International Ophthalmology 15: 335-341, 1991. 9 1991 Kluwer Academic Publishers. Printed in the Netherlands.
Effect of piroxicam on the blood-retina barrier in experimentally induced diabetes in rats Mostafa Bahgat, Hossam H. Anis, Gholam A. Peyman, Hassan G. Farahat, Gareth J. Parry 1 & Bahram Khoobehi LSU Eye Center and 1Department of Neurology, Louisiana State University Medical Center School of Medicine, New Orleans, LA, USA Accepted 31 January 1991
Key words: piroxicam, streptozocin, blood-retina barrier, vitreous fluorophotometry, platelet aggregation Abstract
The effect of piroxicam on the blood-retina barrier was evaluated in rats with experimentally induced diabetes. Diabetes was induced in rats by intraperitoneal injection of streptozocin (STZ). Diabetic rats were divided into two equal groups: those treated with piroxicam, a long-acting platelet inhibitor, and an untreated control group. Vitreous fluorophotometry (VFP) was performed both before and two weeks after induction of diabetes and piroxicam intake. Streptozocin-induced diabetes caused an alteration in the blood-retinal barrier evidenced by an increase in vitreous fluorescein concentration in diabetic rats compared with normal rats. Piroxicam intake did not lead to significant change in vitreous fluorescein concentrations. However, the examination had to be terminated at two weeks because of cataract formation. The piroxicam treated group showed less incidence of lens opacity formation (59.1% compared to 81.8% in the untreated group, p = 0.0006). Piroxicam administration appears to protect the diabetic rat eye against lens opacification.
Introduction
The normal blood-retina barrier prevents the passage of the most intravascular components into the vitreous. The site of that barrier is the retinal pigment epithelium and the endothelial cells of the retinal vessels [3]. In diabetes mellitus, the blood-retina barrier is disrupted and allows fluorescein leakage into the vitreous. The exact nature of the defect responsible for the breakdown of this barrier is not known [21]. However, it occurs before the detection of visible pathologic changes in the retina [7]. Breakdown of that barrier may be functional in nature as it may be improved by insulin therapy [2]. Insulin either affects the decompensated retinal endothelial cells
directly [5] or indirectly by normalization of blood glucose level [22], or improvement of diabetic acidosis [15, 16]. The last possibility is questionable since Vine et al. [20] did not find significant difference in blood pH between normal and diabetic rats. Waltman et al. [22] and Tso et al. [19] postulated that breakdown of blood-retina barrier is functional rather than organic in nature; therefore it could be influenced by different forms of therapies. Piroxicam is a cyclooxygenase inhibitor with antiinflammatory and antiplatelet properties. Like most drugs of its class, it is a weak aldose reductase inhibitor. Parry and Kozu [11] showed that piroxicam slows the rate of progression of neuropathy in diabetic rats, probably by inhibiting platelet aggre-
This work was supported in part by U.S. Public Health Service Grants EY02377, EY07541 and EY08137 from the National Eye Institute, National Institutes of Health, Bethesda, MD and by the Juvenile Diabetes Foundation International and Pfizer, Inc.
336
M. Bahgat et al.
gation. They postulated that pharmacologically inhibiting platelet aggregation should prevent or retard the development of the ischemic component of diabetic neuropathy. These findings encouraged us to study the effect of piroxicam on diabetic retinopathy. Because vitreous fiuorophotometry (VFP) after intravenous injection of fluorescein is a sensitive method of evaluation of blood-ocular barrier, we chose it to assess the effect of piroxicam on experimentally induced diabetes in rats.
Materials and methods
Twenty-six male Long-Evans rats (Harlan SpragueDawly, Hsd/Blu : LE) initially weighing 200-225 g were used. Slit lamp examination was performed using halothane inhalation anesthesia and cyclopentolate 1% and tropicamide 1% eye drops. Thorough examination was performed to assess the clarity of the ocular media. Then each animal was anesthetized by intraperitoneal injection of ketamine hydrochloride (60 mg/kg) and xylazine hydrochloride (6 mg/kg). The pupils were dilated with the same combination used for slit lamp examination. The external jugular vein was exposed and sodium fluorescein was injected intravenously at a dose of 33 mg/kg. Vitreous fluorescein concentrations were measured 5 minutes and 60 minutes after fluorescein injection using the fluorophotometer (Fluorotron | Master, Coherent, Palo Alto, CA). A rat adaptor lens was coupled to the optics of the rat eye without the use of contact lens. The first measurement point of reference was located at a position posterior to the retina, proceeding forward to a position anterior to the cornea. Each measurement of the machine was automatically calibrated to an internal fluorescent glass. Methyl cellulose drops were used to avoid dryness of the cornea. To induce diabetes, the animals were fasted for approximately 19 hours and injected intraperitoneally with streptozocin (STZ) (Sigma, 65 mg/kg) which was dissolved in 0.05 M citrate buffer, pH 4.5. Mean weight prior to fasting was 311 + 32 g. All animals were given free access to water and
food and were kept at the same state of hydration during the study. Non-fasting blood glucose was measured 24 hours after STZ injection using an automated blood-glucose monitor (Boehringer Mannhein Accu-Chek II). Fasting blood glucose (FBG) was measured 48 hours after STZ injection. Diabetes was defined as an FBG value greater than 120mg/dL. Twenty-four animals met the criteria for diabetes. These animals were divided into two equal groups by first ranking the animals according to the fasting glucose values. Pairs were then formed with the animals in each pair matched for FBG. The animal in each pair with the higher FBG value was then alternatively assigned to either the treated or untreated group. Mean FBG was 293.2+ 42.8mg/dL for the treated group and 290.1 + 35.5 mg/dL for the untreated group. This difference was not statistically significant (Student's t = 0.26, P = 0.80). Treatment with piroxicam (Pfizer, 5mg/kg) commenced 48 hours after injection of STZ. The piroxicam was dissolved in a 5% aqueous solution of gum tragacanth (Sigma), and administered orally by garage once a day in the afternoon. Slit lamp examination was performed for both eyes of all rats one week after induction of diabetes and then weekly for six weeks. Grading of lens opacities followed the classification of Poulsom et al. on a seven-point scale of severity: O = clear, no sign of opacity; P = pinpoint central whiteness; T = traces of opacity or central hazeness; Few = a few distinct opacities (small punctate); S = several distinct opacities (small punctate with or without merging); M = mostly opaque (dense white haziness with or without distinct white opacities); C = completely opaque lens [12]. The second VFP was done on the same eye examined prior to induction of diabetes two weeks after the start of piroxicam intake. Two rats (one from each group) died during the VFP examination in spite of subcutaneous saline injection and warming of rats. VFP and slit lamp study were performed by the same workers who did the baseline study, unaware of the treatment condition of the animals. The development of lens opacities made the repetition of VFP difficult after two weeks. The axial length and vitreous chamber depth of
Effect of piroxicam on the blood-retina barrier the rat's eye were found to be variable in previous studies [13-16]. Their records ranged from 5.686.35 mm and 1.11-1.71 mm, respectively. Thus the mean axial length is about 6.0 mm and mean vitreous thickness is about 1.41mm. Midvitreous point is taken at 0.7 mm which coincides with the ninth measurement (after conversion to the Fluorotron printout scale which contains about 72 measurements between chorioretinal and corneal peaks) starting from the top of the chorioretinal peak. The measurements were analyzed following the protocol of the Fluorotron manufacturer. All values are presented as means plus or minus standard error of mean. Comparisons were made between treated and untreated groups using the student's t test. Probability values less than 5% were considered stastistically significant. Animals were sacrificed at the end of the study by overdose of ketamine.
Results
Figure 1A shows the Fluorotron printout of untreated rats before and after induction of diabetes. Figure 1B shows the Fluorotron results of treated rats before and after STZ injection. The mean vitreous fluorescein concentrations at the chosen point (0.7 mm in front of the retina) are shown in Table 1 and Fig. 2. The differences in baseline measurements between the treated and untreated groups were not statistically significant: t / p = -0.70/0.49 after 5 minutes and t / p =
-0.75/0.44 after 60 minutes. After induction of diabetes, vitreous fluorescein concentrations became higher than those of normal rats. The differences were statistically significant after 5 minutes in both piroxicam treated (t/p = -2.62/0.016) and untreated groups (t/p = - 3.10/0.006). However, the increase after 60 minutes was not statistically significant (t/p= -0.87/0.40) for the treated group, but slightly significant (t/p = -2.32/0.03) for the untreated group. Comparison between the treated and untreated groups after induction of diabetes and piroxicam intake showed lower vitreous fluorescein concentrations in the treated group after 5 minutes (t/p --0.84/0.41) and 60 minutes (t/p = -0.64/0.53). However, these differences were statistically insignificant. Correlation between the mean FBG and the midvitreous fluorescein concentrations was not statistically significant in either group after 5 or 60 minutes after fluorescein injection. Slit lamp examination revealed that 18 eyes of the 11 untreated rats developed lens opacities in the form of a few scattered vacuoles as early as one week postinjection of STZ. After a six-week follow-up period, central opacities were the commonest (8/18), followed by combined central and sectorial opacities (6/18). Isolated sectorial opacities were seen in three eyes. Multiple scattered vacuoles were found in one eye only. The remaining four eyes did not show any lens opacities. The frequency of lens opacities was less in the treated group. In the first week, eight eyes showed a few scattered vacuoles. At the end of the six-week
7"able 1. Vitreous fluorescein concentrations for Piroxicam treated and untreated rats. Group/time of VFP
Mean + SEM normal
Mean _+ SEM diabetic
T/P
* Untreated/5 minutes Untreated/60 minutes *Treated/5 minutes Treated/60 minutes
10.75 43.71 14.12 51.05
37.17 69.01 29.52 60.93
-
* Each group contains 11 rats. VFP: Vitreous fluorophotometry. SEM: Standard error of mean. P > 0.05 statistically non-significant. P < 0.05 statistically significant.
+ + + +
2.64 6.28 4.04 6.98
337
_+ + _+ _+
8.11 8.88 4.27 9.01
3.10/0.006 2.33/0.03 2.62/0.02 0.87/0.40
M. Bahgat et al.
338
~ /
l
U5
f /
pre stz
to 2 / /3 t
i t.....~.,-, re}
,..i ',,~^. ,~,.F
I/
U 60
j
p"
1/
u. po,, ,,.
\
\
t
"*'%~
%+
....
~ I r,;h; ' ' I ' ' ' {~ I ~;[;
. . . . . .
2~ I s
. . . . . .
k \
J .
i~ 2
.
.
.
I[~[~ '
post stz
t
/f
L.~
LO1
~
U 6o
pre stz
t
:Lg 2
%
L/O1
-S
/
/
k\
,\
'1
%,
10
"'
'
'
'e'.~.;
'
'
''
'
"tb%r,;
. . . . . .
~'olr,;~;
'"'
. . . .
~e~.;.;
9
'
Fig. 1A. Fluorophotometer printout for untreated rats after 5 and 60 minutes of intravenous fluorescein injection before and after induction of diabetes.
study, 13 of 22 eyes developed lens opacities. Again, central haze was the commonest (9/13), followed by peripheral sectorial opacities (2/13). Combined central haze with peripheral sectors and scattered vacuoles were seen in one eye each. The lens opacities were less severe in the treated group than in the untreated rats. Thus, piroxicam had a significant effect (P = 0.006) in minimizing cataract formation in diabetic rats.
Discussion Piroxicam is one of the commonly used antiinflammatory medications. It was found recently [11] to improve ischemic diabetic neuropathy. We evaluated piroxicam in rats with experimentally induced diabetes to determine its possible effect on the blood-retina barrier.
A successful animal model of diabetic retinopathy is difficult to produce. STZ-induced diabetes in rats leads to the development of lens opacities that interfere with fundus examination and VFP [5, 22]. In addition, induction of diabetes is not always successful and some of the rats may become nondiabetic even without medications [10, 13]. The rats are susceptible to dehydration during general anesthesia and are unable to tolerate low temperatures, requiring subcutaneous saline injections and warming during surgery. The rat's eye is small [9]; the lens is large, globular in shape, and occupies about two-thirds of the axial length of the eyeball [1, 8]. The vitreous chamber is crescentshaped, shallow, and approximately 1.4 mm thick [4]. We induced diabetes successfully in 24 out of 26 rats; none of them recovered during the follow-up period. The mortality rate was low (2 of 24) due to
Effect of piroxicam on the blood-retina barrier
339
183 T 5 pre stz
10 2
/ !
"~
/
,^r ' ' ' hr.;,,; . . . . . .
,'.,,
laa,[ \.
f~,
~
I I
\.,
-~ ' ' '
'~b,;~; . . . . . .
/
] !
...../
k.
,v s
......
~.:~ ~,;~,; . . . . .
s163
......
fe
1~~ //
.~; ......
~'~,;~; '
[
T 60 pre stz
/I
T 60 post I t x
] ,)
Ii 01
/
-x*t "-~
printout for piroxicam-treated
,,f"J~
\
t ~p / :tO 1
-5
Fig. lB. F l u o r o p h o t o m e t e r
\\~
~'\.~ "-~...~,
'0In;r,; ' ' '
1 0 r,,rr,
r a t s a f t e r 5 a n d 60 m i n u t e s o f i n t r a v e n o u s
'20In;n; . . . . . .
ff~lri;r,~'~-
fluorescein injection before and
after induction of diabetes.
short duration of anesthesia and type of anesthetic used [5, 6, 13, 18, 22]. Although many sites have been tested for fluorescein injection (tail vein [20], femoral vein [6, 19]), we found the external jugular vein to be the most accessible. Full pupillary dilatation is mandatory for proper VFP records. Previous studies reported the use of different combinations of medications, such as 2.5% phenylephrine and 1% cyclopentolate [5, 16, 18] or 2.5% phenylephrine with 1% tropicamide [19]. We found a combination of 1% cyclopentolate and 1% tropicamide to be most satisfactory for maximum and long-lasting pupillary dilatation. The mechanism by which STZ induced-diabetes causes alterations in the permeability of the bloodretina barrier is not known. Direct toxic effect on the retina is doubtful since STZ acts on the pancreas, as proved by the late alterations in active transport of organic anions [5] and by the normalization of the blood-retina barrier by pancreatic
islet transplant [22]. Osmotic effect due to hyperglycemia is not an acceptable hypothesis since recovery of blood-retina barrier transport occurs without normalization of blood-glucose levels [5]. Thus alternation of this barrier may be due to a combination of factors. The baseline vitreous fluorescein concentrations were not statistically different between the two groups of this work. After induction of diabetes, vitreous fluorescein concentrations became significantly higher in both the treated and untreated groups if compared to the prediabetic measurements. However, the differences were statistically significant after 5 minutes but not after 60 minutes of fluorescein intake. Thus, our findings showed minimal permeability to fluorescein in the normal rat retina and a marked breakdown of the fluorescein blood-retina barrier in rats with STZ induced diabetes. The defect was detectable 15 days following the chemical induction of diabetes.
340
M. Bahgat et al.
Fluoroscein
~-~-I. -~
concentration
(ng/ml)
T-pre-stz T-post-stz U-pre-stz U-post-stz
20
0
i 5
i
time (minutes)
60
70
Fig. 2. Midvitreous fluorescein concentrations versus time for piroxicam treated and untreated groups before and after induction of diabetes. U-Pre-STZ: Untreated group before streptozocin injection; U-Post-STZ: Untreated group after streptozocin injection; T-Pre-STZ: Treated group before streptozocin injection; T-Post-STZ: Treated group after streptozocin injection.
Three-fold to four-fold increase in the concentration of fluorescein in the vitreous was found 8-10 days after induction of diabetes by Waltman et al. [22] and Tso et al. [19]. They reported that the increase in fluorescein concentration occurs sharply and is a constant change. Other workers [10, 13, 14] related that change to many factors other than diabetes. Smith et al. [18] and Vine et al. [20] showed that mid-vitreous fluorescein concentrations in rats with STZ induced diabetes were significantly lower than in normal rats. However, vitreous fluorescein concentrations, when compared with total plasma fluorescein concentration, was significantly increased for after two weeks of diabetes. They explained this finding by increased renal fiuorescein excretion as a result of diabetic nephropathy. The reason for their opposing results in vitreous fluorescein concentrations is not readily apparent. It may reflect differences in the techniques of these studies. In our study, vitreous fluorescein concentrations showed lower levels of fluorescein in the piroxicam treated group than the untreated group. This difference was noticed after 5 and 60 minutes. However, the change in vitreous fluorescein concentration between the two groups was not statistically significant. This finding indicates piroxicam does not have a significant effect on the permeability of the blood-retina barrier. However, long-term
treatment with piroxicam may be more effective and should be investigated. Follow-up examination for lens opacities agree with other records [12] that cataracts did not always develop at equal rates in left and right eyes of diabetic rats. Our findings differ in the early onset of appearance of multiple scattered vacuoles, which were seen as early as the first week after induction of diabetes. The incidence of lens opacities was much higher in the untreated group (81.8%); all rats were bilaterally affected. For the treated group, lens opacities developed in 59.1%. Both eyes were affected in four rats only. Our findings indicate piroxicam therapy may have an effect on lens opacity formation in diabetic rats, probably due to aldose reductase inhibition [17].
Acknowledgements The authors would like to thank Coherent for generously allowing us to use the Fluorotron. We would also like to acknowledge the help of Mark Hurry, Department of Neurology, in performing the experiments. This work was supported in part by U.S. Public Health Service Grants EY02377, EY07541 and EY08137 from the National Eye Institute, National Institutes of health, Bethesda, MD and by grants from the Juvenile Diabetes Foundation International and Pfizer Inc.
References 1. Block MT. A note on the refraction and image formation of the rat's eye. Vision Res 1969; 9: 705-11. 2. Cunha-Vaz JG. Pathophysiology of diabetic retinopathy. Br J Ophthalmol 1978; 62: 351-5. 3. Cunha-Vaz J, Abreu J R F C~mpos AJ, Figo GM. Early breakdown of the blood-retinal barrier in diabetes. Br J Ophthalmol 1975; 59: 649-56. 4. Hughes A. A schematic eye for the rat. Vision Res 1979; 19: 569-88. 5. Jones CW, Cunha-Vaz J, Zweig KO, Stein M. Kinetic vitreous fluorophotometry in experimental diabetes. Arch Ophthalmol 1979; 97: 1941-3. 6. Kirber WM, Nichols CW, Grimes PA, Winegrad AI, Laties AM. A permeability defect of the retinal pigment epi-
Effect of piroxicam on the blood-retina barrier
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
thelium. Occurrence in early streptozocin diabetes. Arch Ophthalmol 1980; 98: 725-8. Krupin T, Waltman S, Oestrich C, Santiago J, Ratzan S, Kilo C, Becket B. Vitreous fluorophotometry in juvenileonset diabetes mellitus. Arch Ophthalmol 1978; 96: 812-4. Lashley KS. The mechanism of vision. The structure and image-forming power of the rat eye. J Comp Psychol 1932; 13: 173-200. Massof RW, Chang FW. A revision of the rat schematic eye. Vision Res 1972; 12: 793-6. Palestine AG, Brubaker RF. Plasma binding of fluorescein in normal subjects and in diabetic patients. Arch Ophthalmol 1982; 100: 1160-1. Parry GJ, Kozu H. Piroxicam may reduce the rate of progression of experimental diabetic neuropathy. Neurology (in press). Poulsom R, Boot*Handford RP, Heath H. Some effects of aldose reductase inhibition upon the eyes of long-term streptozotocin-diabetic rats. Curr Eye Res 1982/1983: 2: 351-5. Prager TC, Chu HH, Garcia CA, Anderson RE, et al. The influence of vitreous change on vitreous fluorophotometry. Arch Ophthalmol 1982; 100: 594-6. Prager TC, Wilson DJ, Avery GD, Merritt JH, Garcia CA, Hopen G, Anderson RE. Vitreous fluorophotometry: Identification of sources of variability. Invest Ophthalmol Vis Sci 1981; 21: 854-64. Rusin MM, Blair NP, Shakin E. The effect of pH on the transport of fluorescein across the blood-retinal barrier. Invest Ophthalmol Vis Sci 1983; 24 (suppl): 249. Shinowara NL, Grimes PA, Rapoport SI, Laties AM, et al. Acidosis alters fluorescein permeability across the pigment
17.
18.
19.
20.
21.
22.
341
epithelium. Invest Ophthalmol Vis Sci 1983; 24 (suppl): 250. Sharma YR, Cotlier E. Inhibition of lens and cataract aldose reductase by protein-bound anti-rheumatic drugs: Salicylate, indomethacin, oxyphenbutazone, sulindac. Exp Eye Res 1982; 35: 21-7. Smith SS, Ashburn FS, Pilkerton AR, Recant L. Vitreous fluorophotometry in three models of experimental diabetes mellitus. Retina 1982; 2: 121-5. Tso MOM, Cunha-Vaz JG, Shih CY, Jones CW, et al. Clinicopathologic study of blood-retinal barrier in experimental diabetes mellitus. Arch Ophthalmo11980; 98: 203240. Vine AK, Kisly AM, Betz AL, Howatt WF. Vitreous fluorophotometry in rats with streptozocin-induced diabetes. Arch Ophthalmol 1984; 102: 1083-5. Waltman SR, Krupin T, Kilo C, Becker B. Vitreous fluorophotometry in adult-onset diabetes mellitus. Am J Ophthalmol 1979; 88: 342-5. Waltman SR, Oestrich C, Krupin T, Hanish S, Ratzan S, Santiago J, Kilo C. Quantitative vitreous fluorophotometry: A sensitive technique for measuring early breakdown of the blood-retina barrier in young diabetic patients. Diabetes 1978; 27: 85-7.
Address for offprints: G.A. Peyman, LSU Eye Center, 2020 Gravier Street, Suite B, New Orleans, LA 70112, USA
Effect of piroxicam on the blood-retina barrier in experimentally induced diabetes in rats Mostafa Bahgat, Hossam H. Anis, Gholam A. Peyman, Hassan G. Farahat, Gareth J. Parry 1 & Bahram Khoobehi LSU Eye Center and 1Department of Neurology, Louisiana State University Medical Center School of Medicine, New Orleans, LA, USA Accepted 31 January 1991
Key words: piroxicam, streptozocin, blood-retina barrier, vitreous fluorophotometry, platelet aggregation Abstract
The effect of piroxicam on the blood-retina barrier was evaluated in rats with experimentally induced diabetes. Diabetes was induced in rats by intraperitoneal injection of streptozocin (STZ). Diabetic rats were divided into two equal groups: those treated with piroxicam, a long-acting platelet inhibitor, and an untreated control group. Vitreous fluorophotometry (VFP) was performed both before and two weeks after induction of diabetes and piroxicam intake. Streptozocin-induced diabetes caused an alteration in the blood-retinal barrier evidenced by an increase in vitreous fluorescein concentration in diabetic rats compared with normal rats. Piroxicam intake did not lead to significant change in vitreous fluorescein concentrations. However, the examination had to be terminated at two weeks because of cataract formation. The piroxicam treated group showed less incidence of lens opacity formation (59.1% compared to 81.8% in the untreated group, p = 0.0006). Piroxicam administration appears to protect the diabetic rat eye against lens opacification.
Introduction
The normal blood-retina barrier prevents the passage of the most intravascular components into the vitreous. The site of that barrier is the retinal pigment epithelium and the endothelial cells of the retinal vessels [3]. In diabetes mellitus, the blood-retina barrier is disrupted and allows fluorescein leakage into the vitreous. The exact nature of the defect responsible for the breakdown of this barrier is not known [21]. However, it occurs before the detection of visible pathologic changes in the retina [7]. Breakdown of that barrier may be functional in nature as it may be improved by insulin therapy [2]. Insulin either affects the decompensated retinal endothelial cells
directly [5] or indirectly by normalization of blood glucose level [22], or improvement of diabetic acidosis [15, 16]. The last possibility is questionable since Vine et al. [20] did not find significant difference in blood pH between normal and diabetic rats. Waltman et al. [22] and Tso et al. [19] postulated that breakdown of blood-retina barrier is functional rather than organic in nature; therefore it could be influenced by different forms of therapies. Piroxicam is a cyclooxygenase inhibitor with antiinflammatory and antiplatelet properties. Like most drugs of its class, it is a weak aldose reductase inhibitor. Parry and Kozu [11] showed that piroxicam slows the rate of progression of neuropathy in diabetic rats, probably by inhibiting platelet aggre-
This work was supported in part by U.S. Public Health Service Grants EY02377, EY07541 and EY08137 from the National Eye Institute, National Institutes of Health, Bethesda, MD and by the Juvenile Diabetes Foundation International and Pfizer, Inc.
336
M. Bahgat et al.
gation. They postulated that pharmacologically inhibiting platelet aggregation should prevent or retard the development of the ischemic component of diabetic neuropathy. These findings encouraged us to study the effect of piroxicam on diabetic retinopathy. Because vitreous fiuorophotometry (VFP) after intravenous injection of fluorescein is a sensitive method of evaluation of blood-ocular barrier, we chose it to assess the effect of piroxicam on experimentally induced diabetes in rats.
Materials and methods
Twenty-six male Long-Evans rats (Harlan SpragueDawly, Hsd/Blu : LE) initially weighing 200-225 g were used. Slit lamp examination was performed using halothane inhalation anesthesia and cyclopentolate 1% and tropicamide 1% eye drops. Thorough examination was performed to assess the clarity of the ocular media. Then each animal was anesthetized by intraperitoneal injection of ketamine hydrochloride (60 mg/kg) and xylazine hydrochloride (6 mg/kg). The pupils were dilated with the same combination used for slit lamp examination. The external jugular vein was exposed and sodium fluorescein was injected intravenously at a dose of 33 mg/kg. Vitreous fluorescein concentrations were measured 5 minutes and 60 minutes after fluorescein injection using the fluorophotometer (Fluorotron | Master, Coherent, Palo Alto, CA). A rat adaptor lens was coupled to the optics of the rat eye without the use of contact lens. The first measurement point of reference was located at a position posterior to the retina, proceeding forward to a position anterior to the cornea. Each measurement of the machine was automatically calibrated to an internal fluorescent glass. Methyl cellulose drops were used to avoid dryness of the cornea. To induce diabetes, the animals were fasted for approximately 19 hours and injected intraperitoneally with streptozocin (STZ) (Sigma, 65 mg/kg) which was dissolved in 0.05 M citrate buffer, pH 4.5. Mean weight prior to fasting was 311 + 32 g. All animals were given free access to water and
food and were kept at the same state of hydration during the study. Non-fasting blood glucose was measured 24 hours after STZ injection using an automated blood-glucose monitor (Boehringer Mannhein Accu-Chek II). Fasting blood glucose (FBG) was measured 48 hours after STZ injection. Diabetes was defined as an FBG value greater than 120mg/dL. Twenty-four animals met the criteria for diabetes. These animals were divided into two equal groups by first ranking the animals according to the fasting glucose values. Pairs were then formed with the animals in each pair matched for FBG. The animal in each pair with the higher FBG value was then alternatively assigned to either the treated or untreated group. Mean FBG was 293.2+ 42.8mg/dL for the treated group and 290.1 + 35.5 mg/dL for the untreated group. This difference was not statistically significant (Student's t = 0.26, P = 0.80). Treatment with piroxicam (Pfizer, 5mg/kg) commenced 48 hours after injection of STZ. The piroxicam was dissolved in a 5% aqueous solution of gum tragacanth (Sigma), and administered orally by garage once a day in the afternoon. Slit lamp examination was performed for both eyes of all rats one week after induction of diabetes and then weekly for six weeks. Grading of lens opacities followed the classification of Poulsom et al. on a seven-point scale of severity: O = clear, no sign of opacity; P = pinpoint central whiteness; T = traces of opacity or central hazeness; Few = a few distinct opacities (small punctate); S = several distinct opacities (small punctate with or without merging); M = mostly opaque (dense white haziness with or without distinct white opacities); C = completely opaque lens [12]. The second VFP was done on the same eye examined prior to induction of diabetes two weeks after the start of piroxicam intake. Two rats (one from each group) died during the VFP examination in spite of subcutaneous saline injection and warming of rats. VFP and slit lamp study were performed by the same workers who did the baseline study, unaware of the treatment condition of the animals. The development of lens opacities made the repetition of VFP difficult after two weeks. The axial length and vitreous chamber depth of
Effect of piroxicam on the blood-retina barrier the rat's eye were found to be variable in previous studies [13-16]. Their records ranged from 5.686.35 mm and 1.11-1.71 mm, respectively. Thus the mean axial length is about 6.0 mm and mean vitreous thickness is about 1.41mm. Midvitreous point is taken at 0.7 mm which coincides with the ninth measurement (after conversion to the Fluorotron printout scale which contains about 72 measurements between chorioretinal and corneal peaks) starting from the top of the chorioretinal peak. The measurements were analyzed following the protocol of the Fluorotron manufacturer. All values are presented as means plus or minus standard error of mean. Comparisons were made between treated and untreated groups using the student's t test. Probability values less than 5% were considered stastistically significant. Animals were sacrificed at the end of the study by overdose of ketamine.
Results
Figure 1A shows the Fluorotron printout of untreated rats before and after induction of diabetes. Figure 1B shows the Fluorotron results of treated rats before and after STZ injection. The mean vitreous fluorescein concentrations at the chosen point (0.7 mm in front of the retina) are shown in Table 1 and Fig. 2. The differences in baseline measurements between the treated and untreated groups were not statistically significant: t / p = -0.70/0.49 after 5 minutes and t / p =
-0.75/0.44 after 60 minutes. After induction of diabetes, vitreous fluorescein concentrations became higher than those of normal rats. The differences were statistically significant after 5 minutes in both piroxicam treated (t/p = -2.62/0.016) and untreated groups (t/p = - 3.10/0.006). However, the increase after 60 minutes was not statistically significant (t/p= -0.87/0.40) for the treated group, but slightly significant (t/p = -2.32/0.03) for the untreated group. Comparison between the treated and untreated groups after induction of diabetes and piroxicam intake showed lower vitreous fluorescein concentrations in the treated group after 5 minutes (t/p --0.84/0.41) and 60 minutes (t/p = -0.64/0.53). However, these differences were statistically insignificant. Correlation between the mean FBG and the midvitreous fluorescein concentrations was not statistically significant in either group after 5 or 60 minutes after fluorescein injection. Slit lamp examination revealed that 18 eyes of the 11 untreated rats developed lens opacities in the form of a few scattered vacuoles as early as one week postinjection of STZ. After a six-week follow-up period, central opacities were the commonest (8/18), followed by combined central and sectorial opacities (6/18). Isolated sectorial opacities were seen in three eyes. Multiple scattered vacuoles were found in one eye only. The remaining four eyes did not show any lens opacities. The frequency of lens opacities was less in the treated group. In the first week, eight eyes showed a few scattered vacuoles. At the end of the six-week
7"able 1. Vitreous fluorescein concentrations for Piroxicam treated and untreated rats. Group/time of VFP
Mean + SEM normal
Mean _+ SEM diabetic
T/P
* Untreated/5 minutes Untreated/60 minutes *Treated/5 minutes Treated/60 minutes
10.75 43.71 14.12 51.05
37.17 69.01 29.52 60.93
-
* Each group contains 11 rats. VFP: Vitreous fluorophotometry. SEM: Standard error of mean. P > 0.05 statistically non-significant. P < 0.05 statistically significant.
+ + + +
2.64 6.28 4.04 6.98
337
_+ + _+ _+
8.11 8.88 4.27 9.01
3.10/0.006 2.33/0.03 2.62/0.02 0.87/0.40
M. Bahgat et al.
338
~ /
l
U5
f /
pre stz
to 2 / /3 t
i t.....~.,-, re}
,..i ',,~^. ,~,.F
I/
U 60
j
p"
1/
u. po,, ,,.
\
\
t
"*'%~
%+
....
~ I r,;h; ' ' I ' ' ' {~ I ~;[;
. . . . . .
2~ I s
. . . . . .
k \
J .
i~ 2
.
.
.
I[~[~ '
post stz
t
/f
L.~
LO1
~
U 6o
pre stz
t
:Lg 2
%
L/O1
-S
/
/
k\
,\
'1
%,
10
"'
'
'
'e'.~.;
'
'
''
'
"tb%r,;
. . . . . .
~'olr,;~;
'"'
. . . .
~e~.;.;
9
'
Fig. 1A. Fluorophotometer printout for untreated rats after 5 and 60 minutes of intravenous fluorescein injection before and after induction of diabetes.
study, 13 of 22 eyes developed lens opacities. Again, central haze was the commonest (9/13), followed by peripheral sectorial opacities (2/13). Combined central haze with peripheral sectors and scattered vacuoles were seen in one eye each. The lens opacities were less severe in the treated group than in the untreated rats. Thus, piroxicam had a significant effect (P = 0.006) in minimizing cataract formation in diabetic rats.
Discussion Piroxicam is one of the commonly used antiinflammatory medications. It was found recently [11] to improve ischemic diabetic neuropathy. We evaluated piroxicam in rats with experimentally induced diabetes to determine its possible effect on the blood-retina barrier.
A successful animal model of diabetic retinopathy is difficult to produce. STZ-induced diabetes in rats leads to the development of lens opacities that interfere with fundus examination and VFP [5, 22]. In addition, induction of diabetes is not always successful and some of the rats may become nondiabetic even without medications [10, 13]. The rats are susceptible to dehydration during general anesthesia and are unable to tolerate low temperatures, requiring subcutaneous saline injections and warming during surgery. The rat's eye is small [9]; the lens is large, globular in shape, and occupies about two-thirds of the axial length of the eyeball [1, 8]. The vitreous chamber is crescentshaped, shallow, and approximately 1.4 mm thick [4]. We induced diabetes successfully in 24 out of 26 rats; none of them recovered during the follow-up period. The mortality rate was low (2 of 24) due to
Effect of piroxicam on the blood-retina barrier
339
183 T 5 pre stz
10 2
/ !
"~
/
,^r ' ' ' hr.;,,; . . . . . .
,'.,,
laa,[ \.
f~,
~
I I
\.,
-~ ' ' '
'~b,;~; . . . . . .
/
] !
...../
k.
,v s
......
~.:~ ~,;~,; . . . . .
s163
......
fe
1~~ //
.~; ......
~'~,;~; '
[
T 60 pre stz
/I
T 60 post I t x
] ,)
Ii 01
/
-x*t "-~
printout for piroxicam-treated
,,f"J~
\
t ~p / :tO 1
-5
Fig. lB. F l u o r o p h o t o m e t e r
\\~
~'\.~ "-~...~,
'0In;r,; ' ' '
1 0 r,,rr,
r a t s a f t e r 5 a n d 60 m i n u t e s o f i n t r a v e n o u s
'20In;n; . . . . . .
ff~lri;r,~'~-
fluorescein injection before and
after induction of diabetes.
short duration of anesthesia and type of anesthetic used [5, 6, 13, 18, 22]. Although many sites have been tested for fluorescein injection (tail vein [20], femoral vein [6, 19]), we found the external jugular vein to be the most accessible. Full pupillary dilatation is mandatory for proper VFP records. Previous studies reported the use of different combinations of medications, such as 2.5% phenylephrine and 1% cyclopentolate [5, 16, 18] or 2.5% phenylephrine with 1% tropicamide [19]. We found a combination of 1% cyclopentolate and 1% tropicamide to be most satisfactory for maximum and long-lasting pupillary dilatation. The mechanism by which STZ induced-diabetes causes alterations in the permeability of the bloodretina barrier is not known. Direct toxic effect on the retina is doubtful since STZ acts on the pancreas, as proved by the late alterations in active transport of organic anions [5] and by the normalization of the blood-retina barrier by pancreatic
islet transplant [22]. Osmotic effect due to hyperglycemia is not an acceptable hypothesis since recovery of blood-retina barrier transport occurs without normalization of blood-glucose levels [5]. Thus alternation of this barrier may be due to a combination of factors. The baseline vitreous fluorescein concentrations were not statistically different between the two groups of this work. After induction of diabetes, vitreous fluorescein concentrations became significantly higher in both the treated and untreated groups if compared to the prediabetic measurements. However, the differences were statistically significant after 5 minutes but not after 60 minutes of fluorescein intake. Thus, our findings showed minimal permeability to fluorescein in the normal rat retina and a marked breakdown of the fluorescein blood-retina barrier in rats with STZ induced diabetes. The defect was detectable 15 days following the chemical induction of diabetes.
340
M. Bahgat et al.
Fluoroscein
~-~-I. -~
concentration
(ng/ml)
T-pre-stz T-post-stz U-pre-stz U-post-stz
20
0
i 5
i
time (minutes)
60
70
Fig. 2. Midvitreous fluorescein concentrations versus time for piroxicam treated and untreated groups before and after induction of diabetes. U-Pre-STZ: Untreated group before streptozocin injection; U-Post-STZ: Untreated group after streptozocin injection; T-Pre-STZ: Treated group before streptozocin injection; T-Post-STZ: Treated group after streptozocin injection.
Three-fold to four-fold increase in the concentration of fluorescein in the vitreous was found 8-10 days after induction of diabetes by Waltman et al. [22] and Tso et al. [19]. They reported that the increase in fluorescein concentration occurs sharply and is a constant change. Other workers [10, 13, 14] related that change to many factors other than diabetes. Smith et al. [18] and Vine et al. [20] showed that mid-vitreous fluorescein concentrations in rats with STZ induced diabetes were significantly lower than in normal rats. However, vitreous fluorescein concentrations, when compared with total plasma fluorescein concentration, was significantly increased for after two weeks of diabetes. They explained this finding by increased renal fiuorescein excretion as a result of diabetic nephropathy. The reason for their opposing results in vitreous fluorescein concentrations is not readily apparent. It may reflect differences in the techniques of these studies. In our study, vitreous fluorescein concentrations showed lower levels of fluorescein in the piroxicam treated group than the untreated group. This difference was noticed after 5 and 60 minutes. However, the change in vitreous fluorescein concentration between the two groups was not statistically significant. This finding indicates piroxicam does not have a significant effect on the permeability of the blood-retina barrier. However, long-term
treatment with piroxicam may be more effective and should be investigated. Follow-up examination for lens opacities agree with other records [12] that cataracts did not always develop at equal rates in left and right eyes of diabetic rats. Our findings differ in the early onset of appearance of multiple scattered vacuoles, which were seen as early as the first week after induction of diabetes. The incidence of lens opacities was much higher in the untreated group (81.8%); all rats were bilaterally affected. For the treated group, lens opacities developed in 59.1%. Both eyes were affected in four rats only. Our findings indicate piroxicam therapy may have an effect on lens opacity formation in diabetic rats, probably due to aldose reductase inhibition [17].
Acknowledgements The authors would like to thank Coherent for generously allowing us to use the Fluorotron. We would also like to acknowledge the help of Mark Hurry, Department of Neurology, in performing the experiments. This work was supported in part by U.S. Public Health Service Grants EY02377, EY07541 and EY08137 from the National Eye Institute, National Institutes of health, Bethesda, MD and by grants from the Juvenile Diabetes Foundation International and Pfizer Inc.
References 1. Block MT. A note on the refraction and image formation of the rat's eye. Vision Res 1969; 9: 705-11. 2. Cunha-Vaz JG. Pathophysiology of diabetic retinopathy. Br J Ophthalmol 1978; 62: 351-5. 3. Cunha-Vaz J, Abreu J R F C~mpos AJ, Figo GM. Early breakdown of the blood-retinal barrier in diabetes. Br J Ophthalmol 1975; 59: 649-56. 4. Hughes A. A schematic eye for the rat. Vision Res 1979; 19: 569-88. 5. Jones CW, Cunha-Vaz J, Zweig KO, Stein M. Kinetic vitreous fluorophotometry in experimental diabetes. Arch Ophthalmol 1979; 97: 1941-3. 6. Kirber WM, Nichols CW, Grimes PA, Winegrad AI, Laties AM. A permeability defect of the retinal pigment epi-
Effect of piroxicam on the blood-retina barrier
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
thelium. Occurrence in early streptozocin diabetes. Arch Ophthalmol 1980; 98: 725-8. Krupin T, Waltman S, Oestrich C, Santiago J, Ratzan S, Kilo C, Becket B. Vitreous fluorophotometry in juvenileonset diabetes mellitus. Arch Ophthalmol 1978; 96: 812-4. Lashley KS. The mechanism of vision. The structure and image-forming power of the rat eye. J Comp Psychol 1932; 13: 173-200. Massof RW, Chang FW. A revision of the rat schematic eye. Vision Res 1972; 12: 793-6. Palestine AG, Brubaker RF. Plasma binding of fluorescein in normal subjects and in diabetic patients. Arch Ophthalmol 1982; 100: 1160-1. Parry GJ, Kozu H. Piroxicam may reduce the rate of progression of experimental diabetic neuropathy. Neurology (in press). Poulsom R, Boot*Handford RP, Heath H. Some effects of aldose reductase inhibition upon the eyes of long-term streptozotocin-diabetic rats. Curr Eye Res 1982/1983: 2: 351-5. Prager TC, Chu HH, Garcia CA, Anderson RE, et al. The influence of vitreous change on vitreous fluorophotometry. Arch Ophthalmol 1982; 100: 594-6. Prager TC, Wilson DJ, Avery GD, Merritt JH, Garcia CA, Hopen G, Anderson RE. Vitreous fluorophotometry: Identification of sources of variability. Invest Ophthalmol Vis Sci 1981; 21: 854-64. Rusin MM, Blair NP, Shakin E. The effect of pH on the transport of fluorescein across the blood-retinal barrier. Invest Ophthalmol Vis Sci 1983; 24 (suppl): 249. Shinowara NL, Grimes PA, Rapoport SI, Laties AM, et al. Acidosis alters fluorescein permeability across the pigment
17.
18.
19.
20.
21.
22.
341
epithelium. Invest Ophthalmol Vis Sci 1983; 24 (suppl): 250. Sharma YR, Cotlier E. Inhibition of lens and cataract aldose reductase by protein-bound anti-rheumatic drugs: Salicylate, indomethacin, oxyphenbutazone, sulindac. Exp Eye Res 1982; 35: 21-7. Smith SS, Ashburn FS, Pilkerton AR, Recant L. Vitreous fluorophotometry in three models of experimental diabetes mellitus. Retina 1982; 2: 121-5. Tso MOM, Cunha-Vaz JG, Shih CY, Jones CW, et al. Clinicopathologic study of blood-retinal barrier in experimental diabetes mellitus. Arch Ophthalmo11980; 98: 203240. Vine AK, Kisly AM, Betz AL, Howatt WF. Vitreous fluorophotometry in rats with streptozocin-induced diabetes. Arch Ophthalmol 1984; 102: 1083-5. Waltman SR, Krupin T, Kilo C, Becker B. Vitreous fluorophotometry in adult-onset diabetes mellitus. Am J Ophthalmol 1979; 88: 342-5. Waltman SR, Oestrich C, Krupin T, Hanish S, Ratzan S, Santiago J, Kilo C. Quantitative vitreous fluorophotometry: A sensitive technique for measuring early breakdown of the blood-retina barrier in young diabetic patients. Diabetes 1978; 27: 85-7.
Address for offprints: G.A. Peyman, LSU Eye Center, 2020 Gravier Street, Suite B, New Orleans, LA 70112, USA
