Octreotide, a Somatostatin Analogue, Fails to Inhibit Hypoxia-induced Retinal Neovascularization in the Neonatal Rat

Objective: Octreotide, a somatostatin analogue, has been shown to prevent angiogenesis in diverse in vitro models. We evaluated its effect on retinal neovascularization in vivo, using a neonatal rat retinopathy model. Methods: We used, on alternating days, hypoxia (10% O2) and hyperoxia (50% O2) during the first 14 days of neonatal rats, to induce retinal neovascularization. Half of the rats were injected subcutaneously with octreotide 0.7 μg/g BW twice daily. At day 18 the eyes were evaluated for the presence of epiretinal and vitreal hemorrhage, neovascularization and epiretinal proliferation. Octreotide pharmacokinetics and its effect on serum growth hormone (GH) and insulin-like growth factor I (IGF-I) were examined in 28 rats. Results: Serum octreotide levels were 667 μg/1 two hours after injection, 26.4 μg/1 after nine hours and 3.2 μg/1 after 14 hours. GH levels were decreased by 40% (p = 0.002) two hours after injection but thereafter returned to baseline. IGF-I levels were unchanged two hours after injection and were elevated by 26% 14 hours after injection (p = 0.02). Epiretinal membranes were highly associated with epiretinal hemorrhages (p < 0.001), while retinal neovascularization was notably associated with vitreal hemorrhages (p < 0.001). Conclusions: Twice-daily injections of octreotide failed to produce sustained decrease in serum GH, but produced rebound elevation of serum IGF-I. Accordingly, no statistically significant effect of injections on retinal pathology was noted. This finding, however, does not contradict our assumption that GH suppression may decrease the severity of retinopathy.


in treatment
f clinical PDR (McCombe et al., 1991)  and experimental ROP (Smith et al., 1997).

Octreotide, a long-acting somatostatin analo- gue, is a potent inhibitor of GH secretion with a plasma half-life of 90 minutes (Bauer et al.,   1982).Octreotide can be used to lower GH and IGF-I levels (Plewe et al., 1987).In addition, octreotide has been shown to have direct anti- angiogenic properties in human retinal endo- thelial cells culture (Grant, Caballero and   Millard, 1993) and in the vascular growth mod- el of the chick-embryo chorioallantoic mem- brane (Woltering et al., 1991).Octreotide was tested in human clinical trials for treatment of both early (Kierkegaard et al., 1990) and severe PDR (Mallet et al., 1992;Lee et al., 1988).Though several positive trends were shown in these studies, the small size of the study groups and lack of adequate controls do not allow us to draw results.We studied the impact of systemic octreotide treatment on retinal neo- vascularization in a neonatal rat ROP model in which we demonstrated a possible role of the retinal IGF system in ischem a-induced neovas- cularization (Averbukh et al., 1998).


METHODS


Animals

Fifty-six neonatal Sabra rats (provided by the Harlan Laboratories, Jerusalem, Israel) were divided into experimental groups of eight to ten animals from mixed litters, not divided ac- cording to gender.At the age of one day the group with the lactating mother were placed in an incubator with a controlled oxygen con- centration.The lactating rats had free access to standard rat chow and tap water.All the animals wer maintained and treated in ac- cordance with the Association for Research in Vision and Ophthalmology statement on the use of animals in ophthalmic and vision re- search.Retinal neovascularization was induced by alternating oxygen concentrations as pre- viously described (Reynaud and Dorey, 1994;  Penn, Tolman and Lowery, 1993).Briefly, the animals were exposed on alternating days to hypoxia (10% O2-24h) and hyperoxia (50%  O2-24h) during their first 14 days, followed by four days in room air.Hypoxia was produced by mixing air with nitrogen, and hyperoxia by mixing air with oxygen.A minimal flow of two liters per minute was maintained.Half of the rats were injected subcutaneously with octreo- tide 0.7 tg/g BW twice daily, while the other half were injected with saline during the whole period.At day 18 the rats were anesthetized with ether and sacrificed.The eyes were enu- cleated and placed in formaldehyde.To evalu- ate the effect of octreotide injection on GH and IGF-I levels, we checked the serum levels in another 28 animals o

animals
and Evaluation of Neovascularization-associated


Changes

In the neonatal ROP model, high oxygen lev- els impede the normal vascularization process (Chang-Ling et al., 1992) and induce vaso-ob- literation (Pierce, Foley and Smith, 1996), thus creating avascular areas in the periphery of the retina.These areas are prone to ischemia when the oxygen level is lowered.Severe hypoxia is also associated with pre-retinal and vitreal .hem-orrhages in neonatal rats.We observed sev- ere intraocular hemorrhages in neonatal rats after 24 hours' exposure to hypoxia, suggesting that these hemorrhages are primary phenom- ena and not secondary to neovascularization (un- published data).Blood on the retina, as well as ischemia-associated factors, may cause epiretinal proliferation and neovascularization which are typical for many retinopathies, such as ROP and diabetic retinopathy.These morphologic changes were the subject of our study when comparing the treated animals with the controls.

Each formalin-fixed, paraffin-embedded eye- ball was examined for existence of gross pathology.Then it was sliced using coronal cuts starting from the optic nerve and ending close to the lim- bus.If intraocular hemorrhages were noted, cuts were made through them.Otherwise, ten repre- sentative histological slides were prepared skip- ping approximately equal numbers of cuts in between.The slides were stained with hematoxylin-eosin and examined and scored in a masked fashion by one examiner.Each eye was ass

ase.We l
oked for epiretinal hemorrhages (RH) (defined as blood trapped between the retina and the posterior hyaloid face), vitreal hemorrhages (VH) (blood cells dispersed in the vitreous), epireti- nal proliferation (proliferative vitreoretinopathy, PVR) and neovascularization (NV).We defined our findings as NV only when we saw discrete blood vessels on the inner side of the internal limiting membrane (Fig. 1A), while the finding of fibrovascular proliferation and membrane for- mation inside the vitreous cavity (Fig. 1B) was defined by us as PVR.This separation of the proliferative changes into NV and PVR was dic- tated by our suspicion that we are seeing two separate phenomena with different pathogenesis.

Each manifestation was given a score between 0 (non-existent) and 3 (severe) according to the number of slides containing the specific pathology and the subjective perception of the severity of pathology.For examp e, if neovas- cularization was seen only on one slide and was of limited extent it was assigned a score of one.The statistical significance of the difference between the scores was evaluated using the Mann-Whitney Rank Sum Test.Correlation between RH and VH on the one hand and PVR and NV on the other hand was evaluated using the Spearman Rank Order Correlation test.All statistical analysis was performed using Sigma- Stat for Windows software (Jandel Corporation, San Rafael, CA).

Serum Octreotide, GH and IGF-I Levels Serum octreotide, GH and IGF-I blood levels were evaluated 2 hours after subcutaneous in- jection of 10 tg octreotide in five 7-day-old ani- mals with an average weight of 15g (0.7 tg/g BW).Five control animals of the same age and weight were injected with saline.Similarly, the levels were measured in seventeen 14-day-old animals with an average weight of 30 g after in- jection of 20 tg of octreotide.Five animals were sacrificed 9 hours after injection, and four ani- mals 14 hours after injection.Eight controls of similar weight and age were injected with saline.

Average levels of octreotide, GH and IGF-I were calculated for each group.We used a one-tail t-test for statistical analysis.

Serum rat growth hormone (rGH) was meas- ured by r dioimmunoassay (RIA) using a speci- fic polyclonal rabbit rGH antibody and rGH as standard.The ingredients for the RIA were obtained from Amersham (Amersham International, Amersham, Bucks, UK).Intraand
(A) (B) FIGURE
Typical findings in the affected eyes.Slide A illustrates big epiretinal hemorrhage (arrows) with epiretinal membrane formation around it.Some hemosiderin spots (arrow heads) are seen inside the hemorrhage, indicating the longstanding nature of the bleeding.Slide B shows vitreous hemorrhage (arrows) with retinal neovascularization (arrow heads).Hematoxylin-eosin stain, magnification x 10. (See Color Plate I).interassay coefficients of variations were less than 5% and 10% respec

vely.Serum IGF-I was measured after ex
raction with acid-metha- nol (30tl serum and 7501 acid methanol) (Furlanetto and Marino, 1987).The mixture was incubated for two hours at room temperature, centrifuged, and 25 gl of the supernatant was diluted 1:200 before analysis.Serum IGF-I was measured by RIA using a polyclonal rabbit anti- body (Nichols Institute Diagnostics, San Juan Capistrano, Calif., USA) and recombinant hu- man IGF-I as standard (Amersham International, Amersham, Bucks, UK).Mono-idionated IGF-I 125I-(31Tyr)-IGF-I was obtained from Novo- Nordisk A/S Bagsvaerd, Denmark.When expos- ing he serum extract to western ligand blotting (WLB), no IGFBPs could be identified and furthermore, semilog linearity of biosynth tic IGF-I and serum extracts was seen, indicating antigen similarity and the fact that no IGFBPs interfered in the RIA.Intraand interassay co- efficients of variations were below 5 and 10% respectively.

Serum octreotide was measured by RIA as previou frskov, Thomsen and  Yde, 1968).Intraand interassay coefficients of variations were less than 5% and 10% for all assays.

calculate the biological half-life time of octreo- tide in our experiment we compared the octreo- tide level two hours after injection with that nine hours after injection.The serum level was reduced from 667.0 gg/1 to 26.4 gg/1 in these sev- en hours (420 minutes), which is less than five [log (667.0/26.4)=4.66] half-lives, thus, a half- life time of approximately 90 minutes (420min/ 4.66=90.1min.).Similarly, basing the calcula- tion of octreotide half-life time on the difference between the octreotide level at 9 hours and its level at 14 hours results in a half-life time of 98 minutes.

Serum GH levels were decreased by 40% (p 0.002) two hours after injection of octreotide as compared with saline injected animals (Tab.I).However, 9 and 14 hours after injection, GH level was comparable in the octreotide and the control groups.Serum IGF-I level was similar in the treated and the control groups two hours after injection, but was elevated in the treated group 9 and 14 hours after injection (by 19% and 26%, p 0.06 and p 0.02 respectively).


RESULTS

Fifty-two eyes of 26 animals were examined in the treatment group and 59 eyes of 30 animals in the control group (one eye was not suitable for histologic preparation).

Serum Octreotide, GH and IGF-I Levels Octreotide levels were high two hours after injection, but much lower later (Tab.I).To


Histologic Findings

The scores of histologic findings in the eyes are summarized in Table II.Epiretinal and vi- treal hemorrhages were the most frequently ob- served findings in both the octreotide-treated and the control groups with approximately 45% of the eyes affected by each one of them and 70% affected by any of them.PVR was less com- mon, affecting less than 40% of the eyes in each group.NV was found only in 35% of the octreotide-treated and in 27% of the control   5) 15% ( 9) 5% (3) 3 10% ( 6) 12% ( 7) 2% (1) 8% ( 5)

Epiretinal hemorrhage (RH), vitreal hemorrhages (VH), proliferative vitreo-retinopathy (PVR) and neovascular budding (NV).Numbers are presented as percent of total.Absolute numbers are given in parentheses.None of the differences reached a statistical significance.

eyes.The intraobserver error for scoring of dif- ferent findings was 5

0%.

Usi
g the Spearman Rank Order Correlation test we found a strong association between the presence of PVR and epiretinal hemorrhages p < 0.001) but not vitreal hemorrhages p

27).

The score of n
ovascularization-associated changes (PVR and NV) was higher in the octreo- tide group than in the control group (Tab.II); this difference, however, was not statistically sig- nificant.Severe (grade 3) changes were slight- ly more common in the octreotide-treated group (Tab.II).


DISCUSSION

We demonstrated that twice-daily injections of octreotide failed to induce sustained inhibition of GH secretion and a reduction in serum IGF-I in neonatal rats, while pharmacological serum levels of octreotide were achieved.Furthermore, retinal neovascularization was not inhibited.If anything, severe (grade 3) manifestations of re- tinal neovascularization were more common in the octreotide-treated group than in the con- trol group.These findings are in contrast with our hypothesis, based on in vitro studies, that octreotide prevents hypoxia-induced neovascu- larization either directly or by lowering GH and IGF-I levels.

The strong correlation between RH and PVR on the one hand, and between VH and NV on the other hand, is of interest nd may point to some underlying pathophysiology in our mod- el.While NV was observed mostly in the peri- phery of the retina (Fig. 1A), on the border of the area which is typically affected by hypoxia in the ROP model, the PVR was found mainly in the posterior pole, associated with epiretinal hemorrhages.Th s finding supports our theory that the proliferative vitreoretinopathy (PVR) is induced by epiretinal hemorrhage (RH), thus should not necessarily originate from the peri- phery of the retina, while "pure" neovasculari- zation (NV) seen on Figure 1B is probably induced by hypoxia-related gr

th factors
roduced by adjacent avascular retina, typically in the periphery.Bleeding may either cause epiretinal proliferation directly or may be a mar- ker of severe retinal ischemia that is triggering epiretinal proliferation through a variety of growth factors.In any case, the sub-group of eyes with intraocular hemorrhages is specifical- ly prone to epiretinal proliferation.We should stress that our definition of neovasculariza- tion was different from that used by some oth- er groups.We defined neovascularization as a gross histological finding of blood vessels on the inner side of internal limiting membrane, while Smith et al (1997) considered every nucleus of vascular endothelial cell at the same location as evidence of neovascularization.Our way of evaluation is less sensitive, but has a greater clinical implication.

The achieved octreotide half-life of 90 minutes fits well with the previously reported half-life time of the drug (Bauer et al., 1982).We evaluated the effect of octreotide in 7-and 14-day animals, since it is at this age that the avascular retina is reacting to ischemia by up-regulation of IGF- I receptor and IGF binding proteins 2 and 3 (Averbukh et al., 1998).It was impossible to take blood samples from the actual experimental ani- mals without severely affecting them by stress and anemia.On the other hand, taking blood samples at the time of sacrifice (18 days) was probably irrelevant since the course of the neo- vascularization process is determined much earlier.

The primary effect of octreotide on neovascu- larization is mediated, at least to some extent, by GH secretion inhibition.GH levels were sig- nificantly reduced two hours after injection, but returned to normal nine hours after injection.

This very brief suppressive effect on GH lev- els is in accordance with previously published data (Wurzburger et al., 1992).However, even two hours after injection, the achieved high level of octreotide failed to reduce serum IGF-I levels.It is of interest that treatment with octreotide at similar dose prevented kidney hypertrophy in diabetic rats (Flyvbjerg et al., 1989) while failing to change significantly serum levels of G or IGF-I.Smith et al., showed recently that lower- ing of GH by another somatostatin analogue is effective in preventing hypoxia-induced neovas- cularization in the ROP model in mice.However, in spite of the demonstrated inhibition of GH secretion, at higher levels of somatostatin ana- logue there was a paradoxical relative increase in serum IGF-I levels that stayed unexplained (Smith et al., 1997).

Octreotide was shown to have a direct inhi- biting effect on IGF-I production in addition to its effect through the inhibition of GH secretion (Heyer et al., 1989).It was previously reported that in cultured GH stimulated hepatocytes, IGF-I mRNA levels were significantl reduced in the presence of low octreotide concentrations (0.3 and 3gg/1) while high concentrations of octreotide (30 and 300gg/1) had no effect on IGF-I mRNA abundance (Serri et al., 1992).In the present study, octreotide was injected at 12- hour intervals; therefore the lowest serum level of octreotide, just before the next injection, was approximately 6.6 gg/1 (extrapolated) which may be still too high for a direct suppressive effect on IGF-I production but too low for suppression of GH secretion.Moreover, withdrawal of octreo- tide, like somatostatin withdrawal, may generate a pulsatile GH release (Cella et al., 1996) which may explain the elevation of IGF-I levels at 9 and 14 hours after injection.Fluctuations in IGF-I levels, secondary to rebound elevation of GH after a brief suppression, may have a deleterious effect on retinal vasculature, promoting rather than suppressing neovascularization.

Finally, octreotide was found to have direct antiangiogenic properties in human retinal en- dothelial cell (HREC) culture (Grant et al., 1993b)   and in a vascular growth model of the chick- embryo chorioallantoic membrane (Woltering  et al., 1991).In cell culture of HREC stimulat- ed by IGF-I and bFGF, [3H]thymidine incorpo- ration was inhibited by octreotide (Grant et al.,   1993b).The dose-effect curve showed that an octreotide level of 1.6gg/1 caused 15% inhibi- tion, 16 gg/1 caused 35% inhibition and 164 gg/1 caused 60% inhibition.In our experiment, octreo- tide serum level was between a few gg/1 and a few hundreds gg/1 during the whole 12-hour interval between injections.Inhibition of angiogenesis, however, was not achieved.

In conclusion, twice-daily injections of octreo- tide did not prevent hypoxia-related neovascu- larization in our model.It is possible, however, that the therapeutic range of octreotide is of cri- tical importance, and that the great variability of octreotide levels achieved in our study due to twice-daily injections is not suitable for a sustained effect.Furthermore, fluctuations in se- rum octreotide level may have caused the ob- served increase in neovascularization-related changes in the treated group.Continuous infu- sion of octreotide and other somatostatin analogues with a longer alf-life that better suppress GH secretion may still be effective in preventing neovascularization.

TABLE Octreotide ,
Octreotide
GH and IGF-I levels after injection of octreotide
TimeGHIGF-IafterAgeOctreotideGH% ofIGF-I% ofinjectionn(days)ktg/1gg/1controlgg/controlControl6702 hours57667.0 + 42.5*Control81409 hours51426.4 4-5.114 hours4143.2 4-0.7

TABLE II
II
Scoring of findings in the eyes of the octreotide-treated and the control groups
Grade ofseverityRH*VH*PVR*NV*OctreotideNone56% (29)54% (28)61% (32)65% (34)13% (7)19% (10)12% (6)10% (5)212% (6)6% (3)19% (10)12% (6)319% (10)21% (11)8% (4)13% (7)ControlNone54% (32)58% (34)64% (38)73% (43)126% (15)22% (13)19% (11)14% (8)210% (6)8% (
*Values are given as mean 4-SEM; p 0.002; *p 0.02.
AcknowledgmentsSupported by a grant from the Pauline and Joseph Fried Family, and by a grant from the Chief Scientist, Ministry of Health, Israel.
Gene expression of IGF-I, its receptor and binding proteins in retina under hypoxic conditions. E References Averbukh, O Weiss, M Halpert, R Yanko, R Moshe, I Nephesh, A Flyvbjerg, L Yanko, I Raz, Metabolism. 471998

SMS 201-995: a very potent and selective octapeptide analogue of somatostatin with prolonged action. W Bauer, U Briner, W Doepfner, R Haller, R Huguenin, P Marbach, T J Petcher, J Pless, Life Sci. 311982

Pituitary ablation in the treatment of diabetic retinopathy. R F Bradley, S B Rees, S B Fager, Med.