SEARCH

SEARCH BY CITATION

Keywords:

  • autoimmunity;
  • diabetes;
  • inflammation;
  • monocytes/macrophage;
  • NOD mice

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Monocytes infiltrate islets in non-obese diabetic (NOD) mice. Activated monocyte/macrophages express cyclo-oxygenase-2 (COX-2) promoting prostaglandin-E2 (PGE2) secretion, while COX-1 expression is constitutive. We investigated in female NOD mice: (i) natural history of monocyte COX expression basally and following lipopolysaccharide (LPS) stimulation; (ii) impact of COX-2 specific inhibitor (Vioxx) on PGE2, insulitis and diabetes. CD11b+ monocytes were analysed for COX mRNA expression from NOD (n = 48) and C57BL/6 control (n = 18) mice. NOD mice were treated with either Vioxx (total dose 80mg/kg) (n = 29) or methylcellulose as control (n = 29) administered by gavage at 4 weeks until diabetes developed or age 30 weeks. In all groups, basal monocyte COX mRNA and PGE2 secretion were normal, while following LPS, after 5 weeks of age monocyte/macrophage COX-1 mRNA decreased (P < 0·01) and COX-2 mRNA increased (P < 0·01). However, diabetic NOD mice had reduced COX mRNA response (P = 0·03). Vioxx administration influenced neither PGE2, insulitis nor diabetes. We demonstrate an isoform switch in monocyte/macrophage COX mRNA expression following LPS, which is altered in diabetic NOD mice as in human diabetes. However, Vioxx failed to affect insulitis or diabetes. We conclude that monocyte responses are altered in diabetic NOD mice but COX-2 expression is unlikely to be critical to disease risk.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Type 1 diabetes mellitus (T1DM) is due to destruction of the insulin-secreting cells in the pancreatic islets of Langerhans, caused probably by destructive innate and adaptive immune responses, as monocytes, macrophages and T lymphocytes infiltrate the islets at clinical onset of the disease [1]. Both adaptive and innate immune responses are also involved in the non-obese diabetic (NOD) mouse, an animal model of autoimmune diabetes with immunopathological features resembling those of the human disease [2–8]. Activation of monocytes and macrophages induce translocation of the transcription factor nuclear factor kappa B, leading to expression of immune response genes with secretion of prostaglandins (PG), nitric oxide, cytokines and chemokines, which can profoundly affect cellular immune responses [9,10]. Pathways contributing to PG output in antigen-presenting cells, such as macrophages and monocytes, are regulated by the enzyme cyclo-oxygenase (COX, PG G/H synthase).

Cyclo-oxygenase is a key enzyme in PG synthesis and catalyses a rate-limiting step in the conversion of arachidonic acid to PG H [which has at least two isoforms (COX-1 and COX-2), encoded by separate genes][11,12]. COX-1 is believed to be expressed constitutively as a housekeeping gene. In contrast, COX-2 is an inducible enzyme responding to various factors including lipopolysaccharide (LPS) with an increase in proinflammatory PGs (e.g. PGE2) [13–15]. Ingestion of fish oils that contain omega-3 (a polyunsaturated fatty acid) results in a decrease in membrane arachidonic acid levels, and a concomitant decrease in the capacity to synthesize proinflammatory PGs from arachidonic acid [11]. PG metabolism is altered in female NOD mice [16], due possibly to aberrant macrophage COX-2 expression. We and others have reported in man aberrant expression of COX-2 in monocytes in T1DM patients as well as in their non-diabetic identical twins and high-risk siblings [5].

We wondered whether aberrant monocyte COX-2 mRNA expression might occur in NOD mice and, if so, whether diabetes development in NOD mice can be altered by COX-2 specific inhibitors. We therefore studied the natural history of monocyte COX-2 mRNA expression in NOD mice and in a separate cohort used the COX-2 inhibitor Vioxx (Vioxx™, Merck and Co., Inc., Bedford Cross, Feltham, London, UK) to determine whether it affects the rate of insulitis or diabetes. The mechanism of action of Vioxx is similar to that reported for other selective COX-2 inhibitors and is believed to be consistent with the inhibition of COX-2 in a two-step-dependent mechanism, involving the formation of a tightly bound inhibited complex leading to the inhibition of LPS-induced COX-2 derived PGE2 synthesis [17]. We investigated the impact of Vioxx on PGE2, production, insulitis and diabetes risk in NOD mice.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Materials

Phosphate-buffered saline, RPMI-1640, Dulbecco's modified Eagle's medium, l-glutamine, pencillin/streptomycin and fetal bovine serum were obtained from Life Technologies (Paisley, Renfrewshire, UK). CD11b+ magnetic beads for positive selection column were purchased from Miltenyi Biotec Ltd (Surrey, Bisley, UK). Enzyme-linked immunosorbent assay kits for detection of PGE2 were purchased from R&D Systems (Abingdon, Oxfordshire, UK). LPS and red cell lysis buffer were purchased from Sigma Chemical Co. (Poole, Dorset, Buckinghamshire, UK). The bicinchoninic acid (BCA) kit was obtained from Pierce, UK; the Qiagen RNeasy mini-kit was obtained from Qiagen Ltd (Crawley, Sussex, UK); the RiboGreen RNA quantitatation kit was purchased from Invitrogen (Paisley, Renfrewshire, UK); mouse COX-1/2 Amplicon, COX-1/2 primers (R and F) and probes for real-time quantitative reverse transcription–polymerase chain reaction (RT–PCR) came from MWG-Biotech AG (Ebersberg, Germany). Vioxx (Rofecoxib) and methylcellulose (MC) were purchased from Merck.

Animals

The NOD mice used were from the NOD colony at St Bartholomew's Hospital, London. All mice were housed under government guidelines. C57BL/6 control mice were obtained from Charles River Ltd (Margate, Kent, UK).

Vioxx administration by gavage

Two groups of female NOD mice were used for this study. Twenty-nine female NOD mice were treated with Rofecoxib at a final dose of 80 mg/kg body weight; the dose was as recommended by Merck based on in-house results (personal communication). A further 29 mice age-, sex- and litter-matched received 0·5% MC vehicle alone as a control group for the study.

Vioxx or MC were administrated daily by gavage, from weaning at 4 weeks of age, until diabetes diagnosis or in non-diabetic mice which survived until the end of the study in order to determine the effect on diabetes incidence and insulitis. Mice (n = 8 per treatment/control group) were culled at diabetes diagnosis or, for non-diabetic mice, at 30 weeks and had their pancreas removed and stored for islet infiltration scoring. A standard rodent maintenance diet (RM1{E}; Special Diet Services, Witham, Essex, UK) was provided ad libitum. Mice were screened weekly for diabetes from 10 weeks of age by urinary glucose testing strips (Diabur-Test, Boehringer Mannheim, Mannheim, Germany). Mice were designated diabetic when showing glucosuria of 56 mmol/l or above on two consecutive occasions plus blood glucose of 12·0 mmol/l or greater (One Touch II; Lifescan Inc., Milpitas, CA, USA). Body weight, food and water intake were also measured weekly until the end of the study at 30 weeks of age.

Histological studies

Pancreata from non-diabetic NOD mice which reached the end-point of the Vioxx intervention study at 30 weeks of age and mice which developed diabetes were removed by dissection and then processed and examined for stages of insulitis, as described previously [18].

Scoring of islet infiltration

Islet infiltration (insulitis) was assessed and scored arbitrarily according to the level of infiltration: no infiltration (grade 0), peri-insulitis (grade 1), where about 10% of the islets were infiltrated by a peripheral ring of lymphocytes; medium (grade 2) 10–50%; or severe (grade 3). The number of islets were noted for each pancreas and an index calculated by multiplying the number of islets in each category by the grades of infiltration [1–3] and adding these together.

In vitro cell culture and RT–PCR

CD11b+ monocyte isolation.  CD11b+ monocytes were isolated from spleen tissues using Microbead kits following the manufacturer's instructions (Milteyni Biotech) and purity was 95% CD11b+ monocytes. CD11b+ moncytes were cultured overnight at 37°C, 5% CO2 and then left untreated or stimulated with LPS (1 µg/ml) for 3 h (for mRNA) and 24 h (for PGE2) as optimized in in-house time–course experiments. RNA was then isolated using a Dynabeads mRNA DIRECTTM Microkit from 5 × 105 CD11b+ monocytes and quantitative RT–PCR was preformed using the following conditions for COX-1 and COX-2 mRNA expression.

Functional activity of COX-2 expression.  Conversion of PGH2 to PGE2 was used to assess the functional activity of COX-2 expression. The accumulated levels of PGE2 from both mouse CD11b+ monocytes pre- and post-LPS stimulation were measured by a competitive enzyme immunoassay, as specified by the manufacturer (Amersham Pharmacia Ltd, Amersham, Bucks, UK). All samples were batched and PGE2 assay was performed following the manufacturer's instructions and reference standard provided; samples were blinded to the experimenter. The batched assay included known high and low PGE2 in each plate as an in-house control. Furthermore, to correct for PGE2 present in the medium, we assayed baseline samples in duplicate using media alone. The final PGE2 concentration was calculated by subtracting this baseline value from the level detected in the culture supernatant. The limit of detection for PGE2 assay was 36·2 pg/ml. PGE2 concentration in the supernatants was standardized to pg/ml per 5 × 105 cells for reporting. Due to insufficient cells, we were unable to perform this in-vitro suppression assay with Vioxx in cells from the NOD mice.

Detection of monocyte COX-1 and COX-2 mRNA expression by quantitative RT–PCR

After 3 h LPS stimulation of isolated CD11b+ monocytes RNA was extracted using a Qiagen RNeasy mini-kit (Qiagen Ltd). RNA was quantified in triplicate, using Ribogreen quantification kits (Molecular Probes, Leiden, the Netherlands) with minor amendments to the original protocol. Real-time RT–PCR was performed for both COX-1 and COX-2 expression using the TaqMan system, and copy number determined using standard curves for murine COX-1 and COX-2 amplicons and finally expressed copy number per µg total RNA. For details of COX-1 and COX-2 primers and probes sequence see Tables 1 and 2 respectively.

Table 1.  Mouse cyclo-oxygenase-1 (COX-1) and COX-2 primers and probes.
GenesAccession no.Forward primerReverse primerTaqman probe
  1. Mouse COX-1 and COX-2 primers and probes used for real-time quantitative reverse transcription–polymerase chain reaction.

COX-1GI:135427345′-GGGAATTTGTGAATGCCACC-3′5′-GGGATAAGGTTGGACCGCA-3′5′-TCCGAGAAGTACTCATGC GCCTGGTACT-3′
COX-2GI:319815245′-AGCGAGGACCTGGGTTCAC-3′5′-AAGGCGCAGTTTATGTTGTCTGT-3′5′-AAGTCCACTCCATGGCC AGTCCTCG-3′
Table 2.  Mouse cyclo-oxygenase-1 (COX-1) and COX-2 amplicons.
GenesAmplicon sequence
  1. Mouse COX-1 and COX-2 amplicons used for real-time quantitative reverse transcription–polymerase chain reaction.

COX-15′-GGGAATTTGTGAATGCCACCTTCATCCGAGAAGTACTCATGCGCCTGGTACTCACAGTGCGGTCCAACCTTATCCC-3′
COX-25′-AGCGAGGACCTGGGTTCACCCGAGGACTGGGCCATGGAGTGGACTTAAATCACATTTATGGTGAAACTCTGGACAGACAA CATAAACTGCGCCTT-3′

Sequence-specific primers and probes

Intron-spanning primers and probes specific for the mouse COX-1 and COX-2 genes were designed using Primer Express (ABI, Warrington, Cheshire UK) (Table 1). Mouse COX-1 and COX-2 probes had a FAM reporter dye at the 5′ end and a quencher dye (6-carboxy-tetramethyl-rhodamine, TAMRA) on the 3′ end. The levels of mouse COX -1 and COX -2 mRNA were quantitated relative to amplicon-specific standard curves by real-time RT–PCR using 50 ng of total RNA in duplicate and analysed on the ABI Prism 7700-sequence detector. Standard curves were generated by serial dilution single-stranded sense oligonucleotides specifying COX-1 or COX-2 amplicons, as described previously [19]. All serial dilutions were carried out in duplicate. Standard curves in duplicate were repeated three times with duplicate no-template controls included with every RT–PCR run. The mRNA levels were expressed as COX-1 or COX-2 mRNA copy numbers/µg total RNA.

The RT–PCR reaction was carried out in a 25 µl reaction mixture containing 1 × TaqMan EZ buffer, 3 mM Mn(OAc)2, 300 mM dA/dC/dG/dUTP, 2·5 U rTth DNA polymerase, 10 pmol primers (forward and reverse primer sets; Table 1), 5 pmol TaqMan probe, 0·5 U AmpErase UNG and 100 ng of total RNA at 50°C for 2 min, 60°C for 30 min and 92°C for 5 min followed by 40 cycles of 20 s at 92°C and 1 min at 62°C. All samples were run in duplicate with water as a no template control, and known RNA from murine macrophage cell line J774 as a positive control.

Statistical analysis

Data were assembled using Microsoft Excel software and are expressed as mean ± standard deviation (s.d.). Data were analysed using the Prism statistical package (GraphPad Prism, Inc., version 3, La Jolla, CA, USA). All statistics were performed with either Student's t-test (two-tailed unless stated) or Mann–Whitney U-test (two-tailed unless stated), whichever was appropriate. We used a one-tailed test when analysing an expected increased in COX-2 expression in response to LPS. Differences between variables were considered significant at P < 0·05.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Basal monocyte COX-1 and COX-2 expression is normal in NOD mice

Basal monocyte COX-1 and COX-2 mRNA expression as well as PGE2 production were also similar in diabetic and non-diabetic NOD mice and control C57BL/6 mice (Fig. 1a,b). No difference was detected in basal monocyte COX-1 and COX-2 mRNA expression between female diabetic and non-diabetic NOD mice. NOD male mice, whether diabetic or not diabetic, were similar to female NOD mice throughout the study and the data on them is excluded here for clarity.

image

Figure 1. CD11b+ monocyte/macrophage cyclo-oxygenase-1 (COX-1) and COX-2 mRNA expression levels ± standard error of the mean pre- and post-lipopolysaccharide stimulation in female C57BL/6 control, diabetic and non-diabetic non-obese diabetic (NOD) mice by quantitative reverse transcription–polymerase chain reaction. An abnormal COX isoform switch is seen in diabetic NOD mice.

Download figure to PowerPoint

Basal monocyte COX-1 mRNA expression levels declined in female non-diabetic NOD mice from 5 to 10 weeks of age (P < 0·04) and remained lower thereafter (data not shown). In contrast, basal monocyte COX-1 expression remained unchanged in female C57BL/6 control mice throughout the 30 weeks of study. Basal monocyte COX-2 mRNA expression levels increased in female NOD mice from 10 to 15 weeks of age (P < 0·02) and remained higher thereafter (data not shown). In contrast to NOD mice, basal monocyte COX-2 mRNA expression remained unchanged in female C57BL/6 control mice throughout the 30 weeks of study, as it did for COX-1 mRNA expression.

Down-regulation of monocyte COX-1 mRNA after LPS stimulation in non-diabetic mice

When CD11b+ monocytes were stimulated with LPS, those mice with detectable basal monocyte COX-1 mRNA showed a significant decrease of COX-1 mRNA expression in non-diabetic female NOD mice (P = 0·009, = 22) and control C57BL/6 mice (P = 0·0001, n = 18) (Fig. 1a). The decrease in monocyte COX-1 mRNA following LPS stimulation was evident in female NOD mice at 5 weeks of age (7·4 × 105 ± 3·2 × 105versus 4·6 × 107 ± 2·8 × 107 mRNA copy number/µg total RNA; P < 0·01) and thereafter. Similarly, in female C57BL/6 control mice monocyte COX-1 mRNA expression also decreased significantly following LPS stimulation at 5 weeks of age (P < 0·01).

Up-regulation of monocyte COX-2 mRNA expression and PGE2 after LPS stimulation in non-diabetic mice

When CD11b+ monocytes were stimulated with LPS, a significant up-regulation of COX-2 mRNA expression was observed in non-diabetic female NOD mice (P = 0·0001, n = 21) and C57BL/6 control mice (P = 0·0001, n = 18) (Fig. 1b). Concomitantly, PGE2 secretion from these cells also increased significantly with LPS stimulation (basal 1498 ± 992 pg/ml versus LPS 3246 ± 2026 pg/ml; P < 0·006 in all groups). The increase in monocyte COX-2 mRNA following LPS stimulation was evident in female NOD mice at 5 weeks of age (3·8 × 105 ± 2·7 × 105versus 1·8 × 107 ± 1·01 × 107 mRNA copy number/µg total RNA; P < 0·0008) and thereafter. Time–course experiments confirmed that COX-2 mRNA and PGE2 production were increased significantly by LPS at the selected time of study (data not shown).

Altered regulation of monocyte/macrophage COX-1 and COX-2 mRNA expression after LPS stimulation in diabetic NOD mice

In contrast to non-diabetic NOD mice, the diabetic NOD mice did not show either a significant decrease in COX-1 mRNA expression or a significant increase in COX-2 mRNA expression following LPS stimulation (Fig. 1a,b). Furthermore, the magnitude of the COX-2 response to LPS was reduced significantly in diabetic NOD mice compared with non-diabetic NOD mice (P = 0·017) (Fig. 1b). Diabetic NOD male mice showed similar changes to diabetic female NOD mice and these data are excluded here for clarity. The observed down-regulation of COX-1 mRNA expression after LPS stimulation was significantly less in female diabetic NOD mice compared with non-diabetic NOD and C57BL/6 control mice (P < 0·05 for both comparisons) (Fig. 1a). Moreover, up-regulation of COX-2 mRNA expression after LPS stimulation was also significantly less in female diabetic NOD mice compared with non-diabetic NOD and C57BL/6 control mice (P < 0·05 for both comparisons) (Fig. 1b).

Diabetes frequency of NOD mice was not altered following Vioxx administration

The incidence of diabetes in female NOD mice administered Vioxx at 80 mg/kg and those administered MC vehicle alone (0·5%) was similar up to and at 30 weeks of age (Fig. 2). No significant difference in overall diabetes incidence was observed. Of the mice administered Vioxx (n = 29), seven died from causes unrelated to diabetes and of the remaining 22 mice nine developed diabetes by the end of the study; similarly, of the vehicle-administered control group (n = 29), two died from causes unrelated to diabetes and of the remaining 27 mice 10 developed diabetes by the end of the study. There was no delay in onset of the disease caused by the treatment. The Vioxx-treated group did not differ from the control group in body weight, food consumption or water intake. Blood glucose at the end of the study or at the onset of diabetes were unaltered by Vioxx administration (mean blood glucose 137 mg/dl versus 133 mg/ml in Vioxx-treated versus vehicle-treated non-diabetic NOD mice and 413 mg/dl versus 337 mg/ml in Vioxx-treated versus vehicle-treated diabetic NOD mice). Islet lymphocytic infiltration both at the end of the study and at the onset of diabetes was also unaltered by Vioxx administration (e.g. mean insulitis index score 1·97 in Vioxx-treated versus 2.2 in vehicle-treated diabetic NOD mice; Table 3).

image

Figure 2. Graphical representation of the incidence of diabetes following Vioxx (80 mg/kg) and vehicle-treated non-obese diabetic (NOD) mice. Diabetes incidence of NOD mice treated daily with either Vioxx (80 mg/kg) (= 29) or vehicle control (= 29). No significant differences were observed in either diabetes incidence or delay in diabetes onset with Vioxx.

Download figure to PowerPoint

Table 3.  Insulitis grading of Vioxx and vehicle-treated non-obese diabetic (NOD) mice which developed diabetes as determined by histological analysis.
 Grade 0Grade 1Grade 2Grade 3Index
  1. The percentage of islets at grade 1 was greater in Vioxx-treated NOD mice; however, this difference between the Vioxx- and vehicle-treated groups was not statistically significant. Values are represented as percentages ± standard deviation. MC, methylcellulose.

VIOXX15 ± 114 ± 334 ± 1348 ± 252 ± 1
Control (MC)6 ± 312 ± 533 ± 948 ± 112 ± 0·7

No change in monocyte functional response to LPS following Vioxx therapy in vivo

No significant difference was observed in basal monocyte PGE2 levels between Vioxx- and vehicle-treated female NOD mice, irrespective of disease status (for non-diabetic mice Vioxx versus vehicle: 1498 ± 992 versus 1284 ± 686 pg/ml; P = 0·51; diabetic mice Vioxx versus vehicle: 1748 ± 841 versus 1257 ± 739 pg/ml; P = 0·2). Following stimulation with LPS monocyte PGE2 levels increased significantly in both Vioxx- and vehicle-treated female NOD mice irrespective of disease status and did not differ according to either disease or treatment status. PGE2 production in diabetic NOD mice, both Vioxx-treated and vehicle-treated, increased following LPS stimulation (for non-diabetic mice on Vioxx pre- and post-LPS: 1498 ± 992 versus 3246 ± 2026 pg/ml respectively; P = 0·006; on vehicle pre- and post-LPS: 1284 ± 686 versus 2808 ± 1850 pg/ml; P = 0·002) (for diabetic mice on Vioxx pre- and post-LPS: 1748 ± 841 versus 4442 ± 2089 pg/ml respectively; P = 0·0031; on vehicle pre- and post-LPS: 1257 ± 739 versus 3488 ± 1091 pg/ml; P = 0·003).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We have demonstrated for the first time that monocytes from NOD mice and C57BL/6 control mice showed a COX mRNA isoform switch following stimulation with the non-specific antigen LPS, resulting in decreased levels of COX-1 mRNA while COX-2 mRNA increased. A similar COX mRNA isoform switch has been described by us previously in a murine macrophage cell line as well as in man [22]. Thus, both NOD mice and human with autoimmune diabetes show an altered COX mRNA isoform switch following LPS stimulation. Such a COX mRNA isoform switch, although documented previously, has never been demonstrated quantitatively in animals, and only recently shown by us in man [20]. Despite our evidence for abnormal monocyte COX responses in diabetic NOD mice and T1DM, we did not find any benefit from treating the NOD mice from weaning with the COX inhibitor, Vioxx.

There is substantial evidence that macrophages are abnormal in both NOD mice and in human T1DM [16,21,22]. We did not detect any abnormalities in basal monocyte COX-1 and COX-2 mRNA expression up to and including the onset of diabetes in NOD mice. Both in diabetic and non-diabetic NOD mice and non-diabetic C57BL/6 control mice throughout the study, basal monocyte COX-1 and COX-2 mRNA expression were similar. We had previously reported in man, both in controls and in T1DM, normal basal monocyte COX-1 and COX-2 mRNA expression [22]. Similarly, in both diabetic and non-diabetic NOD mice there was no abnormality in basal monocyte COX-2 function, as determined by PGE2 secretion.

We found that monocyte/macrophages from NOD mice and C57BL/6 control mice showed a COX mRNA isoform switch following stimulation with the non-specific antigen LPS, resulting in decreased levels of COX-1 mRNA while COX-2 mRNA increased. Accumulating evidence confirms that COX-1 expression can be modulated and expressed differentially by different cells under varying conditions and LPS can regulate constitutive and inducible transcription factor activity differentially [20,23–25]. We now show that in female NOD mice as well as in control mice this COX isoform switch can be detected at 5–10 weeks of age and that it continues to be detected thereafter. However, once NOD mice developed diabetes there was a reduced down-regulation of COX-1 mRNA response to LPS, and a reduced up-regulation of COX-2 mRNA response to LPS compared with the non-diabetic NOD and control mice. A similar change in COX mRNA isoform switch has been reported recently by us in human T1DM [22]. However, the alteration in COX mRNA isoform switch associated with human T1DM is familial, as it can also be detected in non-diabetic twins genetically at risk of the disease but selected to be at low disease-risk; while in NOD mice the change was observed only in the diabetic mice, not in the non-diabetic mice, and is therefore not familial. The explanation, therefore, for the altered monocyte/macrophage COX isoform switch in diabetic NOD mice and diabetic human twins remains unclear, although it is an abnormality common to both.

Because COX-2 response to LPS is essentially anti-inflammatory, the altered mRNA response we detected could predispose to inflammation and might thereby predispose to chronic inflammatory diseases such as T1DM. Macrophages are abnormal in both NOD mice and in human T1DM, although it remains to be determined whether the macrophage abnormalities, which include eicosanoid imbalance, predispose to the disease or are secondary to it; for example, NOD mice show changes in COX-2 expression in the gut before the onset of insulitis [26,27]. We therefore examined the natural history of monocyte COX mRNA and PGE2 in NOD mice and sought to prevent diabetes using the COX-2-specific inhibitor, Vioxx, by feeding it to a treatment group from weaning at 4 weeks of age until the end-point of the study. Vioxx failed to impact on lymphocytic infiltration in these animals, nor did it affect the time to diabetes onset or the diabetes frequency. Despite the high dose of Vioxx used, it remains possible that it was inactive in our NOD mice given the normal basal monocyte PGE2 levels in Vioxx-treated mice. Yet in a previous study of islet xenografts in NOD mice, Vioxx, at a comparable dose to that used here, failed to prevent graft loss but did prolong mean graft survival time [28]. However, reducing macrophage PGE2 production in-vivo, either by dietary fatty acid manipulation or by treating NOD mice with Indomethacin, which is not a specific COX inhibitor, did reduce diabetes incidence significantly in female NOD mice [29].

Taken together, our observations indicate that there are monocyte abnormalities in COX mRNA expression in diabetic NOD mice but they may not predispose to diabetes, as early treatment with the COX-2 inhibitor Vioxx had no impact on the disease course or onset. Our results do not encourage the use of such therapy in man, but an important caveat here is that there are many differences between autoimmune diabetes in NOD mice and in man and many ways to interfere with PG metabolism [30]. Our present study highlights some of those differences, in that the familial nature of the abnormality in COX mRNA expression in our human studies was not found in these studies of NOD mice.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study was supported by the British Diabetic Twin Research Trust (L. R. B., R. D. G. L.), Diabetes UK (R. D. G. L.) and the Joint Research Board at St Bartholomew's Hospital (H. B., R. D. G. L.). We also thank the late Professor Derek Willoughby for advice.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References