Proc. Nat/. Acad. Sci. USA
`Vol. 93, pp. 14014-14019, November 1996
`Medical Sciences
`
`A novel mechanism of action of tetracyclines: Effects on
`nitric oxide synthases
`ASHOKR. AMIN*t:l=§~, MUKUNDAN G. ATTUR*, GEETAD. THAKKER*, PRAKASH D. PATEL*, PRANAVR. VYAS*,
`RAJESH N. PATEL*, lNDRA VADAN R. PATEL*, AND STEVEN B. ABRAMSON*t
`*Department of Rheumatology, Hospital for Joint Diseases, New York, NY 10003; and Departments of tMedicine and *Pathology, and §Kaplan Cancer Research
`Center, New York University Medical Center, New York, NY 10016
`
`Communicated by H. Sherwood Lawrence, New York University Medical Center, New York, NY, September 23, 1996 (received for review April 10,
`1996)
`
`has been in arthritis, based on studies carried out in animal
`models (5, 6), human osteoarthritis (OA) (7), and RA (8).
`We have recently observed that human OA-affected carti-
`lage can spontaneously release NO under ex vivo conditions in
`quantities sufficient to cause cartilage damage (7). The human
`OA-affected NOS (OA-NOS) is overexpressed in OA-affected
`cartilage and not detectable in normal cartilage. The inducible
`OA-NOS has properties similar to neuronal NOS (based on its
`molecular weight and antibody cross-reactivity among anti-
`NOS antibodies) and the 133-kDa inducible NOS [iN OS;
`sensitive to NF-KB and cycloheximide, up-regulated by inter-
`leukin (IL) 1{3, tumor necrosis factor a, and lipopolysaccharide
`(LPS)]. NO is known to potentiate matrix degradation, which
`includes inhibition of proteoglycan and collagen type II syn-
`thesis (9, II) and up-regulation of metalloprotease activity (10).
`Doxycycline and minocycline are members of the tetracy-
`cline family of broad-spectrum antibiotics. During recent
`years, it has been established that tetracyclines, which are
`rapidly absorbed and have a prolonged half-life, exert biolog-
`ical effects independent of their antimicrobial activity (11-13).
`Such effects include inhibition of matrix metalloproteases
`(MMPs) [including collagenase (MMP-1), gelatinase (MMP-
`2), and stromelysin (MMP-3) activity] and prevention of
`pathogenic tissue destruction (11). Furthermore, recent stud-
`ies have also suggested that tetracyclines and inhibitors of
`metalloproteases inhibit tumor progression (14), bone resorp-
`tion (15), and angiogenesis (16) and may have antiinflamma-
`tory properties (17).
`In RA, these matrix metalloproteases have been identified
`in the synovial tissue, synovial fluids, and the proliferative
`pannus (18). These metalloproteases are also known to be
`up-regulated in OA-affected joints (19, 20). Interestingly, Yu
`et al. (21) have also shown that prophylactic administration of
`doxycycline markedly reduced the severity of OA in dog
`models. In humans, minocycline (a semisynthetic tetracycline)
`has recently been demonstrated to be superior to placebo in
`the treatment of RA (22).
`Since NO is a putative mediator of inflammation that exerts
`catabolic effects on cartilage, including the activation of me-
`talloproteases (10), we evaluated the action of tetracycline
`compounds on the spontaneous release of NO from OA-
`
`Tetracyclines have recently been shown to
`ABSTRACT
`have "chondroprotective" effects in inflammatory arthritides
`in animal models. Since nitric oxide (NO) is spontaneously
`released from human cartilage affected by osteoarthritis (OA)
`or rheumatoid arthritis in quantities sufficient to cause
`cartilage damage, we evaluated the effect of tetracyclines on
`the expression and function of human OA-affected nitric oxide
`synthase (OA-NOS) and rodent inducible NOS (iNOS).
`Among the tetracycline group of compounds, doxycycline >
`minocycline blocked and reversed both spontaneous and
`interleukin lf3-induced OA-NOS activity in ex vivo conditions.
`Similarly, minocycline ~ doxycycline inhibited both lipopoly-
`saccharide- and interferon-y-stimulated iN OS in RAW 264.7
`cells in vitro, as assessed by nitrite accumulation. Although
`both these enzyme isoforms could be inhibited by doxycycline
`and minocycline, their susceptibility to each ofthese drugs was
`distinct. Unlike acetylating agents or competitive inhibitors of
`L-arginine that directly inhibit the specific activity of NOS,
`doxycycline or minocycline has no significant effect on the
`specific activity of iN OS in cell-free extracts. The mechanism
`of action of these drugs on murine iN OS expression was found
`to be, at least in part, at the level of RNA expression and
`translation of the enzyme, which would account for the
`decreased iN OS protein and activity of the enzyme. Tetracy-
`clines had no significant effect on the levels of mRNA for
`/3-actin and glyceraldehyde-3-phosphate dehydrogenase nor
`on levels of protein of /3-actin and cyclooxygenase 2 expres-
`sion. These studies indicate that a novel mechanism of action
`of tetracyclines is to inhibit the expression of NOS. Since the
`overproduction of NO has been implicated in the pathogenesis
`of arthritis, as well as other inflammatory diseases, these
`observations suggest that tetracyclines should be evaluated as
`potential therapeutic modulators of NO for various patholog-
`ical conditions.
`
`Nitric oxide (NO), a multifunctional mediator produced by
`and acting on various cells, participates in inflammatory and
`autoimmune-mediated tissue destruction. NO is produced by
`a family of ubiquitous enzymes, nitric oxide synthases (NOSs).
`The overexpression of NOS in a variety of inflammatory
`tissues has led many to conclude that the modulation of NO
`synthesis and action could represent a new approach to the
`treatment of inflammatory and autoimmune conditions (1, 2).
`Where examined, NO formation is found to be increased in
`autoimmune diseases [rheumatoid arthritis (RA), systemic
`lupus erythematosus, ulcerative colitis, and Crohn disease],
`and several classic inflammatory symptoms (erythema and
`vascular leakiness) are reversed by NOS inhibitors (2-4). The
`most compelling evidence for NO as a mediator of tissue injury
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement" in
`accordance with 18 U.S. C. §1734 solely to indicate this fact.
`
`Abbreviations: NOS, nitric oxide synthase; iNOS, inducible NOS;
`L-NMMA, L-N-monomethyl arginine; OA, osteoarthritis; RA, rheu-
`matoid arthritis; IL-1/3, interleukin lf3; LPS, lipopolysaccharide;
`GAPDH, glyceraldehyde-3-phosphate dehydrogenase; COX-2, cyclo-
`oxygenase 2; MMP, matrix metalloprotease; TGF-/3, transforming
`growth factor f3; RT-PCR, reverse transcription-coupled PCR; IFN-y,
`interferon y.
`lfTo whom reprint requests should be addressed at: Department of
`Rheumatology, Room 1600, Hospital for Joint Diseases, 301 East
`17th Street, New York, NY 10003.
`IICao, M., Westerhausen-Larson, A., l Dr. Reddy's Laboratories, Ltd., et al.
`Georgescu, H. I., Rizzo, C. F., Stefano
`v.
`(1996) 42nd Annual Meeting of the Or
`Galderma Laboratories, Inc.
`533 (abstr.).
`IPR2015-_ _
`Exhibit 1024
`
`14014
`
`

`
`Proc. Natl. Acad. Sci. USA
`Vol. 93, pp. 14014–14019, November 1996
`Medical Sciences
`
`A novel mechanism of action of tetracyclines: Effects on
`nitric oxide synthases
`ASHOK R. AMIN*†‡§¶, MUKUNDAN G. ATTUR*, GEETA D. THAKKER*, PRAKASH D. PATEL*, PRANAV R. VYAS*,
`RAJESH N. PATEL*, INDRAVADAN R. PATEL*, AND STEVEN B. ABRAMSON*†
`*Department of Rheumatology, Hospital for Joint Diseases, New York, NY 10003; and Departments of †Medicine and ‡Pathology, and §Kaplan Cancer Research
`Center, New York University Medical Center, New York, NY 10016
`
`Communicated by H. Sherwood Lawrence, New York University Medical Center, New York, NY, September 23, 1996 (received for review April 10,
`1996)
`
`Tetracyclines have recently been shown to
`ABSTRACT
`have ‘‘chondroprotective’’ effects in inflammatory arthritides
`in animal models. Since nitric oxide (NO) is spontaneously
`released from human cartilage affected by osteoarthritis (OA)
`or rheumatoid arthritis in quantities sufficient to cause
`cartilage damage, we evaluated the effect of tetracyclines on
`the expression and function of human OA-affected nitric oxide
`synthase (OA-NOS) and rodent inducible NOS (iNOS).
`Among the tetracycline group of compounds, doxycycline >
`minocycline blocked and reversed both spontaneous and
`interleukin 1␤-induced OA-NOS activity in ex vivo conditions.
`Similarly, minocycline > doxycycline inhibited both lipopoly-
`saccharide- and interferon-␥-stimulated iNOS in RAW 264.7
`cells in vitro, as assessed by nitrite accumulation. Although
`both these enzyme isoforms could be inhibited by doxycycline
`and minocycline, their susceptibility to each of these drugs was
`distinct. Unlike acetylating agents or competitive inhibitors of
`L-arginine that directly inhibit the specific activity of NOS,
`doxycycline or minocycline has no significant effect on the
`specific activity of iNOS in cell-free extracts. The mechanism
`of action of these drugs on murine iNOS expression was found
`to be, at least in part, at the level of RNA expression and
`translation of the enzyme, which would account for the
`decreased iNOS protein and activity of the enzyme. Tetracy-
`clines had no significant effect on the levels of mRNA for
`␤-actin and glyceraldehyde-3-phosphate dehydrogenase nor
`on levels of protein of ␤-actin and cyclooxygenase 2 expres-
`sion. These studies indicate that a novel mechanism of action
`of tetracyclines is to inhibit the expression of NOS. Since the
`overproduction of NO has been implicated in the pathogenesis
`of arthritis, as well as other inflammatory diseases, these
`observations suggest that tetracyclines should be evaluated as
`potential therapeutic modulators of NO for various patholog-
`ical conditions.
`
`Nitric oxide (NO), a multifunctional mediator produced by
`and acting on various cells, participates in inflammatory and
`autoimmune-mediated tissue destruction. NO is produced by
`a family of ubiquitous enzymes, nitric oxide synthases (NOSs).
`The overexpression of NOS in a variety of inflammatory
`tissues has led many to conclude that the modulation of NO
`synthesis and action could represent a new approach to the
`treatment of inflammatory and autoimmune conditions (1, 2).
`Where examined, NO formation is found to be increased in
`autoimmune diseases [rheumatoid arthritis (RA), systemic
`lupus erythematosus, ulcerative colitis, and Crohn disease],
`and several classic inflammatory symptoms (erythema and
`vascular leakiness) are reversed by NOS inhibitors (2–4). The
`most compelling evidence for NO as a mediator of tissue injury
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked ‘‘advertisement’’ in
`accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`has been in arthritis, based on studies carried out in animal
`models (5, 6), human osteoarthritis (OA) (7), and RA (8).
`We have recently observed that human OA-affected carti-
`lage can spontaneously release NO under ex vivo conditions in
`quantities sufficient to cause cartilage damage (7). The human
`OA-affected NOS (OA-NOS) is overexpressed in OA-affected
`cartilage and not detectable in normal cartilage. The inducible
`OA-NOS has properties similar to neuronal NOS (based on its
`molecular weight and antibody cross-reactivity among anti-
`NOS antibodies) and the 133-kDa inducible NOS [iNOS;
`sensitive to NF-␬B and cycloheximide, up-regulated by inter-
`leukin (IL) 1␤, tumor necrosis factor ␣, and lipopolysaccharide
`(LPS)]. NO is known to potentiate matrix degradation, which
`includes inhibition of proteoglycan and collagen type II syn-
`thesis (9, (cid:219)) and up-regulation of metalloprotease activity (10).
`Doxycycline and minocycline are members of the tetracy-
`cline family of broad-spectrum antibiotics. During recent
`years, it has been established that tetracyclines, which are
`rapidly absorbed and have a prolonged half-life, exert biolog-
`ical effects independent of their antimicrobial activity (11–13).
`Such effects include inhibition of matrix metalloproteases
`(MMPs) [including collagenase (MMP-1), gelatinase (MMP-
`2), and stromelysin (MMP-3) activity] and prevention of
`pathogenic tissue destruction (11). Furthermore, recent stud-
`ies have also suggested that tetracyclines and inhibitors of
`metalloproteases inhibit tumor progression (14), bone resorp-
`tion (15), and angiogenesis (16) and may have antiinflamma-
`tory properties (17).
`In RA, these matrix metalloproteases have been identified
`in the synovial tissue, synovial fluids, and the proliferative
`pannus (18). These metalloproteases are also known to be
`up-regulated in OA-affected joints (19, 20). Interestingly, Yu
`et al. (21) have also shown that prophylactic administration of
`doxycycline markedly reduced the severity of OA in dog
`models. In humans, minocycline (a semisynthetic tetracycline)
`has recently been demonstrated to be superior to placebo in
`the treatment of RA (22).
`Since NO is a putative mediator of inflammation that exerts
`catabolic effects on cartilage, including the activation of me-
`talloproteases (10), we evaluated the action of tetracycline
`compounds on the spontaneous release of NO from OA-
`
`Abbreviations: NOS, nitric oxide synthase; iNOS, inducible NOS;
`L-NMMA, L-N-monomethyl arginine; OA, osteoarthritis; RA, rheu-
`matoid arthritis; IL-1␤,
`interleukin 1␤; LPS,
`lipopolysaccharide;
`GAPDH, glyceraldehyde-3-phosphate dehydrogenase; COX-2, cyclo-
`oxygenase 2; MMP, matrix metalloprotease; TGF-␤, transforming
`growth factor ␤; RT–PCR, reverse transcription-coupled PCR; IFN-␥,
`interferon ␥.
`¶To whom reprint requests should be addressed at: Department of
`Rheumatology, Room 1600, Hospital for Joint Diseases, 301 East
`17th Street, New York, NY 10003.
`储Cao, M., Westerhausen-Larson, A., Niyibizi, C., Kavalkovich, K.,
`Georgescu, H. I., Rizzo, C. F., Stefanovic-Racic, M. & Evans, C. H.
`(1996) 42nd Annual Meeting of the Orthopedic Research Society, p.
`533 (abstr.).
`
`14014
`
`Exh. 1024
`
`

`
`Medical Sciences: Amin et al.
`
`Proc. Natl. Acad. Sci. USA 93 (1996)
`
`14015
`
`affected human cartilage in ex vivo conditions (7) and on iNOS
`in LPS-stimulated murine macrophages. Both these enzyme
`isoforms show distinct susceptibility to pharmacological inter-
`vention by hydrocortisone and transforming growth factor ␤
`(TGF-␤) in vitro (7). In the present study we report that (i)
`doxycycline and minocycline inhibit the activity of murine
`macrophage iNOS (minocycline ⱖ doxycycline) and human
`inducible OA-NOS (doxycycline ⬎ minocycline), (ii) doxycy-
`cline and minocycline inhibit iNOS expression at the level of
`iNOS mRNA and protein expression, thereby down-regulating
`its specific activity, and (iii) unlike acetylating agents or
`competitive inhibitors of L-arginine, doxycycline and minocy-
`cline do not directly inhibit the catalytic activity of iNOS in
`vitro in the L-arginine 3 L-citrulline conversion assay.
`MATERIALS AND METHODS
`Cell Lines and Reagents. Murine macrophage cells (RAW
`264.7) were obtained from the American Type Culture Col-
`lection. Anti-murine iNOS and anti-cyclooxygenase (COX-2)
`antibodies were obtained from Transduction Laboratories
`(Lexington, KY). OA-affected cartilage was obtained from
`OA patients who underwent knee replacement surgery and
`were free of steroidal兾nonsteroidal antiinflammatory drugs
`for at least 2 weeks before surgery. Doxycycline, minocycline,
`hydrocortisone, and LPS were obtained from Sigma, and murine
`interferon (IFN)-␥ and human IL-1␤ were from Promega.
`Assay of OA-NOS in Organ Cultures. This assay was basically
`carried out as described (7). Briefly, OA-affected cartilage was
`cut into 3-mm discs; four to six discs (100–200 mg) were placed
`in organ culture in 2 ml of medium (F-12 with 0.1% BSA) for
`24–72 h in the incubator. The medium was analyzed for nitrite
`accumulation by modified Griess reaction (23).
`Western Blot Analysis. Equal amounts of protein (25–50 ␮g)
`estimated by BCA reagent (Pierce) were loaded onto SDS兾
`PAGE gels and examined by Western blot analysis with a
`specific anti-iNOS or anti-cyclooxygenase 2 (COX-2) murine
`mAb as specified by Transduction Laboratories. Membranes
`with bound antibodies (e.g., iNOS) were stripped by submer-
`sion in stripping buffer (100 mM 2-mercaptoethanol兾2%
`SDS兾62.5 mM Tris䡠HCl, pH 6.7) and incubating at 50⬚C for 30
`min with occasional agitation. Membranes were then washed
`twice for 10 min at room temperature using large volumes of
`wash buffer. This filter could then be reprobed with an
`anti-actin antibody provided by James L. Lessard (Children’s
`Hospital Medical Center, Cincinnati). Blots were developed
`using the ECL Western blot system (Amersham). Quantitation
`of the bands was performed using a densitometer from Mo-
`lecular Dynamics.
`Northern Blot Analysis. Total RNA was isolated using TRI
`reagent (MRC, Cincinnati). Northern blot analysis was carried
`out as described (24, 25). Briefly, 30 ␮g of RNA was subjected
`to electrophoresis and transferred via capillary action onto a
`nylon membrane (Zeta-Probe, Bio-Rad). The membrane was
`hybridized with [32P]dCTP-labeled iNOS cDNA (4-kb SmaI
`fragment), a gift from James Cunningham (Harvard Medical
`School), and the blot was exposed to Kodak x-ray film for
`24–48 h with intensifying screens at ⫺70⬚C. The ␤-actin and
`glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probes
`were purchased from CLONTECH and probed as described
`above. Quantitation was performed using a PhosphorImager
`or Personal Densitometer (SI) (Molecular Dynamics).
`Reverse Transcription-Coupled PCR (RT–PCR) Analysis.
`The presence of iNOS and ␤-actin mRNA in cells was analyzed
`by RT of total RNA followed by PCR amplification of the
`cDNA as described (26, 27). The sense and antisense oligo-
`respectively, 5⬘-ACG-
`nucleotides
`for
`iNOS were,
`GAGAAGCTTAGATCTGGAGCAGAAGTG-3⬘ (nt 142–
`171) and 5⬘-CTGCAGGTTGGACCACTGGATCCTGC-
`CGAT-3⬘ (nt 767–796). The sense and antisense primers for
`␤-actin were 5⬘-TCCTTCGTTGCCGGTCCACA-3⬘ (nt 44–
`
`63) and 5⬘-CGTCTCCGGAGTCCATCACA-3⬘ (nt 534–552),
`respectively. The predicted PCR product of the iNOS cDNA
`was 654 bp; that of the ␤-actin cDNA was 508 bp. The cDNA
`was prepared from equal amounts (1 ␮g) of total RNA using
`SuperScript RNase H reverse transcriptase (GIBCO兾BRL).
`An equal amount of the cDNA was used to amplify the mRNA
`by PCR. PCR amplification was performed in 50 ␮l of a
`solution containing 1.5 mM MgCl2, 500 ng of iNOS primer, 100
`ng of ␤-actin primer, all four dNTPs (each at 0.2 mM), 2.5 units
`of Taq DNA polymerase (GIBCO兾BRL). The cycle conditions
`for amplification of cDNA were 1 min at 94⬚C, 1–2 min at 55⬚C,
`and 3 min at 72⬚C for 30 cycles for both iNOS and ␤-actin.
`Assays for iNOS in Cell-Free Extracts. Specific activity of
`iNOS was determined in cell-free extracts by monitoring the
`conversion of L-[3H]arginine to L-[3H]citrulline as described
`(25, 28). RAW 264.7 cells were induced with LPS (100 ng兾ml)
`in the presence and absence of tetracyclines or hydrocortisone
`for 14–20 h. The cell-free extracts were prepared in Tris buffer
`(10 mM, pH 7.4) containing chymostatin at 10 ␮g兾ml, antipain
`at 10 ␮g兾ml, leupeptin at 10 ␮g兾ml, pepstatin at 10 ␮g兾ml, 1
`mM DTT, and 1 mM phenylmethylsulfonyl fluoride (25). The
`protein was measured by BCA assay reagent using BSA as
`standard (29). The reaction mixture for iNOS assay was as
`described (25, 28). After 20 min, the assays were terminated by
`heating the reaction mixture at 90⬚C for 5 min; 10 ␮l (⬇100,000
`cpm) of the supernatant was spotted on activated Avicel TLC
`plates (Analtech Associates). The TLC plates were developed
`in a solvent system consisting of ethanol兾water兾ammonia,
`80:16:4 (vol兾vol). Quantitation of the spot for L-[3H]citrulline
`was performed by a Bioscan System 200 imaging scanner.
`
`RESULTS AND DISCUSSION
`Effect of Doxycycline and Minocycline on Inducible OA-
`NOS Activity in Human OA-Affected Cartilage. Studies in our
`laboratory have shown (7) that OA-affected cartilage sponta-
`neously releases NO in ex vivo conditions sufficient to cause
`cartilage damage. NO has dual effects on matrix metabolism:
`not only does it potentiate activity of matrix-degrading metal-
`loproteases (10) but also it inhibits synthesis of matrix com-
`ponents such as proteoglycans and collagen (9, (cid:219)) . Therefore,
`we first examined whether doxycycline or minocycline could
`block human OA-NOS activity in ex vivo conditions. Generally
`accepted pharmacologically relevant concentrations were se-
`lected for this study based on previous reports (19, 30–32).
`OA-affected cartilage was obtained from patients with ad-
`vanced OA undergoing knee replacement surgery. OA-
`affected cartilage slices (Fig. 1A) were incubated in 0.1%
`BSA兾endotoxin-free medium with doxycycline or minocycline
`at 5–80 ␮g兾ml for 24, 48, and 72 h in ex vivo conditions. Activity
`of NOS was monitored at different time intervals by estimating
`nitrite, the stable end-product, as described (7). The amount
`of NO spontaneously released (as measured by nitrite accu-
`mulation) at 0, 24, 48, and 72 h was 0, 4.8 ⫾ 0.38, 16.4 ⫾ 0.7,
`and 17.8 ⫾ 0.9 ␮M, respectively. The results showed that
`doxycycline and minocycline significantly inhibited NO pro-
`duction in OA-affected cartilage in a dose-dependent manner
`(Fig. 1). These data also indicate that doxycycline was more
`potent in its ability to inhibit OA-NOS activity. For example,
`at 72 h, the IC50 for doxycycline was 32 ␮g兾ml compared with
`54 ␮g兾ml for minocycline. These experiments further indicate
`that doxycycline and minocycline not only blocked the ongoing
`production of NO by OA-NOS ex vivo but also caused a decline
`in nitrite accumulation in cartilage organ culture for at least
`72 h under conditions in which nitrite continues to accumulate
`in control cultures (Fig. 1). In a separate experiment, OA-
`affected cartilage was also exposed to IL-1␤at 5 ng兾ml, which
`augmented the release of nitrite from 3.6 ⫾ 1.7 to 16.7 ⫾ 2.5
`␮M (P⬍0.0001) at 72 h. Addition of doxycycline at 20 and 40
`␮g兾ml at the time of the IL-1␤ stimulation decreased the
`nitrite levels to 13.9 ⫾ 2.0 ␮M (P⬍0.06) and 8.5 ⫾ 1.3 ␮M
`
`Exh. 1024
`
`

`
`14016 Medical Sciences: Amin et al.
`
`Proc. Natl. Acad. Sci. USA 93 (1996)
`
`incubation in these cells. In the same set of experiments,
`minocycline was also administered at concentrations ranging
`from 5 to 80 ␮g兾ml. The IC50 for minocycline was 17 ␮g兾ml at
`14 h and 12 ␮g兾ml at 20 h of incubation in RAW 264.7 cells
`stimulated with LPS. Although a marginal difference in the
`potency of doxycycline and minocycline was seen at 20 h of
`incubation (based on IC50), significantly higher concentrations
`of doxycycline, as compared with minocycline, were generally
`required to inhibit iNOS by ⬎50% at both time intervals. We
`also examined the effect of minocycline or doxycycline at 20
`␮g兾ml on murine iNOS expression when stimulated with
`IFN-␥at 100 units兾ml for 16 h (which induced 16.4 ⫾ 0.5 ␮M
`nitrite). Addition of minocycline and doxycycline significantly
`decreased IFN-␥-induced nitrite production to 9.7 ⫾ 0.3 ␮M
`(P ⬍ 0.0001) and 13.1 ⫾ 1.1 ␮M (P⬍0.005), respectively.
`These studies indicate that both doxycycline and minocy-
`cline inhibit NO production in murine macrophages stimulated
`with either LPS or IFN-␥. Furthermore, these experiments
`together with our prior observations (7) show that the iNOS
`and inducible OA-NOS have distinct susceptibility to doxycy-
`cline, minocycline, TGF-␤, and hydrocortisone. This is not
`surprising since it has been reported that two different forms
`of collagenase, MMP-8 (IC50, 7–15 ␮g兾ml) and MMP-1 (IC50,
`140 ␮g兾ml), in two different cell types (i.e., neutrophils and
`fibroblasts), show distinct susceptibility to inhibition by tetra-
`cyclines (35). Furthermore, it should be noted that the same
`enzyme expressed in two closely related cell lines can have
`differential susceptibility to tetracyclines. For example, two
`osteoblastic cell lines, UMR 106–01 (IC50, ⬎200 ␮g兾ml) and
`ROS 17兾2.8 (IC50, 20–30 ␮g兾ml), showed differential suscepti-
`bility to doxycycline when evaluated for gelatinase activity (36).
`Finally, another factor that may contribute to the differential IC50
`values of tetracyclines on NOS activity in cartilage slices com-
`pared with macrophage cells is the ability of each drug to
`penetrate the cartilage matrix and act on chondrocytes (37).
`In view of the recent observation that in vivo administration
`of minocycline in rats augments the percentage of splenocytes
`exhibiting a rise in intracellular Ca2⫹ under ex vivo conditions
`upon stimulation with concanavalin A, depending on the
`concentration of Ca2⫹ in the medium, we examined the effect
`of extracellular Ca2⫹ on the influence of iNOS expression
`upon stimulation with LPS (38). RAW 264.7 cells were sup-
`plemented (⫺1 h) with an additional 0, 0.34, and 0.68 mM
`Ca2⫹ above that present in the medium (1.3 mM) and stimu-
`lated with LPS. The nitrite levels at 24 h were 24.5 ⫾ 4.9, 26.6 ⫾
`1.8, and 27.3 ⫾ 0.3 ␮M in medium supplemented with 0, 0.34
`mM, and 0.68 mM Ca2⫹, respectively. Addition of LPS and
`minocycline at 40 ␮g兾ml at the time of the addition of Ca2⫹
`(⫺1 h) reduced the nitrite levels to 8.8 ⫾ 1.04 (P ⬍ 0.0029),
`12.7 ⫾ 0.7 (P ⬍ 0.007), and 12.1 ⫾ 0.9 (P ⬍ 0.006) ␮M, for the
`0, 0.34 mM, and 0.68 mM Ca2⫹-supplemented cultures, re-
`spectively. Thus, there was no significant difference between
`the Ca2⫹-supplemented and nonsupplemented controls, indi-
`cating that supplemental Ca2⫹ does not have a significant
`influence on the tetracycline-dependent NOS inhibition in the
`murine macrophages tested in vitro. It should be noted that
`tetracyclines also inhibit IL-1␤- and IFN-␥-induced NOS
`expression; unlike LPS, these cytokines do not flux extracel-
`lular Ca2⫹ into the cells upon activation (39, 40). These
`experiments indicate that tetracyclines do not exert their
`effects on NOS via the chelation of extracellular Ca2⫹. How-
`ever, it should be noted that intracellular Ca2⫹ is critical for the
`enzyme activity of NOS, which usually appears 3–4 h after LPS
`stimulation (3).
`Based on the above studies, we sought to evaluate the
`mechanism of action of tetracyclines on NOS expression in the
`murine macrophage model where the biochemistry, enzymology,
`and molecular biology of iNOS are well-characterized (3, 34, 41,
`42). We did not perform similar studies of OA-NOS since it was
`not possible to precisely and reproducibly quantitate the expres-
`
`(A) Effect of doxycycline and minocycline on OA-NOS
`FIG. 1.
`expression in human OA-affected cartilage. OA-affected knee artic-
`ular cartilage from one OA-affected patient was placed in organ
`culture in 2 ml of medium in the presence or absence (control) of
`doxycycline and minocycline at 5–80 ␮g兾ml. The spontaneous release
`of nitrite was monitored at different time intervals. Data are expressed
`as micromolar nitrite released (mean ⫾ SD, n ⫽ 3–4). Zero time
`indicates spontaneous release of NO at 24, 48, and 72 h, which was
`4.8 ⫾ 0.38, 16.4 ⫾ 0.7, and 17.8 ⫾ 0.9, respectively. Statistics were
`derived using unpaired Student’s t test. Data represent one of three
`identical experiments with samples from different patients. (B) Effect
`of doxycycline and minocycline on nitrite accumulation in RAW 264.7
`cells stimulated with LPS. Murine macrophage cells (RAW 264.7)
`were incubated with doxycycline or minocycline (5–80 ␮g兾ml) for 1–2
`h followed by addition of LPS at 100 ng兾ml. After 14–20 h of
`incubation, the medium was used to estimate nitrite accumulation by
`the modified Griess reaction (23). Data are expressed as micromolar
`of nitrite accumulated at a given time interval (n ⫽ 3). Statistics were
`derived using unpaired Student’s t test. Data represent one of three
`similar experiments. Graphs were plotted using the DELTA graph
`curve-fitting program: polynomial of degree 3.
`
`(P ⬍ 0.0006), respectively, when compared with the IL-1 ␤-
`stimulated cartilage slices at the same time interval. Similarly,
`addition of minocycline (at 20 and 40 ␮g兾ml) showed 16.6 ⫾
`3.2 (P ⬍ 0.47) and 10.0 ⫾ 2.0 (P ⬍ 0.0024) ␮M of nitrite
`accumulation when compared with the IL-1 ␤-stimulated car-
`tilage slices. These experiments also demonstrated that the
`IC50 levels for tetracyclines to inhibit NOS activity augmented
`by IL-1␤ were similar to the spontaneous release of NO. The
`concentrations of doxycycline that inhibited NO production in
`our studies are comparable to those required for the inhibition
`of matrix metalloproteases (19, 30–33). In those studies,
`doxycycline at 20–50 ␮g兾ml inhibited the activity of proteolytic
`enzymes such as collagenase and gelatinase, blocked proteo-
`glycan degradation, reduced the cell death associated with
`proteoglycan loss, and augmented cartilage growth (31).
`Effects of Doxycycline and Minocycline on iNOS in Murine
`Macrophages. Our recent studies have indicated that human
`inducible OA-NOS is distinct from murine and human iNOS,
`based upon its size, immunoreactivity, and susceptibility to
`TGF-␤ and hydrocortisone (7). Therefore, we also evaluated
`the effect of tetracyclines on the expression of iNOS in murine
`macrophages. RAW 264.7 cells were activated with LPS (100
`ng兾ml) to induce iNOS (34) with and without doxycycline and
`minocycline at 5–80 ␮g兾ml. Fig. 1B shows a concentration-
`dependent inhibition of nitrite accumulation in cells stimu-
`lated with LPS in the presence of doxycycline at 5–80 ␮g兾ml
`at 14 and 20 h of incubation. The IC 50 of doxycycline in this
`experiment was 72 ␮g兾ml at 14 h and 22 ␮g兾ml at 20 h of
`
`Exh. 1024
`
`

`
`Medical Sciences: Amin et al.
`
`sion of OA-NOS directly from the OA-affected cartilage without
`disturbing the architecture of the cartilage, which plays a signif-
`icant role in chondrocyte function.
`Using LPS-induced RAW 264.7 cells, we examined the
`following hypotheses. Tetracyclines may (i) decrease only the
`catalytic activity of iNOS without influencing the expression of
`iNOS protein; (ii) decrease both the catalytic activity of iNOS
`and the expression of iNOS protein, which in turn cumulatively
`leads to decrease in the accumulation of nitrite in the medium;
`or (iii) decrease the expression of iNOS protein and, therefore,
`the production of nitrite.
`Effect of Doxycycline and Minocycline on the Enzyme
`Activity of iNOS in Whole Cells. RAW 264.7 cells were
`exposed to doxycycline or minocycline in the presence of LPS
`for 16–18 h; cell-free extracts prepared at the end of each time
`period were evaluated for iNOS enzyme activity using the
`L-arginine 3 L-citrulline conversion assay in total cell extracts.
`As shown in Fig. 2, preexposure of cells to either doxycycline
`or minocycline inhibits the conversion of arginine to citrulline
`in cell lysates in a dose-dependent manner when compared
`with the control LPS-stimulated activity. Doxycycline at 20, 40,
`and 80 ␮g兾ml significantly reduced citrulline accumulation by
`57%, 72%, and 85%, respectively; similar inhibition was also
`observed for minocycline (45%, 69%, and 69%, respectively).
`As expected, pretreatment of cells with 10 ␮M hydrocortisone
`and 75 ␮M L-N-monomethyl arginine (L-NMMA) blocked
`iNOS activity by 60% and 64%, respectively.
`Direct Effect of Doxycycline and Minocycline on iNOS
`Enzyme Activity in Cell-Free Extracts. Recent studies have
`indicated that tetracyclines inhibit collagenase activity via
`direct effects on the enzyme (30, 43), which could be partially
`reversed by addition of Ca2⫹ to the reaction mixture (44).
`Another mechanism proposed for this phenomenon is that
`procollagenase is reduced to inactive fragments upon activa-
`tion in the presence of doxycycline (43). We have recently
`shown that acetylating agents, such as aspirin and N-
`acetylimidazole (25), as well as competitive inhibitors of
`L-arginine (2–4), inhibit iNOS catalytic activity in vitro. To
`evaluate the direct effect of doxycycline and minocycline on
`iNOS enzyme activity, we induced RAW 264.7 cells with LPS
`for 16 h in the absence of these agents and prepared cell-free
`
`14017
`Proc. Natl. Acad. Sci. USA 93 (1996)
`extracts as a source of iNOS enzyme in L-arginine 3 L-
`citrulline conversion assay. Separate aliquots of equal amounts
`of enzyme (in total cell extracts) were preincubated for 15 min
`with doxycycline at 20–80 ␮g兾ml, minocycline at 20–80 ␮g兾ml,
`1 mM N-acetylimidazole, and 200 ␮M L-NMMA, respectively,
`before the enzyme reaction was initiated by adding the cofac-
`tors. The experiment showed that, unlike N-acetylimidazole or
`L-NMMA, doxycycline兾minocycline had no significant effect
`directly on the units of enzyme activity or specific activity of
`iNOS in cell-free extracts (Fig. 3). These experiments indicate
`that the action of these drugs on iNOS seems to be distinct
`from that reported for metalloproteases such as procollag-
`enases (30, 43, 45). Minocycline and doxycycline could not
`block an ongoing L-arginine 3 L-citrulline reaction catalyzed
`by iNOS in cell-fr