially important, therefore, no matter how hard we would like to believe that oxygen radical scavenging will be a therapeutic breakthrough, that we insist upon tightly designed clinical trials. Keywords--Arthritis, Scavengers, Superoxide dismutase, Catalase
gotten into the act as well, offering SOD tablets via health food stores, despite the obvious irrationality of oral enzyme therapy. The major areas of medical interest are arthritis, pulmonary disease, neuromuscular disorders, canine cataract and dermatoses, amelioration of radiation and/or anti-neoplastic drug toxicity, and, most excitingly in the past year, inhibition of ischemic tissue damage. Many of the disorders for which oxygen radical scavenging therapy (which I will abbreviate "oxyR x " ) is being advocated are disease processes of unknown etiology or mechanism for which conventional therapy leaves much to be desired. The desire to try new agents in hopes of a breakthrough is a natural motivating force behind medical discovery. It is my thesis that our excitement about the prospects for oxy-Rx in clinical situations cannot be used as an argument to abandon the principles of good experimental design. We must insist on controlled trials wherein attention is paid not only to the patient population, concomitant therapy, control regimen, and parameters of assessing response, but also to the nature of the agent used as the control. A great deal of the data in support of oxy-Rx is anecdotal, defective or missing totally. In lieu of the informal propagation of uncontrolled information which now permeates the literature, we must insist on solid data. To illustrate these points, I will
INTRODUCTION
At the time of this writing, there are at least four U.S. pharmaceutical firms working feverishly on the development of superoxide dismutase preparations for medical use. Their products include bovine hepatic enzyme, bovine enzyme conjugated to polyethylene glycol, and human enzyme produced by cloning techniques involving either yeast or bacteria. Similar development of catalase products is also underway, as is testing of low molecular weight compounds for anti-radical activity. It is now becoming standard in the pharmaceutical industry to screen new compounds with apparent anti-inflammatory activity for action against oxygen derived free radicals. ~ s t in chelation therapy is also resurgent, in view of elm increasingly prominent role of ferrous ion as a mediator of tissue damage. Even the proponents of hyperbaric oxygen therapy, a long standing fringe form of medical treatment for neurologic and chronic degenerative diseases, are gaining some respectability by citing the data that oxygen exposure enhances the levels of endogenous scavenging enzymes. In Europe, superoxide dismutase (SOD) preparations are being heavily promoted for treatment of a variety of rheumtic afflictions (Fig. 1). The quacks have 173
174
R.A. GREENWALD
Fig. 1. Advertisementfrom an Italian pharmaceuticalmanufacturer offering SOD for injectionas treatmentof arthritis, bursitis, tendonitis, etc. concentrate mostly on the arthritis literature, as this is my field of expertise. MECHANISTIC BACKGROUND
Inflammatory arthritis seems like the perfect place to look for medical benefits of scavengers. Joint inflammation can be induced by immune complexes, bacterial products, and crystals, all of which are stimuli of oxy radical generation by cells. Polymorphonuclear leukocytes (neutrophils) and monocyte/macrophages are cell types prominent in joint inflammation which can readily be induced to generate oxygen radicals. Hyaluronic acid, the major macromolecular constituent of synovial fluid, is a marvelous substrate for oxy radical degradation; proteoglycans and collagen are also susceptible to radical damage, albeit to a lesser extent. Joint inflammation is usually characterized by an influx of inflammatory cells, and oxy radicals have been implicated in a chemotactic mechanism (although some of this work has proved to be non-reproducible in my own and others' laboratories). Both free and bound iron
are readily available in joint fluid. Oxygen radical products can inactivate protease inhibitors or activate collagenase, thereby augmenting tissue damage. In view of the fact that intravenously injected SOD, unless conjugated or otherwise modified, has an extremely short circulating half life, the inflamed joint, a contained space, would seem to be the perfect place to test oxyRx. A perfect model of human inflammatory arthritis d o e s not exist. Anti-inflammatory drugs are usually first tested in non-specific systems such as carrageenan paw edema, sponge implant, etc. If there seems to be some therapeutic efficacy, and there are several reports of such for SOD, the next step is usually adjuvant arthritis. In susceptible rats, an injection of adjuvant (homogenised cell wall material from mycobacterium species in mineral oil) induces an inflammatory joint disease characterized by swelling, pain, erythema, systemic illness, radiologic, and histologic evidence of tissue destruction. A similar model, now in widespread use, involves injection of Type II collagen. Adjuvant arthritis provides us with our first example of the "missing data" phenomenon which permeates the clinical literature on oxy-Rx. The first mention of SOD used therapeutically as medication is a six sentence abstract that appeared in 1968. The preliminary anture of this report is exemplified by the statement that " . . . Ontosein (orgotein, SOD) is not an enzyme. ''1 Six years later, in 1974, we find a report of a clinical arthritis trial of SOD from Denmark 2 introduced by the statement that SOD has been proven effective in animal models including adjuvant arthritis; the reference cited is the 1968 abstract. If detailed data on adjuvant arthritis was published between 1968 and 1974, I have been unable to find it, although there were reports of several veterinary uses of the agent. Actually, the effectiveness of SOD for adjuvant arthritis can be questioned. The major work cited is that of Borrelli et. al. 3 from 1979, in which orgotein, 1.255.0 mg/kg was injected IP from days 7 through 21 following adjuvant, and comparison was made to phenylbutazone, 25 or 50 mg/kg. The authors reported a dose-related improvement in various parameters (paw swelling, performance time on a ,Jtating bar, sedimentation rate, and plasma fibrinogen levels). The dosage levels used in this report are best characterized as industrial strength, and the rats used, Sprague-Dawleys, are notoriously resistant to adjuvant arthritis. In another report, it was stated that the "effective dose" of SOD for this model was 3 mg/kg, substantially higher than that needed for any other conventional animal test system. Two reports from Germany4'5 revealed that while various scavengers partially ameliorated the initial transient swelling which accompanied adjuvant arthritis,
Oxygen scavengertherapy none of the substances tested (SOD, catalase, GSH, penicillamine, and others) inhibited the arthritic aspect of the disease. Most importantly, heat inactivated enzymes were j u s t as effective as native material. In our own laboratory (in experiments to be reported elsewhere in detail), neither SOD nor catalase, used as the polyethylene glycol conjugate in large doses and administered frequently by the intraperitoneal route, had any beneficial effect on adjuvant arthritis. It should be borne in mind that during the development of a conventional anti-arthritic drug, such as a propionate or oxicam, or any agent designed as a prostaglandin synthetase inhibitor, a failure to ameliorate adjuvant or collagen arthritis would likely be grounds for abandonment of the compound as a systemic agent. Two studies have tested the effect of local SOD on experimentally induced osteoarthritis models in rabbits. Such lesions can be produced by meniscectomy 6 or by immobilization 7. In the former study, both SOD and the saline-sucrose vehicle used for its delivery caused adverse metabolic consequences on the tissue and failed to ameliorate the disease. In the latter study, the SOD caused worsening of the histologic and inflammatory parameters. Except for uncontrolled trials of SOD administration to race horses, reported exclusively in the veterinary literature, animal model substantiation of SOD efficacy cannot be adduced. PHARMACOLOGY
A comprehensive review of the pharmacology of SOD is be~cond the scope of this article. The native enzyme obviously cannot be used for oral administration; someone has even gone to the trouble of proving that feeding SOD to mice did not raise tissue levels 8. It is therefore destined for either intravenous or local injection. In a remarkable article, Baret et. al.9 reported the fate of injected human CuZn-SOD and Mn-SOD in rats. They noted that CuZn-SOD disappeared rapidly, as had been well known, but that its injection seemed to be associated with a prolonged anti-inflammatory effect even though the levels attained by injection were lower than those of the animal's endogenous enzyme. Mn-SOD, on the other hand, persisted in the circulation at appreciable levels much longer, but had absolutely no pharmacologic activity; this was true despite the fact that both preparations had the same enzymatic specific activity. Disparities like this have serious implications in evaluation of the therapeutic efficacy of scavengers. To prolong the circulating life of SOD, the enzyme is often conjugated to either polyethylene glycoP °'~j or albumin 12. Such maneuvers generally prolong the enzyme's half-life while preserving most of its enzymatic and pharmacologic properties. Again, an important dis-
175
parity must be noted. Cleland et. al. ]3 studied albuminSOD conjugates produced with gluteraldehyde and noted that the product had a prolonged circulating half-life and effective anti-inflammatory activity. However, so did the control preparation, namely albumin-albumin conjugates. Despite the total lack of enzymatic activity in the control material, inhibition of rat paw edema was still detected. One thing does stand out about the pharmacology of SOD: it does appear to be without major toxicity. Both in animal and human studies, adverse effects are few and far between. Most of the observed toxicity has been attributed to the vehicle in which it is given, a problem which should be readily soluble by the pharmaceutical industry. The development of human enzyme should obviate many potential adverse reactions, such as the possibility of immune response, although SOD does appear to be a very weak immunogen. HUMAN STUDIES Studies of new agents for arthritis treatment are frought with difficulty. There is a substantial placebo response rate, generally in excess of 40%, even in the most conscientiously designed trials involving frequent clinical and/or laboratory evaluations. An effective drug most provide at least a 65% response rate for its benefit to be detectable over background unless massive numbers of patients are involved. The heterogeneity of arthritic disorders makes such response rates difficult to attain. Actually, the first human trial of a scavenger for arthritis was done with catalase rather than SOD i4. Equine liver catalase was injected in courses of daily or every other day IM injections for a total of 18 doses. Most of the patients had cervical osteoarthritis, with symptoms such as paresthesias, vertigo, or headache. No criteria of disease severity, radiologic corroboration, or definition of improvement were specified. A "global result" such as good, unchanged, or worse was recorded. The authors rated 69% of the patients as better. No control was involved. There is, of course, no conceivable a priori reason why catalase should have any beneficial effect on nerve entrapment syndromes arising from mechanical impingement secondary to degenerative, non-inflammatory changes of the spine. This is the archtypical example of the uncontrolled, haphazard trial of a putative anti-arthritic agent which makes interpretation of subsequent studies so difficult. The data improved only slightly in 1974 with publication of a Danish trial involving 22 patients with osteoarthritis of the knee:. Criteria for diagnosis were not stated, and the treatment regimens were highly variable, ranging from single injections of 2 mg to multiple
176
R.A. GREENWALD
injections totalling 30 mg. No placebo or control treatment of the contralateral or affected side was used. Interestingly, the authors describe the study as "blind," by which they mean that the assessments were made without reference to the prior visit's results. In conventional trials, " b l i n d " means that the assessor does not know if the treatment has been placebo or active agent. By a variety of parameters, the authors reported prolonged improvements lasting up to 90 days after a single injection. Eight of the patients underwent injections ostensibly into their hip joints, a form of treatment not commonly employed in the U.S. On the bright side, no adverse reactions were noted. The scientific rationale for use of SOD in rheumatoid arthritis (RA) is much stronger than for osteoarthritis (OA) since the inflammatory disorder involves processes such as chemotaxis, protease activation, and matrix dissolution, phenomena which can be linked in vitro to oxygen radicals. The first reported series of SOD treatments for RA, from France in 1980 ~5, was a negative report. In 1980, the first double blind results appeared 16. In a 16 week trial patients received SOD or placebo (identity not mentioned) by intramuscular injection for 12 weeks, followed by 4 weeks of additional observation. The treatment program was reported as having produced improvement in sixparameters (pain, use of analgesics, morning stiffness, knee circumference, grip strength, and global evaluation). Of note, however, is the fact that n o n e of these parameters improved on placebo, i.e. there was no evidence of a placebo effect. Other parameters were assessed, but the results were not reported; notably lacking was data on sedimentation rate, hemoglobin, number of painful joints, walking time, and grip strength. Curiously, the author stated that "orgotein change(s) all abnormal disease parameters in the direction of clinical improvement" but she reported only those in which improvement was noted; the others were not tabulated. It is most extraordinary to find an agent which universally improves every parameter of a complicted disease like RA. Nevertheless, the results suggested that the SOD agent was capable of short term improvement in RA. Three subsequent trials of SOD injected into joints have emanated from Germany ~7-19, two for RA and one for OA. In the first report ~7, the control group received the rather bizarre modality of intra-articular aspirin therapy, a procedure unknown to this writer (and 9 of the 15 had severe local painful reactions). Patients were treated with 6 weeks of either SOD or aspirin injections into their knees, and the efficacy of treatment was indistinguishable from the control except at a final evaluation 6 weeks after therapy had stopped, when some benefit was attributed to the SOD. In contrast, rather profound biochemical changes were reported to have
occurred throughout the entire 12 week period, including decreases in synovial fluid IgM, rheumatoid factor, PGE2 and LDH levels. Such data is uninterpretable in view of the bizarre control, disparity between biochemical and clinical changes, and the lack of improvement during treatment. The same authors followed up by comparing SOD to steroid injections, a more conventional modality 18. Once again, treatment was given for 6 weeks and no improvement was noted save at 12 or 24 weeks after injections had ceased, even though various biochemical parameters showed profound changes much earlier. Finally, in a similarly designed trial involving OA patients, 16 mg of orgotein injected four times over 6 weeks was reported to yield sustained improvement, while a dose of 8 mg or a steroid control showed deterioration 19. Although none of these studies can be used in conclusive support of the effectiveness of local SOD therapy, they do suggest that any trials contemplated in the future be designed with a long term followup component. THE PROPER CONTROL
We have no direct evidence that oxygen radicals are indeed involved in inflammatory processes. There exists no probe which can be introduced into an animal (or human) system, either locally or systemically, where it will undergo a specific, irreversible change that can be unequivocally attributed to an oxy radical product, and which can be quantitated upon recovery of the probe. Development of such a probe, which must also be nontoxic, and non-metabolized, would be a gre~at step forward in tracking down the elusive free radicals in clinical medicine. Our evidence is now only indirect: we extrapolate from in vitro studies of cells, we have circumstantial evidence based on effects which appear to be abolished by scavengers, and we invoke indirect data such as detection of products of free radical action, notably lipid peroxidation reactions, isolated from inflammatory sites. Therapeutic trials therefore have a much wider implication than that of solely providing hints about treatment. The results are invoked mechanistically to attempt to implicate oxy radicals in pathophysiology; if something gets better after SOD or catalase is given, a radical mechanism is surmised. In view of the findings cited above about disparities between enzymatic and anti-inflammatory activity, and the clinical discrepancies between biochemical changes and the time course of improvement, such extrapolations to mechanism appear fraught with danger. If SOD improves things by scavenging superoxide and thereby preventing either direct effects of that radical or secondary generation of other radicals, then it
Oxygen scavenger therapy
seems to me that the proper control is inactive enzyme. Since the apoenzyme reconstitutes itself quickly after injection into a mammalian organism, even exhaustive dialysis of CuZn-SOD against a chelator will not create a proper control. However, SOD can be permanently inactivated by the simple expedient of incubation at pH 9.5 in 10 mM hydrogen peroxide for a short period of time. The protein will be unchanged in primary structure or metal content, but it will lose its dismutase activity. This seems to me to be the proper control for studies of the putative therapeutic effect of SOD; until a properly designed trial has been conducted using such a control, all claims for mechanistic treatment with oxyRx are subject to dispute.
8. 9.
10.
11.
12.
REFERENCES
13. 1. W. Huber, T. L. Schulte, S. Carson, R. E. Goldhamer, and E. E. Vogin. Some chemical and pharmacologic properties of a novel anti-inflammatory protein. Toxicol. Appl. Pharmacol. 12:308 (1968). 2. K. Lund-Olesem amd K. B. Menander. Orgotein: A new antiinflammatory metalloprotein drug. Preliminary evaluation of clinical efficacy and safety in degenerative joint disease. Curr. Ther. Res. 16:706-717 (1974). 3. E Borrelli, C. Serafini, and G. Mattalia. Additional pharmacological aspects of orgotein, a metalloprotein with superoxide dismutase activity. ArzneimitteI-Forsch. 29:781-785 (1979). 4. R. Hirschelmann and H. Bekemeier. Effects of catalase, peroxidase, superoxide dismutase and 10 scavengers of oxygen radicals in carrageenin edema and in adjuvant arthritis of rats. Experientia 37:1313-1314 ( 1981 ). 5. Hirschelmann, R., and H. Bekemeier. Influence of treatment with catalase inhibitors on carrageenin paw edema and adjuvant arthritis in rats. Agents Actions 15:52 (1984). 6. I. A. Rosner, V. A. Goldberg, L. Getzy, and R. M. Moskowitz. A trial of intraarticular orgotein, a superoxide dismutase, in experimentally induced OA. J. Rheumatol. 7:24-29 (1980). 7. J. Gialamas, H. Hoger, G. Partsch, J. Neumulelr, and D. Adamiker. Histologische Untersuchungen uber die Wirkung der Su-
14. 15.
16.
17. 18.
19.
177
peroxid-Dismutase (SOD) an einem Arthrosemodell beim Kaninchen. Zeit fur R heumatologie 43:142-147 (1984). S. Cherr, C. L. Keen, B. Lonnerdal, and L. S. Hurley. Dietary superoxide dismutase does not affect tissue levels. Amer. J. Clin. Nutr. 37:5-7 (1983). A. Baret, G. Jadot, and A. M. Michelson. Pharmacokinetic and anti-inflammatory properties in the rat of superoxide dismutases (Cu SODs and Mn SOD) from various species. Biochem. Pharmac. 33:2755-2760 (1984). F. M. Veronese, E. Boccu, O., Schiavon, et. al. Anti-inflammatory and pharmacokinetic properties of superoxide dismutase derivatized with polyethylene glycol via active esters. J. Pharm. Pharmac. 35:757-758 (1983). P. S. Pyatak, A. Abuchowski, and E F. Davis. Preparation of a polyethylene glycol: superoxide dismutase adduct, and an examination of its blood circulating life and anti-inflammatory activity. Res. Commun. Chem. Pathol. Pharmac. 29:113-127 (1980). K. Wong, L. G. Cleland, and M J. Poznansky. Enhanced antiinflammatory effect and reduced immunogenicity of bovine liver superoxide dismutase by conjugation with homologous albumin. Agents Actions 10:231-239 (1980). L. Cleland, J. Bielicki, B. Vernon-Roberts, and W. H. Betts. Superoxide dismutase (SOD) and albumin conjugates with delayed clearance from plasma and body cavities. Is SOD antiinflammatory? In: Oxy Radicals and their Scavenger Systems, Volume II (R. A. Greenwald an G. Cohen, eds.), pp. 268-273 Elsevier, Amsterdam (1983). R. Riu, R. Le Den, and C. Le Mouel. Interet de L'Equilase en O.R.L. Medicine Interne 6:811-815 (1971). J. P. Camus, I. Emerit, A. M. Michelson, A. Prier, A. C. Koeger, and C. Merlet. Superoxyde dismutase et polyarthrite rhumatoide. Revue Rhum. 47:489-492 (1980). K. Menander-Huber, K. Double blind controlled trials in man with bovine copper zinc superoxide dismutase (orgotein). In: Biological and Clinical Aspects of Superoxide and Superoxide Dismutase (W. H. Bannister and J. V. Bannister, eds.) pp. 408423 Elsevier, Amsterdam (1980). K. M. Goebel, U. Storck, and F. Neurath. Intrasynovial orgotein therapy in rheumatoid arthritis. Lancet 1:1015-1017 (1981). K. Goebeland U. Storck. Effect ofintra-articular orgotein versus a corticosteroid on rheumatoid arthritis of the knees. Am. J. Med. 74:124-128 (1983). W. Gammer and L. G. Broback. Clinical comparison of orgotein and methylpredisolone acetate in the treatment of osteoarthritis of the knee joint. Scand. J. Rheumatol. 13:108-112 (1984).