COX-2 inhibition prevents insulin-dependent diabetes in low-dose streptozotocin-treated mice

Insulin-dependent diabetes mellitus (IDDM) is an autoimmune disease believed to be caused by an inflammatory process in the pancreas leading to selective destruction of the beta cells. Inducible cyclooxygenase (COX-2) is expressed under inflammatory conditions and its product prostaglandin E(2) (PGE(2)) is an important inflammation mediator. We report here that administration of the selective COX-2 inhibitor NS-398 prevents the onset of diabetes in mice brought on by multiple low-doses of streptozotocin (STZ). Histological observations indicated that STZ-mediated destruction of beta cells was prevented by NS-398 treatment. Delayed (day 3) administration of NS-398 was also protective in this model. No protective effect was observed when NS-398 was administered prior to a high, toxic dose of STZ. These results demonstrate the critical importance of COX-2 activity in autoimmune destruction of beta cells, and point to the fact that COX-2 inhibition can potentially develop into a preventive therapy against IDDM.     

Protein Kinases Are Involved in Prolonged Acetylcholine Release from Rat Hippocampus Induced by Thyrotropin-Releasing Hormone Analogue NS-3

Abstract: The effects of various protein kinase inhibitors on acetylcholine release from the rat hippocampus induced by the local application of NS-3 (montirelin hydrate, CG-3703), a thyrotropin-releasing hormone analogue, into the medial septum-diagonal band were examined using in vivo microdialysis. Perfusion of NS-3 (1 µ M ) into the medial septum-diagonal band for 20 min produced a pronounced and prolonged increase in the hippocampal acetylcholine efflux. Pretreatment of the medial septum-diagonal band with either K-252a, a nonselective protein kinase inhibitor, or selective protein kinase A inhibitor H-89 almost completely blocked the acetylcholine efflux evoked by NS-3, and selective protein kinase C inhibitor calphostin C inhibited the action of NS-3. On the other hand, NS-3 (0.1–10 µ M ) or TRH (1–100 µ M ) increased the cyclic AMP efflux from the medial septum-diagonal band in a concentration-dependent manner, as measured by microdialysis. These findings suggest that protein kinases A and C in the neurons of the medial septum-diagonal band are involved in the mechanism of the prolonged stimulation of acetylcholine release from the hippocampus induced by thyrotropin-releasing hormone and its analogue, NS-3.     

A comparison of the effects of indomethacin and NS-398 (a selective prostaglandin H synthase-2 inhibitor) on implantation in the rat

Indomethacin (25 microM) inhibited the amount of prostaglandin (PG) E2, PGF2alpha and 6-keto-PGF1alpha synthesized by homogenates of day 5 pregnant rat uterus by 80-92%. In contrast, 25 microM NS-398, a selective inhibitor of prostaglandin H synthase-2 (PGHS-2), inhibited the synthesis of these three prostaglandins by homogenates of the same tissue by only 37-60%. Since it has been reported that idomethacin and NS-398 inhibit PGHS-2 with similar potencies and that indomethacin (unlike NS-398) is a potent inhibitor of PGHS-1, it may be concluded that prostaglandin production by homogenates of the rat uterus on day 5 of pregnancy is due to the activities of both PGHS-1 and PGHS-2. The administration of indomethacin (3 mg/kg) twice daily on days 3 and 4 of pregnancy reduced the implantation rate by 77%, whereas the similar administration of NS-398 (6 mg/kg) had no significant effect on implantation. It may be concluded that either the dose of NS-398 used was too low to affect uterine PGHS-2 activity in vivo sufficiently to prevent implantation, or that inhibiting uterine PGHS-2 activity in vivo has no effect on the implantation mechanisms or is compensated for by increased PGHS-1 activity such that the implantation process is not impaired.     

Influence of schedule on the therapeutic efficacy of chemoimmunotherapy with doxorubicin and interleukin-2

Combinations of chemotherapeutic agents with recombinant interleukin-2 are currently under investigation in Phase I/II clinical trials as a possible means of improving response rates for metastatic melanoma, breast cancer, non small cell lung and head/neck carcinomas. As chemotherapy often induces marked immune suppressionin vivo, the way in which these agents are combined may be of critical consideration to the therapeutic outcome. Using a rat tumour model, this study aimed to define an optimal schedule for the combined administration of doxorubicin (DOX) with interleukin-2 (IL-2). DOX (4.5 mg/kg bolus i.v.) was administered 24 hours before, during, or 24 hours after, IL-2 immunotherapy (1 x 10⁵ Cetus U/rat/day for 5 days continuous i.v. infusion) to WAG rats bearing hind limb solid colonic adenocarcinoma implants. Tumour measurements taken over the 4 week study period revealed that there was no significant difference in tumour growth inhibition between the three schedules. Furthermore, DOX invariably caused a marked suppression in the rebound lymphoproliferation after cessation of IL-2 therapy (P < 0.001). These results demonstrate that the therapeutic efficacy of the DOX/IL-2 combination is not influenced by the schedule for the administration of these agents within the times of administration investigated in this study.     

NS-7 (nervous system antigen-7): a cell surface antigen of mature brain, kidney, and spermatozoa shared by embryonal tissues and transformed cells

Xenogeneic immunization of rabbits with cells of a mouse undifferentiated neural tumor, MUNTAD, produces antisera which after appropriate absorption bind to the cell surfaces of MUNTAD cells and to three normal adult mouse tissues: brain, kidney, and spermatozoa. Any one of these tissues quantitatively absorbs all of the activity for MUNTAD cells from the antisera, indicating that a single antigen or antigens is shared. This is designated NS-7. The adult tissues positive for NS-7 are the same as those that carry NS-6, but NS-7 is distinguished serologically from NS-6 by being present on many embryonic tissues, tumors, and cell lines which completely lack NS-6. Blocking and capping studies show that on the MUNTAD cell surface, some NS-7 antigenic sites and all NS-6 sites are physically related.     

Cyclooxygenase-2 selective and nitric oxide-releasing nonsteroidal anti-inflammatory drugs and gastric mucosal responses.

Occurrence of gastrointestinal damage and delayed healing of pre-existing ulcer are commonly observed in association with clinical use of nonsteroidal antiinflammatory drugs (NSAIDs). We examined the effects of NS-398, the cyclooxygenase (COX)-2 selective inhibitor, and nitric oxide (NO)- releasing aspirin (NCX-4016) on gastric mucosal ulcerogenic and healing responses in experimental animals, in comparison with those of nonselective COX inhibitors such as indomethacin and aspirin. Indomethacin and aspirin given orally were ulcerogenic by themselves in rat stomachs, while either NS-398 or NCX-4016 was not ulcerogenic at the doses which exert the equipotent antiinflammatory action with indomethacin or aspirin. Among these NSAIDs, only NCX-4016 showed a dose-dependent protection against gastric lesions induced by HCl/ethanol in rats. On the other hand, the healing of gastric ulcers induced in mice by thermal-cauterization was significantly delayed by repeated administration of these NSAIDs for more than 7 days, except NCX-4016. Gastric mucosal prostaglandin contents were reduced by indomethacin, aspirin and NCX-4016 in both normal and ulcerated mucosa, while NS-398 significantly decreased prostaglandin generation only in the ulcerated mucosa. Oral administration of NCX-4016 in pylorus-ligated rats and mice increased the levels of NO metabolites in the gastric contents. In addition, both NS-398 and NCX-4016 showed an equipotent anti-inflammatory effect against carrageenan-induced paw edema in rats as compared with indomethacin and aspirin. These results suggest that both indomethacin and aspirin are ulcerogenic by themselves and impair the healing of pre-existing gastric ulcers as well. The former action is due to inhibition of COX-1, while the latter effect may be accounted for by inhibition of COX-2 and mimicked by NS-398, the COX-2 selective NSAID. NCX-4016, despite inhibiting both COX-1 and COX-2, protects the stomach against damage and preserves the healing response of gastric ulcers, probably because of the beneficial action of NO.     

Serological Differentiation between HCV Subtypes 1a and 1b by a Recombinant Immunoblot Assay

Until now, no serological assay has been available for the differentiation of HCV subtypes. Since there is evidence that the subtypes differently influence the clinical course of HCV infection and the outcome of interferon therapy, we established a strip immunoblot assay (NS-4 IBA) with recombinant HCV proteins of the nonstructural 4 (NS-4) region propagated in Escherichia coli. Using this NS-4 IBA, we were able to distinguish HCV subtypes la and 1b, which are the most prevalent subtypes in Europe and the U.S.A. The results of the serotyping assay were compared with those obtained by nucleotide sequencing from the NS-5 region. Concordant results were observed to match 94.9% (n = 100) by the NS-4 IBA and nucleotide sequencing. Discrepant results were obtained in only 5.1% (n = 6). These data indicate that HCV subtypes can be serologically distinguished, providing the possibility for easier identification of infection with different KCV subtypes.     

The structure of NS-398 bound to cyclooxygenase-2

The cyclooxygenases (COX-1 and COX-2) are membrane-associated, heme-containing homodimers that generate prostaglandin H₂ from arachidonic acid (AA) in the committed step of prostaglandin biogenesis and are the targets for nonsteroidal anti-inflammatory drugs (NSAIDs). N-(2-cyclohexyloxy-4-nitrophenyl) methanesulfonamide (NS-398) was the first in a series of isoform-selective drugs designed to preferentially inhibit COX-2, with the aim of ameliorating many of the toxic gastrointestinal side effects caused by conventional NSAID inhibition. We determined the X-ray crystal structure of murine COX-2 in complex with NS-398 utilizing synchrotron radiation to 3.0 A resolution. NS-398 binds in the cyclooxygenase channel in a conformation that is different than that observed for other COX-2-selective inhibitors, such as celecoxib, with no discernible penetration into the side pocket formed in COX-2 by the isoform-specific substitutions of I434V, H513R, and I523V. Instead, the methanesulfonamide moiety of NS-398 interacts with the side chain of Arg-120 at the opening of the cyclooxygenase channel, similar to that observed for acidic, nonselective NSAIDs such as indomethacin and flurbiprofen. Our structure validates inhibitor studies that identified Arg-120 as a molecular determinant for time-dependent inhibition of COX-2 by NS-398.     

Therapeutic Utilization of the Diurnal Variation in Pituitary-Adrenocortical Activity

The degree of pituitary-adrenocortical suppression resulting from exogenous corticosteroid is related to the time of day the steroid is administered. Morning administration has less effect and evening administration a greater effect than do divided doses given over the course of the day.Clinical studies have shown that in the great majority of patients with corticosteroid responsive diseases, an intermittent dosage schedule is at least as effective as is administration of an equal dose in a three or four times a day regimen. Other undesirable side effects of corticosteroid therapy may also be decreased by an intermittent schedule. It is suggested that the customary divided dosage schedule for corticosteroid administration be replaced with an intermittent regimen, the medication being given in the morning. This may be once a day, or, if therapeutic results are satisfactory, once every other day.     

The COX-2-Selective Antagonist (NS-398) Inhibits Choroidal Neovascularization and Subretinal Fibrosis

Choroidal neovascularization (CNV) is an important pathologic component of neovascular age-related macular degeneration (AMD), and CNV lesions later develop into fibrous scars, which contribute to the loss of central vision. Nowadays, the precise molecular and cellular mechanisms underlying CNV and subretinal fibrosis have yet to be fully elucidated. Cyclooxygenase-2 (COX-2) has previously been implicated in angiogenesis and fibrosis. However, the role of COX-2 in the pathogenesis of CNV and subretinal fibrosis is poorly understood. The present study reveals several important findings concerning the relationship of COX-2 signaling with CNV and subretinal fibrosis. Experimental CNV lesions were attenuated by the administration of NS-398, a COX-2-selective antagonist. NS-398-induced CNV suppression was found to be mediated by the attenuation of macrophage infiltration and down-regulation of VEGF in the retinal pigment epithelium–choroid complex. Additionally, NS-398 attenuated subretinal fibrosis, in an experimental model of subretinal scarring observed in neovascular AMD, by down-regulation of TGF-β₂ in the retinal pigment epithelium–choroid complex. Moreover, we cultured mouse RPE cells and found that NS-398 decreased the secretion of VEGF and TGF-β₂ in mouse RPE cells. The results of the present study provide new findings regarding the molecular basis of CNV and subretinal fibrosis, and provide a proof-of-concept approach for the efficacy of COX-2 inhibition in treating subretinal fibrosis.     

Interleukin-2

Between October 1987 and October 1989 we have treated 110 patients with advanced solid tumors with recombinant interleukin-2 (rIL2) based immunotherapy. In renal cell cancer we have studied rIL2 alone, rIL2 combined with rIL2 activated lymphocytes (LAK), and in an ongoing study rIL2 and LAK and alpha-interferon (alpha IFN). There is suggestive evidence of increasingly good results in these consecutive studies. In melanoma the combination of rIL2 and chemotherapy was investigated, followed by an ongoing study of rIL2 and alpha IFN. In these studies rIL2 has been administered as a continuous intravenous infusion of 18 x 10(6) International Units/m2/day for 5 days (18 x 10(6) IU = 3 x 10(6) Cetus Units = 6.9 Biological Response Modifiers Program (BRMP) Units). Patients with non-small cell lung cancer are entered in a phase I-II study of rIL2 and alpha IFN. The rIL2 administration differs from the above mentioned schedule in that rIL2 is given at a maximum dose of 6 x 10(6) IU/m2/day for 28 days on an outpatient basis. In a phase I study we have searched for the maximum tolerated dose of a daily time 4 schedule of rIL2. In the second part of this study a daily time 4 schedule, every week for 4 weeks is being investigated. Finally, we are investigating the safety and efficacy of local regional administration of rIL2 in patients with head and neck cancer, mesothelioma, and liver metastasis of colorectal cancer.     

Selective inhibition of COX-2 improves early survival in murine endotoxemia but not in bacterial peritonitis

Prostaglandins of the E series are believed to act as important mediators of several pathophysiological events that occur in sepsis. Studies were performed to evaluate the effect of cyclooxygenase (COX)-2-specific inhibition on the outcome in murine endotoxemia and cecal ligation and puncture (CLP). We observed a significant time-dependent upregulation of PGE(2) production in both blood and lung homogenates of mice administered lipopolysaccharide intraperitoneally, which was nearly completely suppressed by the administration of the COX-2 inhibitor NS-398. Treatment with NS-398 significantly improved early but not late survival in lipopolysaccharide-challenged mice. On the contrary, elevated PGE(2) levels were found in bronchoalveolar lavage fluid but not in plasma of mice subjected to CLP (21 gauge). Pretreatment with NS-398 failed to significantly improve survival in CLP mice. No significant differences were noted in plasma or lung homogenate proinflammatory cytokine levels or lung neutrophil sequestration between the NS-398-treated and control groups. These results demonstrate that selective COX-2 inhibition confers early but not long-term benefits without affecting the expression of proinflammatory cytokines or the development of lung inflammation.     

Heat shock protein modulation of KATP and KCa channel cerebrovasodilation after brain injury

Fluid percussion brain injury (FPI) impairs pial artery dilation to activators of the ATP-sensitive (K(ATP)) and calcium-activated (K(Ca)) K(+) channels. This study investigated the role of heat shock protein (HSP) in the modulation of K(+) channel-induced pial artery dilation after FPI in newborn pigs equipped with a closed cranial window. Under nonbrain injury conditions, topical coadministration of exogenous HSP-27 (1 mug/ml) blunted dilation to cromakalim, CGRP, and NS-1619 (10(-8) and 10(-6) M; cromakalim and CGRP are K(ATP) agonists and NS-1619 is a K(Ca) agonist). In contrast, coadministration of exogenous HSP-70 (1 mug/ml) potentiated dilation to cromakalim, CGRP, and NS-1619. FPI increased the cerebrospinal fluid (CSF) concentration of HSP-27 from 0.051 +/- 0.012 to 0.113 +/- 0.035 ng/ml but decreased the CSF concentration of HSP-70 from 50.42 +/- 8.96 to 30.9 +/- 9.9 ng/ml at 1 h postinsult. Pretreatment with topical exogenous HSP-70 (1 mug/ml) before FPI fully blocked injury-induced impairment of cromakalim and CGRP dilation and partially blocked injury-induced impairment of dilation to NS-1619. These data indicate that HSP-27 and HSP-70 contribute to modulation of K(+) channel-induced pial artery dilation. These data suggest that HSP-70 is an endogenous protectant of which its actions may be unmasked and/or potentiated with exogenous administration before brain injury.     

Mitochondrial BKCa channels contribute to protection of cardiomyocytes isolated from chronically hypoxic rats

Chronic hypoxia protects the heart against injury caused by acute oxygen deprivation, but its salutary mechanism is poorly understood. The aim was to find out whether cardiomyocytes isolated from chronically hypoxic hearts retain the improved resistance to injury and whether the mitochondrial large-conductance Ca2+-activated K+ (BKCa) channels contribute to the protective effect. Adult male rats were adapted to continuous normobaric hypoxia (inspired O2 fraction 0.10) for 3 wk or kept at room air (normoxic controls). Myocytes, isolated separately from the left ventricle (LVM), septum (SEPM), and right ventricle, were exposed to 25-min metabolic inhibition with sodium cyanide, followed by 30-min reenergization (MI/R). Some LVM were treated with either 30 μM NS-1619 (BKCa opener), or 2 μM paxilline (BKCa blocker), starting 25 min before metabolic inhibition. Cell injury was detected by Trypan blue exclusion and lactate dehydrogenase (LDH) release. Chronic hypoxia doubled the number of rod-shaped LVM and SEPM surviving the MI/R insult and reduced LDH release. While NS-1619 protected cells from normoxic rats, it had no additive salutary effect in the hypoxic group. Paxilline attenuated the improved resistance of cells from hypoxic animals without affecting normoxic controls; it also abolished the protective effect of NS-1619 on LDH release in the normoxic group. While chronic hypoxia did not affect protein abundance of the BKCa channel regulatory β1-subunit, it markedly decreased its glycosylation level. It is concluded that ventricular myocytes isolated from chronically hypoxic rats retain the improved resistance against injury caused by MI/R. Activation of the mitochondrial BKCa channel likely contributes to this protective effect.     

MAPK (ERK2) kinase--a key target for NSAIDs-induced inhibition of gastric cancer cell proliferation and growth.

Limited clinical and experimental studies indicate that nonsteroidal anti-inflammatory drugs (NSAIDs) may inhibit gastric cancer growth. However, the mechanisms involved are not completely understood and cannot be explained by COX-2 inhibition alone. MAPK signaling pathway is essential for cell proliferation, but the effect of NSAIDs on MAPK activity and phosphorylation in gastric cancer has never been studied. Since increased and unregulated cell proliferation and reduced cell apoptosis are important features of cancer growth, we studied whether NS-398, a selective COX-2 inhibitor and/ or indomethacin (IND), a non-selective NSAID: 1) inhibit gastric cancer cell proliferation, 2) whether this inhibition is mediated via MAPK (ERK2), and 3) whether NSAIDs enhance apoptosis in gastric cancer cells. Human gastric epithelial cells (MKN28) derived from gastric tubular adenocarcinoma were cultured and treated with either vehicle, IND (0.25-0.5mM) or NS-398 (50-100 microM) for 6, 16, 24 and 48h. Studies: 1) Cellular proliferation was determined by 3H-thymidine uptake. 2) MAPK activity was measured by incorporation of radiolabeled phosphate into myelin basic protein. 3) Apoptosis was evaluated using TUNEL assay. IND and NS-398 significantly inhibited the proliferation of MKN28 cells at 24h by 3.5 - 5 fold (p<0.002) and at 48h by 2.5 - 10 fold (p<0.02). Both NSAIDs also significantly inhibited ERK2 activity: IND >53% inhibition, NS-398, 100 microM >72% inhibition; all p<0.05. Both IND and NS-398 significantly increased apoptotic index. In conclusion, IND and NS-398 significantly inhibit proliferation and growth of human gastric cancer cell line MKN28. This effect is mediated by NSAID-induced inhibition of MAPK (ERK2) kinase signaling pathway, essential for cell proliferation. NSAIDs also increase apoptosis in MKN28 cells. In addition to inhibiting cyclooxygenase, NSAIDs inhibit phosphorylating enzymes--kinases essential for signaling cell proliferation.     

Characterization and in vitro expression of non-structural 1 protein of canine parvovirus (CPV-2) in mammalian cell line

Parvoviruses are small, 260-A-diameter, icosahedral, non-enveloped, single-stranded DNA viruses with a genome of approximately 5 kb. Non structural protein, (NS-1) is especially relevant, being both essential for virus replication and the main factor responsible for virus pathogenicity and cytotoxicity. This protein has also been reported to possess the property of killing of transformed cells. The present study was carried out to clone, characterize and express the NS-1 gene of canine parvovirus. NS-1 complete CDS 2020bp was amplified, cloned into eukaryotic expression vector pcDNA 3.1(+), sequenced and characterized by in vitro expression analysis. Functional activity of recombinant construct, pcDNA.cpv.NS-1, was evaluated by RT-PCR and flow cytometry for the expression of NS-1 specific mRNA and NS-1 protein, respectively, in transfected HeLa cells. This recombinant plasmid may serve as an important tool to evaluate the apoptotic potential of NS-1 protein of canine parvovirus in cultured HeLa cells.     

A role for cyclooxygenase-2 in lipopolysaccharide-induced anorexia in rats

Because nonselective cycloooxygenase (COX) inhibition attenuated anorexia after lipopolysaccharide (LPS) administration, we tested the ability of resveratrol (2.5, 10, and 40 mg/kg) and NS-398 (2.5, 10, and 40 mg/kg), selective inhibitors of the two COX isoforms COX-1 and -2, respectively, to attenuate LPS (100 microg/kg ip)-induced anorexia. NS-398 (10 and 40 mg/kg) administered with LPS at lights out attenuated LPS-induced anorexia, whereas resveratrol at all doses tested did not. Because prostaglandin (PG) E(2) is considered the major metabolite synthesized by COX, we measured plasma and cerebrospinal fluid (CSF) PGE(2) levels after LPS administration. LPS induced a time-dependent increase of PGE(2) in CSF but not in plasma. NS-398 (5, 10, and 40 mg/kg) blocked the LPS-induced increase in CSF PGE(2), whereas resveratrol (10 mg/kg) did not. These results support a role of COX-2 in mediating the anorectic response to peripheral LPS and point at PGE(2) as a potential neuromodulator involved in this response.     

Myeloma with multiple rearranged immunoglobulin kappa genes: only one kappa gene codes for kappa chains

In many myelomas more than one kappa gene is rearranged (2-5). We are reporting here the results of studies undertaken to determine whether all the rearranged genes are expressed. It was found that in the myeloma NS-1 three different rearranged kappa genes exist. In a subline of NS-1 and several hybridomas produced by fusion of mouse spleen cells with NS-1 it was found that production of NS-1 kappa chains was correlated with the presence of one of the three kappa genes. Loss of this "expressed" gene eliminated the synthesis of the NS-1 kappa chains, loss of one of the other two rearranged kappa genes did not. It is hypothesized, that allelic exclusion (20) of kappa genes generally operates by the functional rearrangement of one kappa gene; other rearrangements are relatively frequent, at least in myelomas, but mostly they are nonfunctional and thus scrambled antibody molecules do not arise.     

NS-3 protein of the Junonia coenia densovirus is essential for viral DNA replication in an Ld 652 cell line and Spodoptera littoralis larvae

The genome of Junonia coenia densovirus (JcDNV) shares with members of the genus Densovirus the property of possessing structural (VP) and nonstructural (NS) genes in opposite orientations. The three NS genes located in the 5' half on one strand encode three NS proteins assumed to be involved in viral DNA replication: NS-1 and NS-2, which are common to all DNVs, and a 28-kDa polypeptide, NS-3, with a unique sequence. Whereas the essential role played by JcDNV NS-1 in viral DNA replication has been clearly established (C. Ding, M. Urabe, M. Bergoin, and R. M. Kotin, J. Virol. 76:338-345, 2002), nothing is known of the biological function(s) of NS-3. To investigate this function, we designed constructs derived from pBRJ, a plasmid encompassing an infectious sequence of JcDNV DNA (M. Jourdan, F. X. Jousset, M. Gervais, S. Skory, M. Bergoin, and B. Dumas, Virology 179:403-409, 1990), with partial or complete deletion of NS-3 sequence or with the ATG initiation codon mutated by site-directed mutagenesis. Transfection of these constructs to sensitive Ld 652 cells or Spodoptera littoralis larvae prevented the accomplishment of a productive cycle. We clearly established that the blocking of the replicative cycle in the absence of NS-3 expression occurred at the level of viral DNA replication. Replication of viral DNA could be restored by cotransfecting Ld 652 cells with a plasmid expressing JcDNV-NS-3 protein in trans. Time course analysis showed that NS-3 is produced early (6 h posttransfection) in the replicative cycle, and its production parallels that of replicative-form viral DNA. Finally, we present evidence that NS-1 and NS-2 proteins are synthesized at apparently the same levels whether or not NS-3 is expressed.